Contribution of a non-β-cell source to β-cell mass during pregnancy

Abstract

β-cell mass in the pancreas increases significantly during pregnancy as an adaptation to maternal insulin resistance. Lineage tracing studies in rodents have presented conflicting evidence on the role of cell duplication in the formation of new β-cells during gestation, while recent human data suggest that new islets are a major contributor to increased β-cell mass in pregnancy. Here, we aim to: 1) determine whether a non-β-cell source contributes to the appearance of new β-cells during pregnancy and 2) investigate whether recapitulation of the embryonic developmental pathway involving high expression of neurogenin 3 (Ngn3) plays a role in the up-regulation of β-cell mass during pregnancy. Using a mouse β-cell lineage-tracing model, which labels insulin-producing β-cells with red fluorescent protein (RFP), we found that the percentage of labeled β-cells dropped from 97% prior to pregnancy to 87% at mid-pregnancy. This suggests contribution of a non-β-cell source to the increase in total β-cell numbers during pregnancy. In addition, we observed a population of hormone-negative, Ngn3-positive cells in islets of both non-pregnant and pregnant mice, and this population dropped from 12% of all islets cells in the non-pregnant mice to 5% by day 8 of pregnancy. Concomitantly, a decrease in expression of Ngn3 and changes in its upstream regulatory network (Sox9 and Hes-1) as well as downstream targets (NeuroD, Nkx2.2, Rfx6 and IA1) were also observed during pregnancy. Our results show that duplication of pre-existing β-cells is not the sole source of new β-cells during pregnancy and that Ngn3 may be involved in this process.

Figure 1. The Cre/loxP-based lineage tracing system in the RIP-Tet-off-tdRFP-positive transgenic mice.

A) Administration of doxycycline will not allow translocation of Cre and hence, no tdRFP expression. Removal of doxycycline from the RIP-Tet-off-tdRFP-positive mice allows nuclear translocation of Cre, resulting in expression of tdRFP in all insulin-producing cells (RIP =  rat insulin promoter). B) A schematic of the sources of new β-cells during pregnancy. On the day when pregnancy was detected (i.e. presence of a vaginal plug in the female), doxycycline is administered to turn off RFP expression. All new β-cells that arose from pre-existing β-cells will be RFP⁺ and insulin⁺ (red filled in green circles). All new β-cells that did not arise from pre-existing β-cells will not express tdRFP (white filled in green circles). Blue outlined circles are BrdU-positive cells.

Figure 2. β-cell lineage tracing during pregnancy.

A) In the non-pregnant RIP-Tet-off-RFP mice, 97.2±1.45% of β-cells express RFP. During pregnancy, only 87.0±5.01% of β-cells express RFP, suggesting a non-β-cells source for the new β-cells. “*”: p<0.05 in comparison to non-pregnant mice by student’s t-test. At least 10,000 cells from 4 separate mice were counted at G0, and the same for G12. B) A representative islet from G0 and G12 is shown. Green = insulin, Red = RFP, White = BrdU, Blue = nuclear staining, C) Percentage of β-cells that are BrdU⁺. At least 5,000 β-cells from 4 separate mice were counted at G0, and the same for G12. Black bars = RFP^- (unlabeled) β-cells. White bars = RFP⁺ β-cells. Comparisons were done by two-way ANOVA with a Tukey multiple comparisons test. “*”: p<0.05 in comparison to G0.

Figure 3. Ngn3-positive cells in pancreatic islets during pregnancy.

A) Percentage of cells in the pancreatic islets of Ngn3:EGFP mice stained positive for EGFP (therefore expressing Ngn3, these cells will be denoted as Ngn3-EGFP⁺ in all images) but negative for endocrine hormones drops during pregnancy. The number of Ngn3-EGFP⁺ cells was expressed as a percentage of all the cells in the islets. Islets (outlined in white) were identified by staining for endocrine hormones. At least 10,000 cells were counted from each mouse. N = 3–4 separate mice were sampled at each gestational stage. “*”: p<0.05 in comparison to the non-pregnant (G0) mice by student’s t-test. B) A representative image of an islet at G0, G8 and G14. C) A representative image of TUNEL staining of an islet (outlined) at G6. Islets at postnatal day 2 (P2) and DNase treated pancreas served as controls. Black arrowheads indicate apoptotic (brown) cells. D) mRNA levels of Ngn3, as measured by qRT PCR in islets isolated from mice on days 0, 6, 9, and 15 of pregnancy expressed as mean ± SEM; N = 6 separate mice at each gestational stage. All expression levels are normalized to that of PGK-1 (phosphoglycerate kinase 1). “*”: p<0.05 and “**”: p<0.01 in comparison to non-pregnant (G0) mice by one-way ANOVA. E) Whole cell lysates were prepared from islets isolated from mice on days 0, 9, and 15 of pregnancy and immunoblotted for Ngn3. The optical density of the Ngn3 band was normalized to that of actin, which was used as the loading control. A representative blot is shown. Results represent mean ± SEM. N = 3–5 mice at each gestational stage from three separate experiments. “*”: p<0.05 in comparison to non-pregnant (G0) mice by one-way ANOVA.

Figure 4. Co-expression of Ngn3 with insulin, glucagon, somatostatin, GLUT2, Ki67, or Pdx-1 in pancreatic islets.

Representative images from G8 Ngn3^EGFP/+ mice are shown immunostained for Ngn3, identified by EGFP-positivity (red), and either insulin (green), glucagon (green), somatostatin (green), GLUT2 (green), Pdx-1 (green), or Ki67 (green) staining. Islets are outlined in white. Red arrows points to Ngn3-EGFP+ cells, green arrows points to hormone-positive cells, yellow arrows points to Ngn3-EGFP+ cells that co-localizes with either a hormone, GLUT2, Pdx-1 or Ki67.

Figure 5. mRNA expression of upstream regulators and downstream targets of Ngn3.

mRNA expression of upstream regulators (Pdx-1, Sox9, and Hes-1,) and downstream targets (Tle3, NeuroD, Nkx2.2, IA1, and Rfx6,) of Ngn3. Islets were isolated from Ngn3^+/+ mice at G0, G6, G9, and G15. Results represent the mean of 3 separate experiments, and expression levels were compared by one-way ANOVA. “*”: p<0.05 by Tukey’s multiple comparison test against G0. N = 6 separate mice at each gestational age.