EGFR signaling promotes β-cell proliferation and survivin expression during pregnancy

Abstract

Placental lactogen (PL) induced serotonergic signaling is essential for gestational β-cell mass expansion. We have previously shown that intact Epidermal growth factor -receptor (EGFR) function is a crucial component of this pathway. We now explored more specifically the link between EGFR and pregnancy-induced β-cell mass compensation. Islets were isolated from wild-type and β-cell-specific EGFR-dominant negative mice (E1-DN), stimulated with PL and analyzed for β-cell proliferation and expression of genes involved in gestational β-cell growth. β-cell mass dynamics were analyzed both with traditional morphometrical methods and three-dimensional optical projection tomography (OPT) of whole-mount insulin-stained pancreata. Insulin-positive volume analyzed with OPT increased 1.4-fold at gestational day 18.5 (GD18.5) when compared to non-pregnant mice. Number of islets peaked by GD13.5 (680 vs 1134 islets per pancreas, non-pregnant vs. GD13.5). PL stimulated beta cell proliferation in the wild-type islets, whereas the proliferative response was absent in the E1-DN mouse islets. Serotonin synthesizing enzymes were upregulated similarly in both the wild-type and E1-DN mice. However, while survivin (Birc5) mRNA was upregulated 5.5-fold during pregnancy in the wild-type islets, no change was seen in the E1-DN pregnant islets. PL induced survivin expression also in isolated islets and this was blocked by EGFR inhibitor gefitinib, mTOR inhibitor rapamycin and MEK inhibitor PD0325901. Our 3D-volumetric analysis of β-cell mass expansion during murine pregnancy revealed that islet number increases during pregnancy. In addition, our results suggest that EGFR signaling is required for lactogen-induced survivin expression via MAPK and mTOR pathways.

Figure 1. E1-DN mice do not increase β-cell mass during pregnancy.

A: The β-cell mass of wild-type mice increases during pregnancy at GD14.5. There is no difference between groups of E1-DN mice, n = 8–13. B: Isolated islets from wild-type (WT) and E1-DN (E1) female virgin mice were cultured for 96 h either with PL 500 ng/ml or in control media (con). BrdU was added to the media for the last 48 h. Quantification of proliferative β-cells, n = 3 in each group. C: Total number of islets per pancreatic area from WT and E1-DN (E1) control and pregnant gestational day 14.5 (GD14.5) mice, n = 4–6 per group. D: The distribution of islet sizes relative to total pancreatic area. Islets were classified into small <12076 μm², medium 12076–35033 μm² and large islets >35033 μm² from wild-type (WT) and E1-DN (E1) control (con) and gestational day 14.5 (GD14.5) pancreata, n = 4–6 per group. E: The islet size distribution corrected for pancreatic weight. AU, arbitrary unit. Bars represent the mean ±SEM for each group. *p<0.05, **p<0.01,***p<0.001, ****p<0.0001 ns = no statistical significance.

Figure 2. OPT analysis of insulin-positive volume and islet number during pregnancy.

A–C: Isosurface rendered OPT images of insulin labeled (red) representative pancreata from control (A), gestational day 13.5 (GD13.5) (B) and gestational day 18.5 (GD18.5) (C) female mice. The pancreas outline is based on autofluorescence of the tissue. D: Graph illustrating the average insulin-positive volume in control (white), GD13.5 (gray) and GD18.5 (black) pancreata, n = 7–8 per group. E: Graph illustrating the average number of islets per pancreas in control (white), GD13.5 (gray) and GD18.5 (black) pancreata, n = 7–8 per group. F–G: The distribution of islet sizes (G) and contribution to total islet volume (F) n = 7–8 per group. Islets were classified into small 0–1000 um³×10³, intermediate 1000–5000 um³×10³ and large islets >5000 um³×10³ from control (white), GD13.5 (gray) and GD18.5 (black) pancreata. Bars represent the mean ±SEM for each group. *p<0.05, **p<0.01, ***p<0.001 ns = no statistical significance.

Figure 3. Validation of OPT method by 2D morphometrical analysis.

The same pancreata that were analyzed by OPT were sectioned and analyzed with routine methods. A: Insulin positive area per total pancreatic area. B: Insulin positive surface area per pancreatic area multiplied by total pancreatic volume. C: Number of islets per pancreatic area multiplied by total pancreatic volume. D: Islets were classified into small <12076 μm², medium 12076–35033 μm² and large islets >35033 μm² and the distribution was calculated per total surface area and further multiplied by total pancreatic weight. AU, arbitrary unit. Bars represent the mean ±SEM for each group, n = 6–8. *p<0.05, **p<0.01. E–F: Representative images of insulin immunohistochemistry (red) from non-pregnant (E) and pregnant GD18.5 (F) pancreas.

Figure 4. Reduced EGFR signaling leads to decreased survivin expression in islets during pregnancy in mice.

A–G: Islets were isolated from wild-type (WT) and E1-DN (E1) virgin (con) and gestational day 13.5 (GD13.5) mice and RT-qPCR performed. Graphs are showing the relative mRNA level of survivin (A), Tph1 (B), Tph2 (C), serotonin receptor Htr2B (D), FoxM1 (E), endogenous mouse Egfr (F) and prolactin receptor (G) n = 4–5 per group. *p<0.05, **p<0.01, ns = no statistical significance.

Figure 5. Survivin immunoreactivity is detectable in islets of pregnant wild-type but not E1-DN mice.

A–H: Double immunohistochemistry for survivin (green) and insulin (red) from wild-type non-pregnant (A), E1-DN non-pregnant (B), wild-type pregnant GD14.5 (C) and E1-DN pregnant GD14.5 (D) mouse. Scale bar 200 μm. E: Magnification of squared area in C. F: Magnification of squared area in D. G: Negative control from wild-type pregnant mouse pancreas with blocking peptide. Scale bar 200 μm. H: Magnification of squared area in G.

Figure 6. PL-induced survivin upregulation is dependent of EGFR, mTOR and MEK.

A–B: Islets were isolated from wild-type mice, stimulated with or without PL (500 ng/ml), gefitinib (2 μM), rapamycin (10 nM), MEK inhibitor PD0325901 (0,5 μM) or EGF (50 ng/ml) and BTC (50 ng/ml) for 96 h and RT-qPCR performed. Graph is showing the relative mRNA level of survivin (A) and Tph1 (B). n = 6–9 per group. Bars represent the mean ±SEM for each group. *p<0.05, **p<0.01 ***p<0.001. C: Proposed model for PL-induced survivin upregulation. Prl-R activation leads to transactivation of EGFR, which activates MAPK and PI3K-Akt-mTOR pathways, that together stimulate survivin gene expression.