Fig 2: Endothelial cell proliferation is diminished and the number of IGF-1-producing macrophages is decreased in the intestine of human neonates with NEC compared to controls.a, b Sections of human NEC (n = 4) and control (n = 5) intestinal tissues were stained for CD31 (green) and Ki-67 (red). 2–3 images per sample from non-necrotic area were taken at ×20 and proliferating endothelial (CD31+Ki-67+) cells were counted. a Typical images are shown here. b Number of proliferating endothelial cells (represented by each dot) per ×20 image (mean ± SEM), scale bar = 100 µm. c Images of control human neonatal intestinal tissue sections subjected to in situ hybridization using a probe for IGF-1 (green) and counterstained for macrophages using anti-CX3CR1 antibodies (red); bar = 25 (top) and 10 µm (bottom). d, e NEC and control tissue sections were stained by immunofluorescence for IGF-1 (green) and the macrophage marker CX3CR1 (red). d Typical images are shown. Overlay image includes DAPI, scale bar = 100 µm. e Graph represents the number of IGF-1+ CX3CR1+ cells per field (mean ± SEM, 3–4 ×20 fields of non-necrotic area were taken from each sample; n = 4 patients/group). P values were calculated using two-sided Student’s t tests (b and e). ****p < 0.0001. Source data are provided as a Source Data file.

Fig 3: IGF-1-deficient macrophages have decreased proangiogenic properties and Igf-1?M? mice are more susceptible to experimental NEC.a, b 4 × 104 intestinal endothelial cells from mT/mG neonatal pups were co-cultured for 4 days on Matrigel with 2 × 104 CX3CR1+ neonatal intestinal macrophages obtained from IGF-1-sufficient (n = 6) or Igf-1?M? (n = 6) pups. a Typical sprouting images are shown. b Sprout length was measured, each data point indicates the length of a single sprout (mean ± SEM). c, d Intestinal tissue VEGF expression was assessed at D1 (24 h of life) by western blot analysis and relative band density to actin was quantified (mean ± SEM), n = 10 and 12 IGF-1-sufficient or Igf-1?M? groups, respectively. P values were calculated using two-sided Student’s t tests (b, d). e, f IGF-1-sufficient or Igf-1?M? pups were submitted to the neonatal NEC model. e Survival curves are shown, n = 48 and 37 in IGF-1-sufficient or Igf-1?M? groups, respectively, P values were calculated using Gehan–Breslow–Wilcoxon test. f Intestinal histological injury scores are shown, n = 46 and 35 in IGF-1-sufficient or Igf-1?M? groups, respectively, P values were calculated using Chi-square analysis. All plots are the results of three separated experiments combined. *p < 0.05, **p < 0.01, ****p < 0.0001. Source data are provided as a Source Data file.

Fig 4: Blocking IGF-1 signaling decreases intestinal endothelial cell proliferation in vivo; exogenous IGF-1 enhances VEGF protein expression, intestinal endothelial cell proliferation and preserves intestinal microvascular density during experimental NEC.a, b 24-hour-old wild-type C57BL/6 pups were treated with the IGF-1R inhibitor PPP (n = 8) or vehicle control (n = 7) or submitted to the NEC protocol for 24 h (n = 7). Intestinal LP cells were isolated and the percentage of proliferating endothelial cells (Ki-67+CD31+CD45-/CD31+CD45-) was determined by flow cytometry. a Typical FACS plots are shown. b Graph represents the results of three experiments combined (mean ± SEM). c–h 1-day-old neonatal mice were injected with IGF-1 (25 µg/kg, i.p., twice) or vehicle control, with a first dose 2 h prior to NEC initiation and a second dose 12 h later (Dam-fed littermates as controls) then submitted to experimental NEC for 24 h (c–f) or 48 h (g, h). c, d Intestinal tissues were assessed for VEGF and VEGFR2 proteins by western blot (c) and band densitometry is shown (d, mean ± SEM), n = 6–9/group (see panel or Source Data file for exact n number). e, f Tissue sections were stained with Ab against Ki-67 and endomucin. e Typical immunofluorescence images. f Graph represents average proliferating endothelial cells (Ki-67+ endomucin+ cells, arrows in e) counted per ×20 field (mean ± SEM, three fields were taken from each sample), n = 9–12/group (see Source Data file for exact n number). g, h pups were perfused intracardially with WGA-Alexa Fluor 647. Intestinal submucosal vascular networks were imaged, and vascular density was assessed using Photoshop (×10), n = 4–5/group (see dots in panel). P values were calculated using one-way ANOVA followed by Turkey–Kramer multiple-comparison test (b, d, f, and h). i, j Neonatal mice was submitted to the NEC model and treated with IGF-1 or control twice daily (see method section). i 60-hour-survival curve is shown (n = 28 in NEC and 30 in NEC/IGF-1 group), P value were calculated using Gehan–Breslow–Wilcoxon test. j Intestinal injury histological scores (n = 28 in NEC and 29 in NEC-IGF-1 group), P value were calculated using Chi-square test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Source data are provided as a Source Data file.

Fig 5: Serum and intestinal IGF-1 are developmentally regulated during the perinatal period in mice.Pups from each litter were randomly assigned to different time-points. Serum and intestinal tissues were collected from mice at embryonic day 17.5 (E17.5), day of life 0, 1, 2, 3, 7, 14, and 21. a IGF-1 serum level was measured by ELISA, n = 4–12/time point (exact n at each time point see dots in the panels and Source Data file.), data represent results of three experiments combined (mean ± SEM). b Intestinal IGF-1 and IGF-1R were examined by western blot analysis. Data represent the results of three experiments. n = 2/time point for every three separate experiments (mean ± SEM). One of the three experiments did not include E17.5 time point. Gel images are included in Source Data file.