Generation and validation of a conditional knockout mouse model for desmosterolosis

Abstract The enzyme 3β-hydroxysterol-Δ24 reductase (DHCR24, EC 1.3.1.72) catalyzes the conversion of desmosterol to cholesterol and is obligatory for post-squalene cholesterol synthesis. Genetic loss of this enzyme results in desmosterolosis (MIM #602398), a rare disease that presents with multiple congenital anomalies, features of which overlap with subjects with the Smith-Lemli-Opitz syndrome (another post-squalene cholesterol disorder). Global knockout (KO) of Dhcr24 in mice recapitulates the biochemical phenotype, but pups die within 24 h from a lethal dermopathy, limiting its utility as a disease model. Here, we report a conditional KO mouse model (Dhcr24flx/flx) and validate it by generating a liver-specific KO (Dhcr24flx/flx,Alb-Cre). Dhcr24flx/flx,Alb-Cre mice showed normal growth and fertility, while accumulating significantly elevated levels of desmosterol in plasma and liver. Of interest, despite the loss of cholesterol synthesis in the liver, hepatic architecture, gene expression of sterol synthesis genes, and lipoprotein secretion appeared unchanged. The increased desmosterol content in bile and stool indicated a possible compensatory role of hepatobiliary secretion in maintaining sterol homeostasis. This mouse model should now allow for the study of the effects of postnatal loss of DHCR24, as well as role of tissue-specific loss of this enzyme during development and adulthood.

In cholesterol biosynthesis, the conversion of lanosterol to cholesterol has been described to occur via two main pathways, the Bloch pathway (1) and the Kandutsch-Russell pathway (2), which differ in whether reduction of the side chain C24-25 double bond occurs "first" or "last." In the Kandutsch-Russell pathway the C24-C25 reduction takes place in the first step, whereas in the Bloch pathway this reduction occurs in the last step of the synthesis. This reaction is catalyzed by 3βhydroxysterol Δ24-reductase (DHCR24, EC 1.3.1.72) and is required for cholesterol biosynthesis (3). Most commonly described as the enzyme catalyzing the conversion of desmosterol to cholesterol, DHCR24 can reduce the C24-25 double bond of any sterol intermediate, effectively switching sterol synthesis from the Bloch pathway to Kandutsch-Russell pathway. This hybrid or modified Kandutsch-Russell pathway is likely how many (if not all) mammalian cells synthesize cholesterol, rather than by using one pathway exclusively (4,5), and modifies the concept of first or last step. Alterations in DHCR24 activity lead to the accumulation of different biosynthetic intermediates (5), which have distinct biological activities and effects on downstream pathways (6). Variations in DHCR24 expression have been linked to different human diseases, including Alzheimer's disease, cardiovascular disease, hepatitis C virus infection, and prostate cancer (3), although none of these have been rigorously tested.
Homozygous or compound heterozygous defects in two terminal enzymes participating in cholesterol synthesis, DHCR7 or DHCR24, cause the dysmorphological conditions Smith-Lemli-Opitz syndrome (OMIM #270400) and desmosterolosis (OMIM #602398), respectively (7,8). Although endogenous cholesterol biosynthesis is lost, the affected individuals are still exposed to some cholesterol via maternal-fetal transfer in utero and subsequently from dietary sources after birth (9). Desmosterolosis is a very rare disease, with only nine published cases, and has an estimated incidence of <1:10,000,000, compared with Smith-Lemli-Opitz syndrome with an observed incidence of 1:20,000 to 60,000 (6,10,11). Patients with desmosterolosis manifest a spectrum of defects affecting craniofacial and neurological development, intellectual disability, and developmental delays, with the most severe cases dying shortly after birth (8,(12)(13)(14)(15)(16). Although the initial publication describing Dhcr24 knockout (KO) mice reported survival of KO into adulthood (17), this knockout line was freely shared with several investigators and all subsequent knockout progeny failed to survive beyond 24 h post birth as a result of a lethal dermopathy (18), a finding confirmed by the original investigators (19). Therefore, the absence of a viable animal model for desmosterolosis limits exploration of the underlying pathological mechanisms resulting from loss of DHCR24 and also prevents the exploration of mechanisms testing the effects of desmosterol in mammalian physiology. We sought to generate a viable mouse model with tissue-specific deletion of Dhcr24 to avoid the lethal dermopathy. In this study, we describe the characterization of a conditional KO of Dhcr24 using the Cre-loxP system and validated this model by creating a liver-specific loss of Dhcr24.

Generation of KO mice
All animal protocols were approved initially by the Medical College of Wisconsin, Milwaukee, WI, and the Clement J Zablocki Veterans Affairs (VA) IACUC, Milwaukee, WI, and subsequently by the University of Cincinnati IACUC, Cincinnati, OH. Live mice, harboring a floxed Dhcr24 allele (Dhcr24 tm1a(EUCOMM)Wtsi ), were imported from the Wellcome Trust Sanger Center, via EUCOMM (20)(21)(22)(23). Subsequent breeding and germline transmission were confirmed using PCR but showed that the LacZ-Neo cassette (Fig. 1A) was interfering with normal Dhcr24 expression as no homozygous mice were observed from heterozygous matings. The lacZ-Neo cassette (Fig. 1A) was removed by breeding with mice expressing FLP recombinase obtained from Jackson laboratories (B6.129S4-Gt(ROSA)26Sor tm1(FLP1)Dym /RainJ; #009086). Subsequent breeding resulted in a viable Dhcr24 flx/flx line (also referred to as CTL, for control), and these mice were normal and fertile, with no elevations of plasma desmosterol, compared with wild-type littermates. Mice heterozygous for the floxed allele were backcrossed for five generations to C57Bl/6J before crossing with Alb-cre mice. Liver-specific KO mice (LKO) were generated by mating with mice expressing Cre recombinase driven by albumin promoter obtained from Jackson laboratories [B6.Cg-Speer6-ps1 Tg(Alb-cre)21Mgn / J; #003574] and are referred to as Dhcr24 flx/flx,Alb-Cre (also referred to as LKO, for liver knockout). For one set of experiments, adult mice with postnatal global loss of Dhcr24 were generated by breeding Dhcr24 flx/flx mice with tamoxifeninducible ER-Cre strain [B6;129-Gt(ROSA)26Sor tm1(cre/ERT)Nat /J; #004847] and are referred to as Dhcr24 flx/flx,Er-Cre . Animal lines were maintained on a regular chow diet (Envigo 7912; Harlan Teklad, Madison, WI) and housed in individually ventilated PIV cages. To induce postnatal global loss of Dhcr24, Dhcr24 flx/flx,Er-Cre mice were intercrossed and progeny injected at 4-5 weeks of age with tamoxifen (T5648; Sigma-Aldrich, St. Louis, MO) dissolved in canola oil (10 mg/ ml) at 1 mg per mouse or vehicle (time 0) and monitored weekly.

Genotyping details
Genotypes of Dhcr24 flx/flx,Alb-Cre mice, Dhcr24 flx/flx,Er-Cre mice, and their respective controls were determined by evaluating homozygosity of flox alleles and the presence of Cre based on PCR of tail snip genomic DNA. The presence of the floxed allele was determined by amplification of a 652 bp Fig. 1. Verification of liver-specific deletion of Dhcr24 gene and growth curves in Dhcr24 flx/flx,Alb-Cre mice. A schematic representation of the targeting strategy used for generation of Dhcr24 conditional KO mice (Dhcr24 flx/flx ) and hepatocytespecific deletion (Dhcr24 flx/flx,Alb-Cre , A). PCR of genomic DNA isolated from liver (L) and brain (B) confirms hepatocytespecific deletion of exon 3 (B). The faint band corresponding to WT exon 3 in the LKO liver lane is likely from nonhepatocyte cells in the liver. Western blot analysis shows the bona fide loss of DHCR24 protein expression in LKO liver (C). The growth velocity of both male (M) and female (F) mice measured from 5 to 10 weeks (wk) of age was comparable between liver-specific knockout (Dhcr24 flx/flx,Alb-Cre , LKO) and control mice (Dhcr24 flx/flx , CTL) (D); for male LKO mice N = 13 in wk 5-8, N = 9 in wk 9, and N = 10 in wk 8; for male CTL mice N = 8 in wk 5, N = 12 in wk 6, N = 15 in wk 7-9, and N = 9 in wk 10; for female LKO mice, N = 17 in wk 5-8, N = 15 in wk 9, and N = 11 in wk 10; for female CTL mice, N = 9 in wk 5, N = 13 in wk 6, N = 18 in wk 7, N = 17 in wk 8-9, and N = 12 in wk 10. There were no differences in liver weights between LKO and CTL mice sacrificed at 10-12 weeks of age (E); N = 9 male CTL, N = 7 male LKO, N = 12 female CTL, and N = 10 female LKO mice. Males are represented with squares, and females with diamonds; closed symbols indicate CTL mice, and open symbols LKO mice. Bars denote mean ± 1 SD. fragment using primers Dhcr24-flp-for (5'-caaagcatacgaaagagcagcac-3') and Dhcr24-3'arm (5'-tcaagctcaggcaacacaggcagg-3'), compared with the wild-type allele of 442 bp fragment using these primers. The presence of Cre was detected through amplification of a 408 bp fragment using 5'-gcattaccggtcgatgcaacgagtgatgag-3' and 5'-gagtgaacgaacctggtcgaaatcagtgcg-3' primers. The presence or absence of exon 3 was detected with primer Dhcr24-flp-for (5'caaagcatacgaaagagcagcac-3') and primer Dhcr24 cre confirmrev (5'-agctcgtaggcagtgcaaat-3') PCR products (intact floxed allele product of 1,457 bp compared with Exon 3 deletion after cre-mediated recombination of 801 bp) of the genomic DNA isolated from tissues by running on a 1.5% agarose gel under standard conditions using a 100 bp DNA ladder for identification (see also supplemental Fig. S1). DNA was stained using SYBR Safe DNA gel stain (S33102, ThermoFisher Scientific, Waltham, MA) and scanned on a gel station (Universal Hood II, Bio-Rad). Protein expression was assessed by Western blot analysis of liver tissue lysate and probing with anti-DHCR24 Antibody (C59D8; Cell Signaling, Danvers, MA).

Plasma, bile, and stool collection
Four-week-old postweaned mice were recruited for weekly body weight measurements. Blood draws were from the submandibular vein. For bile measurements, mice were weighed and, under isoflurane anesthesia, the gallbladder was cannulated after the common bile duct had been ligated and timed bile flow under gravity was determined (24). We maintained a standard 4 h fast period while performing plasma and bile isolation experiments. For stool collection, mice were singly housed in clean cages and fecal pellets were collected and stored at −80 • C for subsequent assays.

Tissue harvesting
Mice 10-12 weeks of age were euthanized under CO 2 anesthesia; liver and brain tissues were collected and immediately flash frozen in liquid N 2 and stored at −80 • C for subsequent estimation of tissue sterols or RNA isolations. For histology, a small portion of liver from male CTL and LKO mice aged 12 weeks was rinsed twice in ice-cold phosphate buffered saline (PBS; pH7.4) and fixed overnight in 10% neutral buffered formalin (5725; Fisher Scientific, Pittsburgh, PA) followed by embedding in paraffin for hematoxylin and eosin staining or directly embedded in OCT compound for oil red O staining (21). Hepatic triglyceride secretion rates were determined, after 4 h of fasting and at ∼11 AM, by collection of blood at 0, 30, 60, 90, 120 min after intraperitoneal injection of poloxamer 407 (P-407, a.k.a. Pluronic® F127) at 1 g/kg dose (P2443; Sigma-Aldrich, St. Louis, MO) (25).

Sterols estimation using gas chromatography/mass spectrometry
Sterol analyses of plasma, bile, liver, brain, heart, and feces were measured using capillary column gas chromatographymass spectrometry (GC/MS) as described previously (21,24). Briefly, post hexane extraction and trimethylsilyl derivatization, test samples were injected into a 30 m low polarity phase crossbond diphenyl dimethyl polysiloxane Restek-5ms column (Restek Cat No. 13423) on a Thermo-Finnegan FOCUS GC/MS (ThermoFisher) measuring the intensities of the 329, 343, 393, 255, and 325 m/z peaks for cholesterol, desmosterol, lanosterol, lathosterol, and 7-DHC, respectively (26). The retention times of different sterols were verified by matching the retention times of the derivatives of the known standard compounds. A known amount of 5α-cholestane was added as an internal standard (24).

Fast-protein liquid chromatography
Lipoprotein separation for total cholesterol and triglycerides estimation in plasma samples was performed using fast-protein liquid chromatography (FPLC) as described elsewhere (27). Briefly, mice were fasted 4 h, and blood was collected in anticoagulant (EDTA)-coated tubes through the submandibular plexus. Samples were centrifuged at 2500 g for 10 min at 4 • C, and plasma was collected and the total cholesterol was measured (using an enzymatic kit) to ensure that pooling of samples used plasma with comparable sterols to exclude any outliers with any dramatic differences between individual mice (although no samples needed to be excluded). Column chromatography (10/300GL, Superose 6) was then used for separation of different lipoproteins by loading a minimum of 100 μl of pooled plasma mixed with 100 μl of PBS, from each group, onto the Akta pure FPLC, and a total of 52 fractions each with volume of 500 μl were collected and analyzed for cholesterol content using Infinity cholesterol (TR13421) and triglyceride (TR22421) calorimetric kits procured from ThermoFisher Scientific, Middletown, VA.

RNA isolation and qPCR assessments
Total RNA was isolated from liver and spleen tissues (n = 4 per group) by the column purification method using a Qiagen RNeasy® Kit (74104; Qiagen Inc., Germantown, MD) and was reverse transcribed to cDNA using the High-Capacity RNAto-cDNA™ Kit (4387406; ABI, Foster City, CA). Quantitative assessment of Nr1h3 and LXR target genes was performed on Applied Biosystems 7300 Real-Time PCR system (Applied Biosystems, Foster City, CA) using commercially available Taqman real-time PCR probes (ABI Biosystems). All reactions were performed in triplicate, and Atp5po was used as a housekeeping gene for data normalization to compensate the variations between input RNA amounts and analyzed using the comparative C T method (21).

RNA-Seq experimental design
RNA-Seq experiments were performed on total RNA isolated from liver and adrenal glands, using the total RNA isolated by the RNeasy Mini kit column method (74104; Qiagen, Valencia, CA). Total RNA from three female mice for each group was pooled together into a single test and control sample for library preparation followed by sequencing. RNA-Seq libraries were prepared using the NEBNext Ultra Directional RNA Library Prep Kit (New England BioLabs, Ipswich, MA) and sequenced using the TruSeq SBS kit on HiSeq platform (Illumina, San Diego, CA) at UC Genomics, Epigenomics and Sequencing Core. FASTQ files thus generated were used for bioinformatics analysis via the DNA Sequencing and Genotyping Core Facility of Cincinnati Children's Hospital Medical Center. In brief, quality control steps were performed on the FASTQ files to determine the overall quality of the reads using FASTQC. FASTQ files were then trimmed to remove adapter sequences and low-quality reads using Trimmomatic. The trimmed reads were then mapped to the mouse (mm10) reference genome using Hisat2. In the next step, transcript/gene abundance was assessed using Kallisto by creating a transcriptome index for Ensemble cDNA sequences from mouse (mm10). This index was used to quantify transcript abundance in raw counts and transcript per million. The raw count matrix from Kallisto was then used to identify differentially expressed genes between Dhcr24 flx/flx,Alb-Cre and Dhcr24 flx/flx using RUVSeq (R package). Differentially expressed genes showing statistical significance were defined using two filters: fold change cutoff of 2 and adjusted P-value/P-value cutoff of ≤ 0.05. Downstream functional annotation of significantly dysregulated genes was determined using gene ontology (cellular components, molecular function, and biological process) and pathway analysis. A detailed functional annotation and pathway analysis was performed using TOPPFUN part of the TOPPGENE suite (28).

Statistical analyses
RNA-Seq data were analyzed as described above. All other data are shown as mean ± SD. Data were assessed using Shapiro-Wilk's/Kolmogorov-Smirnov/D'Agostino and Pearson omnibus normality tests to determine whether it followed a Gaussian distribution prior to analyses. Statistical significance of differences between groups was evaluated by independent Student's t-test or Mann-Whitney U test or multiple t-test where applicable. A variation with P ≤ 0.05 was considered statistically significant.

Verification of liver-specific deletion of Dhcr24
The strategy employed for creating the liver-specific knockout mice of Dhcr24 gene is detailed in Fig. 1A. To verify that there was liver-specific deletion of Dhcr24, we performed PCR and Western blot analyses. PCR using genomic DNA isolated from liver of Dhcr24 flx/ flx,Alb-Cre mice (LKO, Fig. 1) showed a smaller product than that of liver of Dhcr24 flx/flx mice (CTL, Fig. 1), indicating deletion of the Dhcr24 alleles (Fig. 1B). However, PCR products of DNA from the brain of Dhcr24 flx/flx,Alb-Cre and Dhcr24 flx/flx mice were not different, thus confirming the liver-specific deletion of Dhcr24 (Fig. 1B). Western blots confirmed the expression of DHCR24 protein in tissue lysates from livers of Dhcr24 flx/flx mice but not in livers of Dhcr24 flx/flx,Alb/Cre mice (Fig. 1C). Dhcr24 flx/flx,Alb-Cre showed normal growth curves compared with Dhcr24 flx/flx mice (Fig. 1D) and exhibited normal terminal liver and spleen weights ( Fig. 1E and supplemental Fig. S2A).
Dhcr24 flx/flx, Alb-Cre mice have elevated liver and plasma levels of desmosterol There was no difference in the circulating levels of cholesterol ( Fig. 2A) between LKO mice and their sexmatched controls. Plasma desmosterol levels were considerably elevated in LKO mice, whereas desmosterol levels remained undetectable in control mice (Fig. 2B). Levels of hepatic cholesterol (Fig. 3A) were slightly decreased in female Dhcr24 flx/flx,Alb-Cre mice. The difference was small and was not significant in male mice. Levels of hepatic desmosterol (Fig. 3B) were very elevated, and lanosterol levels slightly elevated (Fig. 3C), in Dhcr24 flx/flx,Alb-Cre compared with Dhcr24 flx/flx mice.
There were no obvious differences in liver cell morphology or hepatic lipid accumulation (Fig. 3D, E).
Sterol levels in the brains of Dhcr24 flx/flx,Alb-Cre mice showed no significant differences in the levels of cholesterol (supplemental Fig. S3A), or desmosterol (supplemental Fig. S3B), when compared with Dhcr24 flx/flx mice, showing that plasma desmosterol did not readily cross the blood-brain barrier and accumulate in this organ.
Hepatobiliary excretion of desmosterol in Dhcr24 flx/flx,Alb-Cre mice There were no marked differences in the levels of biliary cholesterol, bile acids, and biliary phospholipids (Fig. 4A, D, E) between LKO mice and controls, suggesting only a minimal effect on bile synthesis (29). We hypothesized that hepatobiliary secretion may play a role in disposing of excess desmosterol to maintain sterol homeostasis in Dhcr24 flx/flx,Alb-Cre mice. Although desmosterol was not detectable in bile or stool from control mice, biliary (Fig. 4B) and stool desmosterol (Fig. 4C) levels were significantly increased in Dhcr24 flx/flx,Alb-Cre mice, suggesting a possible route for excess desmosterol excretion. There were no changes in the levels of stool cholesterol (supplemental Fig. S3C). Similarly, hepatic triglyceride secretion rates (Fig. 4F) and FPLC cholesterol (supplemental Fig. S4A) and triglyceride lipoprotein profiles (supplemental Fig.  S4B) were not significantly different between Dhcr24 flx/flx,Alb-Cre and Dhcr24 flx/flx mice.

Expression of Nr1h3 and LXR target genes
Desmosterol is a known activator of the Liver X receptor (LXR), encoded by Nr1h3, and has been reported to play a role in foam cell formation (6). Gene expression in liver and spleen were analyzed through qPCR as they express LXR and its target genes, Abca1, Abcg1, Fas, and Srebf1 (30). It is surprising that no consistent difference in the expression of these genes was found between KO and control mice livers (Fig. 5A, B) and spleens (Fig. 5C, D).

Gene ontology and pathway analysis
Given the changes in the sterol profile of Dhcr24 flx/flx,Alb-Cre mice, we sought to characterize the  , there was markedly increased excretion of desmosterol into bile and stool of LKO mice (panels B and C), while levels were undetectable in CTL mice. No differences were observed in biliary total bile acid or biliary phospholipid levels (panels D and E). Liver triglyceride secretion rates were not significantly different between LKO and CTL mice (panel F). For bile cholesterol and desmosterol measurements, N=5 male CTL, N=6 male LKO, N=9 female CTL, and N=7 female LKO mice. For bile total bile acid and phospholipid assays, N= 7 male CTL, N=7 male LKO, N=12 female CTL, and N=10 female LKO mice. Bile was collected from mice euthanized at age 10-12wk. For stool desmosterol measurements, N=4 for all groups. For poloxamer assay to assess triglyceride secretion, N=6 male CTL,N=5 male LKO, N=6 female CTL, and N=6 female LKO mice. Males are represented with squares, and females with diamonds; closed symbols indicate CTL mice, and open symbols LKO mice. Bars denote mean ± 1SD. Comparisons were performed only to sex-match controls.
Towards a cholesterol-free mouse broader effects on the differentially regulated transcriptome using RNA-Seq (Fig. 6). We identified 717 differentially expressed genes in the livers of female mice (Fig. 6A). Further analysis revealed association of the differentially expressed genes into 3 upregulated and 16 downregulated pathways. It is surprising that the pathways involved in steroidogenesis and steroid hormone metabolism were enriched by significantly downregulated genes (supplemental Table S1). Notably missing was any major perturbation in the cholesterol synthesis and metabolism pathways. To examine if elevated desmosterol in the blood affected adrenals (a major steroidogenic organ), we performed RNA-Seq on adrenals from Dhcr24 flx/flx,Alb-Cre mice and compared with controls (Dhcr24 flx/flx mice, Fig. 6B). Enrichment of differentially expressed genes in steroid hormone biosynthesis was also noted, but interestingly, the enrichment was through significantly upregulated genes instead of downregulation (supplemental Tables S1 and S2). Top hits of pathways enriched by significantly upregulated and downregulated genes are listed in supplemental Tables S1 and S2, and the complete dataset can be found in NCBI Gene Expression Omnibus (GEO) database (#GSE146524). Pathway analyses failed to show a common theme altered by loss of Dhcr24. A list of the changes in expression of key genes involved in lipid and cholesterol homeostasis is shown in Table 1. Differences were considered significant based on two criteria: logFC ≥ 1 and FDR ≤ 0.05, with up-or downregulated genes highlighted in bold.

Postnatal loss of Dhcr24 resulted in viable mice
Since embryonic global loss of Dhcr24 results in neonatal death of homozygous pups within 10 h of birth [(18) and unpublished observations], we investigated if postnatal global deletion of Dhcr24 would be viable using a tamoxifen-inducible cre line. Dhcr24 flx/flx,Er-Cre pups were viable after birth and had no obvious difference in appearance, mortality, or fertility (unpublished data). At 4-5 weeks of age, we injected tamoxifen or oil (vehicle) in matched littermates and tracked their weights and plasma sterol profiles (Fig. 7). Deletion of Dhcr24 was confirmed by PCR of genomic DNA isolated from liver (supplemental Fig. S1). There were no significant differences noted in the growth curves between vehicle or tamoxifen-treated mice (Fig. 7A). Liver (Fig. 7D) and spleen (supplemental Fig. S2B) weights also remained similar between vehicle and tamoxifen-treated mice. In tamoxifen-treated mice, there was an early decline and stable decrease in plasma cholesterol (Fig. 7B) with a dramatic and very large increase in plasma desmosterol (Fig. 7C). Liver, brain, and heart tissue cholesterol levels remained comparable between vehicle and tamoxifentreated mice (Fig. 8A-C) with significantly elevated levels of desmosterol in these tissues (Fig. 8D-F), as well as in bile (Fig. 7F). In contrast to Dhcr24 flx/flx,Alb-Cre mice, where levels of circulatory cholesterol were normal, Dhcr24 flx/flx,Er-Cre mice possessed markedly lower plasma (Fig. 7B) and biliary cholesterol (Fig. 7E) with a maximal reduction down to 50% of the levels found in controls. The mechanism for the decreased plasma cholesterol is unclear, as the cholesterol content of liver, brain, and heart were not markedly different between tamoxifen and vehicle mice (Fig. 8A-C).
Although the presence of desmosterol was not initially obvious, after we performed sterol analyses, we noted that some vehicle-injected mice had detectable desmosterol levels. Although very low, these levels are still comparatively high, as wild-type mice and Dhcr24 flx/flx mice had undetectable desmosterol levels under our assay conditions. At sacrifice, 8-9 weeks post tamoxifen injection, PCR analyses of both male and female liver samples confirmed that tamoxifeninduced tissues showed distinct Dhcr24 gene deletion (supplemental Fig. S1). Unfortunately, vehicle-injected mice showed variable levels of gene deletion (presence of both WT and KO bands) confirming leaky ER-Cre activity (supplemental Fig. S1). Although the vehicle-injected mice are therefore not ideal controls because of the endogenous leakiness of this particular ER-Cre driver (and we caution against its use), these data remain informative; tamoxifen-injected mice showed dramatic elevations in desmosterol yet continued to show no major phenotypic ill-health.

DISCUSSION
We describe here the creation of a Dhcr24 conditional knockout mouse model and verified its functionality by generating and characterizing an LKO. Although the first report of a global Dhcr24 knockout proposed that a "cholesterol-free" mouse was possible (17), subsequent characterizations of this line with further breeding showed no survival of global Dhcr24 KO pups beyond 24 h and lethality was likely caused by a lethal dermopathy with transepithelial water loss (18). Our initial attempts reproduced the perinatal lethality in Dhcr24 homozygous null mice (unpublished observations). This prompted us to pursue conditional KO models in order to explore the role of Dhcr24 in embryonic development. Dhcr24 flx/flx mice were developed and validated using cre-mediated deletion in liver, as well as postnatal global loss using tamoxifen-inducible ER-Cre. Both the liver-specific and the postnatal global loss of Dhcr24 mice were remarkably normal and showed comparable growth curves and preserved fertility and manifested no overt phenotype. This suggests that almost all the phenotypes observed with humans with desmosterolosis may be as a result of embryonic loss of DHCR24 activity and that loss of DHCR24 postnatally may not be as detrimental; loss of Dhcr24 in mice using a globally expressed ER-cre driver resulted in the absence of endogenous cholesterol synthesis and a dramatic elevation in desmosterol levels but was well tolerated.
Cholesterol homeostasis in mammals is a complex and tightly regulated process that involves a proper balance between its endogenous synthesis, cellular uptake, and efflux. Inherited post-squalene cholesterol biosynthetic pathway disorders show distinct patterns of precursor accumulation and variable tissue cholesterol reductions, depending on the severity of  enzymatic impairments (31,32). Therefore, it is difficult to determine whether excessive accumulation of sterol precursors or a decrease in cellular cholesterol or some combination of these is responsible for the pathogenesis of disease phenotypes (31,32). The liver was chosen as a tissue of interest for targeted deletion of Dhcr24 owing to its primary role in the maintenance of systemic cholesterol homeostasis (33). An accumulation of excess desmosterol in the livers but not in the brains of Dhcr24 flx/flx,Alb-Cre mice confirms the liver-specific knockout of Dhcr24. The maintenance of elevated levels of plasma desmosterol in Dhcr24 flx/flx,Alb-Cre mice indicated that there was secretion of hepatic desmosterol into the circulation (34), and the increased desmosterol in bile and stool of Dhcr24 flx/flx,Alb-Cre male mice suggested that desmosterol can use the hepatobiliary system for excretion from the body. Nevertheless, the ability of Dhcr24 flx/flx,Alb-Cre mice to have a mostly normal phenotype despite accumulation of circulatory and tissue desmosterol in the liver suggests that elevated desmosterol and a loss of DHCR24 enzyme activity are well tolerated. Elevated levels of lanosterol in the liver, when compared with control mice, indicated limited involvement of the Kandutsch-Russell pathway in the liver and predominant use of the Bloch pathway. This is expected as DHCR24 catalyzes the first step of entry into the Kandutsch-Russell pathway and the last step of the Bloch pathway. Of interest, these results suggest also that control animals had minimal or very efficient utilization of the Kandutsch-Russell pathway, since there was little accumulation of intermediates from this pathway, although more sensitive techniques (such as LC-MS/MS) to detect subtle elevations were not employed. Lanosterol, which is an early intermediate in post-squalene biosynthesis, was upregulated in both male and female Dhcr24 flx/flx,Alb-Cre mice, although these mice differed in liver cholesterol levels, suggesting additional sex-specific regulation of sterol homeostasis.
Liver X receptors (LXRs) are nuclear proteins and a member of receptor family transcription factors that are reported to upregulate genes involved in cholesterol efflux, thereby participating in cholesterol homeostasis (35). Two important LXR-regulated genes involved in cholesterol efflux are ATP-binding cassette transporter-A1 (Abca1) and ATP Binding Cassette Subfamily G Member 1 (Abcg1) (35). Earlier studies have suggested that desmosterol, zymosterol, and various oxysterols act as endogenous ligands for activation of these nuclear proteins (LXR) and through this mechanism participate in glucose and lipid homeostasis, steroidogenesis, and immunity (30). However, our data did not show significant upregulation of Nr1h3 and LXR target gene expression in livers of Dhcr24 flx/flx,Alb-Cre mice despite elevated circulatory and tissue levels of desmosterol. This is consistent with a previously published study by Muse et al. (36) that found that desmosterol had almost no activity in hepatocytes. , and hearts (C and F) of Dhcr24 flx/flx,Er-Cre mice harvested 8-9 weeks after injection with tamoxifen (Tam) or vehicle (Veh). As expected, after tamoxifen administration, desmosterol levels were dramatically increased in all three tissues (D-F). It is surprising that, despite the reduction in circulatory cholesterol, there was no reduction in the cholesterol content of tamoxifen-treated liver (A), brain (B), or heart (C). For liver measurements, N = 11 male Veh, N = 10 male Tam, N = 8 female Veh, and N = 9 female Tam mice; for brain measurements, N = 9 male Veh, N = 10 male Tam, N = 8 female Veh, and N = 9 female Tam mice. For heart measurements, N = 5 male Veh, N = 4 male Tam, N = 7 female Veh, and N = 5 female Tam mice. Males are represented with squares, and females with diamonds; closed symbols indicate CTL mice, and open symbols LKO mice. Bars denote mean ± 1 SD. Statistical testing was performed only to sex-matched controls, *P < 0.05.
Notably missing from the list of differentially expressed genes in RNA-Seq analysis were genes involved in cholesterol biosynthesis. It seems that the loss of hepatic Dhcr24 did not result in upregulation of the cholesterol biosynthesis pathway. The accumulation of desmosterol may compensate for the loss of cholesterol, resulting in no detected change in the total sterol content sensed by the hepatocyte, or the liver may receive so much cholesterol from the periphery that unless that is altered, hepatic sterol balance may not be impacted by loss of DHCR24. Postnatal global deletion of Dhcr24 showed that plasma cholesterol fell significantly, while total plasma sterol levels seemed unaltered, as the decrease in cholesterol was matched by the significant rise in desmosterol.
Finally, we also asked if a global deletion of Dhcr24 was lethal. Although vehicle-injected Dhcr24 flx/flx,Er-Cre mice showed leaky expression of Cre recombinase, the tamoxifen-treated mice are nevertheless informative; almost complete whole-body loss of cholesterol synthesis is compatible with life; growth parameters seem relatively normal and the accumulation of pharmacological amounts of desmosterol in the blood (∼60 mg/dl) seems to be well tolerated. This result is expected given that there are some surviving adult patients with desmosterolosis and also because whole-body inhibition of cholesterol synthesis using Triparanol had been used on humans as a treatment of elevated cholesterol in the last century (37,38). Future studies, using a more specific ER-cre, are planned to examine if there is a long-term phenotype, especially on nervous system function as the bloodbrain barrier is a limitation for dietary cholesterol entry.
In conclusion, we demonstrated that Dhcr24 is not necessary for survival after fetal development. Insights derived from the pathways identified by RNA-Seq analysis suggest many potential mechanisms by which desmosterol accumulation may be clinically important. However, expression of genes involved in cholesterol synthesis and metabolism seem not to be affected by the deletion of Dhcr24, although direct measurements of sterol synthesis rates were not performed. Dhcr24 flx/flx mice can serve as a valuable tool to investigate the role of post-squalene synthesis in development, as well as study the long-term effects of Dhcr24 deletion on mammalian physiology.

Data availability
The RNA-Seq data have been deposited with the NCBI Gene Expression Omnibus (GEO) database (#GSE146524), and all primary raw data are available from the corresponding author.