CARBOHYDRATE NANOPARTICLES: A NOVEL DRUG DELIVERY PLATFORM 
FOR THE SYSTEMIC ROUTE 
 
 
Except where reference is made to the work of others, the work described in this 
dissertation is my own and was completed under the supervision of the 
advisory committee.  This dissertation does not include 
proprietary or classified information.   
 
 
 
 
________________________________________ 
Arindam Basu Sarkar 
 
 
________________________                ______________________ 
William R. Ravis                                                          Gregory M. Kochak, Chair 
Professor                                                                       Associate Professor 
Pharmacal Sciences                                                      Pharmacal Sciences 
 
 
 
 
________________________                ______________________ 
Daniel L. Parsons                                                         John Eley 
Professor                                                                       Assistant Professor 
Pharmacal Sciences                                                      Pharmacal Sciences 
 
 
 
 
________________________________                     _____________________________ 
Anwar Ahmed                                                              Stephen McFarland 
Associate Professor                                                      Dean 
Aerospace engineering                                                 Graduate School 
 
 
CARBOHYDRATE NANOPARTICLES: A NOVEL DRUG DELIVERY PLATFORM 
FOR THE SYSTEMIC ROUTE 
 
 
 
Arindam Basu Sarkar 
 
 
 
 
 
A Dissertation 
Submitted to 
the Graduate Faculty of  
Auburn University 
in Partial Fulfillment of the 
Requirements for the  
Degree of 
Doctor of Philosophy 
 
 
Auburn, Alabama 
December 15, 2006 
 
 
  
iii 
CARBOHYDRATE NANOPARTICLES: A NOVEL DRUG DELIVERY PLATFORM 
FOR THE SYSTEMIC ROUTE 
 
 
 
Arindam Basu Sarkar 
 
Permission is granted to Auburn University to make copies of this dissertation at its 
discretion, upon request of individuals or institutions and at their expense.  
The author reserves all publication rights. 
 
 
 
                                                                                 _____________________________ 
                                                                                    Signature of Author 
 
                                                                                 _____________________________ 
                                                      Date of Graduation 
 
 
 
 
 
 
 
 
 
 
 
 
  
iv 
VITA 
Arindam Basu Sarkar was born on the 5th of April, 1968 in Calcutta, West Bengal, 
India to Mr. Suhaskanti Basu Sarkar and Mrs. Mandira Basu Sarkar.  After completing 
his school education from Jadavpur High School, Calcutta in 1986, he commenced 
university education in Department of Pharmacy, Faculty of Engineering, Jadavpur 
University.  He graduated with Bachelor of Pharmacy in 1991 and started working in 
pharmaceutical industry.    Following a two year stint (September 1995 to October 1997) 
in Himalayan Pharmacy Institute as a lecturer, he commenced his Masters in 
Pharmaceutical Engineering in Jadavpur University.  Upon obtaining his Masters in 
1999, he started pursuing graduate education in Pharmaceutics at Harrison School of 
Pharmacy, Auburn University, Alabama.  During his studentship in Auburn University, 
he served as Graduate Teaching Assistant as well as Graduate Research Assistant in the 
pharmacy school.  In May, 2005 he received ?Don Tillery Scholarship? from Harrison 
school of Pharmacy.  In April, 2006 he received ?Outstanding Doctoral Student? award 
from the Graduate School, Auburn University for the academic year 2005-2006.  He is 
also recipient of ?Postgraduate Best Paper Award? in the basic sciences section of 
American Pharmaceutical Association Annual Meeting and Exposition, 2003 held at New 
Orleans, Luisiana. 
 
 
 
  
v 
DISSERTATION ABSTRACT 
 
 
CARBOHYDRATE NANOPARTICLES: A NOVEL DRUG DELIVERY PLATFORM  
 
FOR THE SYSTEMIC ROUTE  
 
Arindam Basu Sarkar 
 
Doctor of Philosophy, December 15, 2006 
(M. Pharm, Jadavpur University, 1999) 
(B. Pharm, Jadavpur University, 1999) 
 
196 Typed Pages 
 
Directed by Gregory M. Kochak 
 
 
With the phenomenal developments in the fields of biotechnology and molecular 
biology, the focus in the medical fraternity is shifting towards cellular level intervention.  
The use of polypeptides, proteins and possibly genes through targeted delivery platforms 
is viable today.  Available particulate delivery platforms are inadequate targeting 
platforms because of the poor physical and chemical properties of the polymers used.  In 
course of current research a novel nanoparticulate delivery platform has been developed 
with hydrophilic carbohydrate polymers, using a water-in-oil emulsion and interfacial 
crosslinking. Synthesis of nanoparticles has been characterized and key parameters 
affecting the reaction identified.  The reaction is dependent on the pH, species of the 
polymer, the polymer load of the dispersed aqueous phase and species of the crosslinker. 
  
 
  
vi 
Smaller starch polymers were found to be most suitable for this purpose because their 
relatively higher aqueous solubility and superior rheological properties.   Trimesoyl 
chloride was found to be the best crosslinker because of better reactivity and the higher 
hydrophilicity imparted to the surface.  Sorbitan mono palmitate (Span 40?) was found to 
be the best surfactant.  The nanoparticles have been chemically characterized by 
determining the degree of substitution (DS).  Analytical methodology was developed for 
quantifying the DS.  The reaction occurs in a wide pH range of 3-11 and was base 
catalyzed at higher and acid catalyzed at lower pHs.  The nanoparticles prepared using 
maltodextrin from amylopectin and trimesoyl chloride had hydrophilic surfaces. The 
particles suspended in aqueous media as monodispersed mononucleated particles.   
The crosslinked membrane of the nanoparticles was largely inadequate in sustaining 
the release of small molecules like tartrazine.   For moderate sized proteins like lysozyme 
though, the membrane was adequate in sustaining the release for weeks.  Diffusion based 
modeling of release data showed two independent modes of release, one fast phase which 
is complete within 3 hr and one slow phase lasting for weeks.  Estimated permeability 
coefficients show that there is a negative correlation between the degree of substitution 
and the permeability coefficient of the slower mode of release.  Further work is required 
to establish the behavior of these nanoparticles in vivo.  Overall, the developed platform 
appears suitable for macromolecular delivery in the systemic circulation.  Future studies 
will investigate potential cellular level intervention. 
 
 
 
 
 
  
vii 
 
 
 
 
 
ACKNOWLEDGEMENTS 
 
I would like to respectfully remember the selfless sacrifices of my parents Mr. 
Suhaskanti Basu Sarkar and Mrs. Mandira Basu Sarkar in bringing me up under trying 
conditions and their contributions in setting the priorities right at an early stage of my 
life.  I am thankful to my wife Mrs. Jayita Basu Sarkar for her wholehearted physical and 
emotional support and understanding.   I am truly indebted to my brother, my sister-in-
law and my nice for their support and encouragement and for making life worth living. 
I would like to acknowledge the invaluable contributions of my graduate advisor Dr. 
Gregory M. Kochak, who has been a fountainhead of ideas and inspiration and a teacher 
in the truest sense of the word.  I extend my sincere gratitude to all the faculty members 
of Harrison School of Pharmacy, specially the members of my Graduate Advisory 
Committee for their help and cooperation.  I am thankful to Dr. Randal Clark, who has 
been a wonderful teacher for me.   
 
 
 
 
  
viii 
 
 
 
 
 
Style manual or Journal used: Carbohydrate Polymers in which the third chapter has been                         
published [59, (2005), 305-312] 
 
Computer software used:        MS Word 2003, Excel 2003 & ChemDraw Standard 10 
 
 
 
 
 
 
 
 
 
 
  
ix
 
 
 
 
 
TABLE OF CONTENTS 
 
 
I.  INTRODUCTION                                                                                     1 
 
II. REVIEV OF LITERATURE                                                                      9 
III. HPLC ANALYSIS OF ALIPHATIC AND AROMATIC                                     75 
     DICARBOXILIC ACID CROLSS-LINKERSHYDROLYZED  
     FROM CARBOHYDRATE POLYESTERS FOR ESTIMATION  
     OF MOLAR DEGREE OF SUBSTITUTION                                
IV. SYNTHESIS OF CARBOHYDRATE NANOPARTICLES                               98 
     WITH HYDROPHILIC CORE 
V. INCORPORATION INTO AND RELEASE OF TARTRAZINE                        135 
     AND YSOZYME FROM CARBOHYDRATE NANOPARTICLES 
VI. SUMMERY AND CONCLUSION                                                                      174 
 
 
  
x
 
LIST OF TABLES 
CHAPTER 3. 
Table 1.    Calibration curve statistics (crosslinker assay)                                 95 
Table 2.    Accuracy and precision (crosslinker assay)           96 
Table 3.    Assay results (crosslinker assay)                                                       97 
CHAPTER 4. 
Table 1.   Degree of substitution comparison in terephthalic acid                     126 
Table 2.   Comparative reactivity of different crosslinkers                                126 
CHAPTER 5. 
Table 1.    Effect of polymer species on the encapsulation efficiency               161 
                   of tartrazine 
Table 2.    Accuracy and precision of lysozyme assay (release studies)            161 
Table 3.    Accuracy and precision of lysozyme assay (encapsulation               161 
                 efficiency studies) 
Table 4.    Effect of pH and solvent on the partition behavior                            162 
                  of lysozyme 
Table 5.    Effect of carbohydrate content on the partition                                 162 
                  behavior of lysozyme 
Table 6.    Encapsulation efficiency and degree of substitution                          162 
                  In representative lysozyme loaded batches 
 
  
xi
 
 
 
LIST OF FIGURES 
 
CHAPTER 3. 
Figure 1.  Phenacyl ester formation kinetics of polycarboxylic acids          92 
Figure 2.  A representative chromatogram showing the separated               93 
                 peaks of  derivatized polycarboxylic acids 
Figure 3.  A chromatogram of the authentic sample of crosslinker              94 
                 derived from hydrolysis of nanoparticles 
      Figure 4.  Chromatogram of different C6 dicarboxylic acids showing         94 
                       utility of the assay method 
  Figure 5.  Optical microphotograph of a representative blank                     95 
                        Nanoparticles 
  CHAPTER 4. 
  Figure 1.  Particle size distribution of two representative nanoparticle       127 
                     Batches 
  Figure 2(a-c).  Optical microphotographs of nanoparticles made               128 
                         from different polymers and crosslinkers 
      Figure 2(d-f).  Optical microphotographs of nanoparticles made               129 
                         from different polymers and crosslinkers 
 
  
xii
  Figure 3 (a & b)   Transmission electron microscopy of a                          130 
                          representative nanoparticle batch 
  Figure 4.   Effect of pH on the degree of substitution                                  131 
  of terephthaloyl chloride and diethylmalonoyl dichloride 
  with maltodextrin 
  Figure 5.  Effect of pH on the degree of substitution                                 131 
  of terephthaloyl chloride with dihydroxypropyl amylose  
  Figure 6.  Effect of pH on the degree of substitution                                 132 
  of trimesoyl chloride with maltodextrin  
  Figure 7.  Effect of pH on the degree of substitution                                 132 
  of diethyl malonyl dichloride with dihydroxypropyl 
   amylose 
  Figure 8.   Effect of buffer strength on the degree of crosslinking               133 
  Figure 9.   Effect of polymer concentration on the degree of                       133 
  substitution 
  Figure 10.  Effect of crosslinker concentration on degree of                        134 
  Substitution 
  Figure 11.  Effect of reaction time on the degree of substitution                  134 
  CHAPTER 5. 
  Figure 1.   Release profile of tartrazine from nanoparticles made                163 
  using terephthaloyl chloride 
   
 
  
xiii 
Figure 2.   Release profile of tartrazine from nanoparticles made                164 
  using trimesoyl chloride 
  Figure 3.   Comparative release profiles of tartrazine from                          165 
  nanoparticles  made using terephthaloyl chloride  
   and trimesoyl chloride 
  Figure 4.   Relationship between degree of substitution and       166 
  concentration of trimesic acid for two lysozyme loadings 
  Figure 5.  Comparative release profile of lysozyme from                             167 
          nanoparticles   made with different crosslinker  
  concentration 
  Figure 6.   Relationship between degree of substitution and extent              168 
  of lysozyme release 
  Figure 7.  Comparative release profile of lysozyme from different sized      169 
  particles 
  Figure 8.  Comparative release profile of lysozyme from nanoparticles        170 
  with different lysozyme loading 
  Figure 9.  Fitting of lysozyme release data with homogeneous                      171 
  Diffusion model 
  Figure 10. Fitting of lysozyme release data with heterogeneous                     171 
  diffusion model (I) 
  Figure 11.  Fitting of lysozyme release data with heterogeneous                     172 
  diffusion model (II) 
 
  
xiv
   Figure 12.  Fitting of lysozyme release data with heterogeneous                     172 
  diffusion model (III) 
  Figure 13.  Fitting of lysozyme release data with heterogeneous                     173 
  diffusion model (IV) 
   Figure 14 (a, b).  Mechanistic contribution of two different modes                 174 
  of lysozyme release (short time and long time view) 
  Figure 15.  Relationship between degree of substitution and                           175 
  Permeability coefficient of the slower mode of release (P2) 
   
 
 
 1 
I.  INTRODUCTION 
 
 
The last two decades have seen a phenomenal growth in the fields of molecular 
biology and bio-technology.  Major advancement, especially in the areas of recombinant 
DNA technology and hybridoma technology has resulted in a deluge of macromolecules 
showing therapeutic potentials which are gradually being introduced to clinical studies.  
Unlike in the past, when polypeptides and proteins from human and animal sources had 
to be carefully isolated and purified at a laboratory scale, it is now possible to produce 
commercial quantities of polypeptides and proteins like interleukins, immunoglobulins, 
growth hormone and insulin, to name just a few, in a high degree of purity, using 
transcriptional and translational methods in bacteria.  Additionally, developments in 
functional genomics and proteomics are likely to expand the therapeutic repertoire even 
further, once the gene-protein and protein-protein relationships are firmly established for 
5,000-10,000 drug targets expected from the human genome map (Drews, 2000).  
Already, by 2002, 130 biotechnology derived protein drugs and vaccines were approved 
by FDA in the US (Biotechnology Industry Organization Site, 2002).  Seventy percent of 
these were approved in the last 6 years. As indicated by the latest available information, 
some 371 biopharmaceuticals were in clinical trials in the USA (New biotechnology 
medicines in development report, 2002).  More than 300 of them were protein based, 
with recombinant vaccines and monoclonal/engineered antibodies representing the two 
 2 
most significant product categories.  Over the coming decade, therefore, approximately a 
dozen new therapeutic proteins should, on average, gain regulatory approval each year. In 
the similar time frame of 1995-2002, 88 recombinant proteins/monoclonal antibody-
based products had gained marketing approval in the European Union (EU). This 
represents 36% of all new approvals in the EU in that period (Walsh, 2003). 
Despite the initial enthusiasm, it remains a fact that no gene therapy based product 
has thus far been approved in either the USA or EU.  Clinical trials have also reported a 
disappointing lack of efficacy and in some cases have raised safety concerns (Stein, A., 
1998; Pfeifer, A., 2001; Phillips, A., 2001; Lebedeva, I., 2001).  The failure of nucleic 
acid based products to meet original expectations in no way reflects a flaw in the 
concepts of gene therapy or antisense technology.  These technologies retain the potential 
to revolutionize medical practice.  Their full potential will however, accrue only when 
several associated technical difficulties pertaining to stability and regulation of 
expression and gene/nucleotide delivery are overcome (Walsh, G., 2003). 
This brief overview of the drug discovery area shows a significant, steady but gradual 
change of trend.    Increasingly, more and more of the new drug molecules approved 
world wide are and will be macromolecular in origin.  These so called 
biopharmaceuticals including polypeptides, proteins, macromolecule-protein conjugates 
and nucleotides present a new and emerging challenge to the science and technology of 
formulation and/or drug delivery. 
Biopharmaceuticals, especially proteins are fundamentally different from low 
molecular weight drugs.   Unlike low molecular weight drugs, the chemical 
 3 
structure of the proteins, as described by their amino acid sequence (primary structure), 
does not  adequately define them.  They have specific native three dimensional structures 
based on secondary structures (like alpha-helices, beta sheets and random coils, etc.), 
tertiary structures (folding of secondary structures into complicated three dimensional 
arrays) and in some cases quaternary structures (where different monomers interact).  The 
secondary, tertiary and quaternary structures of proteins often change with the change in 
environmental conditions like polarity and ionic strength of the media, pH, temperature 
and presence of foreign molecules, often leading to irreversible loss of native structure 
(called denaturation) and thus the activity of the protein.   Accurate and absolute 
quantification of these natural and denatured configurations of therapeutic proteins is 
difficult and often impossible with currently available analytical tools (Crommelin, D. J. 
A., 2003). 
These conformational complexities aside, proteinacious drugs are also complicated in 
terms of their stability and effectiveness.  Peptide bonds are hydrolyzed in extreme pH 
ranges, like the acidic pH of media present in sections (stomach) of the gastro-intestinal 
tract (GIT).  Proteolytic enzymes, present in the GIT, can also break down proteins into 
small polypeptides and constituent amino acids.    Protein drugs are fast metabolized by 
the liver and eliminated from the body.  The biological half-life of protein drugs is 
usually very short, ranging from a couple of minutes to a few hours, while their medical 
applications frequently require prolonged residence time in the body.   There are inter-
species variations in chemical structure of biologically active proteins. The structural 
difference is often enough to cause an immune response.  Purity of proteins is also a great 
 4 
concern.  A small amount of bio-molecular impurity, usually coming from expression 
systems and growth conditions, can impart immunogenicity. 
All of these structural and conformational complexities of the therapeutic 
macromolecules like polypeptides and proteins, set them apart as a completely different 
category of drugs.  They have to be formulated, handled and delivered differently with 
the materials, methods and route(s) of administration carefully chosen to adequately 
address these complexities.  The classical approach to drug therapy, through oral 
administration, is thus not applicable for these macromolecules.   
Peroral delivery of proteins is hampered by several factors. The main impediment is 
that they are rapidly hydrolyzed by proteolytic enzymes and the acidic pH in the gut.  
Even if the proteolysis is controlled and overcome to a manageable extent, the trans-
epithelial absorption of molecules beyond the molecular weight of one thousand Daltons 
is pathetically inadequate.   There are emerging approaches directed towards oral delivery 
of macromolecules.  These include stabilization and protection of the macromolecules in 
the harsh environment of the GIT (Sood, A., 2001) and improving the membrane 
permeability, either by the use of permeability enhancers (Sharma, P., 1999) or by 
exploiting the natural carrier system in the GIT (Orelana, I. G., 1998).  Despite these 
developments, a commercial and successful peroral delivery of proteins is still a distant 
goal. 
The significant obstacles associated with the peroral route have forced the focus to 
other possible routes of drug delivery like transdermal.  While the transdermal route has 
the inherent advantage of simplicity and ease of application, it also presents enormous 
challenges and limitations.  Penetration of drugs through the stratum cornium is 
 5 
inadequate, even for most small molecules.  There seems to be a critical need of optimum 
polarity and molecular weight of the drug molecules for adequate permeation through the 
transdermal route.  This, in turn, limits the applicability of the route for drug delivery.  
Since the introduction of the scopolamine patch in 1983, only eight drugs have been 
commercialized in the next seventeen years, proving that drug delivery through the skin 
has limited utility (Naik, A., 2000).  Chemical and physical enhancement techniques like 
iontophoresis, electroporation, sonophoresis and high speed powder delivery are under 
active investigation to improve the scope of drugs (in terms of size, lipophilicity and 
charge) that can be delivered through the skin (Pilli, O., 2001).  Transdermal delivery of 
proteins and polypeptides with all these chemical and physical enhancement tools are still 
experimental technologies with many unresolved issues like skin safety considerations 
(Panchagnula, R., 2000). They are far from reaching a stage of critical maturity and are 
unlikely to impact macromolecule delivery significantly in the foreseeable future. 
Another promising route for macromolecule delivery is nasal mucosa.  Of all the 
mucosal tissues, nasal mucosa is considered to be most permeable and offers potential for 
peptide delivery (Patridos, C. D., 2000; Brayden, D. J., 2000), particularly for vaccines.  
Already several peptide formulations are on the market and further development will 
depend on minimizing the immunogenicity of the bioactive agents and the long-term 
effect of the permeation enhancers on the nasal mucociliary activity. 
The lung also offers a potential opportunity for systemic delivery.  It has enormous 
absorptive surface area comparable to that of a tennis court and is rich in blood supply 
(Gonda, I., 2000).  This route is targeted through reservoir-type micro/nano particulate 
drug delivery devices. However, the major limitations are high dose requirements, low 
 6 
efficiency of deposition on alveolar surfaces, and the reproducibility of metered doses.  
Although it is possible to design particles close to the desired characteristics for 
aerosolization using supercritical fluid technology (York, P., 1999), further progress 
awaits improved aerosol design.  The new generation of breath activated dry powder 
inhalers are an improvement in that direction (Ashrst, I., 2000).   
Progress is being made in the delivery of macromolecules including the proteins and 
polypeptides.  Nevertheless, in spite of a few marginal exceptions, an overwhelming 
majority of the macromolecules are being administered today by intra-venous bolus 
injections.  Due to the short plasma half-life and the usual need for chronic treatment with 
these drugs, repeated intra-venous bolus injections are used.  This mode of treatment 
suffers from a number of disadvantages including inconvenience, lack of patient 
compliance, need of hospitalization or professional care and most importantly high cost 
of treatment.  The search for alternative routes and delivery systems for proteins and 
polypeptides are all aimed at eliminating or minimizing some or all of these 
disadvantages.  From this point of view, obviously, absorbable oral formulations are the 
ideal answer.  But unfortunately, that is not feasible with the current level of technology.  
Of the other routes, the pulmonary and injectable routes currently show most promise.  In 
both these routes drug release is controlled and sustained by incorporating the drug(s) 
within micro/nano meter sized, reservoir-type, polymeric particles.  This current research 
is aimed at developing the processing technologies for making such micro/nano particles, 
overcoming some of the limitations associated with the current technologies for making 
such particles and establishing their effectiveness with respect to macromolecules and 
investigating their effectiveness with and possible use for small drug molecules.  In the 
 7 
next section, the status of prevalent current technologies, their merits and limitations, the 
choice of appropriate polymers for such technologies and devices will be explored with 
respect to their applicability and usefulness in the current context. 
 
REFERENCES 
 
Ashrst, I., Malton, A., Prime, D.,Sumy, B.  (2000).   Latest advances in the development 
of dry powder inhalers.  Pharmaceutical Science and Technology Today, 3, 246-
256. 
 
Biotechnology Industry Organization Site, 29th August, 2002. 
http://www.bio.org/er/statistics.asp. 
 
Brayden, D. J. (2000). Vaccinology ? the challenges for non injected approaches.  
Pharmaceutical Science and Technology Today, 3, 115-117. 
 
Crommelin, D. J. A., Strom, G., Verrijk, R., Leede, L., Jiskoot, W., Hennik, W. E. 
(2003).  Shifting paradigms: biopharmaceuticals and low molecular weight drugs.  
International journal of Pharmaceutics, 266, 3-6.  
 
Drews, J.  (2000).   Quovadis biotech? (Part I).  Drug Discovery Today, 5, 547-553.  
 
Gonda, I.  (2000).   The ascent of pulmonary drug delivery.  Journal of Pharmaceutical 
Science, 89, 940-945. 
 
Levadova, I. & Stein, C.  (2001). Antisence oligonucleotides: promise and reality.  
Annual  Review of  Pharmacology and Toxicology, 41, 403-419. 
 
Naik, A., Kalia, Y. N., Guy, D. R. (2000).  Transdermal drug delivery: overcoming the 
skin?s barrier function.  Pharmaceutical Science and Technology Today, 3, 318-
326. 
New biotechnology medicine in development report (5th October, 2002).  
http://www.pharma.org. 
 
Orellana, I. G. & Paton, D. R. (1998).  Advances in oral delivery of Proteins.  Expert 
Opinion on Therapeutic patients, 8, 223-224. 
 
 8 
Panchagnula, R., Pillai, O., Nair, V. B., Ramarao, P.  (2000).   Transdermal Iontophoresis 
revisited.  Current Opinion on Chemical Biology, 4, 468-473. 
 
Patridos, C.D.  (2000). Intranasal vaccines: forthcoming challenges.  Pharmaceutical 
Science and Technology Today, 3, 273-281. 
Pfeifer, A. & Verma, I.  (2001). Gene therapy: promises and problems.  Ann. Rev. 
Genomics & Human, Genetics, 2, 177-211. 
 
Phillips, A.  (2001). The challenge of gene therapy and DNA delivery.  Journal of 
Pharmacy and Pharmacology 53, 1169-1174. 
 
Pilli, O., Nair,V.,Jain, A. K., Thomas, N. S., Panchagnula,R.  (2001). Non-Invasive 
transdermal delivery of peptides and proteins through skin.  Drugs Future, 26, 
779-791. 
 
Sharma, P. & Chawla, H. P. S.  (1999).   The role of sorption promoters in increasing the 
bioavailability of drugs in oral preparations.  Drugs Future, 24, 1221-1240.  
 
Sood, A. &Panchagula, R. (2001).  Peroral route: an opportunity for protein and peptide 
delivery.  Chemical Reviews, 101, 3275-3303. 
 
Stein, A.& Kreig, A. (1998).  Applied Antisence Oligonucleotide Technology.  J.  Wally 
and Sons, Chichester,  UK, Pp. 36-71. 
 
Walsh, G.  (2003).  Pharmaceutical biotechnology products approved within the 
European Union.  European Journal of Pharmaceutics and Biopharmaceutics, 55, 
3-10.  
 
York, P.  (1999).   Strategies for particle design using supercritical fluid technologies.  
Pharmaceutical Science and Technology Today, 2, 430-440. 
 
  
 9 
II. LITERATURE   REVIEW   
 
1. INTRODUCTION 
Over the last three decades, particulate drug delivery systems have been the 
central focus of drug delivery research.  An enormous amount of research effort from 
both academia and industry has been devoted towards the development of these 
particulate systems to deliver drugs safely, effectively and in a manner that improves 
the therapeutic efficacy of the drug as well as makes the intervention process more 
patient-friendly.  These delivery systems essentially consist of spherical, 
biodegradable polymeric devices containing the active drug and with size ranging 
from nanometer to a few hundred micrometers. As there has been some confusion 
regarding the terminology, for the purpose of uniformity, spherical particles in the 
sub-micron range are termed nanospheres and those above one micron are termed 
microspheres in this text.   
Particulate drug delivery systems have been prepared from a host of different bio-
degradable and non-degradable polymers, both hydrophilic and hydrophobic.  The 
term biodegradable polymer is defined as a polymer which degrades by either 
hydrolysis or solubilization in the physiological conditions in vivo.  These polymeric 
systems include both non water soluble polymers that are degraded 
 10 
through surface or bulk-erosion,  and water soluble gels that are dissolved and cleared  
slowly from the body without undergoing a decrease in molecular weight.   
Sustained release delivery systems have been successfully prepared in the past using 
non-degradable polymers such as Implanton? from Organon (Tice, T., 2004).  Implanton 
is a single rod shaped formulation made with ethylene vinyl acetate, a non-degradable 
polymer. The long-acting contraceptive formulation releases the steroid molecule 
etonogestrel for three years.  But despite the market success of Implanton, non 
biodegradable polymers are not generally used for the preparation of particulate sustained 
release delivery platforms mainly for two reasons; (1) the need for implantation and 
surgical removal of the implants and (2) the concern about long-term histo-toxicological 
effects of the non degradable platforms.  Consequently, almost all current developments 
in this area are carried out with biodegradable polymeric systems, though there is still 
some interest in non biodegradable polymers in the area of sustained local delivery of 
drugs. 
Significant progress has been made in the sustained release systemic delivery of drugs 
using biodegradable polymers, both through particulates and implants.  All such 
commercial successes involve the use of co-polymers of lactic acid and glycolic acid 
(PLGA) in various combinations of different proportions of the monomers and also the 
molecular weight of the overall polymer.  Currently there are 15 different PLGA depot 
formulations in the US market (Tice, T., 2004), involving 13 drug companies and 10 
different drug molecules.  These products deliver small molecular weight antibiotics 
(doxycycline, minocycline), antipsychotic (resperidone), small polypeptides (LHRH 
analogs triptorelin, luprorelin) and one large molecular weight protein, growth hormone, 
 11 
for a time frame of one week to four months.  Many PLGA depot formulations are being 
developed today.   Products in preclinical and clinical stages include formulations of 
naltrexone and buprenorphene for alcoholism/drug abuse/terminal pain treatment, GLP-1 
peptides for diabetes, r-hFSH for fertility, dopamine for nerve growth, dexamethasone for 
ocular application, melanotan and plasmid DNA for cancer prevention, a variety of 
vaccines and the generics of luprorelin and octreotide.  
As evident here, micro/nano particulate sustained release delivery systems have come 
a long way from being just another new idea in drug delivery research.  The idea has been 
proved to be effective and achievable beyond doubt by the commercialization of the 
products enumerated above.  It is now well accepted that traditional small molecular 
drugs can be effectively incorporated in the particulate devices and released in the body 
in a sustained manner and the release rate can be potentially tailored according to the 
need.  The application of this technique to small molecules will be limited only by 
pharmacological profiles of the small molecules and the clinical needs.  Currently in most 
clinical situations, long term treatment with traditional small molecules is simply 
achieved through multiple oral dosing (though the issue of lack of patient compliance 
remains unresolved).  Development of sustained release formulations of these molecules 
are theoretically possible but in most cases not driven by the market because of cost 
considerations. But there are some special clinical situations like long term de-addiction 
therapy with opiates like buprenorphene, where a sustained release injectable delivery 
system is the only option because of the concerns of drug abuse (Mondal, T., K., 1999).  
Also, injectable particulate delivery systems for sustained release are the only way to 
 12 
achieve long term constant plasma concentration of some small molecular weight drugs 
that show inconsistent and low bioavailability and short half-life. 
Another important application of particulate delivery systems is in the delivery of 
proteins and polypeptides.  Quite a few small polypeptides are currently in the market.  
These polypeptides, mainly leutenizing hormone releasing hormone and follicle 
stimulating hormone analogs, are small in size and thus lack the secondary, tertiary and 
quaternary structures.  Thus they are relatively easier to formulate since the problem of 
denaturation and consequent loss of activity does not have to be dealt with for these 
molecules.  The enormous challenge lies in the delivery of big protein molecules while 
preserving their native structures and activity.   A modest beginning has been made by 
the commercialization of a single growth hormone formulation.  Still generalized use of 
sustained release injectable delivery systems for the delivery of bigger proteins through 
the use of particulates remains elusive.  A more thorough understanding of the micro-
environmental requirements and restrictions imposed on the preparation process and 
delivery platform by the complicated structures of the big proteins are needed, potentially 
in addition to newer approaches to address these issues. 
Particulate delivery systems can also be potentially used in other emerging areas like 
gene therapy.  In this field, disease conditions are approached from the perspective of 
genetic correction of the physiological system through the introduction of a DNA 
fragment into the cells (usually in the form of plasmids with appropriate helper domains).  
But unfortunately there are no effective ways of introducing the plasmids into cells in 
vivo in the clinical environment.  The currently used gene transfection vectors like adeno 
virus, adeno-associated virus (Monahan, P. E., 2000), cationic lipids or lipoplexes (Ferrai, 
 13 
S., 2002) and low molecular weight citosan (Gao, S., 2003) are not effective in gene-
transfection.  In addition, there are other problems peculiar to each of the platforms.  
Particulate delivery systems have been shown to be taken up by different mammalian cell 
lines presumably by endocytosis and fluid phase pinocytosis (Desai, M., 1997; Wilhelm, 
C., 2003).  It is also established that the uptake is size dependent, the smaller the size the 
higher is the efficiency of uptake (Desai, M., 1997; Prava, S., 2002).  Other factors 
influencing the uptake include the surface polarity, surface charge and presence of some 
receptor-specific molecules on the surface of the particulates (Wilhelm, C., 2003).  
Experiments have also proved that uptake of bio-degradable PLGA nanoparticlates into 
the cell does not affect the cell viability (Davda, J., 2002).  Thus particulate delivery 
systems, especially the nanoparticles, have potential as DNA delivery vectors.  This 
application is currently being pursued seriously. 
Particulate delivery systems seem to be fast developing into a basic platform which 
can be used to serve different purposes with appropriate modifications.  Currently long 
circulating reservoir type applications are not technologically feasible.  The successful 
systems involving PLGA as polymers are limited by the surface properties that activate 
the natural immune system of the body and quickly eliminate the particulate drug 
delivery system (DDS) (Stolnik, S., 1995).  This characteristic of the PLGA particulates 
limit their applications to intra-muscular or subcutaneous depot forming formulations.  
However, this property can be exploited in specific cases where the targeted site is the 
immune system itself, for example in acquired-immuo-deficiency syndrome (AIDS) 
treatment.  Long circulating reservoir type application is possible with this platform if the 
surface can be adequately modified to escape the immune system.  Once that is 
 14 
technologically possible, long circulating particulate depots can be targeted passively by 
the use of magnetic field or actively by antibody targeting to the site of action.  This 
platform can also be used in imaging applications by incorporating appropriate 
molecule(s).  It has the promise to affect a paradigm shift in a broad spectrum of medical 
technologies as it matures over time. 
 
2. GENERAL CONSIDERATIONS 
Since the particulate DDS has to function in a dynamic physiological environment, 
enormous complexity is involved in determining the overall applicability, efficiency and 
success of the systems.  The multitude of factors affecting the particulate DDS can be 
grossly classified into two groups; the physical factors and the physiological factors. 
 
2A. PHYSICAL FACTORS 
The physical factors affecting the particulate DDS include the polymer related 
factors, the active ingredient related factors and the process related factors.  It is 
important to note that while some of these factors are independent, most have a very 
complex interdependent relationship with one another. 
  
2A (i) THE POLYMER RELATED FACTORS  
The polymer is the most abundant component of the particulate DDS.  The properties 
of the polymer determine the behavior of the DDS in physiological conditions to a large 
extent.  The chemical structure, stereochemistry, crystallinity and molecular weight are 
important factors that determine the polymer properties in this context. 
 15 
Generally speaking, the more non-polar the constituent monomer, the slower is the 
degradation of the resulting polymer in the physiologic conditions.  A common example 
is the erosion controlled particulate DDS composed of poly(lactic/glycolic) acid.  Lactic 
acid is more hydrophobic than glycolic acid.  The micro/nano spheres fabricated with 
poly(lactic acid) (PLA) and poly(lactic-co-glycolic acid) (PLGA) of comparable 
molecular weight show different rate of polymer degradation accompanied by different 
drug release kinetics.  On exposure to water, the bulk erosion polymers like PLGA and 
PLA undergo random chain scission by simple hydrolysis of the ester bonds following 
hydration of the polymer.   Thus PLA being more hydrophobic than PLGA, the extent of 
hydration and consequently the extent of degradation of the former in physiological 
conditions is less.  This results in a slower release of drugs from PLA spheres than PLGA 
spheres (with polymers of same molecular weight and specific area).  The effect of 
molecular weight and copolymer ratio on the extent and rate of drug/macromolecule 
release has been reviewed (Ogawa, Y., 1988). Since, the drug/macro-molecule has to be 
released into the body-fluid, hydration of the polymeric surface of the devices is 
important not only for degradation of the polymeric spheres but also the actual release 
kinetics.  The more the hydrophobicity, the less the hydration and also the less is the 
release rate.  The polymer polarity also determines the solubility profile of the polymers.  
All the commonly used polymers for sustained release DDS are moderately to highly 
non-polar and thus are soluble in solvents of moderate to high degrees of non-polarity.  
These solvents are generally not compatible with biological macromolecules.  Thus the 
polarity or hydrophobicity of the polymer is a critical constraint in selection of a solvent 
 16 
system which is benign enough with regards to its chemical activity on the bioactive 
molecule to be used to form the DDS.   
In addition to composition and molecular weight, other factors such as crystallinity 
and stereo-regularity of the polymer, affect its polarity and solubility.  While low 
molecular weight polymers tend to be more hydrophilic and degrade faster than high 
molecular weight polymers (Kwong, A. K., 1986) the stereo-regularity of the monomer 
units affects the tightness of packing of polymer chains.  The more the stereo-regularity, 
the tighter is the packing of polymer chains which become more crystalline and less 
amorphous.  The crystallinity of the polymer disfavors its degradation and slows down 
drug release kinetics (Migliaresi, C., 1994).  Thus poly-L-lactide is crystalline while 
poly-D, L-lactide is amorphous and the degradation time of the former polymer is longer 
than the second one (Sinha, V. R.  2003).    
Another important factor in the selection of the polymer is its bio-degradability.  The 
rate of degradation may also be an important factor in the selection process as it might be 
used to control the drug release rate in erosion controlled systems.  There are several 
kinds of labile bonds that are used to form biodegradable polymers.  These bonds may be 
categorized by their relative rates of hydrolysis under physiological conditions based on 
the known hydrolysis rates of low molecular weight analogs (Baker, R., 1987) as shown 
below. 
 
Polycarbonate   >  Polyesters  >  Polyurethanes  >  Polyorthoesters  >  Polyamides 
 
 17 
Since polymer morphology and the presence of substituent groups can influence the 
rates of hydrolysis significantly, the above comparison is only an approximate guide.   
Biodegradable polymers containing almost all of the above mentioned bonds have been 
used to produce particulate DDS in an effort to achieve a required level of balance 
between release rate, biodegradability, biocompatibility, toxicity and encapsulation 
efficiency.  Examples include bulk erosion polymers like PLA, PLGA blends, block co-
polymers of PLGA and PEG (polyethylene glycol), poly(cyanoacrylates), surface erosion 
polymers like poly(anhydrides) and poly(orthoesters), hydrogel systems like polyols 
(pluronics), poly(vinyl alcohol), poly(vinyl pyrrolidone) (PVP), hyaluronic acid 
derivatives, alginate, collagen, albumin, starches and dextrans  (Gombotz, W. R.,  1995).  
Another important issue to be considered is the biocompatibility of the polymers.  
Polymers used for particulate DDS need to exhibit biocompatible characteristics in terms 
of both the polymer?s effect on the organism receiving the DDS and the polymer?s effect 
on the macromolecule or protein incorporated.  Several aspects of polymeric DDS 
ultimately contribute to its overall biocompatibility, or lack thereof.  The polymer itself 
may potentially be antigenic (De Lustro, F., 1990; 1986), carcinogenic (Nakamura, T., 
1994; Weiss, W. M., 1991), or toxic (Yoshida, S. H., 1993; Busch, H.,  1994).  The shape 
of an implanted material has been implicated in its biocompatibility as well. Smooth 
surfaces are less irritating and more biocompatible than rough surfaces (Matlaga, B. F., 
1976).   One key factor influencing the biocompatibility of the polymer is the presence of 
low molecular weight extractables.  These extractables generally come from the un-
reacted monomers and/or polymerization initiators.  The monomers as well as the 
polymerization initiators can be toxic at times.  For example, acrylic acid and methacrylic 
 18 
acid are toxic, while their polymers are relative non-toxic (King, D. J., 1989).  These 
extractables are generally removed by organic solvents prior to manufacture or 
preparation of DDS.  However, the organic solvents used to remove the extractables are 
themselves not biocompatible and can be a cause of concern by themselves.  In some 
cases, soluble polymers or their breakdown products are sequestered within organs, 
resulting in long term adverse effects.  The accumulation of high molecular weight PVP 
in liver following prolonged exposure is one such example (Roske-Nielsoen, E., 1976; 
Meijer, A.E.F.H., 1963).  All these polymer related issues have to be considered for 
selection of the appropriate one for a specific DDS application. 
 
2A (ii). THE MACROMOLECULAR DRUG RELATED FACTORS 
Incorporating low molecular weight drugs in polymeric particulate DDS and attaining 
sustained release has been achieved.  Macromolecular drugs, especially proteins, 
however, behave quite differently in these systems.  Generally they give rise to a 
multitude of additional complications resulting in poor encapsulation efficiency, 
denaturation and concomitant loss of potency, and incomplete release among others.  
There are a number of different properties and factors peculiar to the macromolecules 
that lead to these drawbacks. 
The first factor to be considered is the protein molecular weight.  As the molecular 
weight of the macromolecule increases, the hydrodynamic radius increases.  The 
diffusion coefficient of the solute is inversely related to the hydrodynamic radius of the 
solute molecule.  As the molecular weight and consequently the hydrodynamic radius of 
the solute molecule increases, the diffusion coefficient of the solute decreases 
 19 
logarithmically (Colton, C. K., 1971).  So, as the molecular weight of the protein drugs 
increase, their release rate decreases dramatically. The low diffusivity of high molecular 
weight proteins, while not unduly restrictive, has to be factored into the design of the 
DDS. 
The isoelectric point of the protein is an important parameter determining the 
ionization state of the protein at a particular pH.  The ionization state of the protein 
governs ionic interactions of the individual protein molecule with other protein molecules 
as well as with the polymer in some cases, where the polymer is also capable of being 
ionized.  Thus formation of dimers and oligomers of the proteins and also adsorption of 
the protein to the polymer surface partially depends on its isoelectric point and the pH of 
the microenvironment.  Additionally, at the isoelectric point, the protein has its lowest 
aqueous solubility. This also restricts the pH range of operability.   
Sulphur containing amino acids like cystine is another source of concern for stability.  
Cystine is known to participate in covalent interactions like disulphide bonding both intra 
and inter molecularly.  This interaction is often responsible for the formation of 
oligomers and concomitant denaturization of the molecule and loss of potency.  
Aggregation of lyophilized formulations of albumin, ?-lactoglobulin and glucose oxidase 
is attributed to disulphide interchange (Johnson, O. L., 1999).  Other chemical 
degradations like oxidation (Pearlman, R., 1993) and deamination are also possible.  
Deamination is known to cause reduction of catalytic activity of lysozyme (Manning, M. 
C., 1993) and ribonuclease at high temperatures (Zale, S. E., 1986). 
Protein stability in particulate DDS is partially determined by their primary sequence 
in various ways. The primary amino acid sequence determines the hydrophobic domains 
 20 
of the protein.  The hydrophobic domains are often implicated in protein adsorption on 
the polymer surface.  Human serum albumin is known to undergo multilayer adsorption 
to PLA nanospheres. At least a fraction of the human serum albumin, adsorbed on the 
PLA surface, is known to be adsorbed irreversibly (Verrecchia, T., 1993). Protein 
adsorption to polymeric materials has been widely studied in the context of polymeric 
implants (Horbett, T.A., 1987).  The deleterious effects of protein adsorption to 
polymeric surface are significant in the context of polymeric particulate DDS containing 
proteins as well. Attempts have been made to address the problem of protein adsorption 
on polymeric surfaces by addition of stabilizing excipients (Liu, W. R., 1991; Liu, W. R., 
1995; Constantino, H. R., 1994).  The primary amino acid sequence of the protein is also 
important in that it partially determines the possibility of chemical modification/reaction 
when in association with polymeric materials.    
The presence or absence of carbohydrate in association with proteins is known to 
alter the interactions of proteins with polymer surfaces.  Additionally, the presence of a 
carbohydrate also affects the hydrodynamic volume (Gombotz, R., 1995).   
There are additional factors that are encountered regularly in development of 
particulate DDS.  When particulate DDS containing proteins are administered, the 
incorporated proteins may become hydrated at relatively high concentrations for 
prolonged periods of time.  Proteins in this kind of environment are susceptible to 
denaturation and aggregation (Liu, W. R., 1991).   Additionally, polymer degradation is 
induced.  The degradation products are highly concentrated in the micro-environment.  
These degradation products can be a source of protein instability including degradation 
and/or chemical modification of the proteins, especially in situations where the polymer 
 21 
degradation products change the pH of the microenvironment significantly.  In cases 
where the protein itself has significant moderately hydrophobic domains, in addition to 
adsorption to polymeric surface, the problem of inadequate protein hydration is often 
encountered.  This leads to incomplete protein release as well as protein aggregation and 
denaturation.  This problem has been addressed by addition of solubilizing or hydration 
additives (Ogawa, Y., 1988; Okada, H., 1991). 
In a different approach, modification of the protein molecule itself and the polymer 
by introducing water soluble substituent has been attempted to prevent protein 
aggregation (Delgado, C., 1992) and/or adsorption (Cohn, D, 1988; Sawhany, A. S., 
1994).  Protein modification with PEG has been accomplished utilizing at least two 
proteins in PLGA delivery systems; interleukin-2 (IL-2) (Hora, M.S., 1990) and 
granulocyte colony stimulating factor (C-GSF) (Camble, R., 1994). In both cases, the 
unmodified protein exhibited poor release profile and much of the protein remained 
trapped within the polymer after several weeks of incubation in solution.  The poor 
release was attributed to difficulty in resolubilization of the protein.  The PEG modified 
proteins, however, were released much more readily from the systems, probably due to 
increased solubility, decreased aggregation, and decreased protein adsorption on the 
polymeric surfaces.  All the above mentioned factors related to the macromolecular drug 
contribute to the overall performance of the particulate DDS and thus need to be taken 
into account in their design. 
 
 
 
 22 
2A (iii). THE PROCESS RELATED FACTORS 
The assortment of processes and techniques used to produce micro/nano sized 
particles for drug delivery purposes are generally referred to as microencapsulation.  The 
microencapsulation process is generally dominated by the polymer used for the particle 
formation.  Most polymers used and/or investigated for this purpose are hydrophobic.  
Consequently, a non-polar solvent becomes a requirement for the purpose.  Almost all the 
microencapsulation processes currently in use are variations or modifications of three 
basic techniques: solvent extraction/evaporation, phase separation (coacervation) and 
spray drying (Aftabrouchad, C., 1992).   
Spray drying is the simplest technique among the three mentioned and has the highest 
throughput.  But this process uses elevated temperature and thus is not suitable for 
temperature sensitive molecules.  Additionally, control of particle size is difficult and 
yield for small batches is moderate (Johansen, P., 2000).  Coacervation, though elegant as 
a technique, is frequently impaired by residual solvents and coacervating agents 
remaining in the microsphere (Thomasin, C., 1996).  Also, this technique is not well 
suited for producing microspheres in the low micrometer size range.  To minimize the 
problem of residual solvent and/or coacervating agent, an alternative approach involving 
supercritical fluid has been tried.  This approach has resulted in new processes namely 
?precipitation with compressed antisolvent? (PCA) (Falk, R. F., 1998), ?gas or 
supercritical fluid antisolvent? ((GAS or SAS) and ?aerosol solvent extraction system? 
(ASES) (Jung, J., 2001).  These techniques are in the development stage and are currently 
limited to very small laboratory batch sizes.  Kilo scale or industrial scale production 
using these techniques have not been reported. 
 23 
With the obvious limitations of the techniques discussed above, solvent 
evaporation/extraction seems to be the most appropriate technique to be used for the 
purpose of preparing/fabricating micro/nano particles to be used as drug delivery 
platforms.  Solvent extraction/evaporation has been used for a long time as a preferred 
technique for preparation of polymeric micro/nano particulates.  It does not require high 
temperature or potentially toxic phase separation inducing agents.  Particles with size 
ranging from micro to nanometer have been fabricated using this method, though precise 
control of particle size and size range is still an issue.  Polydispersity is often 
encountered.  Also, careful selection of the encapsulation conditions is required in order 
to achieve a high degree of encapsulation efficiency.  Because of the advantages, this 
technique will be discussed in more detail.  Microencapsulation by solvent 
extraction/evaporation contains four major steps: (a) incorporation of the active 
compound/drug along with the matrix forming polymer in a liquid solvent phase 
(frequently organic as dictated by the solubility profile of the matrix forming polymer), 
(b) emulsification of this first liquid phase into another immiscible  liquid phase 
(frequently aqueous), (c) Extraction of the solvent from the dispersed droplets by the 
continuum phase accompanied by solvent evaporation, (d) harvesting and drying of the 
particles. 
Incorporation of active ingredient 
Lower molecular weight drugs (i.e. molecular weights < 1000 Daltons) have more 
favorable solubility profile.  They are generally of intermediate polarity and can be 
dissolved in a solvent or solvent system with solvents ranging from acetone to hexane.  
The biodegradable matrix-forming polymers commonly used have similar solubility 
 24 
profiles.  As a result, for small molecular weight drugs, the incorporation step is 
relatively easier because both the drug and the matrix forming polymer can be co-
dissolved in the same solvent or solvent system. 
Unfortunately, for the macromolecular bioactive compounds like proteins this step is 
immensely complicated by the very different solubility profiles of proteins.  Proteins are 
soluble and have biological activity only in aqueous systems and denature on exposure to 
most organic solvents especially methylene chloride (Weert, M., 2000; Sah, H., 1999).  
The protein incorporation problem is thus approached by two different methods; 
dispersion of small solid micronized protein particles into the solution of the matrix 
forming polymer; and the dispersion of an aqueous solution of the protein in the organic 
solution of the matrix forming polymer. 
Incorporation of micronized solid protein particles into the matrix forming polymer 
has been successfully attempted.  It is found that encapsulation efficiency is improved 
with the decrease in size of micronized protein particles (Maa, Y. F., 1997; Al-Azzam, 
W., 2002).  Also spherically shaped protein particles are found to be more efficiently 
encapsulated than irregular ones (Maa, Y. F., 1997).  There is concern about protein 
stability during the micronization process itself. Unfortunately very little research has 
been directed towards this problem.  Consequently, little literature evidence has been 
found addressing this concern.  Finally, in this technique the size of the drug delivery 
platform is limited by the size of the solid protein particles. 
Proteins are more commonly microencapsulated in the laboratory by water-in-oil-
water double emulsion technique.  In principle, an aqueous protein solution is emulsified 
in an organic matrix forming polymer solution and this primary emulsion is again 
 25 
emulsified in an external aqueous phase.  These techniques are complicated and difficult 
to control and scale up.  Generally it is found that increasing the volume fraction of the 
internal aqueous phase lowers encapsulation efficiency due to droplet coalescence and 
increased probability of contact between the internal drug solution and the external 
extraction phase resulting in drug loss (Yang, Y.Y., 2000; Crotts, G., 1995).  Increasing 
the volume fraction of the internal dispersed phase of the first emulsion also results in an 
increase in particle porosity and burst release (Yang, Y. Y., 2000; Crotis, G., 1995).    
Emulsification 
Emulsification is the second and probably the most complicated and important step of 
the particle forming process.  Various different techniques have been employed starting 
from the most rudimentary technique of stirring to increasingly more sophisticated 
techniques like homogenization with a rotor-stator arrangement, static mixing and 
extrusion through micro-porous membranes.  All of these techniques produce emulsions 
with different degrees of efficiency, precision and control.   
The most important factors that are affected by the emulsification process are 
encapsulation efficiency, porosity of the particles, denaturation of the macromolecules, 
particle size, size distribution and reproducibility.  These factors affect the release 
properties of the active ingredients from the particles and thus the overall efficacy and 
applicability of the platform. The first three factors are affected by the proper choice of 
polymer, dispersed phase, continuous phase, phase ratio, surfactant and surfactant 
concentration.  
One of the problems of encapsulating macromolecules like proteins is that they often 
have significant surfactant properties.  Most proteins tend to partition to the oil-water 
 26 
interface to a significant degree as they have both hydrophilic and hydrophobic zones.  
When drug free microparticles are prepared from emulsions consisting of plain water and 
PLA dissolved in methylene chloride, increasing amounts of bovine serum albumin 
(BSA) added to the water as a surfactant stabilizes the emulsion and decreases the pore 
size of the resulting microspheres.  The addition of BSA is thought to be responsible for 
the surfactant activity, bringing down the water droplet size of the dispersed internal 
aqueous phase.  As the water is extracted, it leaves behind smaller sized pores in the 
resulting microsphere (Nihant, N., 1994).  Also a significant but varied degree of protein 
loss is observed from the aqueous phase when protein solutions are emulsified with 
organic phases (Sah, H., 1998).  All these observations point to the fact that proteins have 
varied but significant degrees of surfactant activity.  This surfactant activity has a 
deleterious effect on encapsulation not only because it causes a loss of active ingredient, 
but also because the denatured proteins sometimes possess altered antigenic behavior 
Sah, H., 1998).  This problem can be addressed to a great extent by introducing 
competing surface active additives like partially hydrolyzed polyvinyl alcohol (Weert, 
Marco van de, 2000), hydroxypropyl-?-cyclodextrin (Sah, H., 1998), or a combination of 
sorbitan esters and polysorbates (Soriano, I., 1995) in the aqueous phase. 
Among all the emulsification techniques, the most rudimentary and simplest 
technique is stirring carried out in a so called continuously stirred tank reactor (CSTR).  
In this technique the two immiscible phases are forced together by the shearing force of 
an impeller.  The emulsion droplets are formed when the shear force is sufficient and are 
maintained by the action of surfactants.  Homogenization with a rotor-stator arrangement 
is an advanced version of the simple stirring where the shear force is complemented by 
 27 
cavitations and thus is more efficient in particle size reduction.   Obviously, the most 
important independent variable that dictates the particle size is the shear force.  Dispersed 
phase viscosity affects the microsphere size by increasing the shear force requirement to 
bring about comparable size reduction.  The increase in continuous phase viscosity 
results in a decrease in microsphere size as increased viscosity of the continuum phase 
increases the shear force (for a given rate of shear) acting upon the emulsion droplets.  
All available literature points to the simple fact that an increase in the speed of the 
stirrer/rotor results in decreased microsphere mean size (Yang, Y. Y., 2001; Sansdrap, P.,  
1993; Gabor, F., 1999; Mateovic, T.,  2002).  This can be explained by physical 
mechanics in which increased impellor/rotor speed results in higher shear force, which is 
the driving force in the size reduction process.  The impeller size and shape and the size 
ratio of the impeller to the mixing vessel also has impact on the microsphere size 
(Sansdrap, P., 1993), not surprisingly, as these parameters directly affect the shear force.   
Increased polymer concentration and thus increased viscosity of the dispersed phase 
is also known to increase the size of the microspheres produced (Jeyanthi, R.,  1997; 
Marchais, H.,1996; Yang, Y. Y., 2000; Viswanathan, N. B., 1999), as higher viscosity of 
the dispersed phase also requires higher shear force for size reduction.  Increasing the 
continuum phase viscosity by the addition of a viscosity building stabilizer has exactly 
the opposite effect on the microsphere size (Capan, Y., 1999; Yang, Y. Y, 2001; 
Sansdrap, P., 1993; Carrio, A.,  1995) because the increase in viscosity also increases the 
shear force acting upon the dispersed phase droplets at a given shear rate.    
The complex interplay of variables involved in determining the microsphere mean 
diameter in the so called CSTR setup has been mathematically modeled to arrive at an 
 28 
empirical equation (Maa, Y. F., 1996) relating various factors to particle size.   In 
agreement with previous reports, the model showed strong inverse correlation of the 
microsphere mean diameter with the increase of stirring speed, impeller diameter 
(decrease in microsphere diameter) and strong direct correlation with polymer 
concentration/dispersed phase viscosity (increase in microsphere diameter).  It also 
showed moderate inverse correlation with the increase in continuum phase viscosity 
(decrease in microsphere diameter) and moderate direct correlation with the increase in 
interfacial tension (increase in microsphere diameter).  The model reproduced and 
predicted the microsphere diameter with good accuracy for different types of extraction 
fluids and for microspheres with or without protein loading.  However, no prediction 
regarding the range of particle size distributions and polydispersity could be made. 
A relatively newer method of emulsification is through the use of static mixers.  
Static mixers essentially consist of a series of flow obstacles of baffles in a tube through 
which the two phase system flows.  The baffle arrangement ensures repeated splitting and 
recombination of the fluid stream and the size reduction occurs mainly by the turbulence 
created by the baffles.  This process has also been empirically modeled (Maa, Y. F., 
1996).  The correlation with the experimental data as well as the predictive power is 
found to be very good with the calculated mean diameter deviating from the 
experimentally determined diameter by less than 10%.   A close look into the model 
equation reveals that increasing the interfacial tension, polymer concentration in the 
dispersed phase and mixer diameter produce larger microspheres while increasing the 
flow rate through the static mixer, continuous phase viscosity and the length of the mixer 
produce smaller microspheres.  A comparison of the static mixers with the CSTR system 
 29 
for emulsification efficiency reveals that the static mixers produce the same degree of 
mixing at a much lower Reynolds numbers.  The uniformity of the particle size of the 
microspheres is not improved by the use of static mixers.  In terms of mixing efficiency 
and scalability, though, the static mixers have some advantage over the CSTR system 
(Freitas, S., 2005).   
Another more modern method of forming an emulsion is by extrusion.  Extrusion in 
this context means feeding the drug/polymer solution/dispersion through single or a 
plurality of pathways directly into the continuous extraction phase.  Upon leaving the 
pathway(s), discrete droplets of dispersed phase are formed within the slowly flowing 
continuous phase, which also transport the droplets from the site of their formation.  
Extrusion through single pathway systems has been studied for microsphere formation 
(Leelarasamee, N., 1996).  Generally it has been observed that the microsphere size is 
controlled by needle diameter and the flow rate of the continuous phase at the needle tip, 
and the adhesion between the forming droplet and the needle tip.  The resulting 
microspheres are large (> 50 micron) and also very poly-disperse with the coefficient of 
variation for the mean diameter varying from (15-40) %.  By increasing the continuous 
phase velocity, reducing the needle diameter and reducing the adhesion between the 
needle and the forming droplet with PTFE or silicone coated needles the polydispersity 
can be narrowed somewhat (Amsden, B. G., 2001).  Overall these systems are unsuitable 
for producing small microspheres (< 50 micron).  Multichannel extrusion systems 
specifically designed for producing mono-disperse particles has been successful in 
reducing the coeffcient of variation of the mean microsphere diameter to (< 5 %) 
(Kawakatsu, T., 1999).  Still, sub micron sized particles could not be produced, and 
 30 
achievable throughput in these systems is grossly inadequate for them to be practically 
useful beyond the laboratory. 
Membrane emulsification science has significantly evolved since the advent of a 
particular glass membrane called Shirasu porous glass and the subsequent observations 
that highly uniform kerosene-water and water-kerosene emulsions can be made using 
these membranes (Nakashima, T., 1991).  Since then other micro porous membranes of 
both inorganic and organic origin with pore size ranging from 10 nanometers to 150 
microns have been used by researchers for membrane emulsification (Vladislavljevic, G. 
T., 2005).  Products produced by this membrane emulsification through extrusion include 
metal solder particles, solid lipid and polymeric microspheres, and nonporous and porous 
core-shell microcapsules to name a few.  All types of emulsions including water-in-oil, 
oil-in-water, water-in-oil-in-water and oil-in-water-in-oil have been reported to be 
produced.  While virtually mono dispersed particles with good control over particle size 
can be produced by this method, preparing particles in the low sub-micron range is still 
not reported. 
Solvent removal 
The dispersed phase solvent needs to be removed for the condensation of solid 
particles.  Solvent removal is generally carried out by evaporation through the use of 
elevated temperatures, reduced pressure, or by solvent extraction.  Apparently, more than 
the method, the rate of solvent removal affects the performance of the particles produced.  
Fast solvent removal ensures high encapsulation efficiency by rapid solidification of the 
matrix material.  At the same time, fast solvent removal also results in production of 
porous particles thereby increasing the release rate.  The use of elevated temperatures is 
 31 
restricted by the thermo labile nature of a number of drugs and proteins. Moderate 
increase of temperature up to 40?C has been attempted by some researchers resulting in 
hollow microspheres with porous walls (Jayanthi, R., 1996).  Also, there are reports that 
when the solvent removal is achieved slowly at atmospheric temperature and pressure, 
there is a tendency to form a crystalline polymer matrix.  Solvent removal at reduced 
pressure produces amorphous polymer matrix (Isumikawa, S., 1991).  In an amorphous 
state, data indicate a molecular dispersion of the polymer and drug which results in the 
lowering of the release rate of the drug from the microspheres (Freitas, S., 2005).  Liquid-
liquid extraction is also a widely practiced technique.  It is reported that encapsulation 
efficiency can be significantly improved if a large excess of continuous phase than what 
is theoretically necessary to extract the dispersed phase is used (Tice, T. R., 1995).  The 
use of tenfold the amount of continuum phase necessary to extract all the dispersed phase 
solvent is suggested for encapsulating sparingly to freely soluble drug molecules. 
 
2A (iv). PRESENCE OF RESIDUAL COMPOUNDS AND RELATED ISSUES 
It is imperative that the drug delivery systems intended for the injectable route are 
nontoxic in addition to being sterile.  Even though some polymeric materials are known 
to be non toxic, biodegradable and biocompatible as required by the Food and Drug 
Administration (FDA), toxicity can be imparted into the system through the presence of 
trace amounts of residual surfactant, monomer, polymerization initiators and most 
importantly, residual solvents.  Even if the polymer used is not toxic, it can produce 
degradation by products that are toxic.  One such example is that of nanoparticles 
composed of poly(cyanoacrylate).  While poly(cyanoacrylate) itself is not toxic, one of its 
 32 
degradation by-products is formaldehyde (Gombutz, W,. R., 1995), which is a known 
carcinogen (NTP, 81-43, December, 1981). 
It has been shown that surfactants can be adsorbed on the surface of the nanoparticles.  
This adsorption can affect the pharmacokinetic distribution of the nanoparticles (Krueter, 
J., 1997).  As an example, surfactants adsorbed to the surface of butylcyanoacrylate 
nanoparticles containing the peptide dalargin have been shown to transport them across 
the blood-brain barrier (Krueter, J., 1997).  This also has the potential of changing the 
therapeutic efficacy of the delivery platform when the presence of residual surfactants on 
the surface of the particles is unintentional. On the other hand, residual monomers and 
oligomers present in the particles can act as catalyst in the chain scission reactions 
thereby causing the premature release of the drug (Niu, C. H., 1998).  In the same way, 
the presence of polymerization initiators, which are generally toxic, is a cause of concern.  
Since organic solvents of low polarity have to be used in the fabrication of the 
nanoparticles, the presence of residual solvents in the final product is a major concern.  
The most commonly used organic solvent, methylene chloride, is a carcinogen and has 
been shown to induce other forms of toxicities including neurotoxicity or teratogenicity, 
especially following inhalation (Kaushik, K., 2004).  When mice were treated with 1000 
ppm inhaled methylene chloride, increased DNA synthesis and increased cell 
proliferation in lungs was observed (Foster, J. R., 1994).  The USP limits the amount of 
methylene chloride in pharmaceuticals to 500 ppm.  The International Conference on 
Harmonization guidelines for permissible methylene chloride is 6 mg/day unless it can be 
shown that the residual solvent is released in a sustained fashion for several days.  The 
effective limit of methylene chloride is 6 mg/dose (Internal Conference on 
 33 
Harmonization, 1997).  Interestingly, with the conventional methods of particle 
preparation, the residual methylene chloride levels are 3,000-30,000 ppm (Courteille, J. 
P., 1994; Thoma, K., 1992).  The residual methylene chloride can however be reduced to 
below 25 ppm level in PLGA systems by treatment with supercritical carbon dioxide with 
a concomitant increase in particle size (about ten fold), decrease in bulk density (five 
fold) and increase in porosity (Kaushik, K.,  2004).  Even the presence of residual non 
toxic solvents can be a problem as they can possibly change the physico-chemical 
properties of the delivery platform.  High residual ethanol levels may reduce the glass 
transition temperature of PLGA-like polymers, causing the microspheres to aggregate 
under storage and may result in poor injectability (Sinha, V. R., 2003).  Consequently, 
careful design and selection of the manufacturing procedure, selection of the materials 
and solvents are essential for the success of the delivery platform. 
 
2B. PHYSIOLOGICAL RESPONSE TO PARTICULATE DDS 
Once the nanoparticulate drug delivery systems are introduced into the systemic 
circulation, two major events take place; (a) the particles are recognized by the cell 
mediated immune system of the body and promptly eliminated by phagocytosis, (b) the 
surviving particles are subsequently taken up by the cells by endocytosis in varying 
degrees depending on their size and surface properties.  Phagocytotic removal of the 
particulates is undesirable for drug delivery and/or targeting especially with respect to 
macromolecular drugs for a variety of reasons.   
Most proteins including the physiologically active ones have a very short biological 
half life.   Even if these proteins can be released slowly from the particulate DDS injected 
 34 
intramuscularly or subcutaneously, the problem of short half life of these molecules are 
not solved.  Additionally, the uptake of these released proteins into the cells, the most 
likely site of their action, remains a problem.  DNA or genes also are not taken up 
appreciably by cells. That is why, except for very exceptional cases, successful drug 
delivery systems for macromolecules require delivery of the bioactive molecules to their 
site of actions, i.e., into the cells.  For successful cellular delivery and targeting, the first 
and foremost condition to be met is that the delivery device has to remain in systemic 
circulation for a sufficiently long time.  Phagocytotic removal of the traditional drug 
delivery systems from systemic circulation is thus a serious limiting factor. 
The complex cascade of events following the introduction of foreign particles into the 
systemic circulation that leads to their rapid clearance from systemic circulation are better 
understood in recent years.  Conventional colloidal drug carriers are rapidly cleared from 
the blood stream by the mononuclear phagocyte system, mainly represented by the 
Kupffer cells of the liver and spleen macrophages (Stolnik, S., 1995).  This removal from 
the circulation generally occurs through specific recognition by cellular receptors specific 
for the plasma proteins bound to the carriers rather than the carriers themselves.  The 
deposition of the plasma proteins on the surface of the carriers is dependent on the 
surface charge, size, hydrophilicity/hydrophobicity and the conformation of the polymer 
chains among others.   
It is now widely accepted that the opsonization process starts with the activation of 
complement systems, most importantly the C3 system (Artursson, P., 1986; Kazatchkine, 
M. D., 1988; Brazile, D., 1995).  The activation of C3 complement results in proteolytic 
degradation of the C3 molecule producing two fragments C3b and iC3b (Kazatchkine,M. 
 35 
D., 1988; Brazile, D., 1995; Jensen, J., 1996).  C3 products can be generated by the 
classical or alternative pathways of complement activation, requiring Ca2+ and Mg2+ ions.  
The activation of alternative pathway is regulated by the nature of particle surface.  The 
cleavage of C3 by C3 convertase is induced by the functional groups at the surface of the 
foreign particles and results in the deposition of C3b onto them.  Also the surface of 
nanoparticles prepared from biodegradable polyesters, especially PLA, act as activators 
of the complement system (Mosqueira, V. C. F., 2001).  The macrophages are equipped 
with two receptors namely CR1 and CR3 which recognize the complement fragments 
C3b and iC3b.  Thus, following the complement activation, the complement fragments 
generated by the proteolysis are adsorbed onto the surface of the particles which are then 
phagocytosed by the macrophages through the complement fragment recognition 
receptors present on their surface (Mosqueira, V. C. F., 2001 & 1999).  Since the 
activation of the complement system by the alternative pathway is influenced by the 
nature of the particle surface, efforts are made to modify the particle surface in a manner 
that the complement activation, if any, is minimized (Gref, R., 2000; Vittaz, M., 1996).  It 
has been found that the plasma half life of PLA nanoparticles is two to three minutes.  
When they are coated with either human serum albumin or pluronic surfactant F68 
(polyoxyethylene-polyoxypropylene) does not increase the plasma half life.  But when 
polyethylene glycol (PEG) is used to modify the surface, the resulting PLA-PEG 
nanoparticles showed an increased half life of six hours (Verrecchia, T., 1995).  Further, 
it is found that the density of PEG chains grafted to the surface has a reciprocal 
relationship with the C3 cleavage.  The higher the PEG density, the lower is the C3 
cleavage.  Increasing the PEG chain length decreases the C3 cleavage even at a lower 
 36 
density (Mosqueira, V. C. F., 2001).  The longer chains, at the same surface density, were 
hypothesized to cover the surface better with the so called brush or mushroom 
configuration of the chains.  Overall, the hydrophobic surface, mostly dictated by the 
choice of the polymers, of the nanoparticles seems to be a major limitation towards 
development of long circulating particulate delivery systems.   
Endocytotic uptake of the particulates by the non-immune cells, on the other hand, 
can be helpful.  If sufficient control and selectivity can be achieved through suitable 
modifications of the surface of the particulates, endocytosis along with receptor mediated 
uptake can be an effective tool in delivering the active payload by the particulates into the 
target cell. 
It has been known for about one hundred years that intact micro and nanoparticles of 
natural origin can be absorbed across the gut wall at least to some extent.  Raw starch 
granules fed to rats have been known to be absorbed across intestinal mucosa for a long 
time (Verzar, F., 1936).  More recently, a wide variety of particles starting from titanium 
dioxide, to polyvinylpyrrolidine-coated silica (Florence, A. T., 1997), and polystyrene 
nanoparticles (Jani, P., 1989) have been shown in vivo experiments to translocate across 
the gastro-intestinal tract lining.  The extent of absorption following oral administration 
though is believed to be quite small (2-3% of the administered dose) (Florence, A. T., 
1997). 
Efforts have been made to characterize the cellular uptake of particles in vitro.  In 
most cases, cultured epithelial cells of various origin including Caco-2 representing 
intestinal mucosa, human umbilical vein endothelial cells (HUVEC), African green 
monkey kidney endothelial cells (COS-7) and human embryonic kidney epithelial cells 
 37 
(HEK-293) have been used.  Results of all these investigations have consistently 
suggested a few key features.  Firstly, rapid uptake of nanoparticles by endothelial cells 
has been demonstrated (Davda, J., 2002).  Further the uptake is found to be dependent on 
the concentration of the nanoparticles.  The particles are found to be localized into 
cytoplasm.  No adverse side effects have been detected in the cells as a result of the 
nanoparticle uptake, as shown by mitogenic assay 48 hours following the particle intake.  
Secondly, the uptake has been shown to depend on the temperature of incubation with 
relatively lower uptake in 4?C than in 37?C.  This suggests that the process is energy 
dependent and is mediated by endocytosis (Quaddoumi, M. G., 2000).  Further, it is 
found that the endocytotic uptake of the nanoparticles is size dependent (Prabha, S., 
2002; Desai, M.P., 1997).  The efficiency of cellular uptake in Caco-2 cells varied from 
41% for 100 nm particles to 15% for 1 micron particles and 6% for 10 micron particles.  
Similarly, when PLGA nanoparticles containing plasmid DNA encoding luciferase 
protein is used for gene transfection studies, smaller particles with mean diameter of 70 
nm showed 27 fold higher transfection than larger sized nanoparticles with mean 
diameter of 202 nm in the COS-7 cell line and a 4 fold higher transfection in the HEK-
293 cell lines (Prabha, S., 2002). 
In an effort to understand the process of internalization of the nanoparticles better, it 
has been hypothized that the internalization is a two step process.  The first step is the 
adsorption of the particles onto the plasma membrane of the cell surface.  The second 
step is endocytosis.  Both processes occur concomitantly at 37?C, whereas only the first 
step occurs at 4?C.  When anionic maghemite nanoparticles are incubated with human 
ovarian tumor cells (HeLa) and a cultured mouse macrophage cell line (RAW-264.7) at 
 38 
4?C for one hour (endocytosis inhibited), the adsorption of the nanoparticles on the outer 
cell surface is found to be saturable both in terms of time of incubation and extra cellular 
concentration of the nanoparticles (Wilhelm, C., 2003).  When, the same nanoparticle cell 
system is then incubated at 37?C for one hour, again the internalization process showed 
saturable kinetics for both the cell types in terms of time and extra cellular nanoparticle 
concentration.  In both cases, the quantifications are done by two different methods based 
on two different physical properties namely magnetophoresis (MP) and electron spin 
resonance (ESP).  Measurements from both the methods showed remarkable agreement.  
Visualization of the whole process has also been attempted in stages using a transmission 
electron microscope (TEM).  TEM pictures of the HeLa cells fixed after one hour 
incubation at 4?C (endocytosis inhibited), showed the adhesion of anionic nanoparticles 
on the plasma membrane mainly in the form of clusters probably through electrostatic 
interactions.  TEM photographs were also taken of the cells after the same incubation at 
4?C for an hour but followed by ten minutes and an hour of incubation at 37?C (restoring 
endocytotic activity).  After ten minutes of incubation at 37?C, TEM photographs showed 
early events of cell internalization of calthrin coated vesicles containing nanoparticles.  
Following an hour of endocytotic activity, densely confined nanoparticles into endocytic 
organelles with various morphological features and cytoplasmic localizations including 
late endosomes and lysosomes with micromeridic sizes could be identified.  This is by far 
the most comprehensive proof of endocytosis of nanoparticles by cells found by this 
author in contemporary literature.  Solid lipid nanoparticles containing spin labeled 
compounds have also been found to be taken up by leucocytes by endocytosis (Kristl, J., 
2003). 
 39 
Recently, efforts have been made in targeting nanoparticles to different target cells.  
A monoclonal antibody (MAb) with specificity for rabbit M cells was attached covalently 
to one micron polystyrene particles and investigated for cellular uptake.  The MAb 
indeed increased the particle count in rabbit Payer?s patches whereas a nonspecific MAb 
had no effect.  Similarly, tomato-lectin bound to microspheres induced a marked increase 
in systemic uptake, which could only be accounted for by induction of uptake through 
enterocytes as well as lymphoid tissue.  Tomato-lectin coupled microspheres showed 
unusual absorption tendencies.  Single dosing of animals produced no evidence of uptake, 
whereas five to ten days of daily dosing caused uptake of 18% of the dose (Florence, A. 
T., 1996), probably due to the induction of receptor expression by lectin.   Interestingly 
lectins from different sources are known to bind to specific cells in different regions of 
the gastro intestinal tract.  Though the last few examples show systemic absorption, they 
are sited because the systemic absorption of the particulates occurs in these cases through 
activation and targeting of a particular cell type in the gastro intestinal tract. 
 
3. CURRENTLY AVAILABLE PARTICULATE PLATFORMS 
A plethora of different particulate polymeric platforms are currently available for 
drug delivery purposes.  These platforms mainly differ in the physical and chemical 
properties of their constituent polymers and also the change in methodology dictated by 
the polymers.   
 
 40 
3A. HYDROPHOBIC SYSTEMS 
3A (i) BULK EROSION POLYMERIC SYSTEMS 
PLA, PLGA AND ASSOCIATED SYSTEMS  
The most successful and most widely investigated particulate platforms are made up 
of polymers of either lactic acid or both lactic acid and glycolic acid.  The main 
advantage of this class of polymers is their well known biodegradability and 
biocompatibility (Visscher, G. E., 1985, 1987; Yamaguchi, K., 1993).  Due to decades of 
use of these polymers as medical sutures, toxicological, chemical, and 
bio/histocompatability data are readily available. Other attractive features of this class of 
polymers include predictable biodegradation kinetics, commercial availability and most 
importantly the relative ease of regulatory approval (Lewis, D. H., 1990). 
Particles composed of PLGA and PLA undergo homogeneous bulk erosion on being 
exposed to water or physiological fluid (Gliding, D. K., 1981).  Initially there is random 
cleavage of hydrogen bonds due to hydration followed by cleavage of ester bonds leading 
to a decrease in molecular weight of the polymer and production of lactic and glycolic 
acids.  These are metabolized in the body through the Kreb?s cycle and eliminated as 
carbon dioxide and water. The carrier retains its original shape until significant 
degradation has occurred (~90%) (Sinha, V. R., 2003).  It dissolves completely once the 
molecular weight is reduced to the extent that it becomes soluble. The chemical 
composition and ratio of the monomers used in the polycondensation reaction strongly 
influence the degradation and thus the release kinetics as well.  The degradation kinetics 
is also influenced by polymer chain packing (crystallinity), hydrophobicity and 
 41 
sterioregularity.  Since some of these parameters like molecular weight, molar ratio 
between lactic and glycolic acid monomers, stereo regularity, and crystallinity can be 
easily controlled at least partially, resulting polymers with varying degradation time and 
release properties are readily available for use in the particulate DDS. 
While the obvious advantages of this class of polymers encouraged a spurt in research 
activity as sustained release particulate DDS, their success in terms of macromolecular 
delivery is rather limited.  Except Nutropin Depot which delivers growth hormone, all 
such products deliver small molecules including three leutenizing hormone releasing 
hormone analogs (decapeptide or less) and all are either intra-muscular or subcutaneous 
injections, or implantable rods (Tice, T, 2004).  A long circulating reservoir type 
particulate formulation for sustained release of the macromolecule in the systemic 
circulation is still not achievable with the current level of sophistication in the associated 
technology. 
PLA and PLGA particulate products suffer from several limitations.  Their 
hydrophobicity and solubility profiles dictate the choice of solvents ranging from acetone 
to dichloromethane.  The removal of organic solvents from PLA, PLGA and most other 
hydrophobic polymers is problematic at best.  With conventional methods of particle 
preparation, the residual organic solvent levels can be as high as 3,000-30,000 ppm while 
the USP limit is in the vicinity of 500 ppm.  The residual organic solvents can be 
removed by supercritical carbon dioxide processing, but the resulting particles become 
very porous with around a ten fold increase in size (Kaushik, K., 2004). Most biological 
macromolecules are not soluble in these solvents, rather they tend to precipitate and 
aggregate often irreversibly in their presence (Carpenter, J. F., 1989; Sah, H., 1999).  The 
 42 
incorporation of macromolecules into these particles is thus complicated.  It is generally 
done in the form of solid micronized macromolecule crystals suspended in the polymer 
solution or dissolving the macromolecule in the aqueous phase and emulsifying the 
aqueous phase in the polymer solution (Hermann, J., 1998).  Either of these methods 
results in an increase in particle size making them more vulnerable to phagocytosis.  The 
hydrophobic nature of these polymers often cause multilayer adsorption of the protein on 
the polymer surface (Horbett, T. A., 1987; Verrecchia, T., 1993) often resulting in 
incomplete release.  Also these polymers produce the monomeric acids as the particles 
degrade in the body, producing a very acidic microenvironment which is not conducive to 
the stability of a number of macromolecules.   
The most serious limitation of the PLA and PLGA systems is the activation of the cell 
mediated immune system caused by introduction into the systemic circulation.  The 
plasma half life of PLA and PLGA nanoparticles is around two to three minutes.  To 
minimize this problem a new class of polymers has been introduced: PLA-PEG and 
PLGA-PEG which are polyethylene glycol covalently attached to PLA or PLGA chains.  
In the particles prepared from these polymers through similar methodology, the 
hydrophilic PEG chain orients itself appropriately in aqueous suspension such that the 
more hydrophobic PLA and/or PLGA surface is shielded from the aqueous interface.  
With appropriate choice of molecular weight and PEG density on the surface it has been 
possible to reduce the immune system activation resulting in an increased plasma half life 
from two to three minutes to six hours (Mosquerra, V. C. F., 2001; Verrecchia, T., 1995).  
Though this is a significant improvement, clearly this is not adequate for systemic 
delivery.  Perhaps in recognition of this limitation and potential of the activation of the 
 43 
immune system, PLA and PLGA micro/nanosphere research is now focusing more and 
more on vaccine delivery (Jiang, W. L., 2005; Bramwell, V. W., 2005; Waeckerle-Men, 
Y., 2005). 
POLYACRYLATES 
Poly(acrylate) or more appropriately poly(alkylcyanoacrylate)s are another group of 
very well explored polymers for particulate drug delivery platforms.  The main attraction 
of poly(alkylcyanoacrylate)s is the ease of manufacturing of the particles by 
polymerization simultaneously as the particles are formed without the need of toxic 
initiation catalysts.  They undergo spontaneous polymerization at room temperature when 
the monomers are dissolved in water and a hydrophobic interface is provided in the form 
of an organic solvent or when oil is emulsified in the continuum phase.  As the 
polymerization progresses, the molecular weight increases and the polymer appears in the 
organic phase in the form of porous particles.   
The drug can be loaded by introducing the molecule in the continuous phase at the 
beginning or the end of the reaction (Couvreur, P., 1979; Bech, P. H., 1993, 1994).  The 
introduction of the active ingredient at the beginning of the reaction results in its 
incorporation in the core (Bech, P. H., 1993, 1994), while introduction of the active 
ingredient at the end of the reaction results in its adsorption at the surface (Losa, C., 
1991; Bech, P. H., 1993).  The whole incorporation process seems to be adsorption 
controlled.  A number of different molecules like the antibiotic gentamicin ((Zang, Q., 
1998), human immunodeficiency antiretroviral agent stavudine (Kuo, Y. C., 2005), 
analgesic peptide dalargin (Kreuter, J., 1997) and proteins like insulin (Zang, Q., 2001; 
 44 
Sullivan, C. O., 2004) have been successfully incorporated into poly(alkylcyanoacrylate) 
nanoparticles. 
Long term stability of polybutylcyanoacrylate nanoparticles has been studied in 
suspensions in various aqueous based media (Sommerfeld, P., 1997) including dilute 
hydrochloric acid solution, distilled water, phosphate buffered saline and human serum.  
It has been found that the particles tend to agglomerate during separation and recovery 
(Kreuter, J., 1983; Sommerfeld, P., 1997).  An increase in both polydispersity and 
particle size occurred following each step. This demonstrates that the final size of the 
particles is heavily dependent on the recovery process and reproducibility is difficult to 
achieve. Further, purified nanoparticles showed rapid degradation in aqueous suspension 
accompanied by particle size reduction and significant aggregation. The degradation 
process accelerated at higher temperature.  When suspended in phosphate buffered saline 
the nanoparticles aggregated rapidly with sedimentation.  However the particle 
suspension could be preserved in dilute (0.01N) hydrochloric acid solution for over two 
months.  Also, the particle suspension is found to be very stable with no aggregation in 
human serum for five days with significant aggregation appearing only after eight days. 
Polyalkylcyanoacrylates are not toxic as a class of polymers.  However they show 
significant toxicity in vivo because of the toxic degradation by products.  Acrylic acid 
polymers are bulk erosion polymers undergoing random chain scission until they become 
soluble with the formation of the toxic by product formaldehyde.  This reaction is 
strongly dependent on the polymer molecular weight and the pH of the medium 
(Leonard, F., 1966).  The alkylcyanoacrylates also undergo the same reaction along with 
a parallel side chain scission by the blood esterases producing the corresponding alkanol 
 45 
(Lenaerts, V., 1994).     Comparative in vitro cytotoxicity studies have been carried out to 
assess the relative toxicity of different polymers including polyalkylcyanoacrylates and 
PLGA (Zange, R., 1997).  The conclusion of the study is that all alkylcyanoacrylates are 
more toxic than PLGA and generally the toxicity increases with the decrease of the size 
of the alkyl side chain.  The polymer concentration level at which the cytotoxicity is 
observed gives the following increasing order of toxicity starting from the lowest. 
PLGA > BCA > PVC > HCA > ECA > MCA 
In addition to toxicity, the polyalkylcyanoacrylate nanoparticles suffer from the same 
disadvantage as the PLGA nanoparticles in that they are instantaneously detected by the 
reticuloendothelial system of the host body and are promptly eliminated from the blood 
stream (Grislain, L., 1983; Lenaerts, V., 1984; Gibaud, S., 1996).  They accumulate in the 
liver, spleen and bone marrow.  As a consequence these nanoparticles are unsuitable as 
prolonged delivery platform to be used in systemic circulation.  
 
3A (ii) SURFACE EROSION SYSTEMS 
POLYANHYDRIDES 
Polyanhydrides represent a new class of biodegradable polymers meeting some of the 
essential requirements of a sustained release particulate drug delivery system.  They are 
composed of highly water-labile anhydride bonds joining dicarboxylic acids of different 
molecular structures and weights in series to make up the polymer molecule.  The 
polyanhydrides that are important from drug delivery purposes are all made up of bulky 
or hydrophobic dicarboxylic acids.  The hydrophobicity of the monomeric dicarboxylic 
acids decreases the rate of penetration of water into the polymeric device.   Since the 
 46 
anhydride bonds are highly water-labile, on being exposed to water the polymeric 
particles erode from the surface (outside in) as water slowly penetrates into them.  A 
continuous decrease in device thickness throughout the erosion process, maintenance of 
structural integrity and nearly zero order degradation kinetics suggest dominancy of 
heterogeneous surface erosion (Angelova, N., 1999).  As the polyanhydrides can be 
synthesized from a wide range of monomers with varying polarity, their degradation rate 
can be controlled over a wide period of time with over several thousand fold difference in 
rate (Domb, A. J., 1987, 1989).  The aliphatic polyanhydrides erode very fast and thus are 
not very useful for pharmaceutical applications.  The aromatic polyanhydrides degrade 
extremely slowly (Domb, A. J., 1992).  Thus the aromatic and aliphatic copolymers are 
best suited for drug delivery applications with optimum degradation profiles and 
mechanical properties.  The hydrolysis of the anhydride bonds are base catalyzed and the 
degradation of the polymer is dependent on pH and solubility of the degradation products 
in the surrounding media.  Thus the degradation is slower in acidic pH at which the 
monomers are unionized and thus have limited solubility (Jain, J. P., 2005). 
Toxicity associated with polyanhydrides has been investigated in terms of 
cytotoxicity, mutagenicity and local irritation.  For cytotoxicity and mutagenicity studies, 
cultured bovine endothelial and bovine smooth muscle cell models, which are very 
sensitive to the changes in growth media and substrate (Leong, K. W., 1986), have been 
used.  The study showed an absence of acute toxicity of these polymers or their 
degradation products to sensitive mammalian cells.  Since carcinogenesis is usually 
preceded by mutagenesis; mutagenicity testing has been used for rapid screening of 
neoplastic transformation.  The in vitro results show that polyanhydrides, particularly the 
 47 
more popular poly[1,3-bis-(p-carboxy)-poly[1,3-bis(p-carboxyphenoxy)methane]-co-
sebasic anhydride]  [P(CPP-SA)] are non cytotoxic and non mutagenic but have a very 
low teratogenic potential (Leong, K. W., 1986). In vivo studies with poly[1,3-bis?
(carboxyphenoxy)propane] and poly(terephthalic -sebasic anhydride) by implantation of 
polymer samples into the cornea of rabbits has shown no indications of local irritation 
like loss of corneal clarity, inflammation or proliferation of new blood vessels (Leong, K. 
W., 1986) for six weeks.  Similar experiments involving subcutaneous insertion of 
polymer disks in rat models for eight weeks failed to detect any systemic toxicity 
(Laurencin, C., 1990).  It is to be noted that no long term carcinogenicity studies on 
polyanhydrides has been reported to date (Jain, J. P., 2005).   
Polyanhydrides are primarily developed for injection molding, compression molding 
and hot-melt extrusion.  The last two processes require elevated temperature and are thus 
not well suited for thermo labile compounds like proteins (Leong, K., 1986).  
Polyanhydrides had also been reportedly used to develop particulate delivery system 
containing proteins like lysozyme, trypsin, heparinase, immunoglobulin (Tabata, Y., 
1993) and bovine serum albumin (Determan, A. S., 2004) using double emulsion 
technique.  They have been more extensively studied for various applications as localized 
delivery carriers (Jain, J. P., 2005) for various disease conditions including cancer 
(Teomim, A., 1999; Krasco, M. Y., 2003; Li, Y., 2004), glaucoma (Uppal, P., 1994), and 
Alzheimer?s disease (Howard III, M. A., 1989).  Recently, one of the polymers in this 
class, namely [P(CPP-SA)],  has been approved by the FDA for local delivery in the 
treatment of brain cancer (Wang, P. P., 2002).  This is one of the few examples where a 
biodegradable synthetic polymeric implant has been approved for human use by the 
 48 
FDA.  Generally speaking, most of the drug delivery research is specifically focused on 
local delivery matrices like implants, films, and surgical pastes mostly involving highly 
hydrophobic chemotherapeutic agents and other hydrophobic drugs. 
In an overall analysis, polyanhydrides do not seem to be drastically different from the 
more established platforms like PLGA.  Being hydrophobic polymers, particles have to 
be formulated with hydrophobic organic solvents in the internal phase.  So the problems 
like macromolecule denaturation, presence of residual organic solvents and aggregation 
associated with the nature of the polymers and size associated with the necessary double 
emulsion technique still remain.  Additionally and most importantly, polyanhydrides 
being a more hydrophobic class of polymers are expected to activate the cell mediated 
immune system more strongly and thus a free circulating particulate injectable 
formulation does not seem to be a possibility.  No study showing phagocytosis of 
polyanhydrides, or the lack of it, could be found in literature.  Probably because of these 
factors the focus of research involving this platform is more on local delivery of 
hydrophobic small molecules like chemotherapeutic agents.  Lastly, long term safety of 
the polyanhydrides is still to be established. 
3B. HYDROPHILIC SYSTEMS 
A wide variety of hydrophilic polymers have been investigated for drug delivery 
purposes.  Synthetic polymers like poly(ethylene oxide) and poly(propylene oxide) and 
their copolymers, polyvinyl alcohol, polyvinyl pyrrolidone as well as natural polymers 
like starch, dextran, polyanionic carbohydrates like alginic acid, polycationic 
carbohydrates like chitosan and proteins like albumin, gelatin, and collagen have all been 
tried for this purpose and have shown some promise for restricted applications.  One 
 49 
common advantage of the hydrophilic polymer based platforms is that they are all 
aqueous based and thus represent a much more benign microenvironment for the 
bioactive macromolecules like proteins.  Secondly, being hydrophilic they show a much 
reduced complement activation and the resulting immune recognition.  Some of the more 
successful and promising platforms will be considered in more detail here. 
ALGINATES 
Alginates are linear polysaccharides containing varying amounts of 1, 4'-linked ?-D-
mannuronic acid and ?-L-guluronic acid residues.  These two acids are arranged in 
random domains of homopolymer blocks of one kind and the domains do not follow any 
particular pattern.  The exact proportion of guluronic acid and mannuronic acid residues 
vary in alginates according to its natural source with some alginates rich in guluronic 
acids while others rich in mannuronic acid (Huang, A., 1962).  The molecular weight of 
the alginates vary widely depending on its source from 80-290 kilo Daltons and the 
viscosity of alginate solution increases with increasing molecular weight (Berth, G., 
1992).  Two special properties sustained the interest in alginates for drug delivery; poly 
ion complexation and resulting gelation of alginates when in contact with polyvalent 
metal ions or cationic polymers in solution and the mucoadhesive nature of the polymer.  
The mucoadhesion of alginates to living intestinal epithelial lining of rat has been 
investigated and tensile strength quantified (Chickering, D., 1992, 1995).  It is found that 
alginate has the highest mucoadhesive strength when compared to polymers such as 
polystyrene, chitosan, carboxymethylcellulose and poly(lactic acid).   
 50 
A variety of different proteins have been incorporated into alginate micro/nano 
particles like antigens (Bowersock, T. L., 1996). Most of the work, though, seems to be 
directed towards local delivery of growth factors like fibroblast growth factor (Edleman, 
E. R., 1993) for tissue repair, nerve growth factor (Maysinger, D., 1992) for treatment of 
age dependent neuronal atrophy of fore brain and anticancer proteins like tumor necrosis 
factor receptor and interleukin receptors like interleukin-17, 1, 4 receptors (Gombotz, W. 
R., 1998).  In addition, alginate gelation has been successfully used to encapsulate and 
preserve live cells like chromaffin cells (Tsang, P. W., 1996), and islet cells (Lim, F., 
1980), etc., with the overall goal of hormone or neurotransmitter replacement therapy 
through cell transplantation.  It is observed that basic proteins show more prolonged 
release patterns than acidic proteins when encapsulated in alginate particles due to 
obvious ionic interactions between the polymer and the protein in the first case, and the 
lack of it in the later.  Also, basic proteins, like basic fibroblast growth factor is found to 
be largely deactivated due to the ionic interactions with alginates (Mumper, R. J., 1994).   
Alginate micro/nano particles are prepared by poly ion complexation of  sodium 
alginate solution with divalent metal ions like Ca2+, Ba2+, Sr2+ ( Gombotz, W. R., 1998) 
and Al3+ (Al-Musa, 1999).  Monovalent cations and Mg2+ do not induce gelation.  The 
use of other divalent cations such as Pb2+, Cu2+, Cd2+, Co2+, Zn2+, Ni2+ and Mn2+ causes 
gelation but their use is limited by their potential toxicity.  The gelation and crosslinking 
mainly occur by the exchange of sodium ions from guluronic acid residues with the 
divalent cations and the stacking of these guluronic acid groups of two different 
polymeric chains together in a highly cooperative manner.  The size of the cooperative 
units is more than twenty monomer units (Smidsrod, O., 1990).   It is known that the 
 51 
crosslinking reaction is fast and the reaction is complete within thirty minutes. The 
tightness of crosslinking and thickness of the crosslinking layer is dependent on the ionic 
radius of the polyvalent metal ion (Al-Musa, 1999).  Calcium and barium ions having 
greater atomic diameters of 0.95 ? and 1.35 ?, respectively, produce looser crosslinking 
of lesser thickness.  Aluminum ion with an atomic radius of only 0.5? produces tighter 
crosslinking throughout the matrix.  The resultant in vitro release profile also verifies 
these observations by showing varying rates of diffusional release.  
Since the alginate particulate platform is held together by ionic interactions, change in 
ionic environment and pH affects its stability and structural integrity.  The alginate gel 
structure breaks down by removal of polyvalent metal ions from the environment using 
chelating agents.  High concentrations of anions like lactate, citrate and phosphate break 
down the gel structure.  A high concentration of cations like Na+ and Mg2+ in the medium 
also cause a breakdown of the gel structure due to formation of more preferred 
complexation species and breakdown of metal ion-guluronic acid complex (Sutherland, I. 
W., 1991).  Alginates are protonated in acidic medium producing alginic acid and 
breaking down the gel structure (Al-Musa, 1999).  The alginates are also hydrolyzed in 
acidic media leading to a decrease in molecular weight and faster release of the 
encapsulated molecules (Huang, A., 1963).  This acid catalyzed degradation and 
breakdown of poly ion complexation limits the use of monolithic alginate particles for 
oral delivery of drug molecules.  The mucoadhesive properties of alginic acid can be 
utilized if the alginic acid is used in particle formation along with other biodegradable 
polymers. 
 52 
Though the alginates are found to be largely biocompatible, they do produce an 
immunogenic response.  There are numerous reports of alginates producing fibrotic 
reaction (Soon-Sheong, P., 1991; Otterlei, M., 1991; Spargo, B. J., 1994).  The 
immunogenic reactions of alginates seem to be related to its chemical composition and 
mitogenic contaminants (Zimmerman, U., 1995).  Though all kind of alginates activate 
cytokine production upon the introduction of alginate particles into the blood stream, 
alginates rich in mannuronic acid activate cytokine production more than alginates rich in 
guluronic acid (Soon-Sheong, P., 1991; Otterlei, M., 1991; Spargo, B. J., 1994).  Irregular 
surfaces of the particles also contribute to the fibrotic reaction of the alginates (Vos, P.D., 
1996).   
Overall, the alginates appear to be a promising platform for sustained release of 
macromolecules because of the benign microenvironment and relative biocompatibility.  
In vitro studies also show sustained release for large macromolecules. The immunogenic 
reactions caused by the alginates remain a concern and the phagocytotic reaction to 
alginates has not been investigated in depth. Also, alginate hydrogel stability in vivo is 
suspect due to the high ionic strength, presence of enzymes and presence of specific ionic 
species known to affect the stability of the hydrogel structure.  No sustained release 
alginate product has been approved so far after more than two decades of research in this 
field.   
GELATIN 
Gelatin is a natural polymer derived from collagen and is commonly used in 
pharmaceutical and medical applications due to its biodegradability (Balakrishanan, B., 
2005; Yamamoto, M., 2001) and biocompatibility in the physiological environment (Yao, 
 53 
C.H., 2004).  Gelatin has a proven record of safety as a plasma expander and as an 
ingredient in drug formulations and as a sealant for vascular prosthesis (Kuijpers, A. J., 
2000).  Collagen can be pretreated (before extraction) with acid or alkali to produce basic 
gelatin with an isoelectric point (IEP) of 9 or acidic gelatin with IEP of 5 (Ikada, Y., 
1998).  The difference in IEP is because of the fact that acid treated gelatin possesses 
more amino groups than alkali treated gelatin which possesses more carboxylic acid 
groups.  As a result gelatin is a good polymer for polyion complexation either between 
two of its own species to produce a electrovalently bonded structure or with 
macromolecules of different overall charge to increase encapsulation efficiency and 
decrease release rate. 
Because of the good gelation properties, gelatin has been extensively investigated for 
local sustained release applications in the form of block hydrogels and films.  Although 
polyion complexation is important in the context of biomolecule loading and retention, 
the key determinant of controlled release from these constructs is the rate of degradation 
of gelatin carriers (Yang, S., 2005).  Given that gelatin gel undergoes rapid degradation in 
vivo whereas chemically crosslinked gelatin gels are relatively stable, most gelatin-based 
platforms relevant to sustained biomolecule delivery use chemical crosslinking.  In vivo 
models have been used to assess the effect of crosslinking density on the rate of gelatin 
degradation (Tabata, Y., 1999).  Gelatin hydrogels are chemically crosslinked by using 
aldehydes like formaldehyde, glyceraldehyde and gluteraldehyde, polyepoxides and 
isocyanates (Kuijpers, A. J., 2000).  The carboxylic and amino groups of gelatin can also 
be directly crosslinked by using chemical activators like acyl azides and carbodiimides 
(Tabata, Y., 1994).  As expected, crosslinking density in gelatin gels is found to be 
 54 
inversely proportional to hydrated water content of the gel and as also the rate of gel 
degradation in vivo (Tabata, Y., 1999). 
Different growth factors have been incorporated into gelatin hydrogels for hard tissue 
regeneration (Yamamato, M., 2003), soft tissue regeneration (Kimura, Y., 2003) and 
therapeutic angiogenesis (Thomson, J. A., 1988) with pharmacodynamic improvement 
noted in each case for a prolonged period of time.  Gelatin films have been used for local 
delivery of insulin from open wounds (Shinde, B. G., 1992) and also in local delivery of 
basic fibroblast growth factor to the heart following artificial acute myocardial infraction 
in experimental  animal models for neovascularization in the heart (Ueyama, K., 2004).  
Crosslinked gelatin microparticles have also been investigated for sustained local 
delivery of mainly growth factors (Kimura, Y., 2003; Shakakibara, K., 2003).  These 
investigations showed marked pharmacodynamic effect in both cases. 
The most important drawback of gelatin based drug delivery platform is that polyion 
complexation by gelatin is not enough to produce significant sustained release by itself.   
The crosslinking techniques used to slow down gelatin degradation are too generic and 
result in crosslinking in the bulk of the polymer.  Such crosslinking in the bulk can react 
with the bioactive macromolecule resulting in deactivation (Gombotz, W. R., 1995).  
Additionally, gelatin particles have not been reported in the submicron range in size 
thereby making it unsuitable as a systemic delivery vehicle.  Indeed not a single 
investigation has been found aimed at using the gelatin particulate platform for long 
circulating systemic and sustained release.  Though gelatin is a hydrophilic polymer and 
has been used as plasma expander, it does produce activation of RES in particulate form.  
Gelatin is known to adsorb fibronectin, a blood component that provides particle 
 55 
recognition (Tabata, Y., 1989).  In totality, gelatin though largely biocompatible and 
which can effect polyion complexation, is not very useful as a particulate delivery 
platform for sustained release systemic delivery of bioactive macromolecules. 
STARCH, DEXTRAN AND ASSOCIATED SYSTEMS 
Starch, dextran and associated natural polymers and their derivatives have been 
investigated as a sustained release particulate delivery platform for bioactive 
macromolecules for a long time.  The use of biodegradable hydrogels for delivery of 
bioactive macromolecules like proteins is of particular interest due to their 
biocompatibility and their relative inertness towards protein drugs (Park, K., 1993), 
(Domb, A., 1990).   Hydrogels are the only class of polymeric systems that allow proteins 
to permeate through the continuum of the carrier (Gombotz, W.R., 1995).  The hydrogels 
rapidly swell with exposure to water and can rapidly decrease in mechanical strength, and 
often integrity, if they are not reinforced with chemical crosslinking to control swelling.  
The initial rate of release of macromolecules from hydrogel systems are diffusion 
controlled through the aqueous channels of the gel structure.  Generally the rate is 
inversely proportional to the molecular weight of the encapsulated molecule and directly 
proportional to the equilibrium water content characteristic of the gel.   
Crosslinked starch (Degling, L., 1993), dextran (Kim, S. H., 1999), amylopectin 
(Michailova, V., 2001) and cellulose derivative (Bech, S. L., 1990) based hydrogels have 
all been investigated for this purpose.   It has been generally found that carbohydrate 
based hydrogel systems release drug molecules in a few hours (Michailova, V., 2001).  
The drug release roughly corresponds to the time that is required for the hydrogel system 
to attain equilibrium swelling.  In an effort to reinforce the hydrogel structure a wide 
 56 
variety of chemical crosslinking methods have been tried.  Polyfunctional aliphatic acids 
(Sediel, C., 2001), epichlorohydrin, phosphorous oxychloride, methylenebisacrylamide 
(Denizli, B.K., 2004) have all been employed to modify starch and dextran to improve 
the mechanical properties and decrease swelling of the resulting hydrogels.  
Epichlorohydrin crosslinked dextran hydrogel shows the minimum equilibrium swelling 
of 900% while phosphorous oxychloride crosslinked dextran shows the maximum 
equilibrium swelling of 5450%.  Dextran has also been derivatized with the sodium salt 
of acrylic acid to produce dextran acrylate.  The acrylated dextran gel was then photo 
crosslinked by long wave UV irradiation. The resulting product showed relatively lower 
equilibrium swelling of 171% at pH 7 (Kim, S. H., 1999).  Polyacryl starch (Artursson, 
P., 1984) and polyacryl dextran (Edman, P., 1980) microspheres have been investigated 
for immobilization and delivery of several proteins  and also liposomes (Stenekes, R.J.H., 
2000).   
Though these systems have shown some promise as far back as the early 1980s, the 
problems associated with rapid swelling, fast release of the contents or incomplete release 
of proteins could not be solved satisfactorily until now.  Additionally, all these systems 
suffer from the basic disadvantage in that they are crosslinking in the bulk of the polymer 
matrix.  This kind of bulk crosslinking can denature the incorporated macromolecule or 
covalently incorporate the macromolecule in the polymer matrix and thus make the 
macromolecule unavailable for release. 
 
 
 
 57 
4.  CONCLUSION: NEED FOR A NEW APPROACH 
Certain general conclusions can be made from this review of theavailable literature.  
To be effective and clinically useful, long circulating, sustained release, particulate 
delivery systems for bioactive molecules, especially macromolecules, have to fulfill 
certain essential criteria.  In addition to being biodegradable and biocompatible and non 
toxic, they have to be able to protect the stability of the incorporated bioactive 
macromolecule throughout storage and during the actual use.  In this respect, processing 
parameters like the use of organic solvents, and the presence of trace amount of residual 
solvents and their effects on the incorporated macromolecule, is as important as the safety 
and biocompatibility of the polymer used.  The systems must not show irreversible 
interactions like adsorption to the incorporated bioactive molecules, especially proteins.  
The system should be transparent to the immune system and not produce appreciable 
complement activation.   
It is recognized that no current platform completely fulfills all the necessary 
criteria.  The more successful hydrophobic platforms like PLGA and associated systems 
have the advantage of better control of the release rate of the incorporated molecules.  
But they suffer from the disadvantage of greatly activating the immune system and thus 
are not suitable for systemically circulating particles.  Interestingly, not only PLGA, but 
all such hydrophobic platforms show a high degree of immune response and complement 
activation. 
The hydrophilic hydrogel systems, especially the starch derivatives and dextrans 
have the advantage of being biodegradable, biocompatible and also largely transparent to 
the reticuloendothelial systems.  They suffer the disadvantage of not being able to sustain 
 58 
and control release of incorporated small molecules.  When crosslinked in the bulk, they 
show more promise in sustaining the release of incorporated bioactive molecules, but 
bulk crosslinking raises the concern of denaturation and deactivation of delicate 
macromolecules like proteins.  Additionally, this kind of bulk crosslinking may actually 
incorporate the bioactive molecule to the polymer thus making them unavailable for 
release. 
Consequently the best approach seems to be to attempt to achieve the best 
properties of both types of systems.  The inherent advantages of the hydrogel systems in 
terms of hydrophilicity and thereby lack of immune activation can be used as an 
advantage. An attempt to control the diffusional release of the incorporated bioactive 
molecule by formation of a semi permeable rate limiting membrane on the surface of the 
particle by a suitable interfacial crosslinking reaction may have significant potential.  The 
advantage of an interfacial crosslinking reaction over a bulk crosslinking reaction is that 
the former is expected to leave the incorporated bioactive molecule unaffected within a 
benign hydrophilic core. 
 
REFERENCES 
Aftabouchad, C. and Doelker, E.  1992.  Methodes de preparation des microparticles 
biodegradables chargees en principes actifs hydrosolubles.  S.T.P. Pharma 
Sciences, 2, 365-380. 
 
Al-Azzam, W., Pastrana, E. A., Griebenow, K. 2002.  Co-lyophilization of bovine serum 
albumin (BSA) with poly (ethylene glycol) improves efficiency of BSA 
encapsulation and stability in polyester microspheres by a solid-in-oil-in-oil 
technique.  Biotechnology Lettters, 24,  1367-1374. 
 
 59 
Al-Musa, S., Abu, Fara, D. and Badwan, A.A.   Evaluation of parameters involved in 
preparation and release of drug loaded in crosslinked matrices of alginate. Journal 
of Controlled Release, 57, 223-232. 
 
Amsden, B. G. and Liggins, R. T.  2001.  Methods for Microsphere production.  U. S. 
Patient, 6,224,794. 
 
Angelova, N. and Hunkeler, D.  1999.  Rationalizing the design of polymeric 
biomaterials.  Trends in Biotechnology, 17, 411-421. 
 
Artursson, P. and Sjoholm, I.  1986.  Effect of opsonins on the macrophage uptake of 
polyacryl starch microparticles.  International  Journal of Pharmaceutics, 32, 
165-170. 
 
Baker, R.  (1987). Controlled Release Of  Biologically Active Agents.  John Walley & 
Sons, New York, Pp. 81-97. 
 
Balakrishnan, B. and Jayakrishnan, A.  2005.  Self crosslinking biopolymers as injectable 
in situ scaffolds.  Biomaterials, 26, 3941-3951. 
 
Baush, H.  1994.  Silicone toxicology.   Seminars in Arthritis Rheumatism, 24,  11-17. 
 
Bech, P. H., Keruter, J., Reszka, R. and Fichtner, I.  1993.  Influence of 
polybutylcyanoacrylate nanoparticles and liposomes on the efficacy and toxicity 
of anticancer drug mitoxanthrone in murine tumour models.  Journal of  
Microencapsulation,10, 101-114. 
 
Bech, P. H., Kreuter, J., Muller, E. G. and Schatton, W.  1994.  Improved peroral delivery 
of avarol with polybutylcyanoacrylate nanoparticles.  European Journal of 
Pharmaceutics and Biopharmaceutics, 40, 134-137. 
 
Bech, S.L., Chen, T. L., Mikhalauski, P. and Amman, A.  1990.  Recombinant human 
transforming growth factor- beta1 (rhTGF-?1) enhances healing and strength of 
granulation skin wounds.  Growth Factors, 3, 267-275. 
 
Berth, G.  1992.  Methodical aspects of characterization of alginate and pectate by light 
scattering and viscometry coupled with GCP.  Carbohydrate Polymers, 19, 1-9.  
 
Bowersock, T. L., HogenEsch, H, Suckw, M., Porter, R.E., Jackson, R. and Park, K.  
1996.  Oral vaccination withalginate microsphere systems.  Journal of Controlled 
Release, 39, 209-220. 
 
Brazile, D., Prud?Homme, C., Bassoulet, M-T, Marlard, M., Spenlehauer, G. and 
Veillard, M.  1995.  Stealth Me. PEG-PLA nanoparticles avoid uptake by 
 60 
mononuclear phagocytes system.  Journal of Pharmaceutical Sciences, 84, 493-
498. 
 
Camble, R., Timms, D. and Wilkinson, A.J.  1994.  Continuous release of pharmaceutical 
compositions.  U. S. Patent 5,320,840. 
 
 Capan, Y., Woo, B. H., Gebrekidan, S., Ahmed, S. and DeLuca, P. P.  1999.  Influence 
of formulation parameters on the characterestics of poly(D,L-lactide-co-glycolide) 
microspheres containing poly(L-lysine) complexed plasmid DNA.  Journal of 
Controlled Release,  60,  279-286. 
 
Carpenter, J. F. and Crwe, J. H.  1989.  Comparison of Solute-Induced Protein 
Stabilization in Aqueous Solution and in the Frozen and Dried state.  Journal of 
Dairy  Science, 73, 3627-3636. 
 
Carrio, A., Schwach, G., Coudane, J. and Vert, M.  1995.  Preperation and degradation of 
surfactant-free PLGA microspheres.  Journal of Controlled Release, 37,   113-
121. 
 
Chickering, D. and Mathiowitz, E.  1995.  Bioadhesive microspheres:  I. A novel 
electrobalence-based method to study adhesive interactions between individual 
microsphere and intestinal mucosa.  Journal of Controlled Release, 19, 88-89. 
 
Chickering, D., Jacob, P. and Mathiowitz, E.  1992.  A tensile technique to evaluate the 
interaction of bioactive microspheres with intestinal mucosa.  Proceeds of 
International  Symposium Controlled Release of Bioactive Materials, 19, 34, 251-
261. 
 
Cohn, D. and Younes. H.  1988.  Biodegradable PEO/PLA block copolymers.  Journal of 
Biomedical Material Research,  22,  993-1009. 
 
Colton, C.K., Smith, K. A., Merril, E. W. and Ferrel, P. C.  1971.  Permeability studies 
with cellulose membranes.  Journal of Biomedical Material Research, 5,  459-
488. 
 
Constantino, H. R., Langer, R. and Kilbanov, A.M.  1994.  Moisture induced aggregation 
of lyophilized insulin.  Pharmaceutical Research, 11,  21-29. 
 
Courteille, F., Benoit, J. P. and Thies, C.  1994.  The morphology of progesterone loaded 
microspheres.  Journal of Controlled Release, 30, 17-26. 
 
Couvreur, P., Kante, B., Roland, M. and Speiser, P.  1979. Adsorption of antineoplastic 
drugs to polyalkylcyanoacrylate nanoparticles and their release in calf serum.  
Journal of Pharmaceutical Sciences, 68, 1521-1523. 
 
 61 
Crotts, G. and Park, T. G., 1995.  Preperation of porous and nonporous biodegradable 
polymeric hollow microspheres.  Journal of  Controlled Release, 35, 91-105.  
 
Davda, J. and Labhasetwar, V.  2002.  Characterization of nanoparticle uptake by 
endothelial cells.  International Journal of Pharmaceutics, 233, 51-59. 
 
Davda, J., labhasetwar, V.  (2002). Characterization of nanoparticle uptake by endothelial 
cells.  International Journal of Pharmaceutics, 233, 51-59. 
 
De Lustro, F.,  Condell, R.A., Nguen, M. A. and McPherson, J. M.  1986.  A comparative 
study of biologic and immunologic response to medical devices derived from 
dermal collagen Journal of Biomedical Material Research, 20,  109-120. 
 
De Lustro, F.,  Dasch, J.,  Keefe, J. and Ellingsworth, L.  1990.  Immune responses to 
allergenic and xenogeneic implants of collagen and collagen derivatives.  Clinical 
Orthopaedics, 260, 263-279. 
 
Degling, L., Stjarnkvist, P. and Sjoholm, I.  1993.  Interferon-alpha in starch 
microparticles; Nitric oxide-generating activity in vitro and antileshmanial effect 
in mice.  Pharmaceutical Research, 6, 783-790.  
 
Delgado, C., Francis, G.E. and Fisher, D.  1992.  The uses and properties of PEG linked 
proteins.  Critical review of Therapeutic Drug Carrier Systems, 9, 249-304. 
 
Denizli, B. K., Can, H. K., Rzaev, Z.M.O. and Guner, A.  2004.  Preparation conditions 
and swelling equilibria of dextran hydrogels prepared by some crosslinking 
agents.  Polymers, 45, 6431-6435. 
 
Desai, M. P., Labhasetwar, V., Walker, E., Levy, R. J., Amidon, G. L.  (1997). The 
Mechanism of Uptake of Biodegradable Microparticles in Caco-2 Cells Is Size 
Dependent.  Pharmaceutical Research, 14, (11), 1568-1573. 
 
Determan, A. S., Trewyn, B. G. Narasimhan, B.  2004.  Encapsulation, stabilization, and 
release of BSA-FITC from polyanhydride microspheres.  Journal of Controlled 
Release, 100, 97-109. 
 
Domb, A. J.  1992.  Synthesis and characterization of biodegradable aromatic anhydride 
copolymers.  Macromolecules, 25, 12-17. 
 
Domb, A., Davidson, G. W. and Sanders, L. M. 1990.  Diffusion of peptides through 
hydrogel membranes.  Journal of Controlled Release, 14, 133-144. 
 
Domb, A.J. and Langer, R.  1987.  Polyanhydrides. I.  Preparation of high molecular 
weight polyanhydrides.  Journal of Polymer science Part A: Polymer 
Chemchemistry, 25, 3373-3386. 
 62 
 
Domb, A.J. and Langer, R.  1989.  Poly(anhydrides). 3. Poly(anhydrides) based on 
aliphatic-aromatic diacids.  Macromolecules, 22, 3200-3204. 
 
Edelman, E.R., Nugent, M.N. and Karnovosky, M. J.  1993.  Perivascular and 
intravenous administration of basic fibroblast growth factor: Vascular and solid 
organ deposition.  Proceedings of National Academy of Science, USA, 90, 1513-
1517. 
Falk, R. F. and Randolf, T.W., 1998.  Process variable implications for residual solvent 
removal and polymer morphology in the formation of gentamycin-loaded poly (L-
lactide) microparticles,  Pharmaceutical Research, 15, 1233-1237. 
 
Ferrai, S., Geddes, D. M., Alton, W. F. W. E., (2002).  Barriers and new approaches to 
gene delivery in cystic fibrosis.  Advanced Drug Delivery Review, 54, 1373-1393. 
 
Florence, A. T.  1997.  The Oral Absorption of Micro- and Nanoparticulates: Neither 
Exceptional Nor Unusual.  Pharmaceutical research, 14, 1997-2007. 
 
Foster, J. R., Green, T., Smith, L. L., Tittensor, S. and Wyatt, I.  1994.  Methylene 
chloride: an inhalation study to investigate toxicity in the mouse lung using 
morphological, biochemical and Clara cell culture techniques.  Toxicology, 91, 
221-234. 
 
Freitas, S., Merkle, H. P. and Gander, B.  2005.  Microencapsulation by solvent 
extraction/evaporatuion: reviewing the state of the art of microsphere preparation 
process technology.  Journal of Controlled Release, 102,  313-332. 
 
Gabor, F., Ertl, B., Wirth, M. and Mallinger, R.  1999.  Ketoprofen- poly(DL-lactide-co-
glycolic acid)  microspheres: influence of manufacturing parameters and type of 
polymer on the release characteristics.   Journal of Microencapsulation,  16,  1-
12. 
 
Gao,S., Chen, J., Xu, X., Ding, Z., Yang, Y. H., Hua, Z., Zhang, Z. (2003).  Glycosylated 
low molecular weight chitosan as DNA carrier for hepatic targeting.  
International Journal of Pharmaceutics, 255, 57-68. 
 
Gibaud, S., Demoy, M., Andreux, X. P., Weingarten, C.,Gourtin, B. and Couvreur, P.  
1996.  Cells involved in the capture of nanoparticles in haematopoietic organs.  
Journal of Pharmaceutical Science, 85, 944-950. 
 
Gliding, D. K. and Rees, A. M. 1981.  Biodegradable polymers for use in surgery- 
poly(glycolic)/poly(lactic acid) homo and copolymers: 1.  Polymer, 20, 1459-
1464. 
 
 63 
Gombotz, W. R. and Wee, S.F.  1998.  Protein release from alginate matrices.  Advanced 
Drug Delivery Reviews, 31, 267-285. 
 
Gombotz, Wayne R., Petit, Dean K.  1995.  Biodegradable polymers for protein and 
peptide drug delivery.  Bioconjugate Chemistry, 6, 332-351. 
 
Gref, R., Luck, M., Quellec, P., Marchand, M., Dellacherie, E., Harnish, S., Blunk, T. and 
Muller, R.  2000.  Stealth corona-core nanoparticles surface modified by poly 
ethylene glycol (PEG): Influence of corona (PEG chain length and surface 
density) and the core composition on the phagocytic uptake and plasma protein 
adsorption.  Colloidsans Surfaces B:Biointerfaces, 18, 301-313. 
Grislain, L, Couvreur, P., Lenaerts, V., Roland, M., Deprez-Decempeneere, D. and 
Speiser, P.  Pharmacokinetics and distribution of a biodegradable drug carrier.   
International of Pharmaceutics, 15, 335-345. 
 
Harmann, J. and Bodmeir,R.  Biodegradable somatostatin acetate containing 
microspheres prepared by various aqueous and non-aqueous solvent evaporation 
methods.  European Journal of Pharmaceutics and Biopharmaceutics, 45, 75-82. 
 
Haung, A. and Larsen, B.  1962.  Quantitative determination of uronic acid composition 
of alginates,  Acta Chemica  Scandinavica, 16, 1908-1918. 
 
Hora, M.S., Rana, R. K., Nunberg, J. H., Tice, T.H., Gilley, R.M. and Hudson M. E.  
1990.  Controlled release of interleukin-2 from biodegradable microspheres.  
Bio/Technology, 8,  755-758. 
 
Horbett, T. A. and Brash, J. L.  1987.  Proteins at interfaces: current issues and future 
prospect.  Proteins at Interfaces: Physiochemical and Biochemical Studies.  ACS 
Symposium Series 343, American Chemical Society, Washington, D C., 121-125. 
 
Howard III, M. A., Gross, A., Grandy, M. S., Langer, R. S., Mathiowitz, E., Winn, H. R. 
and Mayberg, M. R.  1989.   Intracerebral drug delivery in rats with leasion-
induced memory deficits.  Journal of Neurosurgery, 71, (1), 105-112. 
 
Huang, A. and Larsen, B. The solubility of alginates at low pH.  Acta Chemica  
Scandinavica, 17, 1653-1662. 
 
Ikada, Y. and Tabata, Y.  Protein release from gelatin matrices.  Advanced Drug Delivery 
Reviews, 31, 287-301. 
 
Jain, J. P., Modi, S., Domb, A.J. and Kumar, N.  2005.  Role of polyanhydrides as 
localized drug carriers.  Journal of Controlled Release, 103, 541-553.    
 
 64 
Jani, P, Halbert, G. W.,Langridge, J. and Florence, A. T.  1989.  The Uptake and 
Translocation of Latex Nanospheres and Microspheres after Oral Administration 
to Rats.  Journal of Pharmacy and Pharmacology, 41, 809-812. 
 
Jayanthi, R., Thanoo, B. C., Mehta, R. C. and DeLuca, P. P.  1996.  Effect of solvent 
removal technique on the matrix characteristics of polylactide/glycolide 
microspheres for peptide delivery.  Journal of Controlled Release, 38, 235-244. 
 
Jensen, J., Song, X. and Brooks, D. E.  1996.  Interfacial thickness of liposomes 
containing poly(ethylene Glycol)-cholesterol from electrophoresis.  Biophysical 
Journal, 70, 313-320. 
Jeyanthi, R., Mehtha, R. C., Thanoo, B. C. and DeLuca, P. P.  1997.  Effect of processing 
parameters on the properties of peptide containing PLGA microspheres.  Journal 
of Microencapsulation, 14,  163-174. 
 
Jiang, W., Gupta, R. K., Deshpande, M. C. and Schwendeman, S. P.  2005.  
Biodegradable poly (lactic-co-glycolic acid) microparticles for injectable delivery 
of vaccine antigens.  Advanced Drug Delivery reviews, 57, 391-410. 
 
Jiang-Bramwell, V. W., Eyles, J. E. and Alpar, H. O.  2005.  Particulate delivery systems 
for biodefense subunit vaccines.  Advanced Drug Delivery reviews, 57, 1247-
1265. 
 
Johansen, P., Merkle, H. P. and Gander, B.,  2000.  Technological considerations related 
to up-scaling of protein microencapsulation by spray drying.  European Journal 
of Pharmaceutics and Biopharmaceutics, 50, 413-417. 
 
Johnson, O. L., Tracy, M. A.  Peptide and protein drug delivery.  Encyclopedia of 
Controlled Drug Delivery, (Mathiowitz, E., Ed.), 2. Walley, New York, 816-833. 
 
Jung, J. and Perrut, M..  2001.  Particle design using supercritical fluid: Literature and 
patient survey.  Journal of Supercritical fluids, 20, 179-219. 
 
Kaushik, K. and Kompella, U. B. 2004.  Preparation of Large Porous Deslorelin-PLGA 
Microparticles with Reduced Residual Solvent and Cellular Uptake Using a 
Supercritical Carbon Dioxide Process.  Pharmaceutical Research, 21, 524-535. 
 
Kawakatsu, T., Komori, H., Nakajima, M., Kikuchi, Y. and Yonemoto, T.  1999.  
Production of monodispersed oil-in-water emulsion using crossflow-type silicon 
microchannel plate.  Journal of Chemical Engineering  of Japan., 32,  241-244. 
 
Kazatchkine, M. D. and Carreno, M. P.  1988.  Activation of complement system at the 
interface between blood and artificial surfaces.  Biomaterials, 9, 30-35. 
 
 65 
Kim, S. H. Won, C. Y. and Chu. C. C.  1999.  Synthesis and characterization of dextran 
based hydrogel prepared by photocrosslinking.  Carbohydrate Polymers, 40, 183-
190. 
 
Kimura, Y., Ozeki, M., Inamoto, T. and Tabata, Y.  2003.  Adipose tissue engineering 
based on human preadepocytes combined with gelatin microspheres containing 
basic fibroblast growth factor.  Biomaterials, 24, 2513-2521. 
 
King, D. J. and Noss, R.R.  1989.  Toxicity of polyacrylamide and acrylamide monomer.   
Review of Environment and Health, 8,  3-16.  
 
Krasko, M. Y., Shikanov, A., Ezra, A. and Domb, A. J. 2003.  Poly(esteranhydride)s 
prepared by incersion of ricinoleic acid into poly(sebasic acid).  Journal of 
Polymer Science, 41, 1059-1069.  
 
Kreuter, J.  1983.  Physicochemical characterization polyacrylic nanoparticles.  
International Journal of Pharmaceutics, 14, 43-58. 
 
Kreuter, J., Petrov, V.E., Kharkevich, D.A. and Alyutdin, R. N.  1997. Influence of the 
type of surfactant on the analgesic effects induced by the peptide dalargin after its 
delivery across blood-brain barrier using surfactant coated nanoparticles.  Journal 
of Controlled Release, 81-87. 
 
Kristl, J., Volk, B., Ahlin, P., Gombac, K. and Sentjerc. M. 2003.  Interactions of solid 
lipid nanoparticles with model membranes and leucocytes studied by EPR.  
International Journal of Pharmaceutics, 7416, 1-8. 
 
Krueter, J., Petrov, V. E., Kharkavich, D. A.. and Alyautdin, R. N.  1997. Influence of the 
type of surfactant on the analgesic effects induced by the peptide dalargin after its 
delivery across the blood-brain barrier using the surfactant-coated nanoparticles.  
Journal of Controlled Release, 49, 81-87. 
 
Kuijpers, A. J., Engbers, G. H.,  Krijgsveld, J., Zaat,S.A., Dankert, J. and Feijen, J.  2000.  
Crosslinking and characterization of gelatin matrices for biomedical applications.  
Journal of Biomaterial Science: Polymer Edition, 11, 225-243. 
 
Kuo, Y., C.  2005.  Loading efficiency of stavudine on polybutylcyanoacrylate and 
methylmethacrylate-sulfopropylmethacrylate copolymer nanoparticles.  
International Journal of Pharmaceutics, 290, 161-172. 
 
Kwong, A. K., Chou,S., Sun, A. M., Sefton, M. V. and Goosen, M. F. A.   (1986).  In 
vitro and in vivo release of insulin from poly(lactic acid) microbeads and pellets.  
Journal of  Controlled Release, 4, 47-62. 
 
 66 
Lanearts,V., Couvreur, p., Christiansen-Leyh, D., Joiris, E. and Roland, M. Degradation 
of poly(isobutylcyanoacrylate) nanoparticles.  Biomaterials, 5, 65-68. 
 
Laurencin, C., Domb, A., Morris, C., Brown, V., Chasin, M., McConnel, R., Lange, N. 
and Langer, R.  1990.  Polyanhydride administration in high doses in vivo: studies 
of biocompatibility and toxicology.  Journal of Biomedical Materials Research, 
24, (11), 1463-1481. 
 
Leelarasamee, N., Howard, S.A., Malanga, C. J. and Ma, J. K.  1996.  A method for 
preparation of polylactic acid microcapsules of controlled size and drug loading.  
Journal of Microencapsulation, 13, 419-433. 
 
Lenaerts, V., Nagelkerke, J. F., Van Berkel, T. J. C., Couvreur, P., Grislain, L., Roland, 
M. and Speiser, P.  In vivo uptake of polyisobutylcyanoacrylate nanoparticles by 
rat liver, Kupffer, endothelial and parenchymal cells.  Journal of Pharmaceutical 
Science, 73, 980-983.  
 
Leon. K. W., D?Amore, P., Marletta, M. and Langer, R.  Bioerodable polyanhydrides as 
drug carrier matrices. II. Biocompatibility and chemical reactivity.  Journal of 
Biomedical Materials Research, 20, 51-64. 
 
Leonard, F., Kulkerrny, R. K., Brandes, G., Nelson, J. and Cameron, J. J.  1966.  
Synthesis and degradation of poly (alkyl)?-cyanoacrylates.  Journal of Applied 
Polymer Science, 10, 259. 
 
Leong, K., D?Amroe, P., Marletta, M. and Langer, R.   1986. Bioerodible polyanhydrides 
as drug carrier matrices. I. Biocompatibility and chemical reactivity.  Journal of 
Biomedical Materials Research, 20, 51-64. 
 
Lewis, D. H.  1990. Controlled release of bioactive agents from lactide/glycolide 
polymers.  Biodegradable Polymers as Drug Delivery Systems, Drugs and 
Pharmaceutical Sciences, Chasin, M. and Langer, R. (Eds.), Marcel Dekker, Pp. 
9-41 
. 
Li, Y., Owusu, S. and Langer, R.  2004.  Treatment of intracranial rat glioma model with 
implant of a radiosensitizer and biomodulator drug combined with external beam 
radiotherapy.  International Journal of Radiation Oncology, Biology, Physics, 58. 
(2), 519-527. 
 
Lim, F. and Sun, A. M.  1980.  Microencapsulated islets as bioartificial endocrine 
pancreas.  Science, 210, 908-910. 
 
Liu, W. R. and Park, T.G.  1995.  Protein release from poly (lactic-co-glycolic acid) 
microspheres: Protein stability problems.  PDA Journal of Pharmaceutical 
Science and Technology, 49,  13-19. 
 67 
 
Liu, W. R., Langer, R. and Klibanov, A. M.  1991.  Moisture induced aggregation of 
lyophilized proteins in solid state.   Biotechnology and Bioengineering, 37, 177-
184. 
 
Losa, C., Calvo, P. Castro, E., Vila Jato, J. L.and Alonso, M. J.  1991.  Improvement of 
ocular penetration of amikacin sulphate by association to 
poly(butylcyanoacrylate) nanoparticles.  Journal of Pharmacy and 
Pharmacology, 43, 548-552. 
 
Maa, Y. F. and Hsu, C.  Liquid ?liquid emulsification by static mixers for use in 
microencapsulation.  Journal of Microencapsulation, 13, 419-433. 
 
Maa, Y. F. and Hsu, C.  Microencapsulation reactor Scale-up by dimensional analysis.  
Journal of Microencapsulation, 13,  53-66. 
 
Maa, Y. F. and Hsu, C.C., 1997.  Effect of primary emulsions on microsphere size and 
protein-loading in the double emulsion process.  Journal of Microencapsulation, 
14, 225-241. 
 
Manning, M.C., Patel, K., Borchatdt, R. T.,  1993.  Stability of protein pharmaceuticals, 
Pharmaceutical Ressearch, 6, 903-918. 
 
Marchais, H., Boury, F., Damge, C., Proust, J. E. and Benoit, J. P.  1996.  Formulation of 
bovine serum albumin loaded PLGA microspheres.  Influence of process 
variables on the loading and in vitro release.  S. T. P. Pharmaceutical Science,  6,  
417-423. 
 
Marco van de Weert, Julia Hoechstetter, Wim E. Hennink, Daan J.A. Crommelin.  2000.  
The effect of water/organic solvent interface on the structural stability of 
lysozyme.  Journal of Controlled Release, 68, 351-359. 
 
Mateovic, T., Krizner, V., Bogataj, M. and Mrhar, A.  2002.  The influence of stirring 
rate on biopharmaceutical properties of Eudragit RS microspheres.   Journal of 
Microencapsulation,  19,  29-36. 
 
Matlaga, B. F., Yashenchak, L. P. and Salhouse, T. N.  1976.  Tissue responses to 
implanted polymers:  the significance of shape. Journal of Biomedical Materials 
Research,  24,  391-397. 
 
Maysinger, D., Jalsenjak, I. and Cuello, A.D.  1992.  Microencapsulated Nerve growth 
factor: Effects on the forebrain neurons following devascularizing cortical 
lleasions.  Neuroscience Letters, 140, 71-74. 
 
 68 
Meijer, A.E.F.H. and Willinghagen, R.G.J.  1963.  The activity of glucose-6-phosphate, 
adenosine triphosphatase, succinic dehydrogenase and acid phosphates after 
dextran or polyvinylpyrrolidone uptake by liver in vivo.  Biochemical 
Pharmacology, 12, 973-979.  
 
Michailova, V., Titeva, S., Kotsilkova, R., Krusteva, E. and Minkov, E.  2001.  Influence 
of hydrogel structure on the process of water penetration and drug release from 
mixed hydroxypropylmethyl cellulose/thermally pregelatinized waxy maize starch 
hydrophilic matrices.  International Journal of Pharmaceutics, 222, 7-17. 
 
Migliaresi, C., Fambri, L. and Cohn, D.   1994.  A study on the in vitro degradation of the 
poly (lactic acid).  Journal of Biomaterials Science: Polymer Edition, 5, 591-606. 
 
Monahan, P. E., Samulski, J. R.  (2000). Adeno associated virus vectors for gene therapy: 
More pros than cons?  Molecular Medicine today, 6, 433-440. 
 
Mondal, T., K. (1999).  Development of Biodegradable Drug delivery System to Treat 
Addiction.  Drug Development and Industrial Pharmacy, 25, (6), 773-779. 
 
Mosqueira, V. C. F., Legrand, P., Gref, R., Heurtault, B., Appel, M. and Barratt, G. 1999. 
Interaction between a macrophage cell line (J774A1) and surface modified poly 
(D, L- lactide) nanocapsules bearing poly (ethylene glycol).  Journal of Drug 
targeting, 7, 65-78. 
 
Mosqueira, V. C. F., Legrand, P., Gulik, A., Bourdon, O., Gref, R., Labarre, D. and 
Barratt, G. 2001. Relationship between complement activation, cellular uptake 
and surface physicochemical aspects of novel PEG-modified nanocapsules.  
Biomaterials, 22, 2967-2979. 
 
Mumper, R. J., Hoffman, A.S., Poulakainen, P.A., Bouchard, L. S. and Gombotz, W.R.  
19894.  Calcium alginate beads for oral delivery of transforming growth factor-?1 
(TGF?-1): stabilization of (TGF?-1) by addition of polyacrylic acid within acid 
treated beads.  Journal of Controlled Release, 30, 241-245. 
 
Nakamura,  T., Shimizu, Y., Okumura, N., Matsui, T., Hyon, S. H. and Shimamoto, T.  
1994.  Tumerogenicity of poly-L-lactide (PLLA) plates compared with medical 
grade polyethylene.  Journal of Biomedical Materials Research, 28, 17-25. 
 
Nakashima,T., Shimizu, M. and Kukizaki, M.  1991.  Membrane emulsification by 
microporous glass.   Key Engineering Materials, 61, 513-520. 
 
Nihant, N., Schugens, C., Grandfills, C., Jerome, R. and Teyssie, P., 1994.  Polylactide 
microparticles prepared by double emulsion/evaporation technique: I. Effect of 
primary emulsion stability.  Pharmaceutical Research, 11, 1479-1484. 
 
 69 
Niu, C. H. Chiu, Y. Y., 1998.  FDA perspective on peptide formulation and stability 
issues.  Journal of Pharmaceutical Science, 87, 1331-1334. 
 
NTP, 81-43, December, 1981.  Second Annual Report on Carcinogens, EPA, 134-136. 
 
Ogawa, Y., Yamamato, M., Takada, S., Okada, J. and Shimamato, T.  1988.  Controlled 
release of Luprolide Acetate from polylactic acid or copoly(lactic/glycolic) acid 
microcapsules: influence of molecular weight and copolymer ratio of polymer.  
Chemical & Pharmaceutical Bulletin, 36, 1502-1507. 
 
Okada, H., Heya, T.,Ogawa, Y., Toguchi, H. and Shimamoto, T.  1991.  Sustained 
pharmacological activities in rats following single and repeated administration of 
once-a-month injectable leuprolide acetate.  Pharmaceutical Research, 8, 787-
791.  
 
Otterelei, M., Ostgaard, K., Skjak-Braek, G., Smidsrod, O., Soon-Shiong, P. and 
Espevic., T.  1991.  Induction of cytokine production from human 
monocytesstimulated with algininate.  Journal of Immunotherapy, 10, 286-291. 
 
Park, K., Shalaby, W.S. and Park, H.  1993.  Biodegradable hydrogels for drug delivery.  
Technomic Publishing Co., Lancaster, Pp. 91-115. 
 
Pearlman, R., Bewley, T.A., 1993.  Stability and Characterization of Protein and Peptide 
Drugs: Case Histories, (Wang, Y. J. and Pearlman, R., Ed).  Planum, New York, 
1-57. 
 
Prabha, S., Zhou, W. Z., Panyam, J., Lahhasetwar, V.  (2002). Size-dependency of 
nanoparticle-mediated gene transfection: studies with fractionated nanoparticles.  
International Journal of pharmaceutics, 244, 105-115. 
 
Quaddoumi, M. G., Ueda, H., Yang, J., Davda, J. and Labhasetwar, V.  2000.  Uptake of 
biodegradable nanoparticles in pigmented rabbit conjunctival epithelial cells.  
Investigative ophthalmology & Visual Science, 41, S765. 
 
Roske-Neilsen,  E., Bojsen-Moller, M., Venter, M. and Hansen, J. C.  1976.  
Polyvinylpyrrolidone-storage disease.  Acta Pathology Microbiology Scandinavia, 
Sect. A. 84, 397-401. 
 
Sah, H., 1999.  Stabilization of proteins against methylene chloride/water interface-
induced denaturation and aggregation.  Journal of Controlled Release, 58, 143-
151. 
 
Sansdrap, P. and Moes, A. J.  1993.  Influence of manufacturing parameters on the size 
characteristics and release profiles of nifedipine from poly (DL-lactide-co-
glycolide) microspheres.  International Journal of Pharmaceutics, 98, 157-164. 
 70 
 
Sawhany, A. S., Pathak, C. P., van Rensberg, J. J., Dunn, R. C. and Hubbel, J. A.  1994.  
Optimization of photopolymerized bioerodible hydrogel properties for adhesion 
prevention.  Journal of Biomedical Materials Research, 28,  831-838.   
 
Seidel, C., Kullicke, W. M., Heb, C., Hartmann, B., Lechner, M. D. and Lazik, W.  2001.  
Influence of Cross-linking Agent on the Gel Structure of Starch derivatives.  
Starch Starke, 53, 305-310. 
 
Shakakibara, Y., Tambara, K., Shakaguchi, G., Lu, F., Yamamoto, M., Nashimura, K., 
Tabata, Y. and Komeda, M.  2003.  Toward surgical angiogenesis using slow 
release basic fibroblast growth factor.  European Journal of Cardio Thoracic 
Surgery, 24, 105-111. 
 
Shinde, B. G. and Erhan, S.  1992.  Flexibilized gelatin skin based artificial skin model: 
II. Release kinetics of incorporated bioactive molecules.  Bio-Medical Materials 
and Engineering, 2, 127-131.  
 
Sinha,  V. R. and Trehan, A.   2003.  Biodegradable microsphere for protein delivery.  
Journal of Controlled Release, 90, 261-280. 
 
Smidsrod, O. and, Skjak-Braek, G.  1990. Alginate as immobilization matrix for cells.  
TIBTECH, 8, 71-78. 
 
Sommerfeld, p., Schroeder, U. and Sabel, A.  1997.  Long-term stability of PBCA 
nanoparticles suggests clinical usefulness.  International journal of 
Pharmaceutics, 155, 201-207. 
 
Soon-Shiong, P., Otterlie, M., Skjak-Braek, G., Smidsord, O., Heintz, R., Lanza, R. and 
Espevic, T.  1991.  An Immunologic basis for the fibrotic reaction to implanted 
microcapsules.  Transplantation Proceedings, 23, 758-759. 
 
Soriano, I., Delgado, A., Diaz, R. V. and Evora, C., 1995.  Use of surfactants in 
polylactic acid protein microspheres.  Drug Delivery and Industrial Pharmacy, 
21, 549-558.  
 
Spargo, B. J., Rudolph, A.S., Rollwagen, F.M.  1994.  Recruitment of tissue resident cells 
to hydrogel composites: in vivo response to implant materials.  Biomaterials, 15, 
853-858.  
 
Stenekes, R. J. H., Loebis, A. E., Fernandes, C.M., Crommelin, D.J.A. and Hennik, W.   
2000.  Controlled Release of Liposomes from Biodegradable Dextran 
Microspheres: A Novel Delivery Concept.  Pharmaceutical Research, 17, 690-
695. 
 
 71 
Stolnik, S., Illum, L., Davis, S.S.  (1995).  Long circulating microparticulate drug 
carriers.  Advanced Drug Delivery Reviews, 16, 195-214. 
 
Stolnil, S., Illum, L. and Davis, S.S.  1995.  Long circulating microparticulate drug 
carriers.  Advanced Drug Delivery Reviews, 16, 195-214. 
 
Sutherland, I. W.  1991.  Alginates, in: Biomaterials; Novel Materials from biological 
sources, Stockton Press, New York, 309-331. 
 
Tabata, Y. and Ikada, Y.  1989.  Protein precoating of polylactide microspheres 
containing a lipophilic immunopotentiator for enhancement of macrophage 
phagocytosis and activation.  Pharmaceutical Research, 6, 296-301. 
 
Tabata, Y., Gutta, S. and Langer, R.  1993.  Controlled delivery of proteins using 
polyanhydride microspheres.  Pharmaceutical Research, 10, 487-496. 
 
Tabata, Y., Hijikata, S. and Ikada, Y.  1994.Enhanced vascularization and tissue 
granulation by basic fibroblast growth factor impregnated in gelatin hydrogels.  
Journal of Controlled Release, 31, 189-199. 
 
Tabata, Y., Nagano, A. and Ikada, Y.  1999.  Biodegradation of hydrogel carrier 
incorporating fibroblast growth factor.  Tissue Engineering, 5, 127-138. 
 
Teomim, D., Nyska, A. and Domb, A. J.   1999.  Ricinoleic acid based biopolymers.  
Journal of Biomedical Materials Research, 45, 258-267. 
 
Thoma, K. and Schluetermann, B.  1992.  Relationship between manufacturing 
parameters and pharmaceutical-technological requirements on biodegradable 
microparticles.  Part3: Pharmaceutical and technological investigations on particle 
size distribution of local anesthetics containing biodegradable microspheres.  
Pharmazie, 198-202.   
 
Thomasin, C., Johansen, P., Alder, R., Bemsel, R., Hottinger, G., Altrofer, H., Wright, A. 
D., Wehrli, E., Merkle, H. P. and Gander, B.,  1996.  A contribution to 
overcoming the problem of residual solvents biodegradable microspheres 
prepared by coacervation.  European Journal of Pharmaceutics and 
Biopharmaceutics, 42, 16-24. 
 
Thompson, J. A., Anderson, K D., DiPietro, J. M., Zwiebel, J. A., Zametta, M., 
Anderson, W. F. and Macing T.  1988.  Site directed neo-vessel formation in vivo.  
Science, 241, 1349-1352. 
 
Tice, T. R. and Gilley, R. M.  1995. Microencapsulation process and products, therefrom.  
US Patent  5,407,609. 
 
 72 
Tice, Thomas (2004).  Delivering With Depot Formulations.  Drug Delivery Technology, 
4, (1), 44-47. 
 
Tsang, P. W., Kwan, H. C. and Sun, A. M.  1996.  Striatal transplant of 
microencapsulated bovine cromaffin cells reduces rotational behavior in the rat 
model of Parkinsonism.  Proceedings of International Symposium on Controlled 
Release of Bioactive Compounds, 23, 6-7. 
 
Ueyama, K., Bing, G., Tabata, Y., Ozeki, M., Doi, K., Nashimura, K., Suma, H. and 
Komeda, M.  2004.  Development of biologic coronary artery bypass grafting in 
rabbit model: revival of a classic concept with modern biotechnology.  Journal of 
Thoracic and Cardiovascular Surgery, 127, 1608-1615. 
 
Ullivan, C. O. and Birkinshaw, C.  2004.  In vitro degradation of insulin loaded 
poly(butylcyanoacrylate) nanoparticles.  Biomaterials, 25, 4375-4382. 
 
Uppal, P., Jampel, H. D., Quigley, H. A. and Leong, K. W.  1994.  Pharmacokinetics of 
etoposide delivery by a boerodible drug carrier implanted at glocoma surgery.  
Journal of Ocular Pharmacology, 10, 471-479. 
 
Verrecchia, T., Huve, P., Brazile, D., Veillard, M.,  Stephenhauer, G. and Couvreue, P.  
1993.  Adsorption/desorption of human serum albumin at the surface of poly 
(lactic acid) nanoparticles prepared by a solvent evaporation process.  Journal of 
Biomedical Materials Research, 27, 1019-2028.  
 
Verrecchia, T., Spenlehaur, G., Brazile, D. V., Murry-Brelier, A., Archimbaud, Y. and 
Veillard, M. V.  1995.  Non-stealth [poly (lactic acid/albumin)] and stealth 
(poly(lactic acid-polyethylene glycol)) nanoparticles as injectable drug carriers.  
Journal of Controlled Release, 36, 49-61. 
 
Verzar, F.  1936.  Absorption from the Intestine.  Longmans, Green, London, New York, 
Pp. 89-101.. 
 
Visscher, G. E., Robinson, R. L. and Agrentieri, G. J.  1987. Tissue response to 
biodegradable injectable microcapsules.  Journal of Biomaterials Applications, 2, 
118-131. 
 
Visscher, G. E., Robinson, R. L., Maulding, H. V., Fong, J. W., Pearson, J. E. and 
Agrentieri, G. J.  1985.  Biodegradation and tissue reaction to 50:50 poly(DL-
lactide-co-glycolide)microcapsules Journal of Biomedical Materials Research, 
19, 349-357. 
 
Viswanathan, N. B., Thomas, P. A., Pandit, J. K.,  Kulkerni, M. G. and Mashelkar, R. A.  
1999.  Preperation of non-porous microspheres with high entrapment efficiency of 
 73 
proteins by a (water-in-oil)-in oil emulsion technique.  Journal of Controlled 
Release, 58, 9-20.  
 
Vittaz, M., Brazile, D., Spenlehauer, G., Verracchia, T., Veillard, M., Puisieux, F. and 
Labarre, D.  1996.  Effect of PEO density on the long-circulating PLA-PEO 
nanoparticles which are very low complement activators.  Biomaterials, 17, 1575-
1581. 
 
Vladislavljevic, G. T. and Williams, R. A.  2005.  Recent developments in manufacturing 
emulsions and particulate products using membranes.  Advances in Colloid and 
Interfacial Sciences, 113, 1-20.  
 
Vos, P. D., Hann, B. D., Wolters, G. and Schilfgaarde, R.V.  1996.  Factors influencing 
the adequacy of microencapsulation of rat pancreatic islets.  Transplantation, 62, 
888-893. 
 
Waeckerle-Men, Y. and Goettrup, M.  2005.  PLGA microspheres for improved antigen 
delivery to dendritic cells as cellular vaccines, Advanced Drug Delivery reviews, 
57, 475-482. 
 
Wang, P. P., Frazier and Z., Brem, H.  2002.  Local drug delivery to the brain.  Advanced 
Drug Delivery reviews, 54, 987-1013. 
 
Weiss, W. M., Riles, T. S., Gauge, T. H. and Mizarchi, H.  1991.  Angiocarcinoma at the 
site of a Dacron vascular prosthesis: a case report and literature review.  Journal 
of Vascular Surgery, 14, 87-91. 
 
Wilhelm, C., Billotey, C., Roger, J., Pons, J. N., Bacri, J. C. and Gazaeu, F.  2003.  
Intracellular uptake of anionic superparamagnetic nanoparticles as a function of 
their surface coating.  Biomaterials, 24, 1001-1011.  
 
Wilhelm, C., Billoty, C., Roger, J., Pons, J. N., Bacri, J. C., Gazeau, F.  (2003). 
Itracellular uptake of anionic suparamagnetic nanoparticles as a function of their 
surface coating.  Biomaterials, 24, 1001-1011. 
 
Yamaguchi, K. and Anderson, J.M.  1993.  In vivo biocompatibility studies with 
Medisorb 65/35 D,L-lactide/glycolide copolymer microspheres.  Journal of 
Controlled Release, 24, 81-93. 
 
Yamamoto, M. Takahashi, Y. and Tabata, Y.  2003.  Controlled release by biodegradable 
hydrogels enhances the ectopic bone formation of bone morphologic protein.  
Biomaterials, 24, 4375-4383. 
 
 74 
Yamamoto, M., Ikada, Y. and Tabata, Y.  2001.  Controlled release of growth factos 
based on biodegradation of gelatin hydrogel.  Journal of Polymer Science, 12, 77-
88. 
 
Yang, S., Wong, M., Tabata, Y. and Micos, A. G.  2005.  Gelatin as a delivery vehicle for 
the controlled release of bioactive molecule.  Journal of Controlled Release, 109, 
256-274. 
 
Yang, Y. Y., Chung, T. S. and Ng, N. P.  2001.  Morphology, drug distribution, and in 
vitro release profiles of biodegradable microspheres containing protein fabricated 
by double emulsion solvent extraction/evaporation method.  Biomaterials, 22, 
231-241. 
 
Yang, Y. Y., Chung, T. S., Bai, X. L. and Chan, W. K., 2000.  Effect of preparation 
conditions on morphology and release profiles of biodegradable polymeric 
microspheres containing protein fabricated by double emulsion method.  
Chemical Engineering Science, 55, 2223-2236. 
 
Yao, C.h., Liu, B.S., Hsu, S. H., Chen, Y. S. and Tsai, T. T.  2004.  Biocompatability and 
Biodegradation of bone composite containing tricalsium phosphate and genipin 
crosslinked gelatin.  Journal of Biomedical Materials Research, 69A, 709-717. 
 
Yoshida,  S. H., Chang, C.C., Teuber, S. S. and Gershwin, M. E.  1993.    Silicon and 
silicone:  Theoretical and clinical implications of breast implants.  Regulatory 
Toxicology and Pharmacology,  17,  3-18. 
 
Zale, S.E. and Klibanov, A. M.  1989.  Why does ribonuclease irreversibly inactivate at 
high temperatures?  Biochemistry, 25, 5432-5443. 
 
Zange, R. and Kissel, T.  1997.  Comparative biocompatibility testing of 
polycyanoacrylates and poly (D,L-Lactide-co-glycolide) using different mouse 
fibroblast (L929) biocompatibility test models.  European Journal of 
Pharmaceutics and Biopharmaceutics, 44, 149-157. 
Zhang, Q., Liao, G., Wei, D. and Nagai, T.  1998.  Increase in gentamicin uptake by 
cultured mouse peritoneal macrophages and rat hepatocytes by its binding to 
polybutylcyanoacrylate nanoparticles.  International Journal of Pharmaceutics, 
164, 21-27. 
 
Zhang, Q., Shen, Z. and Nagai, T.  2001.  Prolonged hypoglycemic effect of insulin-
loaded polybutylcyanoacrylate nanoparticles after pulmonary administration to 
normal rats.  International Journal of Pharmaceutics, 218, 75-80. 
 
Zimmerman, U., Klock, G, Grohn, P., Jangaard, N. O. and Wilkinson, C.K. 1995.  Cell 
encapsulation in alginate: Elimination of Foreign Body Reaction and 
Microcapsule volume minimization.  Monsanto Co. Newsletter, 6, 11-19. 
 75 
 
 
 
III. HPLC ANALYSIS OF ALIPHATIC AND AROMATIC DICARBOXYLIC 
ACID CROSS-LINKERS HYDROLYZED FROM CARBOHYDRATE 
POLYESTERS FOR ESTIMATION OF THE MOLAR DEGREE OF 
SUBSTITUTION 
 
ABSTRACT 
Quantitative analysis of carboxylic acids in hydrophilic matrixes remains a challenge 
despite the introduction of derivatization reagents for carboxylic acids nearly three 
decades ago.  Analysis of many bifunctional ester forming polymer cross-linkers are 
additionally complicated by their amphiphatic character.  We report a sensitive and 
selective HPLC method to quantify organic dicarboxylic acids resulting from the 
hydrolytic degradation of cross-linked carbohydrate polyesters.  An important group of 
cross-linkers encompassing C7-C10 aromatic and nonchromophoric aliphatic diacid 
chlorides were investigated.  Specific validation is reported for terephthalic acid, 
diethylmalonic acid and sebacic acid.  Linear calibration curves were obtained in the 0.5 -
15.0 ?g/mL range corresponding to a mass fraction of 0.05 - 1.5% diacid to polyester.  
The accuracy and precision in this range was 94.4 - 114.1% and 6.1 - 24.8% RSD, 
respectively.   This method is suitable for the routine determination of the molar or mass 
degree of substitution in carbohydrate polyesters. 
 
 76 
1. INTRODUCTION  
Carbohydrates are the most abundant natural polymers available on earth.  Among 
them, starch, in all its natural and semi-synthetic forms, is of great interest to food and 
pharmaceutical scientists because of its inherent advantages regarding safety and 
biocompatibility.  Unmodified starch has limited application because of the undesirable 
properties of gelation and retrogradation.  Cross-linking starch polymer molecules with 
bifunctional reagents by far the most important chemical modification  used to control 
and minimize these undesirable properties (Langan, 1986), as well as to be used to create 
desirable properties such as in the development of biodegradable drug delivery devices.    
In this context, material development requires a routine method which is efficient and 
sensitive to quantify organic dicarboxylic acids resulting from modified carbohydrate 
polymers in order to relate the degree of cross-linking to physical, structural or chemical 
properties of the polymer.   
Since most polyesterification reactions result in only a 0.5?2.0% degree of molar 
substitution of cross-linkers in the polymer (Jarowenko, 1986), sensitive methods of 
detection of the diacid cross-linkers are required.  In addition, many dicarboxylic acids 
commonly employed as cross-linkers have limited solubility profiles in aqueous and 
organically based solvent systems.  For instance, terephthalic acid, an important cross-
linker in condensation polymerizations, is practically insoluble in any solvent other than 
aqueous alkali.  
Although many methods have been reported for the analysis of monocarboxylic acids 
(Coenen, A. J. J. M., 1992; Toyo?oka, T., 1995; Mehta, A., 1998), only a few methods 
have been reported for the analysis of dicarboxylic acids (Fischer, K., 1997; 
 77 
Baziramakenga, R., 2001; Grushka, E., 1975) with requisite sensitivities in the upper 
nanomolar to lower micromolar concentration range.  Ion chromatographic methods 
(Fischer, K., Baziramakenga, R.) utilizing ion-exchange separation and conductance 
detection report the analysis of ?C6 aliphatic dicarboxylic acids and a number of aromatic 
dicarboxylic acids with good sensitivity.  The acidic mobile phase in one method 
(Fischer, K.), however, limits the method to only the most soluble of the aliphatic 
dicarboxylic acids.  No accuracy and precision data are given for this method. The other 
method (Baziramakenga, R.), requires two separate chemical analyses using different 
analytical columns for the separation of aromatic and aliphatic dicarboxylic acids.  
Reported accuracy and precision for this method are good.  Both methods require ion 
suppression to achieve optimal sensitivity.  A gas chromatographic method using on-
column trimethylsilane derivatization (Docherty, K. S.), is described for the analysis of 
C2 - C10 dicarboxylic acids with excellent selectivity, but inadequate sensitivity.   
Sensitive and selective HPLC methods for analysis of fatty carboxylic acids are 
reported based on derivatization with fluorescent chromogenic reagents (Naganuma H., 
1989; Yamaguchi, M., 1990).  These methods, however, require synthesis and 
purification of the fluorescent reagents which are not commercially available.  Even with 
the power of tandem mass spectrometry, derivatization is necessary to obtain adequate 
selectivity in complex matrixes (Kushnir,  M. M., 2001).  
Reported reverse-phase HPLC methods (Grushka, E., 1975; Capristo, E., 1999) with 
potentially adequate sensitivity are based on the derivatization reaction of carboxylic 
acids with p-bromophenacyl bromide (Durst, H. D., 1975) or phenacyl bromide (Borch, 
R.F., 1975) both of which are commercially available.  The utility of this reaction was 
 78 
demonstrated for nonchromophoric C4-C6 dicarboxylic acid esters of phenacyl bromide, 
but not p-bromophenacyl bromide (Grushka,E.).  These authors found the p-
bromophenacyl derivatives of some diacids, especially fumaric and succinic, to be 
insoluble in almost all solvents.  Phenacyl bromide derivatives showed improved 
solubility characteristics, but were still compromised.  Chromatographic resolution of the 
dicarboxylic acid derivatives was poor and exhibited peak distortion on a nonyl bonded 
stationary phase column.  Another study (Capristo, E.) showed much improved 
specificity and peak shape of detected dicarboxylic acids resulting from the in vivo 
deesterification of triglyceride esters of sebacic and dodecanedioic acids on an octadecyl 
bonded stationary phase.  Analytical accuracy and precision was only provided with 
respect to the triglyceride esters for this method.  
The purpose of this study was to develop a sensitive and selective HPLC assay for 
routine analysis of organic dicarboxylic acids resulting from the hydrolytic degradation 
of cross-linked polyester carbohydrates in order to determine the molar or mass degree of 
substitution from various polymeric materials.  Particular emphasis was placed on assay 
methodology for an important category of dicarboxylic acid chloride cross-linkers 
encompassing C7-C10 aliphatic and aromatic diacid chlorides.  Validated analytical 
accuracy and precision is reported specifically for terephthalic acid, diethylmalonic acid 
and sebacic acid, although the method is suitable for a variety of C6 diacids. 
 
 
 
 79 
2. EXPERIMENTAL  
2.1 Preparation of reagents and sources 
Acetonitrile was used as a reverse phase HPLC solvent (Optima grade, Fisher 
Scientific/Acros Organics, Pittsburgh PA, USA).  N, N-dimethylformamide (Fisher), 
glacial acetic acid (Fisher), ammonium hydroxide (Fisher) and potassium hydroxide 
(Fisher) were reagent grade.  Terephthalic acid (Acros), phenacyl bromide (2-
bromoacetophenone, Acros), sebacic acid (Sigma-Aldrich/Fluka, St. Louis MO, USA), 
diethylmalonic acid (Aldrich), suberic acid (Aldrich), azelaic acid (Fluka), and 
triethylamine (Sigma) were of purity >98%.   Water for the mobile phase was double 
glass distilled.   For the polymer condensation reactions, terephthaloyl chloride >99% 
(Aldrich), sebacoyl chloride 92% (Acros), and diethylmalonoyl dichloride 98% (Aldrich) 
were used as reagents. 
Analytical derivatization standard solutions were prepared as follows.  A 200 mM (40 
mg/mL) phenacyl bromide solution was made by dissolving 400 mg phenacyl bromide in 
10 mL of acetonitrile.  A 60 mM triethylamine solution was prepared by adding 832 ?L 
of triethylamine to a 100 mL volumetric flask and diluting to volume with acetonitrile.  
 
2.2 Synthesis of carbohydrate polyesters 
Polyesterification of a variety of starches and modified starches with different mass 
percentages of amylose and amylopectin were investigated.  These included maltodextrin 
from corn starch (Maltrin M40, Grain Processing Corporation, Mucatine IA, USA), 
maltodextrin from waxy maize starch (C*Pharm 01980, Cerestar USA, Inc., Hammond 
IN, USA), corn starch (Sigma), high amylose corn starch (C*Amylogel 03003, Cerestar 
 80 
USA, Inc.), waxy maize starch (PFP 2850, Cargill Foods, Minneapolis MN, USA),  
hydroxypropylated starch modified from tapioca starch (C*Aratex 7570, Cerestar USA, 
Inc.), and hydroxyethyl starch modified from corn starch (Sigma).   Polyesterification of 
the starches was accomplished using an emulsion polymerization technique similar to 
that reported by L?vy and Andry (L?vy, M. C.,1992).  Briefly, a specified carbohydrate 
was dispersed in aqueous 200 mM borate buffer, pH 10.  Carbohydrate concentrations 
ranged from 3 to 20% (w/v) for starches and were 50% for maltodextrins.  The dispersion 
was heated to approximately 100?C to produce a solubilized colloidal solution, and then 
allowed to cool to about room temperature.  The aqueous colloid was emulsified with a 
continuum phase in a 1:5 solvent ratio, respectively, consisting of a mixture of 
cyclohexane and chloroform 68:32 (v/v) containing an appropriate surfactant in a 
homogenizer. After homogenization, an aliquot of cylcohexane/chloroform containing 
the diacid chloride is added to the emulsion to give a final concentration of 60 mM in the 
continuum phase.  The reaction was carried out for 3 hours at ambient temperature with 
stirring.  Esterified carbohydrate was precipitated by adding the reacted emulsion to 
excess ethanol while stirring.  The precipitated residue was collected either by 
gravitational settling or centrifugation, decanting, and washed three additional times with 
ethanol.  The product was dried in a vacuum oven at 20-25?C at 20 psi for subsequent 
analysis of cross-linking efficiency and as material to be used to determine the suitability 
of this assay.  
 
 
 
 81 
2.3 Calibration standards, sample preparation, and polyester hydrolysis 
Depending on the application, either terephthalic acid, diethylmalonic acid, or sebacic 
acid was chosen as the internal standard.  The internal standard primary working solution 
was prepared by dissolving 40 mg of the appropriate acid in 100 mL of 5 M ammonium 
hydroxide to make a final concentration 400 ?g/mL.   The other two acids were regarded 
as analytes.  Working solutions were prepared by dissolving 40 mg analyte with an 
appropriate volume of 5 M ammonium hydroxide to give two working solutions with 
final concentrations of 40 ?g/mL and 400 ?g/mL.  Calibration standards were prepared 
by adding 500 ?L of the 40 ?g/mL analyte solution, and 100, 500, 1000 and 1500 ?L of 
the 400 ?g/mL analyte solution, respectively, to 4 mL screw-top glass vials containing 40 
mg of nonesterified carbohydrate.  A 1000 ?L aliquot of the internal standard solution 
was added, and aliquots of 5 M ammonium hydroxide were added to each vial to give a 
total volume of 4.0 mL for each sample.  For carbohydrate polyesters, 40 mg of the 
polymer, 1000 ?L of the internal standard solution, and 3.0 mL of 5 M ammonium 
hydroxide were added to the vial. 
The vials were tightly capped and hydrolyzed at 130?C in a dry bath for 60 min.  
After hydrolysis, the samples were allowed to cool to ambient temperature, and 1.0-mL 
aliquots of hydrolysate were transferred to 15-mL centrifuge tubes.  A 9.0 mL aliquot of 
a 60 mM triethylamine in acetonitrile solution was then added to each tube in order to 
precipitate any remaining insoluble carbohydrate.  The tubes were vortexed and 
centrifuged at 2,500 rpm for 30 min separating any insoluble material from the dissolved 
diacids.  Exactly 1.0 mL of the clear supernatant was retained for subsequent analytical 
derivatization and transferred to a reaction vial.  The theoretical final concentrations of 
 82 
the calibration standards are 0.5, 1.0, 5.0, 10, and 15 ?g/mL, corresponding to 0.5, 1.0, 
5.0, 10, and 15 ?g/mg carbohydrate, or 0.05 to 1.5 % (w/w) respectively.  
 
2.4 Analytical derivatization 
The 1-mL aliquot of retained supernatant from each sample is evaporated until 
completely dry in a vacuum manifold.  It is important that sample is dry and any excess 
ammonium hydroxide is removed by evaporation of the ammonia.  To each sample is 
added 300 ?L dimethylformamide, 100 ?L acetonitrile, 300 ?L 60 mM triethylamine in 
acetonitrile, and 300 ?L phenacyl bromide solution.  The reaction vials are capped and 
heated at 60?C for 60 min in a dry bath.  Following the reaction, the samples are removed 
from the heating block and allowed to cool to ambient temperature before subsequent 
processing. 
 
2.5 Separation of soluble polysaccharides 
It is necessary to reduce the soluble polysaccharide background of the sample 
resulting from the hydrolytic decomposition of polyesters and oxidative decomposition of 
the carbohydrate backbone.  Although residual polysaccharides do not interfere with the 
derivatization, they can effect the chromatography. This separation was achieved by solid 
phase extraction using PrepSep (Fisher) C18 SPE cartridges (500 mg sorbent/6 mL).  The 
SPE cartridge is conditioned by passing 4 mL of distilled water through the cartridge.   
The derivatized sample (1 mL) is diluted with 2 mL of distilled water and drawn through 
the sorbent with a vacuum manifold, washed with 4 mL water, and eluted with 1.0 mL of 
acetonitrile into autosampler vials. 
 83 
2.6 Samples for derivatization reaction kinetics 
Standard solutions of diacids used in the evaluation of the derivatization reaction 
kinetics were prepared by dissolving the diacids in an appropriate volume of 60 mM 
triethylamine in acetonitrile. An aliquot of 1.0 mL was evaporated until dry, and 
derivatized according to the above procedure.  The samples were reacted at 60?C for 
varying periods of time.  At the completion of each reaction period, sample vials were 
placed in a dry ice/acetone slurry in order to quench the reaction.  Samples were retained 
in this bath until injected on-column for chromatographic analysis without further 
cleanup. 
 
2.7 Chromatographic conditions and instrumentation 
The chromatographic system consisted of ThermoSeparation (San Jose CA, USA) 
Model P2000 pumps used in the isocratic mode, Model AS3000 autoinjector, and Model 
UV1000 variable-wavelength UV detector.  Data collection and processing used 
Chromquest v3.0 software and a PC computer. The stationary phase was a 25 cm x 4.6 
mm i.d., 5 ?m, 150 ? pore size, C8 bonded phase on silica chromatographic column (Beta 
Basic 8, ThermoHypersil-Keystone, Bellefonte PA, USA) and a 10 mm x 4.6 mm i.d., 
Beta Basic 8 precolumn.  The mobile phase consisted of 60% acetonitrile, 40% 100 mM 
acetate buffer at pH 4.5 delivered at a flow rate of 1.0 mL/min.  The column temperature 
was ambient.  The detection wavelength for all derivatized analytes was 242 nm.  The 
injection volume was 15 ?L. 
 
 
 84 
3. RESULTS AND DISCUSSION     
3.1 Phenacyl ester formation 
The reaction of carboxylic acids with phenacyl bromide typically require the 
production of a ?naked? carboxylate anion utilizing either triethylamine (Borch, R. 
F.,1995) or crown ethers (Durst, H. D.,1975) as catalysts.  The proper choice of base used 
to hydrolyze carbohydrate polyesters is essential. The high concentration of base required 
for the hydrolysis precludes the use of alkali metal bases which have the potential to 
overwhelm the catalysts competitively with metal cations, as well as provide a large 
excess of hydroxide ions which are good nucleophiles for the production of phenacyl 
alcohols.  Both of these reactions hinder the desired formation of phenacyl esters of the 
analytes.  The use of ammonium hydroxide for hydrolysis was found to overcome these 
deficiencies since excess reagent can be evaporated leaving a stoichiometric 1:1 
ammonium salt with carboxylate anion.  The ammonium cation is displaced with excess 
triethylamine as the catalyst in the reaction solvent.  
Another problem is reaction solvent choice.  Although the aprotic solvent acetonitrile 
is generally reported as the preferred reaction solvent for fatty acid derivatization, 
reaction products from some diacids are insoluble (Grushka, E., 1975).  The phenacyl 
ester derivatives of the diacids investigated in this study were generally found to be 
solubility limited in neat acetonitrile.  Co-solvency with 30% dimethylformamide 
improved solubility of diacids and products in the reaction medium. 
Terephthalic, diethylmalonic and sebacic phenacyl ester formation was facile under 
the prescribed conditions.  The reactions kinetics (Figure1) showed completion of 
derivative formation within 40 min for all three diacids.  The kinetic profiles were 
 85 
monotonically asymptotic.  Chromatographic analyses showed that the retention time (tR) 
of all corresponding products were independent of mobile phase pH in a range from 2.9 
to 7.0.  In addition, phenacyl ester derivative kinetics remained stable beyond 40 min for 
reactions carried out to 2 hr.  Although theoretically a mixture of both monoester and 
diesters products can be formed, no biphasic kinetic profile of the reaction products was 
observed corresponding to monoester formation, typical of A ? B ? C kinetics with B 
corresponding to the monoester and C the diester.  In view of the pH independence of the 
reaction products, these results indicate that the phenacyl derivatives were diesters. 
3.2 Chromatography and selectivity 
Typical chromatograms obtained from spiked standards prepared with Maltrin M40 
as the background, a blank sample with Maltrin M40, and authentic Maltrin M40 
polyester crosslinked with terephthaloyl chloride are illustrated in Fig. 2 and 3.   
Esterification of carboxylic acids with phenacyl bromide, produce a large number of 
reaction by-products, having retention times of < 10 min in this assay.  These peaks are 
present in chromatograms from neat samples without a carbohydrate background as well 
as those with carbohydrates.  Similar results in other matrixes using this reaction have 
been reported (Capristo,E., 1999).  In developing the chromatography, the intent was to 
adequately separate analytes of interest by increasing on-column retention relative to the 
reaction by-products in order to increase selectivity.   
Terephthalic acid, diethylmalonic acid and sebacic acid derivative retention times 
were 12.3 ? 0.11 (n=25), 15.9 ? 0.14 (25), and 19.5 ? 0.20 (25) min, respectively.  The 
blank control sample was clean with respect to the diethylmalonic acid and sebacic acid 
derivatives.  Low level interference with a similar retention time as terephthalic acid was 
 86 
detected in the blank which limited the lower level of quantitation for terephthalic acid to 
about 0.5 ?g/mL.  Qualitatively, chromatographic profiles from blank samples using 
other carbohydrates resulted in virtually identical chromatography.    
The suitability of this assay for other dicarboxylic acids including suberic and azelaic 
acids is demonstrated in Fig. 4.  For ?C5 dicarboxylic acids including succinic acid, 
retention was inadequate (< 10 min) resulting in reaction by-product interference.  Since 
the retention time increases with diacid hydrocarbon length, this assay would be 
considered to provide adequate selectivity for dicarboxylic acids of C6 hydrocarbon 
length.   
Minimizing soluble polysaccharides in the injected samples was essential.  During the 
development of this method and prior to adding an SPE cleanup step, it was noticed that 
analyses of an autosampler tray of samples resulted in progressively longer retention 
times of analytes to values more than double their initial values. Although the Beta Basic 
8 column is specified by the manufacturer to be endcapped with a 7% carbon load, 
significant on-column accumulation of background polysaccharides was found, 
presumably due to their interaction with the silica support.  This accumulation resulted in 
significant changes in chromatography probably due to mixed mode retention of analytes.  
Retention of the background polysaccharides relative to the analytes on a C18 SPE 
cartridge, however, was found to be minimal, and analyte extraction was optimized to 
wash out soluble polysaccharides and elute the purified analytes.  Polysaccharide on-
column buildup was found to be reversible such that columns could be regenerated by 
flushing with 8.5 mM aqueous ammonium hydroxide for 30 min followed by distilled 
water.  Routine column cleanup is recommended for assay ruggedness.  
 87 
3.3 Linearity of calibration curves 
Calibration curves were constructed by statistically analyzing 5 calibration levels with 
5 replications within each level. All samples contained Maltrin M40 as the background 
carbohydrate.  These curves were evaluated with regard to peak area ratio and peak 
height ratio of a designated analyte relative to the internal standard.  The selected internal 
standard was added at a concentration level to attain a final concentration of 10 ?g/mL 
after hydrolysis and carbohydrate precipitation.  For terephthalic and diethylmalonic acid 
standard curves, sebacic acid was chosen as the internal standard.  For sebacic acid 
standard curves, diethylmalonic acid was chosen as the internal standard.  Least-squares 
regression analysis used a model with peak area ratio or peak height ratio as the 
dependent variable and concentration level as the independent variable. Linear calibration 
curves were established for all three diacids in the 0.5 to 15 ?g/mL calibration range 
(Table 1).  Extended calibration ranges from 0.2 to 40 ?g/mL tended to deviate from 
linearity at the upper end, while the 0.2 ?g/mL level was determined to be below the limit 
of quantitation.  
 
 3.4 Accuracy and Precision 
Accuracy of analyte determinations ranged from 94.4 to 114.1% for all three diacids 
with the exception of the 0.5 ?g/mL sebacic acid concentration level (Table 2).  The 
precision of estimates ranged from 6.1 to 24.8 % RSD.  UV response limited the lower 
limit of quantitation for sebacic acid to 1.0 ?g/mL, while terephthalic and diethylmalonic 
acids were sufficiently sensitive to give lower limits of quantitation of 0.5 ?g/mL.   
Although assay reproducibility was considered acceptable, improved precision could be 
 88 
expected based on duplicate sample analysis.  This is particularly relevant with regards to 
diethylmalonic acid. 
The effect of different carbohydrate matrixes was evaluated with regards to accuracy 
based on calibration curves employing Maltrin M40 as the background carbohydrate 
matrix.  Calibration standards containing 10 ?g/mL terephthalic acid and maltodextrin 
from waxy maize starch, high amylose corn starch, waxy maize starch, 
hydroxypropylated tapioca starch, and hydroxyethylated corn starch were analyzed.  The 
relative accuracies compared to the assayed 10 ?g/mL terephthalic acid concentration 
level with Maltrin M40 were 93.6% for waxy maize starch, 95.0% for maltodextrin from 
waxy maize starch, 92.4% for high amylose corn starch, 101.1% for hydroxyethyl corn 
starch, and 105.9% for hydroxypropyl tapioca starch.  All assayed values were within the 
95% confidence interval for terephthalic acid assayed in a Maltrin M40 matrix.  These 
results indicate that the assay is robust with regards to different carbohydrate matrixes 
and polyesters.  
 
3.5 Carbohydrate Polyester Degree of Substitution 
The cross-linking efficiency of various polyesterified starches was analyzed to relate 
molar degree of substitution to material properties required for nanoparticulate drug 
delivery systems, and to evaluate the application suitability of this method.  In general, all 
carbohydrates tested reacted to form three-dimensional structured nanocapsules (Figure 
5).  The molar degree of substitution per anhydroglucose unit ranged from 0.044 to 
8.32% (Table 3).   
 89 
Several structural variables can be identified from this preliminary study, which 
contribute to the efficiency of polyesterification and formation of nanocapsules.  Based 
on the molar degree of substitution per anhydroglucose unit, higher molecular weight 
starches are 2.5 - 14.3 fold more reactive when compared to maltodextrins when reacted 
at equivalent mass concentrations of carbohydrate and using the same cross-linker.  This 
increased reactivity suggests that the tertiary structure of high molecular weight 
polysaccharides contribute to either the accessibility or reactivity of glucose alcoholic 
groups. The more sterically hindered diethylmalonoyl dichloride was 10-fold less 
reactive than terephthaloyl chloride as a cross-linker. 
Of the higher molecular weight starches, no differentiation in gross physical attributes 
was observed between the hydroxyethyl and hydroxypropyl substituted corn starches and 
unmodified corn starch.  
 
4. CONCLUSION 
It can be concluded that a sensitive and selective HPLC method  has been developed 
suitable for the analysis of ?C6 dicarboxylic acids hydrolyzed from carbohydrate 
polyesters.  Both aromatic and nonchromophoric diacids can be analyzed using a single 
assay.  This assay can be used for investigation of detailed structural assessment of 
carbohydrate polyesterifcation requirements for producing nanoparticulate drug delivery 
systems. 
 
 
 
 90 
REFERENCES 
Baziramakenga, R., Simard, R. R., &. Leroux, G. D. (1995), Determination of organic 
acids in soil extracts by ion chromatography.  Soil Biology & Biochemistry 27, 349-
356. 
 
Borch, R. F. (1975), Separation of long chain fatty acids as phenacyl esters by high 
pressure liquid chromatography.  Analytical Chemistry 47, 2437-2439. 
 
Capristo, E., Mingrone, G., DeGaetano, A., Addolorato, G., Greco, A. V., & Gasbarrini, 
G. (1999), A new HPLC method for the direct analysis of triglycerides of 
dicarboxylic acids in biological samples.  Clinica Chimica Acta 289, 11-21.  
 
Coenen, A. J. J. M., Kerkhoff, M. J. G., Heringa, R. M., & van der Wal, Sj. (1992), 
Comparison of several methods for the determination of trace amounts of polar 
aliphatic monocarboxylic acids by high-performance liquid chromatography.  Journal 
of Chromatography 593, 243-252. 
 
Docherty K. S., & Ziemann, P. J. (2001), On-line, inlet-based trimethylsilyl 
derivatization for gas chromatography of mono- and dicarboxylic acids.  Journal of 
Chromatography A 921, 265-275. 
 
Durst, H. D., Milano, M., Kikta, E.J., Connelly, S. A., & Grushka, E., (1975), Phenacyl 
esters of fatty acids via crown ether catalysts for enhanced ultraviolet detection in 
liquid chromatography.  Analytical Chemistry 47, 1797-1801. 
 
Fischer, K., Chodura, A., Kotalik, J., Bieniek, D., & Kettrup, A. (1997), Analysis of 
aliphatic carboxylic acids and amino acids in effluents of landfills, composting plants 
and fermentation plants by ion-exclusion and ion-exchange chromatography. Journal 
of  Chromatography A 770, 229-241. 
 
Grushka, E., Durst, H. D., & Kikta, E. J. (1975), Liquid chromatographic separation and 
detection of nanogram quantities of biologically important dicarboxylic acids. 
Journal of Chromatography 112, 673-678. 
 
 91 
Jarowenko, W. (1986).  Acetylated starch and miscellaneous organic esters. In O.B. 
Wurtzburg (Ed.), Modified Starches: Properties and Uses (pp. 55-77).  Boca Raton 
FL: CRC Press. 
 
Kushnir, M. M., Komaromy-Hiller, G., Shushan, B., Urry, F. M., & Roberts, W. L. 
(2001), Analysis of dicarboxylic acids by tandem mass spectrometry.  High-
throughput quantitative measurement of methylmalonic acid in serum, plasma, and 
urine.  Clinical Chemistry 47, 1993-2002. 
 
Langan, R. E. (1986).  Food industry. In O. B. Wurtzburg (Ed.), Modified Starches: 
Properties and Uses (pp. 199-212).  Boca Raton FL: CRC Press. 
 
L?vy, M. C. & Andry, M. C. (1992), Drug targeting and delivery. In T. L. Whateley 
(Ed.), Microencapsulation of Drugs (Drug Targeting and Delivery Series, Vol 1), pp. 
7-16.  London U.K.: Hardwood Academic Publishers. 
 
Mehta, A., Oeser, A. M., & Carlson, M. G. (1998), Rapid quantitation of free fatty acids 
in human plasma by high-performance liquid chromatography.  Journal of 
Chromatography B 719, 9-23. 
 
Naganuma, H., & Kawahara, Y. (1989), Sensitive fluorescence labelling for analysis of 
carboxylic acids with 4-bromomethyl-6,7-methylenedioxycoumarin. Journal of 
Chromatography 478, 149-158. 
 
Toyo?oka, T. (1995), Use of derivatization to improve the chromatographic properties 
and detection selectivity of physiologically important carboxylic acids.  Journal of 
Chromatography B 671, 91-112. 
 
Yamaguchi, M., Iwata, T., Inoue, K., Hara, S., & Nakamura, M. (1990), 6,7-Dimethoxy-
1-methyl-2(1H)-quinoxalinone-3-propionylcarboxylic acid hydrazide: a highly 
sensitive fluorescence derivatisation reagent for carboxylic acids in high-performance 
liquid chromatography.  The Analyst 115, 1363-1366. 
 
 
 
 92 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 1.  Phenacyl ester formation kinetics for terephthalic acid (TA), diethylmalonic 
acid (DEMA), and sebacic acid (SA) at 60?C.  
0
10
20
30
40
0 30 60 90 120
Time (min)
Pe
ak
 A
re
a (
ar
bi
tra
ry
 u
ni
ts 
x 1
0
-5 )
TA
DEMA
SA
 93 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 2.  Chromatogram of dicarboxylic acid ester derivatives from a sample spiked 
with terephthalic acid (TA), diethylmalonic acid (DEMA), and sebacic acid (SA) 
corresponding to 10 ?g/mL of each in hydrolysate with a Maltrin M40 (maltodextrin 
from corn starch) matrix background (upper trace); blank sample containing Maltrin 
M40  matrix background (bottom trace). 
 
 
 
 
 
TA
DEMA
SA
 94 
 
Figigure 3.  Chromatogram of an authentic sample resulting from the hydrolytic 
degradation of a carbohydrate polyester cross-linked with terephthaloyl chloride; sebacic 
acid (SA) is the internal standard; terephthalic acid (TA) is the diacid hydrolysis product.  
 Figigure 4.  Chromatogram of ?C6 dicarboxylic acids demonstrating utility of method for 
other diacids; terephthalic acid (TA), suberic acid (SU), azelaic acid (AZ), diethylmalonic 
acid (DEMA), and sebacic acid (SA). 
 
 
 
 
TA
SA
TA
SU
AZ
DEMA
SA
 95 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 5.  Optical micrograph of nanocapsules fabricated from the emulsion 
polyesterification of terephthaloyl chloride and corn starch.  Nanocapsules are suspended 
in 100 mM phosphate buffered saline.  Each division of the micrometer is 1 ?m. 
 
Table 1.  Calibration curve statistics.  a  parametric standard error, b  linear correlation 
coefficient.   
  
Analyte  Peak Area 
Ratio 
   Peak Height 
Ratio 
  
        
 Slope (SEa) Intercept 
(SE) 
rb   Slope (SE) Intercept 
(SE) 
r 
        
terephthalic acid  0.2981  
(0.00796) 
0.151 
(0.0667) 
0.992  0.4848  
(0.01131) 
0.175 
(0.0948) 
0.994 
diethylmalonic acid  0.2606  
(0.01067) 
-0.066 
(0.0894) 
0.981  0.3275  
(0.01347) 
-0.062 
(0.1129) 
0.981 
sebacic acid  0.0325  
(0.00099) 
-0.00137 
(0.008380) 
0.989  0.0260  
(0.00071) 
-0.00209 
(0.005965) 
0.992 
        
 
 
 
 
 96 
 
Table 2. Accuracy and precision for the determination of terephthalic, diethylmalonic, 
and sebacic acids hydrolyzed from carbohydrate polyesters.  *  N = 5, calculations based 
on peak height ratio from appropriate calibration curves.  SD, standard deviation; RSD, 
relative standard deviation. 
 
 
Analyte Calibration 
Level 
Assay 
Mean* 
SD Accuracy RSD 
 (?g/mL) (?g/mL)  % % 
  
Terephthalic acid 0.5 0.57 0.123 114.1 21.5 
 1.0 1.09 0.145 109.0 13.3 
 5.0 4.87 0.370 97.3 7.6 
 10.0 9.81 1.116 98.1 11.4 
 15.0 15.16 0.926 101.1 6.1 
     
Diethylmalonic 
acid 
0.5 0.52 0.093 104.2 17.9 
 1.0 1.03 0.154 103.1 14.9 
 5.0 4.91 0.651 98.3 13.2 
 10.0 10.02 1.268 100.2 12.7 
 15.0 15.01 2.314 100.1 15.4 
     
Sebacic acid 0.5 0.65 0.225 130.5 34.5 
 1.0 1.09 0.271 109.3 24.8 
 5.0 4.93 0.737 98.6 15.0 
 10.0 9.44 0.900 94.4 9.5 
 15.0 15.39 1.086 102.6 7.1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 97 
Table 3.   Estimation of bifunctional acid cross-linking efficiency in various 
carbohydrate polyesters. 
 
Carbohydrate Diacid Halide Cross-
linker 
Diacid Mass 
Fraction 
Molar Degree of 
Substitution 
of Polyester Per Anhydroglucose 
Unit 
(%) (%) 
maltodextrin from corn 
starch 
terephthaloyl chloride 0.39 0.58 
maltodextrin from corn 
starch 
diethylmalonoyl 
chloride  
0.041 0.044 
maltodextrin from waxy 
maize starch 
terephthaloyl chloride 0.57 0.84 
waxy maize starch 
corn starch 
terephthaloyl chloride 
terephthaloyl chloride 
1.41 
7.85 
2.09 
8.32 
high amylose corn starch terephthaloyl chloride 2.49 3.69 
hydroxyethyl corn starch terephthaloyl chloride 1.73 2.56 
hydroxypropyl tapioca 
starch 
terephthaloyl chloride 1.40 2.08 
 
 98 
 
IV. SYNTHESIS OF CARBOHYDRATE NANOPARTICLES WITH 
HYDROPHILIC CORE 
 
 
ABSTRACT 
A new novel particulate drug delivery platform has been developed based on a 
fundamentally new approach. Existing particulate drug delivery platforms are all 
composed of hydrophobic polymers requiring the use of organic solvents in the internal 
dispersed phase through emulsion-solvent extraction/evaporation techniques.  These 
solvents are frequently found to be incompatible with bioactive molecules especially 
macromolecules.  Additionally, the surface of these hydrophobic polymers give rise to 
particle aggregation as well as complement activation resulting in fast removal of 
particulates from the body by the macrophages of the reticuloendothelial system.  Thus 
systemic use of these platforms for sustained release and/or cellular level intervention is 
not possible.  Other platforms using hydrophilic polymers like gelatin and also 
carbohydrate polymers have been explored. In these platforms the mechanical strength of 
hydrogels have to be reinforced by crosslinking the polymers throughout the device to 
prevent its disintegration upon exposure to aqueous based solvents.  The crosslinking is 
generally performed with very reactive crosslinkers (Yong, S., 2005; Stenekes, R. J. H., 
2000; Can, H. K., 2005) which can freely react with the bioactive molecule incorporated. 
The novelty of this developed platform is that it is produced using a water-in-oil 
emulsion and interfacial crosslinking reaction (emulsion polymerization). The internal 
 98 
aqueous phase containing carbohydrate polymer represents a benign and stabilizing 
microenvironment even for the most labile of bioactive molecules.  Upon removal of 
water, the alcoholic hydroxyl groups of the carbohydrate polymer are expected to provide 
the hydrogen bonding required for the preservation of native biological activity of many 
macromolecules including proteins.  Problems associated with protein denaturation are 
expected to be at least minimized both during synthesis and in the condensed solid phase.  
Since the polymers used in this platform are hydrophilic, complement activation and fast 
removal of the drug delivery system from the body is not anticipated.  Additionally, the 
crosslinkers for interfacial crosslinking are chosen such that they are inactivated by 
reaction with water present in the dispersed phase.  As a result, the crosslinking reaction 
involved in the preparation of the platform, is expected to be a purely interfacial 
phenomenon, keeping the core of the particles undisturbed and unaffected by the 
crosslinking.  In the course of this investigation, preparation methodology of the 
nanoparticulate platform has been developed.  Physical characterization has been carried 
out using both high resolution optical microscopy and transmission electron microscopy.  
Variables affecting the preparation of the particles have been identified and their 
contribution evaluated.  Factors involved with incorporation of model molecules and their 
release from the platforms will be reported separately. 
 
1.  INTRODUCTION 
The idea of using polymeric micro or nanoparticles as drug delivery platforms is 
reasonably well established.  Delivery of drugs using these platforms have many 
advantages including improved bioavailability, stability, control over plasma 
 99 
concentrations, duration of treatment and patient compliance, altering biodistribution 
(Couvreur, P., 1980) and above all maximizing the effectiveness of the drug at the site of 
action and minimizing the systemic side effects.  Drug delivery research has been 
intensely focused on particulate platforms for more than three and a half decades since 
first observations with regards to attaining sustained release of drugs from polymer 
matrices (Yolles, S., 1971) were reported.  Significant progress has been achieved in this 
field, from conceptualization to commercialization of some products delivering small 
molecules and polypeptides in a sustained manner from hydrophobic poly(lactide-co-
glycolide) (PLGA) matrices (Tice, T., 2004).  
But, with current advances in molecular biology, genetics and biotechnology, the 
focus in drug delivery research is slowly shifting from mere sustained release of 
molecules into the blood stream to cellular level intervention in disease conditions.  This 
kind of intervention using particulate platforms essentially requires the platform to 
remain in the systemic circulation for a long period of time.  Limitations of traditional 
particulate platforms made from hydrophobic polymers in terms of aggregation and 
complement activation and their subsequent removal from the blood stream (Verrecchia, 
T., 1995) have excluded their use as systemically circulating delivery devices. They are, 
therefore not suitable for cellular level intervention which requires transport of the 
devices from the blood stream into target tissues.  Surface PEGylation of nanoparticles 
prepared with hydrophobic polymers has been studied (Gbadamosi, J.K., 2002).  It has 
been found that surface PEGylation frequently produces a heterogeneous population of 
particles with differential surface characteristics which is reflected in their physiological 
performance.  When the heterogeneous populations of particles are sorted according to 
 100 
their surface hydrophobicity by using hydrophobic interaction chromatography, a 
remarkable linear correlation between the particle zeta potential and sequestration by 
phagocytotic macrophage-type cells is found.  The development of so called ?stealth? 
particles by covering hydrophobic surfaces with hydrophilic polyethylene polymers has 
only decreased the magnitude of the problem, but has not eliminated it.   
Hydrogel systems composed of hydrophilic polymers have relatively greater 
biocompatibility.  Complement activation is very low to negligible. Some hydrogels have 
been used as plasma expanders demonstrating their safety.  On being introduced into 
water, hydrogel systems swell rapidly losing their mechanical strength and integrity with 
concomitant rapid release of the encapsulated biomolecules (Gombotz, W. R., 1995).  In 
order to control the swelling, attempts have been made to reinforce the hydrogel structure 
by chemical crosslinking in the bulk.  Polyfunctional aliphatic acids (Sediel, C., 2001), 
epichlorohydrin, phosphorous oxychloride, methylene-bisacrylamide (Denizli, B.K., 
2004), and gluteraldehyde (Tabata, Y., 1989) have all been employed to modify 
hydrogels to improve the mechanical properties and decrease swelling which results in 
quick release of the target molecules from hydrogels.  But crosslinking of hydrogels in 
the bulk, especially when they contain labile bioactive compounds like proteins, poses a 
significant problem of denaturation, aggregation or immobilization as a result of the 
crosslinkers reacting with the incorporated bioactive molecule. 
In the context of current approaches and their limitations, a new and novel approach 
has been conceived as the operating principle of this investigation.  The idea is to 
incorporate the bioactive molecules in the core of hydrogel nanoparticles fabricated 
through interfacial crosslinking carried out in a water-in-oil emulsion.  If the crosslinkers 
 101 
are carefully chosen in such a way that they are soluble in the continuous organic phase 
of the emulsion and are deactivated in the dispersed aqueous phase, the crosslinking 
reaction can only occur at the interface where the hydrophilic polymer and the crosslinker 
comes in contact.   
 
Scheme 1.  An illustration of concentration gradients of different component 
molecules in the emulsion polymerization system across the Gibb?s thermodynamic 
interface between aqueous and organic phases.  Bold type represents high concentration 
and small type represents low concentration. 
 
The emulsion interfacial crosslinking/polymerization conception can be better 
understood with an illustration of the principle of Gibb?s thermodynamic interface 
(Scheme-1). As is evident in the diagram, the carbohydrate polymer and buffer 
concentrations are high in the aqueous internal phase and very low in the organic phase.  
The emulsifier, by its nature is mostly concentrated at the interface. Crosslinker acid 
chlorides are in high concentration in the continuum phase and low in the dispersed 
aqueous phase. Moreover, acid chlorides present in the aqueous phase in low 
concentrations are hydrolyzed into the parent acids.  The interface between the 
 Internal Phase 
 Interface Continuum Phase 
Carbohydrate 
 
Buffer / H2O 
 
Crosslinker Acid 
 
Organic solvent  
  
Polymerized 
carbohydrate 
 
emulsifier 
Organic solvent 
 
Crosslinker acid 
chloride 
 
Buffer / H2O 
 
 102 
continuum and dispersed phases is the only region where both the carbohydrate and acid 
chloride crosslinker come in contact with each other in appreciable concentrations. Thus 
the resulting crosslinking reaction predominates at the interface to produce a crosslinked 
membrane at the surface of the nanoparticle.  The membrane, thus produced is expected 
to improve properties like strength, rigidity, and maintain the integrity of the 
nanoparticles as well as the incorporated bioactive molecules.  Additionally, the 
crosslinked membrane is expected to decrease swelling and the rate of diffusion of the 
incorporated bioactive compound out of the particles when resuspended in aqueous 
biological fluids.   
One class of crosslinking reagents having suitable properties is the di or poly 
carboxylic acid chlorides.  These acid chlorides are soluble in organic solvents used as 
the continuum phase of the emulsion and decompose on exposure to water to the 
corresponding acid and hydrogen chloride.  They can however, react to the ubiquitous 
hydroxyl groups of the carbohydrate at the interface of the emulsion droplets to produce 
ester crosslinks.  In this work, various organic polycarboxylic acid chlorides have been 
investigated as potential crosslinkers. The degree of substitution achieved has been 
determined in each case using high pressure liquid chromatography (HPLC). Three 
crosslinkers, terephthaloyl chloride, trimesoyl chloride and diethylmalonyl chloride have 
been found as potentially ideal candidates for this purpose.  Attempts have been made to 
characterize the reaction for optimal conditions in terms of pH of the aqueous phase, 
concentration of the crosslinker in the organic continuum phase, concentration of the 
polymer in the dispersed aqueous phase, reaction time and molar concentration of the 
buffers used.   
 103 
The influence of different starches or modified starch polymers on the extent of 
crosslinking has also been investigated.  Specifically, commercially available corn starch, 
amylopectin (waxy maize starch), hydroxyethyl starch and maltodextrins from corn and 
waxy maize starches have been investigated as polymers for the drug delivery system.  
An attempt was made to determine which of the two constituent polymers of starch, 
amylopectin or amylose was more suitable for the nanoparticle formation. Amylose was 
derivatized with dihydroxypropyl group to improve its solubility.  The derivatized 
amylose, henceforth referred to as dihydroxypropyl amylose (DHPA), was then used as a 
polymer for nanoparticle synthesis.  The dihydroxypropyl group was specifically chosen 
for derivatization due to their toxicological safety (Langan, R.E., 1986) and favorable 
properties.  
 
2. EXPERIMENTAL 
2.1 Materials 
Corn starch and hydroxyethyl corn starch (Sigma, St. Louis, MO), high amylose corn 
starch (C*Amylogel 03003, Cerestar USA Inc., Hammond, IA), waxy maize starch (PFP 
2850, Cargill Foods, Minneapolis, MN), hydroxypropylated starch modified from corn 
starch (C*Aratex 7570, Cerestar USA Inc., Hammond, IA) and maltodextrin (MA) from 
waxy maize starch (C*Dry 01955, Cerestar USA Inc.), maltodextrin (MCS) from corn 
starch (Maltrin M40, Grain Processing Corporation, Mucatine, IA) and amylose (City 
Chemicals, West Haven, CT) were used as supplied.  Crosslinking acid chlorides include 
terephthaloyl chloride (TC) and diethylmalonyl dichloride (DEMC) (Sigma-Aldrich, St.-
Louis, MO); trimesoyl chloride (TMC), sebacoyl chloride (SC) and oxaloyl chloride 
 104 
(OC) (Acros Organics, Pittsburg, PA).  Other reagents include glycidol, buffer salts 
(Sigma-Aldrich, St. Louis, MO); chloroform, cyclohexane, acetone, ethanol, 
triethylamine (TEA), phenacyl bromide (PB) and dimethyl formamide (DMF) (Fisher 
Scientific, Pittsburg, PA).  All organic solvents were of ACS grade.  Surfactants used 
were Tweens 40, 65, 80, and 85 (Fisher Scientific), nonyl phenyl ethoxylate (IgepalCO-
520) and Spans 20, 40, 60, and 80 (Sigma-Aldrich, St. Louis, MO). 
 
2.2 Synthesis of dihydroxypropyl amylose (DHPA) 
  Amylose derivatization was carried out according to a reported procedure (Wulff, G., 
1998).  Two (2) grams of amylose (average molecular weight 150,000 Daltons) was 
dispersed in 100 mL of 1(N) NaOH in a screw capped 1 L bottle after purging the alkali 
solution thoroughly with nitrogen and the cap tightly closed.  The amylose is stirred on a 
magnetic stirrer for 2 hr or until the suspension becomes translucent.   At this point, the 
bottle was opened, 6 ml of 2, 3-epoxy-1-propanol (glycidol, ~81 mMolar in reaction 
medium) was introduced and the bottle closed after thoroughly purging with nitrogen 
again.  The reaction was allowed to continue overnight at room temperature.  The 
reaction was stopped by neutralizing the base with HCl.  The derivatized amylose was 
then precipitated by pouring the neutralized reaction mixture into 300 mL of ethanol.  
After 30 min, the precipitated derivatized amylose was filtered and redissolved in 
demineralized water.  Alternate dissolution in demineralized water and precipitation in 
ethanol was repeated three times to eliminate the unreacted glycidol.  Finally, the product 
was dried overnight in a vacuum oven at 54?C.   
 
 105 
2.3 Emulsion stability 
In order to obtain a stable emulsion, various water-in-oil systems were investigated.  
Cyclohexane, chloroform and cyclohexane/chloroform (68/32, v/v) mixed solvent 
systems were used as the continuum phase and aqueous buffers as the dispersed phase.  
Emulsion stability was investigated in pH 7.0 and 10.0 and a buffer strength of 200 mM.  
The volume ratio between continuum phase and dispersed was always maintained at 5:1.  
The surfactants Spans 40, 65, 80, 85 and nonyl phenyl ethoxylate were added in the 
continuum phase while the Tweens 20, 40, 60 and 80 were added in the dispersed phase 
because of their water solubility.  Surfactants or surfactant combinations were evaluated 
in a concentration range of 1.0 to 6.0 % (w/v).  The two phases, 20 mL aqueous and 100 
mL organic, were mixed by homogenization (Silverson LR-4) at 5,000 rpm for one 
minute to make the emulsion and left standing at room temperature for stability studies.  
The stability was investigated visually every hour for 6 hr and thereafter at 12 hr and 24 
hr.  The most stable of emulsions was chosen for nanoparticle preparation.  
 
2.4 Synthesis of carbohydrate nanoparticles 
Carbohydrate polymers and oligomers were dispersed in appropriate buffer and 
heated in screw capped bottles to produce carbohydrate colloids for high molecular 
weight polymers and true solutions for maltodextrins.  The concentration of the polymers 
varied from 4.5 to 50% (w/v) according to the specific polymers.  An attempt was made 
in each case to achieve a near saturation concentration of each polymer when cooled to 
room temperature.  The effect of pH on the synthetic reaction was evaluated by buffering 
the colloid/solution to appropriate pHs.  Buffer strength was maintained at 200 mM 
 106 
except where the effect of buffer strength on the degree of substitution was investigated.  
Phosphate, acetate and borate buffers were used for maintaining pHs ranging from 3-11.  
The colloid/solution was allowed to cool to room temperature while stirring before 
further use.   
The continuum phase was composed of 68% cyclohexane and 32% chloroform (v/v).  
Six (6) percent (w/v) of the surfactant Span-40 was dissolved in the continuum phase to 
prepare an organic phase designated as OP-1, which was subsequently used for 
emulsification.  Twenty ml of the aqueous phase was poured into 100 mL of OP-1 and 
homogenized in a Silverson-LR4 homogenizer with a rotor-stator arrangement at 5,000 
rpm for 1 min.   
Terephthaloyl chloride (TC), trimesoyl chloride (TMC), diethylmalonyl chloride 
(DEMC), sebacoyl chloride (SC) or oxaloyl chloride (OC) was added to the 
cyclohexane/chloroform (68/32, v/v) in appropriate concentrations.  Thus, a final 
concentration of 55 mM was achieved to prepare the organic phase designated as OP-2, 
except where mentioned otherwise. The crosslinker concentration was varied from 0.8 to 
161 mM in the continuum phase in batches made for evaluating the effect of crosslinker 
concentration on degree of substitution.  Two hundred (200) mL of OP-2 was added to 
the prepared emulsion as described and the reaction was allowed to continue under gentle 
stirring using a magnetic stirrer for an appropriate time.  The reaction was stopped by 
pouring the reaction mixture into 700 mL ethanol in a 1000 mL beaker while stirring.  
The stirring was discontinued after 30 min and the formed nanoparticles slowly 
precipitated.  Next the supernatant was decanted, the precipitate sedimented for 2 hr, and 
the collected precipitate was washed three times with ethanol followed by three washes 
 107 
with acetone by repeated centrifugation at 1,500 g and resuspension by vortexing.  
Finally, the suspension of nanoparticles in acetone was dried under vacuum at 14.5 psi 
pressure at room temperature overnight. 
 
2.5 Physical characterization 
2.5.1 Particle size 
The particle size was determined using photon correlation spectroscopy with a 
Nicomp model 380 (Particle Sizing Systems, Santa Barbara, CA) particle size analyzer.  
Average particle sizes based on population distributions were determined using Nicomp 
CW380v1.51A software.  Measurements were made with a very dilute nanoparticle 
suspension (~40 ?g/ml) in deionized water.  Mean diameters were calculated from the 
volume and number averaged size based on discrete mono disperse distributions.  Particle 
size measurements corresponded to the swollen mean diameters of the nanoparticles.  
 
2.5.2 Visualization 
Visualization of the nanoparticles and particle morphology was accomplished using 
transmission electron microscopy (TEM) and optical microscopy.  For TEM a very dilute 
suspension (~40 ?g/ml) of nanoparticles was made in deionized water. A drop of the 
nanoparticle suspension was placed on a copper grid covered with a polyvinyl alcohol 
film.  The solvent was evaporated in vacuum.  A microphotograph was then taken with 
an electron microscope (Zeiss, Model EMCR-10, Germany). 
Since the nanoparticles can be visualized by TEM in the dry state only, visualization 
in aqueous suspension was achieved using an optical microscope (Olympus, Model BX-
 108 
51, Japan) used in the bright field mode.  A digital camera (Olympus, Model CMAD-3, 
Japan) was used to capture the image. 
 
2.6 Chemical characterization 
Nanoparticles were chemically characterized in terms of the crosslinker content per 
unit carbohydrate mass and the molar degree of substitution per 100 anhydroglucose 
units.  The molar degree of substitution (DS) was determined using the following 
equation. 
                                         DS CL MolWtCLCL= ?(% / . . )*( % )/ . 100100 16214  
 
Here DS indicates number of crosslinker molecules reacted to each 100 anhydroglucose 
units and % CL is the (w/w) % of the crosslinker determined in blank nanoparticles when 
dry.  Crosslinker di- and tri- carboxylic acids were quantitatively analyzed by an HPLC 
method developed for this purpose (Basu Sarkar, A., 2005) after hydrolysis of the 
crosslinked polymers.  Briefly, required quantities of nanoparticles were suspended in 5 
M ammonium hydroxide solution in a screw capped vial and heated to 130?C in a dry 
bath for 1 hour to hydrolyze the ester bonds between the carbohydrate and the 
crosslinkers.  The vials were then allowed to cool down to room temperature, 1 mL 
aliquot of the hydrolysate was transferred to a 15 mL glass test tube and a 9 mL aliquot of 
60 mM TEA in acetonitrile was added.  After mixing the contents, the carbohydrate was 
precipitated by centrifugation at 2500 g for 30 min.  One (1.0) mL of clear supernatant 
was transferred into a screw capped vial.  Solvent was evaporated and the dry samples 
 109 
used for derivatization.  Derivatization was carried out by adding 100 ?L acetonitrile, 300 
?L of 60 mM TEA solution in acetonitrile, 300 ?L of 200 mM PB solution in acetonitrile 
and 300 ?L DMF and then heating at 60?C for 1 hour.  The derivatized acids were further 
purified in a PrepSep C18 (Fisher) solid phase separation cartridge (SPE) (500 mg 
adsorbent, 6 mL volume).  The SPE cartridge was conditioned by passing 4 mL deionized 
water prior to use. The derivatized sample was diluted with 2 mL deionized water and 
then passed through the cartridge.  The sample was washed with 4 mL of distilled water 
and finally, the sample was eluted with 1 mL of acetonitrile into an autosampler vial for 
HPLC analysis. 
 
3. RESULTS 
3.1 Physical characterization 
3.1.1 Particle Size 
Particle size analysis showed that carbohydrate nanoparticles synthesized by 
emulsion/polymerization were generally in the size range of 100 to 300 nm using the 
specified conditions.  The nanoparticle suspension in deionized water exhibited a single 
peak (Figure 1).  The abscissa in this figure is in log scale.  A close correlation was 
observed among the number weighed, volume weighed and intensity weighed values 
indicating a mono disperse particle distribution of a single population of particles. 
 
3.1.2 Particle Morphology 
Both optical and transmission electron microscopy were employed to investigate 
nanoparticle morphology.  Optical microphotographs (Figure 2a-2f) indicate that 
 110 
nanoparticles can be synthesized with all the carbohydrate polymers investigated.  Fully 
hydrated particles could be seen as distinct structured vesicles with spherical shape.  
When suspended in distilled water, the particles suspended nicely, showing a 
mononucleated suspension.  This observation, coupled with the observation that when 
particles were put into a non polar organic phase like cyclohexane or chloroform, they 
show a highly aggregated dispersion (photograph not shown), indicates that the surface of 
the nanoparticles is hydrophilic in nature as expected.  The nanoparticles remained stable 
in aqueous suspension for at least 4 weeks (Figure 2e and 2f).  Nanoparticles prepared 
from DHPA were initially distinct structured vesicles when suspended in water but 
dissolved completely in two days indicating that the DHPA particles were too water 
soluble.  Another important observation from optical microscopy was that in some cases 
the nanoparticles synthesized from TC batches showed dark crystal-like spots within the 
particles.  Terephthalic acid is known to be very insoluble at neutral pH. It was thus 
postulated that some terephthalic acid, generated as a byproduct of the reaction with 
water, accumulated within the nanoparticles and was not extracted during the washing 
process with ethanol and/or acetone.  This was confirmed by the observation that the dark 
spots disappeared on addition of alkali in the suspension media.  Transmission electron 
microscopy (TEM) (Figure 3a and 3b) shows slightly distorted spheres which result from 
the contraction of carbohydrate polymers in dehydrated particles.  These observations 
also indicate partially hollow sphere cores.  Additionally, the size estimates from TEM 
closely matches with the size range obtained from photon correlation spectroscopy. 
 
 
 111 
3.1.3 Emulsion stability 
Stable emulsions were found using Span 40 and Span 60 at a 3.0 % (w/v) 
concentration in cyclohexane, chloroform and cyclohexane/chloroform (68/32, v/v) as the 
continuum phase.  The cyclohexane/chloroform mixed solvent system was stable over 
days probably because of the density matching with the dispersed phase at the chosen 
proportions of cyclohexane and chloroform. 
 
3.2 Chemical Characterization 
3.2.1 Relationship between polymer species and degree of substitution (DS) of 
TC 
Synthesized nanoparticles, using DHPA, amylomaize starch, corn starch, 
amylopectin, HES, HPS, MCS and MA as condensing polymers, were evaluated.  Each 
polymer or oligomer was reacted with TC for 3 hr at pH 10.0.  The concentration of 
polymers in the aqueous phase was kept near saturation (Table 1).  Initially, pH 10.0 was 
chosen for poly esterification since other workers found pH 9.8 to be optimal for similar 
reactions (Levy, M. C., 1991).   
Results show that nanoparticles can be prepared with all the carbohydrate polymers 
and oligomers including MCS and MA.  It was found that the lower molecular weight 
carbohydrate polymers, like MCS and MA (~3,200 Daltons), resulted in lower degrees of 
substitution with values of 0.58 and 0.84 per 100 anhydroglucose units.   Higher 
molecular weight polymers all showed higher DS with values ranging from 2.08 for HPS 
to 8.32 for corn starch.  Generally, amylose rich starches showed higher DS.  The DS of 
TC with DHPA was 5.57, followed by amylomaize starch (~70% amylose) with a value 
 112 
of 3.69, and waxy maize starch (no amylose) with a value of 2.09.  The primary alcoholic 
groups of HES and HPS did not show any higher reactivity in terms of DS contrary to 
expectations.  The very high reactivity of corn starch (DS of 8.32) could not be explained. 
 
3.2.2 Effect of crosslinker species on DS 
Five different di or tri functional acid chlorides, both aliphatic and aromatic, have 
been investigated for their suitability for this emulsion- interfacial crosslinking reaction.  
The acid chlorides include OC, SC, DEMC, TC and TMC.  For comparison of reactivity 
all crosslinkers were evaluated at a 55 mM concentration in the organic continuous phase.  
When bifunctional linear aliphatic crosslinkers including OC and SC were used, 
nanoparticles could not be prepared since the emulsion broke and the phases separated.  
These results suggest that the surfactant sorbitan monopalmitate, which has alcoholic 
hydroxyl groups, react with the acid chlorides. Esterified surfactant thereby loses its 
amphipathic character bringing about phase separation. When sterically hindered 
aromatic di and tri functional acid chloride were used, emulsions were stable and 
nanoparticles could be formed.   
Structural comparison shows that both the TC and TMC have a bulky benzene ring in 
between the reactive functional groups. From these observations it was hypothized that 
the bulky benzene ring produces enough steric hindrance with surfactant molecules to 
prevent them from reacting and thus prevent phase separation.  In order to validate this 
hypothesis, DEMC, an aliphatic acid chloride crosslinker with two ethyl side chain was 
selected for evaluation.  When DEMC was used as the crosslinker, indeed, no phase 
separation was detected.  Thus, it could be established that one of the structural 
 113 
requirements for the di or tri functional crosslinker to work in this system is that a bulky 
group/side chain has to be present in between the acid chloride groups of the crosslinkers 
to produce enough steric hindrance and prevent the crosslinker from reacting with the 
surfactant. 
 
 
Cl
Cl
O
O Cl (CH2)8 Cl
O O
Cl
C2H5C2H5
Cl
OO
 
          OC                                              SC                                    DEMC 
      Cl
Cl O
O                                      
Cl
O
Cl
Cl
O
O     
                      TC                                                                                  TMC  
In order to assess the relative reactivity of the different crosslinkers, nanoparticle 
batches were made using MCS and DHPA as polymers and TC, TMC and DEMC as 
crosslinkers at pHs of 7.0 and 10.0.  The crosslinker concentration in each case was fixed 
to 55 mM in the external organic phase.  The reaction time in each case was fixed for 3 hr 
except for batches at pH 10.0 using DHPA as the polymer, in which case, the batches 
were reacted for 24 hours.   
 114 
DEMC was found to be the least reactive crosslinker (Table-2).  At pH 7.0, its 
reactivity ranged from 20-fold to 17-fold less for MCS and DHPA, respectively, than TC 
with the same polymers.  At pH 10.0, the difference was less pronounced, such that the 
reactivity of DEMC ranged from about 10-fold and 5-fold less for MCS and DHPA, 
respectively, than TC.  The number of crosslinker molecules reacted to 100 
anhydroglucose units was relatively higher for batches with DHPA at pH 10.0 because 
those batches were reacted for 24 hr instead of 3 hr as in the case of other condensing 
reactions. 
TMC was found to be the most reactive crosslinker.  At pH 7.0, the reactivity of TMC 
was 1.54 to 1.38 times of that of TC with MCS and DHPA, respectively.  The same trend 
in reactivity was also observed at pH 10.0.  At pH 10.0, the reactivity of TMC was about 
1.6 times greater that of TC with MCS and DHPA, respectively.  The reactivity 
difference between the polymers DHPA and MCS were not pronounced at short reaction 
times.  
 
3.2.3 Effect of pH on the degree of substitution 
The effect of pH on the degree of substitution was investigated over a wide range of 
pH from 3 to 11.  To avoid polymer specific trends, two different polymers, MCS and 
DHPA, were used for this purpose.   
Esterification with TC (Figure 4 & 5) was found to be independent of pH between 
values of 7 and 10 exhibiting evidence of base catalysis above pH 10 and acid catalysis 
below pH 7.  Esterification with TMC (Figure 6) was also generally independent of pH 
between pH values of 3 to about 7.  Thereafter base catalysis contributed to increasing 
 115 
DS.  Esterification with DEMC (Figure 7) was independent of pH from values of 3 to 9 
with base catalysis indicated at pH 10.  A precipitous decline in DS was observed at pH 
11 for TMC condensed with MCS, and DEMC condensed with DHPA. These results 
indicate that ester formation is overwhelmed by ester hydrolysis at pH ? 11.  Similar 
results are reported elsewhere (Levy, M. C., 1991), indicating that the optimal pH for 
nanoparticle synthesis is pH 9.8.  This investigation though, shows a much broader range 
of pH for optimal nanoparticle formation.  Both acid and base catalysis, as observed in 
course of this investigation, are typical of ester formation reactions. 
 
3.2.4 Effect of buffer strength on degree of substitution 
Two series of nanoparticle batches were made with TC and TMC (Figure 8) with 
varying buffer strength in the dispersed aqueous phase.  The TMC batches were made at 
pH 7.0 while the TC batches were made at pH 10.0.  It was found that TMC reactivity 
was independent of the buffer strength over a wide range of concentrations.  The 
reactivity of TC was found to be somewhat dependent on buffer strength at the low end.  
The reactivity of TC was found to be 5-fold less at buffer strength of 50 mM than 100 
mM.  In the range of 100 to 200 mM the reactivity of TC remained essentially the same.  
  
3.2.5 Effect of polymer and crosslinker concentrations on the degree of 
substitution 
Nanoparticle batches were made with TMC and DHPA at pH 7.0 and three different 
polymer concentrations of 1.5, 3.0 and 4.5 % (w/v of the aqueous phase).  Results (Figure 
9) show that the number of crosslinker molecules reacted to 100 anhydroglucose units 
 116 
increases from 2 for a polymer concentration of 1.5% (w/v) to 2.9 for a polymer 
concentration of 3% (w/v) and finally to 4.6 for a polymer concentration of 4.5%.  A 
concentration of 4.5% (w/v) is the maximum amount of the dihydroxypropyl derivatized 
amylose that can be dissolved in water. 
Results from experiments carried out to establish the effect of crosslinker 
concentration on the degree of substitution were similar.  Nanoparticle batches were 
made with MA as the polymer and TMC as the crosslinker under identical conditions.  
The only variable was TMC concentration which was varied from 0.8 to 162 mM.  
Results (Figure 10) show that the number of crosslinker molecules reacted to each 100 
anhydroglucose units increased moderately from 0.03 for 0.8 mMolar TMC concentration 
to 1.07 for 80.86 mM TMC concentration and finally to 1.55 for 162 mM TMC 
concentration (Figure 10).  The relationship between DS and crosslinker concentration 
was hyperbolic in the concentration range of 0.8 to 162 mM.  These results indicate that 
carbohydrate polyesterification is dependent on both carbohydrate and crosslinker 
concentrations and suggest that the chemical kinetics is second order.  Furthermore, in 
agreement with Le Chatelier?s principle, the extent of reaction can be forced by 
increasing both crosslinker and carbohydrate concentrations.  In view of the limited 
solubility of the higher molecular weight polymers, maltodextrins are therefore 
considered the preferred form of the carbohydrate reactant species for maximal 
condensation of the carbohydrate polyester.  Nanoparticles formed with condensed 
maltodextrins will also contain the highest degree of carbohydrate at the core and thereby 
maximal solid state stability of the encapsulated macromolecules.  In addition, product 
variability can be minimized by standardizing the reaction with a saturated solution of 
 117 
maltodextrin in the dispersed phase and varying the crosslinker concentration to obtain 
the desired performance characteristics. 
 
3.2.6 Reaction time profile  
The time frame for the completion of the reaction was established using DEMC, TC 
and TMC crosslinkers and MCS as the polymer.  Reactions were carried out under 
identical conditions with a 55 mM crosslinker concentration and 50% (w/v) polymer 
concentration.  The only variable across the batches was reaction time for each series.  
The reaction time was varied from 1 to 48 hr for each series.  The reaction was found to 
be complete in 12 hr for DEMC batches (Figure 11).  For the TC and TMC batches, 
however the reaction was found to be complete in 24 hr (Figure 10).   
 
4.  DISCUSSION 
Starch is a physical mixture of two different polymers; amylose and amylopectin.  
Amylose (Scheme 2) is a linear polymer composed of (1-4)- ?-D glucopyranosyl units 
while amylopectin is a branched polymer consisting of both (1-4)- ?-D glucopyranosyl 
units with occasional branching with (1-6)- ?-D glucopyranosyl units.  Because of the 
presence of a large number of alcoholic groups, starch polymers are hydrophilic (Domb, 
A., 1990).  Additionally, the presence of a large number of reactive hydroxyl groups also 
presents an opportunity to crosslink starch through a multitude of different crosslinkers.  
The biodegradable nature of starches makes them one of the most suitable for particulate 
injectable DDS.   
 118 
 
2 (a).  
 
2 (b). 
Scheme 2.  Representative molecular structure of amylose (a) and amylopectin (b). 
Though very similar in their molecular structure, amylose and amylopectin have very 
different properties. Amylose is sparingly soluble in water and retrogrades with time.  
Synthesis of nanoparticles with amylose is therefore, problematic.  The reactivity and DS 
of this polymer may vary with the state of retrogradation which cannot be controlled.  
Amylose also produces inclusion complexes with a wide variety of molecules (Hui, Y., 
1991; Kubik, S., 1996). The formation of inclusion complexes is not important for the 
synthesis of nanoparticles by itself, but it may have an important effect on encapsulation 
efficiency of different bioactive molecules.    
 119 
Amylopectin on the other hand has relatively greater solubility and a relatively low 
grade of retrogradation.  These properties make amylopectin a more suitable candidate 
for emulsion polymerization reaction.  The fact that amylopectin is a branched polymer 
may also be helpful for the formation of a three dimensional crosslinked mesh membrane 
at the interface.  A lower number of inter chain crosslinks should be required to achieve 
the same level of crosslinking density, as that with amylose.  Crosslinking density is 
important for particle integrity in aqueous suspension and also for control of the release 
rate of incorporated bioactive molecules. 
In addition to the physical integrity of the nanoparticles in physiological fluid, and the 
ability to incorporate and release bioactive molecules from the nanoparticles, several 
other physico-chemical properties are required for their use as circulating nanoparticles.   
For systemic use, the nanoparticles need to have the right surface polarity to be 
suspended in aqueous physiological fluid as discrete, mononucleated units.   Otherwise 
they might aggregate to form bigger particles which might cause thrombosis in the 
smaller blood vessels. The fact that the nanoparticles synthesized in the course of this 
work, can be suspended as discrete, mononucleated particles is a major development 
toward their eventual use in the systemic circulation.  The mono disperse nature of the 
produced nanoparticles is also helpful in controlling various important features including 
release rate.   
The molecular weight of the starch polymers used for the synthesis of the 
nanoparticles is also important.  The high molecular weight starches have low water 
solubility resulting in low polymer load of the dispersed aqueous phase.  A low polymer 
load in the dispersed aqueous phase may affect the integrity of the nanoparticles and 
 120 
more importantly, can potentially affect the encapsulation efficiency as a result of limited 
capacity of physical entrapment of the bioactive molecules by the polymer.  High 
molecular weight starch polymers also show retrogradation.  The effect of retrogradation 
of the starch polymers on their reactivity and DS has not been studied in the course of 
this work.  But as the retrogradation process involves a change of the specific stereo 
chemical structure (Fechner, P.M., 2005) of the polymer, it is anticipated to have some 
effect on the reactivity and DS.  In that case it will be impossible to precisely control the 
reaction with high molecular weight starch polymers if reacted during the retrogradation 
process.  The low molecular weight maltodextrins, on the other hand, have two useful 
properties.  They are much more soluble in water producing a clear true solution in as 
high as 50% (w/v) concentration and they do not retrograde.  These two helpful 
properties outweigh their limitations regarding their relatively lower reactivity.  Among 
the maltodextrins, MA is the more suitable polymer to be used for the synthesis of the 
nanoparticles as they are prepared from waxy maize starch (>99% amylopectin) and thus 
are branched to begin with.  The problems associated with both polymer load and 
retrogradation can thus be addressed by using MA. 
Another important feature of the nanoparticles to be used for systemic circulation is 
the size of the nanoparticles.  As reported, nanoparticles could be prepared in the range of 
100-300 nm (within colloidal size range).  This size range is a very important property of 
the prepared particles for several reasons.  Firstly, at this size range the particles have 
Brownian motion.  This has been observed and recorded in a video.  As a result of this 
Brownian motion the particles are not expected to sediment in circulating fluid.  The 
phagocytotic uptake and removal of foreign particles are known to be size dependent 
 121 
(Prabha, S., 2002) with minimum phagocytotic activity at the 100 nm size range.  Thus 
the produced nanoparticles are expected to stay in the systemic circulation for a long 
period of time.  Also nanoparticles in this colloidal size range can potentially escape the 
endothelial wall of the blood vessels especially in the areas of diseased patho-physiology.  
This particular feature of these nanoparticles exposes them to the possibility of cellular 
uptake through adhesion, cell fusion, passive endocytosis or active targeting. 
The hydrophilic nature of the carbohydrate polymers are expected to be altered by the 
incorporation of the crosslinkers at the surface of the nanoparticles.  In this regard, the 
two chosen crosslinkers have some theoretical difference.  Terephthalic acid is a 
relatively non polar dicarboxylic acid with pKa1 and pKa2 of 3.51 and 4.82, respectively.  
But even at neutral pH, where both the acid groups are ionized, it is sparingly soluble in 
water.  On the other hand, the more polar trimesic acid is a tri carboxylic acid with pKa1 
of 2.98 (other two pKa?s are unknown), and is very soluble over a wide range of pH?s.  
Thus, the nanoparticle batches produced using these two crosslinkers are expected to 
have a difference in polarity.  A difference in polarity and wetting by water is indeed 
observed. Properly suspending the nanoparticles prepared with TC was more difficult; 
they tended to float on water due to lower polarity and higher surface tension.  The TMC 
batches, on the other hand could be suspended readily.  In the case of TMC, the 
probability of all three carboxylic acids crosslinking with carbohydrate is low.  Thus the 
residual unreacted free carboxylic acids are believed to be partially responsible for 
making the surface hydrophilic reducing the interfacial tension in aqueous dispersed 
medium.  The other favorable features of TMC are higher reactivity over a wide range of 
pH and near independence of reactivity on pH. 
 122 
5. CONCLUSION 
Several conclusions can be arrived at from these investigations.  Firstly, carbohydrate 
nanoparticles can be made using the chosen emulsion-interfacial crosslinking method in 
the colloidal range required for their use in the systemic circulation.  The hydrophilic 
carbohydrate polymers are ideally suited for this purpose.  The nanoparticles produced 
have ideal surface properties of polarity for being suspended as discrete, mono dispersed, 
mono nucleated particles in aqueous media. Among all the carbohydrate polymers 
investigated maltodextrin from waxy maize starch has the most favorable properties for 
the preparation of the nanoparticles and their loading.  Among the crosslinkers, only 
terephthaloyl chloride and trimesoyl chloride showed promise.  Trimesoyl chloride is 
better suited because of higher reactivity, reactivity over a wide range of pH and lastly, 
but perhaps most importantly, because of the hydrophilicity it imparts to the surface of 
the nanoparticles. 
 
REFERENCES 
Basu Sarkar, A. and Kochak, G. M.  2005.  HPLC analysis of aliphatic and aromatic 
dicarboxylic acid cross-linkers hydrolyzed from carbohydrate polyesters for 
estimation of molar degree of substitution.  Carbohydrate polymers, 59, 305-312. 
 
Can. H. K., Denizli, B. K., Guner, A. and Rzaev, Z. M. O.  2005.  Effect of functional 
crosslinking agents on preparation and swelling of hydrogels.  Carbohydrate 
Polymers, 59, 51-56. 
 
Couvreur, P., Kante, B., Lanaerts, V., Scailteur, V., Roland, M. and Speiser, P.  1980.  
Tissue distribution of antitumor drugs associated with polyalkylcyanoacrylate 
nanoparticles.  Journal of Pharmaceutical  Science, 69, 199-202. 
Denizli, B. K., Can, H. K., Rzaev, Z.M.O. and Guner, A.  2004.  Preparation conditions 
and swelling equilibria of dextran hydrogels prepared by some crosslinking agents.  
Polymers, 45, 6431-6435. 
 
 123 
Domb, A., Davidson, G. W. and Sanders, L. M.  1990.  Diffusion of peptides through 
hydrogel membranes.  Journal of Controlled Release, 14, 133-144. 
 
Fechner, P. M., Wartewig, S., Kleinebudde, P. and Neubert, R. H.  Studies of the 
retrogradation process for various starch gels using Raman spectroscopy. 
Carbohydrate Research, 340, 2563-2568. 
 
Gbadamosi, J. K., Hunter, A. C. and Moghimi, S. M.  2002.  PEGylation of microspheres 
generates heterogeneous population of particles with differential surface 
characteristics and biological performance.  Federation of European Biological 
Societies Letters, 532, 338-344. 
 
Gombotz, Wayne R., Petit, Dean K.  1995.  Biodegradable polymers for protein and 
peptide drug delivery.  Bioconjugate Chemistry, 6, 332-351. 
 
Hui, Y. and Zou, W.  1991.  Helical inclusion complexes and reactivity of amylose and 
its derivatives.  Schneider, H. J. (Ed).  Frontiers in Supramolecular Organic 
Chemistry and Photochemistry, VCH, Weinheim, Pp. 203-221. 
 
 Kubik, S. and Holler, O., Steinert, A., Tolksdorf, Y., Van der Leek, Y. and Wulff, G.  
1996.  Molecular inclusion within polymeric carbohydrate matrices.  Van Bekkum, 
H., Roper, H. and Voragen, F. (Eds).  Carbohydrates as Organic Raw Materials, 3, 
Pp. 169-187. 
 
Langan, R. E.  1986.  Modified Starches: Properties and Uses. Pp-199-212.  Ed. 
Wurzburg, O. B. CRC Press, Boca Raton, FL, Pp. 56-89. 
 
Levy, M.C. and Andry, M.C.  1991.  Microcapsules with walls made of cross-linked 
starch derivatives.  Microencapsulation of Drugs, Ed.-Whatley, T. L., Drug Targeting 
and Delivery Series, Harwood Academic publishers, Great Britain, Pp. 7-16.   
 
Pariot, N., Edwards-Levy, F., Andry, M. C. and Levy, M. C. 2000. Crosslinked ?-
cyclodextrin microcapsules: Preparations and properties.  International Journal of 
Pharmaceutics, 211, 19-27. 
 
Probha, S., Zhou, W. Z., Panyam, J. and Labhsetwar, V.  Size dependency of 
nanoparticle mediated gene transfection: Studies with fractionated nanoparticles.  
International Journal of Pharmaceutics, 244, 105-115. 
 
Seidel, C., Kullicke, W. M., Heb, C., Hartmann, B., Lechner, M. D. and Lazik, W.  2001.  
Influence of Cross-linking Agent on the Gel Structure of Starch derivatives.  Starch 
Starke, 53, 305-310. 
 
 124 
Stenekes, R. J. H., Loebis, A. E., Fernandes, C. M., Crommelin, D. J. A. and Hennink, 
W. E.  2000.  Controlled release of liposomes from biodegradable dextran 
microspheres: A novel delivery concept.  Pharmaceutical Research, 17, 690-695. 
 
Tabata, Y. and Ikada,Y.  1989.  Synthesis of gelatin microspheres containing interferon.  
Pharmaceutical Research, 6, 422-427. 
 
Tice, Thomas (2004).  Delivering With Depot Formulations.  Drug Delivery Technology, 
4, 44-47. 
Verrecchia, T., Spenlehaur, G., Brazile, D. V., Murry-Brelier, A., Archimbaud, Y. and 
Veillard, M. V.  1995.  Non-stealth (poly(lactic acid/albumin)) and stealth (poly(lactic 
acid-polyethylene glycol)) nanoparticles as injectable drug carriers.  Journal of 
Controlled Release, 36, 49-61. 
 
Wulff, G., Steinert, A. and Holler, O.  1998.  Modification of amylose and investigation 
of its inclusion behavior.  Carbohydrate  Research, 307, 19-31. 
 
Yolles, S., Eldridge, J. E and Woodland, J.H.R.  1971.  Sustained delivery of drugs form 
polymer drug mixtures.  Polymer News, 1, 9. 
 
Young, S., Wong, M., Tabata, Y. and Micos, A. G.  2005.  Gelatin as a delivery vehicle 
for the controlled release of bioactive molecules.  Journal of Controlled Release, 109, 
256-274. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 125 
 
   
Table 1.  Comparison of degree of substitution by terephthalic acid in preparation of  
nanoparticles with different carbohydrate polymers. 
 
*Number of crosslinker molecules per 100 anhydroglucose units 
 
     
Table 2. Comparative reactivity of the crosslinkers TMC, TC, DEMC at pHs 7.0 and 
10.0 with polymers MCS and DHPA 
          
Crosslinker Number of crosslinker molecules reacted to 100 anhydroglucose units  
  POLYMER   MCS POLYMER    DHPA 
  pH7.00 pH10.00 pH7.00 pH10.00* 
TMC 0.52 0.62 0.47 5.13 
TC 0.34 0.39 0.34 3.13 
DEMC 0.017 0.04 0.02 0.62 
*Reacted for 24 Hours, TMC: trimesoyl chloride, TC: terephthaloyl Chloride, DEMC: 
diethyl malonyl chloride, MCS: maltodextrin from corn starch, DHPA: dihydroxypropyl 
amylose. 
  
 
 
 
   
Polymer Polymer Concentration (w/v)%  Degree of substitution * 
   
 Dihydroxypropyl amylose 4.5 5.57 
Amylomaize starch 5 3.69 
Corn starch 4 8.32 
Amylopectin (Waxy maize 
starch) 3 2.09 
Maltodextrin (from Corn starch) 50 0.58 
Maltodextrin (from Waxy Maize 
starch) 45 0.84 
Hydroxyethyl starch (from Corn 
starch) 20 2.56 
Hydroxypropyl starch (from 
tapioka starch) 20 2.08 
 126 
 
Figure 1.  Particle size distribution of two representative nanoparticle batches.  Nicomp? 
distribution is employed. 
 
 
 
 127 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 2a. Optical microphotograph of nanoparticles prepared from corn starch and 
terephthaloyl chloride and suspended in distilled water.  Each division in the scale 
represents 1 ?m.  All other optical photographs are of same magnification.  
 Figure 2b. Optical microphotograph of nanoparticles prepared from waxy maize 
starch and terephthaloyl chloride.  
 Figure 2c. Optical microphotograph of nanoparticles prepared from hydroxyethyl 
corn starch and terephthaloyl chloride. 
 
 128 
 Figure 2d. Optical microphotograph of nanoparticles prepared from maltodextrin 
from corn starch and terephthaloyl chloride. 
 
Figure 2e. Optical microphotograph of nanoparticles prepared from maltodextrin from 
waxy maize starch and trimesoyl chloride. 
 Figure 2f. Optical microphotograph of nanoparticles prepared from maltodextrin from 
waxy maize starch and trimesoyl chloride after 4 weeks in suspension in distilled 
water. 
 
 
 129 
 
Figure 3a. 
 
Figure 3b. 
Figure 3.  A representative transmission electron microphotograph of nanoparticles 
prepared with maltodextrin and trimesoyl chloride.  The sample is dry-mounted on a 
copper grid coated with polyvinyl alcohol film. Magnification was (a) 100,000 
(b)160,000. 
 
  
 130 
0
0.2
0.4
0.6
0.8
1
1.2
5 6 7 8 9 10 11 12
pH
Nu
m
be
r o
f c
ro
ss
lin
ke
r m
ole
cu
les
 re
ac
te
d t
o 1
00
 
an
hy
dr
og
luc
os
e u
nit
s
 DEMA TA
  Figure 4. Effect of pH on degree of substitution with TC and DEMC as crosslinkers and 
maltodextrin (MCS).  Reaction was carried out for 3 hr.   
 
 
 
0
0.2
0.4
0.6
0.8
1
1.2
5 6 7 8 9 10 11 12
pH
Nu
mb
er 
cr
os
sli
nk
er 
m
ole
cu
les
 re
ac
ted
 to
 
10
0 a
nh
yd
ro
gl
uc
os
e u
ni
ts
 Figure-5. Effect of pH on degree of substitution with TC as crosslinker and DHPA.   
Reaction was carried out for 3 hr. 
  
 131 
 
0
1
2
3
4
5
6
2 3 4 5 6 7 8 9 10 11 12
pH
Nu
m
be
r o
f c
ro
ss
lin
ke
r m
ole
cu
les
 re
ac
ted
 
to
 10
0 a
nh
yd
ro
glu
co
se
 un
its
 Figure 6.  Effect of pH on degree of substitution with TMC as crosslinker and 
maltodextrin (MCS). Reaction was carried out for 24 hr.   
 
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
2 4 6 8 10 12
pH
Nu
m
be
r o
f c
ro
ss
lin
ke
r m
ole
cu
les
 re
ac
ted
 
to
 10
0 a
nh
yd
ro
glu
co
se
 un
its
  Figure 7. Effect of pH on degree of substitution with DEMC as crosslinker and DHPA.   
Reaction was carried out for 24 hr. 
 
 
 132 
0
1
2
3
4
5
6
0 50 100 150 200 250
Concentration of the aqueous phase buffer mM
Nu
m
be
r o
f c
ro
ss
lin
ke
r m
ol
ec
ul
es
 
re
ac
ted
 to
 10
0 a
nh
yd
ro
gl
uc
os
e u
ni
ts
TMC
TC
  Figure 8. Effect of buffer strength on degree of crosslinking with TMC and TC as 
crosslinker and DHPA.  The TMC batches were reacted for 24 hr at pH 7.0 while the 
batches with TC were reacted for 3 hr at pH 10.0. 
 
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 1 2 3 4 5
Polymer concentration (w/v)%
Nu
m
be
r o
f c
ro
ss
lin
ke
r m
ole
cu
les
 
re
ac
te
d t
o 1
00
 an
hy
dr
og
luc
os
e u
nit
s
 Figure 9.  Effect of polymer concentration on degree of substitution with TMC as 
crosslinker and DHPA.  Reaction was carried out for 24 hr at pH 7.0. 
 
 
 133 
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 20 40 60 80 100 120 140 160
Crosslinker concentration in organic phase (mMoles)
Nu
m
be
r o
f c
ro
ss
lin
ke
r m
ole
cu
les
 
rea
cte
d t
o 
10
0 a
nh
yd
ro
glu
co
se
 u
nit
s
 Figure 10.  Effect of crosslinker concentration on the degree of substitution with TMC as 
crosslinker and MA as polymer.  Reaction was carried out for 24 hr. 
 
0
1
2
3
4
5
6
0 10 20 30 40 50 60
TIME (Hr)
DE
GR
EE
 O
F C
RO
SS
LIN
KI
NG
   (
w/
w)
%
DEMA
TA
TMA
 
Figure 11.  Effect of reaction time on the degree of substitution with TMC, TC and 
DEMC as crosslinkers and MCS as polymer and at pH 10.0. 
 
 
 134 
V. INCOPORATION INTO AND RELEASE OF TARTRAZINE AND 
LYSOZYME FROM CARBOHYDRATE NANOPARTICLES 
 
ABSTRACT 
The preparation and characterization of carbohydrate nanoparticles with a hydrophilic 
surface has been reported elsewhere.  In the present work, the suitability of nanoparticles 
as a sustained/controlled release platform for bioactive molecules will be investigated in 
detail.  A model small molecule, tartrazine, and a protein, lysozyme, were incorporated 
into nanoparticles.  Encapsulation efficiency and release profiles were determined for 
both molecules.  The results show that the crosslinked membrane of the nanoparticles is 
not likely to produce adequate resistance to diffusional release of small molecules for 
most sustained release applications.  However, the incorporation of small molecules into 
the nanoparticles revealed that high carbohydrate polymer load in the dispersed aqueous 
phase is necessary and sufficient to achieve high encapsulation efficiency of the 
incorporated molecules. 
Pre-formulation studies with lysozyme revealed that when cyclohexane or chloroform 
was used as the continuum phase, high encapsulation efficiency was achieved.  But when 
a combination of these two organic solvents or methylene chloride was used, 
encapsulation efficiency sharply decreased.  Further, lysozyme encapsulation efficiency 
is strongly affected by pH of the dispersed aqueous phase.  The optimum pH for 
encapsulation is 3.0.  Carbohydrate polymer load also improves encapsulation efficiency.  
 134 
Studies prove sustained release of lysozyme is achievable using this platform.  
Limited control over the release rate can also be exercised by varying the degree of 
substitution within a narrow range.  Mathematical modeling of lysozyme release shows 
heterogeneous diffusion with two distinct modes, one slow and one very fast.  The slower 
mode of diffusional release is believed to be the consequence of a diffusional barrier 
provided by the crosslinked membrane at the surface of the nanoparticles.  The cause of 
the faster mode of diffusion in the initial period could not be definitely identified.  
Further research needs to be focused towards identifying and minimizing the initial burst 
release.  Future research should also address the biologic disposition of nanoparticles 
specifically with regards to phagocytosis, in vivo release of bioactive molecules from the 
nanoparticles, particle integrity in physiological fluid, pharmacokinetic distribution and 
plasma half life of the nanoparticles.  
 
1. INTRODUCTION 
Carbohydrate polymers, both charged and uncharged have been used historically in 
drug delivery research to achieve controlled delivery.  Charged carbohydrate polymers, 
like anionic alginates (Gombotz, W. R., 1998), pullulan (Mocanu, G., 2002), and cationic 
chitosan (Richardson, C. W. S., 1999; Mi, F. L., 1999) have all been used to produce 
microspheres as drug delivery platforms.  The use of nonionic carbohydrate polymers 
like starch (Morath, L. P., 1998), high amylose starch (Lenaerts, V., 1998) and dextran 
(Mehvar, R., 2000; Stenkes, R. J. H., 2001; Franssen, O., 1999) for drug delivery has 
been gaining acceptance in recent years.  Microspheres have been prepared with alginic 
acid by utilizing the ionic interactions between the carboxylic acid groups in the polymer 
 135 
and di or trivalent metal ions like Ca+2 or Al+3.  The limitation of these systems is that the 
ionic interaction is reversible and the integrity of the microspheres are compromised in 
the presence of other ionic species like citrates, phosphates, Na+ and Mg+, which are 
normally present in physiological fluids.  The resulting disintegration of nanoparticles on 
exposure to physiological fluids causes rapid release of encapsulated bioactive molecules 
(Gombotz, W. R., 1998).  Polyion condensation has been used to produce 
micro/nanoparticles with chitosan and DNA for potential use as non-viral DNA delivery 
vectors.  Apart from the fact that all cationic polymers including chitosan possess 
considerable toxicity (Richardson, C. W. S., 1999), chitosan microspheres are also made 
through ionic interactions and thus suffer from similar limitations as alginates.  A variety 
of different crosslinking agents have been used to produce micro/nanoparticles by 
crosslinking nonionic carbohydrate polymers throughout the device.  Crosslinking in the 
bulk, especially with strong crosslinking agents like gluteraldehyde (Tabata, Y., 1989) is 
always associated with concerns of reacting the encapsulated bioactive compound itself 
and thus denaturing and/or immobilizing them in the particle.  
In an attempt to address these concerns, carbohydrate nanoparticles have been 
prepared using a novel emulsion/interfacial crosslinking reaction.  The main advantage of 
this platform is that the encapsulated bioactive molecules remain in a benign 
microenvironment at the core of the device unaffected by the interfacial crosslinking or 
other harsh preparative conditions.  Another important aspect of this drug delivery system 
(DDS) is that being made up of carbohydrate polymers, this DDS has a hydrophilic 
surface which is expected to minimize if not eliminate the activation of the 
reticuloendothelial system and thus the removal of the particles from blood stream 
 136 
through phagocytosis when introduced into the body.  In addition, aggregation in aqueous 
supporting matrices is eliminated. 
Having characterized the new carbohydrate nanoparticulate DDS in terms of physical 
and chemical properties, the object of the current research is to investigate complexities 
involved with the incorporation of potential bioactive molecules into the nanoparticles 
and their subsequent in vitro release from the particles in simulated physiological 
conditions.  Two molecules have been chosen for the encapsulation to represent the two 
          
                1a.                                                  1b.                                           1c. 
Scheme-1.  A schematic depiction of the two [(a) and (b)] natural stereo structures of 
lysozyme.  Proc. Natl. Acad. Sci., USA, 1983, (80): 5773. 1(c) is a depiction of the 
molecular structure of tartrazine. 
 
extreme ends of the spectrum of bioactive molecules.  Tartrazine, a hydrophilic small 
molecule, containing strong anionic groups, is chosen to represent traditional small drug 
molecules [Scheme-1(c)].  Chicken egg white lysozyme is a moderately large protein 
(molecular weight 14,700) with native stereo structures [Scheme 1(a) and 1(b)]. It is a 
relatively hydrophobic protein carrying a net cationic charge at pH 9 and below and has 
been chosen to represent the emerging group of polypeptides and proteins that are 
 137 
increasingly being used in the treatment of disease conditions.  Lysozyme has a 
considerably large hydrodynamic diameter of 4.1 nm (Jensen, M., 2002) and slower 
diffusion of the molecule out of the nanoparticles is expected. The main goals pursued in 
this effort are to maximize encapsulation efficiency and quantitatively describe the 
mechanism of release. 
 
2. EXPERIMENTAL 
2.1 Materials 
The carbohydrate polymers were procured from a variety of commercial sources and 
used without further purification.  The polymers include corn starch (CS) (Sigma, St. 
Louis, MO), waxy maize starch (WMS) (PFP 2850, Cargill Foods, Minneapolis, MN), 
maltodextrin (MA) from waxy maize starch (C*Dry 01955, Cerestar USA Inc., 
Hammond, IA), and maltodextrin (MCS) from corn starch (Maltrin M40, Grain 
Processing Corporation, Mucatine, IA).  The crosslinkers include terephthaloyl chloride 
(TC) (Sigma-Aldrich, St.-Louis, MO), and trimesoyl chloride (TMC) (Across organics, 
Pittsburg, PA).  Other chemicals include sorbitan-mono-palmitate (Span-40), buffer salts 
(Sigma-Aldrich, St. Louis, MO), chloroform (CL), cyclohexane (CH), methylene 
chloride (MC), acetone, and ethanol (Fisher Scientific, Pittsburg, PA).  All solvents were 
of ACS grade.  Dimethyl sulfoxide (DMSO), acetonitrile (ACN), tetrahydrofuran (THF), 
and dimethyl formamide (DMF) were used as high pressure liquid chromatography 
(HPLC) solvents.  Triethylamine (TEA), phenacyl bromide (PB), sodium azide (SA) are 
other reagents (Fisher Scientific).  
 138 
Organic phase one (OP-1) was prepared by dissolving 6 % (w/v) Span 40 in a mixture 
of cyclohexane and chloroform (68/32, v/v) as the continuum phase of the emulsion.  
Organic phase two (OP-2) was prepared by dissolving 2.5 % of TC or 2.27 % (w/v) of 
TMC in cyclohexane and chloroform (68/32, v/v) for carrying out the crosslinking. 
 
2.2 Preparation of nanoparticles and incorporation of tartrazine 
Carbohydrate nanoparticles were prepared by an interfacial crosslinking reaction as 
earlier described.  Tartrazine nanoparticles were made using combinations of CS, WMS 
and maltodextrins MCS and MA as polymers with TC and TMC as crosslinkers.   
Briefly, the carbohydrate polymers CS, WMS, MCS and MA were heated in an 
appropriate volume of buffer in screw capped bottles to produce carbohydrate colloids in 
the case of high molecular weight carbohydrates, and true solutions in the case of 
maltodextrins.  The concentration of carbohydrate polymers was maintained close to their 
maximum solubility (Table-1).  The pH of the aqueous phase was maintained at 10.0 by 
using 200 mM borate buffer to maximize the encapsulation of tartrazine. The colloid or 
solution was allowed to cool to room temperature while stirring before tartrazine was 
incorporated at a concentration of 10% (w/w) relative to the polymer.   
Twenty (20.0) mL of the carbohydrate colloid/solution was emulsified in 100 mL of 
OP-1 in a Silverson LR-4? homogenizer at 5,000 rpm for 1 minute.  The emulsion was 
then poured into 100 mL of organic OP-2.   The reaction was continued for 24 hr.  The 
final concentration of both the crosslinkers was ~55 mM in the reaction mixture. The 
reaction was stopped by pouring the reaction mixture into 700 mL of ethanol in a 1000 
mL beaker while stirring.  The precipitated nanoparticles were washed 3 times with 
 139 
ethanol and 3 times with acetone by repeated centrifugation at 1,500 g and vortexing 
following discontinuation of stirring.  Finally, the suspension of nanoparticles in acetone 
was dried under vacuum at 14.5 psi at room temperature (20-25?C) overnight.  
 
2.3 Encapsulation and release studies with tartrazine  
To evaluate the effect of degree of substitution (DS) on the release of tartrazine the 
following protocol was observed.  Two hundred and twenty (220) mg of dry 
nanoparticles with a theoretical loading of 9.09% (w/w) of tartrazine was dispersed in 5 
mL of 100 mM potassium phosphate buffered saline (KBPS) at pH 7.4.  The suspension 
was immediately placed into an 18 cm long dialysis membrane tube (SpectraPor?, 
Spectrum Laboratories Inc., Ranco Dominguez, CA; molecular weight cutoff of 50,000 
Daltons).  One end of the dialysis tube was clamped securely, the suspension added and 
the other end was clamped such that the distance between the two clamps was 5 cm in 
each case.  The lower clamps had a magnetic weight built in so that the formed pouch 
floats upright and remains stable during stirring.  The pouch containing the suspended 
nanoparticles was placed in a 1000 mL beaker containing 495 mL of 100 mM KBPS.  To 
minimize evaporation of the dissolution media, the beakers were covered with a thin 
polyethylene sheet (Clingwrap?). A multiple head magnetic stirrer with precise control 
was used to maintain a stirring speed of 150 rpm.  One (1.0) mL samples were withdrawn 
at predetermined time intervals and the media replenished each time.  Tartrazine was 
quantified by absorbance spectroscopy using a Beckman UV/VIS spectrophotometer at 
445 nm wavelength.  To account for any possible contribution of the dialysis membrane 
to tartrazine delivery, 20 mg of tartrazine was dissolved in 5 mL of 100 mM KPBS (4 
 140 
mg/mL).  The aliquot was placed into dialysis tubing as previously described and used as 
the control.  
At the end of the release studies, the asymptotic level of tartrazine concentration was 
used for encapsulation efficiency based on:   E = (A/T)*100, where E is the encapsulation 
efficiency, A is the actual amount of tartrazine released and T is the amount (9.09 %) of 
tartrazine added. 
 
2.4 Partitioning and interfacial aggregation of lysozyme 
In order to ensure maximum loading of lysozyme in the carbohydrate nanoparticles, 
pre formulation studies were performed.  The effect of organic solvents used as the 
continuum phase, pH, and carbohydrate load of the dispersed aqueous phase on lysozyme 
partitioning was investigated.  Firstly, 4.5 % (w/v) lysozyme was dissolved in deionized 
water.  The pH of the deionized water was adjusted to 3.0, 4.0 and 5.0 by adding 50, 20, 
10 ?L of 10 % (v/v) trifluoroacetic acid.  The final concentration of trifluoroacetic acid 
was 1.345, 0.054 and 0.027 mM, respectively.  Since lysozyme is a basic protein with an 
isoelectric pH of 11.0 (Alderton, G., 1945), the series of pHs chosen are expected to fully 
ionize the amino groups resulting in minimum partitioning into the continuum phase. 
Twenty (20) mL of this aqueous phase was homogenized with 100 mL of methylene 
chloride, cyclohexane, chloroform or cyclohexane/chloroform (68/32, v/v).  Surfactant 
was deliberately omitted from the preformulation study to ensure adequate separation of 
the phases.  The homogenized phase was allowed to separate on standing overnight at 
room temperature.  After the phases separated, 1.0 mL samples were collected from the 
 141 
aqueous phase, centrifuged at 1,500 g to remove any insoluble aggregated lysozyme and 
analyzed by HPLC.  
 
2.5 Incorporation of lysozyme in nanoparticles 
Lysozyme was incorporated in nanoparticles made from maltodextrin obtained from 
waxy maize starch (MA).  The crosslinker used was TMC.  The preparation procedure 
used was similar to the process already described.  Only the modifications will be 
mentioned in this section.  Forty five (45)   percent (w/v) of MA was dissolved in the 
aqueous phase buffered at pH 3.0 by heating.  Unless stated otherwise, 4.5 g of lysozyme 
was added to 50 mL of MA solution upon cooling to room temperature.  The solution 
was slowly stirred until the lysozyme dissolved.  A final theoretical loading of 9.09 % 
(w/w) was obtained.   Nine (9) g of lysozyme was added to the same volume of MA 
solution in some batches to achieve a theoretical lysozyme loading of 16.67 % (w/w).  
The TMC concentration was varied from 0.8 to 486 mM in the continuum phase.   
 
2.6 Lysozyme assays  
2.6.1 Assay of lysozyme from release samples 
Standard protein solutions were made in the concentrations of 1, 5, 10, 20, 30 and 40 
?g/ml in 100 mM (KPBS) containing 0.2 % (w/v) sodium azide.  To account for any 
interference from the nanoparticles in authentic samples, blank nanoparticles were 
suspended in the standard solution in the same concentration used in the release 
experiments.  As a low molecular weight carbohydrate scavenger, 75 ?L of n-butanol 
was added to 1.0 mL of the sample in a 1.5 mL polypropylene micro centrifuge tube.  
 142 
The samples were then centrifuged at 14,000 g for 15 min to separate any precipitated 
carbohydrate residue and the suspended nanoparticles. Approximately 250 ?L of the 
supernatant was then transferred to inserts in autosampler vials. Seventy (70) ?L of the 
sample was injected for HPLC analysis.   Both the lysozyme standards and samples from 
release studies were identical to one another in composition except for the lysozyme 
content. 
The stationery phase was a Zorbax SB300? column with 5 ? particle size, 300 ? pore 
size, C18 packing, and with 4.6 x 150 mm column dimension.  A 4.6 x 10 mm precolumn 
of the same stationary phase composition was used as a protective device.  The mobile 
phase was composed of 57 % (v/v) of solvent A and 43 % (v/v) solvent B pumped at a 
flow rate of 1.0 mL/min.  Solvent A was composed of deionized water containing 0.1 % 
(v/v) trifluoroacetic acid (TFA).  Solvent B was composed of 4 % (v/v) DMSO, 20 % 
(v/v) THF, 76 % (v/v) ACN, and 0.1 % (v/v) TFA.    
The chromatography was carried out at ambient temperature. The chromatographic 
system consisted of a Thermo Separation (San Jose CA, USA) Model P2000 pump used 
in the isocratic mode, a Model AS3000 autoinjector, and a Model FL2000 variable-
wavelength fluorimetric detector.  The detection was carried out at 280 nm excitation and 
340 nm emission wavelengths.  Lysozyme eluted at 4.2 min.  The calibration curve was 
constructed in the concentration range of 1 - 40 ?g/ml and all release study measurements 
were conducted within the range.  
 
 
 
 143 
2.6.2 Content assay of lysozyme 
For quantifying nanoparticle lysozyme content, a slight modification of the HPLC 
sample preparation method was adopted for optimum efficiency.  An appropriate amount 
of dry nanoparticles were dispersed in 1.0 mL of 100 mM KPBS containing 0.2 % (v/v) 
SA. The suspension was diluted with 2.0 mL of deionized water and 1.0 mL of mobile 
phase, mixed for 2 min in a vortex mixer, and kept for 1 hr at room temperature.  A 70 
?L aliquot was injected.  The DMSO present in the mobile phase was found to dissolve 
the nanoparticles.  A calibration curve was constructed and used for this purpose in the 
range of 1-40 ?g/ml.  The sample amount of the nanoparticles was adjusted to maintain 
lysozyme concentrations within the calibration range. 
 
2.7 Crosslinker assay 
The degree of substitution of TC and TMC were measured by a reverse phase HPLC 
method (Basu Sarkar, A., 2005).   
 
2.8 Release studies with lysozyme 
Lysozyme release was evaluated from a dispersion of nanoparticles at pH 7.4 in 100 
mM KPBS.  The dialysis membrane was omitted since prolonged mass transfer was 
observed for non-encapsulated lysozyme solution using a 100,000 Da molecular weight 
cutoff dialysis membrane.  Two hundred and twenty (220) mg of dry nanoparticles 
containing lysozyme in the concentrations of either 9.09 or 16.67 % (w/w) were 
dispersed in 500 mL of 100 mMolar KPBS containing 0.02 % (w/v) of SA in a 1000 mL 
conical flask.  To avoid evaporation of the aqueous media, the flask was covered with a 
 144 
thin polyethylene sheet (Clingwrap?).  Sink conditions were maintained by a volume of 
500 mL of reslese media and stirring the media at 150 rpm.  One (1.0) mL of sample was 
withdrawn at predetermined time points for quantification.  The samples were handled as 
described in the lysozyme assay and 70 ?L of the sample was injected for analysis by 
HPLC. 
 
2.9 Regression analysis and model assessment for lysozyme release 
Nonlinear least squares regression was used to analyze the underlying mechanism of 
lysozyme release based on a spherical diffusion model.  Initially a homogeneous 
diffusion model (Kim, C., 2000) was used as described by the following equation (1). 
M
M n e
t
n
Pnt
? =
?
?= ? ?1 6 1
2 2
1
2
p   P
D
r=
p 2
2 1() 
  Here, ?Mt? is the mass of lysozyme released in time ?t?, ?M?? is the mass of 
lysozyme released at infinite time, ?n? is the summation index varying from one to 
infinity, ?P? is the permeability coefficient, ?D? is the diffusion coefficient of lysozyme, 
?r? is the mean radii of the spheres and the term Mt/M?, denotes fractional release of 
lysozyme at time ?t?.  Best fit statistical analysis demonstrated the discrimination of two 
independent modes of lysozyme diffusional release.  As a consequence, a heterogeneous 
diffusional model was adopted for subsequent analysis as described by equation (2). 
 
 145 
M
M
M
M n e
M
M n e
t
n
Pnt
n
Pnt
?
?
? =
?
? ?
? =
?
?= ??
??
?
?? + ?
?
??
?
??? ?
1
2 2
1
2
2 2
1
1 6 1 1 6 112 2 2, ,p p
 
        P
D
r P
D
r1
1
2
2 2
2
2
2 2= =
p p, ()
 
 
 
Here, M1,? is the mass of lysozyme released by mode 1 at infinite time, M2,? denotes 
the mass of lysozyme released by mode 2 and M? is the total mass of lysozyme released. 
 In both models, a ten-term truncation of the infinites series was found adequate for 
data analysis.  The series coefficient (6/pi2) was appropriately adjusted to a value of 
0.64526 corresponding to the reciprocal sum of the first ten terms of the series at t = 0. 
All curves were analyzed using normalized data where M? was estimated from the 
numerical average of the last 2-3 asymptotic data values.  Residual sum of squares was 
weighted using a factor of 1.0 for regression analysis.  The parametric values of P1, P2, f1, 
and f2 were estimated, where f1 = M1,?/M? and f2 = M2,?/M?.  
 
3. RESULTS  
3.1 Encapsulation and release of tartrazine: The effect of polymers 
Tartrazine was incorporated in nanoparticles synthesized from the carbohydrate 
polymers CS and WMS and the oligomers MCS and MA using both TC and TMC as 
crosslinkers.  Variable encapsulation efficiency was observed (Table 1). Nanoparticle 
batches prepared from carbohydrate polymers CS and WMS showed relatively low 
 146 
encapsulation efficiency of 32 to 39 % of the incorporated tartrazine irrespective of the 
crosslinker used.  Batches made from carbohydrate oligomers MCS and MA showed 
higher encapsulation efficiencies of 73 and 84 %, respectively.   
 Following completion of the crosslinking reaction, the dispersed aqueous phase was 
extracted in a large volume of ethanol resulting in a rapid precipitation of the polymer.  
Though tartrazine is not soluble in any of the solvents, cyclohexane, chloroform or 
ethanol, it is extractable from the nanoparticles in the presence of small amounts of water 
provided by the dispersed phase.  The small amount of   tartrazine remained in the 
nanoparticles prepared from CS and WMS because it was entrapped in the rapidly 
precipitating polymers.  The entrapment effect of the encapsulated tartrazine was low for 
CS and WMS because of the low polymer loads that can be achieved with these polymers 
in the dispersed phase.  In cases of MCS and MA, the entrapment effect of the 
carbohydrate is much more pronounced due to a ten fold higher carbohydrate load [45-50 
% (w/v)], made possible by their higher aqueous solubility. 
Release studies from TC (Figure 1) batches showed significant sustained release from 
the branched polymers WMS and MA as compared to the control. Nanoparticles made 
from the other two polymer species, CS and MCS, showed very little difference in their 
release profiles compared to the control.  Release studies from TMC (Figure 2) shows 
significant sustained release of tartrazine only from the nanoparticles made with WMS.  
A comparison (Figure 3) of the tartrazine release from nanoparticles synthesized with TC 
and TMC shows that the branched polymer amylopectin (WMS) imparts more sustained 
release than regular corn starch for both the crosslinkers. 
 147 
Although some differences in tartrazine release were observed relative to the specific 
carbohydrate used for encapsulation, comparative tartrazine delivery relative to the 
control was unremarkable. These results indicate that the encapsulating membrane offers 
little resistance to tartrazine diffusional mass transfer. 
 
3.2 Lysozyme assay  
3.2.1 Accuracy and precision of release samples assay 
The accuracy of lysozyme determinations (Table 2) ranged from 96 to 103 % 
corresponding to nominal concentrations of 1.0 to 40 ?g/mL.  The precision as described 
by the relative standard deviation (RSD) was found to be < 2.5 % corresponding to 
concentrations from 5 to 40 ?g/mL.  At the 1 ?g/mL concentration level, the precision 
was approximately 20 % of the mean.  This concentration level was considered the lower 
limit of detection.  Chromatography was stable and no distortion of peak shape or 
deviation of retention time was observed between standard and actual samples (data not 
shown). 
 
3.2.2 Accuracy and precision of the content assay 
In the content assay, the accuracy of lysozyme determinations (Table 3) varied from 
90 to 110 % for concentrations ranging from 1 to 40 ?g/mL.   The precision, as described 
by the relative standard deviation was between 1 to 3.5 % for every concentration level 
except 1 ?g/mL which is considered the lower limit of quantification for this assay as 
well.  At this level, the precision was 24 % of the mean.  The chromatography was less 
stable compared to the release sample assay of lysozyme.  Some broadening of the peak 
 148 
was observed though retention time did not change.  Peak height was used instead of 
peak area for constructing calibration curves. 
 
3.3 Partitioning of lysozyme and its encapsulation efficiency  
The results of lysozyme partitioning experiments using methylene chloride, 
cyclohexane, chloroform and cyclohexane/chloroform [68/32, (v/v)] at pHs 3.0, 4.0, and 
5.0, showed that lysozyme partitioning depended heavily upon the exact set of 
conditions.  When MC was used as the continuum phase (Table 4), no lysozyme was 
detected in the aqueous phase following emulsification and phase separation at any pH 
tested.  Insoluble aggregated lysozyme was observed at the interface and also in both 
phases.    Lysozyme partitioning into the organic phase and aggregation at the interface 
were found to be low when CH and CL were used separately as the organic phase.  The 
soluble lysozyme, recovered from the aqueous phase following emulsification and phase 
separation, were 73 and 75 % for CH and CL, respectively, at pH 3.0.  However recovery 
of soluble lysozyme in the aqueous phase was 10 % when CH/CL [68/32, (v/v) %] was 
used as the continuum phase at pH 3.0.  Also, the percentage of soluble lysozyme 
decreased rapidly with the increase in pH in all cases in which lysozyme was detected in 
the aqueous phase.  The observation of progressively lower recovery of lysozyme from 
dispersed phase as the pH increased could be attributed to the higher degree of 
protonation of the amino groups of lysozyme.  Thus, for incorporation of lysozyme, CL 
was selected as the continuum organic phase and the pH of the dispersed aqueous phase 
was selected to be 3.0.  Lysozyme is known to be stable for several weeks at low acidic 
pH. 
 149 
In order to investigate the effect of carbohydrate content on lysozyme partitioning, 
studies were conducted with 15, 30 and 45 % (w/v) of MA in the aqueous dispersed 
phase at pH 3.0 with CL as the organic continuum phase.  The results show a near 
quantitative recovery of soluble lysozyme in the aqueous phase following emulsification 
and phase separation with all levels of carbohydrate in contrast to 75 % recovery from the 
identical phase system without any carbohydrate (Table 5).  These results lead to the 
conclusion that high carbohydrate load cause lysozyme retention in the dispersed phase 
due to carbohydrate protein interaction.  Table 6 shows the actual lysozyme 
encapsulation efficiency measured in representative batches.  The encapsulation 
efficiency averaged 98 ? 10.2 % of the added lysozyme.   
 
3.4 Effect of lysozyme loading on degree of substitution 
The effect of lysozyme loading on the DS was assessed with two different levels of 
lysozyme loading with varying TMC concentrations.  The results (Figure 4) show that the 
DS increases with TMC concentration in the continuum phase as expected in both cases.  
The DS is hyperbolically dependent on TMC concentration.  The observed differences in 
reactivity were associated with lysozyme load when the TMC concentration exceeded 16 
mM.  This observation suggests the presence of lysozyme at the interface and excessively 
large protein loads can interfere with interfacial crosslinking.  This effect probably limits 
protein loading in encapsulated systems. 
 
 
 
 150 
3.5 The effect of crosslinking on lysozyme release 
Nanoparticles were prepared with the same lysozyme loading [9.09 % (w/w) of the 
dry nanoparticles] and 5 different levels of TMC substitution.  The crosslinking was 
varied in terms of DS from 0.03 to 7.09 by varying the initial crosslinker concentration in 
the continuum phase from 0.8 mM for to 485.2 mM.  The release studies show (Figure 5) 
that DS has a profound effect on the extent of lysozyme release as discussed later.   
Nanoparticles with a DS of 0.03 showed instantaneous release of the encapsulated 
lysozyme.  The resistance effect of the crosslinked membrane was completely absent.  
Nanoparticles with DS of 0.13 to 0.16 released 27 and 18 % of encapsulated lysozyme in 
the first hour, 48 and 26 % after 2 days, and 62 and 39 % in 11 days, respectively, and no 
significant release thereafter in the time frame of the release studies. Continuous and 
sustained release characteristics of the nanoparticles are expressed prominently in this 
range of DS.  Nanoparticles with DS between 0.03 and 0.13 showed some sustained 
release for up to 2 days (data not shown).  Further increase of DS to 1.07 resulted in very 
slow release of ~13 % in 1 hr and up to ~20 % in two days and approximately 22, 23 and 
24 % release on days 10, 12 and 15 respectively.  For DS of 7.09 lysozyme release 
virtually stopped with 3 % in 1 hr, ~ 8 % in 2 days and no significant release thereafter.  
An inverse relationship between the extent of lysozyme released in 15 days and the 
DS of TMC was evident (Figure 6).  As the DS increased from 0.032 to 7.094, the extent 
of lysozyme release decreased rapidly up to a DS value of 0.158.   The extent of 
lysozyme release decreased slowly thereafter up to a DS of 7.094.  
It is apparent that there is a narrow range of substitution in which the lysozyme 
release can be controlled. This range corresponds with DS values between 0.13 and 1.07.  
 151 
Another significant observation is that the varied degree of crosslinking affects the extent 
of lysozyme release more significantly than the rate of release. 
 
3.6 Modeling of lysozyme release  
A homogeneous diffusion model initially used to mechanistically explain lysozyme 
release from nanoparticles was found to be inadequate (Figure 7).  As evident from the 
figure, the errors or residuals from the predicted data are not random but systematic.  The 
residuals progressively decrease from a positive value to negative as time increases.  The 
same trend was also observed with the release data from other representative batches 
(data not showed).   
 In an effort to fully capture the complexity of the events, a heterogeneous diffusion 
model was then applied for data analysis.  Resulting fitted curves (Figure 8-11) all show 
high correlation with coefficients (r2) > 0.985 and random residuals, thereby validating 
the model.   
Figure 12a and 12b show the calculated release profiles of lysozyme from 
nanoparticles (DS 0.132) in the time frames of 24 hr and 12 days.  The profiles in each 
figure correspond to the two individual diffusional modes of release characterized by P1 
and P2.   The first mode, characterized by P1, is very fast and is complete by 
approximately three hours in all cases (data not shown).  The second mode, characterized 
by P2, is slower.  The diffusional release due to the second mode continues for 
approximately two weeks.   
The permeability coefficient P2, corresponding to the slower mode of lysozyme 
release, was found to be related to the DS (Figure 13).  This finding suggests that the 
 152 
slower mode of lysozyme release may be due to the resistance provided by the 
polymerized membrane at the nanoparticle surface.  No such relationship was found with 
P1. 
 
3.7 The effect of lysozyme loading on the release of lysozyme 
Nanoparticles with lysozyme loading of 9.09 % (w/w) and 16.67 % (w/w) were 
prepared under identical conditions of crosslinking.  The release profiles showed (Figure 
14) that lysozyme release from nanoparticles with 16.67 % loading was much faster (P2 
0.371) and more complete than from the nanoparticles with lower lysozyme loading of 
9.09 % (P2 0.198).   
 
4. DISCUSSION 
Crosslinking of carbohydrate polymers with the tri functional TMC reagent at the 
emulsion interface is an exceedingly complicated reaction.  Many different products are 
possible.  Mono, di or tri esterification as well as inter chain crosslinking, intra chain 
crosslinking, and chain elongation of the carbohydrate polymers are all possible.  A 
crosslinking reaction involving at least two acid chloride groups is the goal of the 
interfacial esterification reaction.  But it is dependent upon the availability of alcoholic 
hydroxyl  groups  of  the carbohydrate  polymers  in the  right  spatial  arrangement  with 
respect to the crosslinker molecule.  The hydroxyl groups of the carbohydrate polymers 
have to be present at the interface in sufficient concentrations so that the groups are 
available for reaction.  
 
 153 
2(a) Carbohydrate Polyesterification ~ possibilities
O
H
O
H
HO
H
OHH
H
OH
O
H
O
H
HO H
OH
H
H
OH
O
O
H
HO
H
HO
H
OHH
H
OH
O
H
O
H
HO H
HO
OH
H
H
COCl
ClOC COCl
+
aqueous buffer
chloroform
emulsion
O
H
O
H
HO
H
OHH
H
OH
O
H
O
H
HO H
OH
H
H
O
O
O
H
HO
H
HO
H
OHH
H
OH
O
H
O
H
HO H
HO
OH
H
H
COO
COO
O
Amylopectin
(maltodextrin, ~ 20 agu)
Trimesoyl chloride
 
Scheme 2.  A schematic representation of some of the possibilities of the carbohydrate 
crosslinking with trimesoyl chloride.  2(a) shows a mono esterified trimesic acid with two 
free acid groups. 2(b) a di esterified trimesic acid with only one free acid group.    
                           
2(b) Carbohydrate Polyesterification ~possibilities
O
H
O
H
HO
H
OHH
H
OH
O
H
O
H
HO H
OH
H
H
OH
O
O
H
HO
H
HO
H
OHH
H
OH
O
H
O
H
HO H
HO
OH
H
H
COCl
ClOC COCl
+
O
H
O
O
H
OHH
H
OH
O
H
O
H
HO H
H
OH
O
H
HO
H
HO
H
OHH
H
O
O
H
O
H
HO H
HO
OH
H
H
O
O
OOC
aqueous buffer
chloroform
emulsion
H
O
OH
H
Amylopectin
(maltodextrin, ~ 20 agu)
Trimesoyl chloride
 
 154 
When lysozyme is incorporated into the particles, an amide formation as a result of 
the reaction between acid chloride groups of the crosslinker and free amino groups of the 
protein is also possible.    This side reaction may also be aided by the fact that lysozyme, 
like most moderate sized proteins, has some surfactant activity and tends to migrate to the 
interface.  It is expected that the polymerization of lysozyme into the carbohydrate 
polymer would reduce the cumulative extent of lysozyme delivered, but not necessarily 
affect the superficial membrane since the carbohydrate is present in concentrations 
several magnitudes greater than lysozyme. It was also observed that when the lysozyme 
loading was increased a more complete release was achieved.  The partitioning of 
lysozyme is a saturable phenomenon for a fixed surface area and thus the above 
observation can be readily rationalized.  As the lysozyme concentration in the dispersed 
aqueous phase increased, a lower fraction of the incorporated molecule partitioned into 
the interface.   
Mechanistic studies on lysozyme release demonstrated two modes of delivery, one 
relatively slow lasting about 2 weeks, and a fast mode which was complete within 2-3 hr.  
Results showed that P2, the permeability coefficient related to the slower mode of 
lysozyme release, was correlated with the degree of substitution (DS).  P2 decreased with 
increase of DS indicating a dependence on crosslinking density.  These results indicate 
that the slower mode is related to trans-membrane flux of lysozyme from the nanoparticle 
core load.   
One observation of interest is that the initial extent of release varied widely. One 
hundred (100) % of the encapsulated lysozyme was released in 1 hr for the minimally 
crosslinked nanoparticles while for the nanoparticles with highest crosslinking the release 
 155 
was only 3.5 % for the same time frame.  It is well known that all starch polymers 
(Fundueanu, G., 2004) and dextrans (Denizli, B. K., 2004) swell on exposure to water.  It 
is thus possible that the initial swelling on being introduced to the release medium was 
associated with relatively fast release of short duration.  Another possibility is the fast 
dissolution and release of surface associated lysozyme.  It is to be noted that at the end of 
the reaction, water from the dispersed phase was extracted by the introduction of a large 
volume of ethanol. During the extraction of water from the dispersed phase, migration of 
lysozyme to the surface is a possibility. The surface of the nanoparticles is potentially 
anionic as a result of the mono or di esterified TMC. Thus any superficial cationic 
lysozyme can remain electro statically associated with the anionic surface.   This surface 
associated lysozyme can then be quickly released in simulated physiological fluid, 
resulting in the initial fast mode of release. 
 
5. CONCLUSION 
Several conclusions can be arrived at from the results of the present work.  Firstly, 
both hydrophilic small molecules like tartrazine and moderate sized proteins can be 
successfully incorporated into the carbohydrate nanoparticles developed.  For the small 
molecules, the barrier function of the crosslinked membrane at the interface is not 
adequate to achieve sustained release for most applications. For larger protein molecules 
like lysozyme with a hydrodynamic diameter 4.1 nm, the membrane is adequate to result 
in a significant sustained release of up to at least two weeks.  Further work is needed to 
positively identify the mechanism of faster release associated with the observed delivery 
profile.  Reduction in the extent of protein release during the first 2 hr period would be 
 156 
desirable for sustained release applications, but not necessarily a requirement for 
pharmacokinetic distribution and cellular interaction.    Lastly, though the hydrophilic 
carbohydrate nanoparticles are theoretically ideally suited to evade the opsonization 
process and their subsequent elimination from the blood stream, in vivo confirmation is 
needed to establish the integrity, half life of circulation and also the release pattern of 
encapsulated bioactive macromolecules from this platform.  The physiological 
distribution of the prepared nanoparticles needs to be investigated in detail to unravel the 
entire spectrum of patho-physiological situations in which this platform can be a useful 
tool for medical intervention.   
 
REFERRENCES         
Alderton, G., Ward, W. and Febold, H.  1945.  Isolation of lysozyme from chicken egg 
white.  Journal of Biological Chemistry, 157, 43. 
 
Basu Sarkar, A. and Kochak, G. M.  2005.  HPLC analysis of aliphatic and aromatic 
dicarboxylic acid cross-linkers hydrolyzed from carbohydrate polyesters for 
estimation of molar degree of substitution.  Carbohydrate Polymers, 59, 305-312. 
 
Bezmer, J. M., Radersma, R., Grijpma, D. W., Djikstra, P. J., Feijen, J. and van 
Blitterswijk, C. A.  2000.  Zero-order release of lysozyme from poly(ethylene glycol)/ 
poly(butylene terephthalate) matrices.  Journal of Controlled Release, 64, 179-192. 
 
Chung, Y. I., Tae, G. and Yuk, S. H.  2006.  A facile method to prepare heparin 
functionalized nanoparticles for controlled release of growth factors.  Biomaterials, 
27, 2621-2626. 
 
Denizli, B. K., Can, H. K., Rzaev, Z. M. O. and Guner, A.  2004.  Preparation conditions 
and swelling equilibria of dextran hydrogels prepared by some crosslinking agents.  
Polymer, 45, 6431-6435. 
 
Franssen, O., Stenekes, R. J. H. and Hennink, W. E.  1999.  Controlled release of model 
protein from enzymatically degrading dextran microspheres.  Journal of Controlled 
Release, 59, 219-228. 
 
 157 
Fundueanu, G., Constantin, M., Dalpiaz, A., Bortolotti, F., Cortesi, R., Ascenzi, P. and 
Menegatti, E.  Preparation and characterization of starch/cyclodextrin bioadhesive 
microspheres as platform of nasal administration of Gabexate M P1 esylate (Foy?) in 
allergic rhinitis treatment.  Biomaterials, 25, 159-170. 
 
Gombotz, W. R. and Wee, S. F.  1998.  Protein release from alginate matrices.  Advanced 
Drug Delivery Reviews, 31, 267-285. 
 
Hennink, W. E., Fransen, O., van Dijk-Wolthuis, W. N. E. and Talsma, H.  1997.  
Dextran hydrogels for controlled release of proteins.  Journal of Controlled Release, 
48, 107-114. 
 
Jensen, M., Hansen, P. B., Murdan, S., Frokjear, S. and Florence, A. T.  2000. Loading 
into and electro-simulated release of peptides and proteins from condroitin 4-sulphate 
hydrogels.  European Journal of Pharmaceutical Sciences, 15, 139-148. 
 
Kim, C.  2000.  Controlled release dosage form design. Technomic Publishing Company, 
Inc., Lancaster, Pennsylvania, USA., Pg 20-22. 
 
 Kuijpers, A.J., van Wachem, P. B., van Luyn, M. J. A., Engbers, G. H. M., Krijgsveld, 
J., Zaat, S. A. J., Dankert, J. and Feijen, J. 2000. In vivo and invitro release of 
lysozyme from crosslinked gelatin hydrogels: a model system for the delivery of 
antibacterial proteins from prosthetic heart valves.  Journal of Controlled Release, 67, 
323-336. 
 
Lanaerts, V., Moussa, I., Dumoulin, Y., Mebsout, F., Chounard, F., Szabo, P., Mateescu, 
M. A., Cartilier, L. and Marchessault, R.  1998.  Cross-linked high amylose starch for 
controlled release of drugs: recent advances.  Journal of Controlled Release, 53, 225-
234. 
 
Levy, M.C. and Andry, M.C.  1991.  Microcapsules with walls made of cross-linked 
starch derivatives.  Microencapsulation of Drugs, Pp-7-16, Ed.-Whatley, T. L., Drug 
Targeting and Delivery Series, Harwood Academic publishers, Great Britain.  
 
Liaw, Y. I., Brown, M. B. and Martin, G. P. 2001.  Turbidometric and HPLC assays for 
the determination of formulated lysozyme.  Journal of  Pharmacy and Pharmacology, 
53, (4), 549-554. 
 
Mehvar, R.   2000.  Dextrans for targeted and sustained delivery of therapeutic and 
imaging agents.  Journal of Controlled Release, 69, 1-25. 
 
Mi, F. L., Shyu, S. S., Chen, C. T. and Schoung, J. Y.  1999.  Biomaterials, 20, 1603-
1612. 
 
 158 
Mocanu, G., Mihai, D., Picton, L., LeCerrf, D. and Muller, G.  2002.  Associative 
pullulun gels and their interaction with biological active substances.  Journal of 
Controlled Release, 83, 41-51. 
 
Morah, L. P.  1998.  Microspheres as nasal drug delivery systems.  Advanced Drug 
Delivery Reviews, 29, 185-194. 
 
Pariot, N., Edwards-Levy, F., Andry, M. C. and Levy, M. C. 2000. Crosslinked ?-
cyclodextrin microcapsules: Preparations and properties.  International Journal of 
Pharmaceutics, 211, 19-27. 
 
Richardson, S., C. W., Kolbe, H. V. J. and Duncan, R.  1999.  Potential of low molecular 
weight mass chitosan as a DNA delivery system:  biocompatibility, body distribution 
and ability to complex and protect DNA.  International Journal of Pharmaceutics, 
178, 231-243. 
 
Stenekes, R. J. H., Loebis, A. E., Fernandes, C. M., Crommelin, D. J. A. and Hennink, 
W. E.  2001.  Degradable dextran microspheres for controlled release of liposomes.  
International Journal of Pharmaceutics, 214, 17-20. 
 
Tabata, Y. and Ikada,Y.  1989.  Synthesis of gelatin microspheres containing interferon.  
Pharm. Res., 6, 422-427. 
 
Wulff, G., Steinert, A. and Holler, O.  1998.  Modification of amylose and investigation 
of its inclusion behavior.  Carbohydrate  Research, 307, 19-31. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 159 
Table 1.  Effect of carbohydrate polymer on the encapsulation efficiency of tartrazine. 
 
Crosslinker Polymer Encapsulation 
Concentration (w/v)% Concentration (w/v)% Efficiency of Tartrazine 
TMC  (1.635) Corn Starch (4) 39.09 
TMC  (1.635) Waxy maize starch (5) 32.43 
TC     (1.25) Corn starch (4) 39.09 
TC     (1.25) Waxy maize starch (5) 39.6 
TC     (1.25) MCS  (50) 78.36 
TC     (1.25) MA  (45) 83.89 
 
 
 
 
 
 
 
 
 
 
 
 
 
Table 2.  Accuracy and precision for the determination of  
lysozyme released from carbohydrate nanoparticles.  
 
Calibration level Assay  SD RSD Accuracy 
 meana  (%) (%) 
(?g/ml) (?g/ml)       
1 1.02 0.20 19.95 101.56 
5 4.78 0.09 1.83 95.67 
10 10.10 0.22 2.16 100.98 
20 20.70 0.11 0.55 103.48 
30 30.50 0.15 0.49 101.67 
40 38.84 0.23 0.60 97.09 
SD, standard deviation; RSD, relative standard deviation.  a N = 5. 
 
 
 
 
 
 
 
 
 
 
 
 160 
 
Table 3.  Accuracy and precision for the determination of  
lysozyme in encapsulation efficiency studies of carbohydrate nanoparticles.  
 
Calibration level Assay  SD RSD Accuracy 
 meana  (%) (%) 
(?g/ml) (?g/ml)       
1 0.96 0.229 24.03 95.68 
5 4.98 0.064 1.29 99.59 
10 11.01 0.379 3.44 110.13 
20 21.49 0.033 0.15 107.43 
30 29.65 0.542 1.82 98.83 
40 35.97 0.393 1.09 89.93 
 SD, standard deviation; RSD, relative standard deviation.  a N = 3. 
 
 
 
 
 
 
 
 
 
 
Table 4.  Effect of pH and organic phase solvent on the partitioning behavior of 
lysozyme in the pH range of 3.0-5.0.  Water-organic phase ratio was 1:5.  
 
Solvent  Percent of total lysozyme found in   
       aqueous phase  at different pHs 
   pH 3.0 pH 4.0 pH 5.0 
Methylene chloride  *ND ND ND 
Cyclohexane  73.26 44.08 37.4 
Chloroform  74.9 21.26 14.03 
Cyclohexane: Chloroform    9.95 7.16 4.69 
[68:32, (v/v)]        
*ND, not detected. 
 
 
 
 
 
 
 
 
 
 
 161 
 
Table 5.  Effect of carbohydrate content on the partitioning behavior of  
lysozyme at pH 3.0.  The organic phase consisted of chloroform.   
Water/organic phase ratio was 1:5. 
 
Aqueous phase description Lysozyme remaining 
      in aqueous phase (%)  
0% (w/v) polymer (MA) 74.9  
15% (w/v) polymer (MA) 100  
30% (w/v) polymer (MA) 97.3  
45% (w/v) polymer (MA) 95.3   
 
 
 
 
 
 
 
 
 
 
 
 
 
Table 6.  Encapsulation efficiency and degree of substitution  
achieved in some representative nanoparticle batches. 
 
Batch # Encapsulation Degree of substitution* 
 efficiency  
NC-127 91.24 7.09 
NC-136 95.36 0.159 
NC-138 111 0.133 
NC-139 108.97 0.128 
NC-141 84.99 0.106 
NC-142 96.26 0.059 
*Degree of substitution refers to the number of crosslinker  
molecules reacted to 100 anhydroglucose units. 
 
 162 
0
20
40
60
80
100
0 0.5 1 1.5 2 2.5
Time (day)
Cu
mm
ula
tiv
e (
%
) o
f t
art
ra
zin
e r
ele
as
ed
 
Control solution
Corn starch
Waxy maize starch
MCS
MA
 Figure 1. Comparative release profiles of tartrazine from nanoparticles made with TC as 
crosslinker. 
 
 
 
 
 
 
 
 
 
 163 
0
20
40
60
80
100
0 0.5 1 1.5 2 2.5
TIME (day)
Cu
m
m
ul
ati
ve
 %
 o
f t
ar
tra
zin
e r
ele
as
ed
 
 Control solution
CORN
WAXY MAIZE
  
Figure 2. Comparative release profiles of tartrazine from nanoparticles made with TMC 
as cross linker. 
 
 
 
 
 
 
 
 
 
 164 
0
20
40
60
80
100
0 0.5 1 1.5 2 2.5
Time (day)
Cu
m
mu
lat
ive
 %
 o
f t
ar
tra
zin
e r
ele
as
ed
 
Control solution
CORN/TMC
WAXY MAIZE/TMC
CORN/TC
WAXY MAIZE/TC
 Figure 3. Comparative release profiles of tartrazine from nanoparticles made with TC & 
TMC as crosslinkers. 
 
 
 
 
 
 
 165 
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 20 40 60 80 100 120 140 160 180
TMC Conc. in reaction medium (mM)
Nu
m
be
r o
f T
MA
 m
ol
ec
ul
es
 re
ac
ted
 
to
 10
0 A
GU
9 %    Lys
16 %  Lys
 Figure 4.  Relationship between degree of substitution (DS) and trimesoyl chloride 
(TMC) concentration in the continuum phase.  The effect of lysozyme concentration on 
the degree of substitution is demonstrated at 9 and 16 % protein load. 
 
 
 
 
 
 166 
0
20
40
60
80
100
0 2 4 6 8 10 12 14 16
TIME (day)
Cu
m
ul
ati
ve
 %
 ly
so
zy
m
e r
ele
as
ed
 
[0.132] [0.158] [0.032] [1.069] [7.094]
 Figure 5.  Comparative release profiles of lysozyme from nanoparticle batches with 
different TMC concentrations.  The number in parenthesis indicates degree of 
substitution achieved in terms of number of trimesoyl chloride molecules reacted per 100 
anhydroglucose units of the polymer. 
 
 
 
 
 
 
 167 
 
y = 22.063x-0.4441
R2 = 0.9744
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8
DS
Pe
rc
en
t o
f l
ys
oz
ym
e r
ele
as
ed
 Figure 6.  The relationship between the degree of substitution (DS) and the extent of 
release of lysozyme in 15 days from nanoparticles with different levels of crosslinking.   
 
 
 
 
 
 
 
. 
 
 
 
 168 
                      
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 3 6 9 12 15
Time (days)
Ly
so
zy
me
 R
ele
as
ed
 (r
ela
tiv
e)
P = 0.650
R2 =0.9871  
 Figure 7.  Fitting of the lysozyme release profile data from a representative batch with 
homogeneous diffusion model.  P denotes the predicted permeability coefficient. 
 
                      
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 3 6 9 12 15
Time (days)
Ly
so
zy
me
 R
ele
as
ed
 
Relative
Absolute
P1  35.3
f 1  0.48
P2  0.294
f 2  0.52
R2  0.9940
 Figure 8.  Lysozyme release profile from nanoparticles with DS=0.051 and lysozyme 
loading of 9 %.  P1 and P2 are predicted permeability coefficients, f1 and f2 are the 
fractional contributions from each mode. 
 
 
 
 169 
 
 
 
                               
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 3 6 9 12 15
Time (days)
Ly
so
zy
me
 R
ele
as
ed
 
Relative
Absolute
P1  41.8
f 1  0.41
P2  0.198
f 2  0.59
R2  0.9859
 Figure 9 Lysozyme release profile from nanoparticles with DS=0.159 and lysozyme 
loading of 9 %.  P1 and P2 are predicted permeability coefficients, f1 and f2 are the 
fractional contributions from each mode. 
 
 
 
 
 
 
 
 
 170 
                               
0
0.2
0.4
0.6
0.8
1
1.2
0 5 10 15 20
Time (days)
Ly
so
zy
me
 R
ele
as
ed Absolute
Relative
P1  592
f1   0.47
P2  0.239
f2   0.53
r2   0.9914
 Figure 10.  Lysozyme release profile from nanoparticles with DS=0.088 and lysozyme 
loading of 9 %.  P1 and P2 are predicted permeability coefficients, f1 and f2 are the 
fractional contributions from each mode. 
 
 
 
 
 
 
 
 171 
                     
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 3 6 9 12
Time (days)
Ly
so
zy
me
 R
ele
as
ed
 
Relative
Absolute
P1  48.8
f 1  0.60
P2  0.371
f 2  0.40
R2  0.9909
 Figure 11.  Lysozyme release profile from nanoparticles with DS=0.093 and lysozyme 
loading of 16 %.  P1 and P2 are predicted permeability coefficients, f1 and f2 are the 
fractional contributions from each mode. 
 
 
 
 
 
 
 
 172 
                        
Mechanistic contributions
0.0
0.2
0.4
0.6
0.8
0 6 12 18 24
Time (hr)
Ly
so
zy
me
 R
ele
as
ed
 1stterm
2nd
term
P1  35.3
f 1  0.48
P2  0.294
f 2  0.52
 Figure 12a.  Mechanistic contribution of the two diffusional modes on the lysozyme 
release in a representative batch (DS=0.051).  A look at the scenario at a short time 
frame.  
                        
Mechanistic contributions
0.0
0.2
0.4
0.6
0.8
0 3 6 9 12 15
Time (days)
Ly
so
zy
me
 R
ele
as
ed
 1st term
2nd term
P1  35.3
f 1  0.48
P2  0.294
f 2  0.52
 Figure 12b.  Mechanistic contribution of the two diffusional modes on the lysozyme 
release in a representative batch (DS=0.051).  A look at the scenario at a longer time 
frame.  
 
 173 
 
 
 
 
 
y = -2.5297x + 0.5955
R2 = 0.8909
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.05 0.1 0.15 0.2
Degree of Substitution
Pe
rm
ea
bi
lit
y c
oe
ffi
cie
nt
 (P
2)
 Figure 13.  Relationship between the degree of substitution (number of crosslinker 
molecules reacted to each 100 anhydroglucose units) and permeability coefficient, P2. 
 
 
 
 
 
 
 
 
 
 174 
0
0.2
0.4
0.6
0.8
1
1.2
0 2 4 6 8 10 12 14 16
TIME (day)
Fr
ac
tio
n r
ele
as
ed
? (16.67 %)
?(9.09 %)
P1  48.8
f 1  0.60
P2  0.371
f 2  0.40
P1  41.8
f 1  0.41
P2  0.198
f 2  0.59
 Figure 14.  Comparative release profiles of lysozyme batches with different degrees of 
lysozyme loading.  The number in the parenthesis indicates the percent loading of 
lysozyme compared to the dry mass of the spheres. 
 
 175 
VI. SUMMARY AND CONCLUSION 
 
 
A universal drug delivery platform that can be used beyond mere sustained delivery 
of bioactive molecules and for focused cellular level intervention like drug targeting and 
gene transfection, has been a long sought after goal of the pharmaceutical research 
fraternity.  Due to the physiological complexity of the human body and the enormous 
scope of different scientific disciplines involved, the above goal is far reaching.  
Incremental developments in various related fields are continuously contributing to the 
overall progress.  A number of particulate platforms composed of polymers of different 
physical and chemical characteristics have been developed in the last two decades and 
employed as sustained release injectable drug delivery systems (DDS).  These DDS, 
while successful in sustaining the release of the incorporated bioactive compound, were 
found to be largely inadequate in achieving the above mentioned goal.   
For cellular level intervention, injectable particulate platforms must have a long 
circulating half life in human blood plasma.  A review of current literature reveals that 
there are certain conditions that have to be satisfied by the injectable particulates to be 
able to have a long plasma half life.  The required characteristics of the particulate DDS 
cannot be achieved by using the traditional hydrophobic polymers like poly (lactide), 
poly (lactide-co-glycolide), poly (cyanoacrylate) or poly (alkyl cyanoacrylate)s.   
 176 
In addition to being non toxic, biodegradable and biocompatible, the polymers used in 
particulate platform have to be sufficiently hydrophilic to avoid complement activation 
and their subsequent removal by the reticulo-endothelial system from the human body.  
Additionally, nanoparticles have to be monodispersed and mononucleated in an aqueous 
environment and of small size (< 250 nm) to achieve a long circulating half life in blood 
plasma.  The research presented in this dissertation has been focused on developing 
nanoparticles which have the above mentioned characteristics and provide sustained 
release of incorporated bioactive molecules.  The broader issues involved with cellular 
level intervention need to be addressed separately. 
Nanoparticles have been synthesized by crosslinking carbohydrate polymers at an 
emulsion interface with polyfunctional acid chloride crosslinkers.  Adequate 
characterization of the nanoparticles requires both visualization and quantification of the 
crosslinking achieved.  An analytical method using chemical derivatization and high 
pressure liquid chromatographic techniques has been developed to quantify the degree of 
substitution achieved with the crosslinkers.  This method showed adequate sensitivity and 
selectivity for all the crosslinkers investigated.  Further, the repeatability, accuracy and 
precision achieved were satisfactory.   
The appropriate properties of the carbohydrate polymers to be used in drug delivery 
platforms are biodegradability, biocompatibility, high degree of solubility in aqueous 
medium, minimum retrogradation and high reactivity towards the crosslinkers.  All 
common carbohydrate polymers including cellulose, starch, starch derivatives and 
dextrans are hydrophilic and can be potentially used to produce long circulating 
nanoparticles.   But only starch, starch derivatives and dextrans can be hydrolyzed by the 
 177 
amylases present in the body (biodegradable) and therefore are useful for synthesizing 
particulate injectable drug delivery platforms.    
Results of this work show that nanoparticles can be prepared from a broad spectrum 
of starches and starch derivatives using an interfacial crosslinking reaction.  Larger 
polymers and polymer blends like amylopectin (waxy maize starch), corn starch, 
hydroxyethyl starch from corn, hydroxypropyl starch from tapioka suffer from two major 
disadvantages for being used for this purpose; retrogradation and low aqueous solubility.  
They all show retrogradation in varying degrees.  Retrogradation involves changes in 
three dimensional configuration of the polymer and is thus expected to affect the reaction.  
Since the retrogradation process cannot be controlled, there is a possibility of product 
variability due to different states of retrogradation of starch polymers at the time of 
reaction.  Starches also have low water solubility.  Results show that the degree of 
substitution by crosslinkers is directly proportional to the concentration of the starch 
polymer in the aqueous phase.  High polymer load in the dispersed phase has also been 
shown to improve the encapsulation efficiency of the incorporated bioactive molecules 
both for small molecules and large proteins.   
Lower molecular weight derivatives of starch like maltodextrins have more suitable 
properties.  They produce true solutions in water at a high concentration of 45-50 % (w/v) 
and do not retrograde.  As a result, a relatively high degree of substitution with the 
crosslinkers can be achieved.  The encapsulation efficiency of the incorporated bioactive 
molecule is improved.  The possibility of product variability due to retrogradation is also 
eliminated.  Results show that of the two maltodextrins investigated, maltodextrin from 
 178 
waxy maize starch (amylopectin) is better than maltodextrin from corn starch in terms of 
degrees of substitution and also sustaining the release of incorporated molecules. 
The surfactant is another important material component of the drug delivery system 
(DDS).  Of the different surfactant and surfactant combinations investigated, Span 40? 
was found to the most stable water-in-oil emulsion with cyclohexane and chloroform.  
Span 40? is a nonionic surfactant that is non toxic and permitted for pharmaceutical use.  
When aliphatic polyfunctional crosslinkers like oxaloyl chloride or sebacoyl chloride 
were used, they reacted with the surfactant molecules bringing about phase separation 
within minutes.    
Structural requirements were identified for emulsion polymerization systems.  The 
presence of a bulky group like aromatic or tertiary dialkyl is required in between the acid 
chloride functionalities of the crosslinkers to prevent their reaction with Span 40?, 
probably by steric hindrance.  Diethyl malonyl chloride (DEMC), terephthaloyl chloride 
(TC) and trimesoyl chloride (TMC) are three such crosslinkers which have been 
investigated.  The first one DEMC proved to be a poor choice for this purpose because of 
its poor reactivity over a wide pH range of 3-11.  Only at pH 10 did DEMC show any 
appreciable reactivity.   
Both terephthaloyl chloride and trimesoyl chloride have been found to have adequate 
reactivity.  Nanoparticles could be successfully formed with these crosslinkers with a 
wide variety of starch and starch derivatives.   A review of available literature shows that 
both terephthalic and trimesic acids are found in the drinking water supply in the U.S.A. 
as disinfectant byproducts.  A mechanism based structure-activity relationship analysis 
was carried out to establish the carcinogenic potential of more than 200 drinking water 
 179 
disinfectant byproducts (Woo, Y. T., 2002) including terephthalic and trimesic acids.    
The study concluded that both terephthalic and trimesic acids are unlikely to be 
carcinogenic.  The biocompatibility of terephthalic acid containing copolymers like poly 
(ethyleneglycol terephthalate)/poly (butylene terephthalate) (PEGT/PBT) is well 
established both in vivo and in vitro (Radersma, R. D., 2002; Bezemer, J. M., 2000).  
PEGT/PBT also obtained market clearance from the Food and Drug Administration in the 
form of a degradable cement restrictor (Radersma, R. D., 2004).  Less toxicity concerns 
are expected from the very water soluble trimesic acid.   
Results show that between TC and TMC, the later is a more appropriate choice of 
crosslinker for two reasons.  Comparatively higher degrees of substitution can be 
achieved using TMC as a crosslinker and trimesic acid is more hydrophilic and highly 
soluble in water.  Nanoparticles prepared from TC were more difficult to disperse in 
aqueous suspension, probably due to their relatively non polar nature of the surface and 
the resulting high surface tension.  The optimum choice of the polymer, surfactant and 
crosslinker were thus found to be maltodextrin from amylopectin, Span 40? and TMC, 
respectively. 
One significant achievement in the course of this research is that nanoparticles could 
be prepared with properties appropriate for long circulating half life.  The prepared 
nanoparticles have hydrophilic surfaces (especially when prepared with TMC) and the 
aqueous suspensions could be visualized as discrete, monodispersed and mononucleated 
particles.  Traditional particulate drug delivery platforms prepared from hydrophobic 
polymers are reported to form aggregates in an aqueous environment.  The aggregates 
 180 
can potentially negate the advantages of small size of the particles and can lead to 
thrombosis in the blood vessels and their phagocytotic removal from circulation. 
The results of the investigation indicate that the resistance offered by the crosslinked 
membrane to the diffusional release of small molecules like tartrazine is modest.  
Differences could be detected among the release profiles of tartrazine from nanoparticles 
prepared with different polymer/crosslinker combinations. But, overall it can be 
concluded that achieving clinically significant sustained release of small molecules from 
these nanoparticles is difficult and rather unlikely.  The mesh size of the crosslinked 
membrane of the nanoparticles is probably too large to control and/or sustain the release 
of small molecules significantly.  
 The release of larger molecules like lysozyme, on the other hand, can be controlled 
by the mesh size of the crosslinked membrane.  Data of lysozyme release from 
mathematical modeling by spherical diffusion indicate two independent modes of release.  
In the first mode, the release is very fast, the estimated permeability coefficients are high 
and the release is complete within 2-3 hr.  The second mode shows a much more 
sustained release of lysozyme over a period of weeks.  The estimated permeability 
coefficients are much lower.  A negative correlation between the degree of substitution 
and the permeability coefficient could be established.  These results prove that the second 
mode of diffusional release of macromolecules incorporated in the nanoparticles can be 
controlled.  The developed novel nanoparticular drug delivery system is thus, potentially 
useful for sustained release applications of macromolecules.   
The present research represents a significant development towards achieving long 
circulating drug delivery system that can be used for sustained release applications as 
 181 
well as in future efforts towards targeting.  But some questions remain unanswered.  
Further investigations are needed to identify the cause of the fast mode of lysozyme 
release.  The behavior of the nanoparticles in biological systems needs to be established.  
Surface properties of the nanoparticles must be investigated in detail; quantification of 
surface polarity, opsonization of the nanoparticles in circulation, their plasma half life, 
phagocytotic removal by macrophages and the endocytotic uptake by non-phagocytotic 
cells are issues of extreme interest.  The full potential of the developed novel 
carbohydrate nanoparticles can be assessed when all the above mentioned issues are fully 
understood. 
 
REFERRENCES 
Woo, Y. T., Lai, D., McLain, J. L., Manibusan, M. K. and Delllarco, V.  2002.  Use of 
Mechanism Based Structure-Activity Relationship Analysis in Carcinogenic Potential 
Ranking for Drinking Water Disinfection By-Products.  Environmental Health 
Perspective, 110, 75-87. 
 
Radersma, R. D., Hesseling, S. C., Kaim, P. E., Groot, K. and Bezemar, J.M.  2002.  
Biocompatibility and degradation of poly (ether-ester) microspheres: in-vitro and in-
vivo evaluation.  Biomaterials, 23, 4719-4729. 
 
Bezemar, J. M., Radersma, R. D., Grijpma, D. W., Dijkstra, P. J., Blitterswijk, C. A.. and 
Feijen, J.  2000.  Microspheres for protein delivery prepared from amphiphilic 
multiblock copolymers. 2. Modulation of release rate. Journal of Controlled Release, 
67, 249-260. 
 
Redersma, R. D., Wright, S. J., Tayloe, L. M. John, B. A., Groot, K. and Bezemer, J. M.  
2004.  In Vitro/in Vivo Correlation for 14C-Methylated lysozyme release from Poly 
(Ether-Ester) Microspheres.  Pharmaceutical Research, 21, 484-491.