Co-Transplantation of Marginal Mass Allogeneic Islets with 3D-Culture-Derived Adult Human Skin Cells Improves Glycemia in Diabetic Mice

Background. Pancreatic islets transplantation represents a promising therapeutic option for severe type 1 diabetes (T1D). Maintenance of long-term viability of transplanted islets still requires improvement. Stem cell use represents an option to repair and replace damaged islets or, alternatively, b cells in T1D. Mesenchymal stem cells (MSC) have been proposed as adjuvants for islet transplantation, facilitating grafting and improving their functionality. Aggregation of stem cells has gained interest in providing physiological interactions between cells and enhancing the in situ concentration of modulators of inammation and immunity. Methods. We established a hanging-drop culture by the spontaneous aggregation of adult broblast-like cells as spheroids. Adult skin spheroid-derived cells (SphCs) were characterized in vitro and in vivo. We assessed the potential benet of SphCs as adjuvants to improve islet functionality by cotransplantation with a marginal mass of allogeneic islets in an experimental diabetic mouse model. We characterized the secretome of SphCs by mass spectrometry-based proteomics. Conclusions. Hanging-drop cell culture methodology might contribute to the development of an ecient way to improve transplantation outcome by reducing undesirable consequences of pharmacological immunosuppression as well as the number of allogeneic islets required to achieve normoglycemia in T1D transplanted patients. Further studies might determine whether the identied proteins sustain immunomodulation and/or cytoprotective effects in transplanted allogeneic islets. The top hits on the functions displayed by these proteins were related to extracellular matrix remodeling, modulation of apoptosis and cell differentiation. reported that low levels of some matrix metalloproteases might lead to an increased activity of membrane-bound FASL. Consequently, one might expect augmented apoptotic capacity by FASL-expressed SphCs on effector T-cells without affecting the viability of Treg lymphocytes, both contributing to a signicant increase in b cell viability and islet graft lifespan 81). that inhibitors of metalloproteases were upregulated in the SphCs secretome and could, partially, contribute to the less aggressive proinammatory microenvironment created when these cells were cocultured with murine splenocytes. The protein 14-3-3ζ, also upregulated in the SphCs secretome, has been implicated in the inhibition of b cell apoptosis induced by different stimuli (82). Both Gremlin-1 and pentatraxin-3, which are abundant in the secretome of SphCs, have been related to growth and survival in response to inammation and tissue damage There is no knowledge about gremlin-1 in the context of b cell biology. In tubular epithelial cells, gremlin-1 increased TFG-b production through Smad activation, inducing a myobroblast-like phenotype Taken together, these results highlight the underscored ability of SphC to create a favorable niche for allo-islet reception, islet viability


Introduction
T1D arises with progressive dysfunction and loss of insulin-producing β cells due to an autoimmune process; its incidence has increased markedly in recent years (1,2). Insulin remains as a palliative for T1D. Therapeutic interventions involving immunosuppressive drugs failed to achieve lasting remission, and bene cial effects were only observed for short periods of time during treatment (3)(4)(5).
Attention has been focused on the use of cell therapy to repair damaged tissues, including islets and β cells, in T1D (6, 7). Notably, the signi cant advance introduced by the use of potent induction therapy in addition to the management of anti-cytokines before transplantation leads to 5-year insulin independence rates greater than 50%, as reported by the 2014 Collaborative Islet Transplant Registry (CITR) (8, 9).
Stem cells have shown promising experimental results, but their use has faced limited effectiveness owing to the low survival rates and susceptibility to environmental stress after transplantation (10). Multipotent stem cells, including mesenchymal stromal cells (MSCs), differentiate into cell types of a speci c germ layer (11). Embryonic stem (ES) cells and induced pluripotent stem cells (iPSCs) possess the potential to give rise to cells of all three germ lineages (e.g., mesodermal, endodermal, and ectodermal) (12,13). Concerns have arisen regarding the use of ES cells due to teratoma formation through uncontrolled self-renewal and triploblastic differentiation. Although iPSC use resolved the bioethical issues concerning embryonic human origin, the high rate of teratoma formation on transplantation has impeded their use in regenerative medicine.
MSCs can be easily isolated and expanded in vitro and have been employed in T1D clinical trials (14). Although studies have reported a modest preservation of residual β cells after MSCs infusion, prospective long-term studies showing faithful insulin independence are needed (15)(16)(17).
Distinctive pluripotent stem cells have been isolated from adult human tissues. Multilineage-differentiating stress-enduring (Muse) cells are stress tolerant and express pluripotency markers (18,19). They are good candidates for treating in ammation as well as altering Th1 lymphocyte activity in an antigen-speci c way (20). Infusion of Muse cells attenuated hyperglycemia in an experimental mouse model of T1D (21).
Allogeneic islet transplantation in severe T1D has become a reliable therapeutic alternative (22). Islet transplantation should achieve precise glycemic control over insulin administration. However, it remains a restricted therapy due to a shortage of cadaveric donors and lack of adequate immunosuppressive regimens (7). Very few transplant patients achieve insulin independence for periods longer than a year (22). To improve functionality along with insulin independence, the use of MSCs to provide a favorable niche for islet engraftment arose as an option to be tested in T1D models (23). Cotransplantation of MSCs and islets ensures the provision of factors that improve long-term islet viability (24). MSCs increase insulin secretion in vitro (25) and improve transplantation outcomes (23). MSCs modulate immune responses by secreting cytokines and soluble factors in the microenvironment (26,27). Immune modulation is key in the context of pancreatic islet allotransplantation, and MSCs help prevent host rejection and nourish the graft (28, 29).
MSCs grown as spheroids, meaning a 3D MSCs culture, enhance the secretion of downregulators of the immune response (30,31), bringing our attention to using this methodology to improve islet allotransplantation in a diabetic mouse model. We describe that spontaneous formation of spheroids confers adult human skin broblast-like cell characteristics that improve marginal mass allogeneic islet transplantation in a diabetic mouse model.
The quantitative spheroid-derived cells (SphCs) secretome differs from that of its paired skin broblast-like counterparts in containing proteins with putative bene cial effects on islet function and survival. Further studies might determine whether these proteins sustain immunomodulation and cytoprotective effects in transplanted islets. This methodology might contribute to improving transplantation outcome by reducing undesirable consequences of pharmacological immunosuppression as well as the number of islets needed to achieve normoglycemia in T1D recipients.

Materials And Methods
Mice BALB/c, C57BL/6 and BALB/c nude mice (6-9 weeks old) were from the animal care facility (Chemistry Institute; University of São Paulo).

Islets isolation and treatments
Pancreatic islets were isolated from BALB/c male mice (32). Islets were pretreated for 1 h with 20% conditioned medium from skin broblasts or SphCs, followed by exposure to a proin ammatory cytokine cocktail (TNFα 8 ng/ml, IFNγ 4 ng/ml and IL-1β 0.8 ng/mL; PeproTech, USA). After 24 h, islets were labeled with 1 μM Newport Green, 1.5 µM propidium iodide (PI), and 1 µg/ml Hoechst 33342. The death rate of each islet was considered the PI-positive area (nonviable cells) divided by the total area of the islet. Quanti cation was evaluated by ImageJ software. Nano LC-MS/MS analysis Peptides were suspended in 0.1% formic acid (FA) before analysis using a nano ow EASY-nLC™ II system (Thermo Scienti c) coupled to an LTQ-Orbitrap Velos mass spectrometer (Thermo Scienti c). Then, the samples were loaded on an Acclaim PepMap C18 (Thermo Germany) trap column (2 cm x 100 µm; inner diameter 5 µm) and separated onto an Acclaim PepMap C18 (15 cm x 75 µm; inner diameter 3 µm) column. A 70 min gradient was used from 100% mobile phase A (0.1% FA) to 34% phase B (0.1% FA, 95% ACN), 34%-95% at a constant ow rate of 250 nl/min. The mass spectrometer was operated in positive ion mode with data-dependent acquisition. The full scan was acquired in the Orbitrap at 60,000 FWHM resolution in the 400-1600 m/z mass range. Peptide ions were fragmented with CID using a normalized collision energy of 35. Data-dependent acquisition was used to select the 20 most abundant precursor ions for fragmentation. Raw data were accessed in Xcalibur software (Thermo Scienti c).

Database searches and bioinformatics analyses
Raw data were processed using MaxQuant software version 1. Proteins with at least two peptides and two ratio counts were accepted for further validation. Label-free quanti cation was performed using MaxQuant software with the "match between run" feature activated. Statistical analyses were performed on the Perseus version.1.6.10.43 platform. Contaminants and proteins identi ed in the reverse database were excluded before the statistical analyses. Analyses were ltered to include peptides identi ed and quanti ed in 2 replicates in at least one condition.
Enriched gene ontology (GO) terms (biological processes and molecular functions) for statistically regulated proteins found in the SphC secretome were analyzed using Panther (33). Enriched molecular pathways were analyzed by KEGG (34), Reactome (35) and WikiPathways (36) platforms at a threshold of q-value < 0.05, corrected by Benjamini-Hochberg FDR. Tissue expression analysis was performed using the Human Protein Atlas database (37).

Statistical Analysis
Data were rst analyzed for Gaussian distribution. Differences among groups were compared using two-tailed, nonpaired Student's t test or ANOVA followed by Tukey's posttest. Analyses were performed using GraphPad Prism 5 software (GraphPad Software, Inc., CA). Differences were considered signi cant at p 0.05.

Ethics approval and consent to participate
Mouse studies were approved by the Institutional Care and Use Committee (Comissão de Ética em Uso da Animais do Instituto de Quimica, Universidade de São Paulo (CEUA) n°42/2016). All methods were carried out in accordance with relevant guidelines and regulations. All methods are reported in accordance with ARRIVE guidelines.
Human samples were obtained with informed consent according to the Ethics Committee. All methods were carried out in accordance with relevant guidelines and regulations. All experimental protocols were approved by a named institutional and/or licensing Committee. The study protocol was approved by Comité de Ética em Pesquisa do Hospital Universitario da Universidade de São Paulo (CEP-HU/USP 958/09).

Results
Human skin cell culture in hanging drops generates a three-dimensional structure named spheroids Adult human skin cells in adherent plastic showed and maintained a broblast-like morphology ( Figure 1A). A total of 5x10 4 skin broblastlike cells at 30 µl/drop resulted in the optimal initial number of cells for spontaneous spheroid generation. Increasing the cell number generated larger spheroids with necrotic cells inside (data not shown). Figure 1B shows a representative compact spheroid with more than 95% of healthy (Trypan blue negative) cells reaching an average size between 500-700 µm at 72 h. Spheroids were then plated back on adherent plastic dishes, allowing the spontaneous migration of single cells to form a uniform monolayer ( Figure 1C). These cells morphologically resembled the original skin primary cells used for spheroid formation, and we named them spheroid-derived cells (SphCs) ( Figure 1D).

Characterization of Spheroid-derived Cells (SphCs) from adult skin broblast-like cells
SphCs were immunophenotyped by using uorochrome-conjugated antibodies against CD29, CD73, CD90 and CD105 followed by ow cytometry analysis. The expression of CD31, CD34, CD45, CD11b and HLA-DR was also evaluated. SphCs gained the expression of surface CD105 (approximately 40% increment) in comparison with their skin parental cells from which they were generated ( Figure 2A). Skin cells cultured as spheroids in a single culture showed an increase in double-positive expression of CD105 + /CD90 + , CD105 + /CD73 + and CD105 + /CD29 + ( Figure 2B).
SphCs e ciently differentiated into adipocytes, chondrocytes and osteocytes, as revealed by tissue-speci c staining, similar to MSCs ( Figure 3A, B). Considering SphCs' characteristics, such as growing on adherent plastic, expressing speci c surface antigens and showing multipotent differentiation potential, they truly reach the minimal criteria to de ne them as MSC-like cells (38).
To investigate their potential as therapeutic tools, we evaluated SphCs by testing their tumorigenic capacity in a favorable environment, such as that offered by immune-compromised mice. Animals were followed up to 180 days postadministration. Tumorigenic P19 cells generated a tumor at 21 days post i.t (not shown). SphCs did not form tumors ( Figure 3C), nor was any abnormal tissue observed after histochemical evaluation at 180 days (not shown).

Spheroid-derived cell (SphCs) downmodulate immune responses in vitro
BALB/c splenocytes were employed to evaluate SphCs immune regulatory capacity (40). Anti-CD3-stimulated T lymphocyte proliferation was compared under the in uence of SphCs or, alternatively, their skin broblast-like counterparts. There was a reduction in T-CD4 + lymphocyte proliferation when these cells were cocultured with SphCs, while the presence of skin broblast-like cells did not alter proliferation rates ( Figure 4A).
We analyzed the pro le of cytokines secreted by lymphocytes stimulated with anti-CD3 in the presence of SphCs in vitro. Cytokines were quanti ed by ow cytometry in the 36 h-spent supernatant of anti-CD3-primed splenocytes cocultured with either skin broblast-like cells or, alternatively, SphCs. The hallmark Th1 cytokine INF-g was dramatically downregulated by the presence of both SphCs and skin broblastlike cells ( Figure 4B). Relevant cytokines required to mount an effective proin ammatory response, such as IL-2, IL-6, and TNF-a, were less abundant in the supernatant derived from both skin broblast-like cells and SphCs than in the positive control (anti-CD3 alone), and the levels were even lower in the presence of SphCs ( Figure 4C, D, E). IL-4 favors the induction of a Th2 lymphocyte pro le (41). SphCs induced a 4-fold increase in the secretion of IL-4 in comparison with anti-CD3 stimulation, while skin broblast cells did not show any effect ( Figure   4F). Regarding IL-10, a classic immune-modulatory cytokine, SphCs did not stimulate its secretion after CD3 activation ( Figure 4G). There were no differences in the secretion of IL-17A between cocultures and the control (not shown).
Taking into account the proliferation and cytokine pro le of T-CD4 + lymphocytes, SphCs are capable of inducing a bias toward a decrease in the ratio of Th1/Th2 lymphocytes, supporting an anti-in ammatory environment.
Spheroid-derived cell (SphC)-conditioned media attenuates proin ammatory cytokine-induced islet death Islets cultured with a cocktail of proin ammatory cytokines, mimicking the unfavorable microenvironment to which they are exposed during the autoimmune process of T1D, die by apoptosis (42)(43)(44)(45). We assessed whether soluble factors secreted by SphCs might protect isolated islets from the proin ammatory damage exerted by cytokines. After 24 h, the cocktail of proin ammatory cytokines induced a 4-fold increase in islet death relative to the control (islets in optimal conditions, e.g., 10% FBS). SphCs supernatants containing soluble factors reduced islet death in comparison with proin ammatory cytokine stimuli (approximately 50% reduction). There was a lack of bene cial effect on islet survival employing conditioned media from skin broblast-like cells ( Figure 5A).
Conditioned media from both SphCs and skin broblast-like cells did not perturb glucose-stimulated insulin secretion or total islet-insulin content, indicating that soluble factor/s would only affect islet survival ( Figure 5B;C).

Cotransplantation of marginal mass allogeneic islets with SphCs improves glycemic control in diabetic mice
We established the marginal mass of allogeneic islets using 300 IEQ transplanted under the kidney capsule of diabetic C57BL/6 mice. Transplantation of 300 IEQ maintained glycemic levels slightly below those of sham-operated mice ( Figure 6A). As a control, engraftment of 600 IEQ from BALB/c mice into C57BL/6 diabetic mice dramatically decreased glycemia as soon as 1 day postadministration (not shown). Figure 6B shows animal weights throughout the experimental period. The marginal mass of allogeneic islets contributed to the maintenance of body weight, while diabetic surgical sham animals lost signi cant body weight ( Figure 6B). Cotransplantation of the marginal mass of allogeneic islets with SphCs (10 4 ) improved glycemia soon after administration compared with other experimental groups ( Figure 6C).

Spheroid-derived cells (SphCs) protein secretome analysis
Having established that conditioned media of SphCs both immunomodulate T-CD4 + lymphocyte responses and exert cytoprotection on cytokine-challenged islets, we analyzed the protein secretome of SphCs and primary skin broblast-like cells using a comparative proteomic approach. After comparing the abundance of proteins present in both secretomes, we classi ed protein components identi ed in three different groups according to the ratio of abundance in the secretome of SphCs versus the one obtained in the corresponding primary skin culture. We established a cutoff value ≥1.5 or ≤0.65 for more abundant and less abundant proteins, respectively. Tables 1 and 2 (Suppl. material) summarize the protein components of SphCs presenting an increase (upregulated) or decrease (downregulated) in relative abundance when compared with the corresponding skin broblast-like cells. Approximately 33% of the proteins were upregulated in the SphC secretome, while almost 46% of them were downregulated.
Additionally, each set of identi ed proteins was submitted to interactome analysis to understand their putative cellular functions (Figure 7, Suppl. Tables 3 and 4). Among the underexpressed proteins, we identi ed proteins related to extracellular matrix remodeling (e.g., actin, laminin, tubulin, collagen, elastin, lumican, SPARC and thrombospondina-1), collagen formation, complement C1-related proteins, others associated with elastic bers (elastin and other proteins), and glycosaminoglycan metabolism (which have functions in the regulation of cell growth, proliferation, promotion of cell adhesion and anticoagulation).
Among the upregulated proteins, there were Rho GTPases that coordinately activate several transduction pathways, proteins involved in glucose uptake and its metabolism (GLUT4, enolases), regulation of apoptosis (e.g., 14-3-3) and other mechanisms involved in cellular development, cell survival (e.g., gremlin) and stress response (pentraxin-3), and mig lin, which regulates remodeling, morphology and motility processes.
In summary, this protein secretome analysis highlights the expression pattern of proteins secreted by SphCs that might be responsible for islet survival, acceleration of glucose metabolism, and tissue repair.

Discussion
In persons with T1D, most of their b cells are selectively destroyed, and some may remain dedifferentiated (46). Exogenous insulin administration schemes often fail to maintain ne-tuned glycemic control, which may lead to frequent unwarned hypoglycemic/hyperglycemic episodes associated with side effects and death risk. Transplantation of whole solid pancreas or isolated islets from cadaveric donors constitutes valid options to achieve insulin independence and/or diminish glycemia uctuations. Pancreas transplantation is considered invasive and is often associated with comorbidities (47). Therefore, transplantation of isolated islets is presented as an appealing therapeutic option for labile T1D patients. The accessibility of islet transplantation as a wider therapeutic intervention is very limited due to a shortage of cadaveric donors and the existence of few centers worldwide with adequate expertise (https://citregistry.org/home).
Efforts in clinical research have contributed signi cantly to improving islet transplantation, achieving exogenous insulin independence in 50% of patients after 5 years of multidonor infusions. The latter constitutes a success rate similar to that of organ pancreas transplantation (48). These patients manifested a reduced frequency of hypoglycemic events.
Preservation of islet viability during and after isolation remains challenging because of a signi cant loss of functionality during islet preinfusion due to both mechanical and oxidative stress as well as activation of apoptotic pathways (49,50). Therefore, novel procedures toward maintenance of islet viability and normal functionality before and after infusion will improve islet transplantation outcomes to e ciently treat many more people with labile T1D. In this scenario, cotransplantation of stem cells together with isolated islets has been investigated as a putative translational option for T1D. MSCs have been noted because of their anti-in ammatory, angiogenic and immune-regulatory capacities as well as their source of nourishing factors for islets (28, [51][52][53]. Cotransplantation of MSCs with islets improved their survival and function in experimental diabetic mice using syngeneic and allogeneic settings and in immune-de cient mice (21, 28, 54-56).
MSCs possess the capacity to repair or attenuate tissue destruction by paracrine secretions or cell-to-cell contacts modulating in ammatory and immune reactions (11,57). MSCs possess the capacity to repair tissues or attenuate destructive processes by paracrine secretion of cytokines as well as other factors, some of which have already been identi ed, such as tumor necrosis factor-inducible gene 6 (TSG-6) and stanniocalcin-1 (STC-1), which are able to decrease both in ammation and immune responses (30,58). Furthermore, MSCs cultured as spheroids increase the surface expression of CXCR4 and adhesion to endothelial cells (58). Based on these reports and taking into account that obtaining skin tissue from adult humans is safe and minimally invasive, we set out to evaluate whether adult human skin broblast-like cells after aggregation in culture could acquire the aforementioned characteristics and thus contribute to improving islet transplantation outcomes in diabetic mice.
We describe a culture procedure that confers broblast-like cells obtained from adult human skin properties that improve marginal mass allogeneic islet transplantation in experimental diabetic mice. These results highlighted that the bene cial effects of cotransplantation of SphCs with marginal mass allogeneic islets could be due to its in uence on the facilitation of islet engraftment and, inderectly, in the attenuation of the host's alloimmune response. The observed overall increase in islet cell viability nally led to better glycemic control in diabetic transplanted mice.
Adult human skin broblast-like cells cultured for 72 h as hanging drops formed spheroids of ~500/700 µm in diameter and composed of more than 95% living cells, similar to what has been reported using bone marrow (30). These cells spontaneously aggregated under gravity when cultured in hanging drops without the addition of any speci c growth factor. We believe that the phenomenon of aggregation facilitates cell-to-cell contacts that might aid in maintaining the stemness of SphCs, as has already been demonstrated using 3D culture of MSCs (59, 60).
Spontaneously migrating cells from spheroids, e.g., SphCs, expanded in adherent plastic and retained the capacity to differentiate after induction in adipocytes, osteocytes and chondrocytes, similar to conventional MSCs. SphCs gained the expression of several CDs distinctive of MSCs (i.e., CD29/CD73/CD90 and CD105) with very low or absolute lack of expression of CD11b, CD31, CD34, CD45 and CD34 surface molecules (39). Since we found no expression of HLA-DR molecules on the surface of SphCs, they may remain undetectable to allogeneic effector T-CD4 + lymphocytes, as already reported in other studies using different sources of cells (61). After several months of SphCs administration, they did not show proliferation, maldifferentiation, or teratoma formation in immunode cient mice. These last characteristics position SphCs as a very safe alternative with potential therapeutic use.

SphCs have immunoregulatory properties similar to MSCs (62). Our studies offer new insights into the immunomodulatory capacities of
SphCs, focusing on their ability to diminish T-CD4 + cell proliferation and biasing the cytokine pro le of stimulated splenocytes toward an increase in the Th2/Th1 ratio in vitro. We have previously described potent immunomodulatory features of a unique human stem cell population called multilineage-differentiating stress-enduring (Muse) cells derived from adipose tissue (20). Muse cells and ShpCs have the ability to grow in suspension as spheroids. Therefore, aggregation of cells forming clusters might be a key driving force to make cells express cytokines and soluble factors such as TGF-b1 released by Muse cells as prominent mediators of their immunomodulatory actions (20).
In vitro downmodulation of splenocytes by SphCs was much more effective than their primary skin counterparts, which indicates that a single culture step as a spheroid is su cient to acquire such properties. We observed that SphCs signi cantly diminished the secretion of classic proin ammatory cytokines (IFN-g, IL-2, IL-6 and TNF-a) while increasing IL-4 (a Th2-hallmark cytokine) by activated splenocytes.
Comparably, it was reported that MSCs grown as spheroids showed anti-in ammatory capacity and converted macrophages to an M2 phenotype through the expression of TSG-6 or prostaglandin E2 (30, 63, 64). We found that SphC conditioned media promoted cell death inhibition (approximately 50%) in isolated islets when they were challenged by a mixture of proin ammatory cytokines in vitro. Studies have con rmed that growth factors secreted by MSCs, such as VEGF, EGF and HGF (72,73), may conserve islet integrity, growth and functionality. However, the mechanisms by which these factors are protective for b cells under cytokine injury are still unclear. They might act through classic regulators of b cell proliferation, such as Akt and Erk (74). Whether SphCs conditioned media may in uence cell proliferation and/or regeneration of b cells might require prolonged culture periods of islets. Likewise, we did not observe improvements in basal insulin secretion or glucose-stimulated insulin secretion in islets exposed to SphCs conditioned media.
Transplanted islets encounter a rapid attack mediated by allogeneic destructive reactions after infusion (75)(76)(77). For all allogeneic forms of transplantation, antigen-presenting cells (APCs) in ltrating the graft become fully matured by the action of locally released proin ammatory mediators, ischemia and necrotic cells. Matured APCs migrate to secondary lymphoid organs where they stimulate the proliferation of both alloreactive T and recipient T cells recognizing allogeneic MHCs and allopeptides, respectively. These two populations of T cells mount the allo-response against the islets. Thus, the need for lifelong treatment to avoid rejection of allogeneic islets is mandatory and involves di cult management (78). The addition of MSCs at the site of transplantation may help immune tolerance and protect islets, avoiding the administration of high doses of toxic immunosuppressive drugs. We found that cotransplantation of marginal mass allogeneic islets with SphCs improved glycemic control in STZ-induced diabetic mice. We speculate that the success of this procedure is due in part to the local release of pro-survival factors impacting directly on islets, along with creating a favorable anti-in ammatory microenvironment. Added to this last situation is the immune-regulatory activity exerted by SphCs that keep alloreative T lymphocytes under control. All these features are therefore believed to be relevant for the bene cial effects observed in this T1D mouse model (14). The fact that infusing marginal mass allogeneic islets together with SphCs improves glycemia in diabetic mice but does not reach normal glucose levels may be attributable to an insu cient number (10 4 cells) of administered SphCs. Similar experiments employing a mixture of MSCs and islet-derived single cells at a 1:1 ratio transplanted in a mouse model of T1D reached lasting normoglycemia (55). Considering an estimated ~1500 cells in an islet ( 150 m diameter) (79), we cotransplanted SphCs:Islet-cells at a 1:45 ratio. Future experiments, increasing the aforementioned ratio at least 10 times, could improve the results obtained.
As soluble factors secreted by SphC have a favorable impact on the survival and functionality of islets, we hypothesized that proteins secreted differently by cells after aggregation as spheroids could be responsible for the bene cial effects observed on β cell viability.
Therefore, we set out to identify proteins/peptides secreted by them. Protein secretome studies revealed that several biological processes and intracellular pathways were differentially regulated between SphCs and skin broblast-like cell-conditioned media (Figure 7). Within these pathways, we identi ed proteins that could target processes associated with b cell cytoprotection together with modulating immune responses. SphCs regulated approximately 70% of the proteins identi ed in the secretome compared with skin broblast-like cells.
The top hits on the functions displayed by these proteins were related to extracellular matrix remodeling, modulation of apoptosis and cell differentiation. It has been reported that low levels of some matrix metalloproteases might lead to an increased activity of membrane-bound FASL. Consequently, one might expect augmented apoptotic capacity by FASL-expressed SphCs on effector T-cells without affecting the viability of Treg lymphocytes, both contributing to a signi cant increase in b cell viability and islet graft lifespan (80, 81). We found that inhibitors of metalloproteases were upregulated in the SphCs secretome and could, partially, contribute to the less aggressive proin ammatory microenvironment created when these cells were cocultured with murine splenocytes.
The protein 14-3-3ζ, also upregulated in the SphCs secretome, has been implicated in the inhibition of b cell apoptosis induced by different stimuli (82). Both Gremlin-1 and pentatraxin-3, which are abundant in the secretome of SphCs, have been related to growth and survival in response to in ammation and tissue damage (83-86). There is no knowledge about gremlin-1 in the context of b cell biology. In tubular epithelial cells, gremlin-1 increased TFG-b production through Smad activation, inducing a myo broblast-like phenotype (87).
Taken together, these results highlight the underscored ability of SphC to create a favorable niche for allo-islet reception, preserving islet viability and functionality and reducing the need for immune-suppressive drugs.
The possibility of using conditioned media from SphCs or a combination of factors identi ed in the secretome study could provide the exibility of an off-the-shelf product for wider clinical applications. These proteins could be prepared as supplements to include in the cultured media at the time of transplantation and simply comixed with islets for delivery without the need to encapsulate or chemically modify the islets. Further innovations in culture conditions to generate more e cient SphCs as adjuvants of islet engraftment will move the islet transplant eld closer to insulin independence by providing islets with increased cytoprotection and overcoming the requirement of risky chemotherapy for immune suppression.

Conclusion
As islet transplantation for T1D patients remains an ineffective procedure requiring up to 5 different islet infusions associated with loss of functionality and the use of cytotoxic drugs, appropriate employment of cotransplantation with SphCs would be highly advantageous. The use of skin cells to generate SphCs is free of ethical concerns, avoiding the need for genetic cell modi cations associated with considerable technical, ethical and regulatory challenges. Moreover, our results establish evidence of a translatable and potentially more effective immune-modulatory strategy with improved islet viability. Thus, it contributes to tackling the problem of organ donor scarcity.      Cotransplantation of marginal mass allogeneic islets with SphCs improves glycemic control in diabetic mice A.) Establishment of the marginal mass allogenic islets. Both 300 IEQ-transplanted and sham-transplanted diabetic mice showed no difference in blood glucose. Data are presented as the mean ± SEM (n=3 mice per experimental group); *p<0.05 vs. sham and isl-300. B.) 300IEQ recipient diabetic mice did not lose body weight. In contrast, sham-operated mice lost signi cant body weight as of 6 days after the surgical procedure. Data are presented as the mean ± SEM (n=3 mice per experimental group) *p <0.05 vs. isl-300. C.) Cotransplantation of marginal mass allogenic islets with SphCs (1x104) improved glycemia in diabetic mice. The results are shown as plasma glucose concentration in the animals at different time points (left Y axis) and by integrating the area under the curve (AUC) as depicted in the right "Y" axis. Data are presented as the mean ± SEM (n=9 mice per experimental group) *p <0.05 vs. SphCs.

Figure 7
Protein secretome analysis of SphCs compared to that obtained in primary cultures of skin cells A) Biological processes and enriched molecular pathways associated with differentially regulated proteins (q-value <0.05, Bonferroni, ClueGO app) in the secretome of SphCs compared with skin broblast-like cells. The colors red and blue indicate upregulated and downregulated ontologies, respectively. B) Cluster networks of protein-protein interactions corresponding to the main processes described in A were analyzed by the STRING database. A minimum required interaction score of 0.7 was used to build a high con dence interaction network. The line connecting the proteins, depicted as spheres, shows experimental evidence of interaction between them. Each network presents a PPI enrichment value < 1exp-6.