Bone Marrow Mesenchymal Stem Cells Promote the Repair of Islets From

Molecular and Cellular Endocrinology 388 (2014) 41–50

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology

journal homepage: www.elsevier.com/locate/mce

Bone marrow mesenchymal stem cells promote the repair of islets from diabetic mice through paracrine actions ⇑ Xiaodong Gao a,b,1, Lujun Song a,b,1, Kuntang Shen a,b, Hongshan Wang a,b, Mengjia Qian c, Weixin Niu a, , ⇑ Xinyu Qin a,b, a Department of General Surgery, Zhongshan Hospital, Shanghai Medical College, Fudan University, Shanghai, People’s Republic of China b Institute of General Surgery, Fudan University, Shanghai, People’s Republic of China c Experimental Research Center, Zhongshan Hospital, Shanghai Medical College, Fudan University, Shanghai, People’s Republic of China article info abstract

Article history: Transplantation of bone marrow mesenchymal stem cells (MSCs) has been shown to effectively lower Received 22 October 2013 blood glucose levels in diabetic individuals, but the mechanism has not been adequately explained. Received in revised form 18 January 2014 We hypothesized that MSCs exert beneﬁcial paracrine actions on the injured islets by releasing biologi- Accepted 7 March 2014 cally active factors. To prove our hypothesis, we tested the cytoprotective effect of conditioned medium Available online 22 March 2014 from cultured MSCs on isolated islets exposed to STZ in vitro and on mice islets after the experimental induction of diabetes in vivo. We assessed islet regeneration in the presence of conditioned medium Keywords: and explored the possible mechanisms involved. Transplantation of MSCs can ameliorate hyperglycemia Regeneration in diabetic mice by promoting the regeneration of b cells. Both b cell replication and islet progenitors Diabetes Cell therapy differentiation contribute to b cell regeneration. MSC transplantation resulted in increases in pAkt and Paracrine signaling pErk expression by islets in vivo. Treatment with MSC-CM promoted islet cell proliferation and resulted in increases in pAkt and pErk expression by islets in vitro. The MSC-CM-mediated induction of b cell proliferation was completely blocked by the PI3K/Akt inhibitor LY294002 but not by the MEK/Erk inhibitor PD98059. Together, these data suggest that the PI3K/Akt signal pathway plays a critical role in b cell proliferation after MSC transplantation. Ó 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction in vitro transplantation for several months (Kroon et al., 2008; D’Amour et al., 2006). However, the application of hESCs has ethi- Type 1 and type 2 diabetes result from a lack of functional b cal issues and can cause teratoma formation and other problems cells. Pancreas and islet transplantation have proven successful in (Furth and Atala, 2009). These issues remain to be addressed and restoring functional b cells. However, donor organ shortage has require resolution before this approach can become a therapeutic limited the clinical application of these transplantations and option. In addition to identifying an exogenous source of b cells, incentivized efforts to generate insulin-producing cells from other another approach would be to increase the inherent b cells mass sources. Two possible solutions exist: identifying insulin-produc- by promoting the regeneration of a patient’s own b cells. If this ing cells to be used for cell therapy and promoting the regeneration method becomes possible, the limitations of in vivo regeneration of endogenous damaged b cells. could be overcome without requiring immunosuppressive therapy. Considerable efforts have been directed toward the develop- Recently, different types of bone marrow-derived cells have ment of efficient protocols for the differentiation of stem cells of been proposed as potential sources of cells for diabetic cell therapy either embryonic or adult origin into insulin-expressing cells. (Hess et al., 2003; Banerjee et al., 2005; Hasegawa et al., 2007; Lee Heretofore, robust glucose-induced insulin secretion, the main et al., 2006; Zhao et al., 2008; Madec et al., 2009). One strategy uses function of b cell, has only been possible in hESC-derived cells after mesenchymal stem cells (MSCs). Adult bone marrow-derived MSCs are multipotent stem cells that, given their ease of isolation, low immunogenicity, and amenability to ex vivo expansion, are opti- ⇑ Corresponding authors. Address: 180 Road Fenglin, Shanghai, People’s Republic mal candidates for diabetic cell therapy (Pittenger and Martin, of China. Tel./fax: +86 21 64037224 (X. Qin). Tel./fax: +86 21 64048038 (W. Niu). 2004). Indeed, transplantation of bone marrow MSCs has been E-mail addresses: [email protected] (W. Niu), [email protected] (X. Qin). demonstrated to be an effective strategy for the repair of diabetic 1 These authors contributed equally to this work. islets in experimental models (Lee et al., 2006). Seven registered http://dx.doi.org/10.1016/j.mce.2014.03.004 0303-7207/Ó 2014 Elsevier Ireland Ltd. All rights reserved. 42 X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50 clinical trials on type 1 and/or type 2 diabetes in phase I/IIcan be Shanghai, China) between 9:00 and 11:00 a.m. every 3 days after found on the website http://www.clinicaltrial.gov (Si et al., STZ injection and then weekly after bone marrow MSC 2011). In these clinical trials, MSCs exhibited exciting therapeutic transplantation. effects in diabetic volunteers (Jiang et al., 2011). However, the mechanisms underlying these therapeutic effects have not been 2.4. Mesenchymal stem cell purification and transplantation clearly defined. Bone marrow-derived MSCs have been shown to secrete various cytokines, including VEGF and FGF (Madec et al., Bone marrow was collected from the femurs of mice or GFP- 2009). Through these actions, the transplantation of MSCs has been transgenic mice. The mononuclear fraction of bone marrow was experimentally reported to improve diabetic islets and promote isolated from a Ficoll density gradient. The nucleated cells were the expansion of endogenous islet progenitors (Si et al., 2012; Park plated in plastic culture dishes (Corning, NY, USA) and incubated et al., 2010). However, the underlying mechanisms of this phenom- in Dulbecco’s modified Eagle’s medium (DMEM; Sigma–Aldrich, enon are not yet fully understood. We hypothesized that MSCs St. Louis, MO, USA) containing 10% fetal bovine serum (Gibco, exert beneficial paracrine actions on the injured islets by releasing Grand Island, NY, USA) at 37 °C in a 5% humidified CO2 atmosphere. biologically active factors. To prove our hypothesis, we tested the MSCs were selected based on their adherent property of preferen- cytoprotective effect of conditioned medium from cultured MSCs tially attaching to culture dishes. The cells were continuously cul- on isolated islets exposed to STZ in vitro and on mice islets after tured as MSCs until passage 3. The cells were then incubated with the experimental induction of diabetes in vivo and assessed the PE anti-mouse CD117, PE anti-mouse CD29, FITC anti-mouse CD44, regeneration of islets in the presence of conditioned medium. PE-Cy anti-mouse SCA-1, or PE anti-mouse CD34 antibodies (Biolegend, San Diego, CA, USA). Isotype-identical antibodies served as controls. Passage 3 MSCs were incubated to differentiate 2. Materials and methods into adipocytes, osteoblasts in corresponding induction medium for 3 weeks. MSCs adipogenic differentiation medium kit (Cyagen, 2.1. Animals Santa Clara, CA, USA) includes medium A and B. Medium A includes basal medium A 175 ml, FBS 20 ml, Glutamine 2 ml, Insulin 400ul, Six-week-old C57BL/6J mice were purchased from the Shanghai Indomethacin 200ul, Dexamethasone 200ul. Medium B contains Laboratory Animal Center at the Chinese Academy of Sciences. Basal medium 175 ml, FBS 20 ml, Glutamine 2 ml and Insulin Green fluorescent protein (GFP) transgenic mice with the C57BL/ 400ul. For adipogenesis, we followed the instruction of adipogene- 6J background were purchased from the Department of Cell Biol- sis protocol. MSC osteogenic differentiation kit (Cyagen, Santa ogy, Secondary Military Medical University, Shanghai. All animals Clara, CA, USA) contains basal medium 175 ml, FBS 20 ml, Gluta- were maintained in the animal facility of Zhongshan Hospital, mine 2 ml, Ascorbate 400ul, b-Glycerophosphate 2 ml and Dexa- Fudan University, Shanghai. The protocols of animal use complied methasone 20ul. For osteogenic differentiation, we followed the with the principles of laboratory Animal Care (NIH Publication 85- instruction of osteogenesis protocol. Eight hours before transplan- 23) and approved by the Ethics committee of Zhongshan Hospital, tation, recipient mice were irradiated (500 cGy), and MSCs Fudan University. (6.5 106) were transplanted into the diabetic mice through their tail veins. 2.2. Antibodies 2.5. Conditioned medium Guinea pig (GP) anti-mouse insulin (Invitrogen, Carlsbad, CA, USA), rabbit anti-GFP (Millipore, Billerica, MA, USA), goat anti- Conditioned medium was generated as follows: 80% confluent, mouse insulin, rabbit anti-mouse glucagon, rabbit anti-mouse passage-3 MSCs in 10-cm tissue culture dishes (Corning) were insulin, goat anti-mouse nestin, goat anti-mouse PDX-1, goat fed with 5 mL of serum-free DMEM per dish for 12 h in a chamber. anti-Pax4, goat anti-Pax6, goat anti-Nkx6.1, goat anti-Nkx2.2, goat The conditioned medium was further concentrated (50 times) by anti-Ngn3, and goat anti-NeuroD antibodies were all purchased ultra-filtration using 5-kDa cut-off centrifugal filter units (Milli- from Santa Cruz Biotechnology, Santa Cruz, CA, USA. Mouse mono- pore) following the manufacturer’s instructions. clonal anti-BrdU antibody was purchased from Sigma–Aldrich, St. Louis, MO, USA. FITC conjugated chicken anti-mouse secondary 2.6. Isolation of islets antibody was purchased from Santa Cruz. Rhodamine-conjugated donkey anti-GP secondary antibody was purchased from Millipore. After the mice were killed, their pancreata were inflated via the Cy2-conjugated donkey anti-rabbit and Cy3-conjugated donkey pancreatic duct with type V collagenase (0.7 mg/mL in Hank’s anti-goat secondary antibodies were purchased from Jackson balanced salt solution, Sigma), excised, and digested at 37 °C for ImmunoResearch, West Grove, PA, USA. PE anti-mouse CD117, PE 7–10 min. The resulting digest was washed twice with cold Hank’s anti-mouse CD29, FITC anti-mouse CD44, PE anti-mouse SCA-1, balanced salt solution containing 5% bovine serum albumin, and and PE anti-mouse CD34 antibodies were purchased from Bioleg- the islets were separated using a Histopaque density gradient. 0 end, San Diego, CA, USA. 4 -6-diamidino-2-phenylindole (DAPI) The interface-containing islets were removed and washed with was purchased from Sigma–Aldrich. Rabbit antiphosphorylated Hank’s balanced salt solution containing bovine serum albumin, Akt (pAkt, Ser473), rabbit anti-Akt, rabbit antiphosphorylated and the islets were resuspended in RPMI 1640 medium (Gibco, Erk1/2 (pErk, thr202.tyr204) and rabbit anti-Erk antibodies were Grand Island, NY, USA) containing 5% bovine serum albumin, purchased from Cell Signaling, Beverly, MA. 100 U/mL penicillin, and 100 ug/mL streptomycin. Following 2 h of incubation at 37 °C, the islets were handpicked into fresh media 2.3. Induction of diabetes and washed twice in PBS. The islets were fed with RPMI 1640 medium containing 10% fetal bovine serum albumin on plastic dishes at

Mice were injected intraperitoneally (i.p.) with 50 mg/kg of 37 °C in a 5% humidiﬁed CO2 atmosphere. Each dish (30 mm) con- streptozotocin (STZ; Sigma–Aldrich) daily for 7 consecutive days. tained approximately 40 islets. For islet injury, we prepared the STZ was solubilized in sodium citrate buffer, pH 4.5, and injected solution of STZ (102 mmol/L). And each dishes added about within 15 min of preparation. Non-fasting blood glucose was mea- 100 ll solution of STZ. After an hour later, we changed the medium sured using a One Touch Sure Step meter (Johnson & Johnson, and stop the injury of STZ. For MSC-CM treatment, we changed the X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50 43 medium every 3 days and added about 100 ll concentrated MSC- 7.5; 150 mmol/L NaCl; 1 mmol/L EDTA; 1 mmol/L EGTA; 1 mmol/

CM each time. lNa3VO4; and 1 lg/mL leupeptin), and the lysates were centrifuged at 14,000 rpm for 20 min at 4 °C. Protein concentrations were mea- 2.7. BrdU incorporation sured using a bicinchoninic acid assay (Sigma–Aldrich, St. Louis, MO, USA). Ten micrograms of protein were resolved on a 7.5– The thymidine analog 5-bromo-20-deoxyuridine (BrdU; Sigma– 10% SDS–PAGE gel and transferred to nitrocellulose membranes Aldrich) was injected intraperitoneally at 0.05 g per gram of body (Millipore, Billerica, MA). The membranes were then blocked with weight 24 h before harvesting the pancreata. The pancreata were 0.1%Tween-20 Tris-buffered saline containing 5% nonfat dry milk. then isolated and processed for fluorescent immunohistochemical The membranes were then incubated with antibodies against studies. In vitro, BrdU (40 lmol/L) was added to the culture 24 h either pAkt, Akt, pErk, Erk, or GAPDH (1:1000 dilution) overnight before staining. To calculate the number of BrdU-positive cells at 4 °C. The membranes were then incubated with a secondary per- among the islets, we microscopically examined the entire surface oxidase-conjugated antibody (1:5000 dilution) after washing. The of all dishes. immunoreactions were visualized using ECL Plus or ECL Western Blotting Detection. The resulting images were scanned, and the 2.8. Histological and morphological analyses densities were measured using image J software. Protein expression levels were corrected based on GAPDH density. Mice pancreata were excised and fixed overnight in 4% paraformaldehyde at 4 °C. Fixed tissues were processed for paraffin 2.12. Statistical analysis embedding, and serial 5-lm-thick sections were prepared and stained with hematoxylin/eosin (HE; Beyotime Institute of Biology, Data are expressed as the means ± SE. Differences between Suzhou, China) to assess pancreatic islet histology and morphology experimental groups were evaluated using the unpaired Student’s in the experimental animals. Every tenth, twentieth, and thirtieth t test for several independent observations. A value of p < 0.05 was section of the islets were observed. The histology and morphology considered to be statistically significant. The P values indicated in of the islets were scored using a Carl-Zeiss Axiovert 200 micro- the graphs were obtained using Student’s t test. scope (Carl-Zeiss, Jena, Germany) and a computer-assisted image analysis program (AxioVision Ver. 4.0; Carl-Zeiss). 3. Results 2.9. Fluorescent immunohistochemistry of the pancreas 3.1. Characterization of bone marrow mesenchymal stem cells The slides were deparaffinized and rehydrated, and then antigen retrieval was performed using antigen unmasking buffer; the Most cultured adherent cells exhibited a fibroblastic morphol- slides were then blocked for 40 min at room temperature in 5% bo- ogy that is characteristic of MSCs. Fluorescence-activated cell sort- vine serum albumin (BSA) and phosphate-buffered saline (PBS). To ing (FACS) analysis of our bone marrow MSCs showed that they stain for insulin, the sections of paraffin-embedded pancreata were were negative for CD117 and CD34 and that they strongly ex- incubated overnight with GP anti-mouse insulin antibody (1:400) pressed typical surface antigens, such as CD44, CD29, and Sca-1 at 4 °C, following which the sections were incubated in secondary (Fig. 1A). When cultured in adipogenic or osteogenic medium, antibody with rhodamine (1:500) at 37 °C for 30 min. Twenty the cells differentiated into adipocytes (Fig. 1B) or osteoblasts fields were examined for each section, and the number of islets (Fig. 1C), respectively. and insulin-positive cells in each field was determined. For double staining, the sections were incubated overnight at 4 °C with the 3.2. Bone marrow MSCs repair the pancreatic islets of diabetic mice appropriate antibody. Labeled cells were visualized with the appropriate biotin-conjugated secondary antibody with rhodamine Six-week-old C57BL mice were administered STZ, which in- and Cy2. To stain for BrdU, the sections were pretreated with 2 N duced pancreatic injury and hyperglycemia. To assess the effect HCl for 30 min at 37 °C and then incubated overnight with mouse of bone marrow MSC-transplantation on diabetic islets, we ran- monoclonal anti-BrdU (1:50). Labeled cells were visualized with domly divided the diabetic mice into two groups at day 10. One the appropriate secondary antibody with FITC (1:200). Isotype- group (n = 24) received bone marrow MSCs from GFP mice, and matched antibodies and PBS were used as controls for the stained the other group (n = 24) received PBS. During the study period, sections. Nuclear regions were stained by DAPI counterstaining. the untreated mice exhibited persistent hyperglycemia. However, the blood glucose level of the MSC-transplanted diabetic mice 2.10. Fluorescent immunohistochemistry of the islets was significantly lower on days 7, 14, 21, 28, 35, and 42 (Fig. 2A). These results demonstrate that MSC transplantation can amelio- Islets cultured in dishes were fixed with 2% paraformaldehyde rate hyperglycemia in diabetic mice. We analyzed and counted on ice for approximately 20 min; the dishes were then washed the islets and b cells of mice in the two groups on days 7 (n =8) twice, and the islet cells were pretreated with 2 N HCl for 30 min and 42 (n = 16). STZ induced an inflammatory reaction and de- at 37 °C. The dishes were again washed twice and then exposed creased the mass of the pancreatic islets. Microscopic examination to 1% BSA. Anti-insulin antibody (1:1000) and anti-BrdU (1:100) of the hematoxylin-eosin-stained pancreatic sections on day 7 re- antibody were added, and the islets were incubated overnight at vealed that islet size and number were markedly decreased in 4 °C. The dishes were then incubated with secondary fluorescent the STZ-treated mice. After MSC transplantation, both the size antibodies against insulin and BrdU. The dishes were evaluated and number of the islets were partially restored, and the mass of and photographed using a confocal fluorescent microscope (Leica the b cells was partly recovered on day 7. In contrast, the numbers TCS SP2, Heidelberg, Germany). of islets and b cells remained very low in the untreated mice on day 7(Fig. 2B–D). On day 42, blood glucose levels continued to decline 2.11. Western blotting in the MSC-transplanted mice. Furthermore, compared with the untreated mice, the size and number of islets, and the mass of b The isolated islets were analyzed using western blotting. Sam- cells were increased significantly on day 42 after MSC transplanta- ples were lysed in detergent lysis buffer (20 mmol/L Tris–HCl, PH tion (Fig. 2E–G). 44 X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50

Fig. 1. Characterization of bone marrow mesenchymal stem cells. (A) Fluorescence-activated cell sorting (FACS) analysis of bone marrow MSCs. Passage 3 MSCs were analyzed by FACS after staining with FITC- or PE-conjugated antibodies against the indicated cell surface proteins. The proportion of CD34, CD117, CD44, SCA-1, CD29 in MSCs was 0.2 ± 0.1%, 0.2 ± 0.1%, 81.2 ± 15.7%, 92.3 ± 14.2%, 88.1% ± 14.6%. Differentiation of MSCs: Cultured in appropriate differentiation media, MSCs differentiated into adipocytes (indicated by the accumulation of lipid vesicles in the cells (B)) or osteoblasts (indicated by the expression of alkaline phosphatase, as indicated in red (C)). (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

3.3. Transdifferentiation of bone marrow MSCs The bHLH transcription factor Ngn3 is first observed in the embryonic pancreatic epithelium at E9.5, then reaches maximum To investigate whether bone marrow MSCs transdifferentiated expression by E15.5, and declines after birth; Ngn3 is not ex- into insulin-positive cells in our model, pancreata from MSC-trans- pressed in adult mice (Watada, 2004). Ngn3 is considered a marker planted mice were immunostained with anti-insulin antibody; of endocrine progenitor cells (Gu et al., 2002). To assay for putative then, an intensive search for both insulin- and GFP-positive cells pancreatic progenitor cells during pancreatic b cell regeneration, was conducted. GFP-positive cells were located around or in the is- we double stained the samples for insulin and Ngn3. No Ngn3+ lets; however, no double-positive cells were detected. To further cells were found in the untreated mice, but Ngn3+ cells were found confirm this result, we double-stained the sections for GFP and in the islets of the MSC-treated mice (Fig. 4D). PDX-1, Ngn3, Nestin, Nkx6.1, Pax4, NeuroD, and Pax6, all of which are expressed in different stages of islet b cell development (Fig. 3). Again, no double-stained cells were evident. Thus, we suggest that 3.5. Bone marrow MSC-conditioned medium improves function of the observed repair of the diabetic islets in our mouse model was diabetic islets in vivo and in vitro not due to differentiation of the transplanted MSCs. Immunohistochemistry indicated the presence of GFP-positive cells in and around diabetic islets, but these cells were unable to 3.4. The islet regeneration pathway differentiate into insulin-expressing cells. Furthermore, some studies have shown that bone marrow MSCs can produce large In theory, b cell regeneration occurs by the replication of amounts of growth factors (Kinnaird et al., 2004; Gnecchi et al., precursor cells and by neogenesis from precursor cells within the 2005, 2008; Poll et al., 2008; Liu et al., 2006). This led us to suspect pancreas. To test whether the observed regeneration of b cell that a paracrine mechanism underlies the supportive effect of was derived from the replication of b cell, we double-stained for MSCs on islet repair. insulin and BrdU. We observed insulin-positive cells expressing To test our hypothesis, we first studied the effects of condi- BrdU in the MSC-treated mice, but were unable to find these cells tioned medium obtained from cultured MSCs (MSC-CM) in vivo. in the untreated mice (Fig. 4A). Some studies have indicated that Concentrated conditioned medium was injected into the tail veins PDX-1 plays crucial roles in the differentiation of endocrine lin- of diabetic mice (MSC-CM). We found that the blood glucose levels eages, and we noticed that some insulinBrdU+ cells stained posi- of the MSC-CM mice were significantly lower than those of the tive for PDX-1 (Fig. 4B). We also performed double staining using MSC-treated mice on days 17, 24, 31, 38, 45, and 52 (Fig. 5A), antibodies against insulin and glucagon. Cells co-expressing insulin and the numbers of islets and b cells were partly restored in the and glucagon were found in the MSC-treated mice, but not in the MSC-CM mice (Fig. 5C and D). We also found cells that co-ex- untreated mice (Fig. 4C). pressed insulin and BrdU (Fig. 5E). To strengthen these morpholog- X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50 45

Fig. 2. The effect of bone marrow MSCs on pancreatic islets in diabetic mice. Ten days after STZ injection, diabetic mice received 0.5 mL of physiological saline or 5 106 MSCs resuspended in 0.5 mL of physiological saline. (A) Blood glucose levels were determined consecutively using a glucometer. (B) Pancreatic histology was studied using hematoxylin/eosin- and insulin immunofluorescence-stained sections at 7 days after MSC transplantation. Islets and b cells were characterized by immunofluorescence according to the presence and distribution of insulin (red) cells. The number of islets (C) and b cells (D) were quantified in the untreated and MSC-treated diabetic mice at day 7 after MSC transplantation. (E) Pancreatic islet histology of hematoxylin/eosin- and insulin immunofluorescence-stained sections at 42 days after MSC transplantation. The number of islets (F) and b cells (G) were quantified in diabetic mice at day 42 after MSC transplantation. Values of A, C, D, F, and G are means ± SE. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

ical observations and further investigate the protective effect of the tation; we found that the expression levels of pAkt and pErk were conditioned medium, we also studied the effects of the conditioned significantly increased in the islets of MSC-transplanted mice medium in vitro. Isolated islets from normal mice were cultured in (Fig. 6A–D). 30-mm plastic dishes for 2 weeks, and then STZ (5 104 mmol/L) To verify the importance of Akt signaling in vitro, we isolated is- was added to the islets. One hour later, the medium was changed, lets, injured them by exposing them to STZ, and then divided them and the islets were divided into 2 groups, one with normal medium into 2 groups as previously described (termed the untreated and (STZ, n = 6) and the other with concentrated conditioned medium MSC-CM groups). We tested the expression of pAkt and pErk at dif- and normal medium (MSC-CM, n = 6). Ten days later, we found ferent time points after MSC-CM was added to the islets and found more BrdU positive cells in the MSC-CM islets. To further confirm that the expression levels of pAkt and pErk were significantly these results, we double-stained for insulin and BrdU antibodies increased in the MSC-CM islets (Fig. 6E–G). (Fig. 5F). The results showed that there were more insulin-positive To determine the role of the Akt and Erk signaling pathways in cells and doubly stained (insulin and BrdU) cells in the MSC-CM is- pancreatic islet cell proliferation, the effects of LY294002 and lets (Fig. 5G). PD98059 on BrdU incorporation in vitro were examined. As shown in the figures, the MSC-CM islets exhibited significantly increased 3.6. MSCs activate Akt signaling in vivo and in vitro BrdU incorporation compared to the untreated islets; pretreatment with LY294002 (50 nmol/L) completely inhibited the MSC-CM- Recent contradictory results suggest that the signaling pathway mediated BrdU incorporation in pancreatic islet cells (Fig. 7A and of islet regeneration remains unclear (Wang et al., 2004; Schrader B). Conversely, PD98059 (50 nmol/L), a Mek/Erk inhibitor, did not et al., 2007; Chen et al., 2013; Hayes et al., 2013). Some studies attenuate the MSC-CM-mediated BrdU pathway during beta-cell indicate that the Akt and Erk signals may play a part in this path- proliferation. We double-stained for insulin and BrdU in islet cells way. To clarify which signal was activated after MSC transplanta- pretreated with LY294002 and/or PD98059. A significant increase tion, we tested the expression of pAkt in diabetic islets. Protein in cells that exhibited insulin and BrdU double-staining was noted was extracted from islets collected at 6 weeks after MSC transplan- in the MSC-CM-treated islets. Pretreatment with LY294002 also 46 X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50

Fig. 3. Differentiation and engraftment of bone marrow MSCs. At day 42 after MSC transplantation, tissue sections were immunostained with an anti-GFP (green) and anti- insulin (red), anti-PAX4 (red), anti-Nestin (red), anti-NKX6.1 (red), anti-Ngn3 (red), anti-PDX-1 (red), anti-PAX6 (red), and anti-NeuroD (red) antibodies, as indicated. Nuclei (blue) were stained with DAPI. The images show that insulin-positive cells (red) appeared in the diabetic pancreas, especially in and around islets. Almost no cells co- expressed GFP and insulin (1/10,000), and no cells simultaneously expressed GFP and PDX-1, Nestin, Nkx2.2, Nkx6.1, Pax4, Pax6 or Ngn3 in the pancreata of diabetic mice. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.)

Fig. 4. The effect of MSCs on cell proliferation and differentiation at day 42. (A) Pancreatic sections were stained with anti-insulin (red) and anti-BrdU (green) antibodies (magnification 200). (B) Representative visualization of insulin-negative, and BrdU- and PDX-1-positive cells in islets of untreated and MSC-treated diabetic mice (magnification 400). Anti-insulin and anti-PDX-1 antibodies are indicated in red; anti-BrdU antibody is indicated in green. In the merged image, cells expressing both PDX- 1 and BrdU are indicated in yellow. (C) Double immunostaining of pancreatic sections with anti-insulin (red) and anti-glucagon (green) antibodies. Nuclei (blue) were stained with DAPI. In the merged anti-insulin and anti-glucagon image, cells expressing both insulin and glucagon are indicated in yellow. D: Microscopic image demonstrating that Ngn3-positive cells (red) were detected in islets and that these cells did not express insulin (green). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

completely inhibited the BrdU incorporation, whereas pretreat- 2006; Zhao et al., 2008; Madec et al., 2009; Jiang et al., 2011; Si ment with PD98059 did not inhibit this incorporation (Fig. 7C et al., 2012). Heretofore, it was believed that the mechanisms and D). To conﬁrm the speciﬁc inhibitory effects of LY294002, we underlying the therapeutic effects of MSCs on hyperglycemia in- analyzed the MSC-CM-mediated phosphorylation of Akt in the is- volved islet regeneration, immunoregulation, and an improvement lets. Pretreatment with LY294002 completely blocked the MSC- in insulin sensitivity. Different mechanisms have been indicated by CM-mediated phosphorylation of Akt (Fig. 7E). different diabetic animal models, but both type 1 and type 2 diabetes are characterized by b cell loss and dysfunction. Therefore, a major goal of diabetes research is to discover methods of promot- 4. Discussion ing the regeneration of b cells. In this study, we found that MSCs could mobilize into diabetic pancreata and repair diabetic islets Currently, a number of studies and clinical trials have revealed by promoting b cell regeneration in vivo. We excluded the possibil- that the transplantation of MSCs can successfully treat hyperglyce- ity that MSC-derived functional b cells were generated in vivo mia in animals or subjects with type 1 or type 2 diabetes (Lee et al., because we were unable to detect cells co-expressing GFP and X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50 47

Fig. 5. The effect of bone marrow-derived, MSC-conditioned medium on diabetic islets. Ten days after STZ injection, diabetic mice received 0.5 mL of DMEM medium with 5 106 MSCs resuspended in either 0.5 mL of DMEM or 0.5 mL of concentrated MSC-conditioned medium. (A) Blood glucose levels were determined consecutively using a glucometer at various time points (n = 6–16, P < 0.05, P < 0.01). (B) Representative images of diabetic islets of different groups. Pancreatic histology was studied in hematoxylin/eosin- and insulin immunofluorescence (red)-stained sections at 42 days after treatment. The islets (C) and b cells (D) were quantified in the different groups at day 42 after treatment. (E) At day 42 after treatment, tissue sections were immunostained with anti-insulin (red) and anti-BrdU (green) antibodies. (F) Representative confocal microscopy images of STZ-injured islets grown in control medium and MSC-conditioned medium for 10 days after immunostaining for insulin (red) and BrdU (green). BrdU (40 lmol/L) was added 24 h before staining. (G) The number of double-stained cells per dish was determined for various groups (n =6,t test, P < 0.01). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) specific beta cell makers in the pancreas. However, the question of during the process of islet regeneration. Most importantly, several whether b cell regeneration occurs mainly via the replication of lines of evidence in our study indicate that islet progenitor’s preexisting b cells or via neogenesis from stem cells in the islets re- differentiation contributed to the regeneration of b cells. First, mains a major focus of interest. In recent years, doubts about the BrdU+insulin PDX-1+ cells were found in the islets of the existence and importance of b cell neogenesis, especially by stem MSC-treated mice. Previous studies have demonstrated that BrdU+- cells, have been raised. Lineage-tracing studies have only found Insulin PDX-1+ cells can differentiate into b cells in vivo (Duvillié evidence for b cell replication (Dor et al., 2004; Nir et al., 2007), et al., 2003). Second, Ngn3 was activated during the course of b cell and strong evidence to support the presence of a population of regeneration. Ngn3 is a key transcription factor for the differentia- stem cells that gives rise to pancreatic endocrine cell types in the tion of the endocrine pancreas (Watada, 2004). Previous studies adult pancreas remains lacking. Our studies found that the blood have shown that Ngn3+ cells are endocrine progenitor cells and glucose reduction was accompanied by an increase of insulin+ can differentiate into b cells in adults mice (Xu et al., 2008). Third, BrdU+ cells; this finding indicates that b cell proliferation occurred insulin and glucagon double-stained cells were also found during b 48 X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50

Fig. 6. MSC-CM significantly increases the expression of pAkt and pErk in vitro and in vivo. The expression of pAkt in islets was determined using western blotting at day 42 after MSC-CM injection (A and B). The expression of pErk in islets was determined using western blotting on day 42 after MSC injection (C and D). Isolated islets were cultured in dishes for 2 weeks and then injured with STZ (5 104 mmol/L). One hour later, the medium was changed; the islets were then cultured in dishes for 2 days. Phosphorylated Akt (pAkt) and total Akt (Akt) levels in cultured islets in response to MSC-CM were measured using western blot analysis (E). Phosphorylated Erk (pErk) and total Erk (Erk) levels in cultured islets in response to MSC-CM were measured using western blot analysis (F). The histogram represents the densitometric analysis of the bands in Fig. 6E and F after MSC-CM stimulation. Similar results were obtained in 3 independent experiments. cell regeneration, and, in vitro, this staining pattern is only found in similar to that observed with MSC transplantation. Furthermore, embryonic stem cells (D’Amour et al., 2006) and pancreatic ductal the number of islets and b cells in the MSC-CM-treated mice cells (Bonner-Weir et al., 2000) that have differentiated into b cells. were the same as those in the MSC-treated mice. We also ob- A major debate exists concerning the origin of stem cells that can served proliferation of b cells in the MSC-CM-treated mice. We produce b cells via neogenesis. Most reports have suggested that then showed that conditioned medium from MSCs exerts a strik- such cells originate in or adjacent to the pancreatic duct epithelium ing protective effect on isolated islets exposed to STZ in vitro. (Gnecchi et al., 2008; Inada et al., 2008), but a recent study did not Also, conditioned medium from MSCs promoted b cell prolifera- find any sign of b cell through lineage tracing after partial duct liga- tion. These new data strongly support the concept that the ef- tion (Solar et al., 2009). In this study, we noticed that these endo- fects of MSC-transplantation in diabetic mice are, to a great crine progenitor cells were located in the islets and did not find any extent, attributable to paracrine protection and the functional cells expressing stem cell markers in the pancreatic duct. Based on recovery of pancreatic islets. the results of our study, we propose that progenitor cells in the is- Akt and Erk are potent regulators of beta cell proliferation. lets were the main source of beta cell neogenesis in our diabetic Some studies have reported that the overexpression of active mouse model. Akt1 in mouse b cells substantially affects compartment size and In conclusion, we suggest that both the replication of b cell and function (Tuttle et al., 2001). Another study demonstrated that the differentiation of islet progenitors contribute to b cell regener- the anti-proliferative effect of pro-inflammatory cytokines in cul- ation after bone marrow MSC transplantation. Replication and dif- tured b cells is associated with an extracellular signal that is regu- ferentiation are not mutually exclusive. Both processes often occur lated by kinase 1/2 pathway inhibition (Blandino-Rosano et al., simultaneously, and eventually, one or both of these pathways 2008). Thus, which signal is involved in beta cell proliferation re- may be manipulated for the therapeutic treatment of diabetes. mains unknown. This study provides pioneering evidence linking The mechanisms involved in the process of b cell regenera- stem cell paracrine action to the activation of Akt signaling in the tion after MSC transplantation remain unknown. Many studies host. After MSC transplantation into diabetic mice, the expression have confirmed that MSCs can secrete a variety of cytokines, of pAkt and pErk was significantly increased in mouse pancreatic such as VEGF, EGF, and HGF (Gnecchi et al., 2008). Some of these islets; pAkt, in particular, increased by approximately 1.8 times. cytokines, such as EGF and HGF, are known to be crucial for the We also confirmed that MSC-CM could increase the expression of multiplication and apoptosis of b cells. The transplantation of pAkt and pErk in vitro. Meanwhile, the proliferation of b cells MSCs has been shown to improve the repair of infarcted hearts was strengthened accordingly. To further confirm whether the (Gnecchi et al., 2005) and skin wounds (Liu et al., 2006) through Akt and Erk pathways were involved in beta cell regeneration after the secretion of paracrine factors. However, whether MSCs pro- BMSC transplantation, we observed the proliferation of islets after mote the repair of islet function by paracrine stimulation is un- blocking the signal pathway. We found that the islets stopped pro- clear. Support for our paracrine hypothesis was provided by the liferating after the Akt pathway was blocked with LY294002 but results obtained from our MSC-conditioned medium experi- not with PD985009. Therefore, activation of the Akt pathway ap- ments. We first injected concentrated medium from MSCs into pears to be critical for stimulating the proliferation of islets, and diabetic mice and obtained remarkable islet protection in vivo, pancreatic b cell in particular. X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50 49

Fig. 7. LY294002, but not PD98059, blocks cell proliferation in pancreatic islet cells after treatment with MSC-CM. The effect of LY294002 or PD98059 on BrdU incorporation by isolated islet cells was assessed (A). Isolated islets were cultured for 2 weeks and then injured with STZ (5 104 mmol/L). The isolated islets were preincubated with LY294002 (50 lmol/L) and/or PD98059 (50 lmol/L) for 30 min, after which 5% MSC-CM was added. BrdU (40 lmol/L) was added 24 h before staining. The number of BrdU- positive cells was calculated for various groups (B). Values represent the means ± SEM, n = 9. The effect of LY294002 or PD98059 on BrdU incorporation by insulin+ cells was assessed (C). The number of insulin+BrdU+ cells was measured for various groups (D). The effect of LY294002 on PI3K/Akt activation after MSC-CM administration was measured using western blot analysis (E). The islets were preincubated with LY294002 (50 lmol/L) for 30 min prior to 5% MSC-CM treatment and then harvested 1 h later for protein extraction and the determination of pAkt and total Akt expression levels. Similar results were obtained in 3 independent experiments.

In conclusion, our study suggests that MSCs exert direct salu- Acknowledgements tary effects on diabetic islets via paracrine mediators. The therapeutic benefits of MSCs appear to be attributable primarily to This research was supported by a grant from the Research Fund diffusible factors released by cells that promote b cell regeneration. for the Doctoral Program of Higher Education of China The PI3K/Akt pathway plays an essential role in b cell proliferation (20090071110019) and a grant from the Shanghai Municipal Gov- after MSCs transplantation. And MSCs or MSCs derived molecules ernment (09ZR1405900). may promise novel therapeutic approaches in the treatment of diabetes. The future identification of the exact nature and mechanism of action of the secreted factors may have important implications References for the development of islet injury and regeneration. And these new findings would bring new ideas for the treatment of diabetes. Banerjee, M., Kumar, A., Bhonde, R.R., 2005. Reversal of experimental diabetes by multiple bone marrow transplantation. Biochem. Biophys. Res. Commun. 328, 318–325. Blandino-Rosano, M., Perez-Arana, G., Mellado-Gil, J.M., et al., 2008. Anti- Author contributions 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, 35–44. X.G. and L.S.: conception and design, data analysis, manuscript Bonner-Weir, S., Taneja, M., Weir, G.C., et al., 2000. In vitro cultivation of human writing; K.S. and H.W.: collecting of data and data analysis; M.Q.: islets from expanded ductal tissue. Proc. Natl. Acad. Sci. USA 97, 7999–8004. Chen, S.S., Jiang, T., Wang, Y., et al., 2013. Activation of double-stranded RNA- collecting of data; W.N. and X.Q.: conception and design, financial dependent protein kinase inhibits proliferation of pancreatic b cells. Biochem. support, manuscript writing and editing. Biophys. Res. Commun. http://dx.doi.org/10.1016/j.bbrc.2013.12.051. 50 X. Gao et al. / Molecular and Cellular Endocrinology 388 (2014) 41–50

D’Amour, K.A., Bang, A.G., Eliazer, S., et al., 2006. Production of pancreatic hormone- Liu, Y., Dulchavsky, D.S., Gao, X., et al., 2006. Wound repair by bone marrow stromal expressing endocrine cells from human embryonic stem cells. Nat. Biotechnol. cells through growth factor production. J. Surg. Res. 136, 336–341. 24, 1392–1401. Madec, A.M., Mallone, R., Afonso, G., et al., 2009. Mesenchymal stem cells protect Dor, Y., Brown, J., Martinez, O.I., et al., 2004. Adult pancreatic beta-cells are formed NOD mice from diabetes by inducing regulatory T cells. Diabetologia 52, 1391– by self-duplication rather than stem-cell differentiation. Nature 429, 41–46. 1399. Duvillié, B., Attali, M., Aiello, V., et al., 2003. Label-retaining cells in the rat pancreas: Nir, T., Melton, D.A., Dor, Y., 2007. Recovery from diabetes in mice by beta cell location and differentiation potential in vitro. Diabetes 52, 2035–2042. regeneration. J. Clin. Invest. 117, 2553–2561. Furth, M.E., Atala, A., 2009. Stem cell sources to treat diabetes. J. Cell. Biochem. 106, Park, K.S., Kim, Y.S., Kim, J.H., et al., 2010. Trophic molecules derived from human 507–511. mesenchymal stem cells enhance survival, function, and angiogenesis of Gnecchi, M., He, H., Liang, O.D., et al., 2005. Paracrine action accounts for marked isolated islets after transplantation. Transplantation 89, 509–517. protection of ischemic heart by Akt-modiﬁed mesenchymal stem cells. Nat. Pittenger, M.F., Martin, B.J., 2004. Mesenchymal stem cells and their potential as Med. 11, 367–368. cardiac therapeutics. Circ. Res. 95, 9–20. Gnecchi, M., Zhang, Z., Ni, A., et al., 2008. Paracrine mechanisms in adult stem cell Poll, D.V., Parekkadan, B., Cho, C.H., et al., 2008. Mesenchymal stem cell-derived signaling and therapy. Circ. Res. 103, 1204–1219. molecules directly modulate hepatocellular death and regeneration in vitro and Gu, G., Dubauskaite, J., Melton, D.A., 2002. Direct evidence for the pancreatic in vivo. Hepatology 47, 1634–1643. lineage: NGN3+ cells are islet progenitors and are distinct from duct Schrader, J., Rennekamp, W., Niebergall, U., et al., 2007. Cytokine-induced progenitors. Development 129, 2447–2457. osteoprotegerin expression protects pancreatic beta cells through p38 Hasegawa, Y., Ogihara, T., Yamada, T., et al., 2007. Bone marrow (BM) mitogen-actived protein kinase signaling against cell death. Diabetologia 50, transplantation promotes beta-cell regeneration after acute injury through 1243–1247. BM cell mobilization. Endocrinology 148, 2006–2015. Si, Y.L., Zhao, Y.L., Hao, H.J., et al., 2011. MSCs: biological characteristics, clinical Hayes, H.L., Moss, L.G., Schisler, J.C., et al., 2013. Pdx-1 activates islet a- and b-cell applications and their outstanding concerns. Ageing Res. Rev. 10, 93–103. proliferation via a mechanism regulated by transient receptor potential cation Si, Y., Zhao, Y., Hao, H., et al., 2012. Infusion of mesenchymal stem cells ameliorates channels 3 and 6 and extracellular signal-regulated kinases 1 and 2. Mol. Cell. hyperglycemia in type 2 diabetic rats: identiﬁcation of a novel role in improving Biol. 33, 4017–4029. insulin sensitivity. Diabetes 61, 1616–1625. Hess, D., Li, L., Martin, M., et al., 2003. Bone marrow-derived stem cells initiate Solar, M., Cardalda, C., Houbracken, I., et al., 2009. Pancreatic exocrine duct cells give pancreatic regeneration. Nat. Biotechnol. 21, 763–770. rise to insulin-producing beta cells during embryogenesis but not after birth. Inada, A., Nienaber, C., Katsuta, H., et al., 2008. Carbonic anhydrase II-positive Dev. Cell 17, 849–860. pancreatic cells are progenitors for both endocrine and exocrine pancreas after Tuttle, R.L., Gill, N.S., Pugh, W., et al., 2001. Regulation of pancreatic beta-cell growth birth. Proc. Natl. Acad. Sci. USA 105, 19915–19919. and survival by the serine/threonine protein kinase Akt1/PKBalpha. Nat. Med. 7, Jiang, R., Han, Z., Zhuo, G., et al., 2011. Transplantation of placenta-derived 1133–1137. mesenchymal stem cells in type 2 diabetes: a pilot study. Front. Med. 5, 94–100. Wang, Q., Li, L., Xu, E., et al., 2004. Glucagon-like peptide-1 regulates proliferation Kinnaird, T., Stabile, E., Burnett, M.S., et al., 2004. Marrow-derived stromal cells and apoptosis via activation of protein kinase B in pancreatic INS-1 beta cells. express genes encoding a broad spectrum of arteriogenic cytokines and Diabetologia 47, 478–487. promote in vitro and in vivo arteriogenesis through paracrine mechanisms. Watada, H., 2004. Neurogenin 3 is a key transcription factor for differentiation of Circ. Res. 94, 678–685. the endocrine pancreas. Endocr. J. 51, 255–264. Kroon, E., Martinson, L.A., Kadoya, K., et al., 2008. Pancreatic endoderm derived from Xu, X., D’Hoker, J., Stangé, G., et al., 2008. Beta cells can be generated from human embryonic stem cells generates glucose-responsive insulin-secreting endogenous progenitors in injured adult mouse pancreas. Cell 132, 197–207. cells in vivo. Nat. Biotechnol. 26, 443–452. Zhao, M., Amiel, S.A., Ajami, S., et al., 2008. Amelioration of streptozotocin-induced Lee, R.H., Seo, M.J., Reger, R.L., et al., 2006. Multipotent stromal cells from human diabetes in mice with cells derived from human marrow stromal cells. PLoS marrow home to and promote repair of pancreatic islets and renal glomeruli in ONE 3, e2666. diabetic NOD/scid mice. Proc. Natl. Acad. Sci. USA 103, 17438–17443.