Targeting pancreatic islet PTP1B improves islet graft revascularization and transplant outcomes

See allHide authors and affiliations

Science Translational Medicine  19 Jun 2019:
Vol. 11, Issue 497, eaar6294
DOI: 10.1126/scitranslmed.aar6294

Revascularizing eye-lets

Pancreatic islet transplantation is a potentially promising therapy for type 1 diabetes, but poor revascularization hinders islet long-term viability. Figueiredo et al. studied the role of protein tyrosine phosphatase 1B (PTP1B) in regulating islet vascularization. Islets from PTP1B−/− mice retained more endothelial cells during in vitro culture and restored normoglycemia when transplanted into the anterior chamber of the eye in diabetic mice. Short hairpin RNA knockdown of PTP1B in human islets yielded similar effects. Deletion of PTP1B increased vascular endothelial growth factor A secretion by activating peroxisome proliferator–activated receptor γ coactivator 1α and estrogen-related receptor α signaling. Targeting PTP1B may improve islet transplantation.


Deficient vascularization is a major driver of early islet graft loss and one of the primary reasons for the failure of islet transplantation as a viable treatment for type 1 diabetes. This study identifies the protein tyrosine phosphatase 1B (PTP1B) as a potential modulator of islet graft revascularization. We demonstrate that grafts of pancreatic islets lacking PTP1B exhibit increased revascularization, which is accompanied by improved graft survival and function, and recovery of normoglycemia and glucose tolerance in diabetic mice transplanted with PTP1B-deficient islets. Mechanistically, we show that the absence of PTP1B leads to activation of hypoxia-inducible factor 1α–independent peroxisome proliferator–activated receptor γ coactivator 1α/estrogen-related receptor α signaling and enhanced expression and production of vascular endothelial growth factor A (VEGF-A) by β cells. These observations were reproduced in human islets. Together, these findings reveal that PTP1B regulates islet VEGF-A production and suggest that this phosphatase could be targeted to improve islet transplantation outcomes.


One of the main clinical problems related to pancreatic islet transplantation is incomplete graft revascularization, which impairs oxygen and nutrient delivery and hormone and secretagogue modulation (13). These deficiencies contribute toward the primary failure of transplanted islets (4, 5). The native islet architecture is characterized by an ultradense capillary network responsible for the delivery of oxygen, hormones, and nutrients to islet cells and the efficient dispersal of islet hormones into the bloodstream (6). During isolation, islets are severed from their native vascular network (7); thus, after transplantation, the survival and function of islet grafts depend on the reestablishment of vessels within the grafts to derive blood flow from the host vascular system (5, 811).

Intra-islet endothelial cells (IECs) participate in the early stages of angiogenesis and vasculogenesis; however, the exact role of IECs have in improving revascularization outcomes is still unclear (1214). Although the molecular mechanisms underlying islet revascularization remain elusive, numerous factors have been implicated, such as vascular endothelial growth factor A (VEGF-A), a key angiogenic molecule that stimulates extraembryonic blood vessel formation (15, 16). Previous therapies aimed at inducing VEGF-A expression in islets have resulted in the hypervascularization of the islet, compromising islet architecture and inducing the loss of β cell mass (1719). VEGF-A is expressed and secreted in response to hypoxia through the activation of the hypoxia-inducible factor (HIF) signaling cascade (20, 21), or in response to nutrient deprivation or ischemia by the up-regulation of the peroxisome proliferator–activated receptor γ coactivator 1α (PGC1α) (22). The induction of VEGF-A by PGC1α does not involve the canonical hypoxia response pathway or HIF (22). Instead, PGC1α coactivates the orphan nuclear receptor ERRα (estrogen-related receptor α) (22, 23). ERRα is known to interact physically and functionally with PGC1α and is involved in the activation of fatty acid oxidation and oxidative phosphorylation (22, 24).

The protein tyrosine phosphatase 1B (PTP1B) regulates phosphotyrosine signaling in several intracellular transduction pathways involved in cell growth, differentiation, metabolism, apoptosis, and gene transcription (25). In endothelial cells (ECs), tyrosine phosphorylation of both the VEGF receptor 2 (VEGFR2) and the adhesion molecule VE-CADHERIN (vascular endothelial cadherin) constitutes an important branch of the signaling events by which VEGF-A stimulates angiogenesis and vasculogenesis (2628). PTP1B has been shown to inhibit phosphorylation of both proteins, thus negatively affecting these processes (27). However, the role of PTP1B in islet graft revascularization has not been previously described.


Deletion of PTP1B in pancreatic islets decreases the rate of EC loss in vitro

We sought to investigate EC content in pancreatic islets from constitutive PTP1B knockout (PTP1B−/−) mice. Immunofluorescence (IF) staining of the EC marker PECAM-1 (platelet EC adhesion molecule 1) in freshly isolated islets and in islets cultured for 2 days [IECs are lost with culture (14)] revealed that the PECAM+ area was similar to freshly isolated PTP1B−/− and PTP1B+/+ islets (Fig. 1, A and B). However, after 2 days of culture, PTP1B−/− islets exhibited threefold higher PECAM+ area than PTP1B+/+ islets (Fig. 1, A and B), suggesting increased IEC content in PTP1B−/− islets. To corroborate this observation, we quantified expression of genes encoding EC markers PECAM-1 (Pecam1), VEGFR2 (Kdr), and VE-CADHERIN (Cdh5). These genes were all up-regulated in cultured PTP1B−/− islets relative to PTP1B+/+ islets (Fig. 1C).

Fig. 1 PTP1B−/− mouse islets present reduced EC loss in vitro.

Pancreatic islets from adult PTP1B+/+ and PTP1B−/− mice were cultured for 2 days and processed for immunofluorescence (IF) and gene expression analysis. (A) Representative maximum projections of image stacks and (B) quantification of the percentage of PECAM-1+ area relative to total islet area in freshly isolated and 2-day cultured PTP1B+/+ (n = 550 cells × 15 islets) and PTP1B−/− (n = 500 cells × 15 islets) islets. Scale bars, 25 μm. (C) Quantitative real-time polymerase chain reaction (qRT-PCR) of the vascular marker genes Pecam1, Kdr, and Cdh5 in 2-day cultured PTP1B+/+ (n = 9) and PTP1B−/− (n = 12) islets. (D) Representative images of in toto IF staining of cleaved CASPASE-3 (green) and INSULIN (pink) in 2-day cultured PTP1B+/+ and PTP1B−/− islets. Nuclei (blue) were stained using Hoechst 33258. Scale bars, 25 μm. (E) Quantification of cleaved CASPASE-3+ cells relative to total islet cells in PTP1B+/+ (n = 550 cells × 15 islets) and PTP1B−/− (n = 550 cells × 15 islets) islets. (F) qRT-PCR of the apoptotic genes Casp3 and Casp9 in PTP1B+/+ (n = 9) and PTP1B−/− (n = 12) islets. Data presented are means ± SEM. n.s., not significant. **P < 0.01, by one-way analysis of variance (ANOVA) for (B); *P < 0.05, **P < 0.01, by Student’s t test for (C), (E), and (F).

Because PTP1B has been previously shown to modulate apoptosis and insulin secretion in β cell lines and freshly isolated mouse islets (2931), we evaluated whether these processes were affected in PTP1B−/− islets. We found similar Caspase3 and Caspase9 transcript expression and a comparable number of cleaved CASPASE-3+ cells between 2-day cultured PTP1B−/− and PTP1B+/+ islets (Fig. 1, D to F). Similarly, we found no differences in glucose-induced insulin secretion and islet insulin content between genotypes (fig. S1). Together, these observations reveal that PTP1B−/− islets present reduced loss of IECs in culture without apparent alterations in cell survival or insulin secretion.

Transplanted PTP1B−/− islets restore normoglycemia and circulating insulin concentration in diabetic mice

To study the effect of PTP1B loss on the engraftment of islets, we performed a suboptimal allotransplantation of 2-day cultured PTP1B−/− or PTP1B+/+ islets into the anterior chamber of the eye of streptozotocin (STZ)–treated diabetic BALB/c mice (Fig. 2A). The anterior chamber of the eye is a privileged transplantation site, enabling noninvasive in vivo imaging (32, 33), fast islet engraftment by blood vessels of the host iris (a highly vascular site) (3436), and reduced graft early rejection sustained by ocular immune privilege that suppresses immune cell proliferation and purge of the immune cells that enter the eye (37). One week after STZ administration, 85% of the mice presented an average blood glucose of 340 mg/dl; no mice recovered spontaneously from STZ-induced diabetes. STZ-treated diabetic mice were randomly organized into three groups: transplanted with PTP1B−/− islets (PTP1B−/−txbalb-stz), transplanted with PTP1B+/+ islets (PTP1B+/+txbalb-stz), or left untransplanted (nontransplanted). A fourth group of nondiabetic BALB/c mice was monitored in parallel and used to establish normoglycemia threshold. Each transplanted mouse received 200 islets and was monitored for 28 days. Mice in the PTP1B−/−txbalb-stz group showed a reduction in blood glucose concentration and achieved normoglycemia by day 21 (Fig. 2B). By contrast, the PTP1B+/+txbalb-stz and nontransplanted groups exhibited no recovery in nonfasting glycemia (Fig. 2B). All STZ-induced diabetic mice presented with an initial decrease in body weight, due to the effects of STZ (Fig. 2C). However, after transplantation, the PTP1B−/−txbalb-stz group recovered and gained body weight until the end of the experiment (Fig. 2C). The other two diabetic groups did not recover their initial body weight (Fig. 2C).

Fig. 2 Transplanted PTP1B−/− mouse islets restore normoglycemia and circulating insulin concentration in diabetic mice.

(A) Schematic diagram and representative image of islets transplanted into the anterior chamber of a mouse eye (ACE). (B to G) BALB/c mice were divided into four experimental groups: nondiabetic mice (n = 8) and PTP1B+/+txbalb-stz (n = 10), PTP1B−/−txbalb-stz (n = 12), or nontransplanted (n = 10). (B) Blood glucose concentration measured at the indicated times. Median blood glucose concentration at day 0 (transplantation) is 340 mg/dl and is represented as a dashed line. Normoglycemia (120 mg/dl) is set as the average blood glucose of nondiabetic mice during the 28-day follow-up period and is also represented as a dashed line. (C) Body weight measured at the indicated times. (D) Intraperitoneal glucose tolerance test (IPGTT) performed at day 28 after transplant (n = 5). (E) Calculation of the area under curve (AUC) of the IPGTT. a.u., arbitrary units. (F) Plasma insulin concentration during the IPGTT (n = 5). (G) Pancreatic insulin content at day 28 after transplant, expressed relative to nondiabetic mice (normalized as 1; n = 6). Data are presented as means ± SEM. *P < 0.05, **P < 0.01 for PTP1B+/+txbalb-stz versus PTP1B−/−txbalb-stz; #P < 0.05 for PTP1B+/+txbalb-stz versus nontransplanted; n.s. for PTP1B−/−txbalb-stz versus nondiabetic mice; $$$P < 0.001 for nondiabetic versus PTP1B−/−txbalb-stz mice, by two-way ANOVA; ***P < 0.001, by one-way ANOVA in (G).

To assess graft function, we performed an IPGTT on day 28. Results showed that the PTP1B−/−txbalb-stz group presented comparable glucose tolerance to nondiabetic mice, whereas the PTP1B+/+txbalb-stz and nontransplanted groups exhibited severe glucose intolerance (Fig. 2, D and E). Supporting these findings, plasma insulin concentration was similarly increased after the glucose bolus in PTP1B−/−txbalb-stz and nondiabetic groups, whereas the PTP1B+/+txbalb-stz and nontransplanted groups lacked glucose responsiveness (Fig. 2F). At the end of the follow-up period, we harvested pancreata to measure insulin content. All STZ-treated groups, regardless of the type of transplanted islets, presented an average of 10 to 15% remaining insulin content compared to nondiabetic mice (Fig. 2G), thus supporting that amelioration of glucose homeostasis in the PTP1B−/−txbalb-stz group originates from the islets transplanted in the eye.

PTP1B−/− islet grafts exhibit improved revascularization and survival

We next investigated revascularization of the islet grafts transplanted in the eye. We assessed in vivo graft functional revascularization 7, 15, and 28 days after transplantation, using two-photon microscopy after injection of rhodamine B isothiocyanate (RITC)–dextran (Fig. 3A). In agreement with other studies (2, 38), we found that vascular density reached its maximum 15 days after transplant in both experimental groups (Fig. 3B). Nonetheless, PTP1B−/−txbalb-stz grafts presented about 1.5-fold higher vascular density than PTP1B+/+txbalb-stz grafts at all times studied (Fig. 3B). Likewise, the percentage of vascular area was also maximum at day 15 in both groups, and PTP1B−/−txbalb-stz grafts presented a twofold greater vascular area than PTP1B+/+txbalb-stz at all times (Fig. 3C).

Fig. 3 PTP1B−/− mouse islet grafts exhibit improved revascularization and survival without loss of β cell area.

(A) Representative in vivo images of functional vessels (RITC-dextran; red) and cell death [propidium iodide (PI); white] in PTP1B+/+txbalb-stz and PTP1B−/−txbalb-stz eye grafts at 7, 15, and 28 days after transplant (Tx); nuclei are labeled with Hoechst 33258 (blue) and CFDA SE–stained islet cells (green). Scale bars, 25 μm. (B) Relative vascular density at the indicated times. (C) Percentage of vascularization area relative to total islet area at the indicated times. (D) Percentage of PI+ cells at day 28 after Tx (n = 600 cells × 8 islets per animal) in PTP1B+/+txbalb-stz and PTP1B−/−txbalb-stz grafts (n = 8 islets × 10 animals). (E) Representative IF images for cleaved CASPASE-3 (green), GLUCAGON (pink), and INSULIN (pink and green) in eye sections from PTP1B+/+txbalb-stz and PTP1B−/−txbalb-stz animals collected at day 28 after Tx. Scale bars, 25 μm. (F) Percentage of cleaved CASPASE-3+ cells relative to INSULIN+ cells in PTP1B+/+txbalb-stz and PTP1B−/−txbalb-stz grafts (n = 10 islets × 6 animals) grafts. (G) Percentage of INSULIN+ area relative to total graft area in PTP1B+/+txbalb-stz and PTP1B−/−txbalb-stz grafts (n = 5 islets × 6 animals). Data presented as means ± SEM. *P < 0.05, **P < 0.01, by two-way ANOVA in (B) and (C); *P < 0.05, **P < 0.01, ***P < 0.001, by Student’s t test in (D), (F), and (G).

To study graft survival, we assessed in vivo cell death after PI injection and apoptosis by cleaved CASPASE-3+ IF in paraffin sections of the engrafted eyes. Results revealed a significant decrease in the percentage of PI+ nuclei (P < 0.001; Fig. 3, A and D) and of cleaved CASPASE-3+ cells (P < 0.01; Fig. 3, E and F) in PTP1B−/−txbalb-stz grafts when compared with PTP1B+/+txbalb-stz grafts. Last, we measured insulin positive area in engrafted eyes and found that, on average, 76% of PTP1B−/−txbalb-stz graft cells were β cells as compared to 69% in the PTP1B+/+txbalb-stz graft (Fig. 3, E and G). In sum, these data demonstrate that transplanted PTP1B−/− islets exhibit improved vascularization and cell survival as compared with control islets.

IECs are not responsible for improved revascularization of PTP1B−/− islet grafts

Because IECs are known to participate in the early stages of graft revascularization (12, 3941), we examined their involvement in improved vascularization of PTP1B−/−txbalb-stz islet grafts. We depleted the IEC population from the islets to be transplanted by maintaining them in culture for several days (14). PTP1B−/− islets presented a lower rate of IEC loss than PTP1B+/+ islets (Fig. 4, A and B). However, by day 7, we were not able to detect any PECAM+ cells in either PTP1B−/− or PTP1B+/+ islets (Fig. 4B). These results were validated at the gene expression level (fig. S2A). Therefore, PTP1B−/− and PTP1B+/+ islets cultured for 7 days were considered to be free of IECs. Further, islets from both genotypes exhibited comparable glucose-induced insulin secretion (fig. S2B) and showed similar expression of the apoptotic genes Caspase3 and Caspase9 (fig. S2C). Confirming the lack of ECs, vestigial expression of the endothelial markers, Pecam1, Kdr, and Cdh5, was detected, but no differences were observed between genotypes (fig. S2D).

Fig. 4 IECs are not responsible for improved revascularization of PTP1B−/− mouse islet grafts.

(A) Representative maximum projections of image stacks of PECAM-1 IF staining (yellow) in PTP1B+/+ and PTP1B−/− mouse islets cultured for 0, 2, and 7 days; a white line defines islet area. Scale bars, 25 μm. (B) Relative PECAM-1+ area in relation to total islet area (n = 600 cells × 9 islets). (C and D) Blood glucose concentration and body weight of PTP1B+/+ECtxbalb-stz (n = 7) and PTP1B−/−ECtxbalb-stz (n = 7) mice at the indicated times; median concentration of blood glucose at day 0. 347 mg/dl; average concentration of blood glucose of nondiabetic mice. 120 mg/dl (n = 8); diabetes threshold. 250 mg/dl. (E) Representative IF images of PECAM-1 (green), cleaved CASPASE-3 (green), and INSULIN (pink) in grafts 28 days after transplantation; nuclei are labeled with Hoechst 33258 (blue). (F and G) Quantification of PECAM-1+ area relative to INSULIN+ area and cleaved CASPASE-3+ cells relative to INSULIN+ cells in PTP1B+/+ECtxbalb-stz and PTP1B−/−ECtxbalb-stz grafts. Scale bars, 25 μm; n = 550 cells × 7 islets per animal, 5 animals. Data presented as means ± SEM. *P < 0.05, **P < 0.01, by one-way ANOVA in (C) and (D); *P < 0.05, **P < 0.01, by Student’s t test in (B), (F), and (G).

We then transplanted a suboptimal number of IEC-depleted islets into the anterior chamber of the eye of STZ-treated diabetic BALB/c mice and followed them for 28 days. Mice transplanted with IEC-depleted PTP1B−/−islets (PTP1B−/−ECtxbalb-stz) achieved normoglycemia 7 days after transplant, whereas IEC-depleted PTP1B+/+islets (PTP1B+/+ECtxbalb-stz) did not normalize glycemia (Fig. 4C). Moreover, PTP1B−/−ECtxbalb-stz mice recovered and continued gaining weight until day 28, whereas PTP1B+/+ECtxbalb-stz mice did not gain any weight during the follow-up period (Fig. 4D). IF staining detected a 4.2-fold increase in PECAM-1+ area in the PTP1B−/−ECtxbalb-stz grafts when compared with the PTP1B+/+ECtxbalb-stz grafts (Fig. 4, E and F). In addition, PTP1B−/−ECtxbalb-stz grafts showed a 61% decrease in cleaved CASPASE-3+ cells when compared with PTP1B+/+ECtxbalb-stz (Fig. 4, E and G). Together, these results support the notion that donor IECs are not responsible for improved graft revascularization in the absence of PTP1B.

Cultured and transplanted PTP1B−/− islets exhibit enhanced VEGF-A production

To mechanistically understand the increase in revascularization in the PTP1B−/−txbalb-stz grafts, we initially investigated the expression of VEGF-A, a cytokine produced by islet cells that is known to be the principal inducer of angiogenesis (11, 15). IF analysis of paraffin sections of the engrafted eyes using antibodies against VEGF-A and insulin demonstrated that 89% of β cells in PTP1B−/−txbalb-stz grafts expressed VEGF-A, whereas 35% expressed VEGF-A in the PTP1B+/+txbalb-stz grafts (Fig. 5, A and B), thus suggesting that enhanced VEGF-A production may underlie increased vascular network formation in PTP1B−/−txbalb-stz islet grafts. Analysis of insulin/VEGF-A costaining revealed that, in both type of grafts, 90% of the cells expressing VEGF-A were β cells (Fig. 5A).

Fig. 5 PTP1B−/− mouse islets exhibit enhanced VEGF-A production and activation of the PGC1α/ERRα axis.

(A) Representative IF images of VEGF-A (yellow) and INSULIN (pink) staining in PTP1B+/+txbalb-stz and PTP1B−/−txbalb-stz grafts 28 days after transplantation; nuclei are labeled with Hoechst 33258 (blue). Scale bars, 25 μm. (B) Quantification of VEGF-A+/INSULIN+ cells relative to total INSULIN+ cells in PTP1B+/+txbalb-stz (n = 8 islets per animal × 9 animals) and PTP1B−/−txbalb-stz (n = 10 islets × 7 animals) grafts. (C) Vegfa expression in 2-day cultured PTP1B+/+ and PTP1B−/− islets determined by qRT-PCR and expressed relative to PTP1B+/+ (normalized as 1) (n = 9). (D) Representative IF images of INSULIN (pink) and VEGF-A (green) in 2-day cultured PTP1B+/+ and PTP1B−/− islets; nuclei are labeled with Hoechst 33258 (blue). Scale bars, 25 μm. (E) Quantification of VEGF-A+/INSULIN+ cells relative to total INSULIN+ cells in 2-day cultured PTP1B+/+ and PTP1B−/− islets (n = 700 cells × 14 islets). (F) VEGF-A secretion (normalized by content) from 2-day cultured PTP1B+/+ and PTP1B−/− islets (n = 6) determined by enzyme-linked immunosorbent assay (ELISA). (G) qRT-PCR of Hif1a, Ppargc1a, and Esrra in 2-day cultured PTP1B+/+ and PTP1B−/− islets (n = 9). (H) Representative immunoblot for PGC1α (105 kDa) and αTUBULIN (52 kDa) in whole-cell extracts and for HIF1α (120 kDa), ERRα (52 kDa), and LAMIN-B (66 kDa) in nuclear extracts of PTP1B+/+ and PTP1B−/− islets cultured for 2 days in complete medium or Hanks’ balanced salt solution (HBSS) (n = 2). (I and J) qRT-PCR of Vegfa, Ppargc1a, and Esrra in PTP1B+/+ (I) and PTP1B−/− (J) islets cultured in complete medium or HBSS (n = 10). (K) qRT-PCR of Hif1a, Vegfa, Ppargc1a, and Esrra in PTP1B+/+ and PTP1B−/− islets cultured in HBSS (n = 10). (L) VEGF-A secretion (normalized by total protein content) from PTP1B+/+ and PTP1B−/− islets cultured in complete medium or HBSS (n = 6). Data presented as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, by Student’s t test in (B), (C), (E), (F), (G), (I), (J), and (K) and by two-way ANOVA in (L).

To understand why PTP1B−/− islet grafts exhibited higher VEGF-A expression, we studied Vegfa gene expression, VEGF-A protein content, and VEGF-A secretion by PTP1B−/− and PTP1B+/+ islets cultured for 2 days (the conditions under which they were transplanted in Figs. 2 and 3). We found that Vegfa transcript levels were 3.9-fold higher in PTP1B−/− islets relative to controls (Fig. 5C). In agreement with increased gene expression, IF analysis revealed that 92% of β cells from PTP1B−/− islets were positive for VEGF-A as compared to <25% of β cells in control islets (Fig. 5, D and E). PTP1B−/− islets also exhibited augmented VEGF-A fractional secretion (relative to content) versus PTP1B+/+ islets (Fig. 5F). IF analysis of VEGF-A expression in pancreata from PTP1B+/+ and PTP1B−/− mice showed marginal VEGF-A staining in islets in both genotypes (fig. S3), supporting that native PTP1B−/− islets do not express more VEGF-A and that islet VEGF-A production is potentiated after culture.

PTP1B−/− islets present activation of the PGC1α/ERRα axis

On the basis of the previous findings (2022), we reasoned that the increase in VEGF-A production in 2-day cultured PTP1B−/− islets might be due to enhanced hypoxia. However, we found no differences in Hif1a gene expression in knockout relative to control islets (Fig. 5G), thus indicating that HIF1α is unlikely to be responsible for up-regulated Vegfa expression in PTP1B−/− islets. Instead, we observed up-regulation of Ppargc1a (encoding PGC1α) and Esrra (encoding ERRα) (Fig. 5G), revealing a possible modulation of the PGC1α/ERRα pathway by PTP1B. These results were validated at the protein level using immunoblot analysis (Fig. 5H).

Because the genes encoding PGC1α/ERRα and VEGF-A are normally induced by nutrient deprivation (22), we sought to examine the ability of PTP1B to modulate this regulation. We cultured PTP1B+/+and PTP1B−/− islets for 2 days in standard complete medium or in HBSS to mimic nutrient deprivation. Islets cultured in HBSS showed increased Ppargc1a and Esrra mRNA as compared to islets cultured in complete medium, irrespective of the presence or absence of PTP1B (Fig. 5, I and J). However, Vegfa mRNA was only significantly induced by HBSS culture in PTP1B−/− islets (P < 0.001; Fig. 5J). Because gene responses to nutrient deprivation were larger in PTP1B−/− than in PTP1B+/+ islets, transcript levels for these genes were higher in nutrient-deprived PTP1B−/− islets relative to nutrient-deprived PTP1B+/+ islets (Fig. 5K). Note that no difference was observed between PTP1B+/+ and PTP1B−/− islets regarding Hif1a expression, supporting that nutrient deprivation does not activate this gene (Fig. 5K). Last, we measured VEGF-A secretion and confirmed that PTP1B−/− islets secreted more VEGF-A than PTP1B+/+ islets under normal and nutrient deprivation conditions (Fig. 5L). VEGF-A secretion by PTP1B−/− islets was increased in nutrient deprivation, whereas that of PTP1B+/+ islets did not vary (Fig. 5L). Together, these results point to PGC1α, not HIF1α, as the likely mediator of the effects of PTP1B loss on Vegfa expression.

Silencing PTP1B in human islets induces VEGF-A expression and improves graft revascularization

We next asked whether the regulation of VEGF-A by PTP1B was conserved in human islets and, if so, whether loss of PTP1B improved vascularization of human islet grafts. We obtained five different human islet preparations with an average purity of 95.5% (±1.3%) as assessed by dithizone staining and with a mean viability of 85.0% (±7.5%) as assessed by a carboxyfluorescein diacetate succinimidyl ester (CFDA SE)/PI staining (fig. S4).

We used RNA interference to reduce PTP1B in human islets. In a preliminary experiment, we evaluated the uptake of a nontargeting fluorescence-labeled small interfering RNA (siRNA) and confirmed strong green fluorescence in 52 to 87% of islet cells (fig. S5). We then used a pool of siRNAs against PTP1B (PTP1B-siRNA) and corroborated the reduction of PTP1B protein levels by immunoblot analysis (Fig. 6A) and by IF staining (Fig. 6B) in islets transfected with PTP1B-siRNA relative to islets transfected with a nontargeting scrambled-siRNA. Quantification of PTP1B+/INSULIN+ cells revealed a 57% decrease in PTP1B+ β cells in PTP1B-siRNA islets (Fig. 6C). Having validated the reduction in PTP1B protein content, we surveyed gene expression in two independent human islet batches (Fig. 6D). Two days after transfection, the expression of the PTP1B gene (PTPN1) was reduced by 30 and 50%; VEGFA expression was up-regulated 1.4- and 3-fold; PPARGC1A expression was up-regulated 1.2- and 1.6-fold; and ESRRA expression was up-regulated 1.3- and 2.8-fold in islets treated with PTP1B-siRNA, as compared to scrambled-siRNA (Fig. 6D). By contrast, HIF1A mRNA levels remained unmodified (Fig. 6D). In agreement with activation of the VEGFA gene, PTP1B-silenced human islets exhibited sixfold higher VEGF-A secretion than islets transfected with the scrambled-siRNA (Fig. 6E).

Fig. 6 PTP1B silencing in human islets induces VEGF-A expression and improves graft vascularization.

(A) Immunoblot for PTP1B and αTUBULIN using whole-cell extracts from scrambled-siRNA– and PTP1B-siRNA–treated islets 3 days after transfection. (B) Representative IF images of INSULIN (green) and PTP1B (pink) in scrambled-siRNA– and PTP1B-siRNA–treated islets 3 days after transfection; nuclei are labeled with Hoechst 33258 (blue) (n = 8). (C) Percentage of PTP1B+/INSULIN+ cells relative to total INSULIN+ cells in PTP1B-siRNA– and scrambled-siRNA–transfected human islets (n = 8 islets × 700 cells per islet). (D) qRT-PCR for PTPN1, HIF1A, VEGFA, PPARGC1A, and ESRRA in scrambled-siRNA–treated (n = 6) and PTP1B-siRNA–treated (n = 6) islets 3 days after transfection. qRT-PCR for GAPDH in human islets transfected with GAPDH-siRNA or left untransfected (control) is shown as positive control. (E) Secretion of VEGF-A (normalized by total protein content) from scrambled-siRNA– and PTPB1-siRNA–treated islets (n = 4). (F) Fluorescence images of islets infected with sh-PTP1B LV [encoding green fluorescent protein (GFP)] and noninfected (control) islets 2 and 8 days after lentiviral infection. Scale bars, 50 μm. Insets are shown at higher magnification. eGFP, enhanced GFP. (G) Immunoblot for PTP1B and αTUBULIN in cell extracts of human islets 2 and 8 days after or sh-PTP1B LV. (H) Secretion of VEGF-A (normalized by total protein content) from human islets 2 and 8 days after infection with sh-scrambled LV or sh-PTP1B LV (n = 4). (I) Representative in vivo intraocular images of functional vessels using RITC-dextran (red) in sh-scrambled LV and sh-PTP1B LV islet grafts 8 days after transplantation; GFP fluorescence (green) shows human islet cells infected with sh-scrambled LV and sh-PTP1B LV. Scale bars, 25 μm. Data presented as means ± SEM. *P < 0.05, **P < 0.01 by Student’s t test in (C), (D), and (E) and by two-way ANOVA in (H).

After confirming enhanced production of VEGF-A, we investigated whether PTP1B inhibition improved revascularization of human islet grafts. To ensure silencing of PTP1B for longer duration (as required in transplantation experiments), we used lentiviral particles carrying short hairpin RNAs (shRNAs) against PTP1B (sh-PTP1B LV). We first monitored infection efficiency by assessing GFP fluorescence [sh-PTP1B LV carries a cytomegalovirus (CMV)–driven turboGFP (tGFP)] and observed abundant GFP+ cells 2 days after infection and fewer cells 8 days after infection (Fig. 6F). We then measured PTP1B protein content in whole islet extracts by immunoblot analysis and found that it was severely reduced 2 days after sh-PTP1B LV treatment but returned to control values (islets infected with sh-scrambled LV) by day 8 (Fig. 6G). In agreement with PTP1B down-regulation, VEGF-A secretion was 16-fold higher in sh-PTP1B LV islets than in sh-scrambled LV at day 2 (Fig. 6H). Despite VEGF-A secretion having decreased by 76% at day 8, VEGF release by sh-PTP1B LV islets remained higher than sh-scrambled LV islets (Fig. 6H). Last, we transplanted 150 PTP1B-silenced or control human islets into the eyes of nonobese diabetic severe combined immunodeficient interleukin 2 receptor gamma chain null (NSG) mice (n = 3 and n = 2, respectively). Eight days after transplantation, grafts of sh-PTP1B LV islets presented improved functional vascularization compared to control grafts as measured by RITC-dextran and two-photon microscopy (Fig. 6I). When contrasted against the in vitro results, sh-PTP1B LV islets maintained high expression of GFP 8 days after implantation, suggesting that lentiviral expression was sustained longer in vivo than in vitro. These results confirm that inhibition of PTP1B induces VEGF-A production in human islets and improves revascularization of transplanted islets.


Islet transplantation as a potential treatment for type 1 diabetes fails primarily due to poor survival of the islet grafts. Insufficient revascularization is mainly responsible for early graft loss and represents one of the major issues affecting long-term graft survival (7, 39, 42). Searching for new targets to facilitate graft revascularization may lead to improved future outcomes in islet transplantation (41).

Here, we demonstrated that the absence of islet PTP1B increased graft functional vascularization by increasing the number of newly formed vessel branches and total graft vessel area. Several mechanisms could explain improved revascularization in PTP1B−/− islet grafts (11, 14, 41). Because isolated PTP1B−/− islets show an up-regulation of several EC markers and reduced IEC loss in culture, we considered the possibility that donor PTP1B−/− IECs contributed to the increased functional vasculature in grafts (14, 41, 43). However, this hypothesis was refuted by our observation that mice transplanted with IEC-depleted PTP1B−/− islets still exhibited improved graft revascularization. Another possibility was that islets lacking PTP1B released more proangiogenic signals than control islets. Our finding that PTP1B−/− islet grafts expressed and secreted more VEGF-A, an angiogenic cytokine that stimulates extraembryonic blood vessel formation (15, 16, 44), than control grafts favors this notion. One may suggest that a similar effect on revascularization could be obtained using strategies to overexpress or directly administer VEGF-A to islets, as previously described (1719). However, these therapies induced hyper-revascularization of grafts associated with a loss of β cell mass and, consequently, impaired graft function. Thus, our approach differs from previous approaches and offers a major advantage, as we demonstrated that improved revascularization, by modulating endogenous VEGF-A expression, was not associated with deleterious effects on β cell number or function in grafts.

Cells express and secrete VEGF-A in response to different stimuli, such as hypoxia (43, 4547) and nutrient deprivation (22). Our data showed that the up-regulation of VEGF-A in islets lacking PTP1B was not associated with increased HIF1α, a major hypoxia sensor. A study conducted by Arany and colleagues (22) described that, in response to ischemia or nutrient deprivation, activation of VEGF-A could take place through HIF1-independent up-regulation of PGC1α and coactivation of ERRα. Here, we showed that PTP1B−/− islets had higher expression of Ppargc1a and Esrra mRNAs and their respective proteins, suggesting that this axis may be responsible for increased Vegfa expression. Down-regulation of PTP1B had similar effects in human islets as in murine islets, which is in agreement with the predicted conservation of 6 of 11 binding sites recognized by ERRα between the promoters of the mouse and human VEGF-A coding genes (22). We also showed that the up-regulation of PGC1α/ERRα in islets was enhanced in response to nutrient deprivation. In this regard, it is plausible to predict that, before maximum revascularization, islets transplanted into the anterior chamber of the eye are exposed to nutrient deprivation (4749); the aqueous humor of the eye that serves as the graft milieu, although capable of sensing blood plasma changes such as those in glucose, presents lower protein content than plasma (5052).

On the basis of all these data and data in previous published studies, we propose a model whereby VEGF-A expression is enhanced via the PGC1α/ERRα axis in transplanted PTP1B−/− islets (fig. S6). Briefly, under nutrient deprivation conditions, the absence or down-regulation of PTP1B enhances the up-regulation of the gene encoding PGC1α by mechanisms that remain to be elucidated. In turn, PGC1α increases the expression of the orphan nuclear receptor ERRα. It is known that PGC1α/ERRα dimers recognize and bind several conserved sites at the Vegfa gene promoter, thus inducing the expression of this gene (2224). Enhanced Vegfa gene expression results in an increase in both protein and secretion of VEGF-A by islets. Secreted VEGF-A will interact with its receptor, VEGFR2, in ECs, activating the signaling pathway of angiogenesis toward the islets (26, 27).

We propose that better graft revascularization contributes to enhanced survival and restoration of endogenous insulin production in recipients of PTP1B−/− islet grafts. However, because we used a loss-of-function genetic mouse model as islet donor, we cannot rule out that constitutive loss of PTP1B might also have influenced intrinsically the endurance and function of transplanted PTP1B−/− β cells in our study. In any case, if we contemplate the clinical translation of this strategy, then we should aim at inhibiting PTP1B during a limited period of time, for example, at preimplantation or early transplantation stages, when the proangiogenic effects of VEGF are required. In this scenario, the potential effects of PTP1B blockade on β cell physiology, if any, would be also temporary. In this regard, we found that ex vivo treatment of human islets with RNA interference approaches led to effective down-regulation of PTP1B, increased VEGF-A production in vitro, and enhanced early graft revascularization. Future studies will be required to fully evaluate the impact of transient PTP1B inhibition on long-term human islet graft function both in healthy and in diabetic mouse models.

In conclusion, we demonstrate that loss or down-regulation of PTP1B potentiates VEGF-A production in mouse and human islets and that this effect is associated with increased graft revascularization. Our results support the notion that PTP1B may represent a molecular target to improve islet graft revascularization. Our findings represent a proof of concept that could lead to improved future outcomes, eliminating an important stumbling block to islet transplantation as a means of effectively treating type 1 diabetes.


Study design

The primary objective of this study was to determine whether the absence or the down-regulation of PTP1B improved islet graft revascularization and survival. In vivo studies were performed by transplanting islets into the anterior chamber of the eye of diabetic male mice. Animal handling and care procedures were fulfilled by trained scientists of Federation of European Laboratory Animal Science Associations and in accordance with the local ethics committee, the Spanish royal decree 214/1997, and the European Directive 2010/63/EU. We based our group sizes on power analysis to achieve 80% likelihood of detection with a type 1 error of 0.05, an effect size of 50% and a 30% SD. All animals that underwent the diabetes induction protocol that presented glycemia below 250 mg/dl for three consecutive days or presented ≥20% weight loss were excluded from the study and euthanized. We did not use randomization or blinding approaches to allocate the remaining animals into the experimental groups. In vitro experiments were performed by randomly allocating isolated mouse islets (primary culture) into groups according to their genotype. Experiments involving human islets were performed in agreement with the local ethics committee [Centre Hospitalier Universitaire (CHU), Montpellier] and the institutional ethical committee of the French Agence de la Biomédecine [Déclaration de Collection (DC) nos. 2014-2473 and 2016-2716]. Informed consent was obtained for all donors. For each experiment, sample size reflects the number of independent biological replicates and is provided in the figure legend.


Animals were maintained on standard light/dark cycle; food and water were provided ad libitum. PTP1B transgenic mice (wild type, PTP1B+/+; knockout, PTP1B−/−) were obtained from Abbot Laboratories (53). Experiments were performed with 8- and 9-week-old male mice littermates, maintaining the genetic background of 129/SvJxC57Bl6/J. Genotyping was performed as described in Supplementary Methods. Male BALB/c mice were acquired from Charles River Laboratories (London, England) at the age of 6 weeks, and experiments were performed at age of 8 weeks. Diabetes was induced in BALB/c mice by STZ (54, 55) as detailed in Supplementary Methods.

Isolation of mouse pancreatic islets and culture

Islets were isolated by standard procedures using collagenase digestion (see Supplementary Methods). Isolated islets were cultured in RPMI 1640 medium (Sigma-Aldrich) containing 11 mM glucose, l-glutamine (0.3 g/liter), NaHCO3 (2 g/liter), 10% heat-inactivated fetal bovine serum (FBS; HyClone), and penicillin-streptomycin (100 U/ml of penicillin and 100 pg/ml of streptomycin; GE Healthcare Life Sciences). The islets were cultured at 37°C under a 5% CO2 and 95% air-humidified atmosphere. Characterization of PTP1B−/− and PTP1B+/+ islets [gene expression, immunohistochemistry (IHC), immunoblotting, and insulin secretion] was performed using 2-day cultured islets. For transplantation experiments, islets were cultured for 2 days before implantation, where the first day was for islet recovery from isolation and the following day for recovery from labeling protocol (see islet labeling). Because it is known that IECs decrease over time in culture (14, 41), PTP1B−/− and PTP1B+/+ islets were cultured for 7 days after isolation to eliminate IECs. The culture medium used was the same as used in standard culture and was changed every 2 days. Nutrient deprivation conditions (22) involved culturing islets at 37°C in 5% CO2 and 95% air-humidified atmosphere for 2 days in HBSS (Sigma-Aldrich) with glucose (2 g/liter), penicillin (100 U/ml), and streptomycin (100 pg/ml).

Islet labeling and allotransplantation

Diabetic male BALB/c mice were transplanted with 200 islets isolated from PTP1B−/− or PTP1B+/+ mice (suboptimal allotransplantation), in the anterior chamber of the eye. Before transplantation, islets were labeled with the long-term tracer for viable cells CFDA SE (Invitrogen), which passively diffuses through the membrane of viable cells and emits fluorescence after intracellular esterases cleave its acetate groups (56). Islets were washed with Dulbecco’s phosphate buffer saline (PBS; Sigma-Aldrich) with 0.1% bovine serum albumin (BSA) and then incubated at 37°C in a 5% CO2 and 95% air-humidified atmosphere for 15 min in a 10 μM CFDA SE–PBS/BSA dilution. CFDA SE–loaded islets were washed and cultured in standard medium for 1 day before being transplanted. Transplantation was performed as described (33). Detailed procedure is provided in Supplementary Methods.

Physiological studies: Weight, glycemia, and glucose tolerance test

During the 28-day follow-up period, weight and nonfasting blood glucose concentration were measured within the same schedule at the indicated days. Blood glucose was measured by collecting blood from the tail vein directly to the glucometer strap. Glucose tolerance test was performed after a 6-hour fast. Mice were injected intraperitoneally with d-glucose (2 g/kg). Blood glucose was measured 0, 15, 30, 60, 90, and 120 min after the injection; in-parallel blood was collected from the tail vein into a capillary blood collection system (Microvette) to analyze plasma insulin concentration by ELISA (Mercodia).

In vivo revascularization and cell death imaging

Functional graft revascularization and cell death were assessed in vivo 7, 15, and 28 days after transplantation. Briefly, the animals were anesthetized intraperitoneally with a ketamine-xylazine mix (100 and 7.5 mg/kg) and received an intravenous injection of a mix composed of RITC-dextran (100 mg/kg), molecular mass of 70 kDa (Sigma-Aldrich) to assess functional vasculature, PI (250 μg/kg; Invitrogen) to assess cell death, and Hoechst 33258 (12 mg/kg; Invitrogen) as a nuclear marker. Anesthetized animals were transferred to the microscope stage with the operated eye positioned in a cover glass with a drop of carboxymethylcellulose sodium, in the direction of the objective (40× water immersion objective). A microscope incubator chamber maintained the adequate temperature. Images of the grafts were acquired every 0.23 μm in a length of 50 μm using a two-photon laser at 900 nm, with automated motion artifact correction (Leica SP5 TPLSM, Leica Microsystems). Images were collected for posterior analysis with ImageJ v1.50d software [Wayne Rasband, National Institutes of Health (NIH)]. At the end of the experiment, animals were placed under supervision in a warm environment until full recovery or euthanized to perform enucleation of the eye for posterior IHC analysis.

IF staining in paraffinized eye sections and in whole islets

Eyes containing the grafts were fixed overnight in 2% paraformaldehyde and then dehydrated with ethanol gradient, cleared with xylol, and paraffin-embedded. Four-micrometer-thick eye sections (15 μm apart) underwent a standard IHC-IF method for paraffin sections. Images were acquired using a Leica DMR HC epifluorescence microscope (Leica Microsystems). IHC-IF of the whole islet was performed as described (57). For each islet, optical section images starting at the peripheral cell layers were acquired using a Leica TCS SPE confocal microscope (Leica Microsystems). Images were taken every 5 μm (for PECAM-1 labeling) or 10 μm, for a total of 60 μm per islet, using a 40× oil immersion objective. The 405-, 488-, and 532-nm lasers were used, and settings were maintained unaltered between islets and conditions in each experiment. IF studies in whole islets offer several advantages over traditional methods such as constructing a three-dimensional mapping of IECs networks by optically sectioning islets with a confocal microscope (58). Primary antibodies used were as follows: guinea pig anti-insulin (1:1000 dilution; Dako), mouse anti-glucagon (1:100; Dako), rabbit anti-PECAM-1 (1:20; Abcam), rabbit anti–VEGF-A (1:100; Abcam), and rabbit anti-cleaved Caspase-3 (1:400; Cell Signaling Technology). Secondary antibodies used were as follows: Alexa Fluor 555 anti–guinea pig, Alexa Fluor 555 anti-rabbit and Alexa Fluor 288 anti-mouse at 1:500 dilution, and Alexa Fluor 488 anti-rabbit at 1:250 dilution (Thermo Fisher Scientific). Hoechst 33258 (1:500; Invitrogen) or an antifade mounting medium with 4′,6-diamidino-2-phenylindole (Thermo Fisher Scientific) was used to mark nuclei. Positively stained cells and pixel areas were quantified using ImageJ v1.50d software (NIH).

RNA isolation and quantitative PCR analysis

Total RNA was extracted, reverse-transcribed, and mRNA expression–quantified by qRT-PCR as previously described (31, 57). Gene expression was normalized against Tbp or ACTB as endogenous controls for mouse and human islets, respectively. Results were expressed relative to expression in control samples (given arbitrarily the value of 1). Primer sequences are provided in Supplementary Methods.

Immunoblotting and VEGF-A secretion

To prepare whole-cell extracts, islets were lysed in radioimmunoprecipitation assay buffer [tris-HCl (50 mmol/liter) (pH 7.5), EDTA (5 mmol/liter), NaCl (150 mmol/liter), 1% Triton X-100, 0.1% SDS, sodium fluoride (10 mmol/liter), and sodium deoxycholate] supplemented with phosphatase and protease inhibitor cocktails (Roche). Islet lysates were frozen and thawed twice in three cycles, followed by ultrasonication with three short burst cycles of the 30 s at 20 kHz (20,000 cycles/s). Nuclear extracts were obtained by bursting islets with a hypotonic buffer [10 mM Hepes, 10 mM KCl, 0.1 mM EDTA, and 0.1 mM EGTA (pH 8)], with 0.05% NP-40. After centrifugation, the nuclear pellet was incubated with a hypertonic buffer (20 mM Hepes, 400 mM NaCl, 10 mM EDTA, 10 mM EGTA, and 20% glycerol) to obtain nuclear protein extracts. All protein extracts were stored at −80°C until analysis.

Protein was quantified with the Lowry protein assay kit (Bio-Rad). A total of 20 μg of protein extract was used for each replicate. Proteins were separated in a precast 4 to 15% gradient gel (Bio-Rad) and transferred onto a polyvinylidene difluoride membrane. The membranes were blocked for 1 hour with 0.05% Tween 20 and 5% nonfat dry milk and then incubated overnight at 4°C with antibodies against HIF1α (1:1000; Abcam; band, 120 kDa), PGC1α [1:1000; Abcam; band, 92 (105) kDa], ERRα (1:1000; Thermo Fisher Scientific; band, 55 kDa), and PTP1B (1:500; Upstate Biotechnology; band, 49 kDa). LAMIN-B1 (1:1000; Cell Signaling Techonology; band, 66 kDa) and αTUBULIN (1:1000; Cell Signaling Techonology; band, 52 kDa) were used as loading controls. Protein bands were visualized using the Pierce ECL Western blot substrate (Thermo Fisher Scientific) and analyzed using ImageJ v1.50a software.

VEGF-A secretion was measured by culturing islets for 48 hours in complete medium without FBS supplementation or directly under nutrient deprivation conditions (HBSS). Islets and culture medium were collected. The protein content from the culture medium was concentrated by centrifugal ultrafiltration (3 kDa; Merck Millipore). VEGF-A secretion and islet VEGF-A content were quantified by ELISA (Abcam).

In vitro insulin secretion and pancreas insulin content

In vitro glucose–induced insulin secretion was studied in separate batches of eight islets by static incubation assays as previously described (59). To determine total pancreatic insulin content, the harvested pancreata were homogenized in acid alcohol and extracted overnight at 4°C. The solution was then centrifuged to remove tissue in suspension and neutralized. Insulin concentration was measured by ELISA (Mercodia).

Human islet isolation and culture

Human islets were obtained from five cadaveric donors (males and females) with an average age of 60 years (±12 years) and body mass index of 28.8 kg/m2 (±2.5 kg/m2) (table S1). Isolated islets were prepared by collagenase digestion, followed by density gradient purification at the Laboratory of Cell Therapy for Diabetes (Hospital Saint-Eloi, Montpellier, France), as previously described (60). After reception, human islets were maintained in culture, for at least 3 days, using RPMI 1640 with 5.6 mM glucose, 10% FBS, and antibiotics. Assessment of batch purity and viability was performed as described in Supplementary Methods.

RNA interference

For in vitro experiments, PTP1B knockdown in human islets was achieved by transfection of a PTPN1 SMARTpool Accell siRNA following the manufacturer’s instructions with minor modifications. Human islets were preincubated in a dilute solution of trypsin (50 mg/ml) for 1.5 min at 37°C to increase transfection efficiency. In a pilot experiment, diffusion of siRNA throughout the islets was evaluated 72 hours after transfection of a nontargeting Accell siRNA labeled with 6-carboxyfluorescein (6-FAM) (Dharmacon) using confocal microscopy (laser at 488 nm) and bright field. Islet autofluorescence was also visualized and set as a negative control, using the same excitation and emission settings for FAM-siRNA assessment (517 nm). Optical sections (10 μm apart) were acquired and analyzed using ImageJ v1.50d software. After a 3-day culture with siRNA, the islets were collected, and RNA and protein were extracted as described previously. Gene expression was assessed by qRT-PCR, and protein content was assessed by immunoblotting.

For transplantation experiments, lentiviral particles carrying shRNAs were used as this strategy sustains better the down-regulation of a gene in islets (6163), without affecting islet function (6163) when compared with siRNA transfection. A pool of four independent lentiviruses carrying distinct shRNAs specific for human PTP1B (sh-PTP1B LV) and the control lentivirus carrying a nonspecific scrambled shRNA (sh-scrambled LV) were acquired from OriGene Technologies (Rockville, USA). In addition to the shRNA, all lentiviruses carried a CMV-driven tGFP to monitor transduction efficiency. Human islets were pelleted at 50g for 2 min and then incubated with 1000 μl of trypsin (250 mg/liter)–EDTA (0.48 mM) solution for 3 min in a cell culture incubator (37°C in 5% CO2). The islet suspension was carefully pipetted up and down. After adding 1000 μl of complete RPMI 1640 [glucose (1 g/liter), 10% FBS, and penicillin-streptomycin], islets were pelleted by centrifugation at 100g for 1 min. Batches of 150 islets were then resuspended in a serum-free RPMI (final volume, <100 μl) in polystyrene round-bottom tubes. Lentiviruses were added at 20 plaque-forming units (PFUs) per cell, assuming 1000 cells per islet. The final volume did not exceed 300 μl, and virus concentration used was 1.7 × 104 PFU/μl. Polybrene (5 μg/ml) was added for infection. Islets were incubated overnight in a cell culture incubator and then transplanted or incubated in 12-well suspension plates (37°C in 5% CO2) to assess knockdown efficiency and VEGF-A secretion.

Statistical analysis

GraphPad Prism (Prism version 6.00 for Windows, GraphPad Software, was used for analysis. Data are depicted as means ± SEM unless otherwise specified. An unpaired Student’s t test was performed to analyze variances between two populations. Two-way ANOVA (Bonferroni’s post hoc test) was used to compare multiple populations.



Fig. S1. In vitro glucose–induced insulin secretion by PTP1B−/− and PTP1B+/+ mouse islets.

Fig. S2. Characterization of PTP1B−/− and PTP1B+/+ mouse islets after 7 days in culture.

Fig. S3. Immunolocalization of VEGF-A in PTP1B−/− and PTP1B+/+ mouse islets.

Fig. S4. Characterization of human islet preparations.

Fig. S5. Assessment of the uptake of siRNA by human islets.

Fig. S6. Proposed model for the effect of PTP1B inhibition on islet graft vascularization.

Table S1. Human islet donor information.

Data file S1. Raw data.

References (64, 65)


Acknowledgments: We thank Y. Esteban (Institut d’Investigacions Biomèdiques August Pi i Sunyer, IDIBAPS/Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas, CIBERDEM, Barcelona, Spain) for excellent technical support in immunoblotting analysis. We thank K. Katte (IDIBAPS/CIBERDEM, Barcelona, Spain) and B. Sinclair for English proofreading the manuscript. We are grateful to all members of the Gomis-Gasa group for plentiful and rich discussions. Funding: This work has been supported by projects PI13/01500 and PI16/00774 (to R. Gasa and R. Gomis) integrated in the Plan Estatal de I+D+I and cofinanced by ISCIII–Subdirección General de Evaluación and Fondo Europeo de Desarrollo Regional (FEDER, “A way to build Europe”); grant 2014 SGR659 (to R. Gomis) from the Generalitat de Catalunya; Recerca Bàsica grant (to R.Gomis) from the Academy of Medical and Health Science of Catalonia and the Balearic Islands; and Cátedra Astra Zeneca. H.F. was partly supported by DiabetesCero Foundation grant. A.L.C.F. was supported by a Doctoral fellowship from the Consejo Nacional de Ciencia y Tecnologia (CONACYT) from Mexico. Author contributions: H.F. conducted all experiments. A.L.C.F. performed and provided assistance with mouse and human islets experiments. A.G. and R.M. provided assistance with mouse experiments and IHC studies in paraffinized eye sections. C.B. and A.W. performed isolation of human islets and provided assistance with human islet experiments. R.M. and R.F.-R. performed experiments. H.F., R.M., and R. Gomis conceived the project. H.F., R.M., R. Gasa, and R. Gomis discussed the data. H.F., R. Gasa, and R. Gomis wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper and/or the Supplementary Materials.

Stay Connected to Science Translational Medicine

Navigate This Article