Molecular programming modulates hepatic lipid metabolism and adult metabolic risk in the offspring of obese mothers in a sex-specific manner

Male and female offspring of obese mothers are known to differ extensively in their metabolic adaptation and later development of complications. We investigate the sex-dependent responses in obese offspring mice with maternal obesity, focusing on changes in liver glucose and lipid metabolism. Here we show that maternal obesity prior to and during gestation leads to hepatic steatosis and inflammation in male offspring, while female offspring are protected. Females from obese mothers display important changes in hepatic transcriptional activity and triglycerides profile which may prevent the damaging effects of maternal obesity compared to males. These differences are sustained later in life, resulting in a better metabolic balance in female offspring. In conclusion, sex and maternal obesity drive differently transcriptional and posttranscriptional regulation of major metabolic processes in offspring liver, explaining the sexual dimorphism in obesity-associated metabolic risk.

Maternal obesity is a relevant health issue in many countries nowadays. It has been shown in many epidemiological but also animal studies that the offspring is programmed to metabolic disease on the long run. This has also been shown to be sex-dependent, with males usually being the more susceptible sex. Thus, the authors here study the consequences of maternal high-fat feeding and obesity on the offspring's' liver. The experiments are clearly structured. However, neither the experimental setup nor the analyses done are novel but have been performed in similar ways by several laboratories in the last decade, although not to this level of detail. Also the finding are interesting but largely confirmative.
I see two major problems in this study. First, the dams are fed control or high fat diet until weaning, leading to two groups of offspring. However, in the next step all offspring is weaned to high fat diet. Therefore, a comparison to a "real" control group (control diet in pregnancy and control diet for the offspring) is not possible. This limits the value of the experiments. Second, the authors focus on the long-term outcome in relation to sex, which should be attributed to hormonal differences. This neglects the possibility that already the initial programming step (in utero) might be sex dependent. This is especially disappointing as they mention "in utero" in the title of the manuscript. It would have been extremely interesting to see data on these initial steps induced by the high fat diet.
Reviewer #2 (Remarks to the Author): Dear authors, Your study presents an extensive analysis of the effects of maternal obesity in a rodent model and adds a great value to what is already known with regards to sex-specific programming, and sex differences in normal physiology and disease. Combination of various methods used makes your study comprehensive. Especially notable is a thorough lipidomic analysis. For the reviewer results presented are reliable and accountable, however, before publication some focal points should be addressed:  4. Authors state that "Altogether these results indicated that MO alters the liver transcriptome to a much larger extent in females than in males' -however, authors should consider if number of animals analysed (6 females vs 3 males) could have an effect on the outcome of the analysis? 5. For RNA isolation, purity and integrity determination -authors should state how was the integrity determined and state RIN numbers 6. Can authors justify why an FDR of 0.1 rather than 0.05 was used for pathway analysis? How many of genes analysed has passed a cut-off of FDR 0.05? Minor comments 1. Line 85 -comma missing after TG accumulation 2. Line 97 -BW abbreviation needs to be explained here 3. Line 105 -abbreviation maternal obesity (MO) has been introduced earlier 4. Line 108-110 'At MID, F-moHF had higher ratio of total fat (TF) on BW than M-moHF and M-moHF but not F-moHF accumulated less fat compared to moC at MID which vanished at ENDsentence really difficult to understand. 5. Line 117-122 -'Males showed reduced glucose tolerance and insulin sensitivity (high fasting insulin level and during OGTT) compared to females; but females only showed reduced insulin sensitivity by MO during the OGTT (higher insulin levels) (Figs.1h-1i). Surprisingly, MO do not alter glucose disappearance after insulin injection (Fig.1j), which indicated pancreatic disturbances in F-moHF during the OGTT -sentences difficult to understand -consider rewording. 6. Line 126 -in all conditions -specify which conditions 7. Line 127 -physiological changes -can you specify? 8. Line 146 -'highe' typo 9. Line 140 -141 'The four DE genes Pdk1, Lpin1, Prlr and Nox4 have been known to act as key regulators of hepatic insulin sensitivity' -can you provide reference? 9. Line 148-149 -'Interestingly, hepatic ERS1 has been shown to regulate G6pc and Pck1 and to be critical in the regulation of glucose metabolism -can you provide reference? 10. Figure 1q (PCK1) -not consistent with other figures 11. Lines 305-307 -'Interestingly, MO caused the downregulation of the Nsdhl, G6pdx, Rbm3, Ctps2 and Acls4 genes, which are involved in lipid and cholesterol metabolism and the upregulation of the Apex2 gene that is linked to hepatocarcinoma development' -can you provide references? 12. Line 308 -Apex2 and Pgrmc1 -are these genes only involved in hepatocarcinoma? Is there another role/explanation? 13. Lines 313-315 -'We observed that Il13ra1 and Kdm5c encoding for liver lipid homeostasis were higher expressed in females, and Atp11c and Acsl4 encoding for genes connected to lipid disorders and hepatocarcinoma development were higher expressed in males -can you provide references? 14. Lines 355-356 -'reduced expression levels of RNA Binding Motif Protein (Rbm)3, CTP synthase (Ctps)2 and Acsl4 X-linked genes involved in hepatocellular carcinoma development' -can you provide references? 15. Line 368-369 -'Remarkably, we found two key genes (Osgin1 and Stat1) that are known tumor repressors -can you provide reference? 16. Line 402 -'at early life stage of life' -repetition 17. Lines 516-517 -'for HFD mother and HFD offspring) and the group of offspring born from CD fed mother named moC (for CD mother and HFD offspring) -is that correct?

Reviewers' comments:
We appreciate the time and effort that the editor and the reviewers dedicated to providing feedback on our manuscript. We would like to thank the reviewers for their thoughtful comments and efforts towards improving our manuscript. Please find our point-by-point response to the reviewers' concerns below.
Reviewer #1 (Remarks to the Author): Maternal obesity is a relevant health issue in many countries nowadays. It has been shown in many epidemiological but also animal studies that the offspring is programmed to metabolic disease on the long run. This has also been shown to be sex-dependent, with males usually being the more susceptible sex. Thus, the authors here study the consequences of maternal high-fat feeding and obesity on the offspring's' liver. The experiments are clearly structured. However, neither the experimental setup nor the analyses done are novel but have been performed in similar ways by several laboratories in the last decade, although not to this level of detail. Also the finding are interesting but largely confirmative. I see two major problems in this study. First, the dams are fed control or high fat diet until weaning, leading to two groups of offspring. However, in the next step all offspring is weaned to high fat diet. Therefore, a comparison to a "real" control group (control diet in pregnancy and control diet for the offspring) is not possible. This limits the value of the experiments. Second, the authors focus on the long-term outcome in relation to sex, which should be attributed to hormonal differences. This neglects the possibility that already the initial programming step (in utero) might be sex dependent. This is especially disappointing as they mention "in utero" in the title of the manuscript. It would have been extremely interesting to see data on these initial steps induced by the high fat diet.
First: "real" control group Answer: As mentioned in the introduction (line 64) and later discussed (line 413), we have indeed a control group with mothers and postweaning offspring on CD. However, we have already published in a separate paper the results (PMID: 33398027); the current manuscript being a follow-up study. In the current study we aimed to investigate the effect of maternal HFD on offspring metabolic profile when offspring are exposed to an obesogenic post-weaning diet (as often the case in humans). Further, we aimed to investigate whether the early life exposure (in utero and lactation) to HFD (compared to the CD) could accelerate the adverse effects of the post-weaning obesogenic diet in offspring. Therefore, we assumed using the group of offspring born from control diet mothers (moC) as our "control group" in the current manuscript.
Second: long term outcomes and sex hormones Answer: we agree that the long-term outcome in relation to sex could partly be explained by the differences in sex hormones and not solely to in utero programming by the maternal diet. However, a recent paper show sex differences in the metabolic response to maternal obesity at late gestation (E18.5) (PMID: 35025731). Significant differences in the heart lipidome were observed in female fetuses compared to male fetuses born from obese mothers. Another study in humans has shown that exposure to maternal obesity causes sex-dependent alterations in miRNA and gene expression in human fetal liver (PMID: 31852997).
It is rather fascinating how these differences between the sexes can occur given that fetuses of both sexes are exposed to the same maternal metabolic milieu. It is not clear if these sex-specific responses early in life could define an ability of females to adapt to the environment and to protect against longer term detrimental effects. This evidence shows that the intrauterine environment (exposure to obesogenic environment) during the period of cellular differentiation and growth might result in changes in cellular functions that will affect female and male offspring differently in the long term, regardless of their sexual hormones. However, the exact mechanisms that explain the sexual dimorphic response to maternal obesity remains still elusive.
Reviewer #2 (Remarks to the Author): Dear authors, Your study presents an extensive analysis of the effects of maternal obesity in a rodent model and adds a great value to what is already known with regards to sex-specific programming, and sex differences in normal physiology and disease. Combination of various methods used makes your study comprehensive. Especially notable is a thorough lipidomic analysis. For the reviewer results presented are reliable and accountable, however, before publication some focal points should be addressed: 4. Authors state that "Altogether these results indicated that MO alters the liver transcriptome to a much larger extent in females than in males' -however, authors should consider if number of animals analysed (6 females vs 3 males) could have an effect on the outcome of the analysis? Answer: We agree with the reviewer that the differences in the number of animals used for the analysis may have an effect on the outcome. Therefore, we made some additional analysis to look at the homogeneity of our group (tSNE plot).
1. In the unsupervised tSNE clustering plot, the six female samples clearly separate from the three male samples 2. When we use 6 female vs 3 male, the degree of freedom of wald test in Deseq2 is (6-1) + (3-1) = 7. When we randomly select 3 female vs 3 male, the degree of freedom of wald test in Deseq2 is (3-1)+ (3-1) = 4. Under the same expression values, the tested genes with higher degree of freedom tend to be more significant than the ones with lower degree of freedom. 3. Therefore we chose to select six females to increase the degree of freedom to avoid the possible false negative results and also account for outliers.

For RNA isolation, purity and integrity determination -authors should state how was the integrity determined and state RIN numbers
Answer: For the Smart-seq2 experiment (RNA-seq) we focused mostly on the quality of cDNA rather than the quality of RNA since its more important to get a good quality cDNA in order to continue with the sequencing. However, we checked the RIN values of the RNA, and we found an average number of 7.6.
6. Can authors justify why an FDR of 0.1 rather than 0.05 was used for pathway analysis? How many of genes analysed has passed a cut-off of FDR 0.05? Answer: We used a cut-off of FDR < 0.1 to reduce the false negative in our analysis, we did not want to miss any potential causative genes and pathways. Nevertheless, we extracted the DE genes with a cutoff of FDR of 0.05 to make a comparison. Below is the Venn diagram we generated with the genes that passed the cut-off of FDR 0.1 (yellow circle) and FDR 0.05 (blue circle) for the four comparisons. As expected, all genes with a cut-off of FDR at 0.05 are included into the set of genes that are found with a cut-off of FDR at 0.1, without major differences, except for the M-moC versus M-moHF comparison were almost 50% of the DE genes were lost with an FDR cut-off of 0.05 compared to 0.1 (but a very small number of genes are deregulated in males anyway). Minor comments 1. Line 85 -comma missing after TG accumulation Done 2. Line 97 -BW abbreviation needs to be explained here Done 3. Line 105 -abbreviation maternal obesity (MO) has been introduced earlier Done 4. Line 108-110 'At MID, F-moHF had higher ratio of total fat (TF) on BW than M-moHF and M-moHF but not F-moHF accumulated less fat compared to moC at MID which vanished at ENDsentence really difficult to understand. Answer: We have modified the sentence and we hope this is now clearer for the reader. Line 106: "At MID, M-moHF had lower ratio of total fat (TF) on BW than F-moHF due to reduction of fat mass compared to M-moC; this difference disappeared at END" 5. Line 117-122 -'Males showed reduced glucose tolerance and insulin sensitivity (high fasting insulin level and during OGTT) compared to females; but females only showed reduced insulin sensitivity by MO during the OGTT (higher insulin levels) (Figs.1h-1i). Surprisingly, MO do not alter glucose disappearance after insulin injection (Fig.1j), which indicated pancreatic disturbances in F-moHF during the OGTTsentences difficult to understand -consider rewording. Answer: we have rewritten the sentence to clarify the result. Line 115-119: "Males showed reduced glucose tolerance and insulin sensitivity (high fasting insulin level and during OGTT) compared to females. However, only females showed reduced insulin sensitivity by MO during the OGTT (higher insulin levels) (Figs.1h-1i and 1k-1l). Interestingly, insulin tolerance test showed that MO did not alter glucose disappearance in F-moHF (Fig.1j)." 6. Line 126 -in all conditions -specify which conditions Answer: We have changed the sentence. Line 121: "The quantitative insulin-sensitivity check index (QUICKI) indicated that males were more insulin resistant than females at both MID and END, regardless of maternal diet. 7. Line 127 -physiological changes -can you specify? Answer: we have added information to clarify the physiological changes. Line 123: "We tested whether the physiological changes described above (in vivo data) were accompanied by changes in gene expression in the liver." 9. Line 148-149 -'Interestingly, hepatic ERS1 has been shown to regulate G6pc and Pck1 and to be critical in the regulation of glucose metabolism -can you provide reference? Answer: We have added references for the genes listed in the results section. Line 145: "Interestingly, hepatic ESR1 has been shown to regulate G6pc 5 and Pck1 6 and to be critical in the regulation of glucose metabolism" ESR1 + G6pc (PMID: 32150359), ESR1 + Pck1 (PMID: 28490809) 10. Figure 1q (PCK1) -not consistent with other figures Answer: We have modified the Fig.1q to have it consistent with the other figures. 11. Lines 305-307 -'Interestingly, MO caused the downregulation of the Nsdhl, G6pdx, Rbm3, Ctps2 and Acls4 genes, which are involved in lipid and cholesterol metabolism and the upregulation of the Apex2 gene that is linked to hepatocarcinoma development' -can you provide references? Answer: We have provided the references for each of these genes. Line 300-305: "In females, MO reduced the expression of Nsdhl and Acls4 genes, which are involved in lipid and cholesterol metabolism 7,8 . Interestingly, MO altered expression of G6pdx, Rbm3, Ctps2 and Apex2 genes, that are linked to hepatocarcinoma development 9-12 . In males, MO reduced expression of Zdhhc9 gene, involved in cancer and metabolism 13 (Figs.4f-4g). Importantly, expression of X-linked genes Apex2, Acls4 and Pgrmc1, involved in lipid metabolism and hepatocarcinoma was sex-dependent 8,9,14  12. Line 308 -Apex2 and Pgrmc1 -are these genes only involved in hepatocarcinoma? Is there another role/explanation? Answer: We have provided additional references about the role of Apex2 and Pgrmc1 genes. Line 304: "Importantly, expression of X-linked genes Apex2, Acls4 and Pgrmc1, involved in lipid metabolism and hepatocarcinoma was sex-dependent 8,9,14 ." PGRMC1 is involved in many metabolic pathways including fatty liver amelioration (PMID: 30356113), cytochrome activities (PMID: 21825115), drug metabolism, cholesterol synthesis, and steroid synthesis (PMID: 18992768) and hepatic glucose metabolism (PMID: 34970218). 13. Lines 313-315 -'We observed that Il13ra1 and Kdm5c encoding for liver lipid homeostasis were higher expressed in females, and Atp11c and Acsl4 encoding for genes connected to lipid disorders and hepatocarcinoma development were higher expressed in males -can you provide references? Answer: We have provided the references for each of these genes Line: 307-310 "Expression of Il13ra1 and Kdm5c encoding for liver glucose 15 and lipid 16 homeostasis was higher in females, and expression of Atp11c and Acsl4 encoding for genes connected to lipid disorders and hepatocarcinoma development 8,17 was higher in males." Il13ra1 (PMID: 23257358), Acls4 (PMID: 33340617), Atp11c (PMID: 30018401) and Kdm5c (PMID: 32714863)