A Mathematical Model for Predicting In Vitro Stem Cell Proliferation Kinetics.

Abstract

Over 150 active clinical trials of mesenchymal stem cell therapies are underway, testing the efficacy in enhancing solid organ and islet transplant, hematopoetic stem cell transplants or in tissue regenerative strategies with a majority utilizing autologous stem cells. With a number of co-morbidities present in the targeted host, reports have identified adverse effects of these comorbidities such as diabetes, prior radiation, to reduce the ability to sufficiently expand the desired clinical stem cell dose, forcing the recipient to utilize third party donors, carrying the risk of allosensitization. Using mathematical modeling, we developed a method to predict with greater than 90% certainty, the number of cells which would be obtained after full 6-passage ex vivo expansion. Methods: 42 healthy juvenile rhesus macaques, 13 cynomolgus macaques and three human mesenchymal stem cell preparations were ex vivo expanded. A power-law mathematical model was applied to the proliferation data of naïve MSCs in the early days of expansion to predict the fold increase in the stem cell population at a future time point. Based on normal distribution, coefficients in the power law model were determined via the solution of a system of coupled equations. Results: Prospective validation of the predictive model was verified at two different laboratories, (n=7) and (n=6) with healthy macaques. The under-predicting ratios were as follows: at day 20, 8±0.04%, at day 30, 15±0.06%, at day 60, 23±0.13% and at day 90, 39±0.16%, demonstrating a lack of underprediction for determining expanded stem cell doses when inputting day 10 expansion data. Using a confounded data set with nonhuman primates showing comorbidities, the model was also verified in streptozocin-induced diabetic (n=6) and irradiated (n=24) macaques as well as healthy human MSC populations (n=3) with similar predicting power. Conclusion: With early day 10 expansion data, this model consistently predicts the minimum number of MSC obtained over a period of 3-12 weeks in both healthy and diseased hosts. With the ability to predict the numbers of stem cells obtained over time in order to cryopreserve adequate autologous stem cell doses, this method, has the potential to enable autologous therapies for subjects which were previously not considered as viable candidates minimizing the risk of mesenchymal stem cell-based therapies.