The combination of dextran sulphate and polyvinyl alcohol prevents excess aggregation and promotes proliferation of pluripotent stem cells in suspension culture

Abstract Objectives For clinical applications of cell‐based therapies, a large quantity of human pluripotent stem cells (hPSCs) produced in standardized and scalable culture processes is required. Currently, microcarrier‐free suspension culture shows potential for large‐scale expansion of hPSCs; however, hPSCs tend to aggregate during culturing leading to a negative effect on cell yield. To overcome this problem, we developed a novel protocol to effectively control the sizes of cell aggregates and enhance the cell proliferation during the expansion of hPSCs in suspension. Materials and Methods hPSCs were expanded in suspension culture supplemented with polyvinyl alcohol (PVA) and dextran sulphate (DS), and 3D suspension culture of hPSCs formed cell aggregates under static or dynamic conditions. The sizes of cell aggregates and the cell proliferation as well as the pluripotency of hPSCs after expansion were assessed using cell counting, size analysis, real‐time quantitative polymerase chain reaction, flow cytometry analysis, immunofluorescence staining, embryoid body formation, teratoma formation and transcriptome sequencing. Results Our results demonstrated that the addition of DS alone effectively prevented hPSC aggregation, while the addition of PVA significantly enhanced hPSC proliferation. The combination of PVA and DS not only promoted cell proliferation of hPSCs but also produced uniform and size‐controlled cell aggregates. Moreover, hPSCs treated with PVA, or DS or a combination, maintained the pluripotency and were capable of differentiating into all three germ layers. mRNA‐seq analysis demonstrated that the combination of PVA and DS significantly promoted hPSC proliferation and prevented cell aggregation through improving energy metabolism‐related processes, regulating cell growth, cell proliferation and cell division, as well as reducing the adhesion among hPSC aggregates by affecting expression of genes related to cell adhesion. Conclusions Our results represent a significant step towards developing a simple and robust approach for the expansion of hPSCs in large scale.


| INTRODUC TI ON
Human pluripotent stem cells (hPSCs), including human embryonic stem cells (hESCs) and human-induced pluripotent stem cells (hiP-SCs), 1 exhibit great potential in regenerative medicine and pharmaceutical studies. hESCs, derived from the cell mass within the blastocyst before implantation, 2 and hiPSCs generated by forcedexpressing four transcription factors OCT4, SOX2, KLF4 and C-MYC, 3 are regulated by core transcription factors OCT4 and SOX2, as well as NANOG 4 to maintain their pluripotency. Both have the ability for self-renewal with an unlimited proliferation and can be differentiated into almost all types of somatic cells. 1 Therefore, hPSCs and their derivatives hold great potential for applications in the fields of biomedicine such as cell-based therapies in regenerative medicine, and as a source of cells for use in disease modelling and drug discovery. 5 However, before the applications of hPSCs and their derivatives can fulfil their potential, a significant challenge needs to be overcome with regards to culture scalability to produce clinically relevant numbers of cells (eg each treatment of myocardial infarction or diabetes requires 10 9 cardiomyocytes or 10 9 β-cells, respectively). 6 The conventional method of culturing hPSCs is on static twodimensional (2D) systems, which is only suitable for laboratory-scale studies and cumbersome for scale-up due to limited surface area.
In addition, 2D culture systems fail to mimic the physiological environment in vivo or to provide sufficient signalling for stem cell proliferation with high efficiency and quality. 7 For this reason, threedimensional (3D) systems have emerged as a promising strategy for large-scale production of cells.
Currently, some progress has been made towards the development of suitable 3D culture systems for large-scale expansion of hPSCs and their derivatives. One approach is to use microcarriers or microcapsules, which provide an advantage of increased surface area to volume ratio, providing more space for cell attachment and expansion. 8,9 However, this approach is characterized by its difficulty in cell detachment from microcarriers or microcapsules.
Another strategy is microcarrier-free suspension culture that yields a large number of cells. However, hPSCs tend to aggregate during culturing due to intercellular interactions. The control of aggregate size is crucial for maintaining the pluripotency of hPSCs and also for stable and efficient production of hPSCs. Excess aggregates hinder nutrients and oxygen from diffusing towards their interior, leading to hypoxia and central necrosis within cell aggregates and even affecting the pluripotency and differentiation potential. 10 In general, physical and biochemical approaches are available to inhibit excess aggregation. One approach includes the regulation of agitation conditions in the bioreactor in an attempt to obtain size-controlled aggregation. [11][12][13] It is worth noting that an overly high shear stress produced by agitation could affect cell the viability and differentiation of hPSCs. Bauwens et al modulated aggregate size and shape by using microwells which functioned as physical barriers to limit cell movement. However, such a microwell-based approach relies on the size and number of microwells, which is only applicable to laboratoryscale studies. 14 In addition to physical approaches, some chemical molecules or polymers have been used to control the aggregation of hiPSCs. Horiguchi et al demonstrated that KnockOut Serum Replacement (KSR) and lipid-rich albumin are able to reproducibly prevent hiPSC aggregation without influencing pluripotency. 15 Nath and his colleagues established a simple method for hiPSC aggregate break-up by the addition of botulinum haemagglutinin to culture medium. They found that hiPSC aggregates broken up by HA showed a greater cell viability and expansion compared aggregates dissociated with enzymatic digestion; they reached a maximum cell density of 4.5 ± 0.2 × 10 6 cells/mL. 16 Dextran sulphate (DS), a polysulphated compound, has been used to prevent aggregation of cells in biopharmaceutical industry for decades. DS was recently reported to display aggregate control properties in hiPSCs, while not compromising the pluripotency of the cells. 17,18 Although the physical and biochemical approaches mentioned above could prevent cell aggregation and promote cell expansion efficiency to different degrees, none of these Results: Our results demonstrated that the addition of DS alone effectively prevented hPSC aggregation, while the addition of PVA significantly enhanced hPSC proliferation. The combination of PVA and DS not only promoted cell proliferation of hPSCs but also produced uniform and size-controlled cell aggregates. Moreover, hPSCs treated with PVA, or DS or a combination, maintained the pluripotency and were capable of differentiating into all three germ layers. mRNA-seq analysis demonstrated that the combination of PVA and DS significantly promoted hPSC proliferation and prevented cell aggregation through improving energy metabolism-related processes, regulating cell growth, cell proliferation and cell division, as well as reducing the adhesion among hPSC aggregates by affecting expression of genes related to cell adhesion.

Conclusions:
Our results represent a significant step towards developing a simple and robust approach for the expansion of hPSCs in large scale. approaches were employed with the large numbers of cells that are necessary to meet clinical requirements. Poly(vinyl alcohol) (PVA) is a highly biocompatible and non-toxic synthetic polymer that has a wide range of applications in the medical, cosmetic, food and pharmaceutical industries. 19 In a recent study, Wilkinson et al developed a culture system for the long-term ex vivo expansion of functional mouse haematopoietic stem cells (HSC) where serum albumin was replaced with PVA. 20 They demonstrated that using this albuminfree culture system led to a 236-to 899-fold expansion of functional HSC over one month. To the best of our knowledge, the effect of PVA on hPSC expansion has not been investigated yet.
In the present study, we developed a chemical-based approach for ex vivo hPSC expansion by using a combination of PVA and DS.
Our hypothesis was that PVA would promote the proliferation of hPSCs while DS would modulate cell aggregation. We further postulated that the combination of PVA and DS would not only yield a size-controlled aggregate, but also significantly promote the growth of hPSCs in suspension culture. To test this hypothesis, we investigated the effect of DS and PVA as well as their combinations on aggregate formation, cellular proliferation and pluripotency of hPSCs in both static and dynamic suspension cultures ( Figure 1A). Finally, we assessed the possible mechanisms and advantages of this approach.

| hPSCs culture
The hESC line, H9, was purchased from the WiCell Research Institute (STEMCELL Technologies) medium on plates coated with hESCqualified Matrigel (Corning), and incubated at 37℃ in a humidified atmosphere with 5% CO 2 , and the medium was refreshed every day.
After 5-6 days culture, hPSC colonies were dissociated into single cells using Gentle Cell Dissociation Reagent (GCDR, STEMCELL Technologies). Cells were counted using a haemocytometer, and viable cells were identified by exclusion of trypan blue dye (Thermo Fisher). For subculture, cells were replated in a new culture dish at a viable cell density of 1 × 10 4 cells per cm 2 .

| Suspension expansion of hPSCs
For static suspension culture, hPSC colonies were dissociated with GCDR for 5-7 minutes at 37℃ to obtain a single-cell suspension; then cells were seeded into ultra-low-attachment 6-well plates (Corning) at a cell density of 2 × 10 5 cells per ml and cultured in mTeSR1 medium containing 10 μM Y-27632 (STEMCELL Technologies), polyvinyl alcohol (PVA, MW = 31 000-50 000, Hydrolysis: 87%-89%, Sigma-Aldrich) or Dextran Sulphate (DS, MW = 40 000, Sigma-Aldrich), or the combination of PVA and DS. 60% of the culture medium was replaced with fresh medium without Y-27632 each day. The PVA was supplemented during the entire period of the culture, and DS treatment was employed for only the first two days after inoculation.
Cells were harvested by dissociation with TrypLE (Thermo Fisher) treatment at 37℃ for 15 minutes, and cell counts were performed by Trypan Blue exclusion (Thermo Fisher).
For dynamic suspension culture, hPSCs were cultured in disposable stirred bioreactors (Corning) with a maximum volume of 250 ml.
Briefly, hPSC colonies were digested into single cells by using GCDR.
Then, cells were seeded in a bioreactor at a density of 1 × 10 6 cells per ml, and cultured in mTeSR1 media with Y-27632 added until reaching the working volume. The medium was changed after 48 hours to mTeSR1 without Y-27632, and 80% medium was refreshed every day.
DS was supplemented only on day 1 at a concentration of 100 μg/mL while PVA was supplemented every day at a concentration of 1 mg/ mL. Bioreactor cultures were maintained for 7 days, and the stirring speed was continuously maintained at 60 rpm/min.
Aggregate samples were taken on days 5-7 and placed in a 6-well culture dish (Corning) for photomicrographs using a Nikon D5100 camera attached to a Nikon Eclipse TS100 microscope. Three samples for each condition were taken and imaged. Image contrast and brightness were adjusted by ImageJ. ImageJ was used to measure the diameter of imaged cell aggregates.

| Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Total RNA was extracted from the cells using Universal RNA Extraction Kit (TaKaRa). Following quantification in a Nano Drop ΔCt. The RNA levels were calculated as 2 −ΔΔCt .

| FACS analysis
hPSC spheroids were dissociated into single cells by treatment with TrypLE™ Express (Gibco), and cells were fluorescently labelled by

| Embryoid body formation assay
The pluripotent capability and differentiation potential of hPSCs were evaluated in vitro through the formation of the embryoid body (EB), which directly differentiated into all three germ layers: endoderm, mesoderm and ectoderm lineages in EB medium.
Briefly, hPSCs from the Corning stirred bioreactor were replated on 6-well ultra-low-attachment tissue culture plates (Corning) in

| Teratoma formation assay
All animal procedures were approved by the Animal Ethics Committee

| Cell Counting Kit-8 assay
Cell viability was assessed using the Cell Counting Kit (CCK)-8 assay Wells without cells served as blank controls. Each experiment was performed in triplicate.

| Glucose and lactate analysis
Culture supernatants were collected every day prior to and following medium exchange, and centrifuged at 360 g for 10 minutes to remove dead cells and debris. The cell-free supernatants were analysed using an Automatic Biochemistry Analyser (3100, Hitachi) for concentrations of glucose and lactate. The apparent yield of lactate from glucose was calculated for each day as with ΔLac as the production of lactate and ΔGlc as the consumption of glucose during a given day of culture.

| Statistical analysis
Data are expressed as the mean ± standard deviation (n = 3).
Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software, USA), and statistical significance was determined by Student's t Tests. Differences were considered statistically significant where a P value was <.05.

| RE SULTS AND D ISCUSS I ON
In recent years, there has been a strong drive towards translating basic research of hPSCs into industries and clinics. One of the key elements for successful translational applications is the ability to produce hPSCs in a scalable and quality-controlled manner. As such, a variety of approaches have been taken by researchers for largescale expansion of hPSCs. 8,9 Aggregate-based expansion methods hold great promise for scalable expansion of hPSCs due to the relative simplicity and reduced processing steps required. Nevertheless, most aggregate expansion methods to date generate heterogenous and excess aggregate sizes which have negative influence on cellular viability and differentiation potential.

| Effect of aggregate sizes on hPSC quality in suspension culture
Aggregate size control has been recognized as one of critical parameters for mass production of hPSCs using aggregate-based suspension culture systems. 21 However, few studies have described how that if the aggregate diameter was greater than 300 μm which is the diffusion limit of essential factors, the cells inside the centre of the aggregate would suffer from a lack of oxygen and nutrient transport, and thus damage cell pluripotency. 22 Therefore, it appears that it is necessary to control the size of the aggregates by physical methods such as controlling the stirring rate, and chemical methods such as adding small molecules to obtain uniform cell aggregate products with sufficient exposure to oxygen, nutrients and media growth factors.

| Effect of DS on hPSC static suspension culture
DS is a well-characterized polysulphate compound that has been used to prevent cell aggregation in biopharmaceuticals. 23 of DS on hPSCs. 24 Although the DS treatment could lead to uniform and small aggregate formation, the effect of DS on the proliferation of hPSCs is controversial. In our study, we did not observe DS treatment to promote cellular proliferation, as evidenced by the total cell density (2.99 ± 0.23 × 10 6 cells/mL) which is not significantly higher than that of control group (2.92 ± 0.32 × 10 6 cells/mL) ( Figure 2D).
Our results are consistent with the findings by Lipsitz and his colleagues. 17 However, in another study, Nogueira et al reported that the use of mTeSR1 or mTeSR3D media with 100 µg/mL DS led to a 97 or 106% increase in total cell numbers respectively vs the medias without DS. 18 qRT-PCR results demonstrated that in all the investigated concentrations of DS, hiPSCs expressed high levels of pluripotent marker genes comparable to that of control culture ( Figure 2E).
These findings suggest that DS treatment enables the formation of uniform aggregates without losing pluripotency. Previous studies also demonstrated that adding DS during the cultivation process could control the aggregation characteristics of hPSCs without losing pluripotency. 17

| Effect of PVA on hPSC static suspension culture
PVA, a common and cheap synthetic polymer, has been widely used in biomedical applications for its nontoxicity, non-carcinogenic and bioadhesive properties. 26 To investigated the effect of PVA treatment on hPSC expansion, hiPSCs were plated as single cells in low-attachment plates under static condition, and treated with a variety of PVA concentrations ranging from 0.1 to 10 mg/mL over a five-day culture period, we did not observe any change in aggregate uniformity or aggregate size relative to the untreated condition ( Figure 3A, B). Supplementation with 1 mg/mL PVA resulted in the highest cell density (4.88 ± 0.29 × 10 6 cells/mL) representing a 9.76-fold expansion, which was 1.36 times higher than that of control culture (3.59 ± 0.35 × 10 6 cells/mL) ( Figure 3C). A similar finding was observed for hiPSCs in adherent 2D cultures ( Figure S1), and hiPSCs maintained a typical colony morphology after PVA treatment ( Figure S1A). Compared with control culture, a significant increase in total cell numbers was observed in treatment with 0.5 mg/mL and 1 mg/mL PVA, which was further confirmed by cell viability analysis using CCK-8 assay ( Figure S1B, C). qRT-PCR analysis demonstrated that hiPSCs aggregates treated with PVA had similar expression profiles of pluripotent marker genes, including OCT4, NANOG and SOX2 ( Figure 3D). These findings indicate that PVA treatment could significantly promote hPSC proliferation without causing loss of pluripotency.

| Effect of the combination of PVA with DS on hPSC suspension culture
Next, we determined whether the combined addition of DS and PVA could improve cellular proliferation while reducing aggregate size and size variability. To obtain an optimal formulation, we first inves- As shown in Figure 4A-C, after 5 days in static suspension culture, hiPSCs treated with 1 mg/mL PVA and 100 µg/mL DS had an average aggregate size of 207 ± 67 μm which is slightly larger than that of hiPSCs treated with 100 µg/mL DS alone (179 ± 60 μm), but significantly smaller than that of control culture (293 ± 99 μm). In terms of cell yield, hiPSCs treated with 1 mg/mL PVA and 100 µg/ mL DS showed a 36% increase in the numbers of cells vs the control culture ( Figure 4D).
Similar findings were also observed within the H9 cell line ( Figure S4). H9 treated with 1 mg/mL PVA and 100 µg/mL DS had a more uniform, smaller aggregate morphology with an average aggregate size of 191 ± 38 μm ( Figure S4A, B). Compared with control culture, H9 treated with 1 mg/mL PVA and 100 µg/mL DS showed a 53% increase in total cell numbers ( Figure S4C). Additionally, the expressions of pluripotent genes (OCT4, NANOG and SOX2) ( Figure 4E) and flow cytometry analyses ( Figure S4D, E) suggest that it is an appropriate cultivation process without losing pluripotency.
We then employed this approach in scalable stirred bioreactors hiPSCs were comparable to those in control culture ( Figure 5B).
Finally, we examined the pluripotency of treated hiPSCs at day 7 by immunofluorescence and found that these hiPSCs aggregates co-expressed OCT4, SOX2 and NANOG ( Figure 5C). These findings demonstrated that the combination of DS and PVA enabled the formation of uniform aggregates without causing a loss of pluripotency in the dynamic suspension culture.
Harvested hiPSC aggregates after 7 days of culture in stirred suspension were then dissociated into single cells with GCDR and replated on Matrigel-coated 2D tissue culture plates. These cells were able to successfully return to static culture conditions and then form hiPSCs colonies ( Figure S5A). Representative images show strong expression of surface pluripotent markers, SSEA-4 and TRA-1-81 ( Figure S5B).
The differentiation capability of hiPSCs from dynamic suspension culture was also investigated. Embryoid bodies (EBs) generated from hiPSCs were able to spontaneously differentiate into the three germ layers, as differentiated cells were stained positive for specific markers of three germ layers, GFAP (ectoderm) and Brachyury (mesoderm) as well as SOX17 (endoderm; Figure 5D), and then the EBs were transferred onto Matrigel-coated plates. These cells were also stained by immunofluorescence for germ layer markers GFAP, Brachyury and SOX17 ( Figure S5C).
A teratoma assay was utilized to evaluate the capacity for differentiation in vivo. hiPSCs from suspension culture in spinner flasks were injected into immune-deficient mice and were found to form teratomas containing tissues from three germ layers; for example, adipose tissue (mesoderm), intestinal epithelium (endoderm), as well as neuroepithelium and pigment epithelial (ectoderm) ( Figure 5E and Figure S5D, E).

| mRNA-seq analyses
To investigate the mechanisms underlying the aggregation control and growth factor activity, respectively ( Figure 7C). KEGG pathway analysis showed enrichments for glycolysis or gluconeogenesis, biosynthesis of amino acids, PI3K-Akt signalling pathway, carbon metabolism and metabolic pathways ( Figure 7D). Cellular metabolism is fundamental to all biological activities and is now known to play a pivotal role in dictating whether a cell proliferates, differentiates or remains quiescent. 42 Recent studies of metabolism in stem cells have revealed energy metabolism such as glycolysis, oxidation-reduction process and oxidative phosphorylation must occur in order for cells to acquire sufficient nutrients such as glucose, amino acids, lipids and nucleotides that are necessary to support cell proliferation. 43,44 Taken together, these results demonstrated that the use of PVA could significantly promote hPSC proliferation through improving energy metabolism-related processes, regulating cell growth, cell proliferation and cell division. cell junction, growth factor activity, Wnt and BMP signalling pathway ( Figure 8C). KEGG pathway analysis was also carried out, and the result is shown in Figure 8D. The top pathways include signalling pathways regulating pluripotency of stem cells, Hippo signalling pathway, TGF-beta signalling pathway, PI3K-Akt signalling pathway and Wnt signalling pathway. These results indicated that a combined use of PVA and DS significantly promoted hPSC proliferation through improving energy metabolism-related processes, regulating cell growth, cell proliferation, cell division and reducing the adhesion among hPSC aggregates by affecting expression of genes related to cell adhesion.

| CON CLUS ION
In the present work, we developed a chemical-based method to control cell aggregation and also significantly enhance cell proliferation in hPSC suspension culture. Our results identified that supplementation of DS enables formation of reproducible, homogeneous and controlled hPSC aggregates without losing their pluripotency.
Additionally, PVA treatment significantly promoted hPSC proliferation through improving energy metabolism-related processes. A combination of DS and PVA offers two benefits by forming small aggregates and enhancing cell proliferation. This method might be an improvement over other existing methods because the recipes described here are simple and at low cost, facilitating use in largescale suspension cultures. In this present study, the effect of molecular weight of PVA on hPSC proliferation was not directly described, and further study is required to clarify. In addition, improvements might be made by using a more simply basal medium (eg. E8 or E6), instead of mTeSR1 which contains bovine serum albumin making it not a good candidate to see the effects of PVA and DS. In future research, it will be interesting to investigate whether we can integrate cell expansion and differentiation towards specific lineages (eg cardiomyocytes or hepatocyte) in a single spinner.

ACK N OWLED G EM ENTS
This work was supported in part by the National Key Research

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
XLT conceived and designed the experiments, collected and analysed data, and prepared the manuscript. HBW was responsible for the experiment, figures and data analysis. JHX was responsible for the collection and analysis of the data. NW, QCC, ZYZ, YQQ, JW, XJL and PST contributed to the collection of the data. LXL provided materials. MA Z reviewed and revised the manuscript. YYD and HLC conceived and designed the experiments, and provided financial support.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.