High glucose alters fetal rat islet transcriptome and induces progeny islet dysfunction

in Journal of Endocrinology
Correspondence should be addressed to K L Kua: kkua@iu.edu

*(J Casasnovas and Y Jo contributed equally to this work)

Restricted access

Offspring of diabetic mothers are susceptible to developing type 2 diabetes due to pancreatic islet dysfunction. However, the initiating molecular pathways leading to offspring pancreatic islet dysfunction are unknown. We hypothesized that maternal hyperglycemia alters offspring pancreatic islet transcriptome and negatively impacts offspring islet function. We employed an infusion model capable of inducing localized hyperglycemia in fetal rats residing in the left uterine horn, thus avoiding other factors involved in programming offspring pancreatic islet health. While maintaining euglycemia in maternal dams and right uterine horn control fetuses, hyperglycemic fetuses in the left uterine horn had higher serum insulin and pancreatic beta cell area. Upon completing infusion from GD20 to 22, RNA sequencing was performed on GD22 islets to identify the hyperglycemia-induced altered gene expression. Ingenuity pathway analysis of the altered transcriptome found that diabetes mellitus and inflammation/cell death pathways were enriched. Interestingly, the downregulated genes modulate more diverse biological processes, which includes responses to stimuli and developmental processes. Next, we performed ex and in vivo studies to evaluate islet cell viability and insulin secretory function in weanling and adult offspring. Pancreatic islets of weanlings exposed to late gestation hyperglycemia had decreased cell viability in basal state and glucose-induced insulin secretion. Lastly, adult offspring exposed to in utero hyperglycemia also exhibited glucose intolerance and insulin secretory dysfunction. Together, our results demonstrate that late gestational hyperglycemia alters the fetal pancreatic islet transcriptome and increases offspring susceptibility to developing pancreatic islet dysfunction.

Downloadable materials

  • Supplemental Figure 1: GD22 fetal pancreatic sections (n=3 pups/group from 3 mothers of independent infusions) were stained with the aforementioned anti-insulin Ab and DAPI, and the number of nuclei (DAPI) were counted and normalized to 1000 μm2 insulin-positive area.
  • Supplemental Figure 2: 3 GD20 pregnant dams underwent the exact same surgery with saline infused into left uterine artery (SAL). Male pups were cross-fostered and GTT was performed on postnatal day 21 as described in the methods section. The glucose levels between 2 groups are analyzed using the same statistical approach as described in manuscript (two-way ANOVA tests followed by Bonferroni multiple comparison tests)
  • Supplemental Figure 3: Measurement of serum (A) Interferon-gamma, (B) Tnf-alpha, (C) ll-1B, (D) IL-17, and (E) Cxcl10 showing no difference between 2 groups. Each symbol represents one replicate with solid line connecting paired group (n=3-5/group, from three to five mothers). Serum was collected from fetal, neonatal, and adult offspring of Con and HG, then measured using Milliplex Rat Cytokine/Chemokine Magnet­ic Bead Panel - Immunology Multiplex Assay (RECYTMAG-65K, Millipore Sigma, MA) by Indiana University Multiplex Analysis Core. This assay is designed to simultaneously quantify selected rat cyto­kines. The kit contains spectrally distinct antibody-immobilized beads, cytokine standard cocktail, streptavidin-phycoerythrin, assay buffer, wash buffer, serum matrix, and microtitre filter plate. Following the manufacturer's recommendation, 25 ul of samples were diluted (1 :2) and processed, then analyzed using Bio-Plex 200 System with High Throughput Fluidics (HTF) Multiplex Assay Array System (Bio-Rad Laboratories, Hercules, CA)> All the samples were run in duplicate. The detection limits for the measured cytokines are as follow: IFN-Gamma 6.2 pg/ml, TNF-Alpha 1.9 pg/ml, IL-1 B 2.8 pg/ml, IL-17 2.3 pg/ml, Cxcl10 1.4 pg/ml).
  • Supp Table 1: List of Differentially expressed genes in HG GD22 islets
  • Supp Table 2: Pathways enriched by IPA
  • Supp Table 3: EnrichR Analysis of downregulated genes


      Society for Endocrinology

Related Articles

Article Information


All Time Past Year Past 30 Days
Abstract Views 460 460 72
Full Text Views 90 90 10
PDF Downloads 23 23 4



  • View in gallery

    (A) Schematic representing localized fetomaternal hyperglycemia model. (B) Experimental timeline. Fetal pups were exposed to hyperglycemia from GD20 to 22, delivered via Cesarean section and cross-fostered to healthy dams. The cross-fostered pups were evaluated at weaning and at adulthood. (C) Maternal blood glucose was unchanged before (GD20) and during infusion (GD22) (left panel, n = 13 mothers). During infusion, the glucose level in blood returning from the left uterine vein was higher than that of the right uterine vein (right panel, n = 15 mothers). (D) The glucose levels of fetuses residing in left uterine horn (HG) were higher than those of their respective controls (Con) during glucose infusion while the placenta was intact (n = 22–26 fetus from 12 mothers), but lower 30 min after birth (right panel, 12–14 pups from seven mothers). (E) Insulin levels of HG pups were higher as well (n = 5 pups/group from five mothers). (F) Pancreatic beta- and alpha-cell area in the HG fetal pups (n = 5–6 pups/group from five mothers).

  • View in gallery

    (A) Heatmap showing differentially expressed genes regulating diseases and biofunction predicted by IPA (n = 3 fetal islet samples/group from three mothers; each islet samples prepared from pool of one pancreas from each gender). Left panel: 24 genes that enriched diabetes mellitus disease process (adjusted P-value = 2.43 × 106). Right panel: Cell death process was upregulated by 41 genes (adjusted P-value = 2.38 × 104). (B) Upregulated and downregulated DEGs were analyzed separately using PANTHER and enriched GO BP was further summarized using REVIGO with the following parameters – database: whole UniProt; semantic similarity measure: Resnik; similarity allowed: Small (0.5). Note that upregulated genes enriched immune and inflammatory processes (red circles) and downregulated genes enriched different biological processes (blue circles). There were two commonly enriched GO BP (G4 – response to stimulus, purple circle in C1 – immune responses/humoral immune response). The table on right shows the summarized list of GO biological processes (BP) and number/percentage of genes annotated to the GO BP.

  • View in gallery

    Validation of RNA-seq results. (A) Graph correlating four downregulated differentially expressed genes (from left-right: Reg3g: n = 4 fetal islet samples/group, Reg3b: n = 4 fetal islet samples/group, Mmp7: n = 3 fetal islet samples/group, Gabrp: n = 5 fetal islet samples/group) (one technical replicate from RNA-seq experiment was included in Reg3g, Reg3b and Gabrp). (B) Graph showing consistent decrease in area positive for REG3G in pancreatic beta cell area (INS+) (*paired t-test P < 0.05, each symbol represents an average data point obtained from three to five islets per sections, total of two sections per fetus, n = 5 GD22 fetuses/group from five mothers). Internal pairs were connected with solid line. Image panel on right showing representative immunofluorescence images obtained from pancreatic tissue of Con and HG pups’ pancreatic sections on the same slide.

  • View in gallery

    HG weanlings developed glucose intolerance and pancreatic islet insulin secretory defect. (A) 1 g/kg intraperitoneal glucose tolerance testing showing increased blood glucose level at ten timepoint and (B) higher incremental glucose area under the curve (iAUC) (*P < 0.05, n = 7–10 male weanlings from six mothers, internal pairs were connected with dashed lines). (C) 2 g/kg intraperitoneal glucose tolerance testing yielded the same result where HG pups continued to have a higher glucose level at 10 min and (D) a higher incremental glucose AUC (*P < 0.05, n = 7–9 male weanlings/group from four mothers, internal pairs were connected with dashed lines). (E) Alamar blue cell viability assay showing the cell viability of HG islets was decreased (*P < 0.05, 7–8 islets per replicate, n = 4 pups/group from four mothers). (F) Ex vivo static GSIS showing decreased HG islets insulin secretion at 5.6 mM glucose and 16.7 mM glucose phase (*P < 0.05, 20–25 islets/group collected from n = 5 weanlings/group from five mothers). (G) Insulin tolerance testing of weanling males (n = 4–6 HG male/group from three mothers).

  • View in gallery

    (A) 1 g/kg intraperitoneal glucose tolerance testing on 2-month-old adult male offspring showing higher blood glucose level at 10 min timepoint and (B) higher incremental glucose iAUC (n = 6–7 male/group, five mothers). (C) Serum insulin at 0, 10 and 30 min timepoints during GTT showing decreased 10 min serum insulin level in HG male adult (n = 4 males/group, four mothers, *P < 0.05 when statistic was performed on fold change of insulin from baseline). (D) Graph showing consistent trend (P = 0.06) of decrease in beta cell area across adult pancreatic sections, with a (E) decreased in pancreatic weight and ultimately decreased in (F) beta cell mass (n = 4 males/group from four mothers). (G) Offspring weight from birth until adulthood (birth n = 64–71 pups from 14 mothers; 7 d/o n = 7–14 pups from three mothers; 14 d/o n = 12–13 pups from three mothers; 21 d/o n = 17–19 males from six mothers; 2 months n = 6–7 males from five mothers). (H) Fat to lean ratio of weanling (n = 3–5 males/group from three mothers) and 2-month-old adult showing no difference between Con and HG offspring (n = 9–10 males from seven mothers).


Google Scholar