Linking Proliferation Dynamics and Culture Density to Human Pluripotent Stem Cell Differentiation Proclivity

Human pluripotent stem cells (hPSCs) exhibit a shortened G₁ phase and rapid proliferation compared to somatic cells. Latest findings have suggested that hPSCs are permissive to differentiation during the G₁ phase and that commitment is linked to a lengthened cell cycle. We also demonstrated that stochastic partitioning at hPSC mitosis contributes to the heterogeneous response of stem cells in populations to differentiation stimuli. Therefore, hPSC proliferation and cell cycle regulation are critical for fate determination. Yet, how the proliferation dynamics influence commitment at the population level is unclear.

An approach was employed combining experiments and mathematical modeling. The hPSC cycle distribution under different culture conditions was analyzed by flow cytometry (FC) via multiple methods to ensure data accuracy and consistency. Proliferation was monitored both by traditional cell enumeration methods and automated image analysis developed in the laboratory. States of hPSC self-renewal and differentiation were assessed by FC and real-time PCR based on pertinent marker expression. A multiscale population balance model (PBM) was developed comprising integro-differential equations capturing the proliferation and differentiation kinetics of heterogeneous hPSC ensembles.

Human PSC proliferation was markedly reduced with an order-of-magnitude increase in doubling time as the population density surpassed 50x10⁴ cells/cm². This transition was accompanied by a rise in the fraction of G₁ phase hPSCs. Compared to low-density cultures exhibiting 80-90% NANOG⁺ hPSCs, only 50% NANOG⁺ cells were observed in dense cultures after 4 days. The loss of pluripotency was concomitant with the emergence of 10% of the cells entering quiescence. Four-day differentiation of hPSCs (H1, H9 and IMR90) with unconditioned medium (UM) under low or high initial density resulted in significant differences in the NANOG^-cell fractions.

Further analysis was performed by combining these experimental data with the PBM. Because of the observed loss of pluripotency at higher densities, the model was employed to quantify the extent of induction of differentiation or differentiation ‘strength’, k_diff, in media routinely used for hPSC expansion. Upon leave-one-out cross validation of the model, k_diff was predicted for media containing differentiation factors (e.g. BMP4, activin A). The fractions of primed/committed (NANOG^-) cells in various media over a range of initial hPSC densities and differentiation times were calculated by solving the PBM via parallel computing. Simulation results were verified by running experiments for sets of the corresponding conditions. Moreover, the model was applied to hPSC cultures in microcarrier bioreactors. For both static and microcarrier cultures, the estimated k_diff was lower for media used for hPSC propagation than for those utilized in guiding hPSC commitment.  

Our results suggest that the short G₁ phase acts as a safeguard of hPSC pluripotency. We demonstrated that culture density, which is easily adjustable, influences cell cycle length thereby affecting hPSC fate decisions. The approach presented is generally applicable as it allows the prediction of differentiation rates for multiple hPSC lines and culture conditions. Beyond deepening our understanding of mechanisms controlling hPSC self-renewal and differentiation, these findings will contribute to the rational design of niches and differentiation strategies for the generation of stem cell therapeutics.