Transcriptomic responses are sex-dependent in the skeletal muscle and liver in offspring of obese mice

Abstract

Infants born to obese mothers are more likely to develop metabolic disease, including glucose intolerance and hepatic steatosis, in adult life. We examined the effects of maternal obesity on the transcriptome of skeletal muscle and liver tissues of the near-term fetus and 3-mo-old offspring in mice born to dams fed a high-fat and -sugar diet. Previously, we have shown that male, but not female, offspring develop glucose intolerance, insulin resistance, and liver steatosis at 3 mo old. Female C57BL6/J mice were fed normal chow or an obesogenic high-calorie diet before mating and throughout pregnancy. RNAseq was performed on the liver and gastrocnemius muscle following collection from fetuses on embryonic day 18.5 (E18.5) as well as from 3-mo-old offspring from obese dams and control dams. Significant genes were generated for each sex, queried for enrichment, and modeled to canonical pathways. RNAseq was corroborated by protein quantification in offspring. The transcriptomic response to maternal obesity in the liver was more marked in males than females. However, in both male and female offspring of obese dams, we found significant enrichment for fatty acid metabolism, mitochondrial transport, and oxidative stress in the liver transcriptomes as well as decreased protein concentrations of electron transport chain members. In skeletal muscle, pathway analysis of gene expression revealed sexual dimorphic patterns, including metabolic processes of fatty acids and glucose, as well as PPAR, AMPK, and PI3K-Akt signaling pathways. Transcriptomic responses to maternal obesity in skeletal muscle were more marked in female offspring than males. Female offspring had greater expression of genes associated with glucose uptake, and protein abundance reflected greater activation of mTOR signaling. Skeletal muscle and livers in mice born to obese dams had sexually dimorphic transcriptomic responses that changed from the fetus to the adult offspring. These data provide insights into mechanisms underpinning metabolic programming in maternal obesity.

NEW & NOTEWORTHY Transcriptomic data support that fetuses of obese mothers modulate metabolism in both muscle and liver. These changes were strikingly sexually dimorphic in agreement with published findings that male offspring of obese dams exhibit pronounced metabolic disease earlier. In both males and females, the transcriptomic responses in the fetus were different than those at 3 mo, implicating adaptive mechanisms throughout adulthood.

INTRODUCTION

The development of obesity and type 2 diabetes mellitus (T2DM) is linked to an adverse intrauterine environment (1–3). For example, it has been reported that ∼50% of T2DM in young individuals can be attributed to exposure to an in utero metabolic environment perturbed by gestational diabetes mellitus (GDM) or maternal obesity (4). This is daunting given the increasing incidence of GDM and the prevalence of obesity in reproductive age women. Specifically, more than 60% of American women now enter pregnancy either overweight [body mass index (BMI) = 25–29.9 kg/m²] or obese (BMI ≥ 30 kg/m²) (5–7). Pregnancies complicated by obesity are associated with an array of obstetrical complications that increase fetal morbidity and mortality (8–11). Babies born to obese women, in particular if infants are large at birth, have a greater risk to develop heart disease, metabolic syndrome, T2DM, and cancer in adulthood (12–18). Of particular concern is the association between maternal obesity in pregnancy and the increased risk of metabolic dysfunction in childhood, which represents a vicious cycle of detrimental intrauterine transmission of metabolic disease from the mother to her children (17–19). The mechanistic links betweeen exposure to the adverse metabolic environment of the obese mother and metabolic disease later in life remain largely unknown, but an array of diverse adaptations in the offspring, such as altered lipid trafficking, peripheral insulin resistance, dysregulated islet insulin secretion, cardiac remodeling, and hepatic dysfunction have been implicated (8).

Fetal programming from maternal obesity has been extensively studied using animal models of overnutrition, especially diets high in fat (20–29). However, the very-high-fat diets (60%) often used to induce obesity do not accurately model the Western style diet (moderately high fat and high sugar) that most pregnant women consume (30–32). Morever, animal models of maternal obesity generated by high fat are often associated with moderate fetal growth restriction (33–38), in contrast to the well established clincial association between maternal obesity and fetal overgrowth (39–42). We developed a mouse model of obesity utilizing a moderately high-fat and -sugar diet, similar to diets reported in overweight/obese women (30, 31), that recapitulates key features of obesity in human pregnancy, including larger offspring (43). By 3 mo of age, male offspring exhibit greater fat mass, insulin resistance, and hepatic steatosis despite normal chow diet after weaning, with adult female offspring of obese mice having a less pronounced metabolic phenotype (44). These findings are in general agreement with findings in other models of maternal obesity that have been shown to be associated with programming of metabolic dysfunction (21, 45–47) often with male offspring more susceptible than female (25–27, 47, 48).

In humans, the array of programming adaptations observed in children of obese women can interact to worsen outcomes, including the development of nonalcoholic fatty liver disease (NAFLD). In children, NAFLD exacerbates the risk of developing insulin resistance, cardiovascular disease, and diabetes (49, 50). Studies in women with obesity and GDM have demonstrated that maternal BMI strongly correlates with intrahepatocellular lipid storage in the infant and that these early life exposures, independent of breast feeding, are hypothesized to underlie earlier onset of NAFLD (28, 49). Thus, mechanistic studies for understanding programming responses to maternal obesity in the liver and skeletal muscle are critical. In rodents, the development of early liver steatosis in the offspring from obese dams is consistent with the idea that the fetal liver takes up excess fuels in pregnancies complicated by maternal obesity (45), contributing to whole body insulin resistance (51). In the fetal liver, early signs of NAFLD are observed as increased triglyceride content, oxidative stress, and activation of gluconeogenic pathways (45). The livers of adult offspring following maternal overnutrition have increased lipid content (27, 28, 45), increased levels of peroxisome proliferator-activated receptor-gamma (PPARγ), reduced triglyceride lipase, and lowered glycogen and protein content (21, 44). Numerous studies report that oxidative stress markers are elevated (21, 25, 28) and some demonstrate overt mitochondrial dysfunction in the livers of these offspring (21, 27, 45). Likewise, maternal overnutrition in animal models specifically utilizing diets high in fat and sugar have documented changes in skeletal muscle development and metabolism (46, 52), as well as reduced muscle force in the adult offspring, regardless of the postweaning diet (53).

Intrauterine metabolic programming of liver and muscle is of high clinical relevance and warrants the use of comprehensive discovery tools to better understand the mechanisms underlying early dysfunction in these major metabolic tissues. However, the use of “omics” approaches in this area has been relatively limited. Previous reports indicate that maternal high-fat diet caused an inflammatory signature in hepatic gene expression in offspring (54, 55). Lomas-Soria et al. (25) reported that the liver transcriptome in 4-mo-old offspring from obese mice is enriched for insulin signaling and lipid metabolism. In this study, the livers of male offspring displayed more than a 10-fold increase in the number of genes differentially regulated by maternal obesity compared with female livers, consistent with the sex effect observed in physiological experiments (25). Similarly, a separate report showed that only 10% of genes regulated by maternal obesity in the livers of the 2-wk-old offspring overlapped between males and females (56).

The objective of the current study was to identify novel pathways involved in the programming of liver and skeletal muscle metabolism in offspring of obese dams. Our approach was to carry out high throughput RNA sequencing in livers and skeletal muscle of near-term fetuses and 3-mo-old offspring of obese mice.

METHODS

Animals and Ethical Approval

All experimental protocols were approved by the Institutional Review Boards of the University of Texas Health Science Center San Antonio and University of Colorado Anschutz Medical Campus. Twelve-week-old C57BL/6J female mice (The Jackson Laboratory, Bar Harbor, ME) were fed either a control diet (D12489; Research Diets, New Brunswick, NJ) or an obesogenic diet (Western Diet D12089B; Research Diets) consisting of pellets containing 10% and 40% calories from fat, respectively as described previously (43). Female mice fed the obesogenic diet had ad libitum access to 20% sucrose solution supplemented with micronutrients, vitamins (Vitamin Mix V10001; Research Diets), and minerals (Mineral Mix S10001; Research Diets). Females were mated following a 25% gain in weight over the initial body weight in the obesogenic diet group as described previously (39). All males used for mating were fed the control diet. The presence of a postcopulatory plug was defined as embryonic day (E) 0.5. At E18.5 obese dams (n = 10) and control dams (n = 10) were euthanized, and fetal livers and gastrocnemius muscle were collected, snap frozen, and stored at −80°C. To collect tissues from 3-mo-old offspring, separate cohorts of treated and control dams were delivered spontaneously, and litters were culled to equal size (n = 6–8). Dams were maintained on their respective diets throughout pregnancy and lactation. Offspring were fed a control diet after weaning at 3 wk of age. At 3 mo of age, 10 offspring from obese dams (OB-offspring; 5 males and 5 females) and 10 offspring from control dams (CON-offspring; 5 males and 5 females) were euthanized, and livers and gastrocnemius muscle were collected as described for fetuses.

Animal Statistics

Statistical significance of differences in body and organ weights between the fetuses from obese and control dams was assessed using Student’s t test. For offspring, weekly body weights were recorded for each sex independently and the body weights at 3 mo of age were compared with Student’s t test. P < 0.05 was considered significant. Data are presented as means ± SE.

RNA Isolation

Tissue samples (∼10 mg) were homogenized in TRI Reagent and total RNA was purified using the Direct-zol RNA MiniPrep kits (Zymo Research, Irvine, CA), according to the manufacturer’s instructions. RNA was resuspended in 50 µL of DNase/Rnase-free water. RNA quality was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA). RNA concentrations were quantified using Qubit RNA HS assay kits and a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, Wilmington, DE). Total RNA was stored at −80°C.

RNA Sequencing

Messenger RNA was selected and double-stranded cDNA libraries with ligated sequencing adapters were constructed using the KAPA Stranded mRNA-Seq kit (Roche, Wilmington, MA), according to the manufacturer’s instructions. Paired-end reads were aligned to the mouse reference genome GRCm38 or mm10 (https://www.ncbi.nlm.nih.gov/assembly/GCF_000001635.20/) using STAR v2.5.3a with default parameters. The annotation model used was RefSeq Transcripts 86, which had 42,551 transcripts. Paired-end reads for the samples averaged ∼12,000,000 (ranging from 4,000,000 to 16,000,000) and had an overall mapping rate of 95% with 85% concordant pair alignments representing unique mapping (Supplemental Data S1).

Differential gene expression analysis was conducted using edgeR (Bioconductor) in R (v4.0.0) on the full list of annotated transcript abundances in counts per million with the qlf function (57, 58). The primary analysis determined the effect of maternal obesity on transcript expression in the liver and the skeletal muscle at the fetal (E18.5) timepoint and at the adult (3 mo) timepoint. Differential gene expression was performed on OB-fetuses (n = 10) compared with CON-fetuses (n = 10) in the liver, OB-fetuses (n = 10) compared with CON-fetuses (n = 10) in the skeletal muscle, OB-offspring (n = 10) compared with CON-offspring (n = 10) in the liver, and OB-offspring (n = 8) compared with CON-offspring (n = 8) in the skeletal muscle. However, principal component analysis (PCA) was performed on abundance levels to assess the variation across runs. Due to the marked differences between females and males (Supplemental Data S2), separate pairwise comparisons for each sex are presented. For both the liver and skeletal muscle, male OB-fetuses (n = 5) compared with male CON-fetuses (n = 5) were analyzed and then female OB-fetuses (n = 5) compared with female CON-fetuses (n = 5). Subsequently, male OB-offspring (n = 5) compared with male CON-offspring (n = 5) were analyzed and then female OB-offspring (n = 5 for liver and n = 4 for muscle) compared with female CON-offspring (n = 5 for liver and n = 4 for muscle)). Genes were not filtered for a threshold fold change and significance was determined with a more stringent P value of 0.01. Lists of the differentially expressed genes (DEGs) generated for each group were compared between the fetus and adult offspring timepoint to identify persistent programming adaptations in both the liver and skeletal muscle. Lists of DEGs were also compared between males and females to help define sex-independent responses to maternal obesity caused by high fat/high sugar (HF/HS). Data presented here have been deposited in NCBI’s Gene Expression Omnibus and are accessible through accession number XX.

Functional Analysis and Data Visualization

Relevant Gene Ontology (GO) terms and canonical pathways for DEGs were identified with KOBAS [v3.0; (59)] and Ingenuity Pathway Analysis (IPA, QIAGEN Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis). Functional terms were significantly enriched if P < 0.05 following a Fisher’s exact test with Benjamini–Hochberg correction. To summarize GO enrichment and reduce redundant terms, DEGs were further analyzed with online tools in ReviGO (60). Similar to the comparative analysis of DEGs, overlapping functional enrichments were evaluated between males and females as well as between fetuses and 3-mo-old offspring. Venn diagram visualizations were created using Venny 2.1 (https://bioinfogp.cnb.csic.es/tools/venny/). Genemania was used to generate gene networks (61). Heatmaps were generated in R (v4.0.0) with ggplots2 and heatmap functions. All other graphs were generated using Graphpad Prism (v9).

Western Blot Analysis

In brief, equal amounts of protein (5 μg) were loaded into each well and separated on KD Mini-Protean Tris Glycine Pre-cast gels (BioRad, Hercules CA), and electrophoresis was performed at a constant 120 V for 2 h. Proteins were transferred onto polyvinylidene difluoride (PVDF) membranes overnight at a constant 35 V. After the transfer, membranes were blocked in 5% (wt/vol) bovine serum albumin (Bio-Rad) in Tris-buffered saline (TBS; wt/vol) plus 0.1% Tween 20 for 1 h at room temperature. Membranes were incubated with primary antibodies overnight at 4°C. Primary antibodies targeting 4EBP1 (Cell Signaling Technology; Davers, MA; Cat. Nos. 9452, 9459, and 9455), Cpt1a (proteintech 15184-1-AP), Scd1 (Abcam ab39969), and Total Ox Phos antibody cocktail (Abcam ab110413) were used. Subsequently, membranes were incubated with the appropriate secondary peroxidase-labeled antibodies for 1 h. After washing, bands were visualized using chemiluminescence detection reagent SuperSignal West Dura Substrate (Thermo Fisher). Membranes were also stained for total protein using Amido Black Stain (Sigma). Images were captured on a G:Box ChemiXL1.4 (Syngene, Cambridge, UK) and analysis of the blots was performed by densitometry using GeneSys software (Syngene). Target protein expressions were adjusted for amount of total protein (Amido Black stain) to ensure equal loading. The mean value of controls was arbitrarily assigned a value of 1.0 for comparisons between groups. Data are presented as means ± SE. Statistical analysis was performed in Graphpad Prism (v9) using unpaired t test with Welch’s correction. Significance was P < 0.05.

RESULTS

Body Weight in the Offspring of Obese Dams

At E18.5, OB-fetuses and their livers were heavier (P < 0.05) than CON-fetuses (Fig. 1, A and B) but there were no differences between sexes (data not shown) and no difference in litter size (Fig. 1C). In the 3-mo-old offspring cohort, there was a significant interaction between sex and weekly weight gain. Male OB-offspring were heavier than CON-offspring at 6 wk (P < 0.01) and by 3 mo of age, male OB-offspring weighed significantly more than male CON-offspring (36.3 ± 0.5 g vs. 27.2 ± 0.9 g, respectively: P < 0.05: Fig. 1D). Female OB-offspring were not heavier than female CON-offspring (Fig. 1E).

Figure 1.

Animal characteristics. For the fetuses, the weight of total animal (A) and of liver (B) are presented adjacent to mean litter size (C) determined at embryonic day 18.5 (E18.5) from dams fed a control diet (black) and high-fat/high-sugar diet resulting in maternal obesity (gray). Males and females combined, n = 10/group. Values are expressed as means ± SE, *P < 0.05 vs. control; unpaired Student’s t test. For 3-mo-old offspring, weekly weights are plotted for males (D) and females (E) separately. Values are expressed as means ± SE. Controls plotted as black circles (n = 5/sex) and obese plotted as gray squares (n = 5/sex). Means at week 14 analyzed for effect of maternal obesity on offspring body weight. *P < 0.05 within sex, control vs. obese, Student’s t test. CON, control; OB, obese.

The Liver Transcriptome in Fetuses of Obese Dams

Differential transcript abundance in the fetus was analyzed for each sex independently, male OB-fetuses (n = 5/treatment) had 399 DEGs in the liver (P < 0.01; 256 down- and 143 upregulated) compared with male fetuses of control dams. Female OB-fetuses (n = 5/treatment) had 173 DEGs in the liver (P < 0.01; 94 down- and 79 upregulated) compared with their respective controls. In male fetuses of obese dams compared with control male fetuses, transcortin (Serpina6), an acetylglucosaminyltransferase (B3gnt8), and solute transporter for Mg (Slc41a3) were the three most downregulated genes, representing transporters for cations and glucocorticoids (Fig. 2A); whereas, the most upregulated genes included Inhibin β C (Inhbc) and genes that encode enzymes catalyzing the generation of 2-ketoglutrate (d-2-hydroxyglutarate dehydrogenase; D2hgdh) and process steroids (sulfotransferase family 2A member 1; Sult2a1; Fig. 2A). In contrast, in female fetuses lysozyme 1 (Lyz1), ATP synthase F1 subunit γ (Atp5c1), and A kinase anchor protein 5 (Akap5) were the three most downregulated genes; whereas, chymosin (Cym), cathepsin J (Ctsj), apolipoprotein A-V (Apoa5) were the most upregulated genes in response to maternal obesity (Fig. 2B). Only eight genes were differentially regulated by maternal obesity in both males and females (Fig. 2C), with six sharing directionality, including the liver isoform of the fatty acid transporter carnitine palmitoyl transferase 1a (Cpt1a; increased) as well as cyclin dependent kinase 10 (Cdk10; decreased) and progestin and adipoQ receptor family member VII (Paqr7; increased) (Fig. 2D). Interestingly, glycogenin (Gyg1) that encodes for the glycosyltransferase that results in glycogen production was decreased in females but increased in males. Pathway analysis revealed male DEGs were significantly enriched for metabolic and protein processing/export (Table 1); whereas female DEGs were significantly enriched for fatty acid (FA) metabolism and pathways for p53 and PPAR (Table 2).

Figure 2.

Comparison of differentially expressed genes following maternal obesity in the livers of male and female fetuses. A: the expression of the ten most up- and downregulated genes are plotted that have the greatest change (up and down) in expression in obese (OB)-fetuses compared with control (CON)-fetuses in males (A; n = 5) and females (B; n = 5). The x-axis represents the log₂ transformation of fold changes between experimental groups calculated from normalized counts per million. Genes are listed on the y-axis as the official HGNC symbol. C: Venn diagram of genes identified in separate pairwise comparisons in the liver for male OB-fetuses compared with CON-fetuses as well as female. D: the seven genes identified as coregulated in both sexes with corresponding descriptions and fold change in OB-fetuses expressed as log. Genes downregulated in OB-fetuses are negative while genes upregulated are positive.

Table 1.

KEGG pathways enriched in DEGs from the livers of male OB-fetuses

Pathway	Input No.	Total No.	P Value
Protein processing in endoplasmic reticulum	36	163	6.50e-31
Metabolic pathways	63	1494	8.70e-18
Protein export	9	28	5.41e-09
Drug metabolism—other enzymes	11	87	2.23e-07
N-Glycan biosynthesis	8	50	3.99e-06
Steroid hormone biosynthesis	9	89	2.10e-05
Biosynthesis of amino acids	8	78	6.86e-05
Phagosome	10	181	5.88e-04
Pyrimidine metabolism	6	58	8.14e-04
Glycine, serine, and threonine metabolism	5	40	1.34e-03
PI3K-Akt signaling pathway	12	358	6.00e-03
Glutathione metabolism	5	64	7.09e-03
Purine metabolism	7	136	7.36e-03
Amino sugar and nucleotide sugar metabolism	4	49	1.67e-02
ECM-receptor interaction	5	86	1.93e-02
Arachidonic acid metabolism	5	89	2.12e-02
Antigen processing and presentation	5	91	2.28e-02
Retinol metabolism	5	91	2.28e-02
Spliceosome	6	133	2.34e-02
Regulation of actin cytoskeleton	7	217	4.88e-02
Carbon metabolism	5	121	5.10e-02

All significant pathways are presented with the number of differentially expressed genes (DEGs) (Input No.; Gene IDs for each pathway can be found in Supplemental Data) in each pathway’s complete annotated gene number (Total No.). Pathways were filtered by redundancy and only pathways with unique genes are presented. ECM, extracellular matrix; OB, obese. P value is corrected for multiple testing with Benjamini–Hochberg.

Table 2.

KEGG pathways enriched in DEGs from the livers of female OB-fetuses

Pathway	Input No.	Total no.	P Value
Fatty acid degradation	3	50	2.05e-02
Fatty acid metabolism	3	61	2.93e-02
Seleno compound metabolism	2	17	2.98e-02
p53 signaling pathway	3	71	3.92e-02
PPAR signaling pathway	3	87	5.52e-02

DEGs, differentially expressed genes; OB, obese; PPAR, peroxisome proliferator-activated receptor.

The Liver Transcriptome in the 3-Mo-Old Offspring of Obese Dams

When differential liver transcript abundance was analyzed for adult males and females separately, male OB-offspring had 580 DEGs (286 down- and 294 upregulated) as compared with male offspring of control dams (n = 5/group). In female OB-offspring, the liver had 278 DEGs (123 down- and 155 upregulated). In male OB-offspring, the most upregulated transcripts compared with male CON-offspring included genes that encode for proteins involved in fatty acid oxidation such as acyl-CoA thioesterase 8 (Acot8); whereas genes most downregulated encoded for structural proteins (Adamts13, Dlg3, Sybu, and Cib3) and TGFB-induced factor homeobox 1 (Tgif1), a transcription factor binding to retinoic acid responsive element (Fig. 3A). In adult females, the most upregulated DEGs in the liver in response to maternal obesity included Cytochrome P450 (Cyp2c39), calmodulin 1 (Calm1), peroxiredoxin Like 2A (Prxl2a), perilipin 4 (Plin4), apolipoprotein A-II (Apoa2; Fig. 3B) and the most downregulated were also largely related to metabolic processes and included glutathione perioxidase 3 (Gpx3), glutathione s-transferase α 1 (Gsta1), 3-oxoacid CoA-transferase 1 (Oxct1), and protein kinase C and casein kinase substrate in neurons 2 (Pacsin2: Fig. 3B). There were 80 genes in the OB-offspring liver that were differentially regulated in both males and females (Fig. 3C), including congruent increased expression of Plin4, Apoa2, apolipoprotein A-IV (Apoa4), and carnitine palmitoyltransferase 1 A (Cpt1a) as well as decreased expression of stearoyl-Coenzyme A desaturase 1 (Scd1), proteinase 3 (Prtn3), and s100 calcium binding proteins (S100a9 and S100a8). Several of these genes have been associated with the development of NAFLD (62, 63). Ten of the significantly enriched pathways in males (Table 3) and females (Table 4) overlapped between sexes, notably fatty acid and cholesterol metabolism, PPAR signaling, cell cycle, and DNA replication were impacted in both males and females (presented in Table 5 with the corresponding genes). Finally, predictive bioinformatic analysis with IPA identified 139 DEGs in male and 94 DEGs in female 3-mo-old offspring of obese dams that are associated with hepatic hyperplasia and steatosis.

Figure 3.

Comparison of differentially expressed genes following maternal obesity in the livers of male and female offspring. The expression of the ten most up- and downregulated genes are plotted that have the greatest change (up and down) in expression in obese dams (OB)-offspring compared with control (CON)-offspring in males (A; n = 5) and females (B; n = 5). The x-axis represents the log₂ transformation of fold changes between experimental groups calculated from normalized counts per million. Genes are listed on the y-axis as the official HGNC symbol. C: Venn diagram of genes identified in separate pairwise comparisons in the liver for male OB-offspring compared with CON-offspring as well as the female comparison.

Table 3.

KEGG pathways enriched in DEGs from the livers of male OB-offspring

Pathway	Input No.	Total No.	P Value
Metabolic pathways	91	1494	1.71e-25
Cell cycle	18	123	2.53e-10
Antigen processing and presentation	15	91	2.84e-09
Drug metabolism—other enzymes	12	87	8.34e-07
Fatty acid metabolism	9	61	1.92e-05
Cholesterol metabolism	8	49	3.55e-05
PPAR signaling pathway	10	87	3.61e-05
MAPK signaling pathway	17	294	9.87e-05
AMPK signaling pathway	11	126	1.17e-04
Hippo signaling pathway	12	154	1.33e-04
Peroxisome	9	84	1.65e-04
Cellular senescence	13	186	1.69e-04
Retinol metabolism	9	91	2.77e-04
Carbon metabolism	10	121	3.89e-04
Pyruvate metabolism	6	38	5.49e-04
Tight junction	11	167	9.61e-04
Glutathione metabolism	7	64	1.00e-03
Steroid hormone biosynthesis	8	89	1.13e-03
TGF-β signaling pathway	8	95	1.60e-03
Fructose and mannose metabolism	5	35	2.65e-03
Protein processing in endoplasmic reticulum	10	163	2.73e-03
Purine metabolism	9	136	3.09e-03
Pyrimidine metabolism	6	58	3.25e-03
PI3K-Akt signaling pathway	15	358	4.90e-03
Endocytosis	12	270	8.70e-03
Linoleic acid metabolism	5	50	8.83e-03
Cell adhesion molecules (CAMs)	9	171	1.05e-02
Pentose phosphate pathway	4	32	1.15e-02
ECM-receptor interaction	6	86	1.49e-02
Lysosome	7	124	1.97e-02
Glycolysis/Gluconeogenesis	5	67	2.26e-02
p53 signaling pathway	5	71	2.72e-02
Glucagon signaling pathway	6	105	3.09e-02
Ribosome	8	175	3.15e-02
Tryptophan metabolism	4	48	3.38e-02

All significant pathways are presented with the number of differentially expressed genes (DEGs) (Input No.; Gene IDs for each pathway can be found in Supplemental Data) in each pathway’s complete annotated gene number (Total No.). Pathways were filtered by redundancy and only pathways with unique genes are presented. OB, obese; PPAR, peroxisome proliferator-activated receptor. P value is corrected for multiple testing with Benjamini–Hochberg.

Table 4.

KEGG pathways enriched in DEGs from the livers of female OB-offspring

Pathway	Input No.	Total No.	P Value
Metabolic pathways	40	1494	2.40e-10
Steroid biosynthesis	7	19	2.44e-08
PPAR signaling pathway	9	87	1.05e-06
Cholesterol metabolism	6	49	5.19e-05
Cell cycle	8	123	9.24e-05
Circadian rhythm	4	30	1.16e-03
Protein processing in endoplasmic reticulum	7	163	2.48e-03
Neurotrophin signaling pathway	6	121	3.11e-03
Tryptophan metabolism	4	48	4.58e-03
Spliceosome	6	133	4.66e-03
Antigen processing and presentation	5	91	5.51e-03
RNA transport	6	167	1.17e-02
Peroxisome	4	84	2.10e-02
Insulin signaling pathway	5	140	2.26e-02
Arachidonic acid metabolism	4	89	2.44e-02
MAPK signaling pathway	7	294	3.28e-02
Ras signaling pathway	6	233	3.76e-02
Fatty acid metabolism	3	61	4.53e-02

All significant pathways are presented with the number of differentially expressed genes (DEGs) (Input No.; Gene IDs for each pathway can be found in Supplemental Data) in each pathway’s complete annotated gene number (Total No.). Pathways were filtered by redundancy and only pathways with unique genes are presented. P value is corrected for multiple testing with Benjamini–Hochberg. OB, obese.

Table 5.

Overlapping pathways in the livers of male and female offspring

Pathway	Genes in males	Genes in females
Tryptophan metabolism	Kyat, Kynu, Acmsd, Inmt	Kyat1, Acat2, Acmsd, Inmt
Fatty acid metabolism	Acox1, Acaca, Acsl3, Scd1, Elovl6, Acat2, Fasn, Scd3, Fads1	Fads1, Acsl3, Acsl4
PPAR signaling pathway	Cyp7a1, Acsl3, Lpl, Acsl4, Angptl4, Sorbs1, Plin4, Plin5, Apoa2	Acox1, Acsl3, Cyp4a12b, Scd1, Angptl4, Fabp5, Scd3, Plin4, Cyp7a1, Apoa2
MAPK signaling pathway	Stmn1, Sos1, Hspa8, Map3k5, Vegfa, Dusp16, Irak4	Igf2, Hspa2, Stmn1, Casp3, Rasgrp3, Cacnb3, Erbb3, Hspa8, Ntrk2, Hspa1b, Tgfbr2, Arrb2, Relb, Map2k7, Myc, Fgf21, Klk1b4
Cell cycle	Pcna, Cdkn2c, Mcm6, Mcm5, Mcm4, Mcm3, Mcm2, Abl1	Hspa2, H2-T22, Ctsl, Hspa5, Hspa4, H2-Ab1, Calr, Hspa8, H2-Aa, Psme3, H2-Q2, Hspa1b, Ctss, Gm11127, Gm8909
Protein processing in endoplasmic reticulum	Sel1l, Hspa8, Hsph1, Hspa5, Ubqln1, Bak1, Map3k5	Hspa2, Sel1l, Pdia4, Hsph1, Hspa5, Calr, Hspa8, Hspa1b, Derl3, Map2k7
Peroxisome	Nudt12, Acsl3, Acsl4, Mpv17l	Pmp22, Mpv17l, Acsl3, Acox1, Pmvk, Acot8, Slc25a17, Hao2, Pex12
Antigen processing and presentation	Tapbp, H2-K1, Hspa8, H2-T22, Hspa5	Hspa2, H2-T22, Ctsl, Hspa5, Hspa4, H2-Ab1, Calr, Hspa8, H2-Aa, Psme3, H2-Q2, Hspa1b, Ctss, Gm11127, Gm8909
Cholesterol metabolism	Pcsk9, Lpl, Angptl4, Vdac1, Cyp7a1, Apoa2	Lipc, Abcg5, Sort1, Angptl4, Apoa4, Vdac1, Cyp7a1, Apoa2
Metabolic pathways	Atp5c1, Bhmt, Acsl3, Cyp51, Cdo1, Acsl4, Tymp, Oxct1, Cyp26a1, Galnt4, Mtmr4, Fdft1, Acmsd, Inmt, Dhcr7, Gpx3, Phospho1, Fads1, Sat1, Kynu, Idi1, Hyal1, Nampt, Hmgcr, Kyat1, Cyp2c39, Cyp7a1, Plaat3, Msmo1, Car2, Cyp2j9, Gsta1, Nudt12, Selenoi, Lss, B3gat3, Glo1, Sqle, Cndp2, Dhcr24	Fdps, Hmgcr, Acsl3, Uox, Pmvk, Nampt, Cyp2c23, Acly, Smox, Ugt2b1, Entpd4b, Ggct, G6pdx, Pigw, Coq7, Ndufab1, Sat1, Acox1, Gstt2, Pde6d, Upp2, Papss2, Impdh1, Atp5j, Smpd3, Glo1, Nmnat1, Cyp2c39, Bhmt, Sc5d, Fasn, Synj2, Prodh, Adh1, Inmt, Uprt, Tk1, Treh, Acot8, Gstm3, Gpd1, Gal3st1, Kyat1, Cyp7a1, Lipc, Plpp2, Mgst1, Lss, Pkm, Scd3, Scd1, Tymp, Cyp26a1, Pde4a, Acmsd, Bdh1, Cyp3a41b, Cyp3a41a, Cyp4a12b, Nmrk1, Hao2, Atp6v0c, Azin2, Gstp1, Hnmt, Pfkfb1, Acss2, Car1, Elovl6, Got1, B4galt7, Ak2, Pgm3, Acat2, Ipmk, Sord, Phospho2, Phospho1, Fads1, Gpam, Acaca, Ndufb9, Acacb, Pnp2, Gmppa, Pfkp, Aldoc, Cyp2c38, Gsta1, Pnpla3, Acy1

Abl1, c-abl oncogene 1, non-receptor tyrosine kinase; Acaca, acetyl-Coenzyme A carboxylase alpha; Acat2, acetyl-Coenzyme A acetyltransferase 2; Acmsd, amino carboxymuconate semialdehyde decarboxylase; Acot8, acyl-CoA thioesterase 8; Acox1, acyl-Coenzyme A oxidase 1, palmitoyl; Acsl3, acyl-CoA synthetase long-chain family member 3; Acsl4, acyl-CoA synthetase long-chain family member 4; Angptl4, angiopoietin-like 4; targeted mutation 1, Velocigene; Apoa2, apolipoprotein A-II; Arrb2, arrestin, beta 2; Bak1, BCL2-antagonist/killer 1; Cacnb3, calcium channel, voltage-dependent, beta 3 subunit; Calr, calreticulin; Casp3, caspase 3; Cdkn2c, cyclin dependent kinase inhibitor 2C; Ctsl, cathepsin L; Ctss, cathepsin S; Cyp4a12b, cytochrome P450, family 4, subfamily a, polypeptide 12B; Cyp7a1, cytochrome P450, family 7, subfamily a, polypeptide 1; Dusp16, dual specificity phosphatase 16; Elovl6, ELOVL family member 6, elongation of long chain fatty acids (yeast); Erbb3, erb-b2 receptor tyrosine kinase 3; Fabp5, fatty acid binding protein 5, epidermal; Fads1, fatty acid desaturase 1; Fasn, fatty acid synthase; Fgf21, fibroblast growth factor 21; Gm11127, predicted gene 11127; Gm8909, predicted gene 8909; H2-Aa, histocompatibility 2, class II antigen A, alpha; H2-Ab1, histocompatibility 2, class II antigen A, beta 1; H2-K1, histocompatibility 2, K1, K region; H2-Q2, histocompatibility 2, Q region locus 2; H2-T22, histocompatibility 2, T region locus 22; Hao2, hydroxyacid oxidase 2; Hspa1b, heat shock protein 1B; Hspa2, heat shock protein 2; Hspa4, heat shock protein 4; Hspa5, heat shock protein 5; Hspa8, heat shock protein 8; Hsph1, heat shock 105 kDa/110 kDa protein 1; Igf2, insulin-like growth factor 2; Inmt, indolethylamine N-methyltransferase; Irak4, interleukin-1 receptor-associated kinase 4; Klk1b4, kallikrein 1-related pepidase b4; Kyat1, kynurenine aminotransferase 1; Kynu, Kynureninase; Lpl, lipoprotein lipase; Map2k7, mitogen-activated protein kinase kinase 7; Map3k5, mitogen-activated protein kinase kinase kinase 5; Mcm2, minichromosome maintenance complex component 2; Mcm3, minichromosome maintenance complex component 3; Mcm4, minichromosome maintenance complex component 4; Mcm5, minichromosome maintenance complex component 5; Mcm6, minichromosome maintenance complex component 6; Mpv17l, Mpv17 transgene, kidney disease mutant-like; Myc, myelocytomatosis oncogene; Ntrk2, neurotrophic tyrosine kinase, receptor, type 2; Nudt12, nudix (nucleoside diphosphate linked moiety X)-type motif 12; Pcna, proliferating cell nuclear antigen; Pdia4, protein disulfide isomerase associated 4; Pex12, peroxisomal biogenesis factor 12; Plin4, perilipin 4; Plin5, perilipin 5; PPAR, peroxisome ploliferator-activated receptor; Pmp22, peripheral myelin protein 22; Pmvk, phosphomevalonate kinase; Psme3, proteaseome (prosome, macropain) activator subunit 3 (PA28 gamma, Ki); Rasgrp3, RAS, guanyl releasing protein 3; Relb, oncogene related B; Scd1, stearoyl-coenzyme A desaturase 1; Scd3, stearoyl-coenzyme A desaturase 3; Sel1l, sel-1 suppressor of lin-12-like (C. elegans); Slc25a17, solute carrier family 25 (mitochondrial carrier, peroxisomal membrane protein), member 17; Sorbs1, sorbin and SH3 domain containing 1; Sos1, SOS Ras/Rac guanine nucleotide exchange factor 1; Stmn1, stathmin 1; Tapbp, TAP binding protein; Tgfbr2, transforming growth factor, beta receptor II; Ubqln1, ubiquilin 1; Vegfa, vascular endothelial growth factor A; v-rel, avian reticuloendotheliosis viral.

Determination of protein levels in livers from the expanded cohort of 3-mo-old offspring further support the transcriptomic data. Levels of Cpt1a were increased (79%; Fig. 4A) and levels of Scd1 were decreased (40%; Fig. 4B) in OB-offspring in both sexes. Furthermore, the transcriptomic signature suggesting impaired metabolism was supported by the reduced levels of several members of the electron transport chain (ETC). Complex V (Atp5a), complex III (Uqcrc2), and complex II (Sdhb) were decreased in livers from female (Fig. 5A) and male offspring (Fig. 5B). In males, complex I (Ndufb8) was also decreased (20%; P = 0.05).

Figure 4.

Relative protein levels for targets coregulated in livers from 3-mo-old offspring. Representative images of immunoblots are shown for Cpt1a (A) and Scd1 (B) labeled next to their respective bands based on apparent molecular weights. A representative band for obese dams (OB-offspring; OB; n = 5/sex) or control dams (CON-offspring; CON; n = 4/sex) is shown. After normalization to total protein stain (Amido Black), the mean density of control samples was assigned a value of 1; summarized Western blot results shown in histogram (B) as means ± SE and significant differences (*P = 0.05 and **P = 0.01) are indicated.

Figure 5.

Relative protein levels for oxidative phosphorylation in livers from 3-mo-old offspring. Representative images of immunoblots are shown for complexes of the electron transport chain including complex I subunit Ndufb8, complex II subunit Sdhb, complex III-core protein 2 Uqcrc2, complex V α subunit Atp5a, and cytochrome c oxidase complex member Mtco1. The proteins are labeled next to their respective bands based on apparent molecular weights. A, B: representative band for obese female (A) and male (B) offspring. A representative band for obese (OB)-offspring (OB; n = 5/sex) or control (CON)-offspring (CON; n = 5/sex) is presented. After normalization to total protein stain (Amido Black), the mean density of control samples was assigned a value of 1; summarized Western blot results shown in histograms (A and B) as means ± SE and significant differences (*P = 0.05 and **P = 0.01) are indicated.

Persistent Changes in Liver Responses

The offspring liver transcriptome was more markedly altered in males than in females in response to maternal obesity both in the fetus and at 3 mo of age with nearly a twofold increase in DEGs in males (Fig. 6). Interestingly, in both males and females, the transcriptomic responses shared between the late gestation fetus and the adult offspring were minimal (Fig. 6, B and C).

Figure 6.

Venn diagram of differentially expressed genes in the liver following maternal obesity. A: comparison of differentially expressed genes (DEGs) between fetuses and offspring in female (B) and male (C) liver transcripts. For fetuses, n = 10/treatment as five males and five females. Male obese (OB)-fetuses (n = 5) compared with male control (CON)-fetuses (n = 5) were analyzed then female OB-fetuses (n = 5) compared with female CON-fetuses (n = 5). Subsequently, male OB-offspring (n = 5) compared with male CON-offspring (n = 5) were analyzed then female OB-offspring (n = 5 for liver and n = 4 for muscle) compared with female CON-offspring (n = 5 for liver and n = 4 for muscle).

The Skeletal Muscle Transcriptome in Fetuses of Obese Dams

When differential transcript abundance in the fetal muscle was analyzed for each sex independently, male OB-fetuses had 209 DEGs (113 down- and 96 upregulated) and female OB-fetuses had 289 DEGs (148 down- and 141 upregulated) compared with their respective controls. In male fetuses of obese dams compared with control male fetuses, the genes most downregulated were src-like adaptor (Sla), apolipoprotein O (Apoo), and nuclear receptor 1h3 (Nr1h3); whereas, genes most upregulated included renin binding protein (Renbp), protein phosphatase 1 (Ppp1r8), dysbindin (Dbndd2), and transcriptional activators (Fubp1, Eya1, Foxa2; Fig. 7A). In contrast, in muscle from female fetuses discoidin (Dcbld1) and tRNA methyltransferase 1 (Trmt1) were the two most downregulated genes; whereas, enhancer trap locus 4 (Etl4) and secretograninin III (Scg3j) were the two most upregulated genes in response to maternal obesity (Fig. 7B). Only 14 DEGS (2.9%) were common in males and females (Fig. 7C), including upregulated transcripts muscleblind splicing factor 2 (Mbnl2), doublecortin-like kinase 2 (Dclk2), and osteomodulin (Omd) and downregulated transcripts suppression of tumorigenicity (St7) and glutathione synthetase (Gss).

Figure 7.

Differentially expressed genes with the greatest fold change following maternal obesity in the skeletal muscle of male (A) and female (B) obese (OB)-fetuses. A: the expression of the 10 most up- and downregulated genes are plotted that have the greatest change (up and down) in expression in OB-fetuses compared to control (CON)-fetuses in males (A) and females (B). The x-axis represents the log₂ transformation of fold changes between experimental groups calculated from normalized counts per million. Genes are listed on the y-axis as the official HGNC symbol. C: Venn diagram of genes identified in separate pairwise comparisons in the skeletal muscle for male OB-fetuses compared with CON-fetuses as well as the female comparison.

Top enriched pathways in male fetal skeletal muscle of obese dams were predominantly related to metabolism (Table 6). In females, top enriched pathways were RNA synthesis and transport (Table 7). However, despite little overlap in the DEGs, skeletal muscle pathways of both male and female fetuses of obese dams were enriched for MAPK, mechanistic target of rapamycin (mTOR), and PI3K-AKT signaling pathways (Tables 6 and 7).

Table 6.

KEGG pathways enriched in DEGs from the skeletal muscle of male OB-fetuses

Pathway	Input No.	Total No.	P Value
Metabolic pathways	28	1494	1.84e-06
Steroid biosynthesis	3	19	3.05e-03
MAPK signaling pathway	8	294	4.04e-03
PI3K-Akt signaling pathway	8	358	1.16e-02
Purine metabolism	5	136	1.18e-02
Ferroptosis	3	41	1.62e-02
Insulin resistance	4	110	2.76e-02
Thyroid hormone signaling pathway	4	118	3.27e-02
2-Oxocarboxylic acid metabolism	2	19	3.62e-02
Spliceosome	4	133	4.30e-02
Ubiquitin mediated proteolysis	4	139	4.74e-02
Biosynthesis of amino acids	3	78	5.32e-02
EGFR tyrosine kinase inhibitor resistance	3	80	5.55e-02
mTOR signaling pathway	4	155	5.92e-02

All significant pathways are presented with the number of differentially expressed genes (DEGs) (Input No.) in each pathway’s complete annotated gene number (Total No.). Pathways were filtered by redundancy and only pathways with unique genes are presented. P value is corrected for multiple testing with Benjamini–Hochberg. mTOR, mechanistic target of rapamycin; OB, obese.

Table 7.

KEGG pathways enriched in DEGs from the skeletal muscle of female OB-fetuses

Pathway	Input No.	Total No.	P Value
Aminoacyl-tRNA biosynthesis	6	66	3.05e-04
RNA transport	8	167	9.23e-04
MAPK signaling pathway	10	294	1.71e-03
Endocytosis	9	270	3.64e-03
mTOR signaling pathway	6	155	1.22e-02
Protein processing in endoplasmic reticulum	6	163	1.49e-02
Cortisol synthesis and secretion	4	69	1.70e-02
Adipocytokine signaling pathway	4	71	1.81e-02
Ras signaling pathway	7	233	1.85e-02
Biosynthesis of amino acids	4	78	2.30e-02
Regulation of actin cytoskeleton	6	217	4.16e-02
PI3K-Akt signaling pathway	8	358	4.19e-02
Wnt signaling pathway	5	162	4.86e-02

DEGs, differentially expressed genes; mTOR, mechanistic target of rapamycin; OB, obese.

The Skeletal Muscle Transcriptome in 3-Mo-Old Offspring of Obese Dams

Male 3-mo-old OB-offspring had 275 DEGs (115 down- and 160 upregulated) in muscle compared with male CON-offspring. In female offspring, maternal obesity was associated with 580 DEGs of which 305 were downregulated and 275 were upregulated. In male offspring of obese dams, the skeletal muscle genes most downregulated compared with control male offspring encoded structural proteins such as integrin α 3 (Itga3) as well as DNA binding elements (Brcc3, Fubp1, Sting1), a nuclear receptor (Nr4a3), and hexokinase 1 (Hk1; Fig. 8A). Genes most upregulated in male offspring were involved in transcriptional regulation including transcription elongation factor (Tcea1) and metal response element binding factor (Mtf2; Fig. 8A). In contrast, the most downregulated genes in the female offspring of obese dams encode for transporters such as solute carrier (Slc4a1) and lipocalin 2 (Lcn2) as well as calcium binding proteins (s100a8 and s100a9; Fig. 8B). The most upregulated genes in female offspring included dysbindin (Dbndd2), leptin (Lep), and glucokinase (Gck). Interestingly, nuclear receptor Nr4a3 was one of the most downregulated genes in males, but one of the most upregulated genes in females.

Figure 8.

Differentially expressed genes with the greatest fold change following maternal obesity in the skeletal muscle of male (A) and female (B) obese (OB)-offspring. A: the expression of the 10 most up- and downregulated genes are plotted that have the greatest change (up and down) in expression in OB-fetuses compared with control (CON)-fetuses in males (A) and females (B). The x-axis represents the log₂ transformation of fold changes between experimental groups calculated from normalized counts per million. Genes are listed on the y-axis as the official HGNC symbol. C: Venn diagram of genes identified in separate pairwise comparisons in the liver for male OB-offspring compared with CON-offspring as well as the female comparison.

Top pathways enriched in male offspring of obese dams included MAPK signaling, metabolic pathways, specifically arginine and proline metabolism, as well as PI3K-Akt and TGF-β signaling (Table 8). In some cases, genes involved in the same metabolic pathway were regulated in the opposite direction in response to maternal obesity in males, such as increased Akt2 but decreased Akt3 expression. In the female offspring, there were 28 significantly enriched pathways, a longer list compared with male offspring, including the proteasome and endocytosis as well as metabolic signaling pathways such as AMP-activated protein kinase (AMPK), glucagon, insulin, and mTOR (Table 9). The 25 DEGs that overlapped between males and females (Fig. 8C) were significantly enriched for PI3K-Akt signaling, proteasome, and metabolic pathways consistent with the iterative comparative functional analysis for each sex (Fig. 8). In both males and females, metabolic genes such as pyruvate carboxylase (Pcx), 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (Pfkfb3), and myosin 1E (Myo1e) were upregulated in OB-offspring and genes such as lactate dehydrogenase (Ldha) and cationic amino acid transporter 2 (Slc7a2) were downregulated. Adiponectin receptor 1 (Adipor1) expression was significantly increased in both male and female skeletal muscle. Furthermore, analysis of the 25 DEGs regulated in both sexes revealed a tight network of exocytosis-related genes comprised mostly of members of the exocyst complex family (Exoc3), which have been shown to mediate insulin-stimulated GLUT4 exocytosis in skeletal muscle (64).

Table 8.

KEGG pathways enriched in DEGs from the skeletal muscle of male OB-offspring

Pathway	Input No.	Total No.	P Value
MAPK signaling pathway	10	294	1.32e-03
Metabolic pathways	24	1494	4.50e-03
Arginine and proline metabolism	4	50	6.33e-03
Hypertrophic cardiomyopathy	5	91	6.75e-03
Focal adhesion	7	199	7.88e-03
Cellular senescence	6	186	1.93e-02
EGFR tyrosine kinase inhibitor resistance	4	80	2.07e-02
Spliceosome	5	133	2.15e-02
Phospholipase D signaling pathway	5	149	3.00e-02
PI3K-Akt signaling pathway	8	358	3.10e-02
TGF-β signaling pathway	4	95	3.12e-02
Regulation of actin cytoskeleton	6	217	3.21e-02
Human papillomavirus infection	8	361	3.23e-02
AGE-RAGE signaling pathway in diabetic complications	4	101	3.61e-02

All significant pathways are presented with the number of differentially expressed genes (DEGs) (Input No.) in each pathway’s complete annotated gene number (Total No.). Pathways were filtered by redundancy and only pathways with unique genes are presented. P value is corrected for multiple testing with Benjamini–Hochberg. HCM, hypertrophic cardiomyopathy; OB, obese.

Table 9.

KEGG pathways enriched in DEGs from the skeletal muscle of female OB-offspring

Pathway	Input No.	Total No.	P Value
Proteasome	10	46	2.36e-07
mTOR signaling pathway	14	155	5.96e-06
Endocytosis	18	270	9.22e-06
AMPK signaling pathway	11	126	1.01e-04
Insulin signaling pathway	11	140	2.30e-04
AGE-RAGE signaling pathway in diabetic complications	9	101	4.56e-04
Glucagon signaling pathway	9	105	5.75e-04
FoxO signaling pathway	10	132	6.09e-04
Parathyroid hormone synthesis, secretion and action	9	107	6.45e-04
Insulin resistance	9	110	7.66e-04
Longevity regulating pathway	8	90	1.04e-03
Ribosome	11	175	1.17e-03
Metabolic pathways	43	1494	1.18e-03
Cellular senescence	11	186	1.77e-03
MAPK signaling pathway	14	294	2.27e-03
Wnt signaling pathway	10	162	2.32e-03
Adipocytokine signaling pathway	6	71	6.44e-03
Focal adhesion	10	199	8.15e-03
cGMP-PKG signaling pathway	9	173	1.04e-02
ErbB signaling pathway	6	84	1.25e-02
PPAR signaling pathway	6	87	1.42e-02
Protein processing in endoplasmic reticulum	8	163	2.10e-02
Relaxin signaling pathway	7	131	2.23e-02
Purine metabolism	7	136	2.58e-02
HIF-1 signaling pathway	6	114	3.57e-02
Phospholipase D signaling pathway	7	149	3.57e-02
Oxytocin signaling pathway	7	154	4.02e-02
PI3K-Akt signaling pathway	12	358	4.17e-02

DEGs, differentially expressed genes; HIF-1, hypoxia-inducible factor 1; mTOR, mechanistic target of rapamycin; OB, obese; PPAR, peroxisome proliferator-activated receptor.

Determination of protein levels for ETC complexes in an expanded cohort of 3-mo-old offspring further support the transcriptomic data. In males and females, abundance of complex V (Atp5a), complex III (Uqcrc2), and complex I (Ndufb8) proteins were decreased in skeletal muscle from OB-offspring compared with CON-offspring (Fig. 9). Levels of Cpt1a were not different in either sex, consistent with the transcriptomic data (Fig. 10A). Female skeletal muscle was uniquely enriched for mTOR signaling and the levels of mTOR associated proteins suggest increased mTOR signaling in females but not in males. In skeletal muscle from female OB-offspring, total levels of 4EBP-1 (26%: Fig. 10B) and phosphorylated rpS6 (37%; Fig. 10C) were increased while levels of total rpS6 were not different (Fig. 10D).

Figure 9.

Relative protein levels for oxidative phosphorylation in skeletal muscle from 3-mo-old offspring. Representative images of immunoblots are shown for complexes of the electron transport chain including complex I subunit Ndufb8, complex III-core protein 2 Uqcrc2, complex V α subunit Atp5a, and cytochrome c oxidase complex member Mtco1. Complex II subunit Sdhb was included but not detected in this antibody cocktail in skeletal muscle. The proteins are labeled next to their respective bands based on apparent molecular weights. A representative band for obese (OB)-offspring (OB) or control (CON)-offspring (CON) is presented (A, female; B, male). After normalization to total protein stain (Amido Black), the mean density of control samples was assigned a value of 1; summarized Western blot results shown in histograms (A and B) as means ± SE and significant differences (*P = 0.05 and **P = 0.01) are indicated.

Figure 10.

Relative protein levels for targets identified in RNAseq in skeletal muscle from 3-mo-old offspring. Representative images of immunoblots are shown for Cpt1a (A) and mTOR signaling proteins 4EBP-1 (B), P-rpS6 (C), and rpS6 (D). Each is labeled next to their respective bands based on apparent molecular weights. A representative band for obese (OB)-offspring (OB) or control (CON)-offspring (CON) is shown. After normalization to total protein stain (Amido Black), the mean density of control samples was assigned a value of 1; summarized Western blot results shown in histogram (B) as means ± SE and significant differences (*P = 0.05) are indicated.

Persistent Changes in Skeletal Muscle

There were no genes differentially expressed in both sexes and ages (Fig. 11A). Similar to the liver, ∼2% of DEGs in skeletal muscle of fetuses were persistent at 3 mo-old in females (Fig. 11B) and males (Fig. 11C). However, comparison of the enriched pathways for each list of DEGs revealed some commonality to both ages, notably hypertrophic responses in muscle tissues.

Figure 11.

Venn diagram of differentially expressed genes in the liver following maternal obesity (A). Comparison of differentially expressed genes (DEGs) between fetuses and offspring in females (B) and males (C) skeletal muscle transcripts.

DISCUSSION

Using unbiased transcriptomic analysis, we identified differentially expressed genes in liver and skeletal muscle in fetuses and adult offspring of obese dams. These changes were strikingly sexually dimorphic and with marked differences between fetuses and adults. Functional pathways regulated in offspring liver and skeletal muscle by maternal obesity were also distinct in males and females, which broadly included metabolic processes involving fatty acids and lipids, protein processing, and cell proliferation, as well as signaling pathways such as MAPK, AMPK, mTOR, and PI3K-Akt. These differentially expressed genes and regulated pathways are candidates to mediate the programming of hepatic dysfunction and insulin resistance in offspring that we previously reported in this clinically relevant model of maternal obesity (44). Importantly, adult female offspring of obese mice exhibited a less pronounced metabolic phenotype than male offspring (44), in general agreement with the sexual dimorphic liver and muscle transcriptomic response to maternal obesity demonstrated in the current study.

In humans, it is well established that sex influences the onset and development of metabolic disease throughout life, often occurring before or independent of sex hormone signaling (65). For example, hepatic steatosis is more common in men than women (66) and studies show women have higher fatty acid clearance and utiliziation than men (67–69). Furthermore, numerous studies have found that males and females have different responses to an adverse intrauterine environment (70–72) and more recently, sex has been associated with altered responses to maternal obesity throughout the perinatal window (20, 73). Although malesexhibit a more pronounced metabolic phenotype in response to maternal obesity, the mechanisms underlying the increased susceptibility of the male offspring or the conferred protection of the female offspring are poorly understood, but likely begin in utero. Increasingly, published studies in animal models of maternal overnutrition/obesity describe male and female responses independently (21, 33, 44). In the present study, the bioinformatic analysis performed with males and females combined to evaluate the response to maternal obesity revealed few to no differentially expressed transcripts. Only when differential expression testing was performed for males and females separately, we were able to detect hundreds of differentially expressed transcripts that are associated with candidate pathways that mediate metabolic programming in response to maternal obesity. Not surprisingly, comparison of the differentially expressed genes in response to maternal obesity in male and female offspring revealed little overlap (less than 10%), similar to limited previous reports utilizing omics in mouse models of maternal obesity induced by a Western-style diet (56).

A common observation in animal models of maternal overnutrition is the increased risk of the offspring to develop NAFLD and insulin resistance, with males typically more susceptible than females (66). In our mouse model of obesity, dams are fed a diet moderately high in fat and sugar, similar to diets reported in overweight/obese women (30, 31), which results in larger offspring at birth (43) and by 3 mo of age, male offspring exhibit greater fat mass, insulin resistance, and hepatic steatosis despite normal chow diet after weaning (44). These findings are similar to studies of over-nourished pregnant rats reporting greater fat mass, insulin resistance, metabolic dysfunction, and hepatic steatosis in male offspring when challenged with high-fat diet (26). Transcriptomic analysis of male and female livers of the near-term fetuses revealed extensive sex-dependent transcriptional responses. The number of DEGs identified in the livers of the male OB-fetuses was twofold greater than in the female OB-fetuses, indicating male fetal livers were more responsive and, possibly, vulnerable to maternal obesity during pregnancy. This was also the case at 3 mo of age, where males exhibited twice as many significant changes in genes highly enriched in metabolic pathways than females.

Male OB-fetuses and 3-mo-old offspring have increased body weight as compared with their respective controls, and male offspring are insulin resistant and demonstrate overt hepatic steatosis. Thus, it was of particular interest to investigate the metabolic programming that could explain the vulnerability of males to NAFLD, which is driven by accumulation of triglyceride in hepatocytes (43, 74). The current transcriptomic analysis demonstrated upregulation of CPT1a, which encodes for an enzyme required for mitochondrial FA oxidation, in response to maternal obesity in both fetal male and female livers. These findings suggest a common mechanism for male and female offspring of obese dams to increase liver FA oxidation. However, in male fetuses of obese dams, we observed marked reductions in the expression of ATP synthase subunits and members of the electron transport chain consistent with decreased oxidative capacity compared with the fetal female liver. In the fetal male liver, there was also an upregulation of the crucial lipogenic enzyme ATP-citrate lysase (Acly) that has been demonstrated in individuals with NAFLD (75). Together, the data are consistent with the interpretation that OB-fetal livers have increased hepatic triglyceride (TG) and the gene expression changes in males suggest reduced capacity for fatty acids to enter oxidative pathways and increased capacity for lipogenesis which may be the underlying cause of increased lipid content in fetal livers we have previously reported (44).

By 3 mo of age, after being fed a control diet since weaning, hepatic Cpt1a is increased at the transcript and protein level. Furthermore, expression of genes that encode for FA-related proteins such as long chain acyl-CoA synthetases (Acsl3) and fatty acid desaturase 1 (Fads1) were downregulated in liver from both males and females, indicating maternal obesity initiated lasting changes in lipid metabolism relevant to liver dysfunction. In a recent small cohort of children, decreased Fads1 expression was associated with the development of NAFLD in childhood (76). Hepatic expression of Acsl3 is required for fatty acid incorporation into phospholipids and has been linked to lower concentrations of hepatic fatty acid and TG (77). We know the liver displays overt steatosis in the male 3-mo-old offspring of obese dams and while no direct pathways were implicated by the transcriptomic data, downregulation of FA synthesis enzymes stearoyl-coenzyme A desaturases (Scd1 and Scd3) was found only in male offspring. Scd1 abundances were confirmed at the protein level, identifying lower Scd1 in male but not female offspring. Furthermore, lipoprotein lipase (Lpl), the critical enzyme that hydrolyzes TG to release fatty acids, was downregulated in the livers of female OB-offspring and upregulated in males. Lpl is implicated in steatosis and the higher expression in males and lower expression in females suggests a sex-dependent mechanism by which accumulated hepatic lipids are utilized. This is supported by the livers of fetal and adult male offspring having upregulated expression of apolipoproteins A-II (Apoa2) and A-IV (Apoa4) that encode for proteins involved in hepatic VLDL formation and whose expression is strongly correlated with hepatic TG concentrations in mouse models of steatosis (78, 79).

Skeletal muscle represents a major metabolic organ responsible for a majority of whole body insulin-stimulated glucose uptake and therefore, perturbations in development may influence an individual’s long-term ability to maintain glucose homeostasis. Developmental consequences in skeletal muscle have been studied in various models of maternal obesity. In rodent offspring of dams fed a high-fat diet, increased intramuscular fat accretion (46), decreased fiber size in skeletal muscle at weaning (46, 52), and reduced muscle force in the adult offspring, regardless of the postweaning diet (53) have been reported. In the present study, the transcriptomic response in skeletal muscle of OB-fetus was largely in growth-related pathways such as mTOR activation in both sexes but otherwise was distinct between males and females. In males, transforming growth factor β 1 (Tgfb1) and insulin were the primary growth regulators identified by gene expression data, while in females the response involved Wnt signaling, adipocytokine action, and translational mechanisms.

Skeletal muscle in rodents at birth is at a quarter of their mature size and their muscles therefore continue to develop postnatally. Thus, the dramatic changes that persist in the adult offspring of obese dams fed a control diet from weaning suggest on-going programming of muscle metabolism following the intrauterine obesogenic environment. At 3 mo of age, the genes most changed encoded primary metabolic pathways; in both males and female muscle this included increased expression of enzymes in glucose metabolism such as the glycolytic regulating 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (Pfkfb3) as well as pyruvate carboxylase. Adult female skeletal muscle had a gene signature consistent with greater glucose entry as well as increased expression of the rate-limiting glucokinase. This may suggest that skeletal muscle from adult females of obese mothers take up and metabolize glucose more effectively and may explain why females have normal glucose tolerance at 3 mo of age compared with male offspring of obese dams (44). Furthermore, pathways such as mTOR signaling were significantly enriched in both sexes in the fetus; however, enrichment of these pathways was only observed in one sex at 3-mo-old. For example, mTOR signaling was enriched in fetal skeletal muscle and 3-mo-old female offspring, but not in the 3-mo-old male offspring. The protein abundance of mTOR-associated proteins suggests activation of this signaling pathway in female skeletal muscle at 3 mo of age, representing an additional mechanism underlying the normal glucose tolerance observed in female OB-offspring but not in males (44).

In conclusion, maternal obesity causes marked changes in offspring liver and muscle gene expression that differed distinctly between females and males. Our results provide insights into the mechanisms underpinning metabolic programming following exposure to maternal obesity. In both males and females, the transcriptomic responses in the fetus were strikingly different than those at 3 mo, implicating more complicated programming mechanisms throughout development that warrant further study.

GRANTS

Thomas Jansson and Theresa L. Powell are supported by the National Institute of Child Health and Human Development (NICHD) Grant HD065007. Amy C. Kelly is supported by NICHD Grant T32 HD007186. Thomas Jansson is supported by NIH Grant UL1 TR002535.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.C.K., F.R. T.L.P., and T.J. conceived and designed research; A.C.K. and F.R. performed experiments; A.C.K. and J.C. analyzed data; A.C.K. and J.C. interpreted results of experiments; A.C.K. prepared figures; A.C.K. and F.R. drafted manuscript; A.C.K., F.R., J.C., L.A.C., T.L.P., and T.J. edited and revised manuscript; A.C.K., F.R., J.C., L.A.C., T.L.P., and T.J. approved final version of manuscript.

ENDNOTE

At the request of the authors, readers are herein alerted to the fact that additional materials related to this manuscript may be found https://doi.org/10.6084/m9.figshare.20344047.v1. These materials are not a part of this manuscript and have not undergone peer review by the American Physiological Society (APS). APS and the journal editors take no responsibility for these materials, for the website address, or for any links to or from it.