Mesenchymal and induced pluripotent stem cell–based therapeutics: a comparison

Abstract Stem cell–based cell therapeutics and especially those based on human mesenchymal stem cells (hMSCs) and induced pluripotent stem cells (hiPSCs) are said to have enormous developmental potential in the coming years. Their applications range from the treatment of orthopedic disorders and cardiovascular diseases to autoimmune diseases and even cancer. However, while more than 27 hMSC-derived therapeutics are currently commercially available, hiPSC-based therapeutics have yet to complete the regulatory approval process. Based on a review of the current commercially available hMSC-derived therapeutic products and upcoming hiPSC-derived products in phase 2 and 3, this paper compares the cell therapy manufacturing process between these two cell types. Moreover, the similarities as well as differences are highlighted and the resulting impact on the production process discussed. Here, emphasis is placed on (i) hMSC and hiPSC characteristics, safety, and ethical aspects, (ii) their morphology and process requirements, as well as (iii) their 2- and 3-dimensional cultivations in dependence of the applied culture medium and process mode. In doing so, also downstream processing aspects are covered and the role of single-use technology is discussed. Key points • Mesenchymal and induced pluripotent stem cells exhibit distinct behaviors during cultivation • Single-use stirred bioreactor systems are preferred for the cultivation of both cell types • Future research should adapt and modify downstream processes to available single-use devices


Introduction
The application of stem cells in medicine is by no means a modern idea, and with the cell therapy market size expected to grow to $45 billion by 2030 (Vision Research Reports 2022) due to strong financial backing and lower regulatory hurdles, especially in North America (Polaris Market Research 2022), a closer look at the cells driving this trend is definitely warranted. General interest in the regenerative properties of stem cells first began as far back as 1867, when Cohnheim (1867) observed how non-hematopoietic cells migrated to the site of inflammation and differentiated to fibroblasts during wound healing. A century and a half of research finally led to the identification (Rekers 1950) and general characterization (Tavassoli and Crosby 1968;Friedenstein et al. 1970;Owen and Friedenstein 1988;Pittenger et al. 1999;van den Bos et al. 2014) of hMSCs. These cells not only have the ability to replace damaged tissue via differentiation, but also produce and secrete chemo-and cytokines, modulating local immune response and tissue regeneration (van den Bos et al. 2014;Zhou et al. 2019). In this context, hMSC-derived extracellular vesicles have also been the target of recent cell-free therapeutics. A more comprehensive review of these products may be found in the publication by Kou et al. (2022).
Currently, more than 1400 clinical trials have been submitted involving the application of hMSCs as regenerative medicine products or biologics in the USA alone (National Library of Medicine (US) 2022), and more than 27 products containing hMSCs have been approved in several countries worldwide (Table 1). Of these 27 products, 17 are allogeneic, meaning the cells from a single healthy donor are used to treat multiple patients, while only 10 are autologous, i.e., the donor is the patient, displaying a clear trend towards allogenic. The first such product to be granted orphan drug status by the European Union and to receive market approval in 2009 for the treatment of perianal fistulas resulting from Crohn's disease or inflammatory bowel disease was Alofisel (European Medicines Agency 2009). Since then, many more have followed, e.g., for the treatment of osteoarthritis (Ha et al. 2019), graftversus-host disease (Heathman et al. 2015), and spinal cord injury (Cofano et al. 2019).
Compared to hMSCs, hiPSCs are a far more recent discovery (Takahashi and Yamanaka 2006;Takahashi et al. 2007a;Flahou et al. 2021). These cells are characterized by their inherent capacity for indefinite self-renewal and ability to differentiate into all three germ layers (endo-, meso-, and ectoderms), which allows a broader range of indications to be targeted than would be possible with hMSCs. To this end, their suitability for the treatment of diseases currently considered challenging using conventional means, such as macular degeneration, ischemic stroke, and cancer (Shiba et al. 2016;Takagi et al. 2019), has been repeatedly demonstrated. In fact, more than 27 hiPSC-derived cell products are currently undergoing pre-clinical to phase 3 trials (National Library of Medicine (US) 2022), 7 of which are in phases 2 and 3 (Table 1). Of these 7 products, all are based on the allogeneic manufacturing approach. Companies currently leading the development of such hiPSC-based cell therapeutics include Astellas Pharma, Cynata Therapeutics Ltd., and ReNeuron Ltd.
Regardless of whether the cell therapy is based on hMSCs or hiPSCs, it stringently requires the production of clinically relevant cell quantities of between 10 5 and 10 12 per dose while ensuring target cell quality (viability, marker profile, potency), both of which have been linked to therapeutic efficacy and are outlined in more detail elsewhere (Dominici et al. 2006;Sullivan et al. 2018;Scibona and Morbidelli 2019). This review therefore focuses on the upstream and downstream processing for both hMSC and hiPSC-based cell therapeutics. Furthermore, it highlights successfully implemented single-use (SU) devices, while considering cell typical characteristics and requirements.

Origin, safety, and ethical aspects
The production process of hMSC-and hiPSC-based therapeutics always begins with the extraction of tissue from a willing donor (see also Fig. 2). For example, hMSCs, which belong to the group of multipotent adult stem cells, are easily accessible and present in almost every human organ (Audet and Stanford 2009). To date, no safety concerns have been reported regarding their use, likely due to their immune-privileged status (Najar et al. 2022). In addition, their use is not ethically objectionable, as they are obtained from consenting individuals (Cofano et al. 2019). Historically, the most important source of hMSCs has been bone marrow (Audet and Stanford 2009). However, larger quantities of hMSCs can easily be obtained from adipose tissue as a byproduct of liposuction (Timaner et al. 2020) or from the umbilical cord immediately after birth (Audet and Stanford 2009). Consequently, the umbilical cord-derived subtype has also exhibited a superior life span, lower risk of contamination, and better immunological compatibility compared to the bone marrow-derived subtype (Mahmood et al. 2018).
In contrast to hMSCs, hiPSCs are pluripotent stem cells produced by reprogramming (Takahashi et al. 2007a;Hsu et al. 2018) of somatic skin (Takahashi et al. 2007b;Yu et al. 2007) or blood cells (Zeng et al. 2017) in the laboratory. This reprogramming was initially realized through the introduction of 4 factors (Oct3/4, Sox2, Kfl4, and c-Myc) via a viral vector transport vehicle which integrated them into the host cell DNA. The forced expression of these factors returned the cells to an embryonic state, allowing them to once again differentiate into all germ layers (Takahashi et al. 2007a). In this manner, the ethical issues of using promising embryonic stem cell-derived therapeutic products (Menasché 2020) were circumvented, earning Shinya Yamanaka, the researcher who pioneered the method with murine cells in 2006 (Takahashi and Yamanaka 2006), the Nobel Prize in 2012 (The Nobel Prize 2022).
Although the generation of hiPSCs is ethically acceptable and technically simple, it still remains inefficient, time-consuming, and expensive (Borgohain et al. 2019). Currently, it takes several months until sufficient cells are available for experiments. Also, reprogramming efficiencies reported in the literature are generally between 0.001 and 1%, meaning that at most, only 1 out of 100 cells is successfully reprogrammed into a hiPSC (Birbriar 2021).

3
Another important consideration is safety. The integration of the previously mentioned factors into the genome of hiP-SCs during reprogramming increases the risk of interference with other important genes, which may lead to tumor formation in vivo following implantation (Yu et al. 2007), even after directed differentiation (Lee et al. 2013;Kojima et al. 2019). Other risks include the occurrence of genetic abnormalities during ex vivo cultivation and tissue rejection by the patient's immune system, neither of which can be completely precluded at present (Sullivan et al. 2018;Rehakova et al. 2020). Consequently, hiPSCs are considered less safe than hMSCs, which also explains the lack of commercially available hiPSC-based therapeutics to date. Regardless, researchers are working expeditiously towards developing more safe and efficient reprogramming techniques (Rajasingh et al. 2021). These include the use of more responsive and accessible tissues (Rajasingh et al. 2021), more potent and lower risk gene combinations for reprogramming (Okita et al. 2007;Yu et al. 2007;Furukawa et al. 2022), and the use of non-integrating vectors or even completely non-viral methods (Birbriar 2021).

Cell characteristics and impact on cultivation conditions
Once a cell line has been established, a closer look at its characteristics must be taken to ensure optimal cell growth. In this regard, both hMSCs and hiPSCs require a temperature of 37 °C and a pH between 6.9 and 7.5. In addition, both cell types can be cultivated under normoxic as well as hypoxic conditions (Antebi et al. 2018), are considered shear sensitive (Horiguchi et al. 2021;Burns et al. 2021), and are strictly adherent. The latter characteristic describes their need for a planar surface or scaffold to survive and self-renew. These cells are therefore either cultivated in static cultivation systems as a monolayer (2D culture), or in mechanically or hydraulically driven dynamic bioreactors (3D culture), as spheroids (cell aggregates), or on artificial scaffolds, such as membranes, macrocarriers or microcarriers.
In both 2D and 3D cultivations, differences in size, morphology, and motility impact spatial requirements and must therefore be considered. For example, in suspension, single hMSCs and hiPSCs are roughly spherical with a similar diameter of approximately 18 μm (Pittenger et al. 2019) and 16 μm (Lipsitz et al. 2018), respectively, and are able to form and proliferate as spheroids (Allen et al. 2019). However, after attachment to an artificial scaffold, predominantly via integrin (Jin et al. 2012;Isomursu et al. 2019), their morphology and size differ significantly (Fig. 1). While hMSCs adopt a fibroblast-like morphology and require between 450 and 35,000 μm 2 per cell (Haasters et al. 2009), hiPSCs only require ≈43.5 μm 2 due to their epithelial morphology (Wakao et al. 2012;Courtot et al. 2014). Hence, more hiP-SCs can be grown per available surface area before confluency is reached. They also display a high nucleus-to-cytoplasm ratio (Wakao et al. 2012), a characteristic associated with proliferative potential .
The effective use of limited spatial resources can further be optimized through single-cell passaging and by ensuring uniform cell distribution during inoculation. In this regard, hMSCs are less susceptible as they display higher motility (Bertolo et al. 2015;Somaiah et al. 2015) and are able to migrate to sites with lower relative cell occupation, delaying the onset of localized confluency and contact inhibition. In contrast, hiPSCs remain in very close proximity to their point of initial attachment (Zhang et al. 2011) and are therefore more susceptible to inhomogeneous inoculation. Regardless, the migratory capacity of hMSCs has been shown to decrease and their size increase, as replicative senescence sets in (Haasters et al. 2009;Bertolo et al. 2015), Practically, cell cultivation and purification results in cell loss and far lower yields must be expected. Consequently, the manner and duration of these steps have a significant impact on cell yield and quality per batch, following cultivation, harvest, purification, and cryogenic storage, which has prompted the development of bioprocessing solutions to meet these needs.

Manufacturing overview
Typical bioprocessing steps for hMSC-and hiPSC-based therapeutics are outlined in Fig. 2. The manufacturing of these cells begins with the procedures up to and including the cellular expansion or differentiation stage, referred to as upstream processing. While hMSCs only undergo expansion during the upstream process, hiPSCs must also undergo differentiation into the desired cell types, such as cardiomyocytes, insulin-producing pancreatic cells, or neurons, for safety reasons (Kojima et al. 2019). Here cultivation conditions, such as medium composition and shear stress, may be used to realize the differentiation process (Yourek et al. 2010;Gultian et al. 2022). Following the upstream process, downstream processing generally involves cell harvest, clarification, concentration, and washing (Cunha et al. 2015a;Jossen et al. 2018). Thereafter, the cell suspension undergoes formulation, final fill and finish, cryogenic storage, distribution, and patient administration.

Bioreactors and process mode
To maintain target cell quality and growth throughout the entire upstream process, careful attention must be paid to maintaining a well-defined environment with consistent Fig. 2 A simplified workflow for the production of hMSC-based and hiPSC-based therapeutics. The manufacturing process has been divided into (blue) typical upstream and (orange) downstream operations, followed by (green) the formulation to administration steps. Process steps marked with an asterisks only concern hiPSC-based products (created with Biore nder. com) 1 3 mechanical, physical, and chemical cues (García-Fernández et al. 2020). According to literature, this is best ensured through either mechanical or hydraulic agitation (Jossen et al. 2018;Tsai and Pacak 2021) in instrumented bioreactors (see also Table 2), which allow for the regulation and near homogenous distribution of process parameters such as temperature, pH, and dissolved oxygen (Manstein et al. 2021). Our review further revealed that most hMSC and hiPSC cultivations were performed in top-driven stirred bioreactor systems using microcarriers Dufey et al. 2016;Lawson et al. 2017;Pandey et al. 2020;Rotondi et al. 2021). A more detailed overview of microcarrier types suitable for stem cell production is provided in a review by Ornelas-González et al. (2021). Furthermore, while hMSC cultivation up to a pilot scale of 150 L has been described using stirred bioreactors (Jossen et al. 2018), hiPSC expansion has not yet progressed beyond bench-top scale (Pandey et al. 2020).
Independent of scale, one prevailing theme is the use of SU bioreactors for growing hMSCs , hiPSCs (Kwok et al. 2018), and their differentiated progeny (Jiang et al. 2019;Shafa et al. 2019). Such bioreactors, whose sterile plastic containers are used only once, are currently offered by multiple vendors (Eibl and Eibl 2019;Eibl et al. 2022) and are available up to a maximum working volume of 6 m 3 . SU bioreactors are known to reduce the risk of cross-contamination (Jossen et al. 2018) and are approved and even recommended by regulatory agencies for production processes that are subject to stringent safety requirements, i.e., processes whereafter contaminants cannot be easily removed, such as stem cell expansion and differentiation for therapeutic use (Nogueira et al. 2021).
As shown in Table 2, the application of such SU bioreactors has enabled peak cell yields of between 10.9 and 37.5 × 10 9 hMSCs or 10.2 × 10 9 hiPSCs per batch, while still maintaining key cell quality indicators Lawson et al. 2017;Pall Biotech 2020;Huang et al. 2020). Our review has also shown that higher volumetric yields of hiPSCs can be expected per batch compared to hMSCs when using the same system. Another method of improving yield or reducing process time for both cell types was the choice of process mode. Here, the cultivations that achieved the highest hMSC and hiPSC expansion factors were performed either as repeated batch (Kwok et al. 2018;Pandey et al. 2020;Dang et al. 2021) or in perfusion mode (Abecasis et al. 2017;Pandey et al. 2020;Huang et al. 2020;Manstein et al. 2021). For perfusion processes with hMSCs and hiPSCs in stirred bioreactors, filters with either defined pore sizes (dos Santos et al. 2014;Kropp et al. 2016;Abecasis et al. 2017;Huang et al. 2020;Huang et al. 2020;Manstein et al. 2021), settling tubes (Huang et al. 2020;Sion et al. 2021), or acoustic separators (Huang et al. 2020) were used to ensure cell retention during medium replacement. Perfusion rates used for hiPSCs typically ranged from 0.5 to 1.3 vvd (Kropp et al. 2016;Abecasis et al. 2017;Pandey et al. 2020;Huang et al. 2020;Manstein et al. 2021), while for hMSCs, rates of 0.25 to 0.48 vvd (dos Santos et al. 2014;Sion et al. 2021) were used. In this manner, expansion factors of up to 33.6 and 62.6 in 8-9 days could be achieved for hMSCs (Mennan et al. 2019) and hiPSCs (Pandey et al. 2020), respectively. While the N js or N s1 and N s1u criteria have been applied for stirred SU bioreactors operated with microcarriers to support the expansion of hMSCs and their process scale-up Lawson et al. 2017), other criteria, such as critical Kolmogorov length (λ c ), have also been successfully used for the cultivation of hiPSCs as spheroids in both rotating- (Shafa et al. 2019) and vertical wheel-impeller (Dang et al. 2021) bioreactors. Based on Zwieterings (1958) findings, N js or N s1 describes a minimum impeller speed at which solid particles, or in this case microcarriers, are just suspended in a bioreactor's working volume (Hewitt et al. 2011;Rafiq et al. 2013), while N s1u describes a lower limit for N js , where the microcarriers are in contact with the bottom of the bioreactor but not at rest (Kaiser et al. 2013;Jossen et al. 2016). These criteria may be determined visually or using particle image velocimetry (Tsai and Pacak 2021). Application of the N s1u approach when cultivating hMSCs in spinner flasks allowed Jossen et al. (2016) to maintain a mean shear stress ( ) of 4.96 × 10 −3 N m −2 ensuring optimal cell growth and quality. Petry and Salzig (2021) went on to define current estimates for acceptable and energy dissipation rates ( ) in stirred bioreactors to be between 0.01-0.06 N m −2 and 0.2-4.8 mW kg −1 , respectively, for hMSC cultivation. Additionally, they mentioned that the ratio of maximum energy dissipation (ϵ max ) to , or hydrodynamic heterogeneity (Φ) should not exceed 20. While little has been reported on acceptable ranges for hiPSCs, a of 0.3-1.5 mW kg −1 has been suggested (Dang et al. 2021).
These authors also mention the use of λ c as a means of estimating the maximum allowable power input in order to control spheroid size and minimize cell damage (Dang et al. 2021;Petry and Salzig 2021). Accordingly, the higher the impeller speed in a bioreactor, the smaller λ c becomes, with its length ideally being more than two thirds of the diameter of any cell or microcarrier aggregates in the system, to prevent cell stripping at their liquid/aggregate interface (Hewitt et al. 2011;Nienow et al. 2016a;Nienow et al. 2016b). The direct application of λ c does, however, presume fully turbulent conditions (Nienow 2021), which is generally not the case for the cultivation hMSCs or hiPSCs. Regardless, the approach has been readily adopted to limit the size of hMSC and hiPSC spheroids to a critical diameter of 200-300 μm (Sart et al. 2013;Allen et al. 2019;Huang et al. 2020;Petry and Salzig 2021), so that mass transfer is not restricted. These diameter limitations are less critical for microcarrier processes, especially when large spherical non-porous    In this context, the ability of both cell types to detach from populated and reattach to unpopulated microcarriers or form bridges between the two, referred to as bead-to-bead transfer (Badenes et al. 2017;Leber et al. 2017;Rafiq et al. 2018), has also been observed. The use of bead-to-bead transfer during inoculation has been shown to reduce the lag phase associated with the proteolytic treatment of both cell types during single cell transfer from static to dynamic cultivation systems (Badenes et al. 2017;Rafiq et al. 2018), shortening process time.

Culture medium
In addition to the choice of bioreactor, process mode, and surface/scaffold, the culture medium has a decisive influence on the result of cell expansion and differentiation. Table 3 gives an overview of commercially available serum-free, xeno-free, or chemically defined media used for the production of hMSCs and hiPSCs. At this point, it is worth mentioning that while chemically defined media for cell differentiation are available (Gultian et al. 2022), their application remains limited. On the contrary, the more popular approach is to supplement the media either with fetal bovine serum or a substitute, such as KnockOut™ Serum Replacement (Ackermann et al. 2018), human plasma (Sivalingam et al. 2021), or platelet lysate (Mizukami et al. 2018) alongside other recombinant and synthetic components (Olmer et al. 2018;Haack-Sørensen et al. 2018;Yabe et al. 2019;Laco et al. 2020;Jacobson et al. 2021). Correspondingly, these media compositions have facilitated the differentiation of hiPSCs into various cell types, such as hMSCs (Goetzke et al. 2019), cardiomyocytes (Laco et al. 2020), neurons (Silva et al. 2021), definitive endoderm (Jacobson et al. 2021), and hematopoietic cells (Sivalingam et al. 2021). Moreover, by adjusting composition and leveraging changes in intrinsic metabolic requirements during differentiation, selective pressure could be applied, improving target cell purity (Kehoe et al. 2010;Tohyama et al. 2017;Hsu et al. 2021) prior to downstream processing.

Downstream processing
In the context of cell therapeutic production processes, downstream processing has received less attention in peerreviewed literature than its upstream counterpart. It therefore comes as no surprise that there are still challenges and bottlenecks associated with downstream processing during the production of cell therapeutics. These include the short time window available between cell detachment and cryopreservation before quality becomes unacceptable (Viswanathan and Hematti 2017; Scibona and Morbidelli 2019) and the cells shear sensitivity, which restricts rigorous operations in favor of cell recovery (Cunha et al. 2015a;Scibona and Morbidelli 2019). In this context, the first step is the detachment of the hMSCs or hiPSCs from their growth surface as well as from each other. This is usually achieved by adding a proteolytic agent, such as TrypLE or Accutase, to cleave the integrin necessary for cell-tosurface/scaffold attachment, and chelating agents, such as Versene, to scavenge divalent ions required for cell-to-cell junctions (Derakhti et al. 2019). Alternatively, dissolvable scaffolds may be enzymatically digested instead (Rodrigues et al. 2019), or a temperature shift performed to affect cell release if a thermoresponsive coating was used (Narumi et al. 2020). For microcarrier or spheroid-based cultivations, cell recovery rates during detachment may further be improved by increasing fluid dynamic stress and collisions within the system (Nienow 2021), as has been demonstrated in various SU bioreactors with Φ between 10 and 25 at impeller speeds of 2 − 5 • N s1 , corresponding to a ϵ max of 1310-2830 mW kg −1 and a λ c of 24-30 μm (Nienow et al. 2016a;Nienow et al. 2016b). Following complete detachment from carriers, the cells must be separated from any non-dissolvable debris (Viswanathan and Hematti 2017). This so-called clarification is generally realized through dead-end filtration using, for example, SU products such as the Thermo Scientifics' Harvestainer™ BioProcess Container (Jossen et al. 2018), Entegris' Microcarrier and Cell Separation System (Pandey et al. 2020), or Merck's OptiCap ® capsules (Cunha et al. 2015a). Various studies have shown that pore diameters >75 μm resulted in cell recovery rates of >80 %, while ensuring a high cell quality and efficient microcarrier removal Cunha et al. 2015b;Serra et al. 2018).
After clarification cells are concentrated and then washed by diafiltration. Due to time constraints, both of these processes are often integrated (Cunha et al. 2015a;Pandey et al. 2020), e.g., by tangential flow filtration (TFF), using either hollow-fiber modules or flat sheet cassettes (Viswanathan and Hematti 2017; Cunha et al. 2017). Studies with TFFs highlighted that cell recovery rates and protein clearance improved when processing was continuous (Cunha et al. 2015a). Protein clearance could further be improved by adding a negative mode expanded bed adsorption step prior to concentration and washing (Cunha et al. 2016), while cell recovery rates were shown to be dependent on the choice of system, material, and operating parameters (Cunha et al. 2015b;Cunha et al. 2017). SU TFF technologies used for the downstream processing of hMSCs (Cunha et al. 2017) include Asahi Kasei's BioOp-timal™ MF-SL Microfilters (Cunha et al. 2015a), Cytiva's ReadyToProcess™ HF microfiltration cartridges, Merck's Pellicon ® XL Cassettes, and Sartorius' Sartocon ® Slice 200 (Cunha et al. 2017).
An alternative scalable low shear approach capable of integrating cell concentration and washing is continuous centrifugation (counterflow and disk stack centrifugation). While counterflow centrifugation has been used to purify hMSCs and hiPSCs at different scales (Li et al. 2019;Li et al. 2019;Pandey et al. 2020;Li et al. 2020a), there have been no reports using disk stack centrifugation. A design-ofexperiment study performed with a counterflow centrifuge (CTS Rotea) showed that, while neither flowrate, centrifugal force-to-flowrate ratio, serum concentration, nor trypsin concentration impacted live hMSC recovery during centrifugation, the first two did negatively influence cell metabolism (Li et al. 2020a). Further studies showed that these systems not only shorten the process time (Li et al. 2019), but also allow the selective isolation of cell populations based on physical properties, such as size, density, and even viability . Furthermore, the scalability of these systems was demonstrated in a recent study where 3 L cell suspensions with a density of 0.9-3.4 × 10 6 hiPSCs mL −1 were processed within 30 min, achieving a 105-fold concentration and cell recovery rates of up to 99 %, while maintaining target cell quality. The authors did, however, highlight that the chamber capacity would present the main scale-up constraint when moving to 50 L production (Pandey et al. 2020). Commercially available SU continuous centrifugation systems used for the downstream processing of hMSCs and hiPSCs include Sartorius' kSep ® (Pandey et al. 2020), Thermo Scientific's CTS Rotea Counterflow Centrifuge (Li et al. 2020a), and Terumo BCT's Elutra Cell Separation System . After this downstream operation, cells are formulated in preparation for cryogenic storage, filled into vials or bags, visually inspected, and frozen away for transport to the site of administration. A more detailed description of these operations is given elsewhere (Viswanathan and Hematti 2017).

Conclusions and outlook
Allogeneic hMSC-and hiPSC-based therapeutics are on the rise, with a clear trend towards their production in stirred SU bioreactors on microcarriers. Nevertheless, securing clinically relevant cell quantities and quality continues to pose a challenge. In upstream processing, safety issues surrounding hiPSCs need to be addressed and SU perfusion systems adapted to deal with cell shear sensitivity and bioreactor scale. Moreover, suitable chemically defined expansion and differentiation media compositions remain to be developed and commercialized. Here medium compatibility with perfusion mode could be advantageous. In the meantime, the groundbreaking work already done with hMSCs and microcarriers may be adapted to hiPSCs to increase process scale and yield using SU bioreactors, while bearing in mind the inherent similarities and differences between the two cell types. Ensuring a high cell yield during the cultivation process is only the first step, however. Cells must still undergo downstream operations to meet the quality and purity standards set by regulatory bodies. This demands the development of scalable automated systems capable of time sensitive cell processing. SU downstream processing technologies which combine process steps, such as TFF and continuous centrifugation have proven themselves in this regard, yet more research is necessary to fully understand how these systems impact cell quality and recovery. Finally, new technologies with great developmental potential, such as SU acoustic wave separators (Merck's ekko™) and SU disk stack centrifuges (GEAs kytero and Alfa Lavals Culture-One), have only recently become commercially available and remain to be tested for stem cell application. Funding Open access funding was provided by the Zurich University of Applied Sciences, while the manuscript was supported by the Department of Life Sciences and Facility Management of the Zurich University of Applied Sciences.

Declarations
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.

Conflict of interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.