Complete suspension culture of human induced pluripotent stem cells supplemented with suppressors of spontaneous differentiation

Human induced pluripotent stem cells (hiPSCs) are promising resources for producing various type of tissues in regenerative medicine; however, a scalable culture system that can precisely control the cell status for hiPSCs is not developed yet. Utilizing suspension culture without microcarriers or special materials allows for massive production, automation, cost effectiveness, and safety assurance in industrialized regenerative medicine. Here, we found that hiPSCs cultured in suspension conditions with continuous agitation without any microcarriers or extracellular matrix components were more prone to spontaneous differentiation than those cultured in conventional adherent conditions. Adding PKCβ and Wnt signaling pathway inhibitors in the suspension conditions suppressed the spontaneous differentiation of hiPSCs into ectoderm and mesendoderm, respectively. In these conditions, we successfully completed culture processes of hiPSCs including the generation of hiPSCs from peripheral blood mononuclear cells with the expansion of bulk population and single-cell sorted clones, long-term culture with robust self-renewal characteristics, single cell cloning, direct cryopreservation from suspension culture and their successful recovery, and efficient mass production of a clinical-grade hiPSC line. Our results demonstrate that precise control of hiPSC status in suspension culture conditions paves the way for their stable and automated clinical application.


Introduction 35
Human induced pluripotent stem cells (hiPSCs) are promising resources for various types of 36 tissues in regenerative medicine (Takahashi et al, 2007;Yu et al, 2007). To enable cell therapy 37 from hiPSCs, the development of a large-scale manufacturing system is essential because massive 38 cell numbers are required to compose transplantable cells which are enough to rescue the desired 39 physiological function (Chen et  However, utilizing suspension culture without microcarriers or special materials allows for 43 massive production, automation, cost-effectiveness, and safety assurance in industrialized 44 regenerative medicine.  Table 1). These studies have 51 achieved long-term culture and/or mass expansion of hiPSCs and/or human embryonic stem cells 52 (hESCs) in suspension conditions. However, completed processes from clonal hiPSC generation 53 to mass production of hiPSCs based on the precise control of cell status have not yet been 54 achieved. 55 In this study, we have investigated what hampers the stable maintenance of undifferentiated cell 56 states in suspension conditions. HiPSCs cultured in suspension conditions with continuous 57 agitation without any microcarriers or extracellular matrix (ECM) components were more prone 58 to spontaneous differentiation than those cultured in conventional adherent conditions. From 59 screening of candidate molecules to suppress the spontaneous differentiation of hiPSCs, we have 60 identified that inhibitors of PKCβ and Wnt signaling pathways suppress their differentiation into 61 ectoderm and mesendoderm, respectively. In these conditions, we aimed to complete the 62 processes of handling hiPSCs including the generation of hiPSCs with the expansion of bulk 63 population and single-cell sorted clones, long-term culture with robust self-renewal 64 characteristics, single-cell cloning, direct cryopreservation from suspension culture and their 65 successful recovery, and efficient mass production of a clinical-grade hiPSC line.  Furthermore, after adding IWR-1-endo, the inhibitory effect of PAX6 expression was further 145 enhanced, and simultaneously, OCT4 expression was restored to the same level as in the 146 adherent-culture. PAX6 protein expression was also suppressed in hiPSCs treated with IWR-1-147 endo and LY333531 in suspension-culture conditions while its expression increased in 148 suspension conditions with conventional culture medium compared to adherent-conditions 149 ( LY333531 (IWRLY), and adherent conditions (Ad). Hierarchical clustering obtained from these 158 data showed that LY and IWRLY were grouped closely with Ad ( Figure 3C). In contrast, Sus and 159 IWR were both grouped as discrete populations from Ad. Additionally, hierarchical clustering in 160 gene expression among these conditions was supported by principal component 1 (PC1) in the 161 principal component analysis (PCA) ( Figure 3D). In contrast, PC2 represented genes related to 162 the effects of specific inhibitors under these conditions. Next, we investigated the effect of 163 LY333531 and IWR-1-endo in suspension culture. Many genes involved in pluripotency: KLF4 164 and ID1, and epithelial cell-cell interactions: CDH1 (E-cadherin), were significantly up-regulated 165 in IWRLY, while many transcription factors involved in differentiation: PAX2, 3, 5, and 8, SP5, 166 DBX1, and TFAP2B, were down-regulated in IWRLY ( Figure 3E). GSEA and GOEA on down-167 regulated genes in IWRLY showed that the expression of developmentally associated genes, 168 whose expression was elevated in Sus, was generally reduced in IWRLY ( Figure 3F and G).

169
GOEA on up-regulated genes revealed gene sets involved in epithelial cell types ( Figure 3H). As suggest that elevated expression of PKCβ in suspension-cultured hiPSCs could affect the 183 spontaneous differentiation. 184

Simultaneous inhibition of PKCβ and Wnt signals maintains undifferentiated hiPSCs in 185
suspension for long-term culture. 186 We next performed long-term culture for 10 passages in suspension conditions and compared 187 hiPSC growth in presence of LY333531 or Go6983. When hiPSCs were seeded at 4 × 10 5 188 cells/well, the average cell number reached approximately 12-fold after 5 days under both 189 conditions ( Figure 4A and B). After 10 passages, aggregates of hiPSCs cultured in the presence 190 of LY333531 showed a uniform spherical shape, whereas aggregates cultured in the presence of 191 Go6983 were heterogeneously spherical ( Figure 4C). Notably, in LY333531-treated cells,  positive cell numbers were significantly higher than in Go6983-treated samples, as determined by 193 immunostaining ( Figure 4D and E). To evaluate whether hiPSCs cultured in suspension 194 conditions with PKCβ and Wnt signaling inhibitors for 10 passages maintain pluripotency, we 195 performed embryoid body (EB) formation assay. These EBs contained positive cells for TUJ1,196 SMA, and AFP as ectodermal, mesodermal, and endodermal marker, respectively ( Figure 4F). 197 Comparative genomic hybridization (CGH) array analysis showed that hiPSCs cultured long-term 198 in the presence of PKCβ and Wnt inhibitors retained their normal human karyotype ( Figure 4G).

199
These results indicate that, for long-term culture, inhibition of Wnt signaling and PKCβ in 200 suspension conditions was sufficient to maintain the self-renewal, pluripotency, and genomic 201 integrity of hiPSCs. For the clinical applications of hiPSCs, its homogenous mass production is required to obtain 218 sufficient quantities. To test the feasibility of mass production under suspension-culture 219 conditions supplemented with PKCβ and Wnt signal inhibitors, we first performed suspension-220 culture using healthy-donor derived hiPSC lines, 1383D6 and 1231A3, in a 30 mL bioreactor 221 with stirring blade and demonstrated that undifferentiated hiPSCs were efficiently maintained in 222 this condition (  using a perfusion-culture system with 320 mL bioreactor having stirred-wing was performed in 240 GMP-compliant, clinical grade StemFit AK03N medium containing IWR-1-endo and LY333531 241 ( Figure 5A). When the culture was started at 1 × 10 5 cells/mL in 320 mL medium, hiPSCs 242 proliferated nearly 10-fold after 3 to 4 days and reached approximately 1 × 10 9 cells in total to 243 produce ~300 stock vials (1 × 10 6 cells/vial). This large-scale culture was repeated three times 244 (Passages 1, 2, and 3), and hiPSCs obtained at each passage were characterized. These samples 245 showed similar or higher viability (>90%) to that of adherent-culture derived cells ( Figure 5B).

246
When compared to adherent conditions, these samples in suspension-culture showed a similar or 247 higher proliferation rate ( Figure 5C). Cells obtained from large-suspension cultures formed 248 typical hiPSC-like colonies, after vials were thawed and seeded in adherent-culture conditions 249 ( Figure 5D). Flow cytometry detected over 90% of cells positive for pluripotent cell markers 250 expression: TRA-1-60, SSEA4, and OCT4 ( Figure 5E and F). Further, G-band analysis revealed 251 that hiPSCs retained their normal karyotype even after three passages under large-scale 252 suspension conditions ( Figure 5G). When large-scale suspension-culture derived hiPSCs were 253 incubated in each germ layer-specific differentiation medium for 4 -7 days, the expression of 254 early differentiation markers for ectoderm (PAX6 and SOX1), mesoderm (T and PDGFRA) and 255 endoderm (SOX17 and CXCR4) was significantly induced ( Figure 5H). These cells were then 256 directly differentiated into dopaminergic neural progenitors, cardiomyocytes, and hepatocytes to 257 evaluate their differentiation capacity and propensity. There were no differences in the 258 differentiation efficiency toward these lineages ( Figure 5I-M). Hence, these results indicate that 259 the characteristics and quality of clinical-grade hiPSCs cultured in large-scale suspension 260 conditions in the presence of PKCβ and Wnt inhibitors were equivalent to those of hiPSCs 261 maintained under adherent conditions. Taken together, we were successful in mass suspension-262 culture conditions supplemented with Wnt and PKCβ inhibitors. 263 Single cell sorting and expansion of hiPSC subclones cultured in suspension conditions. 264 We sorted a hiPSC line, 201B7, with TRA-1-60 antibody into individual wells of a 96-well plate 265 and attempted to establish subclones by expanding them with serial passages ( Figure 6A). By day 266 28 after this single-cell sorting, we expanded single-cell-sorted hiPSC subclones in suspension 267 culture supplemented with IWR1-endo and LY3333531 ( Figure 6B). On day 28, we examined 268 seven subclones for their cell growth and expression of OCT4 and TRA-1-60. The subclones 269 showed more than three million cells and high ratios of OCT4-and TRA1-60-positive cells 270 ( Figure 6C-E). These results indicate that we have successfully derived single-cell-cloned 271 sublines in suspension culture. 272 Direct freeze and thaw of hiPSCs cultured in suspension conditions. 273 We froze a hiPSC line, 201B7, in suspension culture conditions supplemented with IWR1-endo 274 and LY3333531. Then, we thawed these frozen vials and directly reseeded the cells in suspension that we have successfully frozen and thawed hiPSCs in suspension culture directly. 281 Together, these results suggest that the suspension culture conditions supplemented with Wnt and 282 PKCβ inhibitors are robust in harsh conditions such as single-cell expansion and freezing-and-283 thaw processes. 284

Establishment of hiPSC lines in complete suspension-culture conditions supplemented with 285
inhibitors of PKCβ and Wnt signaling pathways. 286 For the clinical use of hiPSCs, more secure and efficient establishment of hiPSCs in closed 287 systems is strongly desired. To achieve this, we aimed to establish hiPSCs in suspension-culture 288 conditions. Using PBMCs as a starting material, we generated hiPSCs using a novel replication- days. Cell aggregates cultured with IWR1-endo and LY33351 showed uniform spherical structure 294 ( Figure 7B), and most of the cells were positive for OCT4 ( Figure 7C) and TRA-1-60 ( Figure  295 7D) on day 56. These bulk cells were able to differentiate into three germ layers in an in vitro EB 296 formation assay ( Figure 7E) and in teratomas that were transplanted into immunodeficient 297 NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice ( Figure 7F). These bulk-hiPSCs maintained 298 normal karyotype ( Figure 7G). By single cell sorting with fluorescent labeled TRA-1-60 299 antibody, we expanded single-cell-derived clones in suspension-culture conditions. A clone (F-300 10) showed OCT4 and TRA1-60 expressions ( Figure 7H-J). In addition, the established clone 301 showed potency to differentiate into derivatives of three germ layers in vitro as EBs ( Figure 7K) 302 and in vivo as teratoma transplanted into NSG mice ( Figure. 7L). A normal karyotype was 303 observed in this clone ( Figure 7M). SeVdp was nearly extinct in these bulk populations and F-10 304 clone ( Figure 7N). These results demonstrate that, we were successful in establishing transgene-305 free hiPSC lines using SeVdp infection in suspension-culture conditions in the presence of PKCβ 306 and Wnt inhibitors.  In this study, we have developed a series of methods to generate and maintain hiPSCs in 336 suspension conditions. First, we have identified compounds that suppress spontaneous 337 differentiation of hiPSCs in suspension cultures. Based on these findings, we have newly 338 achieved a complete series of culture processes including hiPSC establishment, long-term culture, 339 mass culture, single-cell cloning, and direct freeze and thaw (summarized in Table 1). These 340 methods require no special materials including microcarriers or dialysis bags, and are effective 341 with various existing culture media and many hiPSC lines. Therefore, this suspension-culture of 342 hiPSCs is anticipated to have novel industrial applications, which have been never achieved with 343 conventional methods because suspension-culture conditions, which precisely control cellular 344 states, have advantages in terms of mass culture, automation, and safety assurance. 345 We have identified two compounds as factors that ameliorate the disadvantages of the suspension 346 culture of hiPSCs. Wnt signaling inhibitors specifically inhibited differentiation toward 347 mesendoderm, whereas, the PKCb inhibitor, LY333531, inhibited ectodermal differentiation and, 348 when combined with IWR-1-endo, maintained the undifferentiated nature of hiPSCs with high 349 efficiency. Go6983, a pan-PKC inhibitor, has been used to promote the self-renewal of performed by single-cell sorting with flow cytometry and were successfully characterized for 366 their self-renewal, pluripotency, and genomic integrity. These results indicate that hiPSC lines 367 generated in suspension conditions have the same quality in terms of their pluripotency, self-368 renewal, and genomic integrity as those generated with conventional adherent conditions. Precise 369 control of hiPSC status in suspension culture could pave the way for their stable and automated 370 clinical application toward autologous cell therapy of hiPSCs. 371 372

374
Human subjects 375 The generation and use of hiPSCs were approved by the Ethics Committee of RIKEN 376 BioResource Research Center, Graduate School of Medicine of Kyoto University, and KANEKA 377 CORPORATION.

379
Animal experiments 380 All animal experiments were approved by the Animal Experimentation Committee of the RIKEN 381 Tsukuba branch and were performed according to the committee's guiding principles and the 382 "Guide for the Care and Use of Laboratory Animals" published by the National Institute of 383 Health.

385
Data availability 386 All data are available in the main text or the supplementary materials. RNA-seq data are available 387 in the Gene Expression Omnibus as GSE222833. All the data supporting the findings of this 388 study are available from the lead contact upon reasonable request. Source data is provided with 389 this paper. This paper does not report original code. Ajinomoto, Co., Ltd., Tokyo, Japan). Medium change was performed every day and passaged at 401 80%-90% confluency after 6-7 days of culture. At passage, PBS/EDTA solution (0.5 mM, 402 Cat#06894-14, Nacalai Tesque) was used to dissociate hiPSC colonies, and these cells were 403 seeded at a density of 2,500 cells/cm 2 . 10 µM Y-27632 (Cat#HY-10071, FUJIFILM Wako Pure 404 Chemical Corporation, Osaka, Japan) and 0.25 µg/cm 2 iMatrix-511 silk (Cat#892021, Matrixome 405 Inc, Osaka, Japan) were added to the culture dish on seeding day. 406 407 Suspension-culture of hiPSCs 408 Suspension culture with rotation at 90 rpm was performed with a plate shaker (Cat#WB-101SRC, 409 WAKENBTECH, Kyoto, Japan or #0081704-000, TAITEC, Tokyo, Japan) installed in a CO2 410 incubator (Cat#Steri-Cycle i160, Thermo Fisher Scientific, MA, USA) and operated under high 411 humidity continuously during the whole culture period. To start the culture hiPSCs in suspension 412 conditions, 4 × 10 5 cells were seeded in one well of a low-attachment 6-well plate (Cat#MS-413 80060, Sumitomo Bakelite, Tokyo, Japan) with 4 mL of StemFit AK02N medium, 414 StemFitAK03N (Cat#AK03N, Ajinomoto, Co., Ltd., Tokyo, Japan), or StemScale PSC 415 suspension medium (A4965001, Thermo Fisher Scientific, MA, USA) supplemented with 10 µM 416 Y-27632. This plate was placed onto the plate shaker in the CO2 incubator. The medium without 417 Y-27632 was changed every day unless otherwise specified. On days 3 -5, the hiPSC aggregates 418 were dissociated with Accutase (Cat#12679-54, Nacalai Tesque) at 37°C for 10 min and used for 419 the next suspension-culture. This passage was performed every 5 days unless otherwise specified. 420 To screen for factors that inhibit spontaneous differentiation of hiPSCs, chemicals or recombinant 421 proteins were added to the culture medium (listed in Supplementary Table S1).

423
Bioreactor culture of hiPSCs 424 A frozen stock of hiPSCs was pre-cultured twice in adherent conditions to prepare enough cell 425 numbers. To prepare enough hiPSCs to start large-scale culture, hiPSCs were pre-cultured in system BioFlo320 (Eppendorf, Hamburg, Germany) was used for large-scale hiPSCs culture. To 432 maintain the lactate concentration below a certain level and to regulate the pH, the culture was 433 carried out by increasing the amount of medium perfusion per unit time in accordance with the 434 cell proliferation transition. To prevent pH decrease, CO2 concentration was regulated by 435 feedback control in the reactor system. After 3-4 days of culture, the formed hiPSC aggregates 436 were dissociated with TrypLE SELECT (Cat#A12859-01, Thermo Fisher Scientific, MA, USA) 437 collected for making cell stocks (~300 tubes). This perfusion culture was repeated three times 438 (P1, P2, and P3) and the cells were prepared at each expansion step.

440
HiPSC generation in suspension culture conditions 441 HiPSCs were generated from healthy donor-derived PBMCs (Precision for Medicine). Thawed 442 PMBCs from a vial containing around 1 × 10 7 cells were pre-cultured in one well of low 443 attachment 6-well plate including 4 mL of StemSpan-AOF (Cat#ST100-0130, STEMCELL 444 Technologies, Vancouver, Canada) supplemented with recombinant human IL-6 (100 ng/mL), 445 IL-3 (10 ng/mL), SCF (300 ng/mL), TPO (300 ng/mL), and FLT3 ligand (300 ng/mL) (all from 446 FUJIFILM Wako Pure Chemical Corporation, Tokyo Japan). After 24 h of incubation with 447 continuous stirring at 37°C and 5% CO2, PBMCs were spun down with centrifugation at 200 × g 448 for 10 min at low deceleration speed and resuspended in StemSpan ACF for cell counting. 2.5 × 449 10 6 cells of PMBCs were centrifuged at 200 × g for 10 min with low deceleration speed, and 450 electroporated using Nucleofector 2b device (Lonza, Basel, Switzerland) with iPS cell Generation 451 Episomal vector Mix (Takara Bio, Shiga, Japan) and Amaxa Human CD34 + Cell Nucleofector kit 452 (Lonza, Basel, Switzerland) according to manufacturer's protocol. 5 × 10 5 of electroporated cells 453 were seeded in one well of a low attachment 6-well plate in 4 mL StemSpan ACF with the 454 cytokines mentioned above. Suspension-culture was performed with continuous agitation at 90 455 rpm. Stem span ACF medium were gradually replaced with StemFit AK02N medium. Formed 456 cell aggregates were passaged with Accutase on day 16 in the presence of 10 µM Y27632, and 457 suspension-culture was continued until the cell numbers reached a sufficient amount for 458 characterization. 10 µM IWR-1-endo and 1 µM LY333531 were added from day 3. 459 SeVdp infection was performed with Sendai Reprogramming Kit (CytoTune EX-iPS of 460 virus solution, ID Pharma, Tsukuba, Japan) according to the manufacturer's protocol with some 461 modifications. Briefly, pre-cultured 1 × 10 6 PMBCs were centrifuged at 200 × g for 10 min, with 462 low deceleration speed, and resuspended in 2 mL of StemSpan ACF with cytokines. PMBCs were 463 gently mixed with 2 mL of virus solution prepared at MOI = 5 per 1 × 10 6 cells. 5 × 10 5 infected 464 cells were seeded in one well of low-attachment 6-well plate at a total volume of 4 mL with 465 StemSpan ACF plus cytokines. Suspension-culture was initiated with continuous agitation at 90 466 rpm. Stem span ACFs were gradually replaced with StemFit AK02N as mentioned above, and the 467 cell aggregates were passaged with Accutase on day 16 in the presence of 10 µM Y-27632. 468 Suspension-culture and passages continued until cell reached a sufficient number for 469 characterization. 10 µM IWR-1-endo and 1 µM LY333531 were added from day 23.

593
Karyotyping 594 For virtual karyotyping, genomic DNA (gDNA) was extracted from hiPSCs using a DNeasy 595 Blood & Tissue Kit (Cat#69504, QIAGEN, Hulsterweg, Netherland) and was used for microarray 596 assay. Virtual karyotyping was performed with GeneChip Scanner System 3000 using Karyostat 597 Assay arrays (Cat#905403, both from Thermo Fisher Scientific, MA, US) according to the 598 manufacturer's protocol. Data were analyzed using the Chromosome Analysis Suite (ChAS) and 599 Affymetrix GeneChip Command Console software programs. 600 G-band analysis was performed using the common Giemsa staining method with hiPSCs 601 fixed by Carnoy's fixtative (3:1 ratio of methanol: glacial acetic acid). 602 603 qRT-PCR 604 Total RNA was extracted with a FastGene RNA premium kit (Cat#FG-81250, NIPPON Genetics, 605 Tokyo, Japan) and used for reverse transcription reaction. cDNA was synthesized by using a 606 ReverTra Ace qPCR RT kit (Cat#FSQ-101, TOYOBO, Osaka, Japan) with random primers. 607 Real-Time qPCR reactions were performed with a QuantStudio 3 System (Thermo Fisher 608 Scientific) using THUNDERBIRD Probe qPCR Mix (Cat#QPS-101, TOYOBO) with TaqMan  609 probes (listed in Supplementary Western software program (ProteinSimple). GAPDH was used as the reference for normalization. 646 The primary antibodies used in this study are listed in Supplementary Table 3.

648
Flow cytometry 649 Flow cytometry of self-renewal markers, TRA-1-60 and SSEA4 was performed as described in 650 our previous study. Briefly, adherent-or suspension-cultured hiPSCs were dissociated with 0. The primary and secondary antibodies used in this study are listed in Supplementary Table 3 and  661 4, respectively. Flow cytometry was performed with SH800 cell Sorter (SONY, Tokyo, Japan). 662 663

RNA-Seq 664
Total RNA was extracted using the FastGene RNA premium kit, and strand-specific library 665 preparation was performed. The prepared library was sequenced using a NovaSeq6000 (Illumina,  666 Inc, CA, USA). Sequencing was performed in a 150 bp × 2 paired-end conFigureuration with a 667 data output of about 6 Gb per sample (~ 20 million paired reads). Library preparation and 668 sequencing was performed in GENEWIZ (Azenta, MA, USA     Chromosomal copy numbers detected with CGH array analysis of PAX6-TEZ (top) and SOX17-997