High maternal adiposity during pregnancy programs an imbalance in the lipidome and predisposes to diet-induced hepatosteatosis in the offspring

Non-alcoholic fatty liver disease (NAFLD) is the hepatic manifestation of the metabolic syndrome (MetS) and has grown in prevalence from ∼25% to nearly 40% over the past couple of decades, primarily due to rising obesity rates [1]. Consequent to the declining age of obesity-onset, a greater number of women are obese at the time of pregnancy, approximately 25% of women in the USA [2]. Human studies have established a relationship between maternal obesity and later life development of components of MetS in the offspring, including obesity, insulin resistance and dyslipidemia [3,4]. A recent study in a Swedish cohort of young adults revealed that being born to an obese mother increased the risk for developing NAFLD in adulthood by 3-fold [3]. Therefore, exposure to an abnormal intrauterine milieu may be an important but underappreciated risk factor for NAFLD.

Declining adipose tissue function is a key characteristic in the transition to metabolically unhealthy, hypertrophic obesity. Dysfunctional adipose tissue is characterized by chronic inflammation, insulin resistance, dysregulated lipolysis and a failure to buffer excess lipids, leading to lipid spillover and ectopic deposition in peripheral organs such as the liver. Therefore, adiposopathy is a key trigger for the onset of obesity-associated hepatic steatosis. Recently, we showed that high pre-pregnancy adiposity programs an increased risk for diet-induced adiposopathy in the offspring, in a sex-specific manner [5]. While an effect of the intrauterine environment on diet-induced perturbations in adipose tissue function and lipid homeostasis was observed in both male and female offspring born to overweight dams, females from normal pregnancies were less vulnerable to adverse cardiometabolic effects of high-fat diet. In the present study, we sought to determine the influence of the intrauterine environment and sex on diet-induced hepatic steatosis and lipid metabolism.

All experiments involving animals were conducted at the University of Calgary in Calgary, Canada. All experimental protocols were approved by the University of Calgary Animal Care Committee (AC21-0132) and conducted in accordance with guidelines by the Canadian Council on Animal Care Ethics. Female mice heterozygous for deletion of the leptin receptor (Het_db) (Jackson Laboratory, stock #: 000697) were used as a model of high pre-pregnancy adiposity, as characterized in our previous study. Virgin 14-week old Het_db or wild-type (Wt) C57BL/6J females were mated with Wt males. Following delivery, offspring of Het_db dams were genotyped by qPCR and only Wt offspring used, to isolate the effect of the in utero environment. Pups were weaned on post-natal day (PND) 21. Offspring were randomly assigned to a high-fat/high-fructose (HFFD) (Research Diets Inc. D08040105I) or control diet (CD) (Research Diets Inc. D12450Ki) for 15 weeks. Following the diet period, mice were induced under isoflurane inhalation and euthanized by decapitation. Blood was collected in heparinized syringes and plasma collected by centrifugation. Livers were collected and either flash frozen in liquid nitrogen or fixed in 4% paraformaldehyde for sectioning.

Prior to euthanasia, body composition was measured by time domain nuclear magnetic resonance (TD-NMR) spectroscopy (LF90II, Bruker). Whole-body fat and lean body mass were expressed as a percentage of body weight. Fasting NEFA levels were measured using a colorimetric assay (ab65341, Abcam) according to the manufacturer’s instructions. Samples were run in triplicate and absorbance read on a SpectraMax ID3 microplate reader.

Fixed hepatic tissue was embedded in paraffin and 4µm sections mounted on slides by the University of Calgary Faculty of Veterinary Medicine Diagnostic Services Unit. Sections were stained with hematoxylin and eosin (H&E) or picrosirius red stain (ab150681, Abcam). Steatosis, inflammation, ballooning, and fibrosis were evaluated blind by an experienced histopathologist using the steatosis, activity, fibrosis (SAF) scoring system. Grades 0-3 of steatosis were defined as <5%, 5–33%, 33–66% and >33% steatosis, respectively. Inflammation was scored as 0 (absence of foci), 1 (≤ 2 foci) or 2 (>2 foci). Hepatocyte ballooning was absent (grade 1), apparent by clusters of enlarged hepatocytes with rarified cytoplasm (grade 2) or with more severe hepatocyte enlargement (grade 3). Fibrosis scores 0–3 were defined as the absence of fibrosis, perisinusoidal or periportal fibrosis, perisinusoidal and periportal fibrosis, bridging fibrosis and cirrhosis, respectively. Representative images of H&E stained sections were captured on a Nikon TI Eclipse brightfield microscope at 10× magnification.

Total RNA was isolated from flash frozen liver using the RNeasy Mini Kit (74104, Qiagen), assessed for integrity using a TapeStation assay (University of Calgary Genomic Services), and quantified by nanophotometer (Implen). After digestion of DNA, cDNA was generated from 2 µg RNA using the High-Capacity Reverse Transcription kit (4368814, Applied Biosystems). Samples were assayed in triplicate for RT-qPCR using SYBR Green master mix (A25741, Applied Biosystems). Amplification was performed using the following settings: 50°C × 2 min, 95°C × 2 min, 95°C × 1 s, and 58.9°C × 30 s for 40 cycles on the QuantStudio 5 Real Time PCR System. Primers were designed using NCBI and validated with melt curve analysis (Table 1). Albumin was used as a reference gene for calculation of mRNA expression.

Lipid extraction and thin layer chromatography was performed by the Vanderbilt University Medical Center Lipid Lab. Briefly, thin layer chromatography was used to separate lipids extracted using the Folch method. Individual lipid classes were visualized by rhodamine 6G and methylated for extraction. Methylated fatty acids were identified by comparing the retention times to known standards. Odd chain fatty acid standards were used for lipid quantification.

The Δ9 desaturation index was calculated as the ratio of precursors [palmitic acid (C16:0), stearic acid (C18:0)] to products [palmitoleic (C16:1), oleic acid (C18:1)] and reflects the activity of stearoyl-CoA 9-desaturase-1 (SCD1), an enzyme that catalyzes the rate limiting step in the biosynthesis of substrates (C16:1 and C18:1) for incorporation into triglycerides and cholesterol. The de novo lipogenesis (DNL) index, reflecting the rate of endogenous FA biosynthesis, was calculated as the ratio of palmitic acid (C16:0) to linoleic acid (C18:2n6). The ratio of C18:0-to-C16:0 was calculated to reflect the elongation index, reflecting the chain elongation of C16:0 catalyzed by fatty acid elongase 6 (Elov6). The sums of monounsaturated fatty acids (MUFA), saturated fatty acids (SFA) and polyunsaturated fatty acids (PUFA) were calculated, and levels expressed as a percentage of total lipid species.

Litter effects were controlled by including 1-2 male and female mice per litter in each analysis. Two-way ANOVA was used to determine main effects of diet and intrauterine environment (offspring group), while Sidak’s multiple comparison test in post-hoc analysis determined differences between groups. Normal distribution of the data was assessed by the Shapiro–Wilk test. Non normally distributed data were log transformed and assessed for statistical significance by two-way ANOVA. P-values for transformed data are presented on non-transformed graphs, where indicated in the figure legends. A contingency table was created with histological scoring counts for each group and differences between groups determined using the Chi-squared test. Analyses were performed using GraphPad version 9.0. Comparisons were considered significant when P<0.05 and data are expressed as mean ± SEM.

HFFD was used to simulate the effects of chronic Western Diet consumption and its effect on adiposity in offspring of overweight or lean pregnancies. Following HFFD, whole-body fat mass was similarly increased in males and females (Figure 1A,B). Fat mass was higher (Figure 1A,B) and lean body mass lower (Figure 1C,D) in HFFD fed groups relative to CD fed groups. In females, exposure to the intrauterine environment of an overweight dam led to significantly greater adiposity following HFFD than offspring of lean dams. In male offspring of Wt dams, HFFD led to an increase in fasting FFA, while FFA levels were similar in HFFD-fed male offspring born to Wt dams and Het_db offspring fed a CD or HFFD (Figure 1E). There was no impact of diet or intrauterine environment on fasting plasma FFA levels in females (Figure 1F).

To evaluate the degree of hepatocellular lipid accumulation in offspring of lean or overweight dams, histopathologic sections of offspring liver were evaluated by a trained histopathologist for steatosis, ballooning, and inflammation. In male CD-fed offspring born to Wt or Het_db dams, no steatosis or hepatocellular ballooning was observed. Grade 3 steatosis was observed in 43% and 10% of samples in CD-fed females born to Wt and Het_db dams, respectively, while grade 2 was observed in 10–14% of female CD samples. Grade 1 and 2 steatosis was observed in 14% and 29%, respectively, in HFFD-fed control males; while in HFFD-fed control females, 33% of samples were grade 2 and 50% were grade 3. In Chi-square analyses, HFFD had a significant effect on hepatic steatosis in only offspring born to Het_db pregnancies, and this was the case for both sexes (Figure 2A–C). Hepatocellular ballooning was observed in only HFFD-fed males and HFFD-fed female offspring from Het_db dams, although there were no significant effects of diet or intrauterine environment (Figure 2D,E).

There was no apparent fibrosis in any male groups, whereas grade 1 or 2 fibrosis was noted in a few female samples with no significant differences between groups. Similarly, no differences in scoring for inflammation were noted.

We next examined the effect of intrauterine environment and diet on genes involved in inflammatory and fibrotic processes. In male offspring, there was a significant effect of diet on the expression of transforming growth factor β (Tgfβ1), with a significant difference between Wt males born to Het_db dams on a CD versus HFFD (Figure 3D). In female offspring, there was a significant effect of diet on hepatic expression of pro-collagen I (PcI) and pro-collagen III (PcIII), with a significant difference between Wt females born to Het_db dams on a CD versus HFFD (Figure 3I,L). In males, the mRNA expression of PcI was significantly higher in Het_db offspring on HFFD compared with CD. The expression of tumor necrosing factor α (Tnfα) was not affected by diet or intrauterine environment in either sex (Figure 3E,K). mRNA expression of key lipogenic genes revealed decreased expression of stearoyl-CoA-9-desaturase (Scd1) in male offspring of Wt dams on the HFFD, and to a similar degree in offspring of Het_db dams on the CD with no further decrease in this group with HFFD (Figure 3B). There were no changes to Scd1 in females (Figure 3H). There were no notable changes to sterol regulatory binding element-1 (Srebp1c) in any sex or group, except for an overall main effect of the intrauterine environment in females (Figure 3A,G).

We next carried out lipidomics analysis to evaluate the effect of the intrauterine environment and diet on the composition of FFA and triacylglycerols (TAG) in the liver of adult offspring. Individual lipid species present in abundance in the FFA fraction are shown in Figure 4A,B. Except for 18:1ω9, all major lipid species were affected by the diet [16:0 (P<0.0001); 18:0 (P<0.0001); 16:1 (P=0.0002); 18:1ω7 (P<0.0001); 18:2 (P<0.0001); 20:4 (P=0.023)]. In post-hoc analysis, HFFD increased the abundance of 16:0 (P<0.0001), 18:0 (P<0.001), 16:1 (P=0.0007) and 18:1ω7 (P=0.0005) in Wt offspring born to Wt dams. Similarly, in Wt offspring born to Het_db dams, there was an increase in 16:0 (P<0.0001), 18:0 (P<0.0001) and 18:1ω7 (P<0.001). In both groups 18:2 was decreased by HFFD (P<0.0001). There was a main effect of the intrauterine environment on 16:0 (P=0.0027), 18:1ω7 (P=0.046), 18:2 (P=0.0032) and 20:4 (P=0.018). In CD-fed offspring, 16:0 was increased (P=0.0015) and 18:2 decreased (P=0.0008) in Het_db versus Wt offspring. The increase in 18:0 after HFFD was greater in offspring from Het_db versus Wt pregnancies (P=0.0271). There was a significant interaction between diet and the intrauterine environment for 16:0 (P=0.0117) and 18:2 (P=0.004). For minor lipid species present in low abundance (Figure 4C,D), 18:3ω6, 18:3ω3, 20:3ω6 and 20:5 were all below the detection limit in both HFFD groups. Comparing CD-fed offspring from Het_db versus Wt dams, 18:3ω6 (P=0.0089), 18:3ω3 (P=0.0058) and 20:5 (P=0.046) were all lower in Het_db offspring.

Major individual lipid species in the TAG fraction are shown in Figure 4E,F. Similar to the FFA fraction, diet had an effect on all major lipid species except for 18:1ω9 [16:0 (P<0.0001), 18:0 (P>0.0001), 16:1 (P<0.0001), 18:1ω7 (P<0.0001), 18:2 (P<0.0001) and 20:4 (P<0.0001)]. HFFD increased the abundance of 16:0 (P<0.0001), 18:0 (P=0.001), 16:1 (P<0.0001) and 18:1ω7 (P<0.0001) in offspring from Wt dams. In Het_db offspring, HFFD also increased the presence of 16:0 (P<0.0001), 18:0 (P<0.0001), 16:1 (P=0.0018) and 18:1ω7 (P=0.0002). The abundance of 18:2 was reduced by HFFD in Wt and Het_db offspring (both P<0.0001). Similarly, 20:4 was decreased with HFFD in both Wt (P=0.004) and Het_db (P=0.023) offspring. There was an effect of the intrauterine environment on 18:1ω9 (P=0.028), 18:1ω7 (P=0.0265) and 18:2 (P=0.0126). Post-hoc analysis showed that in CD-fed groups, there was an increase in 18:1ω9 (P=0.0393) and decrease in 18:2 (P=0.0056) in Het_db versus Wt offspring. For minor lipid species (Figure 1G,H), >40% of samples were below the detection limit in the HFFD groups for 18:3ω6 and 18:3ω3 and thus comparisons were not made. In CD-fed offspring, there was a decrease in the abundance of 18:3ω3 in Het_db versus Wt (P=0.0433). There was a significant effect of the diet on the abundance of 22:6 (P<0.0001) with a HFFD-associated decrease in Wt (P=0.0073) and Het_db(P=0.0103) offspring. No comparisons were made for 20:3ω6 due to too many samples below the detection limit, and 20:5 was undetected in the TAG fraction.

Lipidomic profiles were used to indirectly measure the activity of lipogenic enzymes, which are key regulators of one of the major determinants of steatosis, DNL. The Δ9 fatty acid desaturase, SCD1, catalyzes the conversion of palmitic acid (16:0) and stearic acid (18:0) to palmitoleic acid (16:1) and oleic acid (18:1η9), respectively, key substrates used in the biosynthesis of complex lipids. The ratio of products-to-precursors was calculated to reflect the activity of SCD1. There was a significant effect of diet on the Δ9 desaturation index, with 16:1/16:0 increased in HFFD versus CD control offspring, and a decrease in the 18:1/18:0 due to HFFD in both groups of offspring (Figure 5A,B). The 16:0/16:1 ratio was not different between control offspring fed a HFFD and offspring born to Het_db pregnancies fed either a CD or HFFD. The 18:1/18:0 ratio was higher in CD-fed offspring from Het_db dams versus Wt dams and there was a significant effect of intrauterine environment on 18:1/18:0. The rate of DNL was inferred from the abundance of 16:0, the main product of DNL, relative to the essential FA, linoleic acid (18:2). There was a significant effect of diet on the DNL index, with increases in HFFD versus CD offspring from both groups (Figure 5C). Elongation of 16:0 to 18:0, catalyzed by Elov6, was significantly affected by diet, and increased in HFFD versus CD in both groups of offspring (Figure 5E). The elongation index was increased in CD-fed offspring from Het_db dams versus Wt dams.

A common feature of NAFLD is a shift in the composition of FA, namely in the proportions of MUFA, SFA and PUFA. Our lipidomic analysis revealed that the total FFA abundance of MUFA, SFA and PUFA in FFA were significantly affected by diet. SFA were increased and PUFA decreased by HFFD in both groups of offspring (Figure 6B,C). Total MUFA were increased in HFFD versus CD offspring from control pregnancy, while there were no differences in total MUFA between HFFD-fed control offspring and HFFD or CD-fed offspring from Het_db pregnancies (Figure 6A). Total PUFA were lower in CD-fed offspring from Het_db dams compared to CD-fed offspring from Wt dams (Figure 6C). These data are in line with the existing literature on FA composition in NAFLD patients [6].

It is well established that maternal health during pregnancy shapes lifelong risk for chronic disease in the offspring. Several human studies have demonstrated that babies born to overweight or obese mothers are more likely to develop NAFLD and other components of MetS later in life [3,7]. Herein, we used female mice heterozygous for leptin receptor deficiency as a model of maternal metabolic dysfunction. Whereas homozygotes (db/db) are severely obese and infertile, heterozygote females retain fertility and have 2.4-fold higher fat mass compared with age-matched wild-type control females [5]. Thus, this model produces a consistent and reproducible maternal metabolic phenotype that can be used to examine effects of the intrauterine environment by studying Wt offspring. Previously, we showed that Het_db pregnancy accelerates adipose accumulation and adipocyte maturation in early life and predisposes to diet-induced adipose tissue dysfunction [5]. The development of adiposopathy or adipose tissue dysfunction is thought to be a central pathogenic event in obesity-associated metabolic disease. Healthy adipose tissue remodeling, driven by hypertrophy of existing adipocytes and generation of new adipocytes from a resident pool of adipocyte progenitors, ensures adequate buffering of excess energy under obesogenic conditions. A failure in adipose expansion to meet increased demands for lipid storage results in adipocytes that are excessively hypertrophic, inflamed, and which exhibit impaired insulin-mediated inhibition of lipolysis, leading to lipid spillover into the circulation [8]. Therefore, adiposopathy is a key mechanism linking obesity to NAFLD. In the present study, we used the Het_db model of high maternal pre-pregnancy adiposity rather than the common diet-induced obesity (DIO) model. While the DIO model has been shown to result in programmed obesity and metabolic dysfunction in studies of developmental programming of obesity, unpublished evidence from our lab and published data from other groups suggests that this model elicits fetal growth restriction [9,10], which is well known to be associated with rapid catch up growth in the post-natal period. Thus, to eliminate this confounding variable, we used the Het_db model, which not only accurately recapitulates accumulation of body fat observed in infants born to human obese mothers, but generates a consistent, reproducible phenotype of moderate metabolic dysfunction in the dam.

Results of the current study show that female, but not male, offspring born to Het_db dams are more susceptible to diet-induced obesity than those offspring born to lean, Wt dams. Fasting levels of circulating FFA were higher in male offspring from Het_db dams when on a low-fat diet, with no further increase after prolonged high-fat/fructose feeding. In contrast, female offspring from Het_db and normal pregnancies were protected from diet-induced increases in fasting FFA despite exaggerated diet-induced fat gain in Het_db female offspring. These sex differences were also observed in our previous study, which showed that female offspring were protected from increases in post-prandial levels of FFA after exposure to high fat diet [5]. These findings led us to anticipate a sex-specific relationship between maternal adiposity and hepatic steatosis in the offspring. Contrary to our expectations, females and males born to overweight dams were similarly susceptible to diet-induced hepatic steatosis, despite maintenance of normal circulating FFA levels in females. Thus, the effect of an abnormal intrauterine metabolic milieu on risk for diet-induced hepatic steatosis in females appears to be uncoupled from circulating FFA levels, which are strongly correlated to lipogenesis and hepatic lipid accumulation [11–13]. The Western-style diet used in this study contains high levels of both fat (45%kcal) and fructose (35% kcal). Fructose has been demonstrated to drive hepatic DNL, independent of circulating FFA levels [14]. Diets high in fructose have been demonstrated to up-regulate key enzymes involved in the process of DNL, driven in part by the rapid conversion of carbohydrates into acetyl-CoA [14]. In humans, both short- [15] and long-term [16] high-fructose diets have been demonstrated to increase intrahepatocellular lipid. Thus, high dietary fructose plays a significant role in driving hepatic DNL. Interestingly, we demonstrated that even with consumption of a high-fructose diet, offspring of overweight pregnancies demonstrate significantly greater hepatic steatosis, suggesting a synergistic effect of the high-fructose diet with the programming effect of maternal overweight.

Given the lack of sex differences in diet-induced hepatic steatosis, we undertook a lipidomics approach to investigate lipid metabolism in younger male offspring born to overweight dams. Profiles of FA species in FFA and TG extracted from the liver revealed diet-induced changes in the lipidome consistent with NAFLD-associated changes reported in the human and animal literature [13]. Diet-induced increases in the rate of DNL, as reflected by an increased ratio of palmitic acid to linoleic acid, were accompanied by increases in the abundance of SFA (palmitic acid and stearic acid) and MUFA (palmitoleic acid and oleic acid). Diet-induced increases in stearic acid were amplified in offspring from Het_db pregnancies, possibly due to increases in the elongation of palmitic acid observed in this group. The desaturation index (16:1/16:0), reflecting the conversion of SFA to MUFA for incorporation into complex lipids, was higher in high-fat-fed offspring from normal pregnancies. In contrast, HFFD decreased the desaturation of 18:0 to 18:1 in both groups of offspring, which was higher in Het_db offspring on a CD. While both ratios are thought to reflect SCD1 activity, the 16:1/16:0 ratio is more closely associated with DNL and NAFLD. A lower abundance of PUFA is the most common finding in NAFLD [13,17] and was also observed in high-fat/fructose-fed offspring. The intrauterine environment had the greatest impact on PUFA levels, which were decreased in Het_db offspring fed a low fat diet or a high fat diet. PUFA levels are protective against hepatic steatosis and lipotoxicity as they have an inhibitory effect on lipogenesis and inflammation [13,17,18].

Collectively, these results support existing human literature demonstrating high pre-pregnancy BMI as a risk factor for development of NAFLD in the offspring. While an obesogenic phenotype was programmed exclusively in female offspring born to overweight dams, diet-induced obesity in females occurred without changes in circulating FFA levels. In contrast, despite higher FFA levels than controls, male offspring of overweight dams did not demonstrate a programmed effect on adiposity. High maternal adiposity was associated with increased hepatic steatosis in both sexes; however, male offspring displayed a stereotypic pathogenesis of NAFLD, arising from increased circulating FFA, whereas female offspring developed an equivalent degree of hepatic steatosis in the absence of changes to circulating FFA.

• The present study was undertaken to investigate the relationship between high maternal pre-pregnancy adiposity and the risk of NAFLD in the offspring.

• We identified that maternal adiposity programs NAFLD in the offspring, likely driven by high de novo lipogenesis, rather than circulating FFA derived from adipose tissue lipolysis.

• Given the high prevalence of overweight/obesity in the reproductive age population, special attention should be paid to the risk of metabolic disease in the offspring.

The data described in the present study are available from the corresponding author (J.A.T.) upon reasonable request.

The authors declare that there are no competing interests associated with the manuscript.

This work was funded by grants held by J.A.T. from the Canadian Institutes for Health Research (CIHR) [grant number 162452]; the Heart and Stroke Foundation of Canada (HSFC) [grant number G-19-0026536] and the Koopmans Memorial Fund. J.A.T. also received funding by the Libin Cardiovascular Institute. The salary of T.B.S. was supported by the Libin Cardiovascular Institute.

Taylor B. Scheidl: Conceptualization, Data curation, Formal analysis, Visualization, Writing—original draft, Writing—review & editing. Jessica L. Wager: Data curation. Larissa G. Baker: Data curation. Amy L. Brightwell: Data curation. Katrina M. Melan: Data curation. Sebastian Larion: Conceptualization, Data curation. Ousseynou Sarr: Conceptualization. T.R. Regnault: Conceptualization. Stefan J. Urbanski: Conceptualization, Data curation. Jennifer A. Thompson: Conceptualization, Data curation, Supervision, Funding acquisition, Visualization, Writing—original draft, Project administration, Writing—review & editing.

The authors would like to gratefully acknowledge the efforts of the staff at the University of Calgary Faculty of Veterinary Medicine Diagnostic Services Unit for preparing samples for histology. We would also like to thank Dr. Jane Shearer for her invaluable input on the direction of this project. Finally, we would like to thank Dr. Stefan Urbanski for providing expertise in histopathological analyses.

DIO

diet-induced obesity

FFA

free fatty acid

HFFD

high-fat/fructose diet

MUFA

monounsaturated fatty acid

NAFLD

non-alcoholic fatty liver disease

PUFA

polyunsaturated fatty acid

SFA

saturated fatty acid