Chemically Defined Xeno- and Serum-Free Cell Culture Medium to Grow Human Adipose Stem Cells

Adipose tissue is an abundant source of stem cells. However, liposuction cannot yield cell quantities sufficient for direct applications in regenerative medicine. Therefore, the development of GMP-compliant ex vivo expansion protocols is required to ensure the production of a “cell drug” that is safe, reproducible, and cost-effective. Thus, we developed our own basal defined xeno- and serum-free cell culture medium (UrSuppe), specifically formulated to grow human adipose stem cells (hASCs). With this medium, we can directly culture the stromal vascular fraction (SVF) cells in defined cell culture conditions to obtain hASCs. Cells proliferate while remaining undifferentiated, as shown by Flow Cytometry (FACS), Quantitative Reverse Transcription PCR (RT-qPCR) assays, and their secretion products. Using the UrSuppe cell culture medium, maximum cell densities between 0.51 and 0.80 × 105 cells/cm2 (=2.55–4.00 × 105 cells/mL) were obtained. As the expansion of hASCs represents only the first step in a cell therapeutic protocol or further basic research studies, we formulated two chemically defined media to differentiate the expanded hASCs in white or beige/brown adipocytes. These new media could help translate research projects into the clinical application of hASCs and study ex vivo the biology in healthy and dysfunctional states of adipocytes and their precursors. Following the cell culture system developers’ practice and obvious reasons related to the formulas’ patentability, the defined media’s composition will not be disclosed in this study.


S1. Isolation of the Stromal Vascular Fraction (SVF) from human adipose tissue
When the subcutaneous adipose tissue was supplied as a whole thick slice, the first operation was to remove the skin using sterile surgical devices, such as scissors, scalpels, and tweezers. The fat tissue was then washed twicewith DPBS supplemented with Ca2+ and Mg2+ (# 3-05F00-I, BioConcept, Switzerland) in a sterile container (Pot Conique PP 1000 mL 105 x 130 mm Couvercle Sterile, # 080866, Milian AG, Switzerland) and cut into small pieces. Subsequently, the tissue was completely disrupted by an immersion blender, and then the homogenized fat tissue was poured into the 100 mL "extraction syringe" (Omnifix 100mL Syringe, BBraun, Melsungen, Germany). For the homogenization of a small amount of fat tissue, the "ULTRA-TURRAX® Tube Drive System" with its single-use and gamma-sterilized DT-50 tubes (both manufactured by IKA-Werke GmbH & Co. KG, Germany; www.ika.com/en) was found to be very practical. On the other hand, if the adipose tissue was collected during liposuction and was already in a "sampling syringe" used by the surgeon, then it could be quickly and safely transferred to the 100 mL "extraction syringe" using a "Fluid Dispensing Connector" (# 415080, FDC1000, BBraun AG, Melsungen, Germany) as depicted in Figure S1. Indeed, our protocol was based on the utilization of a 100 mL "extraction syringe" (Omnifix 100 mL with Luer Adaptor, and Universal Closing Stopper, #4495101, both from BBraun AG, Melsungen, Germany) as a separatory funnel to exploit the fact that the adipose tissue and the hydrophilic fluid containing the cells spontaneously separates in two phases without the need for centrifugation. The piston of the syringe was used to take in or expel the solutions used to wash the sample, dissociate the suctioned fat (Collagenase), or extract the cells from the dissociated adipose tissue. The syringe was held in a vertical position using a laboratory support stand with support rings ( Figure S2). Therefore, the syringe allowed all the necessary manipulations to be performed to extract hASCs from 50 mL homogenized fat or lipoaspirate. The procedure took about 70 minutes and was done in a laminar flow cabinet for sterile work.  To dissociate the homogenized or suctioned fat, 10 mL DPBS (with Ca 2+ , Mg 2+ ), containing the appropriate amount of Collagenase (# LS004147, Type B, Animal Origin Free, Worthington, Biochemical Corp., Lakewood, NJ) to reach the final concentration of 0.28 Wünsch U/mL enzyme, was added by pulling the piston of the syringe. In liquid aspiration procedures, it is useful to use a needle (e.g., # 4665473, 100 Sterican 14G x 31/8", BBraun, Melsungen, Germany) or, to make the process safe and avoid punctures, one can use a plastic cannula, for example, "Einmalknopfkanüle Steril DM 2.0/1.0 L:100 mm" (# 8572799, Polymed Medical Center, Glattbrugg, Switzerland). After 45 min incubation at 37 0 C under constant but gentle agitation, 30 to 40 mL of a DPBS (without Ca2+, Mg2+, # L0615-500, Biowest, France) solution with 1% injectable human albumin (# 22918180119611, CSL Behring AG, Bern, Switzerland) were aspirated. The syringe was thoroughly shaken to extract the cells. The syringe was then returned to the support stand ring to allow for the separation of the two phases. The lower layer, which contained the SVF, was carefully pushed out into a conical 50 mL centrifuge tube (TPP, Trasadingen, Switzerland, # 91050). The extraction step can be repeated with 30 mL DPBS/1% injectable human albumin solution. Finally, after the sequential filtration, first through a 100 µ m and then a 40 µ m sieve (Cell Strainer, BD Falcon, Basel, Switzerland, # 352360 & # 352340), the SVF was centrifuged (600 g, 5 min). The pellet was resuspended in DPBS (without Ca2+, Mg2+)/1% injectable human albumin or in the tissue culture medium. The steps which followed the digestions with the Collagenase are schematically depicted in Figure S3. Finally, we believe the whole protocol could be easily translated into a GMP-compliant version to produce hASCs for cell therapeutic applications. The utilization of this procedure for commercial purposes is covered by the patent EP2726602B1 [1] Figure S3. Diagram depicting the steps following the digestion with the Collagenase.

S2. Characterization of the cells of the SVF from adipose tissue
The SVF is a heterogeneous mixture of cells isolated by enzymatic dissociation of the adipose tissue. Adipocytes represent roughly two-thirds of the total cell extracted, and the rest are bloodderived cells, vascular cells, endothelial cells, smooth muscle cells, pericytes, fibroblasts, and ASCs. We developed a very informative multiparameter flow cytometry assay to characterize the cells of the SVF. This test allows us to subdivide the CD45-negative cells of the SVF into four fractions/subpopulations (see Figure 1).
This analysis provides the information necessary for the basic operations with these cells, like for example, the determination of the absolute cell number for every subtype (cells/mL) • Crucial for seeding the cells at the optimal density for culture in serum-free conditions • Crucial if one wants to cryopreserve the SVF for later use • Crucial if the cells of the SVF are immediately used after the extraction (e.g., aesthetic surgery treatments).

S3. Initial plating of cells of the SVF for expansion
In serum-free conditions, the initial plating density is often the critical factor that determines whether a primary hASC-culture succeeds in taking root or not. Therefore, we elaborated the following rules to culture SVF cells in UrSuppe medium successfully: With manually Fibronectincoated or commercially available vessels like Corning PureCoatTM ECM Mimetic Fibronectin Peptide, we usually plate: • At least 3 x 10 5 nucleated cells/cm 2 in UrSuppe medium • At least 3 x 10 4 ASCs/cm 2 in UrSuppe medium Knowing the exact number of hASCs (defined as CD45-, CD146-, CD36-, CD34+, and CD73+ cells) contained in the SVF enormously facilitates the growth of these cells in SF conditions and the gain of a first confluent hASCs-culture at passage 0 (P0). If it is impossible to perform an accurate flow cytometric analysis to determine the number of hASCs present in the SVF, then at least 300,000 nucleated cells per cm 2 should be seeded. As shown in Figure S4, we found, on average, the hASCs represent 10% of the cells that compose the SVF. Therefore, with at least 300,000 nucleated cells of the SVF plated, there is the chance to hit the optimal minimal seeding density, which allows for the establishment of a hASCs-culture at P0 in SF conditions. Anyway, when the exact number of hASCs present in the SVF is not known, it is wise to seed 2-3 different cell densities.
After 3 to 5 days of culture, we gently rock the vessels to suspend and aspirate most nonadherent floating cells. After adding fresh UrSuppe medium (we do not recommend washing the cells at this stage), we continue culturing until confluency is reached. From now on, only 50% of the old medium should be replaced with fresh medium. Refer to Figure S5 for images of cultured ASCs.
After extraction of the SVF from human adipose tissue, we do not lyse the erythrocytes present in the cell suspension. This strategy is because we try to avoid further stressing the cells that were subjected to a harsh extraction protocol of at least two hours in the presence of proteases. However, in some preparations, there are indeed too many erythrocytes. In these cases, after two or three days of SVF culture, we gently collect the floating cells and lyse the erythrocytes using standard reagents and protocols. The RBC-free cells can be put back into the starting vessels or seeded in new ones. Therefore, this simple procedure has often augmented the yield of P0 hASCs. Finally, we realized that eliminating erythrocytes is certainly superfluous when the SVF, immediately after extraction, is meant to be cryopreserved. Indeed, when the samples are thawed, all erythrocytes are automatically lysed.

S4. Basic guidelines for working with serum-free (SF) media
The challenge for every serum-free medium is to allow the growth of a primary culture (passage 0, P0) of ASCs. Indeed, after reaching the first confluency, the culture is relatively easy to passage, also in serum-free conditions. From this point on, the strategies for amplifying the cells are similar to those used with FBS-based media. After having placed dozens of different fresh or thawed SVF samples in culture with our serum-free UrSuppe medium, we have identified the critical points which determine whether the primary culture can take root or not. These are: • Human lipoaspirate tissue samples must be processed up to 24 h after surgery • Before processing, store the lipoaspirates at room temperature (not at 4 °C) • Lipoaspirates must be sterile and avoid contamination with common and widespread commensal skin microorganisms • Use commercially available coated vessels: Excellent with Corning PureCoat ECM Mimetic Fibronectin Peptide. Not always good with manually coated vessels • Seeding density: See below, "Initial plating of cells of the SVF for expansion." • Medium changes: At the beginning, not too often and replace only 50% of the old medium with fresh medium (for a possible explanation of this strategy, please see below) • Excellent quality reagents (e.g., LPS-free) and components of the medium should be used, as well as very good quality instrumentation (mostly the incubator) • Skill and training of the operator Taken together, when the points listed above are respected, the success rate with our serum-and xeno-free cell UrSuppe culture medium is very high.
In SF conditions, cells secrete many paracrine factors that complement the synthetic medium and contribute to the cells' survival and growth. Therefore, it is crucial, especially with stem cells' primary cultures, not to completely replace the old medium. Therefore, it is wise to keep 25 to 50% of the old, "conditioned" medium. If any dead cells and debris are present, transfer the old medium to a 15 or 50 mL tube and centrifuge it for a few minutes (600 g) to remove the dead cells. The supernatant is then filtered through a 0.45 µ m filter to remove the small debris. When the cells density is low with primary cultures, do not change the medium too often, the first time after 4 to 5 days of culture. It is wise to replace 80% of the old medium at high cell density every third day.
Serum-free media usually do not contain protease inhibitors. Therefore, it is necessary to wash the cells at least once to remove the protease before reseeding them in SF medium. Without this step, the cells could be irreversibly damaged by the presence of the proteolytic enzymes used to detach them from the culture vessel. The serum contains attachment and spreading factors, so it is usually unnecessary to coat the culture vessels when working with serum-containing media. On the other hand, working in SF conditions involves the use of extracellular matrix (ECM) coated tissue culture plasticware components. We recommend using Fibronectin (FN) since it has been shown that this ECM protein contributes, together with PREF-1, to "sealing" the stemness of the hASCs.

S5. Cryopreservation and thawing of the SVF
Developing effective techniques for the cryopreservation of hASCs could increase the usefulness of these cells in tissue engineering and regenerative medicine. Consistently with what has been done previously with the development of a serum-free medium to culture ASCs, we developed a serumfree cryopreservation medium, and we defined a controlled cooling rate protocol ( Figure S6) for freezing and storing the SVF or ASCs in liquid nitrogen. This is the best procedure for the cryopreservation of cells. However, this procedure is feasible only if one has a controlled rate freezing device.
Our simple but effective cryopreservation medium is composed of: • MEM Alpha w/o Phenol Red (# 1-23F22-I, BioConcept, Switzerland) • 10% Cryosure Dex 40 (55% w/v DMSO USP Grade, 5% w/v Dextran 40 USP Grade, WAK -Chemie Medical GmbH; Germany) • 1% injectable human albumin (# 22918180119611, CSL Behring AG, Switzerland) If you do not have a freezing device available or for practical reasons to simplify your work, you can, of course, freeze down efficiently the cells with traditional systems, for example, using «Mr. Frosty Cryo 1 °C Freezing Container» (Nalgene, Thermo Fisher Scientific, USA) with the freezing medium "Synth-a-Freeze CTS" (Thermo Fisher Scientific, Waltham, USA). A suitable serum-free cryopreservation medium which could be prepared in the lab has been described by Yanela González Hernández and René Fischer. FILOCETH is composed of: • 89% medium (TurboDoma, Cell Culture Technologies, Switzerland) The thawing protocol consists of a dilution system. Slow thaw the vial of cells by dilution with complete UrSuppe medium and subsequently seed the cells without washing or centrifuging. After 5-6 hours or the next day, when the cells have adhered, replace the entire medium.

S6. Evaluation of growth-related parameters
Based on the regular measurements of cell density and substrate/metabolite concentration, growth-dependent parameters were calculated for the planar cultivations as follows: (I) Specific growth rate μ (Eq. 1): Where μ is the net specific growth rate. XA(t) and XA(0) are the cell numbers at the end and the beginning of the exponential growth phase, respectively, and t is the time.
(II) Doubling time td (Eq. 2): Where td is the doubling time, ln(2) the binary logarithm of 2, and μ the specific growth rate.
(III) Population Doubling Level PDL (Eq. 3): Where PDL is the number of population doublings, and XA(0) and XA(t) are the cell numbers at the beginning and the end of the cultivation, respectively.
(IV) Expansion factor EF (Eq. 4): Where EF is the expansion factor, XA(tmax) is the maximum cell number, and XA(t=0) is the inoculated cell number.
(V) Lactate yield from glucose YLac/Glc (Eq. 5): Where YLac/Glc is the lactate yield from glucose, ΔLac is the lactate production over a specific time period. ΔGlc is the glucose consumption over the same time period (= exponential growth phase).
(IV) Specific metabolite flux qmet (Eq. 6): Where qmet is the net specific metabolite consumption or production rate (for Glc, Lac, Amn), μ is the specific cell growth rate, XA(t) is the cell number at the end of the exponential growth phase, Cmet(t) and Cmet(0) are the metabolite concentrations at the end and the beginning of the exponential growth phase, respectively, and t is the time.    Table S6, S7, and S8 provide an overview and short description of genes measured in this study. Hudak et al. [4] Hei et al. [5] SOX9 Sox9 is a member of the HMG-box class DNA-binding proteins and is a Pref1 target. It directly binds to c/EBPb and c/EBPd promoters to suppress their promoter activity, thus repressing adipocyte differentiation.
Wang et al. [6] ZFP521 Zinc Finger Protein 521 is a transcription factor that inhibits adipogenesis.
ZFP521 binds to EBF1 and inhibits its transcriptional activity. This leads to a substantially attenuated expression of the proadipogenic factor ZFP423.
Kang et al. [3] Chiarella et al. [7] WISP2 Wnt1-inducible signaling pathway protein 2 is an endogenous and secreted auto/paracrine non-conventional WNT ligand, promoting the proliferation of precursors cells and inhibiting their adipogenic commitment and differentiation.
Shan et al. [11] Ross et al. [12] DLL1 Delta-like protein 1 is one of the five canonical Notch ligands. It inhibits adipogenesis.
Gupta et al. [19] Gupta et al. [20] RUNX2 Runx2 is a transcription factor that is essential for osteoblast differentiation and chondrocyte maturation.
Toshihisa et al. [21] Komori et al. [22] WISP1 The Wnt1-inducible signaling pathway protein-1 increases during adipocyte differentiation, thus stimulating adipogenesis in humans. It is both an intracellular and a secreted protein found in the extracellular matrix. In mice, the effect of WISP1 on adipogenesis was opposite to what has been reported in humans.
Christodoulide et al. [25] Gustafson et al. [26] CD34 CD34 is a transmembrane phosphoglycoprotein expressed on precursors cells and mature adipocytes. Its function on the adipocyte membrane remains to be determined.
CD36 is a transmembrane glycoprotein classified as a class B scavenger receptor. It imports fatty acids inside cells and plays a functional role in adipocyte differentiation and adipogenesis.
Christiaens et al. [30] Gao et al. [31] CD146 Three forms of this adhesion protein have been described, including two transmembrane isoforms and a soluble protein, detectable in the plasma. Its expression increases during adipogenic differentiation.
Fan et al. [34] Zhang et al. [35] Razzaque et al. [36] β-KLOTHO β-klotho, which is highly conserved and localized to the cell membrane, is expressed predominantly in the liver and white adipose tissue.
Ito et al. [37] FABP4 Fatty acid-binding protein 4, also called aP2 (adipocyte protein 2), is a carrier protein for fatty acids primarily expressed in adipocytes.

Burls et al. [38]
Harms et al. [39] ADIPONECTIN Adiponectin is a protein hormone that modulates several metabolic processes and is secreted from adipose tissue.
Lee et al. [40] Stern et al. [41] Villarroya et al. [42] LEPTIN Leptin is a hormone predominantly made by mature adipose cells regulating energy balance by inhibiting hunger and diminishing fat storage in adipocytes.
Harms et al. [39] Stern et al. [41] Caputo et al. [43] UCP1 Thermogenin is an uncoupling protein found in the mitochondria of brown adipose tissue (BAT). It is used to generate heat by nonshivering thermogenesis.
Villarroya et al. [42] Tamucci et al. [44] Wang et al. [45] PGC1α PGC1α is the master regulator of mitochondrial biogenesis, and it is responsible for β-aminobutyric acid secretion. The effect of βaminobutyric acid in white fat includes the activation of thermogenic genes that prompt white adipose tissue browning.
Merrick et al. [49] Rennert et al. [50] CD142 Tissue factor, also called platelet tissue factor, factor III, or CD142, is a protein present in subendothelial tissue and leukocytes. It is the primary initiator of the blood coagulation cascade, which ensures rapid hemostasis in the case of organ damage. It also has signaling activity and promotes inflammatory responses via protease-activated receptors in concert with other coagulation factors.
Schwalie et al. [51] Chu et al. [52] CD248 CD248, also known as endosialin and tumor endothelial marker 1 (TEM-1), plays a role in cell-cell adhesion processes and host defense. This marker does not have a fully characterized role, but its expression has been associated with angiogenesis in the embryo and uterus and tumor development and growth. CD248 could characterize pro-angiogenic subpopulation of SVF cells.
Merrick et al. [49] Brett et al. [53] S7. Differentiation assay to verify hASC multipotency To confirm that cells isolated from adipose tissue and cultured in UrSuppe maintain their multipotency, we perform a differentiation assay. Cells were expanded until passage 2 (P2), changing the medium every 2 days and splitting them at around 80 % of confluence. After reaching P2 the cells were used for multilineage differentiation assay using a commercial induction medium. Briefly: • hMSC differentiation BulletKit™ adipogenic (Lonza, Basel, Switzerland #PT-3004) were used for adipogenic assay: cells were seeded at 20'000 cells/cm 2 in a 6 wells multiwell in basal medium, after 24 hours induction was started with induction medium. The medium was changed every 7 days, induction continue till day 19.
• hMSC differentiation BulletKit™ osteogenic (Lonza, Basel, Switzerland #PT-3002) was used for the osteogenic assay. The cells were seeded at 3'000 cells/cm 2 in a 6 wells multiwell in basal medium. After 24 hours induction was started with an induction medium. The medium was changed every 3-4 days, induction continue till day 19 • hMSC differentiation BulletKit™ chondrogenic (Lonza, Basel, Switzerland #PT-3003) was used for chondrogenic assay: cells were pelleted at 250'000 cells/pellet in different 15 mL TPP tubes in differentiation BulletKit medium. The medium was changed every 3-4 days, induction continue till day 19. All manipulations were done according to the manufacturer's instructions and recommendations. After 19 days of induction, we performed histochemical assays with the samples. For the adipogenic and osteogenic induced samples: Cells were fixed with a 4% formaldehyde/0.2% glutaraldehyde solution for 20 minutes and washed three times with PBS. After that, the cells were incubated with the specific staining solution: Oil-Red-O (Sigma-Aldrich, cat no. 75087) to detect fat vacuoles in the cells (adipo-induction) and Alizarin-Red S (Sigma-Aldrich, cat. no. 05600) to stain calcium-rich extracellular matrix (osteo-induction). For the chondrogenic induced samples: The highdensity micromass pellets were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into 5μm thick sections with a rotation microtome to examine chondrogenesis. The micro-slices were put onto adhesive microscopic slides, dried on a heating plate for half an hour, and dried overnight at 40°C. Afterward, the microscopic slides were stained with Alcian Blue (Sigma-Aldrich, cat. no. A3157) to highlight the regions saturated with an extracellular matrix composed of acidic polysaccharides that are usually highly expressed in cartilage (chondro-induction). After the stainings, all samples were documented and analyzed, and Figure S7 shows the results of a representative trilineage differentiation assay. These tests confirmed that hASCs expanded with the UrSuppe medium remained undifferentiated and capable of trilineage differentiation.  Table S9. Qualitative evaluation of the factors secreted by undifferentiated hASCs and WAT or BAT induced cells. The qualitative evaluation for each detected factor is indicated by the number of X. Shallow expression is indicated by (X), whereas no expression by (-). Quantitative results are shown in Figure 11.     Lensch et al. [58] Stemline Sigma Xeno-Free None n/a n/a