Scale-down optimization of a robust, parallelizable human induced pluripotent stem cell bioprocess for high-throughput research

Abstract

Human induced pluripotent stem cell (hiPSC) derived therapeutics require clinically relevant quantities of high-quality cell populations for applications in regenerative medicine. The lack of efficacy exhibited across clinical trials suggests deeper understanding of the networks governing phenotype is needed. Further, costs limit study throughput in characterizing the artificial niche relative to outcomes. We present herein an optimized strategy to enable high-throughput hiPSC expansion at <20 mL research scale. We assessed viability of single cell inoculation and aggregate preformation to facilitate proliferation. We modeled aggregate characteristics against agitation rate. Our results demonstrate tunable control with fold expansion comparable to commercial systems. Marker quantification and teratoma assay confirm functional pluripotency. This approach constitutes a scalable protocol to accelerate hiPSC research, and a significant step in advancing the rate of progress in elucidating links to derivative functionality. This work will enable statistically rigorous studies targeting hiPSC and downstream phenotype for clinical manufacturing.

Keywords: Human induced pluripotent stem cells; Optimization; Stem cell bioprocessing; Stirred Tank Bioreactors.

Graphical abstract

Fig. 1

(a) Comparison of scale-down bioreactor outcomes post-inoculation at day 0 and day 4 for single-cell (SC) and aggregate-preformation (APF) conditions. Growth kinetics and viability over culture, alongside day 4 post-harvest measurements (c, d). (e) Morphological comparisons of aggregate preformation in static 12-well plates at t = 12 h for cell concentrations ranging from 1 × 10⁴ – 4 × 10⁴ cells/cm². (f) Aggregate diameter distributions following image processing at 12 h following inoculation. All scale-bars are 500 µm. Values are reported as mean ± standard deviation. The symbols * and ** denote error probability thresholds at 0.05 and 0.01, respectively.

Fig. 2

(D-4 to D1) Pre-expansion protocol, including preparation of preformed aggregates for dynamic inoculation. (D1-D4) Dynamic expansion protocol, including manual sampling strategy for assessment of growth kinetics and morphology over culture time course. All scale-bars are 500 µm.

Fig. 3

(a) Characteristic morphology of static colonies prior to aggregate formation, preformed aggregates inoculated for each agitation rate, and resulting aggregates on day 4. (b) Growth kinetics observed during expansion, coupled with calculated fold expansion. (c) Fold expansion by agitation rate. (d) Aggregate size distributions derived from image-processing by agitation rate on day 4. (e) Cell viability by agitation rate on day 4. All scale-bars are 500 µm. Values are reported as mean ± standard deviation. The symbols * and ** denote error probability thresholds at 0.05 and 0.01, respectively.

Fig. 4

(a) Histological section of excised teratoma. Regions of interest are highlighted: (i) early neural rosette-like formation, (ii) chondrocyte-like cells, (iii) organization resembling glandular epithelium, (iv) organization of intestinal Paneth- and goblet-like tissue. Scale bar is 500 µm. (b) Fluorescence images for lineage differentiation markers of endoderm (GATA4), mesoderm (Brachyury), and ectoderm (TUBB3) in teratoma. All scale bars are 500 µm. (c) Visualization of P1 and P2 gating relative to a single sample. (d) Comparison of marker expression by agitation rate. (e) Comparison of heterogeneity and relative expression of TRA-1–81 against SSEA-4. (f) Comparison of marker expression between pooled dynamic and static single-passage culture. The symbols * and ** denote error probability thresholds at 0.05 and 0.01, respectively.