Preconception Alcohol Exposure Increases the Susceptibility to Diabetes in the Offspring

Abstract

Heavy alcohol drinking alters glucose metabolism, but the inheritability of this effect of alcohol is not well understood. We used an animal model of preconception alcohol exposure in which adult female rats were given free access to 6.7% alcohol in a liquid diet and water for about 4 weeks, went without alcohol for 3 weeks, and then were bred to generate male and female offspring. Control animals were either ad lib–fed rat chow or pair-fed an isocaloric liquid diet during the time of alcohol-feeding in the experimental animals. Our results show that the female rats fed with alcohol in the liquid diet, but not with the isocaloric liquid diet, prior to conception had an altered stress gene network involving glucose metabolism in oocytes when compared with those in ad lib–fed chow diet controls. The offspring born from preconception alcohol-fed mothers showed significant hyperglycemia and hypoinsulinemia when they were adults. These rats also showed increased levels of inflammatory cytokines and cellular apoptosis in the pancreas, altered insulin production and actions in the liver, and a reduced number of proopiomelanocortin neurons in the hypothalamus. Replenishment of proopiomelanocortin neurons in these animals normalized the abnormal glucose to restore homeostasis. These data suggest that preconception alcohol exposures alter glucose homeostasis by inducing proopiomelanocortin neuronal functional abnormalities. Our findings provide a novel insight into the impact of high doses of alcohol on the female gamete that may cause inheritance of an increased susceptibility to diabetes.

According to the Centers for Disease Control and Prevention (CDC), the total number of people diagnosed with diabetes in the United States was 30.3 million in 2017, and 90–95% of them were diagnosed with type 2 diabetes mellitus (T2DM) (1). Type 2 DM can develop due to an unhealthy lifestyle or genetics or both (2). The risk factors related to lifestyle can include obesity, low level of activity, smoking, and alcohol drinking (2, 3). In the United States, approximately 46% of adult women reported drinking alcohol in the last 30 days, and the prevalence of binge drinking, averaging 4 or more drinks in about 2 hours for women, was 15% among nonpregnant women and 1.4% among pregnant women (4, 5). A recent prospective cohort study in Britain showed that nearly two-thirds of women before pregnancy reported alcohol intakes of about 2 units (16 g alcohol)/week (6). Emerging epidemiological research has shown that women who drank more than 2 drinks a day prior to pregnancy were at significantly higher risk of having babies with lower birth weights and birth percentiles (6). Using a rat model of periconceptional (around the time of conception) alcohol exposure in females, offspring displayed significant deficiencies in glucose metabolism when challenged with a high-fat diet (7). Sex-specific alterations in long-term metabolic function have also been reported in studies examining preconception (long-term alcohol drinking until breeding) male exposure (8). This raises the possibility that alcohol drinking by either parent prior to pregnancy may have a long-lasting effect on offspring metabolic health.

Using a rodent model, we have recently shown that alcohol exposure using a free-access liquid diet, which produced blood alcohol levels of 0.08 to 0.13g/dL in rat dams 3 weeks before pregnancy, significantly affected pregnancy outcomes with respect to delivery date and birth weight of the offspring (9). Preconception alcohol exposure (PCAE) also led to the development of increased stress and anxiety behaviors in their offspring, and the effects were linked with epigenetic changes in several stress regulatory genes, including proopiomelanocortin (POMC). The reported effects of alcohol in the offspring suggest that heavy alcohol use can impact the normal physiological and behavioral development of the next generation possibly due to long-lasting alteration in the function of the gamete’s transcriptomes. Research into the heritability of complex diseases has implicated transmission of epigenetic variants and their effect on the development of neuroendocrine phenotypes, including anxiety behaviors and stress responsivity (10, 11). Stress is often connected with increased T2DM (12). Furthermore, stress regulatory POMC-expressing neurons are known to play a significant role in controlling glucose homeostasis and modulating inflammation (13, 14). Hence, using the rat as an animal model, we determined the effects of drinking high doses of alcohol several weeks prior to conception on glucose metabolism and the susceptibility to develop experimental diabetes in the offspring. We also used a recently developed technique of POMC neuronal supplementation via intraventricular administration of nanosphere-attached dibutyryl cyclic adenosine monophosphate (dbcAMP) (15) to determine the role of POMC in PCAE effects on glucose metabolism.

Materials and Methods

Animal feeding and treatments

Sprague Dawley rats were obtained from Charles River (Wilmington, MA) and housed in controlled conditions as previously described (15). Adult (about 90 days old) female rats were assigned to 1 of 3 treatments for a period of 30 consecutive days: (i) the control group which received rat chow and water ad libitum (AD); (ii) the alcohol-fed group (AF) which received a liquid diet containing ethanol (6.7%); or (iii) the pair-fed (PF) group which received an isocaloric liquid control diet (BioServ, Inc., Flemington, NJ). We selected 30 days of ethanol treatment because long-term effects of ethanol have been shown to be achieved following 3 or more weeks of treatments (16-18). After completion of alcohol-feeding, all rats returned to the AD diet for 3 weeks, some of the animals underwent oocyte preparation and collections, and the rest were bred to generate offspring. Litters from each dam were normalized to 8 pups (both sexes were kept) to reduce the confounding variables from litter size. Pups born from the AF group of mothers and control-fed (AD and PF) mothers were defined as PCAE and control offspring, respectively. Pups were left with their birthmother until weaning on Day 23 and then housed by treatment with same-sex siblings until experimentation. Only 1 animal from each litter was used for biochemical and histological measurements. Animal care and treatment were performed in accordance with institutional guidelines, and protocols were approved by the Rutgers Institutional Animal Care and Facilities Committee and complied with National Institutes of Health policy.

Oocyte isolation and mRNA preparation

This was done using published methods (19, 20). After completion of alcohol-feeding and then returning to the AD diet for 3 weeks, animals were treated with pregnant mare’s serum gonadotrophin (PMSG; 30 IU; Sigma Chemical Co., St. Louis, MO, USA) and human chronic gonadotrophin (HCG; 30 IU; Sigma), and cumulus-oocyte complexes from the oviducts were obtained and denuded to obtain oocytes. Oocytes were washed with 1× cold sterile phosphate-buffered saline (PBS) and spun down at 2000 revolutions per minute for 5 minutes before being used for RNA preparation.

Nanosphere injection given to animals

Female offspring animals around 45 days of age were anesthetized with sodium pentobarbital (50 mg/kg body weight, Butler Schein, Columbus, OH) and injected with 10 µL plain nanospheres (control) or nanospheres that deliver 70 nmol dbcAMP (cAMP; Corpuscular Inc., Cold Spring, NY) in 10µL in the third ventricle using a stereotactic instrument as described by us previously (15). We have previously shown that a single administration of 70 nmol of nanosphere-attached dbcAMP increased β-endorphin neuron differentiation as identified by an enhanced number of BrdU-incorporated β-endorphin neurons and total number of β-endorphin neurons by about 30% to 40% in the arcuate nucleus of the hypothalamus. The dbcAMP-differentiated β-endorphin neurons were functional for a long period of time since levels of the β-endorphin protein and the precursor gene POMC mRNA and the levels of stress hormones, which β-endorphin peptide inhibits, were about 2-fold lower in dbcAMP-treated rats than those in control rats for the 2 months studied (15). Six weeks after nanosphere injection, the animals were used in the endocrine study. At the end of the endocrine study, the animals were euthanized and the brains were collected for β-endorphin cell quantification. The site of injection of nanospheres was verified histochemically by localizing fluorescently labeled nanospheres in the third ventricle.

Endocrine studies.

This was done in PCAE offspring animals. Blood glucose level was measured by the AlphaTrak blood glucose monitoring system (catalog number-043244, AlphaTRAK Test strip and catalog number-89422–894, AlphaTRAK Glucose Meter) from Henry Schein Animal Health, Dublin, Ohio. Blood insulin, glucagon, and leptin (were measured using ultrasensitive rat ELISA Kits (rat insulin, catalog number-90060; rat glucagon, catalog number-81519; rat leptin, catalog number-90040); all from Crystal Chem, Elk Grove Village, IL).

Oral glucose tolerance test.

For the oral glucose tolerance test (OGTT), the animals were fasted overnight, followed by oral administration of 2g/kg of glucose solution (Sigma Aldrich). Tail blood samples were collected at 0, 30, 60, 90, and 120 minutes after oral administration of the glucose solution. Blood glucose levels were determined by the AlphaTrak blood glucose monitoring system.

Glucose-stimulated insulin secretion.

For this test, glucose solution (2g/kg) was administered orally into the animal. Approximately 50 μL blood was drawn from the animal tail vein at various time points (0, 30, 60, 90, and 120 minutes after administration). Collected blood samples were centrifuged to obtain plasma and used to measure insulin concentration to evaluate the body glucose-stimulated insulin secretion (GSIS).

Intraperitoneal insulin tolerance test.

For the intraperitoneal insulin tolerance test (IPITT), animals were fasted for one hour, then 1 IU/kg insulin (Humulin R U-100) was injected intraperitoneally and blood glucose level was measured at 30-minute intervals from 0 to120 minutes.

RNA-seq and bioinformatics

Oocytes collected from AF, PF, and AD rats who were given the alcohol or control diet for 4 weeks and then the regular chow diet for 3 weeks were used in this study. Three biological replicates of oocytes from each of the treatment groups (AF, PF, and AD) were used for RNA preparation using the RNeasy Mini Kit (Qiagen, Germantown, MD). RNA quantity was assessed using the Nanodrop ND-100 (Thermo Scientific; Waltham, MA), and RNA quality was verified using the Agilent 2100 Bioanalyzer with the Agilent RNA 6000 Nano Kit (Agilent Technologies; Santa Clara, CA). RNA-seq library preparation and sequencing was conducted by the Columbia Genomic Center (Columbia University, New York). Briefly, ribosomal RNA was depleted with RiboZero Gold (Illumina) and converted to cDNA with the TruSeq Stranded Library Kit (Illumina). The libraries were sequenced on an Illumina 2500 to a depth of 90 million 100 bp paired-end reads per library. After sequencing, the RNA-sequencing data were processed using 2 different analysis pipelines. First, data were cleaned, aligned to Rattus norvegicus reference assembly Rnor_6.0, and counted using HISAT, SAMTOOLS, and CUFFDIFF (20, 21). Second, analysis of the RNA-sequencing data was performed by Genialis, Inc., (Houston, TX) from the FASTQ files received from the Columbia Genome Center. Fold change values were then used to complete analysis by Gene Ontology, DAVID Analysis, or Ingenuity Pathway Analysis (IPA, Qiagen).

Immunohistochemistry

Pancreatic tissues were harvested and kept in neutral formalin for fixation. Pancreatic tissues were dehydrated, cleared, and embedded and completely processed by the Histopathology Core of the Environmental Occupational Health Sciences Institute at Rutgers University. Pancreatic paraffin samples were sectioned at 6 µm. To do immunohistochemistry, slides were incubated overnight at 60 °C, followed by deparafinization in 3 series of xylene (20, 15, and 10 minutes) and then tissue rehydration by using different concentrations of ethanol (from 100% to 50%). Antigen retrieval was performed by boiling the slides in 10 mM sodium citrate (pH 6.0) at 95 °C for 20 minutes followed by cooling at room temperature. Pancreatic sections from different groups were incubated with 2% bovine serum albumin in PBS and then with primary antibodies for insulin (H-86) (catalog number H-86, 1:100, rabbit monoclonal, Santa Cruz Biotechnology, Inc. Dallas, Texas, research resource identification portal [RRID] AB_1123336 (22)); glucagon (FL-180) (catalog number sc-13091,1:300, rabbit monoclonal; Santa Cruz Biotechnology, Inc. Dallas, Texas, RRID:AB_641026 (23)); interleukin-6 (IL-6; catalog number ab9324, 0.125 µg/ml, Mouse monoclonal to IL6, Abcam Plc, Cambridge, MA, RRID:AB_10732306 (24)); cyclooxygenase-2 (COX-2; catalog number ab1872, 1:200, mouse monoclonal, Abcam Plc, Cambridge, MA, RRID:AB_991710 (25)); interferon-γ (IFN-γ; catalog number ab7812, 1:1000, rabbit polyclonal, Abcam Plc, Cambridge, MA, RRID:AB_306100 (26)); CD3 (catalog number, ab5690, 1:100 mouse monoclonal, Abcam Plc, Cambridge, MA, RRID:AB_305055 (27)); Cleaved Caspase-3 (Asp175) (5A1E) (catalog number 9664L, 1:1000, rabbit monoclonal, Cell Signaling Technology, Inc, Danvers, MA, RRID:AB_2070042 (28)); or Ki-67 (catalog number ab16667, 0.5µg/ml, rabbit monoclonal [SP6]; Abcam Plc, Cambridge, MA, RRID:AB_302459 (29)). All antibodies were diluted in 10% normal horse serum and incubated overnight at 4 °C. The following day, the tissues were washed with PBS and incubated with 1:500 secondary antibody (catalog number PI-1000, HRP Goat Anti-Rabbit IgG Antibody Peroxidase, Vector, Burlingame, CA) for 60 minutes at room temperature. Slides were counterstained with DAB (Vector Laboratories SK-4100 DAB Substrate Kit, Brown) followed by mounting with per mount prolong media. By using 20× magnification, 10 pictures were taken randomly per section and the staining intensity was measured by using ImageJ software using ImageJ software (National Institutes of Health, Bethesda, MD).

Immunofluorescence staining for β-endorphin

Frozen brain tissues were sectioned at 30 µm from plate 18 to plate 23 of the stereotaxic atlas (30) to cover the whole hypothalamic arcuate nucleus. Brain slides were fixed with 4% paraformaldehyde for 15 minutes at room temperature, and then washed with PBS twice for 5 minutes each. This step was followed by 15 minutes of fixation by using methanol and 2 washings with 0.05% Triton X100 in PBS followed by 2 washings of PBS. The brain sections were then blocked with 2.5% horse serum for 1 hour in a moisture chamber, and then incubated with primary anti-rat β-endorphin antibody, (catalog number T-4045.0050, Peninsula Laboratories, San Carlos, CA, RRID:AB_518106 (31)) diluted 1:1000 in 10% normal horse serum overnight at 4 ºC. After overnight incubation, brain sections were washed once with 0.05% TritonX in PBS followed by 2 PBS washings at room temperature. Then, brain sections were incubated with secondary antibody, Alexa Fluor 488 Donkey Anti-Rabbit IgG (H+L) secondary antibody (Thermo Fisher Scientific, Grand Island, NY, catalog number R37118, 1:500, RRID:AB_2556546 (32)) for 1 hour at room temperature in a moisture chamber and in dark environment to protect the fluorescence from direct light. Sections were washed with 2 washings with 0.05% Triton X 100 in PBS followed by 2 PBS washings and mounted with DAPI and sealed by commercial nail polish. Pictures were taken by fluorescent microscope with 20× objective lens (Nikon Fluorescent Microscope). Positive cell numbers in each brain section were manually counted and a ratio to the total cell number from all sections of one brain was calculated. The experimenter counting the immunopositive cells was blinded to the experimental conditions and the count is presented as percentage of control.

Western blotting

Protein levels of PDK1, PTEN, IL1-b, IL6, TNF-α, InsR, and β-actin in pancreatic tissue samples were determined by Western blot using routine protocols (9, 15).

Protein from each sample was extracted and estimated and 50μg of protein was run in 4% to 20% SDS PAGE. The primary antibodies used were: PDK1 (catalog number 3062, 1:1000, rabbit polyclonal, Cell Signaling Technology, Inc., Danvers, MA, RRID:AB_2236832 (33)); PTEN (catalog number 9569, 1:1000, rabbit polyclonal, Cell Signaling Technology, Inc., Danvers, MA, RRID:AB_2163438 (34)); IL-1β (catalog number ab9722, 1:100, rabbit polyclonal, Abcam Plc, Cambridge, MA, RRID:AB_302842 (35)); IL-6 (catalog number ab9324, 0.125µg/ml, mouse monoclonal, Abcam Plc, Cambridge, MA. RRID:AB_10732306 (24)); TNF-α, (catalog number ab1793, 1;100, mouse monoclonal [52B83], Abcam Plc, Cambridge, MA, RRID:AB_469254 (36)); and mouse anti-actin antibody (catalog number CP01 [1:2500], mouse monoclonal, EMD Millipore, Billerica, MA. RRID:AB_566293 (37)). The protein band intensities were determined by Image Studio Lite software (LI-COR Biotechnology, Lincoln, NE), and normalized to cellular levels of actin.

Real-time quantitative polymerase chain reaction

Gene expression level of insulin receptors and signaling molecules in the hepatic tissue were measured by quantitative real-time polymerase chain reaction (PCR) (SYBR Green assay) using routine protocols (9). All the primer sequences used for this study are; Glucokinase forward primer (FP), TGGCCTAATGAAAGCTGGGG; Glucokinase reverse primer (RP), ACACACTTATGGCCCGTGTT; GLUT2 FP, GCAACATGTCAGAAGACAAGATCAC; GLUT2 RP, TGGTAGCTCTTCGGTCATC; PEPCK FP, GGATGTGGCCAGGATCGAAA; PEPCK RP, ATACATGGTGCGGCCTTTCA; InsR2 FP, CCCCAGTGTCCCCATCCT; InsR2 RP, TTTCCTGAGAGAGACGTTTTCCA; FOXO1 FP, TACACGTACGGCCAATCCAG; FOXO1 RP, GACTGTTGGGTTGAGCCACT; Glucose-6-Pase FP, AACTCCAGCATGTACCGCAA; Glucose-6-Pase RP, TCCAGCATTCACGCTTTCCT; GAPDH FP, AGACAGCCGCATCTTCTTGT; GAPDH RP, CTTGCCGTGGGTAGAGTCAT; 18S RNA FP, 5’ GTAACCCGTTGAACCCCATT 3’; 18S RNA RP, 5’ CCATCCAATCGGTAGTAGCG 3’; RPL-19 FP, 5’ AATCGCCAATGCCAACTCTCG 3’; RPL-19 RP, 5’ TGCTCCATGAGAATCCGCTTG 3’. The quantity of target genes and 3 reference genes (GAPDH, 18S, RPL19) was measured using the standard curve method. Target gene expression was normalized with the mean level of 3 reference genes and expressed using 2^-(∆∆Ct) methods (38).

Statistical analysis

Graph Pad Prism software was used for statistical analysis of data. Data are presented as mean ±  standard error of the mean (SEM). The significance between treatment groups and controls was assessed using 1-way analysis of variance (ANOVA) between treatment groups and 2-way ANOVA between treatment groups and across days of exposure. The Newman-Keuls posttest was used to determine the differences between groups following the ANOVA test. P < 0.05 was considered significant.

Results

Preconception alcohol drinking affects the oocyte’s transcriptome related to glucose homeostasis

We tested whether high doses of alcohol-feeding alter transcriptomes of oocytes of dams before pregnancy. Our alcohol-feeding (AF) model consisted of providing free access to 2 bottles: one bottle with a 6.7% alcohol-containing liquid diet and the other bottle with drinking water for 30 days. To control for nutritional influence and possible stress of liquid diet feeding, we tested 2 control groups: one involving a pair-feeding liquid diet substituted with calories for alcohol (PF) and another with ad libitum rat chow feeding (AD) to maintain home-cage environments. Both AF and PF animals drank between 50 and 75 mL of liquid diet during a 30-day period (Fig. 1A) and continued to gain weight, as did AD rats (Fig. 1B). Alcohol exposure produced blood alcohol levels between 0.09% and 0.16 g/dL in these animals (Fig. 1C). After 30 days of alcohol- or pair-feeding, all rats were returned to their natural diet with access to ad libitum rat chow and drinking water in a group-housing home-cage-environment. We have shown previously that this model of alcohol-feeding produces no significant withdrawal symptoms in animals following removal of alcohol-feeding and has no effects on maternal behaviors if breeding occurs 3 weeks after withdrawal of alcohol (9). The transcriptional changes induced by alcohol before conception (3 weeks after alcohol withdrawal) in the oocytes of AF, AD, and PF animals were examined using RNA-seq. All genomic data used in this study are located in a digital research materials repository (39). Bioinformatic analyses of RNA-seq data showed that 330 genes were differentially expressed (>2-fold change, false discovery rate [FDR] 0.01), with 251 downregulated and 79 upregulated genes in AF animals compared with AD animals, as shown in this volcano plot (Fig. 2A). The same analysis revealed 238 genes differentially expressed (>2-fold change, FDR 0.01), with 161 downregulated and 77 upregulated genes in AF animals compared with PF animals (Fig. 2B). Only 2 genes were found to be significantly different in PF control animals, with 1 downregulated and 1 upregulated compared to AD animals (Fig. 2C), suggesting very little influence of liquid diet feeding in long-term transcriptional changes in oocytes. The Venn view comparing all groups of differentially expressed genes identified a total of 173 genes, 123 downregulated and 50 upregulated (FDR 0.01), in AF animals compared with both AD and PF groups (Fig. 2D). Ingenuity Pathway Analysis (IPA) identified significant changes in several cell signaling systems, including glucose metabolism (bars in red), immune functions (bars in orange), and stress regulation (bar in green) in AF animals (Fig. 2E). The disease and function specific classification of the 173 genes performed by the IPA also identified an enrichment in apoptosis of beta islets, D-glucose metabolism, glucose metabolism disorder, and diabetes mellitus. These data suggest that chronic alcohol exposure induces transcriptional changes in genes related to glucose metabolism and its regulatory systems (e.g., immune and stress) (40) in female gametes, which exist a few weeks after alcohol withdrawal.

Figure 1.

Showing the amount of daily liquid diet intake by alcohol-fed (AF) and pair-fed (PF) rats (a), changes in body weight of AF, PF, and ad lib chow-fed (AD) rats (b), and the changes in blood concentrations of ethanol in AF rats (c) used in generating offspring for preconception alcohol exposure (PCAE) study. Data shown are mean ± SEM (n = 8).

Figure 2.

Preconception alcohol exposure (PCAE) alters oocyte transcriptomes in adult female rats as determined by RNA-seq analysis. (a-c) Expression of genes significantly downregulated (green) or upregulated (red) in oocytes as identified by RNA-seq analysis (cut-off log₂ fold change ≥ 2; vertical lines) of alcohol-fed (AF) animal oocytes vs ad libitum–fed normal rat chow (AD) animal oocytes (a), AF vs pair-fed liquid diet (PF) animal oocytes (b), or PF vs AD oocytes (c) as plotted against an FDR value < 0.01 (horizontal lines). Each dot represents the transcriptome from 3 biological replicates per group. [d] Venn view of significant downregulated and upregulated genes as determined by RNA-seq analysis in both AF vs AD and AF vs PF. (e) Ingenuity Pathway Analysis (IPA) identified the significantly affected canonical pathways including glucose metabolism (bars in red), immune functions (bars in orange), and stress regulation (bar in green) in AF animals.

Preconception alcohol exposure alters glucose homeostasis in the offspring

Since alcohol-induced changes in the oocyte transcriptome were observed weeks after cessation of the alcohol diet, we tested if some of the transcriptional changes in oocytes were transmitted to the offspring and affected the offspring’s metabolic health. To address this, we first measured the fasting levels of blood glucose, insulin, glucagon, and leptin in the offspring. Both female (AF-F) and male (AF-M) litters (offspring) of AF mothers showed increased basal levels of blood glucose as compared with the female and male litters of AD (AD-F, AD-M) and PF (PF-F, PF-M) mothers (Fig. 3A and 3B). The high blood glucose levels were accompanied by low blood insulin levels in both female and male litter mates of AF rats (Fig. 3C and 3D). The blood glucagon levels were lower in female pups but not in male pups (Fig. 3E and 3F), while levels of leptin, a glucose homeostasis modulator (41), were higher in both male and female pups from AF rats (Fig. 3G and 3H). Oral glucose tolerance test (OGTT) data revealed that both male and female offspring from each litter of AF mothers had increased blood glucose levels (shown in curves and areas under the curve; Fig. 3I-3L), while the glucose-stimulated insulin secretion (GSIS) test, which reflects insulin production in response to a glucose challenge, showed a reduced insulin response to glucose in male and female AF offspring (shown in curves and areas under the curve, Fig. 3M-3P). Insulin tolerance tests (IPITT) also showed a reduced insulin response in male and female litters of AF mothers (Fig. 3Q-3T). These data indicated significant alterations in glucose homeostasis in the offspring of PCAE mothers. We did not find any sex differences in the PCAE effects on glucose homeostasis in the offspring, and therefore remaining studies were done only in female litters.

Figure 3.

PCAE induces abnormalities in glucose homeostasis of female and male rat litters during the adult period. Female (F) and male (M) litters of AD (AD-F, AD-M), PF (PF-F, PF-M), and AF (AF-F, AF-M) mothers were tested for glucose and insulin status during the adult period. Shown in this figure are fasting glucose (a, b), fasting insulin (c, d), fasting glucagon (e, f), fasting leptin (g, h), blood glucose level after oral glucose tolerance test (OGTT) (i, k), area under the curve of blood glucose during OGTT (j, l), glucose-stimulated insulin secretion (GSIS) (m, o), area under the curve of blood insulin during GSIS (n, p), blood glucose during intraperitoneal insulin tolerance test (IPITT) (q,s), and area under the curve of blood glucose during IPITT (r, t). Data are presented as mean ± SEM (n = 7). Data were compared by 1-way analysis of variance (ANOVA) and then the Newman-Keuls posttest. Differences between control and other treatment groups are shown by lines with P values (*, P < 0.05 and **, P < 0.01) above the bar graphs.

To identify the pathophysiological basis of these altered parameters, female litters of AD, PF, and AF mothers were euthanized at 8 weeks of age and their tissues were collected for protein and gene analyses. Immunocytochemical analysis of the pancreas revealed that the level of insulin staining was lower in AF rats (Fig. 4A and 4B), while the level of glucagon staining was higher (Fig. 4C and 4D) in the AF group. These data are consistent with the data showing low insulin response in GSIS test data for AF rats (Fig. 3M-3P) and identifies a possible reduction in pancreatic beta-cell function in PCAE offspring.

Figure 4.

PCAE induces abnormalities in pancreatic cell functions in the litter. Shown here are the immunohistochemical characterizations of the single-labeled hormone insulin (a), glucagon (c), inflammatory markers IL-6 (e), COX-2 (g), IFN-γ (i), and CD3 (k), apoptotic marker Caspase 3 (m), and cell proliferation marker Ki-67 (0) in the pancreas of PCAE and control exposed female litters. Pictures shown are representative photographs. “-” denotes 100 µM. The intensity of protein staining in each section (10 areas/section) was quantitated and presented as mean ± SEM (n = 7) in histograms on the right for insulin (b), glucagon (d), IL-6 (f), COX-2 (h), IFN-γ (j), CD3 (i), Caspase 3 (n), and Ki-67 (p). Data were compared by 1-way analysis of variance (ANOVA) and the Newman-Keuls posttest. Differences between control and other treatment groups are shown by lines with P values above the bar graphs.

Increasing evidence shows that β-cell dysfunction precedes the clinical onset of diabetes. In the case of type 1 diabetes, β-cell loss is often caused by elevated autoimmune function (39). In this study, we found marked changes in immune cell signaling in the oocyte transcriptome following PCAE (Fig. 2). Hence, we determined whether the reduced production of insulin in pancreatic β-cells is associated with increased inflammatory cytokine (IL-6, COX-2, IFN-γ, and CD3) levels in the pancreas of female PCAE offspring. We found significant increases in IL-6 (Fig. 4E and 4F), COX-2 (Fig. 4G and 4H), and IFN-γ (Fig. 4I and 4J) levels in the pancreatic islets from AF rats. CD3, a cluster of differentiation protein, part of the T cell receptor (TCR) complex on a mature T lymphocyte, was also increased in the pancreas of AF animals (Fig. 4K and 4L). Given the possibility that the reduction in insulin-stained cells in the pancreas may have been caused by apoptosis and/or loss of cellular proliferation, we measured the levels of a cellular apoptotic marker Caspase 3 (Fig. 4M and 4N) and cell proliferation marker Ki67 (Fig. 4O and 4P). However, neither of these cell markers showed any significant changes in the pancreases of AF litters. These data support the possibility that pancreatic inflammation may have contributed to the reduced insulin production in the pancreas without affecting cellular toxicity or growth in PCAE litters.

Oocyte transcriptome analysis of PCAE animals also indicated altered glucose metabolism. To determine whether preconception alcohol exposure impacts insulin control of glucose metabolism in the offspring, we measured changes in protein levels of insulin receptors in liver, muscle, and adipose tissues. We found that protein expression of insulin receptors was decreased in the livers of AF litters (Fig. 5A), but not in muscle or adipose tissue (data not shown). Further determination of key insulin signaling molecules in the liver revealed that animals of the AF group also had decreased mRNA levels of insulin receptors (Fig. 5B) and protein kinase B (PKB or AKT; Fig. 5D) and increased mRNA levels of forkhead box protein 01 (FOXO1; Fig. 5F) and glucose-6-phosphatase (G6-Pass; Fig. 5H). We did not find any significant change in glucose transporter 2 (GLUT2; Fig. 5C), glucokinase (Fig. 5E), or phosphoenolpyruvate carboxykinase (PEPCK; Fig. 5G) mRNA levels. The PCAE offspring also had lower protein levels of pyruvate dehydrogenase kinase 1 (PDK1; Fig. 5I), known to facilitate insulin signaling, but they had higher levels of phosphate and tensin homolog (PTEN; Fig. 5J), which is known to suppress insulin signaling. These data suggest that the expression levels of several key insulin signaling molecules and their regulators were changed in the AF group, which may have contributed to the attenuated insulin action and enhancement of glucose production in AF offspring (Fig. 5K).

Figure 5.

PCAE reduces the level of insulin receptors and its signaling molecules in the liver of the litter. Histograms showing the data of insulin receptor proteins (a), insulin receptor mRNA (b), GLUT2 mRNA (c), Akt mRNA (d), glucokinase mRNA (e), FOXO1 mRNA (f), PEPCK mRNA (g), glucose-6-pase mRNA (h), PDK1 (i), and PTEN (j) in AD, PF, and AF hepatic tissues. Data are mean ± SEM (n = 6) and were compared by ANOVA and the Newman-Keuls posttest. Differences between control and other treatment groups are shown by lines with P values above the bar graphs. [k] Schematic diagram illustrating how the observed changes in insulin receptor function affect glucose homeostasis in PCAE offspring. (<-- activation; |-- inhibition).

Preconception alcohol exposure increases susceptibility to develop type 2 diabetes in the offspring

Whether abnormalities in insulin signaling mechanisms increase the susceptibility to develop T2DM in AF offspring rats was tested by feeding litters of AD, PF, and AF mothers a high-fat diet (HFD) (40%) for 2 weeks followed by a single injection of streptozotocin (STZ) (40 mg/kg), which induces many features of human T2DM (42, 43). HFD feeding and STZ injection treatment increased basal levels of blood glucose to 425.5 ± 36.4 mg/dL (n = 6) in control AD rats (nondiabetic value 97.2 ± 3.9mg/dL, n = 7; see Fig. 3A) and to 477.8 ± 66.4mg/dL (n = 6) in PF rats (nondiabetic value 105.0 ± 5.2mg/dL, n = 8; see Fig. 3A), while increasing higher levels of blood glucose to a greater extent, 666.2 ± 36.2mg/dL (n = 6) in the AF group (nondiabetic value 119.0 ± 3.6mg/dL; see Fig. 3A) (Fig. 6A). AF litters also showed lower mean levels of blood insulin (Fig. 6B). Since diabetes is known to increase inflammation in pancreatic tissue (41), levels of various inflammatory cytokines in the pancreas of AF and control groups were compared. Pancreatic levels of IL-1β, IL-6, and tumor necrosis factor-α (TNF-α) were significantly elevated in the AF group as compared to the AD and PF groups (Fig. 6C-6F).

Figure 6.

PCAE increases the susceptibility to develop diabetes in the litter. Histograms show the effects of PCAE on the level of blood glucose (a), blood insulin (b), and various inflammatory cytokines shown as blots (c) or as mean ± SEM (n = 5–7) actin ratios of IL-1β (d), IL-6 (e), or TNF-α (f) in the pancreas after 2 weeks of 40% HFD followed by a single injection of 40mg/kg of STZ treatment. *P < 0.05 and **P < 0.01.

Preconception alcohol exposure decreases POMC neuronal number, while enhancing POMC neuronal cell populations within the hypothalamus reduces hyperglycemia and the susceptibility to develop diabetes in the offspring. Since PCAE offspring are known to have stress hyper-responsiveness and altered hypothalamic expression of POMC genes (9), which has significant glucose and immune regulatory roles (13, 14), we determined whether the loss of this neuronal function has any contribution to the hyperglycemia in the female PCAE offspring. For this, neurons that produced β-endorphin (a peptide product of the POMC gene that is known to act as a ligand for μ- and δ-opioid receptors to regulate consummatory behavior (13); were immunolabeled in the arcuate nucleus, where these cells are primarily located, and counted. In AF litters, the number of these neurons was reduced as compared to AD and PF controls (Fig. 7A, 7C, 7E, and 7G), as demonstrated in a previous study (9). When PCAE offspring were injected with nanosphere-attached cAMP into the CNS (Fig. 7B), which is known to increase the differentiation of in situ neuronal stem cells into POMC neurons (17, 44), the number of β-endorphin cells was significantly increased in the arcuate nucleus of all groups (Fig. 7C-7H). Interestingly, this method of POMC neuronal supplementation produced no significant effect on blood glucose in AD and PF groups, but it significantly reduced the mean blood glucose level of AF litters to levels similar to those in AD and PF groups (Fig. 7I-7K). The neuronal supplementation significantly increased blood levels of insulin in AF rats (Fig. 7I-7N). POMC neuronal supplementation did not affect blood glucagon (Fig. 7O-7Q) or leptin levels (Fig. 7R-7T) in any of the treatment groups. Body weight changes during a period of 6 weeks did not differ between POMC neuronal supplementation and control treatment groups (Fig. 8A-8D). OGTT data revealed that POMC neuronal supplementation produced no significant effect on blood glucose following glucose challenge in AD and PF litters, but it significantly attenuated the blood glucose response levels, as determined by the area under the curve, in the AF group (Fig. 9A-9F). GSIS, which reflects insulin production in response to the glucose challenge, responded positively to POMC neuronal supplementation in AD, PF, and AF groups (Fig. 9G-9L). We also compared the effect of POMC neuronal supplementation in altering HFD- and STZ-induced changes in glucose homeostasis. POMC neuronal supplementation, via nanosphere-delivered cAMP, significantly reduced blood glucose levels (Fig. 9M-9O) and significantly increased blood insulin levels (Fig. 9P-9R) in both AD and AF litters. POMC neuronal supplementation did not affect HFD- and STZ-induced changes in blood glucagon or leptin levels (data not shown) in any of the treatment groups.

Figure 7.

Third ventricular injection of c-AMP-delivering nanospheres increases β-endorphin cell numbers in the hypothalamus and prevents PCAE effects on fasting blood glucose and insulin in the litter. Percentage changes in β-endorphin neuronal number in the arcuate nucleus of AF-F and PF-F rats as compared to AD-F rats (a). Representative stainings of β-endorphin neurons (in green) are shown in d, f, and h. Illustrations showing the site where nanospheres containing cAMP or vehicle are administered (b). Immunofluorescence pictures of β-endorphin neuronal cells in the hypothalamus of AD (d), PF (f), and AF (h) litters not treated, treated with nanospheres containing vehicle (Nano-V-F), or treated with nanospheres containing cAMP (Nano-cAMP-F). Pictures shown are representative photographs. “-” denotes 100 µM. The number of β-endorphin neurons in the hypothalamus of each group was quantitated and the percentage of these neurons in AD-F was calculated and presented as mean ± SEM in histograms on the right (c, e, g). Effects of nanospheres containing cAMP on fasting blood glucose in AD (i), PF (j), and AF animals (k), on fasting blood insulin in AD (l), PF (m), and AF animals (n), fasting blood glucagon in AD (o), PF (p), and AF animals (q) and fasting blood leptin in AD (r), PF (s), and AF animals (t). Data are mean ± SEM (n = 6) and were compared by ANOVA and the Newman-Keuls posttest. Differences between control and other treatment groups are shown by lines with P values above the bar graphs.

Figure 8.

Third ventricular injection of cAMP-delivering nanospheres did not affect changes in the body weight. Changes in body weight following treatment with nanospheres containing vehicle (Nano-V-F), nanospheres containing cAMP (Nano-cAMP-F) or none in AD (a), PF (b) and AF (c). Data are mean ± SEM (n = 6) and were compared by ANOVA and the Newman-Keuls posttest. No treatment difference was found between groups.

Figure 9.

β-endorphin neuronal supplementation via third ventricular injection of cAMP-delivering nanospheres normalizes glucose tolerance (OGTT) and insulin response to glucose (IPITT) and reduces the susceptibility to develop type 2 diabetes in the PCAE litters. OGTT responses in the offspring without treatment, or treated with nanospheres containing vehicle (Nano-V-F), or treated with nanospheres containing cAMP (Nano-cAMP-F) in AD (a), PF (b), and AF (c) groups. The glucose response during the study period was calculated as the area under the curve (AUC) and presented in histograms for AD (d), PF (e), and AF (f) groups. IPITT responses in the offspring without treatment, or treated with Nano-V-F, or treated with Nano-cAMP-F in AD (g), PF (h), and AF (i) groups. The insulin response during the study period was calculated as AUC and presented in histograms for AD (j), PF (k), and AF (l) groups. [m-r] Show the effects of β-endorphin neuronal supplementation on the HFD- and STZ-induced levels of blood glucose and blood insulin in AD (m, p), PF (n, q), and AF (o, r) animals. Data are mean ± SEM (n = 5-7) and were compared by ANOVA and the Newman-Keuls posttest. Differences between control and other treatment groups are shown by lines with P values above the bar graphs.

These data suggest that the negative impact of alcohol exposure before pregnancy in the offspring’s metabolic functions may be caused by a deficiency of hypothalamic, POMC neuronal, glucoregulatory, and anti-inflammatory functions.

Discussion

The study presented here was modeled to assess whether long-term alcohol drinking impacts pregnancy outcomes and the offspring’s health even after withdrawing from alcohol use. Health professionals often recommend women to stop drinking for about 2 to 3 months, about 3 ovulatory cycles, before trying to conceive in order to avoid alcohol-related problems in the offspring. Using the rat animal model, we showed here that chronic alcohol exposure causes significant alterations in the oocyte transcriptome and offspring metabolic health even after alcohol withdrawal. Data presented here show that chronic alcohol exposure via the liquid diet paradigm altered the female rat’s oocyte transcriptome with changes related to glucose metabolism, diabetes, and stress regulation.

We have used a liquid diet paradigm to feed animals with a binge dose of alcohol, since rats do not like alcohol and scheduled access (limited access) of 2 bottles of choice do not always produce a reliable method of feeding consistent doses of alcohol to determine long-term alcohol effects on physiological systems (45). The liquid diet we used is nutritionally complete and has been shown to produce a high level of ethanol consumption (Fig. 1; (45-49)) that mimics clinical conditions and allows experimental duplications of many pathological complications caused by alcohol, such as various alcohol-induced metabolic derangements (50). The PF group we used ensured a reasonable control for nutrition as evidenced by similar body weight gain in AF, PF, and AD rats in this study. The PF group may have also ensured a reasonable control for stress of force-feeding alcohol to the parental rats that could have affected offspring physiology as very small differences have been observed in long-term transcriptional changes in oocytes between AD and PF parental rats in this study and in corticosterone levels in AD and PF offspring rats in a previous report (9). Therefore, it could be interpreted that the observed effects of preconception alcohol on offspring metabolic health is primarily due to long-term changes in parental transcriptomes induced by ethanol. The health outcomes of dams fed with alcohol were not determined in this study and need further studies in the future. There are reports in the literature that high doses of alcohol exposures alter glucose homeostasis and promote the development of diabetes in humans (51, 52) and animals (53, 54).

Our observations indicated that PCAE altered glucose homeostasis status, as determined by fasting glucose levels and reduced OGTT response data, in both male and female offspring. The IPITT data and GSIS results also demonstrated that AF rats showed reduced insulin secretion in response to an oral glucose challenge. These data are in agreement with a report in a rat model of periconceptional alcohol exposures, where a deficiency in glucose metabolism was observed in offspring when challenged with a HFD (9). These observations suggest that ethanol can program an insulin-resistant phenotype, which, similar to alcohol-induced disruptions of the hypothalamic-pituitary-adrenal axis (9, 10), can transmit to the next generation via the female gamete. These data also suggest the possibility that the ethanol effects on gene expression are long-lasting and may be permanent. We have also published data showing a significant association between binge or heavy levels of alcohol drinking and elevated levels of methylation and reduced levels of expression of POMC genes even after withdrawal of alcohol drinking in humans (55). However, additional study is needed to determine if the alcohol-induced changes in gene expression are permanent.

We showed here that PCAE changes the normal functional status of pancreatic tissue, including reduction of insulin production and secretion in the offspring. We also found that the reduced insulin production in β-cells is associated with increased inflammatory cytokines. This is similar to the situation in type 1 diabetes where β-cell functional loss is often caused by elevated autoimmune function (40). In the PCAE offspring, we found high blood levels of leptin that may also reflect increases in body inflammatory status, because high circulating leptin levels are known to initiate nonspecific immune defense and inflammatory activity (56). Plasma glucagon levels did not show consistent responses to PCAE between males and females, and lower levels found in AF female animals may reflect a response to the hyperglycemic state of these animals. We are not certain why ethanol treatment lowered basal glucagon levels in females but not in males, and further research is needed to address this issue. However, a gender difference in alcohol pharmacokinetics has been suggested to explain why women are more vulnerable to ethanol’s toxic effects (57). Also, it has been reported that a gender difference exists in autonomic regulation of the pancreatic alpha-cell, which produces glucagon (58).

We observed that insulin resistance in PCAE offspring was associated with increased gene expression related to glucose homeostasis, such as Foxo1 and G6pc, and decreased insulin receptor and PKB levels, which indicated increased glycogenolysis and decreased insulin function/sensitivity in the liver. Our observations indicated that the decreased insulin response was likely due to reduced sensitivity of the hepatic insulin signaling pathway, but the exact mechanism needs further investigation. We did not find any changes in insulin receptor levels in muscle and adipose tissue, suggesting that PCAE did not cause whole body insulin resistance.

In the present study, we evaluated the effect of PCAE in a diabetic model induced by a combination of HFD and a moderate dose of STZ in rats. Potent hyperglycemic, hypoinsulinemic, and inflammatory effects were observed in HFD+STZ-induced diabetic PCAE rats. These findings together with data showing hyperglycemia and insulin resistance suggest that PCAE programs the offspring to develop diabetes. The inheritability of alcohol-increased diabetes susceptibility in PCAE offspring is consistent with the finding that paternal prediabetes increases the susceptibility to diabetes in offspring through gametic alterations of the transcriptome (8, 59). In the HFD+STZ animal model, we found increased blood glucagon levels in the AF rats, which is in contrast to the decreased blood glucagon levels we found in untreated AF rats. These differences might have been related to the increased HFD’s ability to elevate blood glucagon levels (60), but the exact mechanism is currently unknown.

PCAE offspring also showed significant reductions in POMC neuronal number in the hypothalamus, as determined in this study, and gene expression levels and stress hyper-responsiveness in a previous study (9). We have previously demonstrated that increasing POMC neuronal numbers in the hypothalamus via nanosphere-delivered cAMP enhances body immunity and reduces the body stress response (17, 61). In this study, we showed that nanosphere-delivered cAMP promoted the body’s ability to control glucose homeostasis as well as basal and glucose-induced insulin production in PCAE offspring. This treatment also reduced the sensitivity to develop HFD+STZ-induced diabetes. The particular mechanism for POMC neuronal action on pancreatic beta-cell health is not known, but we have shown previously that a strong relationship exists between POMC neuronal activation and the inhibition of sympathetic activity as well as activation of parasympathetic neurons (61). The postganglionic fibers of the sympathetic nervous system terminate at every organ in the body, including primary and secondary lymphoid tissues (lymph nodes, spleen, and bone marrow) as well as endocrine glands like the pancreas. POMC neurons produce β-endorphin which binds to µ-opioid receptors on postsynaptic parvocellular neurosecretory cells of the paraventricular nucleus. These neurons terminate in the spinal cord or in the rostral ventrolateral medulla where they can directly and indirectly modulate sympathetic nerve activity, respectively (62, 63). Sympathetic innervation of the pancreas inhibits pancreatic beta-cell production and the release of insulin, while parasympathetic innervation stimulates insulin release from pancreatic islets (64). Therefore, POMC neuronal regulation of pancreatic function might involve suppression of sympathetic and activation of parasympathetic neuronal systems. These data suggest the possibility that PCAE programs oocytes to induce stress axis abnormalities, which alters glucose homeostasis and increases the susceptibility to diabetes by increasing pancreatic inflammation and consequently lowering insulin production and action.

The results presented here suggest the possibility that exposure to alcohol produces transmissible long-lasting changes in the genome that result in profound alterations to metabolic function of the offspring. These findings also indicate that chronic alcohol drinking prior to pregnancy may permanently impact the female gamete that causes inheritance of increased susceptibility to develop diabetes. Data presented here show that the ethanol-induced changes in genes related to glucose metabolism and stress regulation in oocytes of mothers are also observed in the offspring during adult life, suggesting that the ethanol effects on gene expression are long-lasting and may be permanent.

Glossary

Abbreviations

AD

ad libitum–fed

AF

alcohol-fed

ANOVA

analysis of variance

COX-2

cyclooxygenase-2

dbcAMP

dibutyryl cyclic adenosine monophosphate

FDR

false discovery rate

HFD

high-fat diet

IFN-γ

interferon-γ

IL-6

interleukin-6

IPA

Ingenuity Pathway Analysis

IPITT

intraperitoneal insulin tolerance test

PBS

phosphate-buffered saline

PCAE

preconception alcohol exposure

PCR

polymerase chain reaction

PF

pair-fed

POMC

proopiomelanocortin

PVN

paraventricular nucleus

SEM

standard error of the mean

STZ

streptozotocin

T2DM

type 2 diabetes mellitus

TNF-α

tumor necrosis factor-α

Additional Information

Disclosure Summary: The authors declare that there is no conflict of interest regarding the publication of this article.

Data Availability

All relevant data are included in this manuscript. Any requests for materials will be fulfilled upon request through appropriate means. Upon publication, RNA-seq data will be uploaded into the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/info/overview.html).