Recent advancements in regenerative medicine have underscored the promise of cellular therapy in restoring compromised tissues and organs (Mao and Mooney, 2015). Mesenchymal stem cells (MSCs) have emerged as promising candidates due to their versatile properties, propelling them into clinical trials for various diseases (Bateman et al., 2018). It appears that doses of several hundred million cells per patient are required to achieve optimal therapeutic effects, which require extensive vitro expansion (Kabat et al., 2020).

Traditionally, MSCs have been expanded in tissue culture polystyrene (TCP) flasks through successive trypsinization procedures upon reaching confluency (Hassan et al., 2020). As a result, most pre-clinical research on MSCs has been conducted on TCP-expanded cells, leading to well-characterized cells (Hoch and Leach, 2014). However, this method is labor-intensive and carries contamination risks, particularly during the subculturing steps (Hoch and Leach, 2014). The Quantum Cell Expansion System by TerumoBCT (United States), known as the hollow fiber bioreactor (HFB) system, has emerged as a functionally closed and automated system designed for reproducibly expanding cells in both good manufacturing practices (GMPs) and research laboratory settings (Quantum Flex, 2025). It is apparent that the HFB system has several advantages over manual cell culture, including reduced laboratory footprint with one bioreactor having a cell growth area of 2.1 m² corresponding to 120 T175 TCP-flasks, significant less labor during cell culture, improved reproducibility and consistency in production as a result of better control over culture conditions such as nutrient concentrations and gas exchange, reduced risk of contamination due to the closed system, and improved scale-up potential (Quantum Flex, 2025; Barckhausen et al., 2016).

Previous studies have compared bone marrow-derived stem cells (BM-MSCs) (Jones et al., 2013) or adipose-derived stem cells (ASCs) (Haack-Sørensen et al., 2016) manufactured on the HFB system versus manual processing on TCP. These studies assessed the expression of surface markers, gene stability, expansion rate, cell yield, purity, and quality, paving the way to explore the automated HFB system as an alternative for efficient clinical-scale cell expansion (Jones et al., 2013; Haack-Sørensen et al., 2016). However, most of the studies failed to take into consideration the standardization of the population doublings (PDs) relative to the differing culture areas of the used expansion systems. This is a crucial consideration, as immunophenotypes can change along with each PD, and cells can exhibit varying growth potentials across PDs (Khong et al., 2017; Yang et al., 2018; Peng et al., 2020). Consequently, to ensure a meaningful comparison between two culture systems with a substantial difference in surface area, it is essential to perform the analysis at equivalent PDs.

Distinct immunophenotypical subpopulations have been shown to exhibit specific functional properties (Mihaila et al., 2013; Yang et al., 2013; Beckenkamp et al., 2018; Smith and Reid, 2018; Peng, 2024). To date, most of these findings are based on MSCs cultured using traditional TCP systems. In contrast, only a limited number of surface markers in ASCs cultured in HFB was reported (Haack-Sørensen et al., 2016), and the presence of immunophenotypical subpopulations in this context remains unexplored. Therefore, this comparative study of TCP and HFB culturing examined cell growth, single marker expression and highlighted the ASC subpopulations based on selected epitopes. The panels for multichromatic flow cytometry (MFC) were adapted from the previous study, including CD73, CD90, CD105, CD166, and CD201 in Panel A and CD34, CD36, D146, CD200, CD248, CD271, CD274, and Stro-1 in Panel B (Peng et al., 2022). To enable a direct comparison between the two culture systems, ASC expansion was carried out to achieve comparable PDs.

Expanding ASCs is crucial for many clinical applications, and several studies have investigated the effect of in vitro expansion on ASCs. Large-scale expansion using more automated systems is a practical necessity for many advanced therapy medicinal product (ATMP) productions. Nevertheless, previous studies have primarily focused on TCP systems, and the effect of HFB expansion on ASCs has been explored less. The TCP and HFB systems differ in intrinsic properties, including media replenishment, surface chemistry, topology, and the growth area as extrinsic factors. The bioreactor procedure requires only one seeding and is much simpler than the TCP protocol which requires multiple passaging. Furthermore, the culture flasks comprise a single small planar area, while the bioreactor consists of many hollow tubes, resulting in an extensive growth area.

It is well known that the immunophenotypes, homogeneity, mutagenic risk, and senescence of primary ASC cultures change with further doublings. Previous comparisons between the two systems were often established by the same passage number. However, the same passage number does not necessarily correspond to the same PD, as PD is related to the seeding density. Therefore, this study was designed to provide a comparable number of PDs during TCP and HFB expansion to allow for a more meaningful comparison. Such an approach was further validated by the result that TCP and HFB cells had a similar PD at the end of culturing. While the individual DT values on SVF are only estimates for cell growth, they still allow for a reliable comparison between the two systems. Previous studies comparing HFB- and TCP-expanded cells after two passages reported that HFB enhanced the expansion rate and yield (Lechanteur, 2014; Haack-Sørensen et al., 2016). Such accelerated cell growth by HFB system could be attributed to the continuous removal of waste materials and the reduction of abrupt fluctuations in PH and local paracrine environments facilitated by the automated feeding system. Despite difference, both systems delivered cells that met the ISCT-release criteria. However, when ensuring similar seeding density and PDs, rather than the number of passages, the present study found that growth rates of ASCs expanded in HFB and TCP were comparable. Collectively, the results indicate that initial seeding density and PD are critical regulators of growth rate.

The immunophenotypic profiles, based on surface markers, can be used as predictive markers for the biological effects of ATMPs (Lv et al., 2014). Increased immunophenotypical heterogeneity (determined by the number of clonal lineages) can lead to less predictable and more variable biological and therapeutic effects (Baer and Geiger, 2012; Prieto González, 2019). It is, therefore, crucial to investigate how these two expansion systems affect the immunophenotypes, including the individual surface markers and the clonal lineages defined by pre-selected surface markers. In this study, ASCs cultured in HFB demonstrated populations with more extensive immunophenotypical heterogeneity than TCP. When analyzing the single surface marker expression of ASCs from our 1:1 expansion comparison, both systems delivered cells that met the ISCT-release criteria for “clinical-grade” ASCs with over 95% expression of CD73, CD90, and CD105 and absent expression of hematopoietic markers like CD34 (Dominici et al., 2006), confirming results of the previous studies (Haack-Sørensen et al., 2016; Haack-Sørensen et al., 2018). Moreover, the previous study that employed both TCP and HFB cells confirmed their trilineage differentiation capacity (Ren et al., 2024).

When analyzing 13 cell surface markers individually, both systems yielded similar ASC populations, except for CD201 and CD274, that are linked to adipogenesis and immune response (Bárcia et al., 2015; Uezumi et al., 2016). When expanded in the TCP system, all cells expressed CD201 and CD274, whereas the CD201 expression was only around 50% for ASCs expanded in the HFB system. In previous studies using the TCP system, CD201 was highly expressed across eight passages of ASCs (Peng et al., 2020), suggesting that this marker may confer direct or indirect functional advantages to ASCs in TCP systems. The CD274 expression observed in this study differed from other studies, which reported expression levels below 50% for most donors (Peng et al., 2020). However, comparing these findings with other studies is somewhat challenging, as previous research has shown that surface marker expression correlates with population age (Peng et al., 2020; Peng et al., 2022). Consequently, it is crucial to consider the specific seeding and harvest designs (Dominici et al., 2006; Haack-Sørensen et al., 2016; Mastrolia et al., 2019).

Co-expression of several surface markers across the two systems showed much larger differences between the populations. For Panel A, the TCP system gave rise only to one significant clone that was positive for all five markers, whereas the HFB system gave rise to another clone that was negative for CD201. Additionally, based on Panel B, the HFB system delivered a much more heterogeneous population with a significant presence of eight individual clones, CLB1-8, compared to only four for the TCP system CLB6-9. Previous studies have demonstrated that surface marker expression in less-expanded stem cell cultures is highly heterogeneous. TCP expansion tends to select a few dominant clones, and efforts to expand specific minor clones typically result in the re-emergence of these characteristic TCP clones (Nielsen et al., 2016). Therefore, we propose that the HFB system is superior in maintaining the naive expression pattern of ASCs, preserving their in vivo expression to a greater extent. And, the clonal distribution was mainly driven by CD201, CD274, and CD146 expression. These altering distributions of clones are thought to support the specific biological effects induced by the donor or the culturing environment. Regarding the latter, the higher immunophenotypic diversity observed in ASCs expanded in automated HFB systems compared to TCP may result from multiple microenvironmental and mechanical factors inherent to the bioreactor design. Continuous perfusion and shear flow could provide mechanical stimuli that influence cell phenotype, while agitation or dynamic culture conditions may prevent local depletion of nutrients or accumulation of secreted factors that help to preserve subpopulation diversity. Higher initial seeding densities may enhance cell-cell interactions, and the chemical or physical properties of the hollow fiber scaffold or medium formulation may further modulate ASC behavior. Together, these factors are likely to contribute to the maintenance of a broader range of phenotypes in HFB-expanded ASCs, although the relative contributions of each remain to be experimentally determined. These hypotheses provide a framework for future studies to dissect the mechanisms underlying the observed heterogeneity.

Functional studies have previously revealed that ASCs descending from the CD274⁺CD146⁺ subpopulation were superior in terms of wound healing capacity, CD34/CD90 double-positive cells enhanced the tissue reconstruction, and CD248⁺ cells accelerated the healing process (De Francesco et al., 2009; Brett et al., 2017; Peng et al., 2022). Taking a specific example here, CD201+ cells have been reported to favor adipogenesis of mesenchymal progenitors (Uezumi et al., 2016). Therefore, it is plausible to infer that the differential proportion of HFB and TCP systems may indicate their substantial distinctions in terms of the adpogeneic potential. Collectively, this indicates distinct immunophenotypes within ASC cultures that are better suited for specific clinical scenarios than others. Of note, although phenotypic heterogeneity was identified among these two systems, functional assays were not performed to directly assess the biological properties of these subpopulations. As a result, no conclusions can be drawn regarding differences in cellular potency, differentiation capacity, immunomodulatory function, or therapeutic efficacy, etc., between those subsets harbored in the two systems. Secondly, the observed phenotypic differences were evaluated at the population level, and their stability over prolonged expansion or following in vivo transplantation was not examined. Finally, while prior studies have suggested potential functional relevance of these markers, the present findings are descriptive and should be interpreted as hypothesis-generating, which warrant further investigation in future.

In a pervious study (Ren et al., 2024), we reported comparable effects of ASC harvested from these two expansion systems on tri-lineage differentiation, colony-forming capacity, proliferation rate, as well as fibroblasts migration and proliferation. Building on these findings, future studies incorporating more potency testing will be required to establish whether the observed phenotypic differences translate into meaningful biological or clinical effect.

This study provides valuable insights into the impact of bioprocessing on ASCs during expansion in HFBs compared to TCP flasks. Our findings revealed that HFB-expanded ASCs demonstrated superior immunophenotypical heterogeneity. Regarding the impact of the culture system on the selection of specific lineages, the CD201⁺ and CD274⁺ clones were favored in the TCP conditions. In contrast, the HFB environment supported both CD274 positive and negative populations. Nevertheless, both systems met ISCT criteria for the single surface marker profiles of ASCs. These results contribute to the growing body of knowledge on large-scale cell expansion methods, specifically in regenerative medicine and cellular therapy. The study highlights the need for further exploration of the HFB system for efficient clinical-grade cell expansion, emphasizing the need for potency assays and a deeper understanding of the link between surface markers and the functionalities of ASCs in the context of ATMP productions.