|dc.description.abstract||Pancreatic digestion is the primary step in preparing islets of Langerhans for clinical transplantation for restoring euglycemia in Type-1 diabetes patients. The yield of islets from the traditional method developed by Ricordi et al. tends to be variable and depends on several parameters. Operator function (shaking the vessel by hand) also produces variability from case to case and can cause damage to the recovered islets. The purpose of this dissertation is to design an automated digestion unit through the incorporation of an automated horizontally rotating bioreactor (the modified Dynacult reactor or MDR) consisting of a rotating outer cylindrical shell and a counter-rotating core both with hemispherical baffles. The new design is capable of producing reproducible operating parameters for the consistent preparation of islets.
Using computational fluid dynamics (CFD), a discrete phase model (DPM) is used to predict the flow patterns in the MDR and Ricordi chamber. CFD analysis and comparison of flow variable contours between above two digesters indicates that MDR has predictable, controllable fluid shear environment and optimized differential rotation mode at 120 ml/min velocity inlet and 100 rpm rotation speed. The superiority of the MDR model is demonstrated through the comparison of simulation data with that of Ricordi chamber, in which are found higher and lower extremes of shear rate, Reynolds’ number and energy dissipation rate. These characteristics of the newly designed digester can be applied to an automatic mode (controlled rotation speed) instead of manual (shaking) operation.
The particle dispersion test reveals that the particles will move along a certain path controlled by the action of different hydrodynamic forces. Localized turbulent flow can be seen through the transparent shell wall. The complete mixing process occurs in the intermediate region between adjacent vanes and double walls without any sedimentation of particles. In particle tracking tests, a new application in the Image-Pro Premier software called “Fiber Separation and Measurement App” is introduced to accomplish the automatic identification of particle streaks that are formed using digital photography with a slow shutter speed such as 1/40 s. High shutter speed has little impact on the distribution of streak lengths while the lower shutter speed will cause streaks that allow measurements in the low velocity zone of a log scale histogram. Almost 100 per cent of particle velocity determined by the experiment is distributed in the numerical range of velocity 0.01 m/s to 0.2 m/s at 1/40s, chosen as the optimal shutter speed. Lengths of particle streaks give good agreement between experimental and numerical results.
A pig pancreas infused intraductally with collagenase enzyme blend is digested by using both MDR and Ricordi chamber in a parallel test in order to compare the islet yield and other relevant parameters between them. Then, the experimental data such as mean islet diameter and islet equivalent quantification (IEQ) are categorized and presented as two-series histograms for further analysis and comparison. In the example test the donor sow had an exceptionally large pancreas and smaller islets than typical one (many of them are less than 100 μm). The MDR digestion proceeded more rapidly and uniformly than that in the Ricordi chamber, and there is clear evidence that excessive digestion occurred in the MDR. In addition, the digestion process in the MDR is more extensive than in the Ricordi chamber. Exocrine tissue is digested to smaller fragments and there are no islets attached to exocrine tissue in the MDR while 32.61 per cent of the islets are still embedded in or mantled by larger tissue pieces in the Ricordi chamber digests. The main advantages of the MDR digester are recovery of cleaner islets that are easier to purify from the surrounding tissue, higher average shear rate for more rapid, uniform digestion, and lack of extremes in shear rate and energy dissipation rate.||en_US