Fig 2: Targeting IGFBP3/TMEM219 signaling protects beta cells in the NOD mouse in vivo.a Bar graph depicting Tmem219 mRNA expression analyzed by qRT-PCR in pancreata of NOD mice (n = 3, 5, and 4 per group). b Representative immunohistochemical TMEM219 staining in serial pancreatic islet tissue sections from NOD mice (n = 3). ×40 original magnification, scale bar, 50 µm. c, d Bar graph depicting peripheral IGFBP3 (c) and IGF-I (d) levels measured in plasma obtained from 8-week-old B6 (n = 10) mice, in pre-diabetic (4-week and 10-week-old, n = 4) and diabetic (hyperglycemic, n = 5) NOD mice, in mice resistant to the development of autoimmune diabetes (NOR, n = 3) and in long-term normoglycemic (LtNglc, 26-week-old, n = 4) NOD mice. e Ten-week-old NOD mice were treated with ecto-TMEM219 (0.1 mg/mouse/day for 10 days or 0.1 mg/mouse/day for 10 days and twice per week or were left untreated (n = 20/group), and the incidence of diabetes was then compared using the log-rank (Mantel–Cox) test (**P < 0.01; ***P < 0.001). f Bar graph depicting semi-quantitative analysis of islet area in pancreatic sections obtained from ecto-TMEM219-treated and untreated NOD mice after 14 weeks of treatment (n = 4/group). Data are expressed as a score ranging between 0 and 4. g Bar graph showing reduction of peripheral IGFBP3 levels in ecto-TMEM219-treated and untreated NOD mice after 14 weeks of treatment (n = 5). h Bar graph depicting peripheral insulin levels measured in ecto-TMEM219-treated (n = 4 and 7, respectively) and untreated NOD mice after 14 weeks of treatment (n = 4). i Bar graph showing peripheral IGF-I levels in ecto-TMEM219-treated and untreated NOD mice after 14 weeks of treatment (n = 5). j Bar graph representing islet infiltration detected in pancreatic sections obtained from ecto-TMEM219-treated as compared to untreated NOD mice after 14 weeks of treatment (n = 3/group). Data are expressed as a score ranging between 0 and 4. k Representative H&E staining in serial pancreatic islet tissue sections obtained from 24-week-old NOD mice treated with prolonged ecto-TMEM219 or from untreated mice shown as control (n = 3/group). ×10 original magnification, scale bar, 300 µm. l Bar graph showing IFN-?-producing cells upon in vitro re-stimulation with BDC2.5 or IGRP peptides detected by ELISpot in ecto-TMEM219-treated as compared to untreated NOD mice after 14 weeks of treatment (n = 10/group, n = 6 in the short course ecto-TMEM219-treated group). m, n, o Bar graphs representing the percentage of CD4+CD44hiCD62Llo (n = 4/group), CD8+CD44hiCD62Llo (n = 4/group), and CD4+CD25+Foxp3+ cells (n = 3/group) detected in splenocytes of ecto-TMEM219-treated as compared to untreated NOD mice after 14 weeks of treatment. p, q Newly hyperglycemic NOD mice were treated with Ecto-TMEM219 or were left untreated and the incidence of diabetes was then compared using the log-rank (Mantel–Cox) test (***P < 0.001). r Representative H&E staining in serial pancreatic islet tissue sections obtained from newly hyperglycemic NOD mice, treated with ecto-TMEM219 or from untreated mice shown as control (n = 3/group). ×20 original magnification, scale bar, 100 µm. Data are expressed as mean ± standard error of the mean (SEM) unless otherwise reported. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 by one-way ANOVA followed by Bonferroni’s post hoc test, log-rank Mantel–Cox test, Kruskal–Wallis test adjusted for multiple comparisons. mRNA expression was normalized to Gapdh. The data shown in (b), k1–k2, r1–r2 are the representative result from three independent experiments. Source data are provided as a Source Data file. NOD nonobese diabetic mice, LtNglt long-term normoglycemic NOD mice, NOR nonobese diabetes-resistant mice, hyper hyperglycemic, mAb monoclonal antibody, Arb. units arbitrary units.

Fig 3: IGFBP3/TMEM219 signaling triggers beta-cell apoptosis.a Bar graph quantifying apoptosis in human islets cultured with/without IGFBP3 and in the presence/absence of ecto-TMEM219 (n = 3). The experiment was run in duplicate. b, c Bar graph and representative flow plots quantifying cell death of insulin-positive cells in human islets cultured with/without IGFBP3 and in the presence/absence of ecto-TMEM219 (n = 3). d Bar graph (percentage of double-positive cells analyzed per field) depicting apoptosis (green, Apoptag) of insulin-positive (red) cells in human purified islets cultured with/without IGFBP3 and in the presence/absence of ecto-TMEM219 (n = 5). e, f Anecdotical picture of INS (red, left panels) and APOPTAG (green, middle panels) immunofluorescence in purified human islets of non-diabetic donors cultured with/without IGFBP3. Merge images are in right panels. Original magnification ×40, scale bar 25 µm. g Scatter plot representing the transcriptome profile of relevant-apoptotic genes in human islets cultured with/without IGFBP3 (n = 3). The experiment was run in triplicate. The complete dataset of differentially expressed genes is reported in Supplementary Table 5. h Bar graph representing normalized mRNA expression of CASP8 quantified by qRT-PCR in flow-sorted insulin-positive cells obtained from human islets cultured with/without IGFBP3 (n = 3). i Bar graph representing CASP8 mRNA expression quantified by RNAseq in laser-captured islets obtained from human non-diabetic subjects (CTRL, n = 18), patients with T1D (n = 20), patients at risk for T1D (AutoAb+, n = 12) and patients with T2D (n = 8). mRNA expression is shown as fold change as compared to CTRL. j Bar graph quantifying Cleaved Caspase 8 by ELISA in human islets cultured with/without IGFBP3 and in the presence/absence of ecto-TMEM219 (n = 3). k Grouped graph depicting targeted gene expression analysis of insulin secretory machinery using qRT-PCR in insulin-positive cells flow-sorted from human purified islets (n = 3) cultured with/without IGFBP3. mRNA was normalized to ACTB, and the experiment was performed in triplicate. l Bar graph depicting baseline insulin secretion measured in the supernatant of human islets cultured with/without IGFBP3 and in the presence or absence of ecto-TMEM219 (n = 4). m Bar graph showing the stimulation index calculated by measuring insulin secretion in the supernatant of human islets upon glucose stimulation and cultured with/without IGFBP3 and in the presence or absence of ecto-TMEM219 (n = 4). n, o Bar graphs confirming the absence of TMEM219 mRNA (n) and protein (o) expression upon transient silencing in human islets (n = 3). The experiment was performed in duplicate. p, q Bar graphs depicting normalized expression (fold change) of CASP8 (p) and INS (q) quantified by qRT-PCR in purified human islets, in which TMEM219 was targeted via siRNA delivery, compared to islets in which TMEM219 targeting was not performed (n = 4 samples per group of treatment). r Heatmap showing fold activation of AKT, Caspase 8, and ERK1/2 in human islets cultured with/without IGFBP3 and/or in which TMEM219 was targeted via siRNA delivery (n = 3 samples/group). Fold activation was assessed by immunoblotting and normalized to that measured in medium-cultured islets. s, t Bar graphs depicting normalized expression (fold change) of CASP8 (s) and INS (t) quantified by qRT-PCR in purified human islets exposed to medium or to sera of CTRL, T1D, and T2D patients (pooled from n = 5 donors per group) and in the presence/absence of ecto-TMEM219 (n = 3 islet preparations per group of treatment). Experiments were performed in duplicate. u Representative picture of confocal microscopy analysis (scale bar 10 µm, ×63 original magnification) depicting colocalization of TMEM219 (green) and serum IGFBP3 (red) in a human beta-cell line (Blox-5) cultured with T1D serum. Cells were stained with DAPI (blue) and immunolabeled with anti-TMEM219 (green) and anti-IGFP3 Abs (red). v Representative blot showing TMEM219 immunoprecipitation (IP) in human beta-cell line (Blox-5) cultured with serum naturally enriched in IGFBP3 (T1D serum) pooled from n = 5 donors. Expression of IGFBP3 is shown. Lane 1: rh-IGFBP3. Lane 2: IP with beads alone. Lane 3: IP with TMEM219 Ab. Data are expressed as mean ± standard error of the mean (SEM) unless otherwise reported. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 by one-way ANOVA followed by Sidak/Bonferroni’s post hoc test, two-sided t test, Kruskal–Wallis test adjusted for multiple comparisons. mRNA expression was normalized to ACTB. The data shown in (e, f, u, v) are the representative result from three independent experiments. Source data are provided as a Source Data file. T1D type 1 diabetes, T2D type 2 diabetes, AutoAb+ positive for autoantibodies, INS insulin, CASP8 Caspase 8, siRNA small interfering RNA, ecto-TMEM219 newly generated recombinant protein based on TMEM219 extracellular portion, Arb. Units arbitrary units.

Fig 4: Targeting IGFBP3/TMEM219 signaling protects beta cells in preclinical models of diabetes in vivo.a Bar graph showing TMEM219 protein expression (ELISA) assessed in the whole pancreas and dissociated pancreatic islets obtained from B6 mice (n = 4). The experiment was performed in duplicate. b Representative immunoblot of TMEM219 protein expression in murine islets (ß-actin was used as a control, n = 3). c Bar graph comparing Tmem219 and other IGFBP3 receptors (Lrp1, Tgf-ß R1, and R2) gene expression quantified by qRT-PCR in purified murine islets of B6 mice (n = 3). d Quantitative bar graph (percentage of double-positive cells analyzed per field) depicting apoptosis in purified murine (B6) islets cultured with/without IGFBP3 in the presence or absence of ecto-TMEM219 (n = 3). e, f Bar graphs representing Ins (e) and Casp8 (f) mRNA expression analyzed by qRT-PCR in murine islets cultured with/without IGFBP3 and in the presence or not of ecto-TMEM219 (n = 4 in (f) and n = 5 in (e)). The experiment was performed in duplicate. g Bar graph depicting fasting blood glucose levels measured in IGFBP3-treated and untreated B6 mice (n = 5/group, day 7). (h). Bar graph depicting peripheral IGFBP3 levels measured in plasma obtained from 8-week-old B6 mice fed a high-fat diet (B6-HFD, n = 10) or left untreated (B6, n = 8). i Box plots representing blood glucose levels measured in B6 mice fed a high-fat diet (B6-HFD) treated with (red lines)/without (black lines) ecto-TMEM219 (n = 8–10/group). j Box plots representing body weight measured in B6 mice fed a high-fat diet (B6-HFD) treated with (red dots)/without (black dots) ecto-TMEM219 (n = 8–10/group). In (i) and (j) box plots include the median line, the whiskers indicate the minimum and maximum value, and the box of the box plot illustrates the upper and lower quartile. k Bar graph representing peripheral IGFBP3 levels in ecto-TMEM219-treated and untreated B6-HFD mice (n = 8 and 10, respectively). l Line graph showing blood glucose level measured in B6 mice injected with multiple low-dose of streptozotocin (ldSTZ, 50 mg/Kg) and treated with ecto-TMEM219 or left untreated (n = 5). m Bar graph showing blood glucose area under the curve (AUC) measured at the intraperitoneal glucose (1 g/Kg) tolerance test (IPGTT) in B6 mice injected with multiple low-dose of streptozotocin and treated with ecto-TMEM219 or left untreated at day 10 (n = 5 and 4, respectively). n Bar graph showing IGFP3 plasma levels measured in B6 mice injected with multiple low-dose of streptozotocin and treated with ecto-TMEM219 or left untreated at day 10 (n = 5). o Representative H&E staining in serial pancreatic islet tissue sections obtained from B6 mice injected with multiple low dose of streptozotocin and treated with ecto-TMEM219 or left untreated (n = 3/group). ×20 original magnification, scale bar, 100 µm. p Targeting strategy to generate the TMEM219fl/fl mouse by using the Cre-loxP strategy. q Line graph showing blood glucose level measured in Beta-TMEM219-/- and in wild-type (WT) mice injected with multiple low-dose of streptozotocin (n = 5). r Bar graph showing blood glucose area under the curve (AUC) measured at the IPGTT in Beta-TMEM219-/- and in wild type (WT) mice injected with multiple low-dose of streptozotocin (n = 5). s Bar graph showing IGFP3 plasma levels measured in Beta-TMEM219-/- and in wild-type (WT) mice injected with multiple low doses of streptozotocin (n = 4 and 5, respectively). t Representative H&E staining in serial pancreatic islet tissue sections obtained from WT and Beta-TMEM219-/- injected with multiple low dose of streptozotocin (n = 3/group). ×20 original magnification, scale bar, 100 µm. In (i, j, l, q) statistical analysis compared blood glucose levels or weight expressed as mean ± SEM (l, q) between the two groups at each timepoint. Data are expressed as mean ± standard error of the mean (SEM) unless otherwise reported. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 by one-way ANOVA followed by Bonferroni’s post hoc test, two-way ANOVA, two-sided Mann–Whitney U test or two-sided t test. mRNA expression was normalized to Gapdh. The data shown in o1–o2, t1–t2 are the representative result from three independent experiments. Source data are provided as a Source Data file. B6 C57BL/6J mice, HFD high-fat diet, hyper hyperglycemic, mAb monoclonal antibody, Arb. units arbitrary units, Beta-TMEM219-/- mice in which TMEM219 was deleted in beta cells, WT wild-type mice in which TMEM219 has not been genetically deleted, STZ streptozotocin, ldSTZ low-dose streptozotocin 50 mg/Kg injected for 5 days.

Fig 5: The death receptor TMEM219 is expressed in islet beta cells.a The transcriptome profile of genes encoding islet surface proteins was screened using RNA-seq of islets isolated from non-diabetic donors (n = 4). A full list of genes analyzed is reported in Supplementary Data 2. b Quantification of TMEM219 relative expression as compared to other relevant receptors in human islets by targeted qRT-PCR (n = 4). c Representative image of TMEM219 (green) and INS (red) co-expression in purified human islets of non-diabetic donors. Original magnification ×40, scale bar 25 µm. Merge is in the left panel. d Bar graph representing the expression of TMEM219 mRNA on flow-sorted insulin-positive and -negative cells obtained from purified human islets isolated from non-diabetic donors (n = 3). e Bar graphs comparing expression of TMEM219 and other IGFBP3 putative receptors (LRP1, TGF-ß R1, and TGF-ß R2) analyzed by qRT-PCR in purified human islets (n = 3). f Representative immunoblot of TMEM219 protein expression in human islets (with ß-actin as a control). g, h Representative flow plot and a quantitative bar graph showing TMEM219 expression in insulin-positive and -negative cells detected using flow cytometry in healthy human islets (n = 3). Data are expressed as mean ± standard error of the mean (SEM) unless otherwise reported. **P < 0.01 by two-sided t test. mRNA expression was normalized to ß-actin (ACTB). Experiments were performed in duplicate. The data shown in (c) and in (f) are the representative result from three independent experiments. Source data are provided as a Source Data file. RNAseq RNA sequencing analysis, RPKM reads per kilobase per million, INS insulin, qRT-PCR quantitative real-time polymerase chain reaction.