1. Human Biology and Medicine
Download icon

Obesity-linked suppression of membrane-bound O-acyltransferase 7 (MBOAT7) drives non-alcoholic fatty liver disease

  1. Robert N Helsley
  2. Venkateshwari Varadharajan
  3. Amanda L Brown
  4. Anthony D Gromovsky
  5. Rebecca C Schugar
  6. Iyappan Ramachandiran
  7. Kevin Fung
  8. Mohammad Nasser Kabbany
  9. Rakhee Banerjee
  10. Chase K Neumann
  11. Chelsea Finney
  12. Preeti Pathak
  13. Danny Orabi
  14. Lucas J Osborn
  15. William Massey
  16. Renliang Zhang
  17. Anagha Kadam
  18. Brian E Sansbury
  19. Calvin Pan
  20. Jessica Sacks
  21. Richard G Lee
  22. Rosanne M Crooke
  23. Mark J Graham
  24. Madeleine E Lemieux
  25. Valentin Gogonea
  26. John P Kirwan
  27. Daniela S Allende
  28. Mete Civelek
  29. Paul L Fox
  30. Lawrence L Rudel
  31. Aldons J Lusis
  32. Matthew Spite
  33. J Mark Brown  Is a corresponding author
  1. Cleveland Clinic, United States
  2. University of Cincinnati, United States
  3. Brigham and Women’s Hospital, Harvard Medical School, United States
  4. University of California, Los Angeles, United States
  5. Ionis Pharmaceuticals, Inc, United States
  6. Bioinfo, Canada
  7. Cleveland State University, United States
  8. University of Virginia, United States
  9. Wake Forest University School of Medicine, United States
Research Article
  • Cited 1
  • Views 1,179
  • Annotations
Cite this article as: eLife 2019;8:e49882 doi: 10.7554/eLife.49882

Abstract

Recent studies have identified a genetic variant rs641738 near two genes encoding membrane bound O-acyltransferase domain-containing 7 (MBOAT7) and transmembrane channel-like 4 (TMC4) that associate with increased risk of non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH), alcohol-related cirrhosis, and liver fibrosis in those infected with viral hepatitis (Buch et al., 2015; Mancina et al., 2016; Luukkonen et al., 2016; Thabet et al., 2016; Viitasalo et al., 2016; Krawczyk et al., 2017; Thabet et al., 2017). Based on hepatic expression quantitative trait loci analysis, it has been suggested that MBOAT7 loss of function promotes liver disease progression (Buch et al., 2015; Mancina et al., 2016; Luukkonen et al., 2016; Thabet et al., 2016; Viitasalo et al., 2016; Krawczyk et al., 2017; Thabet et al., 2017), but this has never been formally tested. Here we show that Mboat7 loss, but not Tmc4, in mice is sufficient to promote the progression of NAFLD in the setting of high fat diet. Mboat7 loss of function is associated with accumulation of its substrate lysophosphatidylinositol (LPI) lipids, and direct administration of LPI promotes hepatic inflammatory and fibrotic transcriptional changes in an Mboat7-dependent manner. These studies reveal a novel role for MBOAT7-driven acylation of LPI lipids in suppressing the progression of NAFLD.

https://doi.org/10.7554/eLife.49882.001

eLife digest

Non-alcoholic fatty liver disease, or NAFLD for short, is a medical condition that develops when the liver accumulates excess fat. It can lead to complications such as diabetes and liver scarring. In humans, mutations that inactivate a protein called MBOAT7 increase the risk of fat accumulating in the liver.

Genetic studies suggest that low levels of MBOAT7 in a human’s liver cells increase the severity of NAFLD. Yet the links between MBOAT7, NAFLD and obesity are not well understood. Helsley et al. used data from humans and from obese mice that had been fed a high-fat diet to investigate the relationship between NAFLD and MBOAT7. This revealed that people who are obese have lower levels of MBOAT7 in their livers. Next, obese mice were genetically manipulated to produce less MBOAT7, which led them to develop more severe NAFLD.

Helsley et al. then grew human liver cells in the laboratory and lowered their levels of MBOAT7, which led to excess fat accumulating in the cells. This increase in fat accumulation was, at least in part, due to how these cells metabolize fats when MBOAT7 is reduced: they start making more new fats and consume fewer lipids to produce energy.

These findings provide a link between obesity and liver damage in both humans and mice, and show how a decrease in MBOAT7 levels causes changes in fat metabolism that could lead to NAFLD. The results could drive new approaches to treating liver damage in patients with mutations in the gene that codes for MBOAT7.

https://doi.org/10.7554/eLife.49882.002

Introduction

Non-alcoholic fatty liver disease (NAFLD) is an increasingly common condition that affects roughly one-third of adults in the United States (Cohen et al., 2011; Machado and Diehl, 2016; Rinella and Sanyal, 2016; Wree et al., 2013). The expansion of adipose tissue in obese individuals is strongly associated with the development of NAFLD, yet mechanisms linking obesity to NAFLD and more advanced forms of liver disease such as NASH and cirrhosis are not well understood (Cohen et al., 2011; Machado and Diehl, 2016; Rinella and Sanyal, 2016; Wree et al., 2013). Genome-wide association studies (GWAS) provide a powerful unbiased tool to identify new genes and pathways that are involved in human disease, allowing for pinpoint accuracy in identification of new drug targets. This is exemplified by the recent success story of human genetic studies leading to rapid development of monoclonal antibodies targeting proprotein convertase subtilisin/kexin type 9 (PCSK9) for hyperlipidemia and cardiovascular disease (Hess et al., 2018). Within the last two years, several independent GWAS studies have identified a liver disease susceptibility locus (rs641738) within a linkage-disequilibrium block that contains genes encoding MBOAT7 and TMC4 (Buch et al., 2015; Mancina et al., 2016; Luukkonen et al., 2016; International Liver Disease Genetics Consortium et al., 2016; Viitasalo et al., 2016; Krawczyk et al., 2017; Thabet et al., 2017). Strikingly, the rs641738 variant is associated with all major forms of liver injury including NAFLD, alcoholic-liver disease (ALD), and viral hepatitis-induced fibrosis (Buch et al., 2015; Mancina et al., 2016; Luukkonen et al., 2016; International Liver Disease Genetics Consortium et al., 2016; Viitasalo et al., 2016; Krawczyk et al., 2017; Thabet et al., 2017). The rs641738 variant is associated with a C > T missense single nucleotide polymorphism (SNP) within the first exon the TMC4 gene, yet available data suggest that TMC4 is not abundantly expressed in human liver (Mancina et al., 2016). Based on eQTL studies in the liver it has been suggested that instead reduced expression and activity of MBOAT7 may be mechanistically linked to liver disease progression (Mancina et al., 2016; Luukkonen et al., 2016; International Liver Disease Genetics Consortium et al., 2016; Viitasalo et al., 2016; Krawczyk et al., 2017; Thabet et al., 2017).

The MBOAT7 gene encodes an acyltransferase enzyme that specifically esterifies arachidonyl-CoA to lysophosphatidylinositol (LPI) to generate the predominant molecular species of phosphatidylinositol (PI) in cell membranes (38:4) (Gijón et al., 2008; Zarini et al., 2014; Lee et al., 2012; Anderson et al., 2013). Given this biochemical activity, MBOAT7 is a unique contributor to the Lands’ cycle, which is a series of phospholipase-driven deacylation and lysophospholipid acyltransferase-driven reacylation reactions that synergize to alter phospholipid fatty acid composition, creating membrane asymmetry and diversity (Shindou and Shimizu, 2009; Shindou et al., 2009). It is important to note that MBOAT7, unlike other lysophospholipid acyltransferases, only diversifies the fatty acid composition of membrane PI species and not phospholipids with other head groups (Gijón et al., 2008; Zarini et al., 2014; Lee et al., 2012; Anderson et al., 2013). Despite recent progress in characterizing the selective biochemistry of MBOAT7 (Gijón et al., 2008), and the clear genetic links to liver disease (Buch et al., 2015; Mancina et al., 2016; Luukkonen et al., 2016; International Liver Disease Genetics Consortium et al., 2016; Viitasalo et al., 2016; Krawczyk et al., 2017; Thabet et al., 2017), there is no information regarding how MBOAT7 activity or its substrate (LPI) or product (PI) lipids impact liver disease progression. Here, we demonstrate that MBOAT7 expression is suppressed in obese humans and rodents. Furthermore, we show that Mboat7, but not Tmc4, loss of function in mice is sufficient to drive NAFLD progression, and show that Mboat7 substrate lipids (LPIs) may be critical mediators of obesity-linked liver disease progression.

Results

MBOAT7 expression is suppressed in obese humans and rodents

NAFLD progression is commonly associated with obesity, and it has even been suggested that obesity is a causative factor in the pathogenesis of NAFLD (Younossi et al., 2018). Although, there is now strong support that the common MBOAT7 SNP (rs641738) is associated with NAFLD, it is not known whether MBOAT7 expression is significantly altered in obese people. To investigate this, we obtained wedge liver biopsies from sequentially consenting obese bariatric surgery patients and normal weight controls (Schugar et al., 2017), and examined the expression levels of MBOAT7 mRNA. We found that hepatic MBOAT7 expression was dramatically reduced in obese people, when compared to normal weight controls (Figure 1A). It is important to note that this striking reduction in MBOAT7 expression was not due to presence of the rs641738 MBOAT7 SNP, as we found the minor allele frequency was quite similar in lean, obese, and severely obese subjects (Figure 1A). To follow up on our findings in human obesity, we also examined hepatic expression levels of Mboat7 in obese leptin-deficient mice and high fat diet-fed rats. Mboat7 expression was reduced in ob/ob mice compared to lean WT controls (Figure 1B). In agreement, we found that high fat diet-induced obesity in Sprague-Dawley rats was also associated with significant reductions in hepatic Mboat7 expression (Figure 1C). Collectively, these results demonstrate that hepatic expression of MBOAT7 is suppressed in obese humans and rodents.

Obesity suppress MBOAT7 expression in humans, mice, and rats.

(A) To investigate the impact human obesity has on MBOAT7 expression, we obtained wedged liver biopsies from sequentially consenting bariatric surgery patients and normal weight controls. We used these biopsies to isolate RNA and measure MBOAT7 expression via qPCR. The data are stratified based on BMI. A total of 52 liver biopsies were analyzed. Lean (BMI <30; n = 10), Obese (30 < BMI > 40; n = 13), and Severely Obese (BMI >40; n = 29) ***p≤0.001 compared to the lean group. Status of the rs641738 SNP is shown in the top inset. (B) Liver MBOAT7 expression was measured by QPCR from wild type and ob/ob mice 16–20 weeks of age (n = 8; **p≤0.01; two-tailed t-test; Data are presented as mean ± S.E.M.) (C) Liver MBOAT7 expression was measured by QPCR from Sprague Dawley rats fed a low fat or high fat high diet for 24 weeks (n = 8; **p≤0.01; two-tailed t-test; Data are presented as mean ± S.E.M.). (D–F) We used a systems genetics approach to examine links between MBOAT7 expression and metabolic traits in mice from the hybrid mouse diversity panel (HMDP). To induce obesity, all mouse strains represented in the HMDP were fed an obesity-promoting high fat and high sucrose diet. Across the different strains in the HMDP, we found that the expression of MBOAT7 in adipose tissue was strongly negatively correlated with body fat percentage (E), and HOMA-IR (F), while liver MBOAT7 expression was also negatively correlated with fat mass (D).

https://doi.org/10.7554/eLife.49882.003

Next, we used a systems genetics approach to examine links between Mboat7 expression and adiposity traits in mice by leveraging data generated using the hybrid mouse diversity panel (HMDP) (Ghazalpour et al., 2012). To induce obesity, all HMDP mouse strains were fed an obesity-promoting high fat and high sucrose diet (Parks et al., 2013). Across the different strains in the HMDP, we found that Mboat7 expression in the liver was only modestly correlated (r = −0.244, p=0.01) with adiposity (Figure 1D). However, Mboat7 expression in white adipose tissue (WAT) was strongly negatively correlated (r = −0.68, p=2.3e-15) with both fat pad weight and % body fat in male and female mice (Figure 1E). Given the fact that obesity is commonly associated with insulin resistance and type 2 diabetes mellitus (Cohen et al., 2011; Machado and Diehl, 2016; Rinella and Sanyal, 2016; Younossi et al., 2018), we also examined the relationship between Mboat7 expression and insulin sensitivity in the HMDP. We found that expression of Mboat7 in adipose tissue was negatively (r = −0.38, p=0.0002) associated with indices of insulin sensitivity (Figure 1F). Collectively, these data suggest that MBOAT7 may be mechanistically linked to the well-known association between obesity, insulin resistance, and NAFLD progression.

Mboat7 loss of function in mice is sufficient to promote NAFLD progression

To test whether MBOAT7 impacts obesity-linked NAFLD progression we utilized an in vivo knockdown approach in high fat diet-fed C57BL/6 mice. Metabolic phenotyping of global Mboat7-/- mice has been limited due to the fact that these mice are only viable for several weeks after birth due to the critical role Mboat7 plays in cortical lamination and neuronal migration in the brain (Lee et al., 2012; Anderson et al., 2013). To overcome this barrier, we have generated second-generation antisense oligonucleotides (ASOs), which predominately target liver, adipose tissue, and cells within the reticuloendothelial system to selectively knock down Mboat7 expression in adult mice using methods we have previously described (Schugar et al., 2017). This ASO approach allows us to circumvent the postnatal lethality of global Mboat7 deletion (Lee et al., 2012; Anderson et al., 2013), and permits investigation into obesity-linked liver disease progression with near complete loss of function of Mboat7 in the liver. Mboat7 ASO treatment resulted in significant reductions of Mboat7 mRNA and protein in the liver (Figure 2A), and white adipose tissue (Figure 2—figure supplement 1), without altering Mboat7 expression in the brain, spleen, heart, or skeletal muscle (Figure 2—figure supplement 2). Although Mboat7 ASO treatment significantly reduced adipose Mboat7 expression (Figure 2, Figure 2—figure supplement 1), this was not associated with alterations in body weight (Figure 2B), fat mass (Figure 2—figure supplement 1L), adipose gene expression (Figure 2—figure supplement 1A–K), food intake (Figure 2C), energy expenditure (Figure 2D, and Figure 2—figure supplement 3), or physical activity (Figure 2—figure supplement 3). Despite this lack of phenotypic differences in adipose tissue, Mboat7 knockdown resulted in large alterations in the liver lipid storage. Mboat7 ASO treatment promoted an increase in liver weight (Figure 2F) and striking hepatic steatosis in HFD-fed, but not chow-fed mice (Figure 2E,I). Importantly, Mboat7 ASO-driven hepatic steatosis was consistently seen with two distinct ASOs targeting different regions of the Mboat7 messenger RNA (Figure 2—figure supplement 4). Mboat7 ASO-driven hepatic steatosis was characterized by accumulation of triglycerides, free cholesterol, and cholesterol esters only in high fat fed cohorts (Figure 2J–M). Despite these significant alterations in hepatic lipids, Mboat7 knockdown did not dramatically alter the levels of triglycerides or cholesterol in circulating lipoproteins (Figure 2—figure supplement 5). Mboat7 knockdown was also associated with hepatocyte injury as indicated by elevated liver enzyme levels (AST and ALT), but this only occurred in high fat-fed mice (Figure 2G,H).

Figure 2 with 5 supplements see all
Mboat7 knockdown promotes hepatic steatosis and worsen liver injury.

(A and B) C57BL/6 mice were fed chow or HFD with concurrent Control and Mboat7 ASO Injections at 12.5 mg/kg week for 20 weeks. Liver Mboat7 expression was measured via qPCR (A); n = 8–10; *p≤0.05, ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test). The inset show a representative Western blot for hepatic MBOAT7 protein levels, which was replicated in n = 4 mice. (B) Body weight was measured weekly. (C) Food intake was measured in C57BL/6 mice at 12 weeks of diet and ASO injections (n = 6). (D) Oxygen consumption was measured by indirect calorimetry in C57BL/6 mice after 12 weeks of diet and ASO injections (n = 6). (E) Representative liver hematoxylin and eosin stained sections. 20x magnification. (F) Liver-to-body weight measurements from mice fed Chow and HFD with Control and Mboat7 ASO Injections for 20 weeks (n = 8–10; ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test). (G and H) Plasma ALT (G) and AST (H) levels were measured after 20 weeks of diet feeding and ASO injections (n = 8–10; ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test). (I) Percent steatosis quantified by a blinded board certified pathologist at the Cleveland Clinic (n = 3; ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test). (J–M) Hepatic triglycerides (J), hepatic cholesterol (K), hepatic esterified cholesterol (L), and hepatic free cholesterol (M) were measured colorimetrically (n = 8–10; **p≤0.01, ***p≤0.001, ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test). All data are presented as mean ± S.E.M., unless otherwise noted.

https://doi.org/10.7554/eLife.49882.004

To more comprehensively understand the global effects of Mboat7 knockdown on liver function, we performed unbiased RNA sequencing experiments in control versus Mboat7 ASO-treated mice (Figure 3). Principal component analysis of RNA expression profiles showed separation by ASO group according to principal component analysis and hierarchical clustering analyses (Figure 3A,B). ASO groups were also partitioned by unsupervised hierarchical clustering (data not shown). Mboat7 knockdown resulted in a number of differentially expressed genes (DEGs), with 124 DEGs being suppressed and 887 DEGs being upregulated by Mboat7 ASO treatment (top 50 DEGs shown in Figure 3B and overall changes in Figure 3C). To understand the major pathways affected by Mboat7 knockdown, we performed Gene Ontology Molecular Function enrichment analysis, and found that many of the pathways that were significantly enriched are mechanistically linked to liver injury such as leukocyte extravasation, monocyte/macrophage activation, and fibrosis/hepatic stellate cell activation (Figure 3D). As predicted Mboat7 mRNA levels were dramatically suppressed by Mboat7 ASO treatment, but quite unexpectedly we also saw that the expression of Tmc4 was reduced in Mboat7 ASO-treated mice (Figure 3B,E). These data shown for the first time that ASO-mediated knockdown of Mboat7 results in coordinated suppression of its neighboring gene Tmc4 (the rs641738 polymorphism is located in exon 1 of the TMC4 gene), indicating some potential cross talk that deserves further exploration (Figure 3E). Quantitative real time PCR (qPCR) analysis also confirmed the RNA sequencing results showing that Mboat7 knockdown resulted in elevated expression of gene associated with inflammation (Tnfa and Il1b) and early fibrosis (Col1a2 and Acta2) in high fat fed mice (Figure 3E). In addition, Mboat7 knockdown was associated with altered immune cell populations in the liver. Mboat7 knockdown increased Cd8+ T lymphocytes, while decreasing the total number of Cd11b+ macrophages in the liver (Figure 3—figure supplement 1). Of the hepatic macrophages that were present, more were skewed towards the proinflammatory M1 (Cd11c+) state, and less were skewed towards the alternative M2 (Cd206+) phenotype (Figure 3—figure supplement 1). Collectively, these data suggest that specifically under high fat feeding conditions, Mboat7 loss of function is associated with dysregulated immune cell homeostasis, inflammation, and gene expression signatures that are consistent with early activation of pro-fibrotic programs in the liver (Figure 3).

Figure 3 with 1 supplement see all
Mboat7 inhibition exacerbates inflammation and fibrotic gene expression in the liver.

(A–C) C57BL/6 mice were fed a HFD with concurrent Control and Mboat7 ASO Injections at 12.5 mg/kg week for 20 weeks. Liver RNA was used for RNA-sequencing. ASO groups clustered by principal component analysis (A). Row-normalized expression for the top 50 DEGs are shown by heat map (B) while the volcano plot (C) summarizes log2 fold changes vs significance in response to Mboat7 inhibition (n = 5; genes with FDR < 0.05 and fold change >2 were considered significantly differentially expressed). (D) Summary of significantly differentially regulated pathways in mice treated with a non-targeting Control ASO vs the Mboat7 ASO. (E) qPCR validation of selected gene expression changes discovered in the RNA-sequencing analysis in both Chow and HFD-fed mice receiving Control or MBOAT7 ASOs for 20 weeks (n = 8–10; *p≤0.05, ***p≤0.001, ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test). All data are presented as mean ± S.E.M., unless otherwise noted.

https://doi.org/10.7554/eLife.49882.010

Genetic deletion of Tmc4 does not result in hepatic steatosis

Given the fact that the rs641738 polymorphism is located in exon 1 of the TMC4 gene, and we unexpectedly found that Mboat7 ASO treatment also reduces Tmc4 expression (Figure 3B,E), we wanted to examine whether alteration in Tmc4 may also be a key regulator of hepatic steatosis. To understand the role of Tmc4 in hepatic steatosis, we generated global Tmc4 knockout mice using CRISPR-Cas9-mediated gene editing and examined hepatic lipid levels under high fat feeding conditions (Figure 4). Global Tmc4 knockout mice have marked reductions in hepatic Tmc4 mRNA (Figure 4A) and protein (Figure 4B), yet importantly Mboat7 expression is unaltered (Figure 4A). In contrast to the striking hepatic steatosis seen with Mboat7 loss of function (Figure 2E, Figure 2—figure supplement 4), Tmc4 null mice show similar levels of hepatic lipids when fed a high fat diet (Figure 4). These data strongly implicate Mboat7, but not Tmc4, as the primary mediator of hepatic steatosis seen with the rs641738 variant or in Mboat7 ASO-treated mice.

Tmc4 knockout mice do not exhibit fatty liver.

Male wild type (Tmc4+/+) or Tmc4 global knockout mice (Tmc4-/-) were fed a high fat diet for 2 weeks. (A) Hepatic gene expression for Tmc4 and Mboat7 measured by qPCR. (B) Western blot of TMC4 protein levels in the liver and kidney. (C) Representative H and E-stained liver sections from Tmc4+/+ and Tmc4-/-. (D) Hepatic triglyceride (TG) levels in Tmc4+/+ and Tmc4-/- were quantified biochemically. Data shown represent n = 5–6 mice per group, ***p<0.001; two-tailed t-test relative to Tmc4+/+ mice.

https://doi.org/10.7554/eLife.49882.012

Mboat7 knockdown promotes profound hyperinsulinemia and impairs hepatic insulin action

The abnormal accumulation of lipids (i.e., lipotoxicity) in tissues such as the pancreas and liver is frequently associated with the pathogenesis of type 2 diabetes (Unger and Scherer, 2010; Samuel and Shulman, 2018). Given that Mboat7 knockdown promoted striking hepatic steatosis (Figure 2E), and the fact that adipose tissue expression of Mboat7 is negatively correlated with insulin sensitivity across strains of the HMDP (Figure 1F) we examined glucose homeostasis in Mboat7 ASO-treated mice fed a high fat diet (Figure 5). Mboat7 knockdown was associated with significantly impaired systemic glucose tolerance (Figure 5A). Mboat7 ASO-treated mice also exhibited profound hyperinsulinemia in the fasted state and throughout an intraperitoneal glucose tolerance test (Figure 5B). In parallel, glucose-stimulated C-peptide release was elevated in Mboat7 ASO-treated mice, potentially indicating overproduction of insulin in the pancreatic β-cell (Figure 5C). During an intraperitoneal insulin tolerance test, Mboat7 ASO-treated mice showed blunted plasma glucose lowering (Figure 5D). To more directly measure tissue-specific insulin action in Mboat7 knockdown mice we examined the acute phosphorylation of the insulin receptor (IR) and protein kinase B (Akt) in response to a single insulin injection (Figure 5E–G). Mboat7 ASO-treated mice had reduced insulin-stimulated phosphorylation of IR and downstream Akt in the liver (Figure 5E). However, Mboat7 ASO treatment did not significantly alter insulin signal transduction in either skeletal muscle or white adipose tissue (Figure 5F,G). These data suggest that ASO-mediated knockdown of Mboat7 promotes hyperinsulinemia, yet impairs insulin action specifically in the liver.

A reduction in Mboat7 expression exacerbates hepatic insulin resistance in HFD-fed mice.

(A–C) C57BL/6 mice were fed a HFD with concurrent Control and Mboat7 ASO Injections at 12.5 mg/kg week for 20 weeks. Plasma glucose (A), plasma insulin (B), and plasma C-peptide (C) levels were measured throughout the course of an intraperitoneal glucose tolerance test (n = 4–5; *p≤0.05, **p≤0.01; two-tailed t-test relative to control ASO at same time point). (D) Plasma glucose was measured after an intraperitoneal insulin injection in mice fed HFD for 20 weeks with Control and Mboat7 ASO injections (n = 4–5; *p≤0.05, **p≤0.01; two-tailed t-test relative to control ASO at same time point). (E–G) Mice were fed a HFD for 20 weeks while injected with Control or Mboat7 ASOs, then saline or insulin (0.35 U/kg) was injected into the portal vein for 5 min. The insulin signaling pathway was examined via western blot in tissues: liver (E), skeletal muscle (F), and white adipose tissue (G) (n = 5) each sample represents an independent mouse; western blot is quantified by densitometry on the right; *p≤0.05; Two-way ANOVA with Tukey’s post-hoc test). All data are presented as mean ± S.E.M., unless otherwise noted.

https://doi.org/10.7554/eLife.49882.013

ASO-mediated knockdown of Mboat7 results in tissue-specific alterations in LPI and PI lipids

Given the fact that Mboat7 is known to catalyze the selective esterification of arachidonyl-CoA to LPI lipids in neutrophils (Gijón et al., 2008), we wanted to examine its lysophosphatidylinositol acyltransferase (LPIAT) activity in the liver and also examine both substrate LPI and product PI species across a range of tissues. First, we isolated hepatic microsomes from control and Mboat7 ASO-treated mice and assayed LPIAT enzymatic activity using saturated and monounsaturated LPI substrates (Figure 6A). In chow fed mice Mboat7 knockdown only modestly decreased hepatic LPIAT activity toward 18:1 LPI (Figure 6A), yet in high fat diet-challenged mice Mboat7 knockdown resulted in a highly significant ~50% reduction in LPIAT activity using either 16:0, 18:0, or 18:1 LPI substrates (Figure 6A). These data suggest that Mboat7 is a significant contributor to total hepatic LPIAT activity, especially under high fat feeding conditions. Aligned with these alterations in enzyme activity, knockdown of Mboat7 in high fat-fed mice resulted in alterations in LPI and PI lipids in a highly tissue-specific manner. Mboat7 ASO treatment did not significantly alter LPI levels in the circulation, but was associated with selective reduction in the 38:3 and 38:4 species of circulating PI lipids (Figure 6B, Figure 6C, and Figure 6—figure supplements 12). In contrast to effects in the circulation, Mboat7 knockdown was associated with significant accumulation of 16:0 and 18:1 LPI species in the liver (Figure 6D and Figure 6—figure supplement 2). Furthermore, Mboat7 ASO-treatment resulted in reduced 38:3 and 38:4 PI levels, while increasing more saturated (34:1, 34:2, 36:1, and 36:2) PI species in the liver (Figure 6E and Figure 6—figure supplement 1). In white adipose tissue, Mboat7 knockdown did not significantly alter LPI levels (Figure 6—figure supplement 2), but promoted marked accumulation of more saturated PI species both in chow and high fat fed mice (Figure 6—figure supplement 1). Also, Mboat7 ASO treatment resulted in significant reductions in several PUFA-enriched PI species (38:3, 38:4, 38:5, and 38:6) in adipose tissue (Figure 6—figure supplement 1). In contrast to alterations in LPI and PI levels in liver and adipose tissue (Figure 6 and Figure 6—figure supplements 12), Mboat7 ASO treatment did not significantly alter LPI or PI levels in the brain (Figure 6—figure supplement 3) or pancreas (Figure 6—figure supplement 4). In contrast to marked changes in inositol-containing phospholipids, it is also important to note that Mboat7 knockdown did not significantly alter the hepatic levels of other major phospholipids including phosphatidylcholines, lysophosphatidylcholines, phosphatidylethanolamines, lysophosphatidylethanolamines, phosphatidylserines, phosphatidylglycerols, or phosphatidic acids (Figure 6—figure supplement 5). Collectively, these data show that ASO-mediated knockdown of Mboat7 primarily alters LPI and PI levels in the liver and white adipose tissue, which creates the potential to induce an imbalance of local lipid mediators that originate from PI metabolism.

Figure 6 with 6 supplements see all
HFD-feeding and Mboat7 inhibition results in an accumulation of LysoPIs.

Male C57BL/6 mice were fed either chow or a HFD for 20 weeks along with receiving a control non-targeting ASO or an ASO targeting the knockdown of Mboat7. (A) Liver microsomes were isolated and assayed for lyso-phosphatidylinositol acyltransferase (LPIAT) activity, using [14C]-arachidonyl-CoA and either 16:0 LPI, 18:0 LPI, or 18:1 LPI as substrates. Plasma LPI (B) or PI (C) species were measured by LC-MS/MS (n = 5; *p≤0.05; two-tailed t-test). Liver LPI (D) and PI (E) species were measured by LC-MS/MS (n = 5–10 per group; *p≤0.05, **p≤0.01, ***p≤0.001; two-tailed t-test). All data are presented as mean ± S.E.M.

https://doi.org/10.7554/eLife.49882.014

LPI lipids increase hepatic inflammatory and fibrotic gene expression in an Mboat7-dependent manner

The major enzymatic product of MBOAT7 (38:4 PI) is a potential reservoir for the generation of arachidonic acid-derived immunomodulatory lipid mediators (Serhan et al., 2015; Dennis and Norris, 2015). Therefore, we initially hypothesized that Mboat7 loss of function might limit the pool of arachidonic acid in PI lipids available for arachidonic acid-derived lipid mediator production. However, when we examined arachidonic acid-derived pro-inflammatory (LTB4, PGE2, PGD2, PGF2α, and TXB2) and pro-resolving (15R-LXA4) lipid mediators there was no apparent alteration in the liver of Mboat7 ASO-treated mice (Figure 6—figure supplement 6). Collectively, these data do not support a rate-limiting role for MBOAT7 in arachidonic acid-derived lipid mediator production in mice. Given the fact that Mboat7 knockdown is associated with increases in LPI lipids, we next hypothesized that MBOAT7 substrate LPI lipids themselves may be the main drivers of liver disease progression under conditions of HFD-induced obesity and Mboat7 loss of function. In further support of this concept, we found that saturated LPI lipids are significantly elevated in the circulation of humans with advanced fibrosis compared to healthy controls (Figure 7).

Circulating levels of saturated lysophosphatidylinositol (LPI) lipids are increased in humans with advanced fibrosis.

Plasma from healthy subjects with no fibrosis (Ishak score = 0) or pathology proven advanced fibrosis (Ishak score = 4) were analyzed by LC-MS/MS to quantify MBOAT7 substrates (16:0 LPI, 18:0 LPI, and 18:1 LPI) and product (38:4 PI) lipids. All data are presented as mean ± S.E.M from n = 10 subjects per group, and group differences were determined using Mann-Whitney testing.

https://doi.org/10.7554/eLife.49882.021

To more directly test the hypothesis that LPI lipids may promote liver disease in a Mboat7-dependent manner, we directly treated mice with LPI lipids to transiently increase circulating LPI levels within a physiologic level (~2 fold) (Figure 8—figure supplement 1). Injection of 16:0, 18:0, or 18:1 LPI into control ASO-treated mice did not alter hepatic inflammatory or fibrotic gene expression (Figure 8 and Figure 8—figure supplement 2). This is not surprising, given that it was previously reported that exogenously provided LPI is rapidly esterified into membrane PI pools (Darnell and Saltiel, 1991; Darnell et al., 1991; Jackson and Parton, 2004). In other words, under normal conditions MBOAT7 activity rapidly acylates LPI to divert this signaling lipid into a membrane PI storage pool. We next postulated that under conditions such as obesity where Mboat7 activity is diminished, LPI-driven signaling can be sustained and promote liver injury. In support of this concept, direct injection of exogenous 18 carbon LPIs can rapidly (24 hr) increase the expression of genes characteristic of hepatic inflammation (Cd11c, Tnfa, IL1b) and early fibrosis (Desmin, Col1A1, Col1A2, and Acta2) in mice with Mboat7 knockdown (Figure 8 and Figure 8—figure supplement 2). These data suggest that 18 carbon LPI lipids can acutely induce hepatic inflammatory and fibrotic gene expression programs, but only when MBOAT7 function is compromised, as is seen in obesity or with loss of function variants like rs641738. It is important to note that 18 carbon LPI lipids can only significantly elicit such pro-inflammatory and pro-fibrotic effects in high fat fed mice (Figure 8 and Figure 8—figure supplement 2), as we did not see this same effect in chow fed cohorts (not shown). Although direct administration of 18 carbon LPI lipids in chow fed mice did not significantly alter the same pro-inflammatory or pro-fibrotic genes as we found with Mboat7 knockdown (Figure 3), we did find that a small subset of acute phase response genes (serum amyloid A genes, Saa1 and Saa2) and other immunomodulatory genes (Gdf15, Ly6d, Lcn2, Socs2, etc.) were altered by 18:0 LPI treatment in a Mboat7-dependent manner (Figure 8—figure supplement 3). Interestingly, when liver triacylglycerol levels were examined in this LPI injection experiment, we only found significantly elevation in the saline and 18:1 LPI injection groups, but not in the 16:0 LPI injection group (Figure 8I). Collectively, these results suggest that 18 carbon LPI lipids can alter hepatic inflammatory transcriptional programs, but this is highly reliant on both diminished Mboat7 expression/activity and high fat diet feeding.

Figure 8 with 3 supplements see all
Lysophosphatidylinositol (LPI) lipids stimulate pro-inflammatory and pro-fibrotic gene expression in a Mboat7-dependent manner.

Male C57BL/6 mice were treated with either a non-targeting control ASO or Mboat7 ASO while being fed a high fat diet for 11 weeks. Thereafter, mice were injected IP with saline (vehicle) or 12.5 μg of the Mboat7 substrate lipids 16:0 or 18:1 LPI (2 injections at 7 am and seven pm), and liver was collected 12 hr later. (A–H) Liver mRNA levels were measured by qPCR. (I) Liver triacylglycerol (TG) levels. Data are presented as mean ± S.E.M. (n = 5; *p≤0.05).

https://doi.org/10.7554/eLife.49882.022

Steatosis Seen with Mboat7 Knockdown is Associated with Reorganization of the Hepatic Lipid Droplet Lipidome and Proteome in vivo, and Alterations In Lipogenesis and Fatty Acid Oxidation in Cultured Cells

Mboat7 loss of function results in a striking accumulation of neutral lipids including triglycerides and cholesteryl esters (Figure 2E–I), yet the mechanism(s) behind this are poorly understood. We therefore examined several potential mechanisms driving the mixed hepatic steatosis in Mboat7 ASO-treated mice. First, we examined the expression of genes involved in lipogenesis, fatty acid oxidation, and cholesterol sensing and export under both fed and fasted conditions (Figure 9). Unlike many other models of hepatic steatosis, there were no significant alterations in lipogenic gene expression either in the fed or fasted state with Mboat7 knockdown (Figure 9A). Unexpectedly, the expression of carnitine palmitoyl transferase 1 (Cpt1) was significantly elevated with Mboat7 knockdown (Figure 9A), but this would not be expected to promote the accumulation of neutral lipids. The expression of enzymes involved in cholesterol biosynthesis (Hmgcr and Hmcgs1) were modestly reduced in Mboat7 ASO-treated mice, but only in the fasted state (Figure 9A). In addition, expression of the cholesterol efflux regulator Abca1 was elevated in Mboat7 ASO-treated mice, but only in the fasted state. Altogether, these minor differences in hepatic gene expression are unlikely to be driving the lipid accumulation in Mboat7 ASO treated mice. Next, we evaluated whether Mboat7 may influence the export of neutral lipids via packaging on nascent very low density lipoproteins (VLDL). However, Mboat7 ASO-treated mice did not have significant differences in VLDL-TG secretion during a detergent block (Figure 9B). Furthermore, steady state plasma levels of triglycerides and total cholesterol were not significantly altered in Mboat7 ASO-treated mice other than a minor reduction in cholesterol content in both low density lipoproteins (LDL) and high density lipoproteins (HDL) only in chow-fed cohorts (Figure 2—figure supplement 5). Another common cause of fatty liver is increased delivery of adipose-derived fatty acids to the liver, as is commonly seen with prolonged fasting and certain types of lipodystrophies. However, Mboat7 knockdown did not alter basal or catecholamine-stimulated adipocyte lipolysis in vivo (Figure 9C), ruling out a role for altered adipose lipolysis as a contributing factor. Finally, we examined the possibility that MBOAT7 may be a determinate of metabolism locally at the surface of cytosolic lipid droplets to regulate hepatic steatosis, given that one recent study reported that MBOAT7 can localize to cytosolic lipid droplets (Mancina et al., 2016). We confirmed that MBOAT7 can indeed be found in lipid droplets isolated by sucrose gradient fractionation (Figure 9F), and plays a regulatory role in both the lipidome and proteome of cytosolic lipid droplets. Cytosolic lipid droplets isolated from Mboat7 ASO-treated mice showed marked accumulation of 18:0 LPI and 20:4 LPI, but not 16:0 LPI or 18:1 LPI (Figure 9D). This is in stark contrast to what is seen in the whole liver, where Mboat7 knockdown instead promotes accumulation of 16:0 LPI and 18:1 LPI (Figure 6D). Also, lipid droplets isolated from Mboat7 ASO-treated mice have a reduced level of 38:4 PI but a reciprocal increase in 36:3 and 38:3 PI (Figure 9E), which is quite different when to compared to effects in whole liver (Figure 6E). In addition to alterations in the lipid droplet lipidome, Mboat7 knockdown is associated with accumulation of several proteins involved in lipid synthesis and storage on isolated lipid droplets (Figure 9F). Both perilipin 2 and 3 (PLIN2 and PLIN3) were much more abundant on lipid droplets isolated from Mboat7 ASO-treated mice, as were the critical lipid synthetic enzymes CTP:phosphocholine cytidylyltransferase α (CCTα) and glycerol-3-phosphate 4 (GPAT4). Given that recent reports have shown that lipid droplet targeting of CCTα and GPAT4 are critical regulators of the overall size and triglyceride storage capacity of lipid droplets (Guo et al., 2008; Krahmer et al., 2011; Wilfling et al., 2013), these proteomic alterations at the lipid droplet surface may to contribute to the hepatic steatosis seen in Mboat7 ASO-treated mice. In order to understand cell autonomous effects of MBOAT7 lipid metabolism we generated MBOAT7-deficient cells via CRISPR-Cas9-mediated genome editing. Huh7 lacking MBOAT7 have increased lipid droplets upon fatty acid loading (Figure 10A,B). This lipid accumulation was due in part to increases in de novo lipogenesis rates (Figure 10C), and reciprocal decreases in fatty acid oxidation rates (Figure 10D). However, MBOAT-deficient Huh7 cells did not have altered turnover of stored triacylglycerol (Figure 10E) or cholesterol ester (Figure 10F).

Mboat7 knockdown promotes hepatic steatosis by altering the lipidome and proteome of the cytosolic lipid droplets.

Male C57BL/6 mice were fed either chow or a HFD for 12–20 weeks along with receiving a control non-targeting ASO or an ASO targeting the knockdown of Mboat7. (A) Mice were necropsied either in the fed or fasted (12 hr fast) state, and the hepatic expression of genes involved in lipogenesis, fatty acid oxidation, or sterol sensing analyzed using qPCR. (B) Plasma triglyceride (TG) secretion rate in mice administered Triton WR-1339 (tyloxapol). (C) Fed mice were treated with either saline vehicle or the β3-adrenergic receptor agonist CL316243 (0.1 mg/g body weight) to stimulate adipocyte lipolysis, and plasma glycerol and NEFA were measured after 15 min. Panels D–F), ASO-treated mice were fasted for 4 hr (from 9:00 a.m. to 1:00 p.m.) prior to necropsy, and hepatic lipid droplets were isolated using sucrose gradient fractionation to analyze both the lipidome and proteome of cytosolic lipid droplets. Lipid droplet fractions were extracted and analyzed by LC-MS/MS to quantify either lysophosphatidylinositol (D) or phosphatidylinositol (E) species on isolated lipid droplets. (F) Lipid droplet fractions were analyzed using Western blotting for MBOAT7, glycerol 3-phosphate dehydrogenase (as a non-lipid droplet fraction control), perilipin 2 (PLIN2), perilipin 3 (PLIN3), CTP:phosphocholine cytidylyltransferase α (CCTα), or glycerol-3-phosphate 4 (GPAT4); four individual mice per group are shown. Data are presented as mean ± S.E.M, (n = 4–6 per group; *p≤0.05, **p≤0.01, ***p≤0.001, ****p≤0.0001; Two-way ANOVA with Tukey’s post-hoc test).

https://doi.org/10.7554/eLife.49882.026
Genetic deletion of MBOAT7 in Huh7 cells results in cell autonomous alterations in lipogenesis and fatty acid oxidation.

CRISPR-Cas9 gene editing was used to delete the MBOAT7 gene from Huh7 hepatoma cells: (A), Immunoblot confirmation of genetic deletion of MBOAT7 (MBOAT7Δ): Microsome were isolated from WT and MBOAT7Δcells, and MBOAT7 protein levels were detected via Western blotting. (B) Fluorescent microscopy image of control and MBOAT7Δ-Huh7 cells were supplemented with 400 µM oleic acid for 24 hr. Following 24 hr of lipid loading, cells were washed with PBS and stained with Bodipy (493/503). Scale bar = 10 µm. (C) Effect of MBOAT7Δ on de novo lipogenesis: Control and MBOAT7Δ-Huh7 cells were incubated [14C]acetate in the presence of broad lipase inhibitors. Cells were harvested at various time point, lipids extracted and the incorporation of [14C]acetate into [14C]triacylglycerol. (D) Effect of MBOAT7Δ on complete fatty acid oxidation: The graph shows the amount of 14C-palmitic acid oxidized to 14C-CO (Mancina et al., 2016). (E–F), Effect of MBOAT7Δ on triacylglycerol (E) and cholesterol ester (F) hydrolysis: Control and MBOAT7Δ-Huh7 cells were incubated with 100 µM oleate complexed to bovine serum albumin + 1 µCi[3H]-cholesterol + 0.5 µCi-[14C]-oleate for 24 hr to label cellular cholesteryl ester and triacylglycerol pools to steady state. The supplemental fatty acids were withdrawn, and the cells were incubated with 6 µM triacsin C to inhibit triacylglycerol synthesis and re-esterificaiton. Cells were harvested, lipids were extracted and separated by thin layer chromatography to determine the turnover of cholesteryl esters and triacylglycerols. Data shown are mean ± S.E.M. and are representative of two separate experiments; ***p<0.0002, ****p<0.0001; two-way ANOVA with Bonferroni’s multiple comparisons test.

https://doi.org/10.7554/eLife.49882.027

Discussion

Given that several recent studies have found a strong link between the common rs641738 variant allele and liver disease progression, there is considerable interest in identifying the causative gene within this locus. Although the rs641738 polymorphisms maps to the first exon of the poorly annotated gene TMC4, this study is the first to demonstrate that genetic deletion of Tmc4 does not result in hepatic steatosis. Instead, selective loss of function of the neighboring gene Mboat7 is sufficient to sensitize mice to high fat diet-driven liver disease progression. The major findings of the current study include the following: (1) Genetic deletion of Tmc4 does not alter hepatic lipid storage, (2) Hepatic expression of MBOAT7 is reduced in obese humans and rodents, independent of rs641738 status, (3) Mboat7 expression in mouse liver and adipose tissue is negatively correlated with obesity and insulin sensitivity across the strains represented in the Hybrid Mouse Diversity Panel, (4) Mboat7 knockdown promotes hepatic steatosis, hepatocyte death, inflammation, and early gene expression profiles consistent with fibrosis, but this only happens when mice are challenged with a high fat diet, (5) Mboat7 loss of function promotes striking hyperinsulinemia and hepatic insulin resistance, (6) Mboat7 knockdown results in tissue-specific reorganization of its substrates (LPI) and product (PI) lipids, (7) LPI lipids can rapidly induce hepatic inflammatory and fibrotic gene expression programs in an Mboat7-dependent manner in mice, and (8) the hepatic steatosis seen with Mboat7 knockdown is not related to differences in lipogenic gene expression, VLDL secretion, or tissue lipolysis, but instead may be related to alterations in hepatic lipid droplet accumulation of lipogenic enzymes (CCTα and GPAT4) that allow for expansion of large lipid droplets (Guo et al., 2008; Krahmer et al., 2011; Wilfling et al., 2013), and increased rates of de novo lipogenesis and reduced fatty acid oxidation. Collectively, these data support a role for Mboat7-driven acylation of LPI lipids as a key protective mechanism against obesity-linked NAFLD progression.

The MBOAT family of enzymes are critical players in determining the composition of fatty acids in cellular membranes, and are emerging as key players in cardiometabolic disease (Rong et al., 2013; Harayama et al., 2014). It is important to note that several other lysophospholipid acyltransferase enzymes have been linked to inflammatory diseases (Rong et al., 2013; Harayama et al., 2014). In particular, lysophosphatidylcholine acyltransferases (LPCAT1, LPCAT2, and LPCAT3) are known to regulate tissue inflammation and endoplasmic reticulum stress by altering arachidonic acid availability as well as determining membrane phosphatidylcholine saturation (Rong et al., 2013; Harayama et al., 2014). In fact, the lysophospholipid acyltransferase family sits at a critical signaling nexus, given that they can play key roles in the generation of arachidonic acid-derived lipid mediators as well as regulating the levels of lysophospholipid signaling lipids (Shindou and Shimizu, 2009; Shindou et al., 2009). Given the fact that MBOAT7 preferentially generates PUFA-enriched PIs it also has the unique potential to also impact phosphorylated PI species (PtdInsP, PtdInsP2, and PtdInsP3, etc.) (Anderson et al., 2013). In agreement with our findings with PI lipids (Fil) in Mboat7 ASO-treated mice, global genetic deletion of Mboat7 selectively lowers the 38:4 molecular species of PIPs, while increasing more saturated species (Anderson et al., 2013). Therefore, we cannot rule out that alterations in PIP-dependent signal transduction may also play a role in MBOAT7 loss of function-driven liver disease progression.

Unexpectedly, Mboat7 knockdown did not significantly alter hepatic lipid storage or inflammation in chow-fed mice. Instead there is a clear unmasking of Mboat7-dependent fatty liver phenotypes only when mice are fed with a high fat diet. This could potentially indicate that unknown dietary factors play a regulatory role and limit hepatic LPIAT activity via unknown mechanisms (Figure 6A). Our data also support the idea that when mice, rats, and humans become obese MBOAT7 expression is reduced to varying degrees (Figure 1A–C), indicating regulation either at the transcriptional of post-transcriptional level. Future work should focus on understanding how dietary factors and adiposity itself impacts MBOAT7 expression and activity, and under what dietary and BMI conditions are MBOAT7 polymorphisms predicted to be most deleterious. In humans, the rs641738 variant has been clearly linked to end stage fibrotic liver disease (Buch et al., 2015; Mancina et al., 2016; Luukkonen et al., 2016; International Liver Disease Genetics Consortium et al., 2016). However, an important limitation of our studies is that high fat diet feeding in mice is not sufficient to drive bridging fibrosis. In future studies, it will be important to examine how Mboat7 loss of function impacts the development of fibrosis in appropriate fibrosis-prone animal models, and whether Mboat7 expression in hepatic stellate cells plays a regulatory role in the progression from NASH to cirrhosis. From this study it is clear that MBOAT7 can diversify the inositol-containing phospholipids and the associated proteome on cytosolic lipid droplets, and this could in part explain the large lipid droplets that accumulate in Mboat7 ASO-treated mice. It is interesting to note that the well-known PNPLA3 variant associated with fatty liver disease (I148M) accumulates on lipid droplets, and similarly reorganizes the lipidome and proteome to promote liver disease progression (BasuRay et al., 2019; Wang et al., 2019; Mitsche et al., 2018; BasuRay et al., 2017). Our findings here with MBOAT7-driven restructuring of the lipid droplet surface, and those recently published with PNPLA3 (BasuRay et al., 2019), suggest that alterations in lipid modifying enzyme access to the surface of cytosolic lipid droplets may be a common mechanism by which human fatty liver develops. Furthermore, we show that genetic deletion of MBOAT7 in Huh7 hepatoma cells can promote cell autonomous increases in cytosolic lipid droplets, and this enhanced lipid storage is associated with augmented lipogenesis rates and reduced fatty acid oxidation rates. Our studies provide new clues into new therapeutic leads for advanced liver diseases.

Another particularly striking finding from the current study is that Mboat7 knockdown promotes severe hyperinsulinemia (Figure 5B). Based on our results examining both insulin and C-peptide release during a glucose tolerance test (Figure 5B,C), it is reasonable to assume that Mboat7 knockdown is enhancing glucose-stimulated insulin secretion (GSIS) in pancreatic β cells. However, it is important to note that Mboat7 ASO treatment did not alter Mboat7 expression (data not shown) or LPI/PI lipids in the pancreas (Figure 6—figure supplement 4). Therefore, it is tempting to speculate that ASO-mediated knockdown of Mboat7 facilitates the production of an endocrine signaling lipid (possibly LPI) that may impact β cell GSIS. Interestingly, there are reports demonstrating that LPI lipids can stimulate β cell GSIS (Metz, 1986; Metz, 1988). It is interesting to note that expression of Mboat7 in white adipose tissue is modestly negatively correlated with HOMA-IR in the hybrid mouse diversity panel (Figure 1F), yet ASO-mediated knockdown of Mboat7 did not alter insulin signaling in adipose tissue (Figure 5F). These data suggest that additional studies are needed to clarify a potential role for Mboat7 in adipose tissue insulin sensitivity. Collectively, our data suggest that under conditions where MBOAT7 activity is suppressed (obesity or with the rs641738 variant) inefficient acylation of LPI substrate allows for these lipids to accumulate and initiate autocrine, paracrine, and potentially endocrine signaling that impact the progression of NAFLD and insulin resistance. Therefore, further investigation into the receptor system(s) that sense LPI lipids could hold therapeutic promise in liver disease and other associated metabolic diseases. Collectively, these studies identify MBOAT7-driven acylation of LPI lipids as an important modulator of both liver disease progression and associated type 2 diabetes.

Materials and methods

Key resources table
Reagent type
(species) or resource
DesignationSource or ReferenceIdentifiersAdditional
Information
Genetic reagent (M. musculus)C57BL/6NJ - C57BL/6NJ-Tmc4em1(IMPC)J/MmjaxJackson LaboratoryStock #: 46062-JAX; RRID:MGI:5882504;
RRID: IMSR_JAX:032275
Sequence-based reagentMboat7SigmaF: CATGCGGTACTGGAACATGA
R: CCAGTAGGCGCTCAGCAG
Sequence-based reagentCyclophilin ASigmaF: CCTGCTCTCCTCTCACCTCCT
R: TAGAAGCGCATGCGGAAGG
Sequence-based reagentMboat7SigmaF: atggcatgcgatactggaac
R: agcatggtccaggcactc
Sequence-based reagentTmc4SigmaF: CGGTACATCCACAAACATGG
R: GAGATCGTTCAGCACAGACG
Sequence-based reagentTnfαSigmaF: CCACCACGCTCTTCTGTCTAC
R: AGGGTCTGGGCCATAGAACT
Sequence-based reagentIL-1βSigmaF: AGTTGACGGACCCCAAAAG
R: AGCTGGATGCTCTCATCAGG
Sequence-based reagentCol1a2SigmaF: GCAGGTTCACCTACTCTGTCCT
R: CTTGCCCCATTCATTTGTCT
Sequence-based reagentα-SMASigmaF: GTCCCAGACATCAGGGAGTAA
R: TCGGATACTTCAGCGTCAGGA
Sequence-based reagentCd11cSigmaF: gagccagaacttcccaactg
R: tcaggaacacgatgtcttgg
Sequence-based reagentSrebp1cSigmaF: tctcactccctctgatgctac
R: gcaaccactgggtccaatta
Sequence-based reagentCol1a1SigmaF: ATGTTCAGCTTTGTGGACCTC
R: CAGAAAGCACAGCACTCGC
Sequence-based reagentFasSigmaF: GCTGCGGAAACTTCAGGAAAT
R: AGAGACGTGTCACTCCTGGACTT
Sequence-based reagentAcc1SigmaF: CCTGAGGAACAGCATCTCTAAC
R: GCCGAGTCACCTTAAGTACATATT
Sequence-based reagentScd1SigmaF: TTCCCTCCTGCAAGCTCTAC
R: CAGAGCGCTGGTCATGTAGT
Sequence-based reagentCpt1aSigmaF: TCCATGCATACCAAAGTGGA
R: TGGTAGGAGAGAGCAGCACCTT
Sequence-based reagentCgi58SigmaF: gcggtgatgaaagcgatg
R: caccctgtcagccatcctg
Sequence-based reagentAtglSigmaF: CTTCCTCGGGGTCTACCACA
R: GCCTCCTTGGACACCTCAATAA
Sequence-based reagentCd68SigmaF: ATCCCCACCTGTCTCTCTCA
R: ACCGCCATGTACTCCAGGTA
Sequence-based reagentF4/80SigmaF: GGATGTACAGATGGGGGATG
R: CATAAGCTGGGCAAGTGGTA
Sequence-based reagentDesminSigmaF: GTGGATGCAGCCACTCTAGC
R: TTGCCGCGATGGTCTCATAC
Sequence-based reagentRgs16SigmaF: AGGGCTCACCACATCTT
R: AGGTTTGTCTTGGTCAGTTC
Sequence-based reagentLcn2SigmaF: ccatctatgagctacaagagaacaat
R: tctgatccagtagcgacagc
Sequence-based reagentSaa1SigmaF: CCAGGATGAAGCTACTCACCA
R: TAGGCTCGCCACATGTCC
Sequence-based reagentSaa2SigmaF: ACTATGATGCTGCCCAAAGG
R: CTCTGCCGAAGAATTCCTGAAA
Sequence-based reagentLy6dSigmaF: CAACTGTAAGAACCCTCAGGTC
R: CACTCTTTCCTCACCAGGTTC
Sequence-based reagentOnecut1SigmaF: atcctcatgcccacctga
R: cctgaattacttccattgctga
Sequence-based reagentFasnSigmaF: GTCACCACAGCCTGGACCGC
R: CTCGCCATAGGTGCCGCCTG
Sequence-based reagentMe1SigmaF: ggagctccaggtccttagaata
R: tctgtcttgcaggtccattaac
Sequence-based reagentAclySigmaF: CTCACACGGAAGCTCATCAA
R: TCCAGCATTCCACCAGTATTC
Sequence-based reagentPparαSigmaF: GGAGGCGTTTCCTGAGACC
R: CAGCCACAAACGTCAGTTCAC
Sequence-based reagentLCADSigmaF: ccggttctttgaggaagtgaa
R: agtgtcgtcctccaccttctc
Sequence-based reagentHmgcrSigmaF: CTTGTGGAATGCCTTGTGATTG
R: AGCCGAAGCAGCACATGAT
Sequence-based reagentHmgcs1SigmaF: GCCGTGAACTGGGTCGAA
R: GCATATATAGCAATGTCTCCTGCAA
Sequence-based reagentAbca1SigmaF: GGGCTGCCACCTCCTCAGAGAAA
R: CACATCCTCATCCTCGTCATTC
Sequence-based reagentAbcg5SigmaF: TCCTGCATGTGTCCTACAGC
R: ATTTGCCTGTCCCACTTCTG
Sequence-based reagentCycloASigmaF: gcggcaggtccatctacg
R: gccatccagccattcagtc
Sequence-based reagentMBOAT7SigmaMBOAT7-E5-Nick-5FCACCGTCCATCAGGGAGGGCACGTC
Sequence-based reagentMBOAT7SigmaMBOAT7-E5-Nick-5RAAACGACGTGCCCTCCCTGATGGAC
Sequence-based reagentMBOAT7SigmaMBOAT7-E5-Nick-3FCACCGCAGCTACAGCTACTGCTACG
Sequence-based reagentMBOAT7SigmaMBOAT7-E5-Nick-3RAAACGTAGCAGTAGCTGTAGCTGC
AntibodyAnti-MBOAT7 (Rat monoclonal)PMID: 23097495RRID:AB_28138511:1000
AntibodyAnti-pIRβY1150/1151(Rabbit monoclonal)Cell SignalingRRID: AB_3312531:1000
AntibodyAnti-TMC4 (Rabbit polyclonal)Thermo FisherCat#: OSR00225W; RRID: AB_22041901:500
AntibodyAnti-IRβ (Rabbit monoclonal)Cell SignalingCat#: 3025S; RRID: AB_22804481:1000
AntibodyAnti-pAktS473(rabbit polyclonal)Cell SignalingCat#: 9271S; RRID: AB_3298251:1000
AntibodyAnti-pAktT308(Rabbit monoclonal)Cell SignalingCat#: 13038S; RRID: AB_26294471:1000
AntibodyAnti-Akt (Rabbit polyclonal)Cell SignalingCat#: 9272S; RRID: AB_3298271:1000
AntibodyAnti-PLIN2 (Rabbit polyclonal)Novus BiologicalsCat#: NB110-40877; RRID: AB_7879041:1000
AntibodyAnti-PLIN3 (Rabbit polyclonal)Novus BiologicalsCat#: NB110-40764; RRID: AB_7151161:1000
AntibodyAnti-CCTα (Rabbit monoclonal)Cell SignalingCat#: 6931; RRID: AB_108300581:1000
AntibodyAnti-GPAT4 (Rabbit Polyclonal)Novus BIologicalsCat#: NB100-2390; RRID: AB_22738111:1000
AntibodyAnti-GAPDH-HRP (Rabbit monoclonal)Cell SignalingCat#: 8884; RRID: AB_111298651:5000
AntibodyAnti-β-Actin-HRP (Mouse monoclonal)ProteintechCat#: HRP-60008; RRID: AB_22892251:5000
AntibodyAnti-CD3e-APC-Cy7 (Hamster monoclonal)BD Pharm; clone: 145–2 C11Cat#: 557596; RRID: AB_3967591:100
AntibodyAnti-CD4-Alexa Fluor 700 (Rat monoclonal)BD Pharm; clone: RM4-5Cat#: 557308; RRID: AB_3966341:100
AntibodyAnti-CD8a-PE-Cy7 (Rat monoclonal)eBiosciences; clone: 53–6.7Cat#: 25-0081-81; RRID: AB_4695831:100
AntibodyAnti-Cd11b-PE-Cy7 (Rat monoclonal)BD PharmCat#: 552850; RRID: AB_3944911:100
AntibodyAnti-Cd11c-PE-CF 594 (Hamster monoclonal)BD PharmCat#: 562454; RRID: AB_27376171:100
AntibodyAnti-CD206 Alexa Fluor 647 (Rat monoclonal)BiolegendCat#: 141712; RRID: AB_109004201:100
AntibodyAnti-rabbit IgG HRP (Donkey)GE-HealthcareCat#: NA934-100UL; RRID: AB_7722061:5000
AntibodyAnti-rat IgG HRP (Goat)Santa CruzCat#: SC-2006; RRID: AB_11252191:5000
AntibodyIRDye 800 anti-Rabbit IgG (Goat)LiCorCat#: 926–32211;
RRID: AB_621843
1:10,000
Sequence-based reagentMboat7Ionis
Commercial assay or kitSupersignal West Pico Plus substrateThermo Fisher34577
Commercial assay or kitAST kitSekisui Diagnostics319–30
Commercial assay or kitALT kitSekisui Diagnostics318–30
Commercial assay or kitLiver TGWako994–02891
Commercial assay or kitFree CholesterolWako993–02501
Commercial assay or kitPhospholipid CWako433–36201
Commercial assay or kitTotal CholesterolFisher SciTR134321
Commercial assay or kitInsulin ELISAMilliporeEZRMI-13K
Commercial assay or kitC-Peptide ELISACrystal Chem90050
Chemical compound, drug16:0 LPIAvanti Polar
Lipid, Inc
850102P50 μM
Chemical compound, drug18:0 LPIAvanti Polar Lipid, Inc850104P50 μM
Chemical compound, drug18:1 LPIAvanti Polar Lipid, Inc850100P50 μM
Chemical compound, drugArachidonyl Coenzyme A [arachidonyl-1–14C];
50–60 mCi/mmol1.85–2.22 GBq/mmol
American Radiolabeled
Chemicals
ARC 05190.025 µCi
OtherTLC silica gel Plate 60 F254Millipore Triton WR-13391055540001reagent
Chemical compound, drugΒeta-3 AR AgonistSigmaC5976-5mg
Chemical compound, drugTyloxapolSigmaT0307-10G500 mg/kg
Peptide, recombinant proteinInsulinSigmaI26430.35 U/kg

Human MBOAT7 expression levels in lean and obese subjects

Request a detailed protocol

The majority of subjects recruited to examine MBOAT7 expression levels were morbidly obese bariatric surgery patients, but we were able to obtain liver biopsies from 10 subjects with a BMI under 30 as normal weight controls. For recruitment, adult patients undergoing gastric bypass surgery at Wake Forest School of Medicine were consented via written consent and enrolled by a member of the study staff following institutionally approved IRB protocols as previously described (Shores et al., 2011). Exclusion criteria included: positive hepatitis C antibody, positive hepatitis B surface antigen, history of liver disease other than NAFLD, Childs A, B, or C cirrhosis, past or present diagnosis/treatment of malignancy other than non-melanocytic skin cancer, INR greater than 1.8 at baseline or need for chronic anticoagulation with warfarin or heparin products, use of immunomodulation for or history of inflammatory diseases including but not limited to malignancy, rheumatoid arthritis, psoriasis, lupus, sarcoidosis and inflammatory bowel disease, and greater or equal to seven alcohol drinks per week or three alcoholic drinks in a given day each week. In addition to bariatric surgery patients, a small number of non-obese subjects (body mass index <30.0) consented to liver biopsy during elective gall bladder removal surgery (n = 10). Each subject was assigned a unique identifier which was used throughout the study and did not include any identifiable information about the patient such as name, address, telephone number, social security number, medical record number or any of the identifiers outlined in the HIPAA Privacy Rule regulations. Only the principal investigator had access to the code linking the unique identifier to the study subject. Basic clinical information was obtained via self-reporting and a 15 ml baseline blood sample was obtained at the time of enrollment. A subset of this cohort has been previously described (Schugar et al., 2017; Shores et al., 2011). At the time of surgery, the surgeon collected a roughly 1-gram sample from the lateral left lobe. Wedge biopsies were rinsed with saline and immediately snap frozen in liquid nitrogen in the operating room before subsequent storage at −80°C. For data shown in Figure 7 showing levels of MBOAT7 substrate and product lipids, de-identified patient samples from the Cleveland Clinic hepatology clinic (IRB # 10–947) were analyzed. These patients had biopsy proven Ishak fibrosis scores (Ishak et al., 1995) of 0 (normal) or 4 (advanced fibrosis). For analysis of hepatic MBOAT7 expression, RNA isolated from liver biopsies were used for quantitative real time PCR (qPCR) as described below.

Rat studies of diet-induced obesity

Request a detailed protocol

Sprague Dawley Rats were received at 12 weeks of age and were housed in individual cages, kept at a constant temperature and ambient humidity in a 12-h light/dark cycle. Animals were then randomly assigned to either a standard chow diet or a high-fat diet (D12492, 60% fat, Research Diets, New Brunswick, NJ, USA) ad libitum to establish diet-induced obesity as previously described (Schugar et al., 2017). After 6 months of HFD-feeding, livers were excised for standard qPCR analysis of Mboat7 expression.

Hybrid Mouse Diversity Panel

Request a detailed protocol

92 inbred strains of 8-week-old male mice (180 individual mice) were fed a high fat, high sucrose diet (D12266B, Research Diets, New Brunswick, NJ) for 8 weeks before tissue collection (Parks et al., 2013). Gene expression of Mboat7 in white adipose tissue and liver were measured and correlated with obesity related traits using biweight midcorrelation analysis as previously described (International Liver Disease Genetics Consortium et al., 2016).

Mouse studies of Mboat7 loss of function

Request a detailed protocol

To explore the role of Mboat7 in diet-induced obesity, NAFLD progression, and insulin resistance, we utilized an in vivo knockdown approach in 8 week old adult mice. Selective knockdown of Mboat7 was accomplished using 2’-O-ethyl (cET) modified antisense oligonucleotides (ASO). All ASOs used in this work were synthesized, screened, and purified as described previously (Crooke et al., 2005) by Ionis Pharmaceuticals, Inc (Carlsbad, CA). For Mboat7 knockdown studies, adult (8 week old) male C57BL/6 mice were purchased from Jackson Labs (Bar Harbor, ME USA), and maintained on either a standard rodent chow diet or a high fat diet (HFD, D12492 from Research Diets Inc) and injected intraperitoneally biweekly with 12.5 mg/kg of either non-targeting control ASO or one of two independent ASOs directed against murine Mboat7 for a period of 20 weeks. Similar results were seen with two independent ASOs targeting different regions of the Mboat7 mRNA, hence key data using one Mboat7 ASO are shown. All mice were maintained in an Association for the Assessment and Accreditation of Laboratory Animal Care, International-approved animal facility, and all experimental protocols were approved by the Institutional Animal Care and use Committee of the Cleveland Clinic (IACUC protocols # 2015–1519 and # 2018–2053).

Studies in global Tmc4 knockout mice

Request a detailed protocol

Global Tmc4 knockout mice (C57BL/6NJ-Tmc4em1(IMPC)J/Mmjax) were provided by the Knockout Mouse Phenotyping Program (KOMP) at The Jackson Laboratory using CRISPR technology. Briefly, guide RNAs (GGAACCAGACCTTTTCCCAA and GAGTCAGCGTCAGAAAATGA) were designed to insert create a 277 bp deletion in exon 3 of the transmembrane channel-like gene family 4 (Tmc4) gene beginning at Chromosome 7 position 3,675,326 bp and ending after 3,675,602 bp (GRCm38/mm10). The mutation is predicted to delete ENSMUSE00001301550 (exon 3) and 131 bp of flanking intronic sequence, including the splice acceptor and donor, and is predicted to cause a change of amino acid sequence after residue 87 and early truncation six amino acids later. Guide RNAs and Cas9 nuclease were introduced into C57BL/6NJ-derived fertilized eggs with well recognized pronuclei. Embryos were transferred to pseudopregnant females. Correctly targeted pups were identified by sequencing and further bred to C57BL/6NJ (Stock No. 005304) to develop the colony. Once the stock mice arrived at the Cleveland Clinic heterozygous mice were intercrossed to generate additional heterozygous (Tmc4+/-) and homozygous wild type (Tmc4+/+) and knockout (Tmc4-/-) progeny. In addition to wild type mice on the mixed C57BL/6NJ background, we also studied parallel wild type mice on a pure C57BL/6J strain to increase sample size of experimental controls. Experimental mice were fed a high fat diet for 2 weeks prior to necropsy for analysis of hepatic steatosis.

Standardized necropsy conditions

Request a detailed protocol

To keep results consistent, the vast majority of experimental mice were fasted for 4 hr (from 9:00 a.m. to 1:00 p.m.) prior to necropsy. For the fasting versus fed experiments, fed mouse tissue were collected at the beginning of the light cycle (7:00 a.m.) in ad libitum fed mice, whereas the fasted group had food removed at 7:00 p.m. and were necropsy after a 12 hr fast (7:00 a.m.). At necropsy, all mice were terminally anesthetized with ketamine/xylazine (100–160 mg/kg ketamine-20–32 mg/kg xylazine), and a midline laparotomy was performed. Blood was collected by heart puncture. Following blood collection, a whole body perfusion was conducted by puncturing the inferior vena cava and slowly delivering 10 ml of saline into the heart to remove blood from tissues. Tissues were collected and immediately snap frozen in liquid nitrogen for subsequent biochemical analysis or fixed for morphological analysis.

Plasma ALT and AST analysis

Request a detailed protocol

To determine the level of hepatic injury in mice fed chow and HFD with ASO treatment for 20 weeks, plasma was used to analyze aminotransferase (AST) and alanine aminotransferase (ALT) levels using enzymatic assays (Sekisui Diagnostics, Lexington, MA, USA).

In vivo measurement of adiposity and lean mass

Request a detailed protocol

Quantitation of lean and fat mass were done using an EchoMRITM-130 Body Composition Analyzer (EchoMRI International).

Isolation and characterization of lipid droplets from mouse liver

Request a detailed protocol

Hepatic LDs were isolated by sucrose gradient centrifugation as we have previously described (Ferguson et al., 2017). Approximately 100 mg of tissue was minced with a razor blade on a cold surface. Minced tissue was transferred to a Potter-Elvehjem homogenizer, and then 200 μl of 60% sucrose was added to the tissue sample and incubated on ice for 10 min. Next, 800 μl of lysis buffer was added and mixed, and then incubated on ice for 10 min. Samples were homogenized with five strokes of a Teflon pestle and transferred to a 2 ml centrifuge tube. Lysis buffer (600 μl) was carefully layered on top of homogenate and centrifuged for 2 hr at 20,000 g at 4°C. The tube was then frozen at −80°C and cut at the 1,000 μl mark. The bottom piece of the centrifuge tube contained the non-LD fraction, which was allowed to thaw before being transferred to a new tube. The LD fraction was collected by cutting an ∼4–6 mm piece from the top of the ice cylinder and placing it in a new 2 ml tube. To increase the purity of the LD fraction, this process was repeated once more. Briefly, 200 μl of 60% sucrose was added to the LD fraction. Next, 800 μl of lysis buffer was added and mixed followed by careful layering with 600 μl of lysis buffer and then centrifugation for 2 hr at 20,000 g at 4°C. After freezing at −80°C, the tube was cut and the LD fraction was collected by cutting an ∼4–6 mm piece from the top of the ice cylinder and placing it in a new tube. Protein analysis was performed using the modified Lowry assay, as previously described (Ferguson et al., 2017), and Western blotting was performed as described below. MBOAT7 substrate (lysophosphatidylinositol) and product (phosphatidylinositol) lipids in the LD fraction were extracted and quantified using the targeted LC-MS/MS methods described below.

Quantitation of phosphatidylinositol (PI) and lysophosphatidylinositol (LPI) lipids

Request a detailed protocol

A targeted lipidomic assay for LPI and PI lipids was developed using HPLC on-line electrospray ionization tandem mass spectrometry (LC/ESI/MS/MS). Plasma and tissues (liver, brain, pancreas) from mice fed chow or HFD with ASO injections for 20 weeks were analyzed for precise detection of LPI and PI species (Figure 6; Figure 6—figure supplements 14). Moreover, plasma LPI levels were measured over time after intraperitoneal injections of 18:0 and 18:1 LPI (Figure 8—figure supplement 1). Standard Solutions: The standards used in this assay were all purchased from Avanti Polar Lipids (LPI-16:0, LPI-18:0, LPI-18:1, LPI-20:4, PI-38:4). Internal standards used for these analyses were LPI-17:1, PI-34:1-d31; all of which were purchased from Avanti Polar Lipids. Standard LPI and PI species at concentrations of 0, 5, 20, 100, 500 and 2000 ng/ml were prepared in 90% methanol containing two internal standards at the concentration of 500 ng/ml. The volume of 5 μl was injected into the Shimadzu LCMS-8050 for generating the internal standard calibration curves. HPLC Parameters: A silica column (2.1 × 50 mm, Luna Silica, 5 μm, Phenomenex) was used for the separation of PI and LPI species. Mobile phases were A (water containing 10 mM ammonium acetate) and B (acetonitrile containing 10 mM ammonium acetate). Mobile phase B at 95% was used from 0 to 2 min at the flow rate of 0.3 ml/min and then a linear gradient from 95% B to 50% B from 2 to 8 min, kept at 50% B from 8 to 16 min, 50% B to 95% B from 16 to 16.1 min, kept 95% B from 16.1 to 24 min. Mass Spectrometer Parameters: The HPLC eluent was directly injected into a triple quadrupole mass spectrometer (Shimadzu LCMS-8050) and the analytes were ionized (ESI negative mode). Analytes were quantified using Selected Reaction Monitoring (SRM) and the SRM transitions (m/z) were 571 → 255 for LPI-16:0, 599 → 283 for LPI-18:0, 597 → 281 for LPI-18:1, 619 → 303 for LPI-20:4, 885 → 241 for PI-38:4, 583 → 267 for internal standard LPI-17:1, and 866 → 281 for internal standard PI-34:1-d31. Data Analysis: Software Labsolutions LCMS was used to get the peak area for both the internal standards and LPI and PI species. The internal standard calibration curves were used to calculate the concentration of LPI and PI species in the samples. All plasma LPI and PI species were normalized to the PI-34:1-d31 internal standard, while all tissue LPI species were normalized to the 17–1 LPI internal standard and all tissue PI species were normalized to the PI-34:1-d31 internal standard.

Shotgun lipidomics to examine alterations in major phospholipid species

Request a detailed protocol

To more broadly examine the molecular lipid species in Mboat7 knockdown livers we utilized a shotgun lipidomics method for semi-quantitation of multiple lipid species using a method we have previously described (Gromovsky et al., 2018). All the internal standards were purchased from Avanti Polar Lipids, Inc (700 Industrial Park Drive, Alabaster, Alabama 35007, USA). Ten internal standards (12:0 diacylglycerol, 14:1 monoacylglycerol, 17:0 lysophosphatidylcholine, 17:0 phosphatidylcholine, 17:0 phosphatidic acid; 17:0 phosphatidylethanolamine, 17:0 phosphatidylglycerol, 17:0 sphingomyelin, 17:1 lysosphingomyelin, and 17:0 ceramide) were mixed together with the final concentration of 100 μM each. Total hepatic lipids were extracted using the method of Bligh and Dyer, with minor modifications. In brief, 50 μL of 100 μM internal standards were added to tissue homogenates and lipids were extracted by adding by adding MeOH/CHCl3 (v/v, 2/1) in the presence of dibutylhydroxytoluene (BHT) to limit oxidation. The CHCl3 layer was collected and dried under N2 flow. The dried lipid extract was dissolved in 1 ml the MeOH/CHCl3 (v/v, 2/1) containing 5 mM ammonium acetate for injection. The solution containing the lipid extract was pumped into the TripleTOF 5600 mass spectrometer (AB Sciex LLC, 500 Old Connecticut Path, Framingham, MA 01701, USA) at a flow rate of 40 μL/min for 2 min for each ionization mode. Lipid extracts were analyzed in both positive and negative ion modes for complete lipidome coverage using the TripleTOF 5600 System. Infusion MS/MSALL workflow experiments consisted of a TOF MS scan from m/z 200–1200 followed by a sequential acquisition of 1001 MS/MS spectra acquired from m/z 200 to 1200. The total time required to obtain a comprehensive profile of the lipidome was approximately 10 min per sample. The data was acquired with high resolution (>30000) and high mass accuracy (~5 ppm RMS). Data processing using LipidView Software identified 150–300 lipid species, covering diverse lipids classes including major glycerophospholipids and sphingolipids. The peak intensities for each identified lipid, across all samples were normalized against an internal standard from same lipid class for the semi-quantitation purpose.

Identification and quantification of AA-Derived lipid mediators by LC-MS/MS

Request a detailed protocol

In order to quantitate the abundance of lipid mediators generated from arachidonic acid, livers collected from C57BL/6 mice exposed to each treatment condition were subjected to solid phase extraction (SPE) and targeted liquid chromatography-tandem mass spectrometry (LC-MS/MS) as previously described in detail (Dalli et al., 2018) Briefly, tissue was minced in ice-cold methanol containing internal deuterated standards (d4-LTB4, d4-PGE2 and d5-LXA4) used to assess extraction recovery. Samples were loaded on C18 SPE cartridges and eluted methyl formate fractions were concentrated under a gentle stream of N2 gas. Samples were then resuspended in methanol:water (50:50) and analyzed using a high performance liquid chromatograph (HPLC, Shimadzu) coupled to a QTrap5500 mass spectrometer (AB Sciex) operating in negative ionization mode. Lipid mediators were identified using specific multiple reaction monitoring transitions, retention time, and diagnostic fragmentation spectra as compared with authentic standards. Quantification of lipid mediators was accomplished using calibration curves constructed with external standards for each mediator.

RNA sequencing and analysis

Request a detailed protocol

RNA was isolated via RNAeasy lipid tissue mini kit (Qiagen) from livers in which mice were fed a high-fat diet and treated with ASOs for 20 weeks. RNA samples were checked for quality and quantity using the Bio-analyzer (Agilent). RNA-SEQ libraries were generated using the Illumina mRNA TruSEQ Directional library kit and sequenced using an Illumina HiSEQ4000 (both according to the Manufacturer’s instructions). RNA sequencing was performed by the University of Chicago Genomics Facility. Raw sequence files will be deposited in the Sequence Read Archive before publication (SRA). Single-ended 50 bp reads were trimmed with Trim Galore (v.0.3.3, http://www.bioinformatics.babraham.ac.uk/projects/trim_galore) and controlled for quality with FastQC (v0.11.3, http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc) before alignment to the Mus musculus genome (Mm10 using UCSC transcript annotations downloaded July 2016). Reads were aligned using STAR in single-pass mode (v.2.5.2a_modified, https://github.com/alexdobin/STAR) (Dobin et al., 2013) with standard parameters but specifying ‘--alignIntronMax 100000 --quantMode GeneCounts’. Overall alignment ranged from 88–99% with 61–75% mapping uniquely. Transcripts with fewer than one mapped read per million (MMR) in all samples were filtered out before differential expression (DE) analysis. The filtering step removed 12,692/24,411 transcripts (52%). Raw counts were loaded into R (http://www.R-project.org/) (R Development Core Team, 2015) and edgeR (Robinson et al., 2010) was used to perform upper quantile, between-lane normalization, and DE analysis. Values generated with the cpm function of edgeR, including library size normalization and log2 conversion, were used in figures. Heat maps were generated using pheatmap (Kolde, 2015). DAVID (v.6.8) (Huang et al., 2009) was used to identify enriched functional annotations in DE gene ID lists relative to the set of ‘expressed’ genes (defined as having a median count across samples > 1 read per million mapped). Functional annotation to gene ontology was also performed using Ingenuity-IPA software (Ingenuity Systems, Inc Redwood City, CA.) as previously described (Thomas and Bonchev, 2010).

Glucose and insulin tolerance tests

Request a detailed protocol

Glucose and insulin tolerance tests were conducted as previously described (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009; Helsley et al., 2016) in male mice following 20 weeks of concurrent chow and HFD-feeding with ASO treatment. During the glucose tolerance test plasma insulin and C-peptide levels were measured via ELISA following manufacturer’s instructions (Millipore).

Plasma lipid and lipoprotein analyses

Request a detailed protocol

Total plasma triacylglycerol levels (L-Type TG M, Wako Diagnostics) and total plasma cholesterol levels were quantified enzymatically (Infinity Cholesterol Reagent, Thermo/Fisher). The distribution of cholesterol across lipoprotein classes was performed by fast-protein liquid chromatography (FPLC) using tandem superose-6 HR columns coupled with an online enzymatic cholesterol quantification as previously described (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009; Helsley et al., 2016).

In vivo insulin signaling studies

Request a detailed protocol

WT mice were fed a HFD for 20 weeks while injected with control or Mboat7 targeted ASOs. At 20 weeks, mice were fasted overnight (7 p.m. to 7 a.m.) and were injected with saline or 0.35 U/kg insulin into the portal vein for 5 min, as previously described (Lord et al., 2016). After 5 min, mice were euthanized and tissues were collected for protein isolation and western blot analysis.

Measurement of hepatic Very Low Density Lipoprotein (VLDL) Triglyceride Secretion in vivo

Request a detailed protocol

Control and Mboat7 ASO-treated mice were fasted overnight, and the following morning rec were injected with 500 mg/kg Triton WR-1339 (tyloxapol, Sigma). and blood was collected every 90 min thereafter. To measure hepatic VLDL secretion, serum TG levels were measured using enzymatic biochemical assays at each time point as described below.

General biochemical plasma and hepatic lipid analysis

Request a detailed protocol

Extraction of liver lipids and quantification of total plasma and hepatic triglycerides, cholesterol, and cholesterol esters was conducted using enzymatic assays as described previously (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009).

Lysophosphatidylinositol acyltransferase biochemical assay

Request a detailed protocol

[1-14C]-arachidonyl-CoA was obtained from American Radiolabeled Chemicals. Lysophosphatidylinositol (LPI) substrates (16:0 LPI, 18:0 LPI, and 18:1 LPI) and lipid standards used in the enzyme assays were obtained from Avanti Polar Lipids. Liver microsomes isolated from Mboat7 ASO and control ASO treated mice on both chow and high fat diet was used to measure LPIAT activity. The assay buffer contained 50 mM Tris-HCL (pH 8.0), 150 mM NaCl, 50 μM 18:0-LPI, 20 μM [1-14C]arachidonyl-CoA (0.025 µCi), and 15 µg of the microsomes in a total volume of 100 µL. Substrate was prepared in CHAPS (0.01 mM final concentration). The assay mixture was incubated for 30 min at 37°C, and the reaction was stopped by the addition of 1:2:1 (v/v/v) chloroform:methanol:2% orthophosphoric acid. The lipids were extracted and separated on a silica–TLC plate using chloroform/acetone/acetic acid/methanol/water (50:20:15:10:5, v/v) as the solvent system. The individual lipid molecules were identified by migration with respect to standards. Enzymatic products were monitored by autoradiogram, corresponding spots were scraped from the TLC plate, and the radioactivity was quantified with a liquid scintillation counter.

Immunoblotting

Request a detailed protocol

Whole tissue homogenates were made from tissues in a modified RIPA buffer as previously described (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009), and protein was quantified using the bicinchoninic (BCA) assay (Pierce). Proteins were separated by 4–12% SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membranes, and proteins were detected after incubation with specific antibodies as previously described (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009). Quantification of blots was performed using ImageJ software (National Institute of Health). Information of antibodies is provided in the key resource table.

Real-Time PCR analysis of gene expression

Request a detailed protocol

Tissue RNA extraction was performed as previously described for all mRNA analyses (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009). qPCR analyses were conducted as previously described (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b), and mRNA expression levels were calculated based on the ΔΔ-CT method. qPCR was conducted using the Applied Biosystems 7500 Real-Time PCR System. Primers used for qPCR are listed in the Key Resource Table.

Histological analysis

Request a detailed protocol

Hematoxylin and eosin (H and E) staining of paraffin-embedded liver sections was performed as previously described (Warrier et al., 2015; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b; Izem and Morton, 2009). Histopathologic evaluation was scored in a blinded fashion by a board-certified pathologist with expertise in gastrointestinal/liver pathology (Daniela S. Allende – Cleveland Clinic).

Quantification of hepatic immune cell populations by flow cytometry

Request a detailed protocol

After 20 weeks of HFD-feeding and ASO injections, livers were excised, washed with 1X PBS, and immediately placed into RPMI with Type IV Collagenase (Sigma Aldrich, St. Louis, Missouri, Lot# 087K8630) and DNase I (Roche, Mannheim, German) for 45 min at 37° C. Digested clumps of liver were pressed through a 70 um strainer and washed with RPMI with 10% FBS. Cells were centrifuged at 50 g for 10 min; supernatant was then centrifuged at 50 g for 7 min. To pellet the NPC fraction, cells were centrifuged at 300 g for 7 min. Cells were resuspended in BD Pharm Lyse (BD Biosciences, San Jose, California) for 5 min on ice. Cells were washed with RPMI with 10% FBS and centrifuged at 300 g for 10 min. (Gibbons MA, American Journal of Respiratory and Critical Care Medicine, 184, 2011). For flow cytometry, cells were resuspended in FACS buffer (1x PBS, 1% BSA, 0.05% sodium azide). Cells were aliquoted into 96 well plates at a concentration of ~1×106 cells/mL. Cells were centrifuged at 830 x g for 4 min, resuspended in 50 µl FACS buffer containing 0.5 ug of Fcγ Block (clone 93, eBioscience, San Diego, California), and incubated for 15 min at room temperature. After blocking, cells were stained with a fluorochrome-conjugated antibody panel CD206, Ly6c, CD3, CD4, CD8, CD11b, and CD11c. – all described in the Key Resource Table) for 30 min at 4° C in the dark. Cells were washed and centrifuged at 830 x g for 4 min twice with FACS buffer. Stained cells are resuspended in 200 µl of 1% paraformaldehyde and kept in the dark at 4°C overnight. Stained cells were centrifuged at 830 x g for 5 min. Stained cells were resuspended in 300 µl of FACS buffer, and data were collected on a LSRII flow cytometer (Becton Dickinson Immunocytometry systems, Mountain View, CA). Data collected on the LSRII were analyzed using FlowJo software (Tree Star, Inc, Ashland Oregon).

Cell culture and generation of MBOAT7-Deficient Hepatoma Cells

Request a detailed protocol

Mycoplasma-tested hepatocellular carcinoma cells (Huh7) were cultured under standard conditions in Dulbecco-modified Eagle's minimum essential medium (D-MEM) (GIBCO, Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS, GIBCO), 1% l-glutamine, 1% penicillin-streptomycin and 1% nonessential amino acids in a 5% CO2-humidified chamber at 37°C. CRISPR-Cas9 genome editing was accomplished using methods previously described (Ran et al., 2013). MBOAT7 sgRNAs were designed by an online tool (https://www.benchling.com/) and cloned into the Lenti-CRISPER v2 vector (Addgene (Ran et al., 2013) with D10A nickase version of Cas9 (Cas9n)). Primers used in this study are listed in Supplementary Table X. MBOAT7 KO cell lines were generated following lentiviral transduction of the Lenti-CRISPER v2-Cas9 D10A- MBOAT7 sgRNA in Huh7 cells. MBOAT7 KO single cell clones were isolated and expanded following FACS sorting. MBOAT7 KO cells were validated by analyzing the expression of MBOAT7 by real-time PCR and Western blot. Primers used for gene editing are shown in the key resource table with the following primer names: MBOAT7-E5-Nick-5F, MBOAT7-E5-Nick-5R, MBOAT7-E5-Nick-3F, MBOAT7-E5-Nick-3R.

Measurement of fatty acid oxidation rates in wild type and MBOAT7Δ Huh7 cells

Request a detailed protocol

Complete fatty acid oxidation to CO2 was measured using methods we have previously described (Brown et al., 2003). In brief control and MBOAT7Δ HUH7 cells were grown to a confluency of 70–80% in DMEM. Upon reaching confluence, one million cells were seeded for both WT-HUH7 and MBOAT7Δ (n = 3) for 24 hr. After 24 hr the medium was replaced with fresh serum free DMEM medium with 0.4 µci of [1-14C]-palmitic acid in 0.3% BSA/100 µM cold palmitate to each plate, and the plate was quickly placed in an airtight CO2 collection chamber (60 ml Nalgene jar with a fitted rubber stopper, and hanging center-well collection bucket containing What-man filter paper soaked with 50 μl benzothonium hydroxide). Substrate incubation was carried out with different time points (0 min, 30 min, 60 min and 120 min) at 37°C. Following each incubation, the reaction was terminated by the addition of 100 µL of 0.5 M H2SO4 to the cells using syringe injection. Liberation of [14C]-CO2 was allowed to proceed for an additional 30 min, and then the center-well collection bucket was cut out of the collection chamber and delivered to a liquid scintillation vial. Production of [14C]-CO2 from [14C]-palmitic acid was determined by liquid scintillation counter. Radiation count was normalized to amount of protein, as quantified by BCA assay (Pierce).

Measurement of triacylglycerol and cholesterol ester hydrolysis rates in cells

Request a detailed protocol

Measurement of neutral lipid mobilization was accomplished using a pulse-chase approach essentially as described previously (Brasaemle et al., 2000). Both control and MBOAT7Δ Huh7 cells were grown till 80% confluent, one million cells were seeded in 6-well plates (WT-Huh7 and MBOAT7 KO, the next day fresh DMEM medium was added and the cells were incubated with 100 µM cold oleate complexed to bovine serum albumin + 1 µCi [3H]-cholesterol + 0.5 µCi [14C]-oleate for 24 hr. After 24 hr of pulse labeling-cells to reach steady state, cells were chased for different time points (0, 1 hr, 2 hr, 4 hr, 8 hr, 24, 48 hr) in media containing 6 µM triacsin C (from 1 mg/ml stock in Me2SO) without supplemental fatty acids to examine turnover without the re-esterification (triacsin C blocks acyl-CoA synthetase activity). Cells were rinsed with phosphate buffered saline and harvested by scraping at various times; lipids were extracted as previously described and separated by thin layer chromatography using hexane:diethyl ether:acetic acid (80:20:1) as a solvent system. Triacylglycerol (TAG) and cholesterol ester spots were scraped off the plate, and the incorporation of radioactivity into TAG was determined by liquid scintillation counter. Radiation count was normalized to amount of protein, as quantified by BCA assay (Pierce).

Measurement of de novo Lipogenesis in Cells

Request a detailed protocol

Measurement of de novo lipogenesis rates was accomplished by tracing [14C]-acetate into [14C]-triacylglycerol essentially as described previously (Brasaemle et al., 2000). Confluent cells of both wild type and MBOAT7Δ Huh7 cells were maintained in DMEM medium, from these 1 million cells of both WT and MBOAT7Δ were seeded to 6-well plates for 12 hr, after 12 hr the medium was replaced by serum free DMEM. To measure triacylglycerol synthesis, without the complication of parallel triacylglycerol hydrolysis, cells were treated with lipase inhibitors of triacylglycerol and cholesterol ester such as diethyl-p-nitrophenyl phosphate (E-600, 500 µM) and diethylumelliferyl phosphate (DEUP, 0.35 µM) for 1 hr. After 1 hr of lipase inhibitor treatment, the cells were incubated with human recombinant insulin (100 nM) as a stimulator for de novo lipogenesis for 30 min. After 30 min of insulin stimulation, the cells were incubated with 1 µCi [14C]acetate and cells were harvested at various time points (30 min, 60 min, 120 min and 240 min) post substrate addition. From each time point cells were rinsed with PBS (twice), lipids were extracted as described above and separated by thin layer chromatography using using hexane:diethyl ether:acetic acid (80:20:1) as a solvent system. Triacylglycerol (TAG) spot was scraped off of the plate, and the incorporation of [14C]-acetate into [14C]-triacylglycerol was determined by liquid scintillation counter. Radiation count was normalized to amount of protein, as quantified by BCA assay (Pierce).

Fluorescence microscopy

Request a detailed protocol

In preparation for fluorescence microscopy, 5 × 104 cells of WT and MBOAT7Δ were plated onto 22 mm square glass coverslips in 4-well cell culture chamber slide containing growth media supplemented with 400 µM oleic acid. After 24 hr, medium was removed, washed with PBS (twice) and cells were fixed in 4% paraformaldehyde for 30 min. After 30 min of incubation, fixed cells were washed two times with PBS and then BODIPY (1 mg/mL) was added incubated for additional 30 min in dark. After incubation the cells were washed with PBS (twice) followed by mounting the slides with ProLong Gold antifade reagent with DAPI (4′,6- Diamidine-2′-phenylindole dihydrochloride). Images were acquired using a Leica DMIRB upright microscope (Leica Microsystems, GmbH, Wetzlar, Germany) equipped with a Retiga SRV camera and QCapture Pro software (QImaging, Surrey, BC, Canada).

Western blot analysis

Request a detailed protocol

Microsome were isolated from single clones of MBOAT7 Δ and WT-HUH7 cells using microsome isolation kit from Abcam. Proteins were quantified from isolated microsome using BCA method, 30 μg of protein were loaded on SDA-PAGE gels, the expression of MBOAT7 is confirmed by using MBOAT7-rat primary antibody (1:1000) and anti-rat secondary antibody (1:5000).

Statistical analysis

All experiments consisted of a minimum of three replicates and data are presented as mean ± SD. GraphPad Prism 8.1.1 was used for data analysis. Two‐way ANOVAs with Bonferroni's multiple comparison tests were used to determine significant differences (p-value<0.0002(***),<0.0001(****).

Statistical analysis

Request a detailed protocol

Human data analyses were performed using R 3.1.0 (Vienna, Austria) and p<0.05 was considered statistically significant. All mouse data were analyzed using either one-way or two-way analysis of variance (ANOVA) where appropriate, followed by either a Tukey’s or Student’s t tests for post hoc analysis. Differences were considered significant at p<0.05. All mouse data analyses were performed using Graphpad Prism 6 (La Jolla, CA) software.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
    pheatma: Pretty Heatmaps
    1. R Kolde
    (2015)
    R Package Version.
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
    R: A Language and Environment for Statistical Computing
    1. R Development Core Team
    (2015)
    R Foundation for Statistical Computing, Vienna, Austria.
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62

Decision letter

  1. Peter Tontonoz
    Reviewing Editor; University of California, Los Angeles, United States
  2. Anna Akhmanova
    Senior Editor; Utrecht University, Netherlands
  3. Dennis Vance
    Reviewer; U Alberta, Canada

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "Obesity-Linked Suppression of Membrane-Bound O-Acyltransferase 7 (MBOAT7<) Drives Non-Alcoholic Fatty Liver Disease" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife. Although all reviewers believed that the study addressed an important problem, they also felt that the present data did not provide a sufficiently definitive mechanistic link between MBOAT7 and the development of hepatic pathology at this stage. Further, the extent of revisions required is beyond that expected for an eLife revision. However, the editors and reviewers are open to the possibility of evaluating a suitably revised manuscript as a de novo submission should you be able to substantially extend the current findings and respond to the reviewers concerns.

Reviewer #1:

Mboat7 is a lysoPI acyltransferase that has been identified as a candidate gene in GWAS studies for NAFLD. In this report, Helsley provide evidence that reduced Mboat7 activity drives the development of obesity-induced NAFLD in mouse models. They knocked down Mboat7 expression with an ASO in high-fat diet fed mice and found that Mboat7 deficient mice developed steatosis and inflammation in mouse liver, and insulin resistance. With respect to the mechanism involved, they showed that the levels of several LPIs and PIs were changed by Mboat7 knockdown. Finally, they showed that LPI treatment increased the expression of inflammatory genes in the livers of Mboat7-deficient mice, but not in control mice. Although the studies are of interest and obvious clinical relevance, the mechanistic insight into how Mboat7 affects hepatic lipid metabolism is limited at this stage.

Major points:

1) Using more than one Mboat7 ASO would increase the strength of the conclusions.

2) What is the mechanism causing lipid accumulation in the livers of ASO-treated mice? Is the expression of Srebp1c and lipogenic genes increased? The authors showed cholesterol in plasma lipoprotein particles was not altered, how about the TG profile in plasma lipoprotein particles? Is there any defect in VLDL secretion?

3) The diet regimen employed does not typically lead to fibrosis in mice. The authors showed increased expression of inflammatory and fibrotic genes, but to establish fibrosis they need to show Trichrome or Sirius staining. Similarly, they claim that stellate cells were activated, but need more direct data to support this.

4) The authors report that 16:0 and 18:1 LPI was increased upon Mboat7 suppression, but the levels of these two LPIs (~0.05 pmol/mg) are very low compared to 18:0 LPI (~5 pmol/mg). It is difficult to see how this change could be driving the phenotype. Is the total LPI level significantly increased? Is there any change in total PI levels? Why did the authors inject mice with 18:0 LPI, which was not altered by Mboat7 suppression, but not 16:0 or 18:1 LPI that was increased by Mboat7 knockdown? Do different LPI species have different effects?

5) For LPI treatment experiment, did the authors inject HFD fed mice or chow diet fed mice? Will an increase in LPI in chow diet fed mice induce a phenotype similar to HFD fed mice?

6) Does LPI treatment also affect other genes identified by RNA seq in figure 3? In other words, does LPI contribute to other phenotypes including steatosis and insulin resistance?

Reviewer #2:

Major comments::

The main focus of this paper is that reduced expression of MBOAT7 in the liver drives NAFLD (per the title) and NAFLD consequences. In my view, it's a quite interesting paper but is fundamentally flawed, as the ASO targeting MBOAT7 in murine liver also reduces expression of TMC4, the other candidate gene. Given this, it's not possible to determine which effects are due to MBOAT7 depletion, which is the focus of the paper.

Figure 1A-D; need more controls (positive and negative). Units for x axes in D-F are not clear. Also, In the Discussion section the authors state a major finding is that hepatic expression of MBOAT7 is reduced in obese human and rodents independent of rs641738 status, so would be nice to see this data.

Figure 2A; pertinent negative controls should be shown. The ASO used targets liver, adipose and reticuloendothelial system. Figure 2A and Figure 2—figure supplement 1 only show mRNA data for liver, brain, spleen, heart, skeletal muscle. It would be good to include data in adipose. All this should be in main figures. Also, given the main thrust of the study is NAFLD/NASH, the authors should show collagen staining and fibrosis pathology scores.

Figure 3E presents a major problem as the MBOAT7 ASO lowers mRNA for TMC4, making the rest of the experiments (most of the rest of the paper) not possible to interpret for MBOAT7 depletion. If the point of the paper is that both genes are coordinately regulated and somehow both function in these phenotypes, then more work needs to be done to understand this. However, the current paper is about MBOAT7.

Figure 5; results for other phospholipids and lysophospholipids should be shown. Is this defect specific for PI/LPI?

Figure 6; it's not clear exactly how LPI is given, how much, and whether it is "physiological".

Reviewer #3:

This is the first characterization of MBOAT7 knockdown mice. The characterization is good and complete with interesting findings. There are several important weaknesses- they don't actually establish whether MBOAT7 or TMC4 are independently or cooperatively involved in the resulting phenotype. I don't think with the experiments presented they have firmly established with the alterations in lipid species found are responsible for the inflammation and stellate cell activation. It is important they address my third major point prior to any publication as it has important implication if true. Finally, they don't address why TGs accumulate in liver. The latter seem pretty basic given this is how the region was identified in the first place. Overall, I think this is marginal for acceptance, but certainly more experiments would be required.

Major points:

1) The effect of the ASO on FAT Mboat7 expression should be included.

2) Regarding the experimental design of Figure 6. It is not clear whether the amount of 18:0 LPI injected resulted in a 2-fold increase in the total LPI concentration or in the amount of 18:0. The later would be appropriate for these experiments. It seems to directly link the LPI and the function of MBOAT7, only the species that were altered should induce the inflammation and stellate cell activation. There is a lack of controls in this important experiment, which should include species that were not altered by MBOAT7 knockdown.

3) What is the mechanism by which the MBOAT7 knockdown results in TG, CE and FC accumulation in liver?

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Obesity-linked suppression of membrane-bound O-Acyltransferase 7 (MBOAT7) drives non-alcoholic fatty liver disease" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Anna Akhmanova as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Dennis Vance (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The reviewers and editors agree that the manuscript has been improved since the original submission and several issues have been clarified. However, the mechanism underlying the steatosis is still unclear. The new data regarding alterations to the liver lipid droplet proteome are interesting but do not establish whether these changes are a cause or a consequence of steatosis. Given that steatosis is the focus of the paper, the consensus is that this issue must be addressed prior to publication.

Essential revisions:

1) Direct evidence of how TG, CE and free cholesterol accumulates in Mboat7-deficient livers is not provided. A higher abundance of GPAT4 in the lipid droplet fraction was observed, but it is unclear where the acyl-CoA substrates for GPAT are coming from if there is no increase in de novo lipogenesis or uptake from circulation. The abundance of PLIN2, PLIN3, GPAT4, and CCTα on lipid droplets may merely reflect the presence of more lipid droplets. Increased abundance of GPAT4 and CCTα on lipid droplets does not explain why the livers of MBOAT7-deficient mice contain more CE and free cholesterol. mRNAs linked to lipogenesis were not measured in conditions that stimulate lipogenesis (i.e. re-feeding) but were instead measured after fasting or at the start of the light cycle.

The following experiments could be performed in the given time frame to address this central issue:

- Measure lipogenic mRNAs in re-fed conditions (see papers by Horton and colleagues on SREBP1-dependent lipogenesis).

- The rate of liver lipid synthesis can be assessed by i.p injection of radiolabeled glycerol and measurement of glycerol incorporation into lipid classes after separation by TLC as in Miyazaki et al., 2001 or Millar et al., 2006.

- Alternatively, if they believe their data is consistent with an impairment in TG mobilization from LDs, then the rate of liver LD lipolysis can be assessed as in Schweiger et al., 2014. These measurements could be done in primary hepatocytes/cell culture models if that is easier to achieve.

- de novo lipogenesis as a contributor to fat accumulation could be assessed by injecting mice with 14C-acetate (after they had been fasted O/N and re-fed for 6 hours) as in Lian et al., 2018.

2) In the experiment in which LPI species were administered to mice, only mRNAs related to inflammation were assessed. Analysis of hepatic lipid metabolism should be included in this experiment, especially since the focus of the manuscript has shifted from fibrosis/inflammation to hepatic steatosis.

https://doi.org/10.7554/eLife.49882.036

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

Mboat7 is a lysoPI acyltransferase that has been identified as a candidate gene in GWAS studies for NAFLD. In this report, Helsley provide evidence that reduced Mboat7 activity drives the development of obesity-induced NAFLD in mouse models. They knocked down Mboat7 expression with an ASO in high-fat diet fed mice and found that Mboat7 deficient mice developed steatosis and inflammation in mouse liver, and insulin resistance. With respect to the mechanism involved, they showed that the levels of several LPIs and PIs were changed by Mboat7 knockdown. Finally, they showed that LPI treatment increased the expression of inflammatory genes in the livers of Mboat7-deficient mice, but not in control mice. Although the studies are of interest and obvious clinical relevance, the mechanistic insight into how Mboat7 affects hepatic lipid metabolism is limited at this stage.

Major points:

1) Using more than one Mboat7 ASO would increase the strength of the conclusions.

Given this valid concern we have now used two independent ASOs targeting different regions of the Mboat7 messenger RNA, and see consistent effects on hepatic lipid accumulation. We have added a new supplemental figure (Figure 2—figure supplement 4) in response to this concern.

2) What is the mechanism causing lipid accumulation in the livers of ASO-treated mice? Is the expression of Srebp1c and lipogenic genes increased? The authors showed cholesterol in plasma lipoprotein particles was not altered, how about the TG profile in plasma lipoprotein particles? Is there any defect in VLDL secretion?

To address this, we have completed a number of additional experiments to understand the mechanism by which MBOAT7 loss of function promotes hepatic steatosis. First, we have repeatedly seen that Mboat7 knockdown does not significantly alter Srebp1c or lipogenic gene expression. We have also consistently seen no alterations in steady state plasma triglyceride levels or VLDL secretion rates. These new data for lipogenic gene expression and VLDL secretion are now included in a new main figure (Figure 9). After ruling out these common causes of hepatic steatosis, we instead focused our attention on whether MBOAT7 knockdown could drive steatosis by locally altering the lipid and proteome composition of cytosolic lipid droplets (i.e. impacting lipid storage at the lipid droplet interface). The rationale for pursuing this possibility was that a recent manuscript (Mancina et al., 2016) suggested that MBOAT7 can co-purify with cytosolic lipid droplets in subcellular fractionation experiments. Using sucrose gradient lipid droplet fractionation protocols, we have now confirmed that MBOAT7 can be found on cytosolic lipid droplets in control mice but is essentially absent on lipid droplets from Mboat7 ASO-treated mice (Figure 9F). In parallel with MBOAT7 localization to lipid droplets, we found the lipid droplets isolated from Mboat7 ASO-treated mice show striking increases in MBOAT7 substrates (18:0 LPI and 20:4 LPI) and reduction in MBOAT7 products (38:4 PI). In fact, the differences seen in LPI and PI lipids are much more apparent in isolated lipid droplets (Figure 9) than what is seen in whole liver homogenates (Figure 6). Aligned with alterations in LPI and PI lipids, lipid droplets isolated from Mboat7 knockdown livers had a clear rearrangement of lipid droplet-associated proteins. Lipid droplets isolated from Mboat7 ASO-treated livers have increased abundance of several lipid droplet-associated proteins including perilipin 2, perilipin 3, as well as the lipid synthesizing enzymes CTP:phosphocholine cytidylyltransferase α (CCTα) and glycerol-3-phosphate acyltransferase 4 (GPAT4) (Figure 9F). Aligned with these new findings we have added the following paragraph to the Discussion section of the revised manuscript:

“Steatosis Seen with Mboat7 Knockdown is Associated With Reorganization of the Hepatic Lipid Droplet Lipidome and Proteome. Mboat7 loss of function results in a striking accumulation of neutral lipids including triglycerides and cholesteryl esters (Figure 2e-2l), yet the mechanism(s) behind this are poorly understood. We therefore examined several potential mechanisms driving the mixed hepatic steatosis in Mboat7 ASO-treated mice. First, we examined the expression of genes involved in lipogenesis, fatty acid oxidation, and cholesterol sensing and export under both fed and fasted conditions (Figure 9). Unlike many other models of hepatic steatosis, there were no significant alterations in lipogenic gene expression either in the fed or fasted state with Mboat7 knockdown (Figure 9a). Unexpectedly, the expression of carnitine palmitoyl transferase 1 (Cpt1) was significantly elevated with Mboat7 knockdown (Figure 9a), but this would not be expected to promote the accumulation of neutral lipids. The expression of enzymes involved in cholesterol biosynthesis (Hmgcr and Hmcgs1) were modestly reduced in Mboat7 ASO-treated mice, but only in the fasted state (Figure 9a). In addition, expression of the cholesterol efflux regulator Abca1 was elevated in Mboat7 ASO-treated mice, but only in the fasted state. Altogether, these minor differences in hepatic gene expression are unlikely to be driving the lipid accumulation in Mboat7 ASO treated mice. Next, we evaluated whether Mboat7 may influence the export of neutral lipids via packaging on nascent very low density lipoproteins (VLDL). However, Mboat7 ASO-treated mice did not have significant differences in VLDL-TG secretion during a detergent block (Figure 9b). Another common cause of fatty liver is increased delivery of adipose-derived fatty acids to the liver, as is commonly seen with prolonged fasting and certain types of lipodystrophies. However, Mboat7 knockdown did not alter basal or catecholamine-stimulated adipocyte lipolysis in vivo (Figure 9c), ruling out a role for altered adipose lipolysis as a contributing factor. Finally, we examined the possibility that MBOAT7 may be a determinate of metabolism locally at the surface of cytosolic lipid droplets to regulate hepatic steatosis, given that one recent study reported that MBOAT7 can localize to cytosolic lipid droplets2. We confirmed that MBOAT7 can indeed be found in lipid droplets isolated by sucrose gradient fractionation (Figure 9f), and plays a regulatory role in both the lipidome and proteome of cytosolic lipid droplets. Cytosolic lipid droplets isolated from Mboat7 ASO-treated mice showed marked accumulation of 18:0 LPI and 20:4 LPI, but not 16:0 LPI or 18:1 LPI (Figure 9d). This is in stark contrast to what is seen in the whole liver, where Mboat7 knockdown instead promotes accumulation of 16:0 LPI and 18:1 LPI (Figure 6d). Also, lipid droplets isolated from Mboat7 ASO-treated mice have a reduced level of 38:4 PI but a reciprocal increase in 36:3 and 38:3 PI (Figure 9e), which is quite different when to compared to effects in whole liver (Figure 6e). In addition to alterations in the lipid droplet lipidome, Mboat7 knockdown is associated with accumulation of several proteins involved in lipid synthesis and storage on isolated lipid droplets (Figure 9f). Both perilipin 2 and 3 (PLIN2 and PLIN3) were much more abundant on lipid droplets isolated from Mboat7 ASO-treated mice, as were the critical lipid synthetic enzymes CTP:phosphcholine cytidylyltransferase α (CCTα) and glycerol-3-phosphate 4 (GPAT4). Given that recent reports have shown that lipid droplet targeting of CCTα and GPAT4 are critical regulators of the overall size and triglyceride storage capacity of lipid droplets30-32, these proteomic alterations at the lipid droplet surface are likely to contribute to the hepatic steatosis seen in Mboat7 ASO-treated mice.”

3) The diet regimen employed does not typically lead to fibrosis in mice. The authors showed increased expression of inflammatory and fibrotic genes, but to establish fibrosis they need to show Trichrome or Sirius staining. Similarly, they claim that stellate cells were activated, but need more direct data to support this.

This is a valid point that high fat diet feeding does not typically lead to measurable fibrosis in mice. Since our original submission we have done a number of experiments to test whether MBOAT7 knockdown indeed stimulates bridging fibrosis as is seen in advanced human liver disease, but as brought forth by this astute reviewer we failed to find pathological or biochemical support of fibrosis in this mouse model. First, we had a board-certified liver pathologist (Dr. Daniela Allende, Cleveland Clinic) blindly score H&E-, pirosirius Red, and trichrome-stained liver sections from HFD-fed control and Mboat7 ASO-treated mice, and in all cases there was no measureable fibrosis scoring above background. To indirectly quantitate fibrosis potential in a complimentary way we analyzed liver hydroxyproline content biochemically, and again found no indication that hydroxyproline levels were altered in Mboat7 ASO-treated mice. Finally, we performed some preliminary studies in primary stellate cells treating with MBOAT7 substrate lipids (LPIs) and did not see any obvious signs of pro-fibrotic responses in primary stellate cells. Given that we do not see any signs of even early stage fibrosis in Mboat7 ASO-treated mice, we have removed any discussion regarding effects on fibrosis and instead focus the discussion on the striking hepatic lipid accumulation seen with MBOAT7 loss of function in high fat fed mice. Although beyond the scope of the current study, we are now generating in parallel either hepatocyte, macrophage, or stellate cell specific Mboat7 knockout mice and will subject them to models of fibrosis induction (carbon tetrachloride, methionine/choline-deficient diet, bile duct ligation, etc.) to understand the cell autonomous effect of Mboat7 in pro-fibrotic responses.

To make this issue clear for the reader we have removed any claims that fibrosis was found in our studies, and have added the following sentences to the Discussion section:

“However, an important limitation of our studies is that high fat diet feeding in mice is not sufficient to drive bridging fibrosis. In future studies it will be important to examine how Mboat7 loss of function impacts the development of fibrosis in appropriate fibrosis-prone animal models, and whether Mboat7 expression in hepatic stellate cells plays a regulatory role in the progression from NASH to cirrhosis.”

4) The authors report that 16:0 and 18:1 LPI was increased upon Mboat7 suppression, but the levels of these two LPIs (~0.05 pmol/mg) are very low compared to 18:0 LPI (~5 pmol/mg). It is difficult to see how this change could be driving the phenotype. Is the total LPI level significantly increased? Is there any change in total PI levels? Why did the authors inject mice with 18:0 LPI, which was not altered by Mboat7 suppression, but not 16:0 or 18:1 LPI that was increased by Mboat7 knockdown? Do different LPI species have different effects?

This is a very insightful comment that was generally shared by all three reviewers, so we felt it was important to perform additional experiments to address the concern. We originally injected mice with 18:0 LPI because it by far is the most abundant LPI in the circulation, and we followed a similar logic as this astute reviewer that the most abundant species is the most likely culprit as a signaling lipid. We also knew from the literature that 18:0 LPI was previously described as a bona fide MBOAT7 substrate in biochemical assays from the Voelker laboratory (Gijón et al., 2008). We also now show that the acylation of 18:0 LPI is lower in Mboat7 ASO-treated mice (Figure 6A), particularly when challenged with a high fat diet. However, we agree that the primary changes seen with Mboat7 ASO treatment in the liver is instead the more minor species 16:0 and 18:1. To better understand how these more minor species of LPI (16:0 and 18:1) could impact liver disease in the context of Mboat7 loss of function, we performed new experiments examining whether direct injection of 16:0 or 18:1 LPI could induce liver injury in control or Mboat7 ASO-treated mice. Results from these new experiments are now included as a new main figure (Figure 8), and the results from the 18:0 LPI injections are included as an online supplemental figure to provide a comprehensive look at LPI subspecies effects (Figure 8—figure supplement 4). Interestingly, both 18:1 and 18:0 LPI induced expression of pro-inflammatory and fibrosis marker genes in a MBOAT7-dependent manner, but this was not apparent with 16:0 LPI administration. These new data provide additional evidence in support of substrate-specific effects in vivo, and highlight the importance of 18 carbon LPIs as potential proinflammatory and pro-fibrotic signaling lipids when MBOAT7 activity is compromised.

5) For LPI treatment experiment, did the authors inject HFD fed mice or chow diet fed mice? Will an increase in LPI in chow diet fed mice induce a phenotype similar to HFD fed mice?

To help get at this question, we have now performed LPI injection experiments in either chow-fed or HFD-fed cohorts of mice, and these new experiments are described in the revised manuscript. We have found that the ability of 18 carbon LPIs to induce pro-inflammatory and fibrosis related gene expression in Mboat7 ASO treated mice really only occurs when mice are fed a high fat diet and not seen in chow-fed mice. We have included a new supplementary figure to show the modest transcriptional changes following 18:0 LPI administration in chow-fed mice, but also show that some of these same genes are altered with Mboat7 knockdown and high fat diet feeding (Figure 8—figure supplement 2).

6) Does LPI treatment also affect other genes identified by RNA seq in figure 3? In other words, does LPI contribute to other phenotypes including steatosis and insulin resistance?

In parallel experiments described in question 5 above, we have injected LPI into control and Mboat7 ASO-treated mice maintained on a chow diet, and performed metabolic phenotyping including hepatic lipid quantification and unbiased RNA sequencing to address this question. In chow-fed mice we did not find that LPI injection promoted steatosis in either control or Mboat7 ASO-treated mice. We have included a new supplementary figure to show the modest transcriptional changes following 18:0 LPI administration in chow-fed mice, but also show that some of these same genes are altered with Mboat7 knockdown and high fat diet feeding (Figure 8—figure supplement 2).

Reviewer #2:

Major comments:

The main focus of this paper is that reduced expression of MBOAT7 in the liver drives NAFLD (per the title) and NAFLD consequences. In my view, it's a quite interesting paper but is fundamentally flawed, as the ASO targeting MBOAT7 in murine liver also reduces expression of TMC4, the other candidate gene. Given this, it's not possible to determine which effects are due to MBOAT7 depletion, which is the focus of the paper.

We have now clarified the role of Mboat7, but not Tmc4, in hepatic steatosis by generating Tmc4-/- mice. Importantly, global deletion of Tmc4 does not result in hepatic steatosis, but Mboat7 ASO treatment promotes striking steatosis when mice are fed a high fat diet (Figure 4).

Figure 1A-D need more controls (positive and negative). Units for x axes in D-F are not clear. Also, In the Discussion section the authors state a major finding is that hepatic expression of MBOAT7 is reduced in obese human and rodents independent of rs641738 status, so would be nice to see this data.

We apologize for this confusion. The units for panels D-F were relative quantitation of mRNA abundance from a previously performed microarray analysis (Affymetrix HT_MG-430A) on tissue isolated from the hybrid mouse diversity panel by the Lusis laboratory (Parks et al., 2015). As requested, we have now fixed these issues with this figure, and have included the rs641738 genotypes within each group as requested.

Figure 2A; pertinent negative controls should be shown. The ASO used targets liver, adipose and reticuloendothelial system. Figure 2A and Figure 2—figure supplement 1 only show mRNA data for liver, brain, spleen, heart, skeletal muscle. It would be good to include data in adipose. All this should be in main figures.

We have seen that our control ASO behaves very similar to a saline control group, and have now included a supplemental figure showing that two distinct ASO targeting different regions of the Mboat7 mRNA have similar effects on hepatic steatosis (Figure 2—figure supplement 4). To be more comprehensive in showing the reader where the ASOs target knockdown, we have now included figures showing the knockdown efficiency of Mboat7 ASO in liver (Figure 2), adipose (Figure 2—figure supplement 1), and other tissue including brain, spleen, heart, and skeletal muscle (Figure 2—figure supplement 2). It is important to note that despite significant knockdown of Mboat7 in Mboat7 ASO-treated mice, there are no differences in adiposity (Figure 2—figure supplement 1), basal and catecholamine-stimulated lipolysis (Figure 9c), energy expenditure (Figure 2—figure supplement 3), and only very minor differences in adipose tissue gene expression (Figure 2—figure supplement 1). Therefore, our data do not support a strong role for adipose tissue function per se in the observed fatty liver phenotype in Mboat7 ASO-treated mice.

Also, given the main thrust of the study is NAFLD/NASH, the authors should show collagen staining and fibrosis pathology scores.

As described above in response to reviewer # 1, we have tested a number of approaches to identify fibrosis in our mice, but under the high fat diet conditions we used we simply do not see fibrosis. In full disclosure we have made this clear to the reader. Since our original submission we have performed a number of experiments to test whether MBOAT7 knockdown indeed stimulates bridging fibrosis as is seen in advanced human liver disease, but as brought forth by this astute reviewer we failed to find pathological or biochemical support of fibrosis in this mouse model. First, we had a board-certified liver pathologist (Dr. Daniela Allende, Cleveland Clinic) blindly score H&E-, pirosirius Red, and trichrome-stained liver sections from HFD-fed control and Mboat7 ASO-treated mice, and in all cases there was no measureable fibrosis scoring above background. To indirectly quantitate fibrosis potential in a complimentary way we analyzed liver hydroxyproline content biochemically, and again found no indication that hydroxyproline levels were altered in Mboat7 ASO-treated mice. Finally, we performed some preliminary studies in primary stellate cells treating with MBOAT7 substrate lipids (LPIs) and did not see any obvious signs of pro-fibrotic responses in primary stellate cells. Given that we do not see any signs of even early stage fibrosis in Mboat7 ASO-treated mice, we have removed any discussion regarding effects on fibrosis and instead focus the discussion on the striking hepatic lipid accumulation seen with MBOAT7 loss of function in high fat fed mice. Although beyond the scope of the current study, we are now generating in parallel either hepatocyte, macrophage, or stellate cell specific Mboat7 knockout mice and will subject them to models of fibrosis induction (carbon tetrachloride, methionine/choline-deficient diet, bile duct ligation, etc.) to understand the cell autonomous effect of Mboat7 in pro-fibrotic responses.

To make this issue clear for the reader we have removed any claims that fibrosis was found in our studies, and have added the following sentences to the Discussion section:

“However, an important limitation of our studies is that high fat diet feeding in mice is not sufficient to drive bridging fibrosis. In future studies it will be important to examine how Mboat7 loss of function impacts the development of fibrosis in appropriate fibrosis-prone animal models, and whether Mboat7 expression in hepatic stellate cells plays a regulatory role the progression from NASH to cirrhosis.”

Figure 3E presents a major problem as the MBOAT7 ASO lowers mRNA for TMC4, making the rest of the experiments (most of the rest of the paper) not possible to interpret for MBOAT7 depletion. If the point of the paper is that both genes are coordinately regulated and somehow both function in these phenotypes, then more work needs to be done to understand this. However, the current paper is about MBOAT7.

We completely agree with this reviewer concern, and this was shared by several reviewers. Therefore, we felt it was imperative to address the issue with additional studies in Tmc4 deficient mice. We generated a global Tmc4 knockout mice using CRISPR-Cas9-mediated gene editing and examined hepatic lipid levels under high fat feeding conditions (Figure 4). Global Tmc4 knockout mice have marked reductions in hepatic Tmc4 mRNA (Figure 4A) and protein (Figure 4B), yet importantly Mboat7 expression is unaltered (Figure 4A). In contrast to the striking hepatic steatosis seen with Mboat7 loss of function (Figure 2E, Figure 2—figure supplement 4), Tmc4 null mice show similar levels of hepatic lipids when fed a high fat diet (Figure 4). These data strongly implicate Mboat7, but not Tmc4, as the primary mediator of hepatic steatosis seen with the rs641738 variant or in Mboat7 ASO-treated mice.

Figure 5; results for other phospholipids and lysophospholipids should be shown. Is this defect specific for PI/LPI?

To address this, we have quantified levels of most major species of phospholipids and lysophospholipids in the liver or Mboat7 ASO-treated mice, and have included these data in a new supplementary figure. Mboat7 knockdown did not significantly alter the hepatic levels of other major phospholipids including phosphatidylcholines, lysophosphatidylcholines, phosphatidylethanolamines, lysophosphatidylethanolamines, phosphatidylserines, phosphatidylglycerols, or phosphatidic acids (Figure 6—figure supplement 5). This is consistent with previously published findings from global Mboat7-/- mice (Lee et al., 2012). Collectively, our data suggest that ASO-mediated knockdown of Mboat7 primarily alters LPI and PI levels in the liver and white adipose tissue.

Figure 6; it's not clear exactly how LPI is given, how much, and whether it is "physiological".

Supplementary Figure 13 shows that with our dosing regime we were able to see a rapid transient ~2-fold increase in either 18:0 LPI or 18:1 LPI. We feel that this is well within the physiological range of these lipids, given that we have found human plasma levels to be much higher than what is found in mice (Figure 7).

Reviewer #3:

This is the first characterization of MBOAT7 knockdown mice. The characterization is good and complete with interesting findings. There are several important weaknesses- they don't actually establish whether MBOAT7 or TMC4 are independently or cooperatively involved in the resulting phenotype. I don't think with the experiments presented they have firmly established with the alterations in lipid species found are responsible for the inflammation and stellate cell activation. It is important they address my third major point prior to any publication as it has important implication if true. Finally, they don't address why TGs accumulate in liver. The latter seem pretty basic given this is how the region was identified in the first place. Overall, I think this is marginal for acceptance, but certainly more experiments would be required.

As described above in responses to reviewers 1 and 2, we have now added data to directly address the role of Mboat7, but not Tmc4 in hepatic steatosis. We also provide new data showing that the hepatic steatosis seen in Mboat7 ASO-treated mice is driven in part by reorganization of the proteome of cytosolic lipid droplets.

Major points:

1) The effect of the ASO on FAT Mboat7 expression should be included.

Given this is a shared concern by all 3 reviewers, we have now included several additional experiments examining the effects of Mboat7 ASO treatment on adipose tissue biology. Our revised manuscript shows that despite significant knockdown of Mboat7 in adipose tissue, there are no difference in adiposity (Figure 2—figure supplement 1), basal and catecholamine-stimulated lipolysis (Figure 9c), energy expenditure (Figure 2—figure supplement 3), and only very minor differences in adipose tissue gene expression (Figure 2—figure supplement 1). Therefore, our data do not support a strong role for adipose tissue function in the observed fatty liver phenotype in Mboat7 ASO-treated mice.

2) Regarding the experimental design of Figure 6. It is not clear whether the amount of 18:0 LPI injected resulted in a 2-fold increase in the total LPI concentration or in the amount of 18:0. The later would be appropriate for these experiments. It seems to directly link the LPI and the function of MBOAT7, only the species that were altered should induce the inflammation and stellate cell activation. There is a lack of controls in this important experiment, which should include species that were not altered by MBOAT7 knockdown.

This concern was shared by reviewer # 1, so we performed additional experiments and now show data with all three potential LPI substrates (16:0, 18:0, and 18:1). Since our original submission, we have now directly injected LPI species that are altered by Mboat7 ASO treatment (18:1 LPI and 16:0) and ones that are not altered by Mboat7 ASO treatment (18:0 LPI). It is important to consider that although only 16:0 and 18:1 LPI accumulate in the liver of Mboat7 ASO-treated mice (Figure 6D), we did find reduced hepatic LPIAT activity with all three potential substrates (Figure 6A). Furthermore, cytosolic lipid droplets isolated from Mboat7 ASO-treated mice showed marked accumulation of 18:0 LPI, but not 16:0 LPI or 18:1 LPI (Figure 9D). This is in stark contrast to what is seen in the whole liver, where Mboat7 knockdown instead promotes accumulation of 16:0 LPI and 18:1 LPI (Figure 6D). Functionally, we found that injection of either 18 carbon LPI (18:0 or 18:1), but not 16:0 LPI, was able to increase pro-inflammatory and pro-fibrotic gene expression, but only in the Mboat7 ASO-treated group. These new data are shown in main Figure 8 and Figure 8—figure supplement 1.

3) What is the mechanism by which the MBOAT7 knockdown results in TG, CE and FC accumulation in liver?

As described above in response to reviewer 1, we have performed a number of additional experiments to understand the mechanism(s) by which Mboat7 knockdown promotes hepatic steatosis. First, we have repeatedly seen that Mboat7 knockdown does not significantly alter Srebp1c or lipogenic gene expression, either under fed or fasted conditions (Figure 9A). We have also consistently seen no alterations in steady state plasma triglyceride levels (Figure 2—figure supplement 5) or VLDL triglyceride secretion rates (Figure 9B). After ruling out these common causes of hepatic steatosis, we instead focused our attention on whether MBOAT7 knockdown could drive steatosis by locally altering the lipid and proteome composition of cytosolic lipid droplets (i.e. impacting lipid storage at the lipid droplet interface). The rationale for pursuing this possibility was that a recent manuscript (Mancina et al., 2016) suggested that MBOAT7 can co-purify with cytosolic lipid droplets in subcellular fractionation experiments. Using sucrose gradient lipid droplet fractionation protocols, we have now confirmed that MBOAT7 can be found on cytosolic lipid droplets in control mice but is essentially absent on lipid droplets from Mboat7 ASO-treated mice (Figure 9F). In parallel with MBOAT7 localization to lipid droplets, we found the lipid droplets isolated from Mboat7 ASO-treated mice show striking increases in MBOAT7 substrates (18:0 LPI and 20:4 LPI) and reduction in MBOAT7 products (38:4 PI). In fact, the differences seen in LPI and PI lipids are much more apparent in isolated lipid droplets (Figure 9) than what is seen in whole liver homogenates (Figure 6). Collectively, our results support a role for MBOAT7 in regulating local LPI acylation at the lipid droplet surface, which is associated with marked accumulation of lipogenic enzymes (CCTα and GPAT4) on lipid droplets, which may contribute to the large lipid droplets seen Mboat7 ASO-treated mice. These new data are discussed in the revised manuscript as follows:

“From this study it is clear that MBOAT7 can diversify the inositol-containing phospholipids and the associated proteome on cytosolic lipid droplets, and this could in part explain the large lipid droplets that accumulate in Mboat7 ASO-treated mice. It is interesting to note that the well known PNPLA3 variant associated with fatty liver disease (I148M) accumulates on lipid droplets, and similarly reorganizes the lipidome and proteome to promote liver disease progression35-38. Our findings here with MBOAT7-driven restructuring of the lipid droplet surface, and those recently published with PNPLA335-38, suggest that alterations in lipid modifying enzyme access to the surface of cytosolic lipid droplets may be a common mechanism by which human fatty liver develops. Therefore, rigorous studies of the lipid droplet lipidome and proteome in freshly isolated human liver specimens may provide clues into new therapeutic leads for advanced liver diseases.”

[Editors' note: the author responses to the re-review follow.]

Summary:

The reviewers and editors agree that the manuscript has been improved since the original submission and several issues have been clarified. However, the mechanism underlying the steatosis is still unclear. The new data regarding alterations to the liver lipid droplet proteome are interesting but do not establish whether these changes are a cause or a consequence of steatosis. Given that steatosis is the focus of the paper, the consensus is that this issue must be addressed prior to publication.

To address the mechanism underlying steatosis associated with MBOAT7 loss of function we have generated MBOAT7 deficient human hepatoma cells and performed several new metabolic tracer studies to further clarify the mechanism. MBOAT7-deficient Huh7 cells showed no significant difference in triacylglycerol hydrolysis rates, but instead showed modest increases in de novo lipogenesis (14C-acetate incorporation into 14C-triacylglycerol) and marked reductions in fatty acid oxidation (14C-palmitate incorporation into 14C-CO2). These new data are included in new Figure 10, and provide new information by which MBOAT7 control fatty acid metabolism and triacylglycerol storage in a cell autonomous manner.

Essential revisions:

1) Direct evidence of how TG, CE and free cholesterol accumulates in Mboat7-deficient livers is not provided. A higher abundance of GPAT4 in the lipid droplet fraction was observed, but it is unclear where the acyl-CoA substrates for GPAT are coming from if there is no increase in de novo lipogenesis or uptake from circulation. The abundance of PLIN2, PLIN3, GPAT4, and CCTα on lipid droplets may merely reflect the presence of more lipid droplets. Increased abundance of GPAT4 and CCTα on lipid droplets does not explain why the livers of MBOAT7-deficient mice contain more CE and free cholesterol. mRNAs linked to lipogenesis were not measured in conditions that stimulate lipogenesis (i.e. re-feeding) but were instead measured after fasting or at the start of the light cycle.

The following experiments could be performed in the given time frame to address this central issue:

- Measure lipogenic mRNAs in re-fed conditions (see papers by Horton and colleagues on SREBP1-dependent lipogenesis).

When we received this original comment, we ascertained from our Institutional Animal Care and Use Committee (IACUC) that in fact this study would take us roughly 4 months to complete. For us to do the in vivo ASO studies looking at fasting/re-feeding along with radioisotope (14C-acetate and 3H-glycerol) we would need to amend our Institutional Animal Care and Use Committee (IACUC) protocol to include radiotracers that would take about 1.5 months before we could even get approval to start the experiments. In typical experiments we treat with ASO for 8 weeks to achieve good knockdown, which would put us another 2 months out. Therefore, any in vivo experiments would require about 4 months of time. To be responsive within the time frame, we instead focused our efforts on measuring rates of de novo lipogenesis rather than gene expression in hepatoma cell lines that we genetically deleted MBOAT7 (see comments below). Although the in vivo fasted and re-feeding experiments would provide another layer of information, we felt it is likely to be incremental in nature to the data we already show. It is our experience that if de novo lipogenic gene expression is different between groups of mice, this can be easily seen under the conditions we collected tissues (fasted vs. fed at the start of the light cycle).

- The rate of liver lipid synthesis can be assessed by i.p injection of radiolabeled glycerol and measurement of glycerol incorporation into lipid classes after separation by TLC as in Miyazaki et al., 2001 or Millar et al., 2006.

As discussed above, to be responsive within the time frame, we instead focused our efforts on measuring rates of de novo lipogenesis in hepatoma cell lines where we genetically deleted MBOAT7. To address this in cultured cells we generated Huh7 cells genetically lacking MBOAT7 using CRISPR-Cas9 gene editing. These cells accumulate large lipid droplets much like those seen in mouse liver in vivo, so we feel this is an appropriate model system to understand cell autonomous mechanisms by which MBOAT7 alters storage of triglyceride and cholesterol. MBOAT7-deficient Huh7 cells in fact showed modest increases in de novo lipogenesis (14C-acetate incorporation into 14Ctriacylglycerol) and marked reductions in fatty acid oxidation (14C-palmitate incorporation into 14C-CO2). These new data are included in new Figure 10, and provide new mechanistic information by which MBOAT7 control fatty acid metabolism and triacylglycerol storage in a cell autonomous manner.

- Alternatively, if they believe their data is consistent with an impairment in TG mobilization from LDs, then the rate of liver LD lipolysis can be assessed as in Schweiger et al., 2014. These measurements could be done in primary hepatocytes/cell culture models if that is easier to achieve.

As suggested, to test whether MBOAT7 loss-of-function is associated with impairment in triacylglycerol or cholesteryl ester mobilization, we performed pulse-chase studies in wild type or MBOAT7-deficient Huh7 cells using both 14C-oleic acid and 3H-cholesterol to label triacylglycerol and cholesteryl ester pools to steady state. Thereafter, cells were chased in the presence of the acyl-CoA synthetase inhibitor triacsin C (effectively blocking re-esterifcation of hydrolysis products). Using this method, we found that the turnover/mobilization of both triacylglycerol and cholesteryl esters was not altered in MBOAT7-deficient cells. These data are included in new Figure 10.

- de novo lipogenesis as a contributor to fat accumulation could be assessed by injecting mice with 14C-acetate (after they had been fasted O/N and re-fed for 6 hours) as in Lian et al., 2018.

Although we were unable to assess the effects of Mboat7 knockdown on de novo lipogenesis in vivo in the time frame, our studies clearly show that acetate incorporation into cellular triacylglycerol is increased in MBOAT7-deficient Huh7 cells compared to wild type Huh7 cells (see new Figure 10).

2) In the experiment in which LPI species were administered to mice, only mRNAs related to inflammation were assessed. Analysis of hepatic lipid metabolism should be included in this experiment, especially since the focus of the manuscript has shifted from fibrosis/inflammation to hepatic steatosis.

This is an excellent point, and we have now included data showing hepatic triacylglycerol levels in the LPI injection experiments in revised Figure 8I.

https://doi.org/10.7554/eLife.49882.037

Article and author information

Author details

  1. Robert N Helsley

    1. Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    2. Department of Internal Medicine, University of Cincinnati, Cincinnati, United States
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5000-3187
  2. Venkateshwari Varadharajan

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  3. Amanda L Brown

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Anthony D Gromovsky

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  5. Rebecca C Schugar

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  6. Iyappan Ramachandiran

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
  7. Kevin Fung

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
  8. Mohammad Nasser Kabbany

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  9. Rakhee Banerjee

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  10. Chase K Neumann

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  11. Chelsea Finney

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  12. Preeti Pathak

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  13. Danny Orabi

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  14. Lucas J Osborn

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  15. William Massey

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  16. Renliang Zhang

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  17. Anagha Kadam

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  18. Brian E Sansbury

    Center for Experimental Therapeutics & Reperfusion Injury, Department of Anesthesiology, Perioperative and Pain Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, United States
    Contribution
    Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  19. Calvin Pan

    1. Department of Medicine, University of California, Los Angeles, Los Angeles, United States
    2. Department of Microbiology, University of California, Los Angeles, Los Angeles, United States
    3. Department of Human Genetics, University of California, Los Angeles, Los Angeles, United States
    Contribution
    Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  20. Jessica Sacks

    Department of Pathobiology, Cleveland Clinic, Cleveland, United States
    Contribution
    Investigation, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  21. Richard G Lee

    Cardiovascular Group, Antisense Drug Discovery, Ionis Pharmaceuticals, Inc, Carlsbad, United States
    Contribution
    Resources, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    employee at Ionis Pharmaceuticals, Inc
  22. Rosanne M Crooke

    Cardiovascular Group, Antisense Drug Discovery, Ionis Pharmaceuticals, Inc, Carlsbad, United States
    Contribution
    Resources, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    employee at Ionis Pharmaceuticals, Inc
  23. Mark J Graham

    Cardiovascular Group, Antisense Drug Discovery, Ionis Pharmaceuticals, Inc, Carlsbad, United States
    Contribution
    Resources, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    employee at Ionis Pharmaceuticals, Inc
  24. Madeleine E Lemieux

    Bioinfo, Plantagenet, Canada
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared
  25. Valentin Gogonea

    Department of Chemistry, Cleveland State University, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  26. John P Kirwan

    Department of Pathobiology, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Methodology, Writing—original draft
    Competing interests
    No competing interests declared
  27. Daniela S Allende

    Department of Anatomical Pathology, Cleveland Clinic, Cleveland, United States
    Contribution
    Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  28. Mete Civelek

    Department of Biomedical Engineering, University of Virginia, Charlottesville, United States
    Contribution
    Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  29. Paul L Fox

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Supervision, Investigation, Methodology
    Competing interests
    No competing interests declared
  30. Lawrence L Rudel

    Department of Pathology, Section on Lipid Sciences, Wake Forest University School of Medicine, Winston-Salem, United States
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Writing—original draft
    Competing interests
    No competing interests declared
  31. Aldons J Lusis

    1. Department of Medicine, University of California, Los Angeles, Los Angeles, United States
    2. Department of Microbiology, University of California, Los Angeles, Los Angeles, United States
    3. Department of Human Genetics, University of California, Los Angeles, Los Angeles, United States
    Contribution
    Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  32. Matthew Spite

    Center for Experimental Therapeutics & Reperfusion Injury, Department of Anesthesiology, Perioperative and Pain Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, United States
    Contribution
    Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  33. J Mark Brown

    Department of Cardiovascular and Metabolic Sciences, Cleveland Clinic, Cleveland, United States
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    brownm5@ccf.org
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-2708-7487

Funding

National Heart, Lung, and Blood Institute (R01-HL122283)

  • Jonathan Mark Brown

National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK120679)

  • Jonathan Mark Brown

National Institute on Alcohol Abuse and Alcoholism (P50-AA-024333)

  • Daniela S Allende
  • Jonathan Mark Brown

National Heart, Lung, and Blood Institute (P01 HL029582)

  • Paul L Fox

National Institute on Alcohol Abuse and Alcoholism (U01-AA021893)

  • Daniela S Allende

National Institute of Diabetes and Digestive and Kidney Diseases (U01-DK061732)

  • Daniela S Allende

National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK103637)

  • Daniela S Allende

National Heart, Lung, and Blood Institute (P01-HL49373)

  • Lawrence L Rudel

National Heart, Lung, and Blood Institute (P01-HL30568)

  • Aldons J Lusis

National Heart, Lung, and Blood Institute (R00-HL12172)

  • Mete Civelek

National Heart, Lung, and Blood Institute (R01-HL106173)

  • Matthew Spite

National Heart, Lung, and Blood Institute (F32-HL136044)

  • Brian E Sansbury

All coauthors are responsible for the content of this work, and different aspects of this work was funded by the National Institutes of Health (NIH) and the American Heart Association (AHA).

Acknowledgements

This work is dedicated in loving memory to the late Dr. Lawrence ‘Larry’ Rudel, who passed away during the revision of this work (deceased, August 29th, 2019). Larry’s passion for science inspired many in the field of lipid and lipoprotein metabolism. We sincerely thank Dr. Hiroyuki Arai (University of Tokyo) for generously providing the MBOAT7 antibody used in this work. This work was supported by National Institutes of Health grants R01 HL122283 (JMB), R01 DK120679 (JMB), P50 AA024333 (JMB, DSA), P01 HL49373 (LLR), P01 HL30568 (AJL), U01 DK061732-15 (DSA), U01 AA021893-05 (DSA), P01 HL029582 (PLF), R00HL12172 (MC), R01 DK103637 (DSA), R01 HL106173 (MS), P01 GM095467 (MS), F32 HL136044 (BES), R21 AR067477 (JPK), and the American Heart Association (Postdoctoral Fellowships 17POST3285000 to RNH,15POST2535000 to RCS, and 19POST34380725 to IR). Development of lipid mass spectrometry methods reported here were supported by generous pilot grants from the Clinical and Translational Science Collaborative of Cleveland (4UL1TR000439) from the National Center for Advancing Translational Sciences (NCATS) component of NIH and the NIH Roadmap for Medical Research, the Case Comprehensive Cancer Center (P30 CA043703), the VeloSano Foundation, and a Cleveland Clinic Research Center of Excellence Award.

Ethics

Human subjects: Human MBOAT7 Expression Levels in Lean and Obese Subjects The majority of subjects recruited to examine MBOAT7 expression levels were morbidly obese bariatric surgery patients, but we were able to obtain liver biopsies from 10 subjects with a BMI under 30 as normal weight controls. For recruitment, adult patients undergoing gastric bypass surgery at Wake Forest School of Medicine were consented via written consent and enrolled by a member of the study staff following institutionally approved IRB protocols as previously described41. Exclusion criteria included: positive hepatitis C antibody, positive hepatitis B surface antigen, history of liver disease other than NAFLD, Childs A, B, or C cirrhosis, past or present diagnosis/treatment of malignancy other than non-melanocytic skin cancer, INR greater than 1.8 at baseline or need for chronic anticoagulation with warfarin or heparin products, use of immunomodulation for or history of inflammatory diseases including but not limited to malignancy, rheumatoid arthritis, psoriasis, lupus, sarcoidosis and inflammatory bowel disease, and greater or equal to 7 alcohol drinks per week or 3 alcoholic drinks in a given day each week. In addition to bariatric surgery patients, a small number of non-obese subjects (body mass index <30.0) consented to liver biopsy during elective gall bladder removal surgery (n=10). Each subject was assigned a unique identifier which was used throughout the study and did not include any identifiable information about the patient such as name, address, telephone number, social security number, medical record number or any of the identifiers outlined in the HIPAA Privacy Rule regulations. Only the principal investigator had access to the code linking the unique identifier to the study subject. Basic clinical information was obtained via self-reporting and a 15 ml baseline blood sample was obtained at the time of enrollment. A subset of this cohort has been previously described20,41. At the time of surgery, the surgeon collected a roughly 1-gram sample from the lateral left lobe. Wedge biopsies were rinsed with saline and immediately snap frozen in liquid nitrogen in the operating room before subsequent storage at -80°C. For data shown in Figure 7 showing levels of MBOAT7 substrate and product lipids, de-identified patient samples from the Cleveland Clinic hepatology clinic (IRB # 10-947) were analyzed. These patients had biopsy proven Ishak fibrosis scores42 of 0 (normal) or 4 (advanced fibrosis). For analysis of hepatic MBOAT7 expression, RNA isolated from liver biopsies were used for quantitative real time PCR (qPCR) as described below.

Animal experimentation: Rat Studies of Diet-Induced Obesity Sprague Dawley Rats were received at 12 weeks of age and were housed in individual cages, kept at a constant temperature and ambient humidity in a 12-hlight/dark cycle. Animals were then randomly assigned to either a standard chow diet or a high-fat diet (D12492, 60% fat, Research Diets, New Brunswick, NJ, USA) ad libitum to establish diet-induced obesity as previously described20. After 6 months of HFD-feeding, livers were excised for standard qPCR analysis of Mboat7 expression. Hybrid Mouse Diversity Panel 92 inbred strains of 8-week-old male mice (180 individual mice) were fed a high fat, high sucrose diet (D12266B, Research Diets, New Brunswick, NJ) for 8 weeks before tissue collection44. Gene expression of Mboat7 in white adipose tissue and liver were measured and correlated with obesity related traits using biweight midcorrelation analysis as previously described4. Mouse Studies of Mboat7 Loss of Function To explore the role of Mboat7 in diet-induced obesity, NAFLD progression, and insulin resistance, we utilized an in vivo knockdown approach in 8-week old adult mice. Selective knockdown of Mboat7 was accomplished using 2'-O-ethyl (cET) modified antisense oligonucleotides (ASO). All ASOs used in this work were synthesized, screened, and purified as described previously45 by Ionis Pharmaceuticals, Inc (Carlsbad, CA). For Mboat7 knockdown studies, adult (8 week old) male C57BL/6 mice were purchased from Jackson Labs (Bar Harbor, ME USA), and maintained on either a standard rodent chow diet or a high fat diet (HFD, D12492 from Research Diets Inc) and injected intraperitoneally biweekly with 12.5 mg/kg of either non-targeting control ASO or one of two independent ASOs directed against murine Mboat7 for a period of 20 weeks. Similar results were seen with two independent ASOs targeting different regions of the Mboat7 mRNA, hence key data using one Mboat7 ASO are shown. All rodents were maintained in an Association for the Assessment and Accreditation of Laboratory Animal Care, International-approved animal facility, and all experimental protocols were approved by the Institutional Animal Care and use Committee of the Cleveland Clinic (IACUC protocols # 2015-1519 and # 2018-2053).

Senior Editor

  1. Anna Akhmanova, Utrecht University, Netherlands

Reviewing Editor

  1. Peter Tontonoz, University of California, Los Angeles, United States

Reviewer

  1. Dennis Vance, U Alberta, Canada

Publication history

  1. Received: July 3, 2019
  2. Accepted: October 11, 2019
  3. Accepted Manuscript published: October 17, 2019 (version 1)
  4. Version of Record published: November 12, 2019 (version 2)

Copyright

© 2019, Helsley et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,179
    Page views
  • 266
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Human Biology and Medicine
    2. Microbiology and Infectious Disease
    Mark S Ladinsky et al.
    Research Article Updated
    1. Cell Biology
    2. Human Biology and Medicine
    Prashant Rajbhandari et al.
    Research Article Updated