Introduction

The prevalence of obesity (body mass index (BMI) ≥ 30) is increasing at an alarming rate in both Europe and the US1-3. Obesity is strongly heritable in humans and mice, with estimates ranging from 50% to 90%4,5. Consumption of energy-dense diets rich in fat and refined carbohydrates and a sedentary lifestyle, are the key environmental contributors to the current obesity epidemic6. The obesity epidemic is associated with a rise in several cardiometabolic diseases, including type 2 diabetes, coronary heart disease, and non-alcoholic fatty liver disease (NAFLD)7-9. NAFLD affects two billion adults globally, whereof 20% of these cases are expected to develop into non-alcoholic steatohepatitis10,11.

Sex differences in body fat distribution and susceptibility to obesity, NAFLD, and other cardio-metabolic traits, have been described in both mice and humans, and are driven by genes as well as sex hormones8,12,13,14,15. Premenopausal females typically have greater subcutaneous fat depositions, whereas males and postmenopausal females tend to accumulate more visceral fat12. Studies on the menopausal transition in mice and humans, suggest that estrogen may increase energy expenditure and adipose tissue mitochondrial function12,13. The genetic contribution to body fat distribution have been shown in large-scale genome-wide association studies (GWAS) identifying several sex dimorphic loci13,16.

We have previously identified an area on chromosome 4 exhibiting a genome-wide significant association with diet-induced subcutaneous fat accumulation in ~ 100 inbred male mice, but not in female mice or humans13. One of the candidate genes in the identified region, which also has been implicated in regulation of lipid metabolism, was UbiA prenyltransferase domain containing 1 (Ubiad1)13,17. Ubiad1 is a transmembrane enzyme in the Golgi apparatus and the endoplasmic rectum (ER) membrane18. Ubiad1 catalyzes the conversion of phylloquinone-derived menadione (vitamin K3) and forms the vitamin K subtype menaquinone-4 (MK-4)19. Loss of Ubiad1 reduces the cytosolic synthesis of coenzyme Q-10, which leads to increased reactive oxygen species-mediated lipid peroxidation in vascular cells18. Studies of Schnyder corneal dystrophy patients, a disease linked to a mutated UBIAD1gene, showed increased accumulation of cholesterol and phospholipids in the cornea20,21. This condition is characterized by the accumulation of crystals and lipids in the cornea, which can lead to blurred vision and reduced visual acuity. UBIAD1 enzymatic activities regulate ER-associated degradation (ERAD) of ubiquitinated HMG-CoA reductase (HMGCR)22,23. HMGCR is the rate-limiting enzyme of the mevalonate pathway that produces cholesterol and essential non-sterol isoprenoids. Accumulation of HMGCR due to mutated UBIAD1 leads to increased intracellular cholesterol in the cornea22,23. Unfortunately, whole body Ubiad1gene-deficiency in mouse seems to be embryonically lethal due to accelerated ER-associated degradation of HMGCR and dysregulated sterol synthesis24, which complicates studies of the actual protein. We hypothesized that reducing Ubiad1 expression in mice would impact lipids associated with inflammation or cholesterol accumulation in obesity.

It is unclear if Ubiad1 affects diet-induced obesity, and it is not described in detail how Ubiad1 might affect accumulation of specific liver lipids in NAFLD. Thus, to evaluate the role of Ubiad1 in obesity and hepatic lipid metabolism, we associated adipose tissue Ubiad1 expression with adiposity in cohorts of inbred mouse strains and humans, and investigated the role of Ubiad1 in heterozygous Ubiad1 gene deficient mice (Ubiad1+/−) fed a high-fat diet (HFD).

Results

Ubiad1 expression correlates negatively with adiposity in mice and humans

We previously identified a region on chromosome 4 that associated strongly with inguinal subcutaneous fat pad weight in ~ 100 male inbred strains fed a high fat/high sucrose (HF/HS) diet for 8 weeks13. The presence of a local expression quantitative trait locus (cis-eQTL) of a candidate gene, i.e., significant association of local single nucleotide polymorphism (SNP) with expression levels of that gene, would suggest that mRNA of this candidate gene could be a determinant of the trait. Our data indicate that adipose tissue Ubiad1 expression show a strong cis-eQTL in both male (Fig. 1A) and female (Fig. 1B) mice. Furthermore, the subcutaneous adipose tissue Ubiad1 mRNA level was negatively correlated with subcutaneous fat pad weight in male (Fig. 1C) and female (Fig. 1D) mice. We next used the DAVID GO bioinformatics resource25,26 for enrichment analysis on the 500 adipose genes correlating most strongly with Ubiad1 mRNA expression levels among the mice strains. The most significantly enriched pathways in male and female mice were protein transport (Fig. 1E) and mitochondrial respiration (Fig. 1F), respectively. Furthermore, Ubiad1 mRNA expression showed a positive association with expression of mitochondria-related genes in both sexes (Fig. 1E, F).

Fig. 1
figure 1

Ubiad1 expression associates with subcutaneous adipose tissue weight. (A, B) Locus plot showing the associations between single nucleotide polymorphisms (SNPs) and subcutaneous adipose tissue Ubiad1 expression on chromosome 4 in (A) male and (B) female HMDP mice. The significant associations between SNPs and Ubiad1 mRNA expression levels in subcutaneous white adipose tissue (WAT) are depicted as dots and refer to the p-value of the association (-log10 of p-value), and location on chromosome 4. (C, D) Ubiad1 mRNA expression is significantly correlated with subcutaneous WAT mass in (C) male and (D) female mice. (E,F) Pathway enrichment analysis using the top 500 genes that were strongest correlated with Ubiad1 mRNA expression levels in the HMDP of (E) male, and (F) female mice. GO Biological process (GO BP) pathways were performed with positively correlated mRNAs (red) and negatively correlated mRNAs (blue) individually. The top 10 enriched pathways are presented. X-axis: P-value (-log10) of enriched terms. (G, H) Fold changes of UBIAD1 mRNA expression in the (G) adipocyte fraction and (H) stromal vascular fraction (SVF) of subcutaneous WAT of lean or obese human subjects. UBIAD1 mRNA was measured by Illumina microarrays. Data were normalized against the average UBIAD1 expression of lean individuals. Data in graphs are shown as means ± SD (n = 12 per group). P-values were calculated with unpaired t test. **p < 0.01. (I) Pathway enrichment analysis of the mRNAs that correlated strongest (366 positively and 264 negatively correlated; correlation coefficient >0.6) with UBIAD1 mRNA expression levels in adipocyte from subcutaneous WAT. PANTHER GO-Slim Biological Process pathway analyses39 were performed individually with positively correlated mRNAs (red) and negatively correlated mRNAs (blue). The top 10 enriched pathways are presented. X-axis: p-value (-log10) of enriched terms.

Full size image

We followed up these findings in mice by investigating a bariatric surgery cohort to evaluate if the correlation between UBIAD1 mRNA and obesity holds up in humans. UBIAD1 mRNA in the adipocyte fraction of human subcutaneous adipose tissue was significantly reduced in obese individuals (p = 0.0082, Fig. 1G), supporting that UBIAD1 expression levels are linked to body weight. We found no difference of the UBIAD1 mRNA expression in the stromal vascular fraction of the subcutaneous adipose tissue between lean and obese individuals (p = 0.93, Fig. 1H), suggesting that fat content is correlated to UBIAD1 mRNA expression in adipocytes but not adipose tissue stromal vascular cells. We next performed enrichment analysis with genes correlating the most strongly with UBIAD1 mRNA levels in the adipocyte fraction of subcutaneous white adipose tissue (WAT). Using a correlation coefficient cut-off of more than 0.6, we identified 366 and 264 genes that correlated positively and negatively with UBIAD1 expression, respectively. In accordance with the mice data (Fig. 1E, F), we observed that expression of lipid- and mitochondria-related genes correlated positively with UBIAD1 expression levels in human adipocytes (Fig. 1I).

Deletion of Ubiad1 is embryonically lethal in C57BL/6 N mice

To investigate directly the role of Ubiad1 in obesity, we generated whole body knockout for Ubiad1 in C57BL/6 N mice using CRISPR/Cas9 (Fig. 2A). Pairs of sgRNAs binding to the Ubiad1 exon 1 were co-injected together with Cas9 RNA in embryos prior to implantation into foster mothers (Materials and Method). Among the generated founders, we identified an individual with a 29 bp deletion generating a frameshift resulting in a disrupted Ubiad1 gene (Fig. 2A). Mating of heterozygous mice for the Ubiad1 deletion (Ubiad1+/−) produced only wild type (Ubiad1+/+) and Ubiad1+/− offspring (Table 1), as observed by PCR analysis (Fig. 2B). In agreement with a previous report24, we did not identify any Ubiad1−/− offspring, observed quite small litter sizes, and concluded that homozygous deletion of the Ubaid gene is embryonically lethal in the C57BL/6 N strain.

Fig. 2
figure 2

Generation of Ubiad1+/- mice by the CRISPR Cas9 system. (A) A schematic drawing of the Ubiad1 locus, binding sites for the two sgRNAs, and the generated disrupted Ubiad1 allele. The 29 base pair deletion generating a frameshift mutation in the Ubiad1 gene is marked on the WT allele. Amino acids marked with red color show the expected codon alterations generated due to the introduced frameshift mutation. (B) Migration of DNA fragments obtained with PCR amplification of a normal and mutated Ubiad1 alleles. The predicted sizes of the DNA fragments are 265 bp for the Ubiad1+/+ allele and 236 bp for the Ubiad1+/- allele. The gel image was cropped and original gels are presented in Supplementary Figure 5A. (C, D) Quantification of wild type and mutated Ubiad1 mRNA transcripts, versus only wild type mRNA in (C) subcutaneous WAT and (D) liver in 20 weeks old mice fed HFD for 10 weeks. (n = 11−13 per group in males, n = 9 per group in females). Data were normalized against the expression of Tbp and Rplp0, respectively. (E) Western blot of Ubiad1 and beta-actin on whole liver proteins from Ubiad1+/+ and Ubiad1+/- mice (n=6 per group). One sample from Ubiad1+/+ male was excluded from the quantification analysis due to saturated band. The blots were cropped and original blots are presented in Supplementary Figure 5B. Data represent means ± SD. P-values were calculated using unpaired t-test. *p < 0.05, **p < 0.01, ***p< 0.001.

Full size image
Table 1 The number of offspring from the breeding of Ubiad1+/− X Ubiad1+/− or Ubiad1+/+ X Ubiad1+/−.
Full size table

To be able to determine expression levels of mRNAs produced by the Ubiad1 gene, we developed a two-pronged RT-qPCR assay. One component is designed specifically to detect only the wild-type Ubiad1 mRNA (wt-transcript), whereas the other component is unable to distinguish between wild-type and the 29-bp mutated Ubiad1 mRNAs (wt/null-transcripts). Whereas expression of wild type and mutated Ubiad1 mRNAs was similar in Ubiad1+/+ and Ubiad1+/− mice, expression of wild type Ubiad1 mRNA was reduced ~ 50% in both subcutaneous WAT and liver of Ubiad1+/− mice (Fig. 2C, D). This suggests that the mutated Ubiad1+/− transcript is as stable as the wild type transcript, and that there is no compensating transcriptional regulation of Ubiad1 mRNA expression in Ubiad1+/− mice. We next compared hepatic Ubiad1 protein levels between Ubiad1+/+ and Ubiad1+/− mice using immunoblotting (Fig. 2E). We detected two bands in the predicted size range for the Ubiad1 protein (50 − 40 kDa). Whereas the strongest band at 50 kDa showed lower expression in Ubiad1+/− mice of both sexes, the 40 kDa showed similar expression between the genotypes in females.

Comparable body weight gains in Ubiad1 +/+ and Ubiad1 +/− mice with HFD

We exposed Ubiad1+/+ and Ubiad1+/− mice to a HFD and compared their weight gain after 10 weeks of diet intervention. All groups exhibited an increase in body weight and fat mass from the baseline due to the HFD (Fig. 3A-B). The relative weight gain in male Ubiad1+/+ and Ubiad1+/− mice was 44.11 ± 19.09% and 46.08 ± 17.70%, respectively. Among female mice, the relative weight gain was 55.58 ± 21.40% in Ubiad1+/+ and 49.51 ± 20.58% in Ubiad1+/−. However, we observed no significant differences between the groups in terms of body weights (Fig. 3A), body compositions (Fig. 3B), liver weights (Fig. 3C), inguinal (Fig. 3D) and gonadal fat pad weights (Fig. 3E), or non-fasting plasma cholesterol levels (Fig. 3F) in either sex.

Fig. 3
figure 3

No effect of lowered Ubiad1expression on obesity-related traits in mice on HFD. (A-B) (A) Total body weights and (B) body composition prior to (0 wks) and post (10 wks) HFD intervention in male and female Ubiad1+/+ and Ubiad1+/- mice (n= 11−13 males, n = 9 females). (C-E) The weight of (C) liver, (D) inguinal WAT, and (F) plasma cholesterol levels at the end of the experiment. The mice were euthanized at 20 weeks of age in the morning (8−11 a.m.). Data represent means ± SD. P-values were calculated using Two-way ANOVA, with Fisher LSD test used for post-hoc multiple comparisons (A, B) or unpaired t test (C-F). ***p < 0.001.

Full size image

In accordance with unaltered adiposity, the expression of inguinal subcutaneous white adipose tissue mRNAs important for lipid metabolism was similar in Ubiad1+/+ and Ubiad1+/− mice (Supplementary Fig. 1).

Increased expression of liver lipid-related genes in female Ubiad1 +/− mice

Ubiad1 has been linked to Schnyder corneal dystrophy27, a condition characterized by the progressive accumulation of cholesterol and lipids in the cornea. Disruption of the UBIAD1 enzyme’s function leads to this abnormal accumulation, contributing to the disease’s symptoms. Thus, we investigated the effect of reduced Ubiad1 mRNA on the expression of genes relevant for hepatic lipid metabolism. Using qPCR, we found that mRNA expression levels of Hmgcr (Fig. 4A), Ldlr (Fig. 4B), Pcsk9 (Fig. 4C), Srebf2 (Fig. 4D), and Pparg (Fig. 4E) were elevated in livers of female Ubiad1+/− mice as compared to Ubiad1+/+ mice. Although the mRNA expression of most of the genes showed similar trends in male and female mice, none of the mRNAs were significantly altered in male Ubiad1+/− mice (Fig. 4A-E). The expression of Fasn showed a tendency to be increased in Ubiad1+/− mice of both sexes (Fig. 4F).

Fig. 4
figure 4

The effect of Ubiad1 mRNA expression on hepatic lipid metabolism. (A-F) Quantification of (A) Hmgcr, (B) Ldlr, (C) Pcsk9, (D) Srebf2, (E) Pparg, and (F) Fasn mRNAs in Ubiad1+/+ and Ubiad1+/- livers of mice after 10 weeks of HFD. Data were normalized against Rplp0 expression levels. Data represent means ± SD (n = 11−13 males, n = 9 females). P-values were calculated with unpaired t test. *p < 0.05, **p < 0.01, ***p < 0.001.

Full size image

Alterations in liver lipid metabolic pathways in Ubiad1 +/− mice

To identify which mRNAs and pathways are influenced by reduced Ubiad1 expression, we conducted RNA sequencing to analyze global gene expression in the liver tissues of both male and female Ubiad1+/+ and Ubiad1+/− mice. In livers of Ubiad1+/− males, 87 genes were increased, and 52 genes were decreased as compared to Ubiad1+/+ (Fig. 5A and supplementary Table 1). In livers from female Ubiad1+/− mice, 68 mRNAs were significantly increased, whereas 16 mRNAs were reduced as compared to Ubiad1+/+ mice (Fig. 5B and supplementary Table 2). Pathway analysis of the differently expressed genes between Ubiad1+/− and Ubiad1+/+ showed significant effect on several pathways in both sexes (Fig. 5C, D). In particular, the pathways related to cholesterol biosynthesis, isoprenoid metabolism, farnesyl diphosphate metabolism, secondary alcohol biosynthesis, and sterol biosynthesis, were significantly enriched among the hepatic mRNAs that were increased in Ubiad1+/− as compared to Ubiad1+/+ in male mice (Fig. 5C). In female Ubiad1+/− mice, lipid and triglyceride biosynthetic/metabolic processes were the most enriched up-regulated pathways (Fig. 5D). Overall, several hepatic lipid-related genes exhibited higher expression levels in Ubiad1+/− mice, leading to the upregulation of lipid metabolic pathways in both sexes.

Fig. 5
figure 5

Lower Ubiad1 mRNA expression affects lipid biosynthetic pathways in male and female mice. (A,B) Heatmap of differentially expressed hepatic genes in Ubiad1+/+ compared to Ubiad1+/- (A) male and (B) female mice (n = 8 per group) as determined by RNA sequencing (FDR < 0.05, FC > 1.5). (C, D) The 10 most significantly different hepatic mRNA expression pathways between Ubiad1+/+ and Ubiad1+/- (C) male and (D) female mice, identified by enrichment analysis. The GO biological process pathway analysis was performed with differentially expressed genes. All the top 10 enriched pathways were increased in Ubiad1+/- mice.

Full size image

Specific hepatic lipids are affected differently in male and female Ubiad1 +/− mice

To determine if the observed differences in mRNA expression linked to lipid metabolism in Ubiad1+/− mice affected accumulation of lipids, we performed detailed lipid analysis with HPLC-qTOF/MS on the livers of Ubiad1+/+ and Ubiad1+/− mice of both sexes (Fig. 6A-P and Supplementary Fig. 2).

Although female Ubiad1+/− mice exhibited enrichment of several lipid biosynthetic pathways (Fig. 5D), only liver ceramide levels were increased in Ubiad1+/− mice as compared to Ubiad1+/+ mice (Fig. 6A). In livers from male Ubiad1+/− mice, PG (Fig. 6K), LPE (Fig. 6P), and PI (Fig. 6N) levels were decreased in Ubiad1+/− mice as compared to Ubiad1+/+ mice. Conversely, levels of DAG (Fig. 6B) and FFA (Fig. 6L) tended to be higher in male Ubiad1+/− mice, although this increase was not statistically significant (p = 0.08 and 0.06, respectively). Liver CE (Fig. 6O) levels were not significantly affected in Ubiad1+/− mice of either sex. Taken together, reduced Ubiad1 expression led to increased hepatic ceramide levels in females, whereas it decreased several phospholipids in males.

Fig. 6
figure 6

Sex differences in hepatic lipid species with altered Ubiad1 mRNA expression (A-Q). Concentration of (A) Cer, (B) DAG, (C) LPC, (D) PC, (E) SPM, (F) PC(O), (G) PE, (H) TAG, (I) PC(p), (J) CL, (K) PG, (L) FFA, (M) HexCer, (N) PI, (O) CE, and (P) LPE in the liver of male and female Ubiad1+/+ and Ubiad1+/− mice fed HFD for 10 weeks. Data represent means ± SD (n = 8 males; n = 6−7 females). p-values were calculated with unpaired t test. *p < 0.05, **p < 0.01.

Full size image

The fact that cholesterol storage was unaltered despite that several genes related to cholesterol biosynthesis were increased in the livers of Ubiad1+/− mice (Fig. 5C), suggests that compensating mechanisms contracts the altered gene expression in this lipid biosynthetic pathway of Ubiad1+/− mice. This is corroborated by Western blot analysis of membrane fractions from female livers, which show no significant differences between Ubiad1+/− and Ubiad1+/+ mice in the cholesterol metabolism-related proteins Hmgcr, Ldlr, and Srebp2 (supplementary Fig. 4). Although the Srebp2 protein levels showed a slight tendency to increase in Ubiad1+/− mice, this trend corresponds with the levels of Srebp2 mRNA. Alternatively, extending the duration of the HFD intervention might be necessary to observe the impact of Ubiad1 on cholesterol accumulation and NAFLD in Ubiad1+/− mice. Combining extending the duration of the HFD intervention with histology, could reveal potential liver-related diseases caused by perturbed lipid accumulation.

Discussion

We have previously shown that Ubiad1 reside in an area of chromosome 4 that has a genome-wide significant association with subcutaneous fat pad weight in male mice13. Here, we show that adipose tissue Ubiad1 expression has a strong cis-eQTL in both male and female mice, and that its Ubiad1 mRNA expression levels associate negatively with adiposity in both mice and humans. Studies on Ubiad1+/− mice fed a HFD revealed no effect on adiposity or subcutaneous fat pad weight but showed effects on mRNA transcripts of several liver lipid genes and sex-specific effects on accumulation of liver lipids. Each of these points are discussed below. Future research should explore the impact of Ubiad1 on traits related to insulin sensitivity.

Role of Ubiad1 in obesity

The local genetic regulation of Ubiad1 (cis-eQTL) and the associations between adipose tissue Ubiad1 expression and adiposity in both mice and humans, support a role of Ubiad1 in development of obesity. This was further supported by a positive association between adipose tissue Ubiad1 expression and expression of genes encoding mitochondrial synthesis and respiration (Fig. 1). On the other hand, the fact that we found no difference in adiposity or subcutaneous fat pad weight between Ubiad1+/+ and Ubiad1+/− mice after HFD feeding for 10 weeks, suggests that a 50% reduction in Ubiad1 mRNA expression is insufficient to cause an effect on obesity development. This is in line with a previous study reporting no difference in the total body weight of Ubiad1+/+ and Ubiad1+/−mice on a regular chow diet24. It is also possible that alternative genes or pathways may compensate for the reduced expression of Ubiad1 in Ubiad1+/− mice. For example, enzymes involved in cholesterol synthesis or lipid processing, such as HMG-CoA reductase or other prenyltransferases, might increase their activity to counterbalance diminished Ubiad1 function. This compensatory upregulation could obscure the complete effects of Ubiad1. It may also be necessary to extend the duration of dietary interventions to elucidate fully the role of Ubiad1 in lipid metabolism. Prolonged dietary regimens could increase metabolic stress, potentially unveiling more pronounced phenotypic differences and allowing for the observation of the long-term effects of Ubiad1 deficiency. Such studies would help determine whether the subtle changes currently observed are early indicators of more significant metabolic disturbances that take time to develop fully. Finally, other genes in the chromosome 4 locus might explain the subcutaneous fat pad association independent of Ubiad1. For example, adipocyte-specific deletion of the gene located next to Ubiad1, mechanistic target of rapamycin (Mtor, encoding the mTOR protein), has been suggested to inhibit adipose tissue development28.

Sex-specific effect of Ubiad1 in hepatic lipid accumulation

The Ubiad1 enzyme has several different functions including inhibition of ER-associated degradation of ubiquitinated HMGCR22, and stimulation of menadione19and CoQ-10 synthesis18. Thus, we analyzed the effect of globally reduced Ubiad1 expression (Ubiad1+/−) on hepatic mRNA expression, in particular lipid synthesis, using RNA sequencing in mice of both sexes. The number of altered mRNAs in Ubiad1+/− mice were higher in males than females (Fig. 5). However, in females, a larger proportion of the genes affected in Ubiad1+/− mice encoded proteins belonging to lipid metabolic pathways. Whereas cholesterol biosynthesis was the main pathway regulated in male Ubiad1+/− mice, lipid biosynthesis was the most regulated pathway in female Ubiad1+/− mice. Intriguingly, our detailed hepatic lipid quantification with HPLC-qTOF/MS showed that the lipid species altered in Ubiad1+/− mice also were different between the sexes. Whereas levels of LPE, PG, and PI were reduced in Ubiad1+/− male livers, ceramide levels were increased in Ubiad1+/−females. Interestingly, a mouse line carrying Ubiad1 N100S point mutation with hyper-reflective dot-like deposits in the anterior cornea, show an increase in corneal PG without any changes in corneal cholesterol content21. Thus, Ubiad1 might have a stronger effect on accumulation of certain phospholipid species than cholesterol in livers of male mice. In theory, a reduction in Ubiad1 protein levels should lead to an accumulation of Hmgcr protein and cholesterol in the livers of Ubiad1+/− mice via ERAD-associated pathways. This would typically suggest a downregulation of Hmgcr transcript through feedback mechanisms. However, our findings in female Ubiad1+/− mice revealed decreased Ubiad1 protein levels, and unexpectedly showed increased Hmgcr mRNA levels (Figs. 4A and 5B) and no significant change in Hmgcr protein levels (supplementary Fig. 4) or liver CE (Fig. 6O). A previous study with a chow diet found slightly reduced Hmgcr protein levels in Ubiad1+/−mice24, aligning with our conclusion that HMGCR protein levels do not increase in these mice. The observed increase in Hmgcr transcription suggests a complex regulatory role of UBIAD1 in the liver, suggesting activation of positive feedback mechanisms.

We have previously shown a clear sex difference in the accumulation of hepatic ceramides and how they are regulated between the sexes29. The observed sex difference in hepatic ceramide levels was probably due in part to testosterone-mediated repression of the ceramide synthase 6 (Cers6) gene in male mice29. Here, we show that the mRNA levels of Ubiad1 might affect hepatic ceramide levels in female mice but not in male mice. This may indicate a general trend of increased inflammation in the livers of female Ubiad1+/− mice compared to Ubiad1+/+ mice (supplementary Table 2). Additionally, we observed higher mRNA expression levels of Cers5 and interleukin-1b (Il1b) in the livers of Ubiad1+/− females using qPCR (supplementary Fig. 3), with RNA sequencing data showing increased expression for Il1b only (supplementary Table 2). Increased ceramide levels in female Ubiad1+/−mice could potentially lead to hepatic insulin resistance with prolonged HFD intervention30.

Taken together, we have demonstrated significant effects on lipid-related genes in both male and female Ubiad1+/− mice for the first time. Notably, female Ubiad1+/− mice exhibited increased hepatic ceramide levels, whereas PG, PI, and LPE levels were reduced in male Ubiad1+/− mice. Extending the duration of a HFD feeding could uncover the potential clinical implications for fatty liver-related diseases. Future research should also explore the molecular mechanisms underlying the sex-specific role of Ubiad1 in hepatic lipid metabolism.

Conclusion

In conclusion, we have shown that Ubiad1 expression in adipose tissue correlates with adiposity in both humans and mice, although there seems to be no difference in HFD-induced obesity between Ubiad1+/+ and Ubiad1+/− mice. In livers of Ubiad1+/− mice, we observed increased expression of many cholesterol biosynthetic genes in male mice, and lipid biosynthetic genes in female mice. Detailed lipid quantification with HPLC-qTOF/MS analyses showed that hepatic ceramides were induced in female Ubiad1+/− mice, whereas several phospholipids were decreased in male Ubiad1+/− livers. These results suggest a sex-specific effect of Ubaid1 in hepatic lipid metabolism.

Materials and methods

Human cohort

The human “ADIPO cohort”31, comprising 12 lean individuals (mean ± SD BMI 22.8 ± 2.2 kg/m2; weight 68.6 ± 10.9 kg; height 1.73 ± 0.10 m; age 43.5 ± 12.1 years) undergoing elective surgery and 12 individuals with obesity (mean ± SD BMI 43.7 ± 5.4 kg/m2; weight 128.4 ± 16.1 kg; height 1.72 ± 0.11 m; age 43.1 ± 10.4 years) undergoing bariatric surgery, was approved by the Regional Ethics Committee in Western Norway (REC West, approval number 2018/2404). Each subject gave written informed consent before the start of the study, and the methods were carried out in accordance with relevant guidelines and regulations.

Adipocytes and stromal vascular fractions were obtained by digesting 700–800 mg of subcutaneous adipose tissue with collagenase and thermolysin (Liberase Blendzyme 3, Roche, Basel, Switzerland). The tissues were cut into smaller pieces and digested in 4 mL of preheated Hanks’ balanced salt solution with 2.6 Wunch units of Liberase Blendzyme 3 at 37 °C. After 30 min the samples were filtered through 25-mm and 200-mm mesh polypropylene filters (Spectrum Laboratories, Rancho Dominguez, CA, US). The filters were placed in an Easy Pressure syringe filter holder (PALL 4320, Life Sciences, Batavia, IL, US) and connected to a 10 mL Omnifix syringe (Braun Melsungen, Melsungen, Germany) to allow application of weak pressure. Following two rinses of the dissolved samples in Hanks’ balanced salt solution/5% bovine serum albumin (A1311, US Biological, Swampscott, MA, US) and a few minutes time for floatation, the adipocytes were transferred to a 2 mL Eppendorf tube and rinsed once more. The non-floating cells were centrifuged at 400 g for 5 min and lysed in 1 mL Qiazol (Qiagen, Venlo, Netherlands), before freezing in liquid nitrogen and stored at − 80 °C. For gene expression analysis, RNA with high integrity was extracted from the adipocytes and stromal vascular fraction and converted to biotin-labeled cRNA, with subsequent hybridization of the cRNA to HumanHT-12v.3 Illumina Sentrix BeadChip microarrays before analysis in the Illumina iScan system (Illumina Inc., San Diego, CA, US).

Hybrid mouse diversity panel (HMDP)

The study was commenced in 2010. The use of animals and all experimental procedures were approved by the Institutional Care and Use Committee of the University of California, Los Angeles (UCLA) and conducted in accordance with the animal care guideline set by UCLA and the ARRIVE guidelines (https://arriveguidelines.org). The about 100 mouse strains from the HMDP panel included in the study, have been described in detail5,13. Briefly, male and female mice were maintained on a chow diet (Ralston Purina Co., MO, USA) until 8 weeks of age before switching to a HF/HS diet (#D12266B, Research Diet, NJ, USA) for 8 weeks. The macronutrient composition of the HF/HS diet was as follows: 16.8 E% from protein, 51.4 E% from carbohydrates, and 31.8 E% from fat. Mice were euthanized by cervical dislocation, and collected blood and tissues were snap frozen in liquid nitrogen. Subcutaneous (inguinal), mesenteric, gonadal, and retroperitoneal adipose tissues were weighed and flash-frozen in liquid nitrogen immediately after harvesting and stored at −80 °C. All mice were housed with a stable light/dark cycle (07 AM–07 PM), with 55 ± 5% relative humidity at 22 ± 2 °C in a specific pathogen-free animal unit.

Generation of Ubiad1 deficient mice

All animal experiments in Ubiad1 deficient mice were performed according to experimental procedures approved by the Norwegian Animal Research Authority (Mattilsynet, approvals FOTS id: #10902 and # 20822). Experiments conformed to the ARRIVE guidelines, Norwegian ethical guidelines (FOR-2015-06-18–761), and guidelines in the Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. Mice were housed with 12 h light/dark cycle (07 AM to 07 PM), 55% relative humidity at 22 °C in a specific pathogen-free animal unit. Mice had free access to rodent chow (62 E% carbohydrate, 11 E% fat, 27 E% protein; #RM3A/SDS RM3, Scanbur, Denmark), and water. The presence of pathogens was monitored quarterly according to Federation of European Laboratory Animal Science Associations (FELASA) recommendations. Animals used in these experiments were specifically pathogen-free (SPF status) according to FELASA recommendations.

Ubiad1-deficient mice were generated at the Norwegian Transgenic center using CRISPR-Cas9 genome engineering32. The CRISPR Design tool was used to identify optimal guide RNA (gRNA) binding sites in the mouse Ubiad1 gene (Ubiad1 gRNA1: 5’-GGGACCCGCAGAAGAACGAA-3’, Ubiad1 gRNA2: 5’-GGCTTCCCGAACGATCCTGG-3’, Ubiad1 gRNA3: 5’-CCTCAGGGCCAACACGTAGG-3’). Oligoes corresponding to the above gRNAs (gRNA1-fw: 5’-CACCGGGACCCGCAGAAGAACGAA-3’, gRNA1-rev: 5’-AAACTTCGTTCTTCTGCGGGTCCC-3’, gRNA2-fw: 5’-CACCGGCTTCCCGAACGATCCTGG-3’, gRNA2-rev: 5’-AAACCCAGGATCGTTCGGGAAGCC-3’, gRNA3-fw: 5’-CACCGGCCTCAGGGCCAACACGTAGG-3’, gRNA3-rev: 5’-AAACCCTACGTGTTGGCCCTGAGGCC-3’) were phosphorylated by T4 polynucleotide kinase (New England Biolabs, MA, USA), annealed and cloned into a BbsIdigested pX330 vector32. Cloned vectors were used as templates to PCR-amplify sRNAs with binding sites for T7 RNA polymerase (gRNA1-T4_fw: 5’-TTAATACGACTCACTATAGGGACCCGCAGAAGAACGAA-3’, gRNA2-T4_fw: 5’-TTAATACGACTCACTATAGGCTTCCCGAACGATCCTGG-3’, gRNA3-T4_fw: 5’-TTAATACGACTCACTATAGGCCTCAGGGCCAACACGTAGG-3’, gRNA-T4_rev: 5’-AAAAGCACCGACTCGGTGCC-3’). The obtained PCR products were purified with agarose gel electrophoresis and used as templates for in vitro transcription of the gRNAs using the MEGAshortscript T7 Transcription Kit (#AM1354, Thermo Fisher, MA, USA).

Four weeks old female C57BL/6 N mice (Janvier Labs, Le-Genest-Saint-Isle, France) were super-ovulated (injection of 5 U PMSG followed by injection of 5 U hCG 45 h later), mated over night with C57BL/6 N male, and euthanized in the morning the following day to harvest fertilized 1-cell embryos. Cas9 mRNA (Sigma-Aldrich, St. Louis, MO, USA) and pairs of gRNAs were co-injected into embryos, followed by implantation of injected embryos into two pseudo-pregnant RjOrl: SWISS foster mothers (Janvier Labs, France). DNA extracted from ear biopsies of offspring was used to PCR amplify the targeted Ubiad1 region (Ubiad1_fw; 5’-GCTTCCATGGCTGCGGTACA-3’, Ubiad1_rev; 5’-CTGCACAACCCAACAACAGCCTG-3’). F0 founders carrying a mutated Ubiad1 allele were identified by confirmation of heterogeneous DNA hybrids among amplified PCR products using heteroduplex DNA polyacrylamide gel electrophoresis. PCR fragments from promising founders were re-amplified with the following primers (5’-GGGGACAAGTTTGTACAAAAAAGCAGGCTGCTTCCATGGCTGCGGTACA-3’ and 5’-GGGGACCACTTTGTACAAGAAAGCTGGGTCTGCACAACCCAACAACAGCCTG-3’) and cloned into the Gateway pDONR 221 vector (Thermo Fisher, MA, USA). 10 individual clones from each founder were sequenced to characterize Ubiad1 allelic variants in the founder. One F0 founder was identified with a 29 bp deletion as a result of double cleavage by gRNA2 and gRNA3 in the exon 1 of the Ubiad1 gene. This 29 bp deletion was predicted to result in the production of a truncated, nonfunctional Ubiad1 protein. The F0 founder was mated to produce F1 offspring carrying the Ubiad1-deficient allele, and F1 offspring subsequently back-crossed with C57BL/6 N mice for an additional generation to minimize the presence of possible off-target cleavages. The Ubiad1 genotype was determined by PCR from ear tissue using the Ubiad1_fw and Ubiad1_rev primers (Ubiad1 WT allele: 265 bp, Ubiad1 KO allele: 236 bp).

Experimental animals were first generated by mating of heterozygous littermates (Ubiad1+/). To determine if Ubiad1−/− mice were viable, three breeding cages with one male and two female Ubiad1+/− mice were arranged. Genotyping of obtained litters revealed no live Ubiad1−/− mice and reduced number of litters compared to the expected litter size (Table 2). After confirming the expected embryonal lethality24, breeding cages were subsequently set up with Ubiad1+/+ and Ubiad1+/− mice to generate animals for phenotypic analyses. Larger litter sizes were obtained from this heterozygous breeding (Table 2). A total of 23 male and 18 female Ubiad1 littermates were produced to characterize the Ubiad1-deficient model in these studies.

Diet intervention and body composition

At 10 weeks of age, mice were fed a high-fat diet (HFD) with 60 E% from fat, 20 E% from carbohydrates, and 20 E% from protein (#D12492, Research Diet) for 10 weeks. Body weight and body composition were monitored at the start and the end of diet intervention using a Minispec LF90II TD-NMR machine (Bruker, Billerica, MA, USA). Mice were euthanized by cervical dislocation, and collected blood and tissues (liver, and perigonadal and inguinal subcutaneous white adipose tissue) were snap frozen in liquid nitrogen. A total of 41 mice were included in this study. No mice were excluded from the analyses, unless stated otherwise.

Reverse transcriptase quantitative PCR (RT-qPCR)

Liver and inguinal subcutaneous white adipose tissue (WAT) were homogenized in Qiazol reagent (Qiagen) by vigorous shaking with glass beads (Precellys 24, Bertin Technologies, Montigny-le-Bretonneux, France). Total RNA was subsequently isolated using a NucleoSpin RNA isolation kit (Macherey–Nagel, PA, USA), with a modified pre-preparation method33. RNA purity and concentration were measured on a NanoDrop ND-100 Spectrophotometer (Thermo Fisher Scientific, MA USA).

First strand cDNA was synthesized with random hexamers using a High-Capacity cDNA Reverse Transcription Kit (#4368814, Thermo Fisher Scientific) on an ep Gradient S Eppendorf Mastercycler (Eppendorf AG, Hamburg, Germany) with the following settings: 25 °C for 10 min, 37 °C for 120 min, 85 °C for 5 min and 4 °C on hold.

Real-time quantitative PCRs were performed with assay primers (200 nmol/L each) and Bio-Rad SsoAdvanced™ Universal SYBR® Green Supermix (10 µL reaction, 95 °C for 3 min, followed by 44 cycles of 95 °C for 10 s and 60 °C for 20 s) on a CFX96 Touch instrument (Bio-Rad, Hercules, CA, USA). Primers are listed in Table 2. Data were analyzed according to the ΔΔCt method, with normalization against mRNA expression of Rplp0 (Liver) or Tbp (Subcutaneous WAT).

Table 2 List of primer sequence for RT-qPCR.
Full size table

Immunoblotting

Liver was homogenized with lysis buffer (1% SDS, 10 mM Tris-HCl, 10 mM EDTA) containing complete Proteinase Inhibitor Cocktail (#11836170001, Roche) and phosphatase inhibitors (#P0044, Sigma-Aldrich) using glass beads (Precellys 24, Bertin Technologies). Samples were mixed with Leammeli buffer, heated at 95 ºC for 5 min, and loaded into Criterion™ TGX™ 4–20% gels (Bio-Rad). Gels were transferred to Nitrocellulose using Trans-Blot Turbo RTA Transfer kit (Bio-Rad). Primary antibodies against rabbit anti-Ubiad1 (#ab191691, abcam, Waltham, MA, USA) and mouse anti-beta-actin (#A3854, Sigma-Aldrich).

Band intensities were quantified using Image Studio Software (LI-COR, Lincoln, NE, USA). Immunoblot signals were normalized to beta-actin levels.

RNA sequencing of liver

RNA sequencing of livers was performed as previously described34. Briefly, RNA integrity was determined using Bioanalyzer RNA 6000 Nano (Agilent Technologies, CA, USA), according to the manufacturer’s instruction. The RNA samples had RIN values between 8.1 and 8.7. Samples were sequenced at the Norwegian Sequencing Centre (Oslo University Hospital, Ullevål, Oslo). Total RNA samples were subjected to Strand-Specific TruSeqTM mRNA-seq library preparation, and 50 bp paired-end reads were sequenced on an Illumina Novaseq 6000 instrument). To remove/trim low-quality reads and adapter sequences we used BBDuk (BBMap v38.86). Reads were mapped to the Mus musculus reference genome (ENSEMBL release 108) using HiSat2 v.2.1.135. Read counting was done with FeatureCounts v2.0.136. The average read count was 40 million/sample.

Measurement of plasma cholesterol

Plasma cholesterol was measured using colorimetric enzymatic detection kits for detection of total cholesterol (#C75101L, Pointe Scientific, MI, USA) according to manufacturer instructions.

Lipid quantification with HPLC-qTOF/MS

Liver samples (about 100 mg) from 7 to 8 mice in each group were used for lipid analyses using high-performance liquid chromatography (HPLC) coupled to time of flight-mass spectrometry (TOF-MS)33. This platform was used to determine separate lipid species such as ceramides (Cer), diacylglycerol (DAG), lysophosphatidylcholine (LPC), phosphatidylcholine (PC), sphingomyelin (SPM), ether-linked phosphatidylcholine (PCO), phosphatidylethanolamine (PE), triacylglycerol (TAG), phosphatidylcholine plasmalogen (PCP), cardiolipins (CL), phosphatidylglycerol (PG), free fatty acids (FFA), hexosylceramides (HexCer), cholesterol (Chol), Cholesteryl esters (CE), lysophosphatidylethanolamine (LPE), and phosphatidylinositol (PI). Detected lipid signals were normalized against liver weights and internal standards. Internal standards used were Cer-35:1, DAG-30:0, LPC-17:1, PC-28:0, SPM-35:1, PE-28:0, TAG 39:0, CL-56:0, PG-30:0, FFA-17:1, HexCer-35:1, Chol-d47, CE19:0, LPE and PI 31:1 (Sigma-Aldrich). The internal standard PC-28:0 was used as a standard for PC, PCO, and PCP.

One sample from Ubiad1+/+ female was excluded from the analysis due to very deviant lipid values compared to the other 6 Ubiad1+/+ mice.

Association analysis

Correlations were calculated with the biweight midcorrelation (Bicor) in the HMDP as described13. Bicor calculates biweight mid-correlations, which are a robust (less sensitive to outliers) alternative to Pearson correlations. Genotypes for the mouse strains were obtained from Jackson Laboratory using the Mouse Diversity Array37. Around 200,000 single nucleotide polymorphisms (SNPs) were left after excluding SNPs with a minor allele frequency of less than 5% and a missing genotype rate of less than 10%. Genome-wide association study (GWAS) for gene expression was performed with FAST-LMM13, and defined as cis if peak SNP mapped within 1 Mb of gene position (threshold, p < 6.1 × 10−6).

Data analyses

For analyses of RNA sequencing, we used R and integrated Differential Expression and Pathway analysis (iDEP.96), which encompasses many R and Bioconductor packages38. Differential expression analyses were done using the DEseq2 pipeline on background filtered read counts. False discovery rate (fdr) < 0.05 and fold change (FC) > 1.5 was considered statistically significant. PCA plots were generated after variance-stabilizing transformation in DEseq2. Enrichment analyses were done using GAGE with GO biological process pathway annotations. For other analyses, we used Graphpad Prism (GraphPad Software, San Diego, CA, USA) software (version 10).