Abstract Objective Dysregulation of hepatic cholesterol metabolism can contribute to elevated circulating cholesterol levels, which is a significant risk factor for cardiovascular disease. Cholesterol homeostasis in mammalian cells is tightly regulated by an integrated network of transcriptional and post-transcriptional signalling pathways. Whilst prior studies have identified many of the central regulators of these pathways, the extended supporting networks remain to be fully elucidated. Methods Here, we leveraged an integrated discovery platform, combining multi-omics data from 107 strains of mice to investigate these supporting networks. We identified retinol dehydrogenase 11 (RDH11; also known as SCALD) as a novel protein associated with cholesterol metabolism. Prior studies have suggested that RDH11 may be regulated by alterations in cellular cholesterol status, but its specific roles in this pathway are mostly unknown. Results Here, we show that mice fed a Western diet (high fat, high cholesterol) exhibited a significant reduction in hepatic Rdh11 mRNA expression. Conversely, mice treated with a statin (3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGCR) inhibitor) exhibited a 2-fold increase in hepatic Rdh11 mRNA expression. Studies in human and mouse hepatocytes demonstrated that RDH11 expression was regulated by altered cellular cholesterol conditions in a manner consistent with SREBP2 target genes HMGCR and LDLR. Modulation of RDH11 in vitro and in vivo demonstrated modulation of pathways associated with cholesterol metabolism, inflammation and cellular stress. Finally, RDH11 silencing in mouse liver was associated with a reduction in hepatic cardiolipin abundance and a concomitant reduction in the abundance of proteins of the mitochondrial electron transport chain. Conclusion Taken together, these findings suggest that RDH11 likely plays a role in protecting cells against the cellular toxicity that can arise as a by-product of endogenous cellular cholesterol synthesis. Keywords: Cholesterol metabolism, Lipid metabolism, Oxidative stress, Inflammation, Lipidomics, Mitochondrial biology, Systems genetics Highlights * • Use a unique and powerful systems genetics platform to identify novel genes associated with hepatic cholesterol metabolism. * • These approaches identify RDH11 as a novel gene highly correlated with the cholesterol biosynthesis pathway. * • In vitro and in vivo manipulation of RDH11 demonstrated roles for this enzyme in cholesterol abundance and cellular stress pathways. * • Loss of RDH11 in mice led to alteration in oxidative stress and changes to mitochondrial readouts. * • These data suggest RDH11 likely plays a role in protecting cells against the cellular toxicity associated with endogenous cellular cholesterol biosynthesis. 1. Introduction Cholesterol homeostasis is critical for the normal functioning of mammalian cells. Specifically, cholesterols contribute to the maintenance of plasma membrane structure, act as signalling molecules and as precursors for numerous metabolites including steroid hormones [[45]1]. In mammals, the primary site of endogenous cholesterol synthesis is the liver. This process is principally driven by two transcription factors, sterol response element binding protein (SREBP) 2 and liver X receptor (LXR), which act in an opposing manner to promote cholesterol synthesis and uptake, and cholesterol efflux, respectively [[46]2,[47]3]. Dysregulation of hepatic cholesterol homeostasis can precipitate elevated serum cholesterol levels, known as hypercholesterolemia, which is a significant risk factor for cardiovascular diseases (CVD) [[48]4]. The strong association between hypercholesterolemia and CVD has fuelled interest in understanding the regulatory pathways that contribute to cholesterol homeostasis, with the hope of identifying novel targets for therapeutic intervention. One approach to identify novel drivers of complex traits such as hypercholesterolemia is to utilise genetic reference panels (GRPs). GRPs exploit model organisms such as mice, fruit flies or roundworms, and investigate how genetic variance influences phenotypic outcomes, a process analogous to human genome wide association studies. GRPs have been utilised to identify numerous novel loci implicated in conditions associated with lipid dysregulation, including obesity [[49]5]. Indeed, our group has previously demonstrated the utility of a mouse GRP known as the Hybrid Mouse Diversity Panel (HMDP) to map genes important in the regulation of lipid metabolism [[50]6]. Specifically, by integrating plasma and liver proteomic and lipidomic analyses, we could successfully identify both known and novel regulators of hepatic lipid metabolism [[51]6]. Here, we utilise these previously generated datasets to identify novel proteins associated with cholesterol metabolism, including retinol dehydrogenase 11 (RDH11). RDH11 is a microsomal short chain dehydrogenase/reductase that has oxido-reductive affinity for all trans and cis retinols, retinals and aldehyde species [[52]7]. Initial RDH11 characterisation studies demonstrated that it was highly expressed in the human prostate epithelium [[53]8]. These studies also identified RDH11 to be responsive to synthetic androgens. Further studies have suggested that mouse RDH11 could react with toxic C[2]–C[12] fatty aldehyde species, including 4-hydroxy-2-nonenal, to produce non-toxic fatty alcohol species [[54]9]. More recently, studies in retinal pigment epithelial cells have implicated RDH11 in maintaining healthy visual function, acting to replenish all-trans and 11-cis aldehyde isomers [[55]10]. Brown and Goldstein have previously suggested that RDH11 expression is induced by SREBPs [[56]9]. Similarly, expression array analysis utilising human immortalised lymphoblastoid cell lines exposed to simvastatin demonstrated that RDH11 correlated with 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (HMGCR) [[57]11], the rate-limiting enzyme in the cholesterol biosynthesis pathway. Moreover, analysis of multi-species transcriptomic data by Gene-Module Association Determination, and Module-Module Association Determination, suggested that the function of RDH11 may relate to cholesterol metabolism and be conserved across humans and rodents [[58]12]. Very recently, RDH11 was further identified through a genetic screen of 35 strains of mice to be co-correlated with cholesterol biosynthetic genes [[59]13], and subsequently suggested in that study to be involved in the modulation of endogenous cholesterol levels in cultured cells. Here, we examined the role of RDH11 both in vitro and in vivo in the context of cholesterol metabolism, to build on its recognised functions as well as its potential roles in inflammatory and oxidative pathways. 2. Results 2.1. Identification of RDH11 utilising the hybrid mouse diversity panel To identify potential novel regulators of hepatic cholesterol metabolism we examined a multi-omics dataset derived from the hybrid mouse diversity panel (HMDP) [[60]6]. Briefly, 107 strains of inbred mice from the HMDP (n = 2–3 mice per strain; 307 mice in total) were fed a chow diet for approximately 8 weeks. Mice were fasted overnight, then blood and tissue collected for proteomic and lipidomic phenotyping ([61]Figure 1A). Specifically, liver proteomic analysis detected 8370 proteins, of which 4311 proteins were present across at least 50 strains. Lipidomic analysis of the liver and plasma quantified 313 and 307 lipid species, respectively. From these data we constructed protein–protein co-correlations across the HMDP using known regulators of cholesterol metabolism. This included NAD(P) Dependent Steroid Dehydrogenase-Like (NSDHL), Acyl-coenzyme A synthetase short-chain family member 2 (ACSS2), Farnesyl pyrophosphate synthase (FDPS), Cytochrome P450 Family 51 Subfamily A Member 1 (Cyp51a1), Transmembrane 7 Superfamily Member 2 (TM7SF2), Isopentenyl-Diphosphate Delta Isomerase 1 (IDI1) and 24-Dehydrocholesterol Reductase (DHCR24). This approach was performed on the premise that proteins with related functions are likely to be co-regulated, and thus high concordance in their correlation structure will provide an informative insight into proteins that participate in related pathways, and indeed, we showed this to be true for many pathways in our previous study [[62]6]. This analysis identified both known and novel proteins that co-correlated with the aforementioned established regulators of cholesterol metabolism ([63]Figure 1B–D; [64]Supp Figure 1A-D). RDH11 was amongst the top 25 statistically significant proteins that correlated with all seven cholesterol biosynthetic proteins for which protein–protein correlation analyses were performed. We also performed protein-lipid correlation analyses utilising liver and plasma lipidomic analyses. Despite the co-correlation with known regulators of cholesterol metabolism, analyses revealed that RDH11 did not significantly correlate with hepatic or plasma total cholesterol ester (CE) or hepatic free cholesterol (COH) abundance across the HMDP ([65]Supp Figure 1E-G; [66]Supp Tables 2 and 3), however, there was a significant correlation with plasma COH levels (p < 0.05; [67]Supp Figure 1H). Figure 1. [68]Figure 1 [69]Open in a new tab Co-correlation analysis of liver proteomics data from the hybrid mouse diversity panel (HMDP) identified retinol dehydrogenase 11 (RDH11) as a potential novel regulator of cholesterol metabolism (A) Schema of the Hybrid Mouse Diversity Panel (HMDP) discovery platform published in Parker et al. [[70]6]; (B–D) Linear corrected hepatic protein co-correlation networks of top 25 proteins significantly associated with known regulators of cholesterol metabolism (white nodes; centre) including (B) NAD(P) dependent steroid dehydrogenase-like (NSDHL), (C) acyl-coA synthetase short-chain family member 2 (ACSS2) and (D) farnesyl diphosphate synthase (FDPS) across at least 50 strains of mice from the HMDP, with RDH11 highlighted in black oval; (E) Linear corrected hepatic protein co-correlation network of top 50 proteins significantly associated with RDH11 across at least 50 strains of mice from the HMDP. B-E, line width (black > grey) is indicative of statistical significance following Benjamini Hochberg correction; Colour of node indicates bicorrelation (bicor) direction and strength as indicated in key. (F) Pathway enrichment analysis of the 422 proteins significantly correlated with RDH11 across the HMDP following Benjamini Hochberg correction. (For interpretation of the references to