Passage- and serum-dependent changes of adipose-derived stromal cells in vitro: a discrepancy of stemness factors regarding mesenchymal surface markers and expression of stemness-related genes

Isolation and cell culture

After receiving approval from the Ethics Committee of Hannover Medical School (reference no. 1569-2012) and the donors’ signed written consent, human adipose tissue was harvested from elective dermolipectomy and the stromal vascular fraction containing adipose-derived stem cells was isolated. The donors (n=5; male=2, female=3) had no significant medical history other than obesity and were aged between 28 and 55 years (mean=35.6 years). After the surgical procedure of dermolipectomy, the fat tissue was removed from the graft, minced, and transferred into sterile multipurpose screw-capped cups (size 50 mL). Each cup was filled with a 20 mL fraction of minced fat and 0.2 % (w/v) of collagenase (C1-22, Biochrom) in Hank’s buffered salt solution to digest the extracellular matrix. The sample was then vortexed and incubated for 1 h at 37 °C with gentle motion agitation at 120 rpm. The suspension was then dispersed, washed with Hank’s buffered salt solution, and centrifuged for 5 min at 30 ×g. The oily upper fraction and the pellet at the bottom were transferred and filled with Hank’s buffered salt solution and 2 g/L bovine serum albumin and centrifuged for 10 min at 300 ×g. The liquid fraction was discarded, and the pellet was resuspended with Hank’s buffered salt solution, followed by centrifugation for 10 min at 300 ×g and siphoning of the liquid fraction. Each pellet, containing the stromal vascular fraction, was then resuspended in a 150 cm² flask (TPP 90151, Sigma Aldrich, Munich, Germany) with 20 mL DMEM/F-12 (Biochrom, FG 4815; Berlin, Germany) contained either 20 % fetal calf serum (FCS; Biochrom, S0615; Berlin, Germany) or 20 % human serum. Apart from the different serum, both media contained similar 50 U/mL penicillin/streptomycin (Biochrom, A2212; Berlin, Germany), 0.173 mM ascorbate-2-phosphate (Sigma Aldrich, A8960; Munich, Germany). Flasks were incubated at 37 °C, 5 % CO₂ and water-saturated atmosphere. Cultures were detached with 0.25 % Trypsin-EDTA at approximately 70 % confluence, split 1:3 and passaged to passage 5.

Human serum preparation protocol

Human serum was obtained from residual blood samples of healthy donors from the Hannover Medical School Blood Bank. In the donation procedure, a small amount of donor blood often remains unused and is usually discarded. These residuals were collected anonymously in tubes and allowed to clot. After 30–60 min at room temperature, the tubes were centrifuged at 2000 ×g for 10 min to separate the serum from the clot. The serum was immediately pipetted into new tubes and heat-inactivated in a water bath at 56 °C for 30 min. The pooled mixed-gender serum was stored at −80 °C until use.

Metabolic activity

To investigate the effect of media on viability, we performed CellTiter-Blue^® Cell Viability Assay (Promega, Fitchburg, USA) according to the manufacturer’s protocol. Fluorescence was measured using the Tecan GENios multi-well reader (Männedorf, Switzerland). About 500 cells per well were seeded on 96-multi-well plates. Both groups (HS, FCS) received daily changes of media and were incubated as described above. Cell viability was determined at 24, 48, 72, 96, 120, and 144 h. The assay is based on the metabolic function of viable cells to reduce the used dye into a fluorescent substrate, which can be detected by a spectrophotometer.

Cell confluency

Cell cultures of 150 cm² flask of both groups were monitored daily by light microscopy (Olympus CK-40, Shinjuku, Japan). Increasing confluence was photodocumented with 40× enlargement. At approximately 70 % confluence, cells were detached and passaged at a 1:3 ratio. We monitored the maintenance of the typical fibroblast-like morphology of the ASCs as well as the monolayer architecture and the cell density. The time to reach app. 70 % confluence (tc70) by subjective evaluation was documented and compared between both groups at each passage until P5. Media changes were performed three times per week when the color of the media indicated an acidic pH, using 20 mL of human-serum supplemented vs. fetal calf-supplemented media per 150 cm² flask.

Flow cytometry analysis

According to the minimal criteria definition for multipotent mesenchymal stem cells [4], we investigated the presence of specific surface markers CD 73 and CD 90 and the absence of CD 11b, CD 31, CD 34, CD 45, and CD 105 were examined by flow cytometry analysis from passage 0 to 5. The investigation was performed with a Beckman Coulter FC500 CellSorter (Brea, USA). The antibodies for CD 34 (Beckmann Coulter; A21691) and CD 73 (Biolegend; 344003) were PE-based/conjugated. The antibodies used for CD 31 (Beckmann Coulter; IM1431U), CD 45 (Beckmann Coulter; IM2652U), and CD 105 (Beckmann Coulter; PN A07414) were FITC-based/conjugated. The antibody used for CD 11b (eBioscience; 25-0118) was PC7-based/conjugated and that used for CD 90 (Beckmann Coulter; PN IM3703) was PC5-based/conjugated. Cells were detached at 70 % confluency for each passage. We incubated the cells with 10 µL of each antibody and 10 µL of the isotype for 45 min at room temperature in the dark. Each tube was washed with 1 mL of PBS and centrifuged at 300 ×g for 5 min. The liquid fraction was discarded, and the remaining cell pellet was suspended in 500 µL Isoton solution for immediate cytometry analysis.

RT-PCR

Cells from P0-P5 were detached from the flasks when they reached approximately 70 % confluence, evaluated by light microscopy using 0.25 % Trypsin-EDTA (TE). After washing with PBS (wo), the cells were detached from the flask, centrifuged at 300 ×g for 5 min, and the enzymatic reaction was stopped by the addition of RA1-Lysis buffer. RNA isolation was performed with NucleoSpin RNAII-Kit (Macherey-Nagel, Düren, Germany) following the standard manual instructions. The quality and quantity of the extracted RNA was verified by using a NanoDrop 2000/2000c spectrophotometer (ThermoFisher, Waltham, Massachusetts, USA). RNA purity and concentration, as well as possible protein contamination, were tested before any further measurements. The integrity of the isolated RNA was verified by gel electrophoresis. cDNA synthesis was performed using the iScript^tm Synthesis Kit (BioRad, Hercules USA) according to standard manual instructions. RT-PCR was performed by using an iCycler real-time polymerase chain reaction cycler (BioRad Laboratories, Hercules USA). The genes analyzed were MCAM, cKIT, MYC, and OCT4 as shown in Table 1. All target gene expression levels were normalized with housekeeping genes (ywhaz, RPL 32, RPL 37, TBP, and β2-microglobulin) and quantified relative to the 2-ΔΔCT method [18]. Each RT-PCR plate layout included positive controls (pooled cDNA from N1-5 in P0, HPLC-water, EvaGreen, housekeeping gene-primer), negative controls (HPLC-water, EvaGreen, Housekeeping gene-Primer), and a blank control (HPLC-water and EvaGreen). The quality of selected primer pairs was validated by melting curve analysis and verification of product size on 2 % agarose gel electrophoresis (Biozym, 840004; Oldendorf, Germany).

Statistical analysis

Statistical analysis was performed using SPSS 16.0 software. All measurement data were expressed as mean±standard error of the mean (SEM). Comparison between groups was performed using Student’s t-test. A p-value of less than 0.01 was considered to be statistically significant.

Cultivation

After the isolation process, the cell pellet, containing the stromal vascular fraction, was seeded into 150 cm² flask, where the cells exhibited ASC typical growth behavior within 24 h. In both groups, cells developed a homogeneous monolayer with the characteristic fibroblast-like morphology and adherence to plastic. Furthermore, cell morphology and size remained unchanged throughout all passages up to P5 within all donor cells and both supplementation groups as shown in Figure 1.

Figure 1:

Typical fibroblast-like morphology and plastic adherence of hASCs from the same donor. (A) & (B) Day 5 after initial seeding (in P0) in different culture conditions: (A): 20 % FCS-supplemented media group, (B): 20 % HS-supplemented media group: a higher cell density can be seen in (B). Images (C) & (D) show the same hASCs in P2 15 days after initial seeding. (C): FCS group with app. 20 % confluence in P2, (D): HS group in P2 with app. >90 % confluence. An expansion of the monolayer structure was regarded as an increase in proliferation.

Cell viability

As shown in Figure 2, metabolic activity was determined in both supplementation groups with hASCs from the same donor at passage 1 (P1) using the Cell Titer Blue assay. Significant differences (p-value=˂0.01 in student’s t-test) between the groups were already detectable after 24 h hASCs with HS supplementation displayed an almost 2-fold higher activity after 24 h and maintained their superiority throughout the measurement period. In the subsequent comparison of the FCS vs. HS-group, cell viability was significantly higher in the HS group at 24, 48, 72, 96, 120, and 144 h. The metabolic activity of the HS group was 2-fold higher than that of the FCS group at both 96 and 120 h. Viability, and therefore proliferation rate, remained significantly increased in human serum supplemented hASCs from day 1 (24 h) to day 6 (144 h).

Figure 2:

Cell viability evaluated by Cell Titer Blue assay. hASCs (P1) from the same donor were used in both groups (HS-supplemented and FCS-supplemented media). After 24, 48, 72, 96, 120, and 144 h metabolic activity was compared between both groups using Student’s t-test. hASCs in HS-supplemented media showed significantly higher cell viability at all times compared to hASCs cultured in FCS-supplemented media (*indicates statistical significance, p-value=<0.01). Relative cell viability was normalized to the highest viability achieved at 144 h in the HS supplemented media is shown.

Cell confluency

While HS-supplemented hASCs reached monolayer confluence of 70 % at a mean of 4.08 days (SEM±0.09), hASCs in FCS-supplemented media had a mean of 7.49 days (SEM±0.29), as shown in Figure 3. The time to reach approximately 70 % confluence appeared to be constant and comparable for all hACS from 5 different donors (N1-5) through all passages (P0–P5) of both groups (HS vs. FCS supplemented media). The characteristic fibroblast-like morphology appeared unchanged. Consistent confluence assessment and cell documentation was performed using a light microscope with 50 µm enlargement (Olympus CK40, Tokyo, Japan) as shown in Figure 1. Adherent fibroblast-like shaped cells are shown in pictures A and B on day 5 after seeding of hASCs from the same donor (P0). Image A shows hASCs cultured in FCS-supplemented media; B shows the HS-supplemented group. A higher cell density can already be seen in B. Images C and D show the same hASCs in passage 2. Image C visualizes the approximately 20 % confluency in the FCS-supplemented group, while image D shows the HS-supplemented group with >90 % confluence. The cell flasks we used to image cell confluence were used solely for confluence documentation and discarded from further examination as 70 % confluence as defined in the study protocol.

Figure 3:

In addition, cell proliferation was determined as days to reach 70 % confluence. hASCs from 5 donors (N1-5) were examined through all passages (P0-P5) in both groups (HS vs. FCS supplemented media). Initial seeded cell number was 2 × 103 cells/cm². The data show approx. 70 % confluence at a mean of 4.08 days (SEM±0.09) for hACS with HS-supplemented media, whereas hASCs in FCS-supplemented media had a mean of 7.49 days (SEM±0.29). Cell proliferation was compared in HS-group vs. FCS-group separately for each donor (N1-5) using Student’s t-test, revealing significantly higher cell proliferation in all HS-groups (*indicates statistical significance, p-value=<0.01).

Expression of surface markers

Flow cytometry was used to determine the expression of the surface markers CD 11b, CD 31, CD 34, CD 45, CD 73, CD 90, and CD 105 in both groups (HS vs. FCS) at P0 and P5. There was no significant difference in surface marker expression related to either passage or supplementation. The absence of certain stemness-related surface markers such as CD 11, CD 31, CD 34, CD 45, and CD 105 was demonstrated by a percentage of ˂5 % and the presence of stemness-related surface markers such as CD 73 and CD 90 was >95 % (Table 2). The markers remained stable, and the standard error of the mean (SEM) showed homogeneity in the present data.

Gene expression in RT-PCR

The expression of stemness-related genes, which are critical for the oncogenic transformation of human cells, was compared at P0, P2, and P5 by RT-PCR. Human cells were cultured in either FCS-supplemented or HS-supplemented media. Statistical comparison was made between conditions at the same time points. In both treatment groups, cKIT showed the highest downregulation (p<0.01) from P0 to P5. Similarly, MCAM expression decreased continuously during long-term culture from P0 to P5 (p<0.01). Interestingly, the initial (P0) expression levels of cKIT and MCAM were higher in FCS-treated cells compared to the initial levels in HS-treated cells. The expression levels of stemness-related genes are shown in Figure 4.

Figure 4:

The relative change in the expression level of stemness-related genes (cKIT, MCAM, MYC, OCT4) in hASCs at P0, P2, P5 cultured in FCS-supplemented or HS-supplemented media. Pooled cells from 5 donors were evaluated. Means are expressed in relative units ±SEM. Statistical comparison between groups was performed using Student’s t-test; *p<0.01.

MYC and OCT4 showed no significant decrease from P0 to P5 in all treatment groups, although there was a slight tendency for MYC expression to decrease. When compared between passages P0 to P5, hASCs maintained in FCS-supplemented media had a stable level of OCT4 and MYC expression, whereas expression in HS group increased unpredictably at P2. However, treatment conditions (FCS vs. HS) had no significant effect on the changes in OCT4 and MYC gene expression.

Summarized, the data showed that certain regulatory genes (cKIT, MCAM) showed significant decrease in long-term culture from P0 to P5 in both treatment groups (FCS and HS). Yet notably, in all groups treated with FCS, gene expression levels were higher at P0, compared to P0 in HS media group.

SVF

The isolated stromal vascular fraction consists of a heterogeneous cell population of ASCs, fibroblasts, preadipocytes, vascular endothelial cells, cytokines, growth factors, and yet only inadequately represents the diversity of the in vivo stem cell niche. Unlike the physiological niche, the SVF lacks key components such as cell-to-cell-contact, which can only be found within the original anatomical structures established by a heterogeneous population of progenitor and lineage-committed cells [21]. When cells are selected by the cell culture regime, in addition to ASCs, a selection of other adherent cohabitants of the stem cell niche, such as fibroblasts, are initially seeded. Thus, under in-vitro-conditions, cell–cell interaction, cell–extracellular matrix interaction, and secretion of signaling molecules secretion are lost as unique and important characteristics of the stem cell micromilieu [22].

Stem cell niche vs. in vitro cell culture

Over the past decades, studies on bone marrow-derived stem cells have described the similar principles of a stem cell niche in the bone marrow and its relevance to the maintenance of stemness capacity within the cells. Schofield et al. described the interdependent homeostasis of the bone marrow stem cell niche as early as 1978 [23]. Their findings suggest that a specific milieu of cell–cell interactions, cytokine levels, and cell–extracellular matrix contacts is critical for the maintenance of stemness capacity of the mesenchymal stem cells. The preservation of the stem cell niche is mandatory for the regulation of stemness factors in the ACS. Moreover, it is key to regulatory processes in mesenchymal stem cells due to a plethora of cellular mechanisms [24, 25]. Voog et al. identified niches in various mammalian epithelial tissues [26]. Fuchs et al. described stem cell niches, e.g., in ovarian, testine, dermal, intestine, subventricular, and bone-marrow tissue in their review [27]. Stem cells in these niches have individual microenvironments that allow them to thrive and function as it is individually required [28, 29].

ASCs secrete a unique cytokine profile that allows them to induce cell proliferation and differentiation, promote angiogenesis, and reduce inflammation. In this regard, a state of constant calibration is required to maintain this delicate homeostasis. In vivo, this is regulated by the microenvironment of the ASC, the stem-cell niche. Alterations such as hypoxic conditions in cell culture would lead to distinct changes in the cytokine secretion profile. Rehman et al. investigated the secretion of angiogenic and antiapoptotic factors in ASCs and found that when ASCs were cultured under hypoxic conditions, e.g., vascular endothelial growth factor (VEGF) secretion increased 5-fold [5, 30].

Vice versa, hASCs maintained in coculture with other cells also affect these other cells in a variety of ways. For example, depending on the function of their neighbors, they may support hematopoiesis by releasing M-CSF in vitro in coculture with umbilical cord-derived CD34 cells. Furthermore, they seem to expand the amount of myeloid and lymphoid progenitor cells [6]. However, when maintained in coculture with T47D breast cancer cells, Kuhbier et al. (2014) revealed for the first time that ASCs fuse and initiate vesicular exchange with cancer cells [31]. Additionally, an increase in the expression of transcriptional genes for typical malignant markers was shown to be much higher in cocultures of ASCs and breast cancer cells than in single cultures [31].

Stemness vs. malignancy

Since each stem cell divides into a progenitor cell and renews itself with each passage, it is understandable that the overall stemness capacity decreases overall during the time of cell cultivation. Progenitor cells lack multipotency and can only differentiate into one type of tissue, such as adipose tissue or osteoblasts. In addition, the renewed ASCs become more senescent with each cell division. Previous studies have shown that this leads to a dysregulation of proliferative activities and an overall decreasing stemness capacity [32]. These aged and dysfunctional stem cells accumulate toxic metabolites that lead to DNA damage and consequently the production of damaged proteins [33]. It is also shown in previous studies that the microenvironment surrounding the cells can either promote or retard cancer development [11, 25]. Plaks et al. described that stem cells that escape regulation within their niche can develop into a cancer stem cell [25]. Given their capacity to self-renewal and clonal expansion, the switch of a stem cell to its malignant counterpart leads to uncontrollable cancer development. Assuming similar mechanisms in the maintenance of stemness or deregulation in vitro, cell identification is as crucial as it is delicate. Prior to clinical use of ASCs expanded through multiple passages in cell culture for, the cells must be identified as nonmalignant to ensure a safe procedure.

Surface markers and stemness-related genes

Regarding surface markers measured by flow cytometry, hASCs showed the required absence and presence of characteristic markers according to the original minimal criteria definition by Dominici et al. (2006) and the joint statement of the International Federation for Adipose Therapeutics and Science (IFATS) with the International Society for Cellular Therapy (ISCT) (2013) [4, 17]. However, RT-PCR showed a significant decrease in the stemness-related genes cKIT and MCAM from passage P0 to P5 in all treatment groups. More interestingly, and not yet described in previous studies, the initial levels of both markers were significantly higher at P0 in cells treated with FCS-supplemented media compared with cells cultured in HS. A decrease of these two stemness markers across all passages indicates a reduction in the signaling pathways involved in multipotency and consequently stemness capacity overall. Additionally, unpredictable variations in expression levels, such as those seen for OCT4 and MYC, indicate a currently unknown impact on the expression of regulatory genes within the cells. Moreover, transcriptional decrease of MCAM indicates loss of angiogenic function, as MCAM plays a fundamental role in mediating angiogenesis [34].

Media influence on proliferation

Lindroos et al. reported that a higher percentage of human serum supplementation in the media results in a higher proliferation rate, which proportionally inclined due to the percentage ranging from 5 to 20 % [11]. However, these results were observed up to day 4 of cell culture. When the proliferation rate of ASCs in 20 % human serum supplemented media was compared with 10 % fetal calf serum media under the same culture conditions, it was equally high [11]. Our data on cell proliferation, as measured by cell viability showed comparable results in terms of the correlation between high serum percentages (20 %) and the attainment of app. 70 % confluence (FCS mean of 7.49 days (SEM±0.29), HS mean of 4.08 days (SEM±0.09)). It is noteworthy that ASCs cultured with human serum tend to grow faster than those cultured with fetal serum. A shorter in vitro cultivation time is beneficial not only for obvious reasons such as efficiency, but it also leads to less risks of contamination and loss of cells due to infection, simply because of the shorter in vitro exposure.

The stemness marker transcriptome shows variabilities from P0 to P5 with no predictable pattern in neither the HS-supplemented nor the FCS-supplemented group. Nevertheless, we found comparable pattern tendencies within some markers, such as cKIT and MCAM as well as OCT4 and MYC. cKIT and MCAM showed a constant decrease in expression in the FCS and HS cell population from P0 to P5.

The impact of human serum

Human serum from blood donors contains a variety of allogenic growth factors, whereas the commonly used FCS is xenogenic to human ASCs due to its origin. However, the use of serum in cell culture is essential for the cell nutrition, and it promotes cell growth and consequently cell survival. Similar to other human blood derivatives, serum contains numerous growth factors to maintain ASC stemness, senescence, and differentiation and to promote overall growth [35]. Heat inactivation of human serum destroys complement activity but does not necessarily affect the growth properties and potential pathogens. Additionally, other protein components of the serum remain active and potentially harmful as pathogens of RNA or DNA origin [36]. This represents a potential risk of infection that can be considered a relevant uncertainty when using allogenic HS for human ACS cell culture.

Human blood donors and their blood samples must undergo an extensive screening process before being accepted for donation, so the risk of infection must be comparable to the general incidence of infection from blood transfusion in the country concerned. The same risks apply to FCS. However, to date there are no reliable data on the level or quality of infection in cell cultures due to the use of either FCS or HS.

Notably, a variety of advantages have been reported for the use of human serum in cell culture: Tallheden et al. showed that articular chondrocytes cultured in autologous human serum showed up to seven times higher proliferation than FCS-supplemented chondrocyte cultures [37]. Furthermore, Kobayashi et al. demonstrated that bone marrow derived MSCs cultured with autologous human serum exhibited greater cell motility than MSCs expanded in the presence of FCS [38].