Bioreactor Suspension Culture: Differentiation and Production of Cardiomyocyte Spheroids From Human Induced Pluripotent Stem Cells

Abstract

Human induced-pluripotent stem cells (hiPSCs) can be efficiently differentiated into cardiomyocytes (hiPSC-CMs) via the GiWi method, which uses small-molecule inhibitors of glycogen synthase kinase (GSK) and tankyrase to first activate and then suppress Wnt signaling. However, this method is typically conducted in 6-well culture plates with two-dimensional (2D) cell sheets, and consequently, cannot be easily scaled to produce the large numbers of hiPSC-CMs needed for clinical applications. Cell suspensions are more suitable than 2D systems for commercial biomanufacturing, and suspended hiPSCs form free-floating aggregates (i.e., spheroids) that can also be differentiated into hiPSC-CMs. Here, we introduce a protocol for differentiating suspensions of hiPSC spheroids into cardiomyocytes that is based on the GiWi method. After optimization based on cardiac troponin T staining, the purity of hiPSC-CMs differentiated via our novel protocol exceeded 98% with yields of about 1.5 million hiPSC-CMs/mL and less between-batch purity variability than hiPSC-CMs produced in 2D cultures; furthermore, the culture volume could be increased ∼10-fold to 30 mL with no need for re-optimization, which suggests that this method can serve as a framework for large-scale hiPSC-CM production.

A three-dimensional culture system for generating cardiac spheroids composed of cardiomyocytes, endothelial cells, smooth-muscle cells, and cardiac fibroblasts derived from human induced-pluripotent stem cells

“…Apoptotic cells also tended to be located toward the center of C1 spheroids, which suggests that the vascular cells and CFs present in C4 spheroids facilitated access of the culture medium to the spheroid interior. These findings are consistent with previous work supporting the crucial role that non-myocytes play in promoting CM survival particularly where nutrient and oxygen diffusion is limited through release of paracrine factors (Ye et al, 2014;Hodgkinson et al, 2016;Giacomelli et al, 2017;Giacomelli et al, 2020;Munarin et al, 2020;Pretorius et al, 2021).…”

“…The inclusion of hiPSC-ECs, -SMCs, and CFs in CM spheroids increased spheroid size and improved cell viability hiPSCs were differentiated into CMs, ECs, SMCs, and CFs via published protocols or the use of commercially available kits (Yang et al, 2016;Zhang et al, 2019;Kahn-Krell et al, 2021), and flow-cytometry analyses with lineage-specific antibodies [CMs: cardiac troponin T (cTnT), ECs: CD144 and CD31, SMCs: smooth-muscle actin (SMA), CFs: TE-7] confirmed that the purity of each differentiated cell population exceeded 95% (Supplementary Figures S1-S4). Spheroids containing CMs alone (C1) and spheroids containing a 4:2:1 ratio of CMs, ECs, and SMCs (C3) or a 4:2:1:1 ratio of CMs, ECs, SMCs, and CFs (C4) were produced by culturing the indicated proportions of cell types in low-attachment 96-well U plates for 1 week to promote aggregation and fusion (Figures 1A,B), and then in low-attachment 6-well plates on an orbital shaker for the remainder of the culture period (Figures 1C,D).…”

“…Furthermore, whereas apoptotic cells became increasingly common from D7 to D60 in C1 spheroids and C3 spheroids, the proportion in C4 spheroids remained stable throughout the culture period, which suggests that C4 spheroids are sufficiently durable for long-term studies. Notably, all four cardiac-cell types were differentiated from the same line of hiPSCs and, consequently, had the same genetic background, and the ratio of CMs, ECs, SMCs, and CFs (4:2:1:1) in C4 spheroids roughly followed the distribution trends found in human myocardium (Banerjee et al, 2007;Pinto et al, 2016;Gao et al, 2018;Arai et al, 2020;Beauchamp et al, 2020;Daly et al, 2021;Pretorius et al, 2021), but whether this ratio also produces the most native-like cardiac-cell phenotypes in cultured spheroids has yet to be conclusively demonstrated. The present study illustrates that additional optimization of this seeding ratio may be necessary to account for cell loss and restructuring that occurs and techniques for determining the composition in the fully matured tissue may be necessary.…”

“…Future iterations of the cardiac spheroid model should consider incorporating data from both early and late time points for all analyses to provide greater longitudinal understanding. Additionally, the media composition chosen was based on previous work with cardiac cell coculture (Giacomelli et al, 2017;Gao et al, 2018;Noor et al, 2019;Pretorius et al, 2021) but was not optimized or fully defined. The use of FBS, although routine, has previously been shown to lead to dedifferentiation of ECs (Kaupisch et al, 2012;Sriram et al, 2015) and an alternative media formulation may be considered for future model iterations that is cGMP compliant (Zimmermann, 2020).…”

“…When cultured in a three-dimensional (3D) environment, rather than as two-dimensional monolayers, CMs coalesce into spheroids (Fischer et al, 2018;Chang et al, 2020) and tend to develop a more mature phenotype (Jha et al, 2016;Sacchetto et al, 2020;Kai-Li Wang et al, 2021). ECs and SMCs also promote CM maturation (Pinto et al, 2016;Ayoubi et al, 2017;Giacomelli et al, 2017;Zhu et al, 2017;Kwong et al, 2019) while facilitating oxygen and nutrient delivery, which improves cell survival (Garzoni et al, 2009;Zhang et al, 2021a;Pretorius et al, 2021), and CFs contribute to myocardial development both by producing components of the extracellular matrix (ECM) and by forming gap junctions with CMs to support electronic signal transduction (Zhang et al, 2019;Beauchamp et al, 2020;Giacomelli et al, 2020;Li et al, 2020;Pretorius et al, 2021). A range of ratios between these different cell types have been explored in previous studies but a relationship of CM:EC:SMC:CF of 4:2:1:1 is predominantly used (Gao et al, 2018;Arai et al, 2020;Beauchamp et al, 2020;Daly et al, 2021;Pretorius et al, 2021) as it roughly recapitulates the relationships found in native myocardium of myocytes predominating with endothelial cells compromising the greatest non-myocyte population (Banerjee et al, 2007;Pinto et al, 2016).…”