Host defense against viral infection involves interferon mediated down-regulation of sterol biosynthesis

Abstract

Little is known about the protective role of inflammatory processes in modulating lipid metabolism in infection. Here we report an intimate link between the innate immune response to infection and regulation of the sterol metabolic network characterized by down-regulation of sterol biosynthesis by an interferon regulatory loop mechanism. In time-series experiments profiling genome-wide lipid-associated gene expression of macrophages, we show a selective and coordinated negative regulation of the complete sterol pathway upon viral infection or cytokine treatment with IFNγ or β but not TNF, IL1β, or IL6. Quantitative analysis at the protein level of selected sterol metabolic enzymes upon infection shows a similar level of suppression. Experimental testing of sterol metabolite levels using lipidomic-based measurements shows a reduction in metabolic output. On the basis of pharmacologic and RNAi inhibition of the sterol pathway we show augmented protection against viral infection, and in combination with metabolite rescue experiments, we identify the requirement of the mevalonate-isoprenoid branch of the sterol metabolic network in the protective response upon statin or IFNβ treatment. Conditioned media experiments from infected cells support an involvement of secreted type 1 interferon(s) to be sufficient for reducing the sterol pathway upon infection. Moreover, we show that infection of primary macrophages containing a genetic knockout of the major type I interferon, IFNβ, leads to only a partial suppression of the sterol pathway, while genetic knockout of the receptor for all type I interferon family members, ifnar1, or associated signaling component, tyk2, completely abolishes the reduction of the sterol biosynthetic activity upon infection. Levels of the proteolytically cleaved nuclear forms of SREBP2, a key transcriptional regulator of sterol biosynthesis, are reduced upon infection and IFNβ treatment at both the protein and de novo transcription level. The reduction in srebf2 gene transcription upon infection and IFN treatment is also found to be strictly dependent on ifnar1. Altogether these results show that type 1 IFN signaling is both necessary and sufficient for reducing the sterol metabolic network activity upon infection, thereby linking the regulation of the sterol pathway with interferon anti-viral defense responses. These findings bring a new link between sterol metabolism and interferon antiviral response and support the idea of using host metabolic modifiers of innate immunity as a potential antiviral strategy.

Figure 1. Regulation of the cholesterol pathway upon mCMV infection.

(A) The Sterol biosynthesis pathway shown in KEGG notation with abbreviated metabolites (abbreviations listed in Text S1). The geranylgeranylation pathway responsible for GGPP synthesis is shown in the dashed box. (B) Heat map of the cholesterol biosynthesis temporal genes' expression during the first 12 h of mCMV infection (left panel) or IFNγ treatment (right panel). Each time point corresponds to one independent biological sample, and columns indicate time in hours. Fold changes of expression levels are represented on a Log2 scale compared to mock-treated cells, ranging from a 0.8× lower expression (dark blue) to a 1.2× higher expression (bright yellow). (C–H) Expression analysis measured by qRT-PCR of Hmgcs1, Hmgcr, Idi1, and Sqle genes in BMDM infected with mCMV(24 hpi) (C) or treated for 24 h with IFNγ (10 and 100 U/ml) (D), IFNβ (10 and 25 U/ml) (E), IL6 (10 and 25 U/ml) (F), IL1β (10 and 100 U/ml) (G), or TNF (10 and 100 U/ml) (H). Graphs show levels of mRNA expression of the respective genes either infected or cytokines-treated relative to mock samples. Bars represent the means ± SD of five independent experiments with biological triplicates for each experiment. *p<0.05, **p<0.01, ***p<0.001, determined with an unpaired Student's t test.

Figure 2. Effect of a coordinated reduction in multiple enzymes on sterol biosynthesis.

(A) Comparison by Western blot analysis of HMGCS1, HMGCR, and SQLE protein levels in mCMV infected (24 hpi) or mock-treated BMDM. Infection was measured by detection of the IE1 mCMV antigen. Intensity values relative to β-actin calculated by densitometry show a decrease of the total amount of protein in the mCMV-infected BMDM compared to the mock-treated samples of 64% for HMGCS1, 50% for HMGCR, and 85% for SQLE. Graphs are representative of two independent experiments with biological duplicates and triplicates, respectively. (B–C) Free cholesterol concentration was determined experimentally by enzymatic assay (Materials and Methods) at 0, 6, 24, 48, and 72 hpi in BMDM (B) and NIH/3T3 cells (C). Cholesterol content is presented as the percentage of free intracellular cholesterol concentration from infected cells compared to mock treatment. Graphs represent the means ± SD of three independent experiments with biological quadruplicates for each experiment. (D) Free cholesterol concentration in BMDM cultures treated with varying concentrations of IFNγ, IFNβ, TNF, IL1β, or IL6. The cholesterol concentration was measured as mentioned above after 48 h post-cytokine treatment. Bars represent means ± SD of two independent experiments with biological quadruplicates for each experiment. *p<0.05, **p<0.01, ***p<0.001, determined with an unpaired Student's t test.

Figure 3. Schematic of the mevalonate-isoprenylation branch point of the sterol biosynthesis pathway.

Metabolites (shown in inverse print) and inhibitor (Simvastatin) (shown in grey) used to dissect the pathway are indicated: Simvastatin inhibits HMGCR and prevents the synthesis of mevalonate and downstream lipids.

Figure 4. Effect of statins on mCMV growth in vitro and in vivo.

(A) NIH/3T3 cells were infected with mCMV-GFP (MOI of 0.2) and subsequently treated with varying concentrations of Simvastatin or Gancyclovir immediately after infection. GFP expression was measured to monitor the level of infection (Materials and Methods). Graph represents the percentage of viral inhibition as a function of drug treatment. Data points represent mean ± SD of two independent experiments with six biological replicates for each experiment. (B) Mice were fed with simvastatin (50 mg/kg/mice) daily for 5 d by gavages and at day 1 post-treatment, were challenged with 2×10⁶ PFU of mCMV by intraperitoneal injection, and sacrificed; at 4 dpi, viral titers in different organs were measured by plaque assay and are expressed per gram of tissue. Data points represent mean ± SD of two independent experiments with five mice per group for each experiment. *p<0.05, **p<0.01, ***p<0.001, determined with a Mann-Whitney U test.

Figure 5. Metabolic investigation of the sterol pathway in infection.

(A) NIH/3T3 cells were infected with mCMV-GFP (MOI of 0.2) and subsequently treated with Simvastatin (SMV) (2.5 µM) and mevalonate (MEV) (300 µM) or geranylgeraniol (GGOH) (15 µM) or farnesol (FOH) (15 µM) or squalene (SQE) (15 µM) or of water soluble cholesterol (complexes of cholesterol with methyl-β-cyclodextrin, CHO/MCD) (5 µg/ml) for 72 h. The level of infection was determined by measuring GFP fluorescence at 76 hpi (Material and Methods). Graph represents the relative level of infection compared to the untreated cells, and bars represent mean values ± SD of three independent experiments with five biological replicates for each experiment. (B) NIH/3T3 cells were transfected for 48 h with either non-targeted, eGFP, Hmgcs1, Hmgcr, Sqle, Fdft1, or Dhcr7 On-target plus siRNA smart pool and then infected with mCMV-GFP (MOI of 0.2). (C) NIH/3T3 cells were transfected for 48 h with either Risc Free, M54, and M86 (knocking down mCMV viral genes), Hmgcr, Dhcr7, Fdps, Fntb, Pggt1b, or Rabggtb On-target plus siRNA smart pool, and then infected with mCMV-GFP (MOI of 0.2). The level of non-targeted siRNA (B) and Risc Free (C) treated cells was used as a baseline estimate for the cutoff point (two standard deviations and a p value <0.001 (determined with an unpaired Student's t test) as significant). Bars represent means ± SD of two independent experiments with three biological replicates for each experiment. (D) NIH/3T3 cells were incubated with various doses of IFNβ for 18 h in the presence or absence of 15 and 150 µM GGOH. The graph represents the inhibition of viral replication (in percentages) as a function of drug concentration. Bars represent mean ± SD of biological triplicates for each experiment. *p<0.05, **p<0.01, ***p<0.001, determined with an unpaired Student's t test.

Figure 6. Alteration of gene expression upon HSV1, SFV, VV, adenovirus, and non-infectious mCMV in primary macrophages.

(A) Heat map of expression levels of a set of genes after 24 h mock treatment, infection with Herpes simplex virus 1 (HSV1), Semliki forest virus (SFV), Vaccinia virus (VV), or Adenovirus (Ad) in BMDM (Text S1). Genes represent the innate immunity activation, the MHC class II antigen presentation, and the cholesterol and unsaturated fatty acids biosynthesis. Each square represents a single biological replicate. Fold changes of expression levels are represented on a Log2 scale compared to mock-treated cells, ranging from a 0.4× lower expression (dark blue) to a 1.6× higher expression (bright yellow). (B) Expression analysis measured by qRT-PCR of Hmgcs1, Hmgcr, Idi1, and Sqle genes in BMDM after 24 h mock treatment, mCMV, or mCMVdie3 infection, respectively. Graphs show the level of expression of the indicated genes relative to mock-treated samples and bars represent mean ± SD of two independent experiments with triplicate biological measurements for each experiment. (C) BMDM were infected with mCMV or mock treated, and supernatant was collected after 8 h and directly added to fresh BMDM. After 24 h, RNA from these cultures was collected and Hmgcs1, Hmgcr, and Sqle expression were measured by qRT-PCR. To test for the presence of any detectable virus, an aliquot of the supernatant was used to perform a standard plaque assay (no infectious virus detected, unpublished data). Graphs show the level of expression of the indicated genes relative to mock-treated samples and bars represent means ± SD of three independent experiments with triplicate biological measurements for each experiment. *p<0.05, **p<0.01, ***p<0.001, determined with an unpaired Student's t test.

Figure 7. Contribution of type I interferon response in the regulation of sterol biosynthesis genes upon infection.

(A–C) Wild type BMDM or BMDM from IFNβ−/− knockout mice or from IFNAR1−/− knockout mice were mock treated, infected with mCMV, or treated with IFNβ (10 U/ml) for 24 h. RNA was collected and the gene expression of Hmgcs1, Hmgcr, Idi1, and Sqle was measured by qRT-PCR. Graphs show the level of expression of the indicated genes relative to mock-treated samples. Bars represent the mean ± SD of biological quadruplicates. (D) Wild type BMDM or BMDM from IFNAR1−/− knockout mice were infected with mCMV or treated with IFNβ (10 U/ml). After 48 h, free cholesterol concentration was measured by enzymatic assay (Materials and Methods). Bars represent the mean ± SD of biological quadruplicates. *p<0.05, **p<0.01, ***p<0.001, determined with an unpaired Student's t test.

Figure 8. Measurement of de novo mRNA synthesis of sterol biosynthesis genes upon viral infection.

Wild Type or Tyk2^−/− BMDM were infected with mCMV at an MOI of 1 for 1 h. De novo RNA was labeled between 360 and 390 min post-infection, isolated, and hybridized to Affymetrix Gene 1.0 ST microarrays (Materials and Methods). After scanning and data capture, gene expression in mock-infected or infected cells was analyzed, and for the purposes of presentation, gene expression values from control (mock infected) BMDM (black) were adjusted to a value of 1. Values for expression in infected cells (white) were then expressed as a number relative to the control.

Figure 9. Regulation of SREBP2 by mCMV infection and IFNβ treatment.

(A) Comparison of cleaved SREBP2 protein in mock-infected (lane 1), mCMV-infected (MOI of 1) (lane 2), mock-treated (lane 3), IFNβ- (50 U/ml) (lane 4), or IFNγ- (50 U/ml) treated (lane 5) BMDM for 24 h by Western blot analysis using YY1 as a loading control. Arrow indicates SREBP2 cleaved form that is induced upon lovastatin and ezetimibe treatment from liver extracts of cholesterol-fed mice (see Figure S8). The blot is representative of two independent experiments with biological triplicates for each experiment. (B) Wild type BMDM were infected with mCMV for 1 h. De novo RNA was labeled between 360 and 390 min post-infection, isolated, and hybridized to Affymetrix Gene 1.0 ST microarrays (Materials and Methods). After scanning and data capture, gene expression in mock-infected or infected cells was analyzed, and for the purposes of presentation, Srebf2 gene expression values from control (mock-infected) BMDM (black) were adjusted to a value of 1. Values for expression in infected cells (white) were then expressed as a number relative to the control. (C) BMDM from wild type or IFNAR1−/− knockout mice were treated with 10 U/ml of IFNβ or infected with mCMV. After 24 h, RNA was collected and the gene expression of Srebf2 was measured by qRT-PCR. Results show the level of gene expression of the treated or infected samples relative to the mock-treated samples. Bars represent the mean ± SD of biological quadruplicates. *p<0.05, **p<0.01, ***p<0.001, determined with an unpaired Student's t test.

Figure 10. Proposed model for down-regulation of the sterol synthesis by type I interferon response to viral infection.