, Volume 60, Issue 10, pp 1987–1998 | Cite as

Proteomic characterisation reveals active Wnt-signalling by human multipotent stromal cells as a key regulator of beta cell survival and proliferation

  • Miljan Kuljanin
  • Gillian I. Bell
  • Stephen E. Sherman
  • Gilles A. LajoieEmail author
  • David A. HessEmail author



Novel strategies to stimulate the expansion of beta cell mass in situ are warranted for diabetes therapy. The aim of this study was to elucidate the secretome of human bone marrow (BM)-derived multipotent stromal cells (MSCs) with documented islet regenerative paracrine function. We hypothesised that regenerative MSCs will secrete a unique combination of protein factors that augment islet regeneration.


Human BM-derived MSCs were examined for glucose-lowering capacity after transplantation into streptozotocin-treated NOD/severe combined immunodeficiency (SCID) mice and segregated into samples with regenerative (MSCR) vs nonregenerative (MSCNR) capacity. Secreted proteins associated with islet regenerative function were identified using stable isotope labelling with amino acids in cell culture (SILAC)-based quantitative proteomics. To functionally validate the importance of active Wnt signalling, we stimulated the Wnt-signalling pathway in MSCNR samples during ex vivo expansion using glycogen synthase kinase 3 (GSK3) inhibition (CHIR99201), and the conditioned culture media (CM) generated was tested for the capacity to support cultured human islet cell survival and proliferation in vitro.


MSCR showed increased secretion of proteins associated with cell growth, matrix remodelling, immunosuppressive and proangiogenic properties. In contrast, MSCNR uniquely secreted proteins known to promote inflammation and negatively regulate angiogenesis. Most notably, MSCR maintained Wnt signalling via Wnt5A/B (~2.5-fold increase) autocrine activity during ex vivo culture, while MSCNR repressed Wnt signalling via Dickkopf-related protein (DKK)1 (~2.5-fold increase) and DKK3 secretion. Inhibition of GSK3 activity in MSCNR samples increased the accumulation of nuclear β-catenin and generated CM that augmented beta cell survival (13% increases) and proliferation when exposed to cultured human islets.


Maintenance of active Wnt signalling within human MSCs promotes the secretion of matricellular and proangiogenic proteins that formulate a niche for islet regeneration.


Angiogenesis Diabetes Inflammation Islets Mass spectrometry Multipotent stromal cells Proteomics Wnt signalling 



7-Aminoacetinomycin D


A disintegrin and metalloproteinase with thrombospondin motifs


Bone marrow


Bone morphogenetic protein


Conditioned culture media


Dickkopf-related protein


Endothelial basal media


Extracellular matrix




Endothelial growth medium


Fibroblast growth factor


Fibroblast growth factor receptor


Vascular endothelial growth factor receptor 1/4






Glycogen synthase kinase 3


Human microvascular endothelial cells


Integrated Islet Distribution Program


Vascular endothelial growth factor receptor 2


Label-free quantification


Matrix metalloproteinase


Multipotent stromal cell


Nonregenerative MSC


Regenerative MSC


Platelet-derived growth factor


Severe combined immunodeficiency


Secreted frizzled-related protein 1


Stable isotope labelling with amino acids in cell culture




Metalloproteinase inhibitor


Vascular endothelial growth factor receptor


Wnt-inducible signalling pathway protein 2


In 2013, WHO reported that over 347 million people worldwide lived with diabetes, and estimated that this number will increase to 552 million by 2030, making diabetes a worldwide epidemic [1]. The Edmonton Protocol provided proof-of-concept that islet transplantation can temporarily reduce insulin dependence in individuals with type 1 diabetes [2]. Although islet replacement is an attractive strategy to combat diabetes, a critical shortage of donor islets and eventual rejection by continuing autoimmunity prevent the widespread application of this approach [3]. Bone marrow (BM)-derived stem cell transplantation has also been reported to promote endogenous islet regeneration in preclinical models, and represents a promising alternative strategy to combat diabetes [4].

Human multipotent stromal cells (MSCs) have become a focal point in recent clinical trials for tissue repair [5]. MSCs are readily available from autologous or allogenic donors, they are efficiently expanded in culture and they home to damaged tissues to initiate innate repair mechanisms [6]. MSCs exert their therapeutic effects primarily through the secretion of trophic signals within damaged organs, or impact tissue repair from distant sites by secretion of regenerative effectors into the circulation within exosomes or microvesicles [7]. Exosomes harvested from MSCs have been shown to contain proangiogenic proteins and promote the healing of ischaemic tissue [8, 9]. MSCs also secrete a wide variety of immunomodulatory molecules that dampen autoimmunity via modulation of immune cell functions [10, 11]. Currently, the mechanisms underlying the regenerative and immunomodulatory effects of MSCs remain poorly understood, and better understanding of MSC protein secretion is required to harness the true regenerative capacity of MSCs.

In the context of diabetes, transplanted human BM-derived MSCs have been shown to promote repair of pancreatic islets and renal glomeruli in NOD/severe combined immunodeficiency (SCID) mice after beta cell ablation [6]. Although MSC conversion into insulin-expressing beta-like cells has been demonstrated after stable induction of PDX1 [12], minimally manipulated MSCs did not adopt a beta cell phenotype after transplantation in vivo. In contrast, transplanted MSCs initiated endogenous islet recovery via paracrine stimulation [13]. In a series of publications, we have shown that human BM-derived MSCs stimulated the emergence of small, recipient-derived islet-like structures associated with the ductal epithelial niche within 7 days of injection into streptozotocin (STZ)-treated hyperglycaemic NOD/SCID mice [14, 15]. Unfortunately, MSC samples showed donor-dependent variability in the capacity to improve glycaemia and prolonged expansion ex vivo reduced islet regenerative prowess [14]. Thus, detailed proteomic analyses of the islet regenerative MSC secretome remain the key to understanding which protein signals promote islet regeneration in situ.

We compared the secretory protein profile of human BM-derived MSCs with or without islet regenerative capacity by performing global proteomic analysis of conditioned culture media (CM) after stable isotope labelling with amino acids in cell culture (SILAC) [16]. We hypothesised that islet regenerative MSCs will secrete a combination of unique protein factors that augment islet regeneration. In order to confirm our findings we also examined the effects of exogenous stimulation of Wnt signalling in MSCs for the survival and proliferation of human islet-derived beta cells in vitro. To our knowledge, this is the first proteomic study linking the human MSC secretome profile with beta cell regenerative function and it highlights the importance of active Wnt signalling in the maintenance of MSC-induced islet regeneration.


MSCR reduced hyperglycaemia after transplantation in vivo

Human BM was obtained from healthy donors after informed consent at the London Health Sciences Centre (London, ON, Canada). All studies were approved by the human ethics committee at Western University (2015–003). The hyperglycaemia-lowering capacity of six MSC samples was assessed after tail vein injection of 500,000 cells into STZ-treated NOD/SCID mice as previously described [14]. Blood glucose concentrations were monitored weekly for 42 days and samples were segregated into regenerative (MSCR) or nonregenerative (MSCNR) based on the ability to reduce blood glucose compared with PBS-injected control mice.

MSC culture and SILAC labelling

MSCs were expanded in AmnioMax media (Invitrogen Carlsbad, CA, USA). At passage 3 MSCs were switched to custom AmnioMax that contained no l-arginine or l-lysine. Heavy [13C6 15N4] l-arginine and [13C6 15N2] l-lysine were added into SILAC media at 87.8 mg/l and 52.2 mg/l, respectively (Silantes Germany). Excess unlabelled l-proline was added to prevent conversion of heavy arginine into heavy proline [17]. MSCs were grown in SILAC media for 9 days to achieve >95% label incorporation.

Generation of labelled CM and proteomic workflows

MSCs were re-plated in AmnioMax without supplement for 24 h to collect proteins secreted by MSCs. CM was collected in biological and technical duplicate, filtered and concentrated using 3 kDa cut-off filter units (Millipore, Bedford, MA, USA). Total protein (150 μg) was subjected to 1D SDS-PAGE fractionation using 12% gels and quantified using the bicinchoninic acid assay (BCA). All MS data was collected using a QExactive (ThermoFisher Scientific, Waltham, MA, USA). Full MS parameters are outlined in the electronic supplementary material (ESM) Table 1. Data analysis was performed using PeaksStudio 7.5 and MaxQuant See ESM Methods for complete details.

HMVEC tubule forming assay

To assess CM influence on endothelial cell function in vitro, 120,000 human microvascular endothelial cells (HMVECs) were cultured on growth factor-reduced Geltrex (Life Technologies, Carlsbad, CA, USA) in endothelial basal media (EBM-2; Lonza, NJ, USA) conditioned by MSCR and MSCNR for 24 h. As a positive control, HMVECs were also grown in Geltrex bathed in complete endothelial growth medium (EGM-2 = EBM-2 + 5% FBS + IGF, basic fibroblast growth factor [FGF], EGF, vascular endothelial growth factor [VEGF]). Tube formation was quantified by counting the number of complete tubule branches in four fields of view using ImageJ software (NIH, Bethesda, MD, USA).

qPCR for β-catenin

Real-time quantitative PCR (qPCR) was performed using SYBR Green along with the human WNT Signalling Pathway PCR Array, which contained 84 genes related to Wnt signal transduction (Qiagen, Hilden, NW, USA) using mRNA isolated from MSCR vs MSCNR. See ESM Methods for complete details.

Confocal microscopy and flow cytometry for total β-catenin

MSCs were cultured in six-well plates on glass cover slips and treated with glycogen synthase kinase 3 (GSK3) inhibitor CHIR99021 (EMD Millipore, Billerica, MA, USA) and stained for nuclear β-catenin. Flow cytometry was used to quantify total cytosolic β-catenin levels after treatment with CHIR99201. See ESM Methods for complete details.

Human islet culture with MSC CM

Human islets from five donors were obtained from the Integrated Islet Distribution Program (IIDP). Upon arrival, 200 islet equivalents were plated in RPMI media without serum (Invitrogen). MSC CM was collected after 24 h from MSCs treated with DMSO (basal CM) or with 10 μmol/l of CHIR99021 (GSK3i CM). CM was concentrated using 3 kDa molecular mass cut-off filters, and ~1.0 μg/μl total protein was added to human islet culture for 1, 3 or 7 days (ESM Fig. 1). After islet harvest and dissociation, beta cell content was estimated using FluoZin-3 (Flz3) (ThermoFisher Scientific) and apoptosis was quantified using 7AAD and Annexin-V. To detect islet cell proliferation, 500 nmol/l of 5-ethynyl-2′-deoxyuridine (EdU) was added to islet culture 24 h prior to harvest. Islets were fixed and permeabilised using 10% formalin and saponin buffer, stained for insulin using a PE-conjugated insulin antibody (R&D Systems, Minneapolis, MN, USA), and nuclear EdU incorporation was detected using the Click-It flow cytometry assay (Life Technologies). Flow cytometry data were analysed using FloJo software (Treestar, Ashland, OR, USA).

Statistical analysis

Statistical analysis was performed using GraphPad Prism version 6.01 (GraphPad, La Jolla, CA, USA) by ANOVA with Tukey’s post hoc test or by multiple t tests. Data are expressed as means (SEM).


Regenerative capacity of human MSCs was donor specific

Hyperglycaemic (15–25 mmol/l), STZ-treated (35 mg/kg per day, days 1–5) NOD/SCID mice were i.v. injected on day 10 with MSCs (5 × 105) [14, 18], and blood glucose levels were monitored for 42 days. Compared with PBS-injected control mice, MSC samples (n = 3) that showed significant reduction in systemic blood glucose from days 17–42 were termed MSCR and MSC samples (n = 3) that did not show blood glucose reduction (>25 mmol/l) were termed MSCNR (Fig. 1a). There was a significant reduction in blood glucose AUC for MSCR vs MSCNR samples (Fig. 1b). Out of a total of 18 MSC lines tested, six (~33%) were characterised as MSCR, four (~22%) demonstrated intermediate regenerative capacity and eight (~44%) were characterised as MSCNR. MSC lines with intermediate regenerative capacity were not used in this study. All MSCR and MSCNR lines possessed multipotent differentiation potential into adipogenic and osteogenic lineages in vitro [14, 15]. Cell surface phenotype showed that both MSCR and MSCNR expressed the stromal markers CD90 and CD105 (>95%) without expression of the pan-leucocyte marker CD45 (data not shown). MSC donor information, including age, sex, weight and BMI, is shown in Fig. 1c.
Fig. 1

Functional characterisation of islet MSCR and MSCNR. (a) MSCs from six donors were injected into STZ-treated (35 mg/kg per day, days 1–5) NOD/SCID mice on day 10, and blood glucose was monitored weekly until 42 days to segregate MSCR (blue) vs MSCNR (red) samples compared with PBS-injected control mice (black). Blood glucose concentrations were lower after injection of MSCR compared with MSCNR samples. (b) Blood glucose AUC was significantly reduced with MSCR (n = 3) (blue) vs MSCNR (n = 3) (red) samples. (c) Donor characteristics were similar between MSCNR and MSCR samples. Data are presented as mean ± SEM. ***p < 0.001

MSCR exclusively secreted proteins associated with active Wnt signalling

We have previously shown that MSCR demonstrated increased transcription of matrix metalloproteases, EGF-family ligands and downstream products of Wnt signalling [15]. Here, we sought to confirm and identify secreted protein targets that contribute to islet regeneration by comparing the secretome of MSCR and MSCNR samples using global MS-based proteomics coupled with SILAC. CM was harvested from each MSC line (24 h) and analysed in biological and technical duplicates. The total number of proteins detected for each MSC sample is shown in Fig. 2a. Proteins from each MSC line were combined based on regenerative capacity and complete lists of secreted proteins detected exclusively in CM from MSCR or MSCNR, or proteins common to both groups, were generated (ESM Tables 24). Of 4665 total proteins detected (false discovery rate 1%), 3023 (~65%) were produced by both MSCR and MSCNR, 850 (~18%) were produced by MSCR exclusively, while 792 (~17%) were produced by MSCNR exclusively (Fig. 2b). Gene ontology using cellular component analysis revealed that MSCR CM contained qualitatively more proteins with extracellular localisation (21% vs 14%), while MSCNR CM contained more membrane-bound proteins (23% vs 19%) (Fig. 2c). Moving forward, protein lists were filtered to include only extracellular (secreted) and membrane-bound proteins, and were further analysed for known biological functions and signalling family association.
Fig. 2

Qualitative analyses of proteins exclusively secreted by MSCR or MSCNR. (a) Total proteins identified for MSCR (blue) and MSCNR (red) lines using SILAC. (b) Venn diagram showing that out of the 4665 total proteins detected, 850 were unique to MSCR (blue) and 792 were unique to MSCNR (red). (c) MSCR demonstrated a higher proportion of proteins with extracellular (secreted) localisation compared with MSCNR. (d) MSCR secreted more proteins associated with the activation of Wnt signalling and angiogenesis. MSCNR secreted more proteins associated with EGF, FGF and Notch signalling

MSCR demonstrated increased representation of secreted proteins associated with angiogenesis and activation of the Wnt-signalling pathway, while MSCNR showed increased secretion of proteins associated with EGF, FGF and Notch signalling pathways (Fig. 2d). MSCR exclusively secreted several proteins associated with cell survival and growth, such as FGF7 and bone morphogenetic protein (BMP)4. Proangiogenic factors vascular endothelial growth factor receptor 2 (KDR), vascular endothelial growth factor receptor 4 (FLT4) and regulators of angiogenesis A disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS)13 and 18 were also unique to the MSCR secretome [19]. Notably, MSCR secreted potentiators of Wnt signalling (spondins), ligand activators of Wnt signalling (WNT5B) and downstream products of Wnt signalling that modify the extracellular matrix (ECM) (Wnt-inducible signalling pathway protein 2 [WISP2/CCN5]) (Table 1) [20, 21, 22]. In contrast, proteins exclusively detected in MSCNR CM included positive and negative regulators of angiogenesis (FGF receptor (FGFR)4, FLT1) [19, 23], and an abundance of proinflammatory cytokines (IL-6, IL-8) and C-X-C motif chemokines (CXCL2, 3, 5) (Table 2) [24]. Collectively, these findings suggested MSCR actively secreted cell growth supportive factors with reduced secretion of common proinflammatory signals, and only MSCR showed production of multiple effectors associated with active Wnt signalling.
Table 1

Exclusively detected proteins in CM from MSCR involved in angiogenesis, inflammation, matrix remodelling and Wnt signalling

Gene symbol

Protein name

Biological function


A disintegrin and metalloproteinase with thrombospondin motifs 13

Matrix metalloproteinase


A disintegrin and metalloproteinase with thrombospondin motifs 18

Matrix metalloproteinase


Epidermal growth factor

Regulator of cell growth, proliferation and differentiation


Fibroblast growth factor 7

Tissue repair, mitogenic cell survival


Vascular endothelial growth factor receptor 3

Development of vascular networks via secretion of VEFGA/C


Insulin-like growth factor 2

Growth promoting hormone


Mast/stem cell growth factor receptor Kit

Stem cell maintenance, migration, survival and proliferation


Vascular endothelial growth factor receptor 2

Regulation of angiogenesis and vascular development



Matrix metalloproteinase


Macrophage metalloelastase

Matrix metalloproteinase



Matrix metalloproteinase


Platelet-derived growth factor D

Regulation of cell proliferation, migration and chemotaxis



Activator of Wnt signalling


Metalloproteinase inhibitor 4

Matrix metalloproteinase inhibitor


Wnt1-inducible-signalling pathway protein 2

CCN5, regulator of cell growth


Protein Wnt5b

Activator of Wnt signalling

Table 2

Exclusively detected proteins in CM from MSCNR involved in angiogenesis, inflammation, matrix remodelling and Wnt signalling

Gene symbol

Protein name

Biological function


C-X-C motif chemokine 2

Inflammatory response


C-X-C motif chemokine 3

Inflammatory response


C-X-C motif chemokine 5

Inflammatory response


Growth/differentiation factor 15

Regulation of inflammation and apoptosis


Vascular endothelial growth factor receptor 1

Negative regulator of VEGFA signalling


Fibroblast growth factor receptor 4

Regulation of angiogenesis and vascular development



Wnt signalling receptor



Wnt signalling receptor



Wnt signalling receptor


Interleukin-1 beta

Inflammatory response



Inflammatory response



Inflammatory response



Matrix metalloproteinase


Protein NOV homologue

CCN3, promoter of cell differentiation


TGF-beta receptor type-2

Regulator of cell growth, proliferation and differentiation

Quantification of Wnt, matrix remodelling and proangiogenic proteins

To quantify proteins secreted by both MSCR and MSCNR, quantitative proteomics was performed using label-free quantification [16, 25] to identify proteins differentially secreted into serum-containing media. Lists were filtered to include only secreted and membrane-bound proteins, and 1038 common proteins were quantified to generate lists of differentially expressed proteins (ESM Table 5). We did not observe any significant difference in the total number of proteins quantified between samples (Fig. 3a), and 468 proteins were differentially expressed between MSCR vs MSCNR samples (p < 0.05, Fig. 3b). Upregulated proteins in MSCR CM included matrix remodelling proteins (matrix metalloproteinase (MMP)1, MMP3, ADAMs and BMPs) [26, 27], effectors of Wnt signalling (WNT5A, secreted frizzled-related protein 1 [SFRP1]) [21], additional targets implicated in angiogenesis (angiogenin [ANG], angiopoietin [ANGPT]1, ANGPTL2, TGFB1, TGFB2, platelet-derived growth factor receptor [PDGFR]A, B) [23, 28] and chemokines involved in cell recruitment (CXCL12) (Fig. 3c) [28, 29]. Thus, the activation of Wnt and proangiogenic signalling again correlated with islet regenerative potency. Conversely, the negative regulator of Wnt signalling Dickkopf-related protein (DKK)1 was increased ~2.5-fold in MSCNR CM, and negative regulators of angiogenesis (ADAMTSs, thrombospondin [THBS]1 and THBS2) and inhibitors of matrix remodelling proteins (metalloproteinase inhibitor [TIMP]3) were increased in MSCNR CM [30]. HMVECs were cultured on growth factor-reduced Geltrex to assess the angiogenic potential of MSC CM in vitro (Fig. 3e). A significant increase in the tubule formation was observed when HMVECs were cultured using CM from MSCR compared with MSCNR. Representative photomicrographs of tube formation after 24 h are shown in ESM Fig. 2. Finally, we identified that pro-differentiation proteins associated with Notch signalling were upregulated in MSCNR (NOTCH2, JAG-1) [31], suggesting that inhibited Wnt and angiogenic signalling and activated Notch signalling was associated with diminished islet regenerative capacity.
Fig. 3

Quantitative analyses of proteins secreted by both MSCR and MSCNR. (a) Total number of proteins quantified within all three MSCR (blue), MSCNR (red) samples and total number of quantified proteins (grey). (b) Representative volcano plot of differentially expressed secreted proteins. A change greater than twofold is represented outside the green boundaries. Red, p < 0.01; blue, p < 0.05; grey, p > 0.05. (c) MSCR showed increased secretion of Wnt activators (Wnt5A), while MSCNR showed increased secretion of Wnt inhibitors (DKK1). (d) MSCR demonstrated increased secretion of proangiogenic proteins, TGFB1 and SDF-1 (CXCL12). (e) Spontaneous tubule formation of HMVECs on growth factor-reduced Geltrex was increased when cultured using CM generated from MSCR compared with MSCNR. Data are presented as mean ± SD. **p < 0.01

MSCR increased expression of Wnt pathway mRNA

To confirm the activation of Wnt signalling in MSCR, we assessed mRNA expression of 84 genes related to Wnt pathway signal transduction, using three MSCR and three MSCNR lines performed in duplicate. Transcripts with differential expression greater than twofold between MSCR vs MSCNR revealed 18 significantly changing genes (ESM Fig. 3a,b), normalised to the geometric mean of three housekeeping genes: ACTB, B2M and HPRT1, that did not show significant variation across samples. Notably, all differentially expressed mRNAs (18 genes) were upregulated in MSCR, and no significantly changing genes were downregulated. Upregulation of WNT5A/B in MSCR were confirmed at the mRNA level. Other upregulated mRNA in MSCR included transcription factors (FOSL1, JUN) and receptors associated with Wnt signalling (FZD5) [20, 32]. Taken together, upregulation of Wnt signalling at the mRNA and protein level strongly suggests active Wnt signalling during expansion was a unique and conserved characteristic that correlated with the reduction of hyperglycaemia after transplantation of MSCR in vivo.

MSCs show activation of Wnt signalling via accumulation of nuclear β-catenin

Canonical Wnt signalling converges on the actions of β-catenin, a transcription factor that increases the expression of downstream effectors of Wnt signals [33]. We mimicked Wnt signalling in MSCNR and MSCR using a small molecule (CHIR99201) inhibitor of GSK3. GSK3 actively phosphorylates β-catenin, which marks it for ubiquitination and degradation. Therefore, inhibition of GSK3 leads to stabilisation and accumulation of free β-catenin [20]. MSC samples showed maximal increases in total β-catenin using 5–10 μmol/l CHIR99201 for 24–48 h. We then used confocal microscopy to visualise the accumulation of nuclear β-catenin (Fig. 4). Compared with DMSO control cells (Fig. 4a), MSCNR and MSCR stimulated with 10 μmol/l CHIR99201 showed increased nuclear β-catenin localisation (Fig. 4b, c). Next, MSCNR and MSCR samples were analysed in quadruplicate to quantify total β-catenin levels by flow cytometry (Fig. 4d). Representative dot plots are shown in ESM Fig. 4. Compared with DMSO control cells, both MSC subtypes treated with CHIR99201 showed significantly increased total β-catenin. Thus, inhibition of GSK3 with CHIR99201 mimicked activated Wnt signalling in MSCR and MSCNR.
Fig. 4

MSC treatment with GSK3 inhibitor induced nuclear β-catenin localisation. Representative photomicrographs showing MSCR and MSCNR stained for β-catenin (green) and DAPI (blue) after treatment with (a) DMSO vehicle, (b) CHIR99201 (24 h) MSCNR, (c) CHIR99201 (24 h) MSCR. White arrows indicate examined nuclei with nuclear staining and arrowheads indicate examined nuclei without nuclear staining. Scale bars, 50 μm. (d) Intracellular β-catenin was increased in both MSCR and MSCNR treated with CHIR99201 (grey bars) or DMSO control (white bars). Data are presented as mean ± SEM. **p < 0.01, ***p < 0.001

GSK3 inhibition in MSCNR generates CM that improves human beta cell survival in vitro

To assess whether MSCR CM could improve beta cell survival in vitro, we cultured primary human islet preparations for up to 7 days in CM from MSCR and MSCNR, and performed multiparametric flow cytometry to analyse beta cell survival and proliferation. Human islets were obtained through the IIDP. Compared with MSCNR CM, human islets grown in MSCR CM showed increased total number of beta cells (Fig. 5a) and an increased frequency of live beta cells (Fig. 5b) after 7 days of culture. To further assess the influence of active Wnt signalling on islet regenerative paracrine function, we also assessed whether CM generated by MSCNR treated with a GSK3 inhibitor (CHIR99201) during expansion could promote human beta cell survival or proliferation in cultured human islets. CM recovered from MSCNR cultured with vehicle control (DMSO) only (basal CM) was compared with CM recovered from MSCNR treated with 10 μmol/l CHIR99201 (GSK3i CM) (Fig. 5c). Five independent human islet samples were cultured as indicated in ESM Fig. 1. At each time point, human islets were harvested, dissociated with trypsin, and stained using Flz3 to estimate beta cell frequency, 7AAD to determine cell viability and Annexin-V to determine apoptosis rates by multiparametric flow cytometry. Representative dot plots are shown in ESM Fig. 5. Compared with basal CM, human islets grown in CM generated by GSK3-inhibited MSCNR showed increased total cell numbers (Fig. 5d), beta cell numbers (Fig. 5e) and the proportion of live (7AAD−) beta cells at 7 days culture (Fig. 5f). However, no significant changes in the frequency of apoptotic beta cells were observed (Fig. 5g).
Fig. 5

Culture of human islets in GSK-inhibited CM increased beta cell number. Human islets cultured in CM generated from MSCR showed (a) increased beta cell number and (b) increased proportion of live beta cells after 7 days of culture. (c) Human islets were cultured in CM generated by MSCNR treated with DMSO (basal CM) or CHIR99201 (GSK3i CM). Compared with human islets cultured in basal CM (red), islets cultured in GSK3-inhibited CM (blue) showed increased (d) cell number, (e) beta cell number and (f) proportion of live beta cells at day 7. (g) Human islets cultured in MSCNR CM did not alter apoptotic beta cell frequency. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001

To determine whether increased beta cell number was augmented by increased beta cell proliferation in vitro, we measured EdU incorporation in insulin + beta cells. At each time point, human islets were harvested, permeabilised, stained for intracellular insulin and proliferation was detected by EdU incorporation using the Click-It system. Representative dot plots showing EdU+ cell detection for each condition are shown in ESM Fig. 6. Compared with basal CM, islet exposure to CM generated by GSK3-inhibited MSCNR increased the overall frequency of proliferating cells (Fig. 6a) and proportion of total proliferating beta cells (Fig. 6b). Finally, islet donor variability was documented by donor information including age, sex, BMI and average blood glucose levels (Fig. 6c). Collectively, these data suggest that CM generated from MSCNR treated with CHIR99201 promoted beta cell survival and induced beta cell proliferation in vitro.
Fig. 6

Culture of human islets in GSK-inhibited CM increased beta cell proliferation. (a, b) Compared with human islets cultured in basal CM (red), islets cultured in GSK3i CM (blue) showed increased proportion of insulin+/EdU+ beta cells at day 3. (c) Islets were obtained through the IIDP from five independent donors with variable weight, BMI and average blood glucose (Avg.BG) values. Data are presented as mean ± SEM. *p < 0.05


We used comprehensive SILAC-based proteomic analyses to identify MSC-secreted factors that correlated with the capacity to lower circulating blood glucose after transplantation into STZ-treated NOD/SCID mice. MSCR, demonstrated glucose-lowering capacity after transplantation, secreted proteins associated with cell growth, matrix remodelling, immunosuppressive and proangiogenic properties. In contrast, MSCNR, which lacked islet regenerative functions, secreted proteins involved in the initiation of inflammation and the negative regulation of angiogenesis. Notably, MSCR consistently demonstrated mRNA and protein expression associated with active Wnt signalling, a novel signature that correlated with islet regenerative capacity. Inhibition of GSK3 activity with CHIR99201 mimicked Wnt signalling in MSCNR and resulted in the generation of CM that supported beta cell survival and proliferation within cultured human islets in vitro. Thus, we report a central role for Wnt signalling in the establishment of an islet regenerative secretory profile in MSCR, and improve our understanding of MSC-secreted signals governing islet regeneration. We predict this dataset will aid in the development of future therapies to augment islet regeneration during diabetes.

Functional beta cells secrete proangiogenic proteins responsible for recruiting circulating or tissue resident progenitor cells to islets after damage [34]. However, during autoimmunity or after islet isolation and transplantation these critical functions are likely to be dysregulated, resulting in transient ischaemia that significantly impairs islet function [35]. Increased production of potent proangiogenic factors, such as FGF7, PDGF and VEGFA, by MSCR underscore the functional capacity to generate a proangiogenic microenvironment. Collectively these proteins potently stimulate human islet vascularisation in vitro [36]. To validate the functional relevance of proangiogenic secretory patterns, MSCR CM showed increased capacity to induce spontaneous tubule formation by HMVEC compared with MSCNR CM. Interestingly, MSCNR primarily secreted negative regulators of angiogenesis such as FLT1 and FGFR4, which have both been shown to inhibit the signal cascade mediated by VEGFA [19]. We have previously shown that intrapancreatic transplantation of proangiogenic haematopoietic progenitor cells induced islet revascularisation and beta cell proliferation and augmented systemic insulin release in STZ-treated NOD/SCID mice [15]. Therefore, our data suggest MSCR may also formulate a proangiogenic microenvironment to support beta cell survival and function within endogenous or transplanted islets.

The prevention of proinflammatory states within damaged islets may also be relevant in the context of MSC therapy for type 1 diabetes. The presence of central mediators of inflammation in CM from MSCNR, such as IL-1β, IL-6 and IL-8, suggests that MSCNR may contribute to a proinflammatory cascades. In people with type 1 diabetes, beta cell destruction is initiated by a combination of proinflammatory cytokines, including IL-1β and IL-6 [37], and CXCL family chemokines, leading to recruitment of immune effectors that mediate beta cell destruction [38]. Furthermore, high levels of IL-8 have been linked to elevated NEFA, which can signal inflammatory cascades in the pancreas [39]. In contrast, these cytokines and chemokines were not secreted by MSCR. Rather, MSCR secreted cytokines that could potentially reduce inflammation. For example, TGF-β suppresses the secretion of various inflammatory cytokines/chemokines, and induces cytokine secretion patterns that balance local immunity [40]. Stromal cell-derived factor 1 (CXCL12), also upregulated by MSCR, has been shown to directly promote beta cell survival through the activation of AKT [41]. Collectively, we propose MSCR generate a niche with reduced inflammation, improving beta cell survival, while MSCNR contribute to a more deleterious proinflammatory microenvironment.

Inadequate beta cell mass leads to hyperglycaemia in both type 1 and type 2 diabetes. Currently, there is significant interest in restoration of beta cell mass through induction of endogenous regenerative mechanisms in situ using cellular or protein therapies. Our previous transplantation studies suggested that human BM-derived MSCs stimulate an islet regeneration programme with neogenic characteristics. Mice transplanted with MSCR, demonstrated improved glycaemic control via the emergence of small beta cell clusters associated with the ductal epithelium [14]. Several identified effectors may act in synergy to mediate islet regenerative processes. First, MSCR showed increased secretion of EGF previously shown to increase beta cell mass in rodents by stimulating beta cell replication [42]. However, EGF receptor signalling was required for expansion of murine beta cell mass in response to a high-fat diet, but was not crucial for neoislet formation after pancreatic ductal ligation [43]. Second, the CCN family of extracellular matrix-associated, heparin-binding proteins has been shown to modulate cell growth and repair in many tissues by increasing the bioavailability of BMPs, VEGF, Wnt and TGF ligands [44]. More specifically, connective tissue growth factor (CTGF) or CCN2 has been widely studied during beta cell development [45]. Within our dataset, MSCR exclusively secreted CCN5). Previously, we have identified WISP1 (encoding CCN4) and WISP2 (encoding CCN5) mRNA as being upregulated in MSCR [15]. Importantly, CCN4 and CCN5 secretion is directly linked to active Wnt signalling. CCN5 also represents a link between insulin and IGF-I regulation in islet function, and CCN5 overexpression leads to increased islet cell growth in vitro [46]. Interestingly, MSCNR exclusively secreted nephroblastoma overexpressed protein or NOV isoform (CCN3), which has been shown to impair beta cell proliferation and inhibit glucose-stimulated insulin secretion in vitro [45]. Thus, we predict modification of the ECM via Wnt-inducible CCN protein regulation is a critical step in islet regenerative processes.

One of the key differences between MSCR and MSCNR was the activation of Wnt signalling. Wnt signalling is highly conserved in primitive MSCs and is involved in multiple developmental processes, including cell proliferation, growth and fate determination. Aly and colleagues have shown that delivery of WNT3A and R-spondin to cultured islets increased the proliferation of adult human beta cells [47]. Our data suggests that MSCR propagate Wnt signalling primarily by autocrine secretion of WNT5A/B ligands with potential regulation by SFRP1. In contrast, MSCNR expressed proteins that inhibit Wnt signalling, namely DKK1 and DKK3. Although more research is required to determine potential effects of Wnt5 signalling on human beta cells, it is clear that MSCR maintain active canonical Wnt signalling and increase secretion of downstream effectors that promote beta cell regeneration. By inhibiting GSK3 activity using CHIR99201, we have shown that the activation of canonical Wnt pathway signals in MSCNR can mimic Wnt signalling allowing accumulation of nuclear β-catenin. Human pancreatic islets grown in media conditioned by MSCNR treated with CHIR99201 showed increased beta cell survival and proliferation during 7 days of culture. Manipulation of Wnt signalling through the inhibition of GSK3 has been previously suggested to directly increase beta cell proliferation. One study treated diabetic neonates with LiCl, a known inhibitor of GSK3, and doubled beta cell mass in rat models of diabetes [48]. Others have inhibited GSK3 using small molecules to regulate islet cell survival and proliferation in vitro [49]. Although, the mechanisms by which inhibition of GSK3 affects beta cell survival and proliferation in vivo is still unclear, our results suggest that activation of Wnt signalling in pancreas-resident stromal cells may have practical application in beta cell regenerative therapies. Nonetheless, activation of Wnt signalling during MSC culture increased the regenerative capacity of MSCNR by altering downstream secretory patterns.

In summary, our data outlines several dynamic and complementary pathways that formulate a regenerative microenvironment applicable to the development of islet expansion therapies for diabetes using MSCs or their secretory products. The proteomic data reported in this study will be used in future studies to characterise functional mechanisms relevant to islet regeneration, or as screening technology to select MSC subpopulations that possess augmented capacity to regenerate islets in situ.



We wish to acknowledge P. Pittock (Don Rix Protein Identification Facility, Department of Biochemistry, Western University) for troubleshooting, maintenance and care of the MS instrumentation and for insightful input on method optimisation.

Data availability

Proteomic raw data files that support the findings of this study are available from the corresponding author upon reasonable request.


This study was supported in part by a Natural Sciences and Engineering Research Council of Canada (NSERC, R-3095A03) grant to GAL, and a JDRF USA Strategic Research Agreement on Optimizing Beta Cell Regeneration (JDRF, 2-SRA-2015-60-Q-R) grant to DAH.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Contribution statement

MK performed experiments, presented and analysed the data, and wrote and critically revised the manuscript and approved the final version. GIB performed in vivo characterisation experiments, collected and assembled data, and provided critical revision of the manuscript and final approval. SES performed confocal imaging, collected and assembled data, and provided critical revision and final approval of the manuscript. GAL and DAH designed the study, provided financial support and data interpretation, wrote and reviewed the manuscript and provided final approval. DAH is responsible for the integrity of the work as a whole.

Supplementary material

125_2017_4355_MOESM1_ESM.pdf (716 kb)
ESM (PDF 715 kb)
125_2017_4355_MOESM2_ESM.xlsx (156 kb)
ESM Tables (XLSX 156 kb)


  1. 1.
    Olokoba AB, Obateru OA, Olokoba LB (2012) Type 2 diabetes mellitus: a review of current trends. Oman Med J 27:269–273CrossRefPubMedPubMedCentralGoogle Scholar
  2. 2.
    Ryan EA, Paty BW, Senior PA et al (2005) Five-year follow-up after clinical islet transplantation. Diabetes 54:2060–2069CrossRefPubMedGoogle Scholar
  3. 3.
    Shapiro AMJ, Ryan EA, Lakey JRT (2001) Islet cell transplantation. Lancet 358:S214CrossRefGoogle Scholar
  4. 4.
    Qi C, Yan X, Huang C et al (2015) Biomaterials as carrier, barrier and reactor for cell-based regenerative medicine. Protein Cell 6:638–653CrossRefPubMedPubMedCentralGoogle Scholar
  5. 5.
    Pomeransky AA, Khriplovich IB (1999) Equations of motion of spinning relativistic particle in external fields. Surv High Energy Phys 14:145–173CrossRefGoogle Scholar
  6. 6.
    Lee RH, Seo MJ, Reger RL et al (2006) Multipotent stromal cells from human marrow home to and promote repair of pancreatic islets and renal glomeruli in diabetic NOD/scid mice. Proc Natl Acad Sci 103:17438–17443CrossRefPubMedPubMedCentralGoogle Scholar
  7. 7.
    Lai RC, Arslan F, Lee MM et al (2010) Exosome secreted by MSC reduces myocardial ischemia/reperfusion injury. Stem Cell Res 4:214–222CrossRefPubMedGoogle Scholar
  8. 8.
    Anderson JD, Johansson HJ, Graham CS et al (2016) Comprehensive proteomic analysis of mesenchymal stem cell exosomes reveals modulation of angiogenesis via nuclear factor-Kappab signaling. Stem Cells 34:601–613CrossRefPubMedGoogle Scholar
  9. 9.
    Meyerrose T, Olson S, Pontow S et al (2010) Mesenchymal stem cells for the sustained in vivo delivery of bioactive factors. Adv Drug Deliv Rev 62:1167–1174CrossRefPubMedGoogle Scholar
  10. 10.
    Madec AM, Mallone R, Afonso G et al (2009) Mesenchymal stem cells protect NOD mice from diabetes by inducing regulatory T cells. Diabetologia 52:1391–1399CrossRefPubMedGoogle Scholar
  11. 11.
    Berman DM, Willman MA, Han D et al (2010) Mesenchymal stem cells enhance allogeneic islet engraftment in nonhuman primates. Diabetes 59:2558–2568CrossRefPubMedPubMedCentralGoogle Scholar
  12. 12.
    Karnieli O, Izhar-Prato Y, Bulvik S, Efrat S (2007) Generation of insulin-producing cells from human bone marrow mesenchymal stem cells by genetic manipulation. Stem Cells 25:2837–2844CrossRefPubMedGoogle Scholar
  13. 13.
    Urbán VS, Kiss J, Kovács J et al (2008) Mesenchymal stem cells cooperate with bone marrow cells in therapy of diabetes. Stem Cells 26:244–253CrossRefPubMedGoogle Scholar
  14. 14.
    Bell GI, Broughton HC, Levac KD et al (2012a) Transplanted human bone marrow progenitor subtypes stimulate endogenous islet regeneration and revascularization. Stem Cells Dev 21:97–109CrossRefPubMedGoogle Scholar
  15. 15.
    Bell GI, Meschino MT, Hughes-Large JM et al (2012b) Combinatorial human progenitor cell transplantation optimizes islet regeneration through secretion of paracrine factors. Stem Cells Dev 21:1863–1876CrossRefPubMedGoogle Scholar
  16. 16.
    Ong S-E (2002) Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol Cell Proteomics 1:376–386CrossRefPubMedGoogle Scholar
  17. 17.
    Bendall SC, Hughes C, Stewart MH et al (2008) Prevention of amino acid conversion in SILAC experiments with embryonic stem cells. Mol Cell Proteomics 7:1587–1597CrossRefPubMedPubMedCentralGoogle Scholar
  18. 18.
    Hess D, Li L, Martin M et al (2003) Bone marrow-derived stem cells initiate pancreatic regeneration. Nat Biotechnol 21:763–770CrossRefPubMedGoogle Scholar
  19. 19.
    Shibuya M (2011) Vascular endothelial growth factor (VEGF) and its receptor (VEGFR) signaling in angiogenesis: a crucial target for anti- and pro-angiogenic therapies. Genes Cancer 2:1097–1105CrossRefPubMedPubMedCentralGoogle Scholar
  20. 20.
    Gordon MD, Nusse R (2006) Wnt signaling: multiple pathways, multiple receptors, and multiple transcription factors. J Biol Chem 281:22429–22433CrossRefPubMedGoogle Scholar
  21. 21.
    Yang Y (2003) Wnt5a and Wnt5b exhibit distinct activities in coordinating chondrocyte proliferation and differentiation. Development 130:1003–1015CrossRefPubMedGoogle Scholar
  22. 22.
    Bakondi B, Shimada IS, Peterson BM, Spees JL (2011) SDF-1α secreted by human CD133-derived multipotent stromal cells promotes neural progenitor cell survival through CXCR7. Stem Cells Dev 20:1021–1029CrossRefPubMedGoogle Scholar
  23. 23.
    Huang Z, Bao SD (2004) Roles of main pro- and anti-angiogenic factors in tumor angiogenesis. World J Gastroenterol 10:463–470CrossRefPubMedPubMedCentralGoogle Scholar
  24. 24.
    Aziz M, Jacob A, Yang W-L et al (2013) Current trends in inflammatory and immunomodulatory mediators in sepsis. J Leukoc Biol 93:329–342CrossRefPubMedPubMedCentralGoogle Scholar
  25. 25.
    Cox J, Hein MY, Luber CA, Paron I (2014) Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ. Mol Cell 13:2513–2526Google Scholar
  26. 26.
    Rohani MG, Parks WC (2015) Matrix remodeling by MMPs during wound repair. Matrix Biol 44–46:113–121CrossRefPubMedGoogle Scholar
  27. 27.
    Vadon-Le Goff S, Hulmes DJ, Moali C (2015) BMP-1/tolloid-like proteinases synchronize matrix assembly with growth factor activation to promote morphogenesis and tissue remodeling. Matrix Biol 44–46:14–23CrossRefPubMedGoogle Scholar
  28. 28.
    Boomsma RA, Geenen DL (2012) Mesenchymal stem cells secrete multiple cytokines that promote angiogenesis and have contrasting effects on chemotaxis and apoptosis. PLoS One 7:e35685CrossRefPubMedPubMedCentralGoogle Scholar
  29. 29.
    Ho TK, Tsui J, Xu S et al (2010) Angiogenic effects of stromal cell-derived factor-1 (SDF-1/CXCL12) variants in vitro and the in vivo expressions of CXCL12 variants and CXCR4 in human critical leg ischemia. J Vasc Surg 51:689–699CrossRefPubMedGoogle Scholar
  30. 30.
    Pepper MS (1996) Positive and negative regulation of angiogenesis: from cell biology to the clinic. Vasc Med 1:259–266CrossRefPubMedGoogle Scholar
  31. 31.
    Blanpain C, Lowry WE, Pasolli HA, Fuchs E (2006) Canonical notch signaling functions as a commitment switch in the epidermal lineage. Genes Dev 20:3022–3035CrossRefPubMedPubMedCentralGoogle Scholar
  32. 32.
    Yokoyama N, Yin D, Malbon CC (2007) Abundance, complexation, and trafficking of Wnt/β-catenin signaling elements in response to Wnt3a. J Mol Signal 2:11CrossRefPubMedPubMedCentralGoogle Scholar
  33. 33.
    Meijer L, Flajolet M, Greengard P (2004) Pharmacological inhibitors of glycogen synthase kinase 3. Trends Pharmacol Sci 25:471–480CrossRefPubMedGoogle Scholar
  34. 34.
    Jin DK, Shido K, Kopp H et al (2006) Cytokine-mediated deployment of SDF-1 induces revascularization through recruitment of CXCR4+ hemangiocytes. Nat Med 12:557–567CrossRefPubMedPubMedCentralGoogle Scholar
  35. 35.
    Brissova M, Shostak A, Shiota M et al (2006) Pancreatic islet production of vascular endothelial growth factor-A is essential for islet vascularization, revascularization, and function. Diabetes 55:2974–2985CrossRefPubMedGoogle Scholar
  36. 36.
    Luo JZQ, Xiong F, Al-Homsi AS et al (2011) Human BM stem cells initiate angiogenesis in human islets in vitro. Bone Marrow Transplant 46:1128–1137CrossRefPubMedGoogle Scholar
  37. 37.
    Blandino-Rosano M, Perez-Arana G, Mellado-Gil JM et al (2008) Anti-proliferative effect of pro-inflammatory cytokines in cultured beta cells is associated with extracellular signal-regulated kinase 1/2 pathway inhibition: protective role of glucagon-like peptide-1. J Mol Endocrinol 41:35–44CrossRefPubMedGoogle Scholar
  38. 38.
    Eizirik DL, Colli ML, Ortis F (2009) The role of inflammation in insulitis and beta-cell loss in type 1 diabetes. Nat Rev Endocrinol 5:219–226CrossRefPubMedGoogle Scholar
  39. 39.
    Shi H, Kokoeva MV, Inouye K et al (2006) TLR4 links innate immunity and fatty acid – induced insulin resistance. J Clin Invest 116:3015–3025CrossRefPubMedPubMedCentralGoogle Scholar
  40. 40.
    Sanjabi S, Zenewicz LA, Kamanaka M, Flavell RA (2009) Anti-inflammatory and pro-inflammatory roles of TGF-β, IL-10, and IL-22 in immunity and autoimmunity. Curr Opin Pharmacol 9:447–453CrossRefPubMedPubMedCentralGoogle Scholar
  41. 41.
    Yano T, Liu Z, Donovan J et al (2007) Stromal cell-derived factor-1 (SDF-1)/CXCL12 attenuates diabetes in mice and promotes pancreatic beta-cell survival by activation of the prosurvival kinase Akt. Diabetes 56:2946–2957CrossRefPubMedGoogle Scholar
  42. 42.
    Rooman I, Bouwens L (2004) Combined gastrin and epidermal growth factor treatment induces islet regeneration and restores normoglycaemia in C57BI6/J mice treated with alloxan. Diabetologia 47:259–265CrossRefPubMedGoogle Scholar
  43. 43.
    Hakonen E, Ustinov J, Mathijs I et al (2011) Epidermal growth factor (EGF)-receptor signalling is needed for murine beta cell mass expansion in response to high-fat diet and pregnancy but not after pancreatic duct ligation. Diabetologia 54:1735–1743CrossRefPubMedGoogle Scholar
  44. 44.
    Zuo GW, Kohls CD, He BC et al (2010) The CCN proteins: important signaling mediators in stem cell differentiation and tumorigenesis. Histol Histopathol 25:795–806PubMedPubMedCentralGoogle Scholar
  45. 45.
    Paradis R, Lazar N, Antinozzi P et al (2013) Nov/Ccn3, a novel transcriptional target of FoxO1, impairs pancreatic β-cell function. PLoS One 8:1–8CrossRefGoogle Scholar
  46. 46.
    Chowdhury S, Wang X, Srikant CB et al (2014) IGF-I stimulates CCN5/WISP2 gene expression in pancreatic β-cells, which promotes cell proliferation and survival against streptozotocin. Endocrinology 155:1629–1642CrossRefPubMedGoogle Scholar
  47. 47.
    Aly H, Rohatgi N, Marshall CA et al (2013) A novel strategy to increase the proliferative potential of adult human β-cells while maintaining their differentiated phenotype. PLoS One 8:e66131CrossRefPubMedPubMedCentralGoogle Scholar
  48. 48.
    Figeac F, Uzan B, Faro M et al (2010) Neonatal growth and regeneration of -cells are regulated by the Wnt/ -catenin signaling in normal and diabetic rats. AJP Endocrinol Metab 298:E245–E256CrossRefGoogle Scholar
  49. 49.
    Mussmann R, Geese M, Harder F et al (2007) Inhibition of GSK3 promotes replication and survival of pancreatic beta cells. J Biol Chem 282:12030–12037CrossRefPubMedGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  1. 1.Don Rix Protein Identification Facility, Department of Biochemistry, Schulich School of Medicine and DentistryWestern UniversityLondonCanada
  2. 2.Department of Physiology and Pharmacology, Schulich School of Medicine and DentistryWestern UniversityLondonCanada
  3. 3.Molecular Medicine Laboratories, Krembil Centre for Stem Cell BiologyRobarts Research InstituteLondonCanada

Personalised recommendations