Adipose‐derived stem cell spheroids are superior to single‐cell suspensions to improve fat autograft long‐term survival

Abstract Autologous fat transplantation is a widely used procedure for surgical reconstruction of tissues. The resorption rate of this transplantation remains high and unpredictable, reinforcing the need of adjuvant treatments that increase the long‐term stability of grafts. Adipose‐derived stem cells (ASC) introduced as single cells in fat has been shown clinically to reduce the resorption of fat grafts. On the other hand, the formulation of ASC into cell spheroids results in the enhancement of their regenerative potential. In this study, we developed a novel method to produce highly homogeneous ASC spheroids and characterized their features and efficacy on fat transplantation. Spheroids conserved ASC markers and multipotency. A regenerative gene expression profile was maintained, and genes linked to autophagy were upregulated whereas proliferation was decreased. Their secreted proteome was enriched in comparison with single‐cell ASC suspension. Addition of spheroids to fat graft in an animal model of transplantation resulted in a better graft long‐term stability when compared to single ASC suspension. In conclusion, we provide a novel method to manufacture homogenous ASC spheroids. These ASC spheroids are superior to ASC in single‐cell suspension to improve the stability of fat transplants, reinforcing their potential in reconstructive surgery.

types and, importantly, produce a large spectrum of factors involved in tissue regeneration and healing. The International Society for Cellular Therapy and International Fat Applied Technology Society established the criteria, which determine ASC in vitro 5 : (i) adherence to plastic surface in standard culture conditions (ii) defined expression of surface antigens while hematopoietic markers must be absent (iii) ability to multilineage differentiation into adipocytes, osteocytes and chondrocytes in vitro. The ability of ASC to produce factors involved in tissue regeneration makes them highly attractive for cell therapy applications. Although their differentiation capacities observed in vitro are now considered to be absent in vivo after injection, 6,7 therapeutic effects are attributed to their secretome favouring local healing. The ASC secretome has been extensively studied 8 and contains factors enhancing the regenerative process.
Recent studies have shown that fat grafts enriched with ex vivo expanded ASC in single-cell suspension markedly improved residual graft volume and histological appearance 9,10 due to their regenerative secretome. Plastic surgeons introduced clinically the technique of fat graft enrichment with ASC in single-cell suspension. Compared with the control grafts, the ASC-enriched fat grafts (injected in the posterior part of upper arms) had significantly higher residual volumes. 11 The procedure had excellent safety, reinforcing the prospect of ASC use in clinical settings.
On the other hand, cell spheroids have gained considerable popularity. [12][13][14] These structures allow a 3D cell-to-cell and cell-to-matrix interactions which guarantee cell functions, nutrient and oxygen gradients closer to the in vivo situation, as compared to 2D cell monolayers growing on plastic or single-cell suspensions. Numerous reports demonstrate that the aggregation of mesenchymal stem cells or ASC into spheroids results in the enhancement of their therapeutic potential. 13,15,16 Compared to their single-cell suspensions equivalents, MSC or ASC spheroids showed increased regenerative properties with a higher production of angiogenic and immunomodulatory factors. [17][18][19][20][21] In vivo, by using models of kidney degeneration, 22 hepatic fibrosis 23 or failure, 24 urinary tract degeneration, 25 wounds 18,26,27 and lungs disease, 28 MSC or ASC spheroids showed a higher regenerative potential than single-cell suspensions. They notably displayed better survival in ischaemic conditions 29 and higher resistance to oxidative stress-induced apoptosis. 30 In this study, we developed a novel method to manufacture stable and highly homogeneous human ASC spheroids suitable for cell therapy and compared the regenerative effects of single ASC suspensions versus spheroids on fat transplantation.

| ASC culture and differentiation
Different human ASC lines were prepared from the fat of patients as described (Uckay, I., et al., J Stem Cell Res Ther, 2019. 9). Cells were cultured in Dulbecco's Modified Eagle Medium DMEM (4.5 g/L glucose, L-Glutamine) supplemented with 10% of human platelet lysate (Stemulate, Cook Regentek), 1% penicillin and streptomycin (Thermo Fisher) at 37°C and under 5% CO2. This study was conducted according to the approval by local ethical committee of the

| Moulding and 3D printing of pads
Pads were created by moulding Aggrewell-800 plates (Stemcell) with a Sylgard 184 (Sigma-Aldrich) silicone elastomer kit according to the supplier's instructions (Dow). In other experiments, the pads were 3D-printed. They were designed by using an open-source parametric 3D modeler (FreeCAD 0.18). Generated designs were then exported as STL format files and transduced into a G-code file with the software Autodesk Netfab2020. The resulting file was imported into a digital light processing (DLP) 3D printer (Rapidshape P30 series, Straumann ® ) with a UV385 nm high power led and a HD resolution of 1920 × 1080 px. An acrylic resin (Sheraprint model plus sand UV) was used to print the models with a layer thickness of 50 µm. Once printed, the models were washed two times for 6 min in isopropanol in an ultrasonic bath (28-34 kHz) at room temperature. Once dried, the models were further cured in a curing unit (Shera flashlight-+3D): 2 cycles of 2000 flashes (100 W/280-700 nm/10 flashes per seconds) before sterilization and moulding.

| Spheroids formation
For the moulding of microwells in agarose, heated ultrapure agarose (LE, Analytical Grade, Promega) at 2.5% in ultrapure water was deposited on Milicell inserts (Merckmillipore, 0,4 µm). The pad was then immediately placed in the heated agarose. After 10 min for complete polymerization of agarose, the pad was gently removed.
To form the spheroids, the ASC suspension in their culture medium was filtered through a cell strainer (Corning, 40 µm) and 1 ml of suspension was deposited on agarose/inserts in 6-well plates. The plate was then centrifuged at 100 g for 5 min to force aggregation of ASC in microwells. To establish air/liquid interface conditions, the residual medium in the insert was removed by aspiration with a 0.5 × 16 mm needle and 1 ml of fresh medium were added in each well of the 6-well plate.

| Determination of spheroids viability
The viability of the spheroids was controlled with a fluorescent labelling kit (Ibidi) with fluorescein diacetate (FDA) and propidium iodide (PI), according to the supplier's instructions.

| Immunofluorescence, histological colorations
Spheroids were washed in PBS and then fixed with a 4% paraformaldehyde solution for 20 min at room temperature. The spheroids were then rinsed in PBS and replaced in PBS supplemented with bovine serum albumin 1%, Triton X-100 0.3%. The spheroids were then incubated overnight at +4°C with the primary anti-Stro1 Mouse IgM (Life Technologies). After three washes in PBS, incubation with Alexa555 secondary anti-Stro1 Mouse IgM (Invitrogen) was performed for 90 min. Again, a step of three washes was performed before the nucleus staining with DAPI (300 nM). After three successive washes in PBS, cells were rinsed in pure water and fixed in FluorSave mounting medium (Calbiochem). Histological coloration with hemalun-eosine was done according to the standard protocol.

| Molecular biology
Microarray analysis was performed with an Affymetrix chip target- with statistical significance (FDR < 0.05) were taken into account (Table S1). Firefly luciferase was introduced in lentivectors under the control of the ubiquitous promoter of short elongation factor (EFS), and cells were transduced, as described previously. 31

| Mass spectrometry
Two million cells in 2D or spheroids were washed with a serum-free DMEM (Thermo Fisher) and cultured overnight at 37°C in 1 ml of medium. Medium was collected and centrifuged at 500 g for 10 min to remove residual cells and free nuclei. Proteins were precipitated, digested and peptides were analysed by nanoLC-MSMS using an easynLC1000 (Thermo Fisher) coupled with a Q Exactive HF mass spectrometer (Thermo Fisher). Database searches were performed with Mascot (Matrix Science) using the Human Reference Proteome database (Uniprot). Data were analysed and validated with Scaffold with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at >22.0% probability to achieve an FDR < 1.0% and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm. 33 Proteins were annotated with GO terms from NCBI. 34 Experiments were performed in biological duplicates, established from 2 independent ASC lines from two different patients. Functional protein class analysis was done by PANTHER (Protein ANalysis THrough Evolutionary Relationships, www.panth erdb.org) classification system and a stringDB analysis (using STRING database, www.strin g-db.org) of known and predicted protein-protein interactions. Data analysis was done with the FlowJo software.

| Manufacturing of individualized, homogeneous ASC spheroids by forced aggregation in agarose-moulded microwells combined with air/liquid interface conditions
To manufacture ASC spheroids suitable for cell therapy in a large scale, forced aggregation is the most attractive method. However, ASC are highly adhesive cells and spheroids generated with this method attach to cell microplates, strongly limiting their homogeneity. To overpass these limitations, we generated non-adherent micro- These larger microwells allowed the compaction of a high number of ASC, having tested 5000 cells and 10,000 cells, respectively ( Figure S1D). We also manufactured spheroids using pads with small pyramids ( Figure S1E), producing small microwells in agarose ( Figure   S1F,G). Smaller spheroids ranging from 250 cells to 1000 cells were manufactured in these smaller microwells ( Figure S1H). In this study, focusing on the efficacy of ASC spheroids on fat transplantation, we made the choice to characterize spheroids containing 1000 ASC.
Indeed, this number of cells created an optimal size suitable for injection through surgical needles and mixing with fat tissue.
Adipose-derived stem cells spheroids after 48 h did not attach to the microwells and harboured a round shape ( Figure 1D). The size of 30 spheroids per line was measured using ImageJ software, and the data were analysed (Table 1). For the all tested lines, the standard deviation was very low, under 5% of the diameter. Shape descriptors were analysed on 4 different ASC lines by the Fiji ImageJ software where six parameters were measured (n ≥ 28). All the parameters, including circularity, perimeter, aspect ratio, Feret's diameter, solidity and area, were highly homogeneous within the same line and among the different lines ( Figure S2) with homogeneous values and low standard deviations. As ASC spheroids containing 1000 cells were stable and homogeneous after two days, and the importance to reduce time of manufacturing in clinical-grade conditions, we characterized and transplanted ASC spheroids of 2 days.

| Characterization of ASC spheroids and comparison with ASC in single cells.
To study the viability of spheroids made in these air-liquid inter- week spheroids ( Figure 2B). On the other hand, a full gene expression profile was also performed in spheroids and compared with ASC in 2D conventional culture conditions. mRNA corresponding to ASC markers was found in both spheroids and 2D conditions, with the exception of CD44 ( Figure 2C). Complete data of microarrays analysis are shown in Table S2. As CD44 expression was seen in flow cytometry ( Figure 2B), we concluded to a weakness of CD44 de-  Table S3.
Together, these observations show that the secreted proteome of ASC grown in spheroids was enriched in their diversity. Of note, some proteins were found from ASC in 2D but not secreted by the spheroids. To understand which proteins were not secreted after 3D formulation, a functional protein class analysis was done by the PANTHER (Protein ANalysis THrough Evolutionary Relationships) classification system. Among the proteins classes not found in the spheroid secretome, those from the extracellular matrix were mostly represented, probably highly mobilized in the 3D structure between compacted cells ( Figure 4B). Other classes were also represented in this regulation, including 'protein modifying enzymes' and 'protein binding activity modulator'. The reduced extracellular matrix pro-

| Spheroids increase the engraftment of fat tissue in an animal model of lipofilling
The been demonstrated that human collagen inhibits MSC proliferation, 37 and having confirmed the increased collagen production by ASC in spheroids, we hypothesized that the decreased proliferation was linked to cell-to-cell or cell-to-matrix interactions. Moreover, and more generally, cells in more hypoxic zones of spheroids have a decreased proliferation. 38 Stemness markers (Nanog, Sox-2, Oct-4) were however not altered in ASC spheroids (data not shown), making the selection of stem cell niches in these conditions unlikely.
Transcriptome and secretome data suggest an increase of cell autophagy in spheroids compared to monolayers. Autophagy generally helps the cells to adapt to stress conditions and maintain their homeostasis, 39 thereby enabling the tissues to maintain a controlled growth and development. Hence, cores in ASC spheroids could induce adaptative autophagic response to an hypoxic or metabolic stress to maintain the cell integrity and viability. Autophagy is a major inducer of vessels formation 40 and was shown to protect MSC from apoptosis under hypoxic conditions. 41 Moreover, although not confirmed in our study, autophagic MSC were described in some studies to accelerate regeneration. 40 The regenerative secretome of ASC spheroids is thought to play a critical role in cell therapy applications. The first observation is that ASC spheroids, compared to the same number of cells in monolayer, have quantitatively a reduced total secreted proteome outside the spheroids. However, the protein diversity is higher. The  43 We hypothesize that the therapeutic effect of ASC spheroids is mediated by their secretome and possibly production of extracellular matrix. Regarding the clear advantage of MSC spheroids for regeneration in many preclinical reports, 13,15,16 further studies are required to better understand the mechanisms of action explaining their advantage.

| CON CLUS ION
In conclusion, ASC spheroids have a clinical potential to increase the engraftment of fat tissue more efficiently than ASC formulated in single-cell suspension. Geneva, for their technical support.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.
Stephane Durual: Conceptualization (equal); Methodology (equal).  The size of fat transplants (n ≥ 3) was monitored after transplantation for 16 weeks by high resolution micro-CT scan imaging. Two-way ANOVA test statistics are indicated for each time point. Letters indicate significant differences between groups with a: comparison between Control and ASC 2D, b: comparison between ASC 2D and ASC spheroids and c: comparison between Control and ASC spheroids. Corresponding p values: a,b,c, p < 0.05; aa,bb,cc, p < 0.01; aaa,bbb,ccc, p < 0.001 and cccc, p < 0.0001