Hypoxia-altered cholesterol homeostasis enhances the expression of interferon-stimulated genes upon SARS-CoV-2 infections in monocytes

Abstract

Hypoxia contributes to numerous pathophysiological conditions including inflammation-associated diseases. We characterized the impact of hypoxia on the immunometabolic cross-talk between cholesterol and interferon (IFN) responses. Specifically, hypoxia reduced cholesterol biosynthesis flux and provoked a compensatory activation of sterol regulatory element-binding protein 2 (SREBP2) in monocytes. Concomitantly, a broad range of interferon-stimulated genes (ISGs) increased under hypoxia in the absence of an inflammatory stimulus. While changes in cholesterol biosynthesis intermediates and SREBP2 activity did not contribute to hypoxic ISG induction, intracellular cholesterol distribution appeared critical to enhance hypoxic expression of chemokine ISGs. Importantly, hypoxia further boosted chemokine ISG expression in monocytes upon infection with severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2). Mechanistically, hypoxia sensitized toll-like receptor 4 (TLR4) signaling to activation by SARS-CoV-2 spike protein, which emerged as a major signaling hub to enhance chemokine ISG induction following SARS-CoV-2 infection of hypoxic monocytes. These data depict a hypoxia-regulated immunometabolic mechanism with implications for the development of systemic inflammatory responses in severe cases of coronavirus disease-2019 (COVID-19).

Keywords: COVID-19; SREBP2; cholesterol; hypoxia; immunometabolism; systemic inflammation.

Figure 1

Hypoxia induces cholesterol biosynthesis enzymes. (A) RNA expression in THP-1 cells exposed to hypoxia (H; 1% O₂) for 8 h (acute hypoxia = AH) or 72 h (chronic hypoxia = CH) was determined relative to normoxic THP-1 cells (N; 21% O₂) by RNA-Seq (n = 3). Differentially expressed genes (padj < 0.05, |log2FC| > 0.3) were visualized in a heatmap (z-score normalized counts) and categorized by k-means clustering. Cholesterol biosynthesis genes are highlighted. (B) Top two functional annotation clusters of down- or upregulated transcripts identified by DAVID (26, 27). (C) Schematic representation of the cholesterol biosynthesis cascade. Oxygen-demanding steps are highlighted by red arrows and selected inhibitors as well as their target enzymes are marked in red. (D, E) THP-1 cells were incubated under N (grey) or H (red) for the indicated times (n = 4). (D) LSS and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. (E) LSS protein expression was determined by Western blot analysis. β-tubulin served as loading control. The blot is representative of four independent experiments. All data are means ± SEM and were statistically analyzed using one-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (*p < 0.05, **p < 0.01).

Figure 2

Hypoxia increases interferon (IFN) signaling. (A) IFN-stimulated genes (ISGs) within the transcripts constituting the functional annotation cluster “immune cell activation” were identified using Interferome v2.01 (28). Venn diagram depicts proposed ISGs regulated by type I, II, and/or III IFNs according to Interferome v2.01. (B) Heatmap representing differentially expressed, hypoxia-induced ISGs (z-score normalized count data) under normoxia (N), acute hypoxia (AH; 8 h), and chronic hypoxia (CH; 72h) (n = 3; padj < 0.05, |log2FC| > 0.3). Selected ISGs are highlighted. (C) THP-1 cells were incubated under N (grey) or H (red) for the indicated times (n = 4) or (D) treated with 5 µg/mL α-IFNAR2 antibody or an IgG2a-isotype control and incubated under N (grey) or H (red) for 24 h (n = 3). OAS1, IRF7, IFI6, and IFNB1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. All data are means ± SEM and were statistically analyzed by one-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 3

Hypoxic ISG induction is not directly affected by cholesterol biosynthesis intermediates. (A) THP-1 cells were incubated under N (grey) or H (red) for the indicated times (n = 5). Sterol levels were measured by GC-MS. (B, C) THP-1 cells were pre-incubated for 1 h with 1 µM simvastatin, 10 µM NB-598, 10 µM ketoconazole, or DMSO, prior to incubation under N (grey) or H (red) for 24 h (n = 5). (B) Cellular sterol levels were measured by GC-MS. (C) OAS1, IRF7, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. (D, E) THP-1 cells were pre-incubated for 1 h with (D) the early cholesterol precursors mevalonate (300 µM) or geranylgeraniol (15 µM) (n = 3), (E) the late cholesterol precursors lathosterol, 7-dehydrocholesterol, or desmosterol (each 5 µM) (n = 3), or appropriate solvent controls, prior to incubation under N (grey) or H (red) for 24 h. OAS1, IRF7, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. All data are means ± SEM and were statistically analyzed using one-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (A), or two-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (B–E) (*p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ##p < 0.01, ###p < 0.001 (compared to respective solvent controls)).

Figure 4

Altered cholesterol homeostasis affects hypoxic ISG induction. (A) Overview of the mechanisms of activation of the cholesterol-associated transcription factors LXR and SREBP2. The LXR agonist T0901317 as well as the SREBP2 inhibitors PF-429242 and fatostatin are highlighted in red. (B) THP-1 cells were incubated under N (grey) or H (red) for 8 h (n = 3). Immunofluorescence staining for SREBP2 (green) was performed. Nuclei were counterstained with DAPI (blue). Mean fluorescence intensity was quantified for nuclear SREBP2. Images are representative of three independent experiments. (C–E) THP-1 cells were pre-incubated for 1 h with (C) 1 µM PF-429242, (D) 10 µM fatostatin, (E) 1 µM T0901317, or appropriate solvent controls, prior to incubation under N (grey) or H (red) for 24 h (n = 3). OAS1, IRF7, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. Data are means ± SEM and were statistically analyzed using two-tailed paired t-test (B), or two-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (C–E) (*p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ###p < 0.001 (compared to ctrl or DMSO, respectively)).

Figure 5

Intracellular cholesterol distribution determines hypoxic chemokine ISG induction. (A) Overview of the lysosomal cholesterol distribution machinery. Relevant interventions are indicated in red. (B) THP-1 cells were pre-incubated for 1 h with 5 µM U18666A or DMSO in medium containing high or low levels of FBS prior to incubation under N (grey) or H (red) for 24 h (n = 11). (C) THP-1 cells were pre-incubated for 1 h with 5 µM U18666A or DMSO in medium containing low levels of FBS ± 0.25 mg/mL methyl-β-cyclodextrin-complexed cholesterol prior to incubation under N (grey) or H (red) for 24 h (n = 7). (D) THP-1 cells were incubated under N (grey) or H (red) for 24 h in medium containing low levels of FBS prior to infection with SARS-CoV-2 (strain FFM1) under N. RNA was isolated 1 hour post infection (n = 5). OAS1, IRF7, CCL2, CXCL10, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. All data are means ± SEM and were statistically analyzed using two-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ##p < 0.01, ###p < 0.001 (compared to to FBS high/DMSO (B), FBS low/DMSO (C), or FBS low/ctrl (D), respectively)).

Figure 6

TLR4 signaling contributes to hypoxic ISG induction. (A) Overview of toll-like receptors (TLRs), adaptor proteins, and kinases regulating ISG induction. Relevant inhibitors are highlighted in red. (B) THP-1 cells were incubated under N (grey) or H (red) for 24 h in medium containing high or low levels of FBS (n = 3). TLR4 and TLR7 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. (C–E) THP-1 cells were pre-incubated for 1 h with (C) 10 µM TJ-M2010-5 (MyD88 inhibitor), (D) 0.5 µM BX-795 (TBK1/IKKϵ inhibitor), 0.1 µM IKK-16 (IKKα/β inhibitor), (E) 10 µM TAK-242 (TLR4 inhibitor), 0.1 µM enpatoran (TLR7/8 inhibitor), or DMSO in medium containing high or low levels of FBS prior to incubation under N (grey) or H (red) for 24 h (n = 3). OAS1, IRF7, CCL2, CXCL10, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. All data are means ± SEM and were statistically analyzed using two-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ##p < 0.01, ###p < 0.001 (compared to to FBS high (B) or FBS high/DMSO (C–E), respectively)).

Figure 7

Hypoxic priming increases production of chemokine ISGs after SARS-CoV-2 infection via TLR4 activation. (A) THP-1 cells were pre-incubated for 1 h with 10 µM TAK-242, 10 µM fatostatin, or DMSO in medium containing low levels of FBS prior to incubation under N (grey) or H (red) for 32 h 5 µg/mL SARS-CoV-2 spike protein or glycerol were added for the last 8 h (n = 4). CCL2, CXCL10, OAS1, IRF7, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. (B, C) THP-1 cells were pre-incubated for 1 h with 10 µM TAK-242, 10 µM fatostatin, or DMSO in medium containing low levels of FBS prior to incubation under N (grey) or H (red) for 24 h Subsequently, cells were infected with SARS-CoV-2 (strain FFM1) under N (n = 4). (B) RNA was isolated 6 hours post infection. CCL2, CXCL10, OAS1, IRF7, and MSMO1 mRNA expression was analyzed by RT-qPCR and normalized to GUSB expression. (C) Secreted CCL2 and CXCL10 protein levels were determined by ELISA in supernatants 6 hours post infection. All data are means ± SEM and were statistically analyzed using two-way repeated measures ANOVA with Holm-Šídák’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ##p < 0.01, ###p < 0.001 (compared to to FBS low/DMSO)).

Figure 8

Model of the impact of hypoxia on IFN responses to SARS-CoV-2 infection in monocytes. Hypoxia alters cholesterol biosynthesis flux and subcellular cholesterol dynamics, consequently enhancing TLR4/MyD88-dependent chemokine ISG production in response to SARS-CoV-2 infection.