Cutting edge: Vitamin D regulates lipid metabolism in Mycobacterium tuberculosis infection

Abstract

Vitamin D has long been linked to resistance to tuberculosis, an infectious respiratory disease that is increasingly hard to treat because of multidrug resistance. Previous work established that vitamin D induces macrophage antimicrobial functions against Mycobacterium tuberculosis. In this article, we report a novel, metabolic role for vitamin D in tuberculosis identified through integrated transcriptome and mechanistic studies. Transcriptome analysis revealed an association between vitamin D receptor (VDR) and lipid metabolism in human tuberculosis and infected macrophages. Vitamin D treatment of infected macrophages abrogated infection-induced accumulation of lipid droplets, which are required for intracellular M. tuberculosis growth. Additional transcriptomics results showed that vitamin D downregulates the proadipogenic peroxisome proliferator-activated receptor γ (PPARγ) in infected macrophages. PPARγ agonists reversed the antiadipogenic and the antimicrobial effects of VDR, indicating a link between VDR and PPARγ signaling in regulating both vitamin D functions. These findings suggest the potential for host-based, adjunct antituberculosis therapy targeting lipid metabolism.

Figure 1. Transcriptomic analyses of in vitro THP-1 infection and of clinical data

(A) DNA-binding proteins implicated by genes upregulated in M. tuberculosis-infected THP-1 cells. Gene sets were defined as genes annotated as bound by each transcriptional modulator (16). Differential expression between sample classes for these gene sets was determined by a non-parametric test (CERNO (17)). The resulting P-values were plotted onto the x-axis. Of 60 gene sets at FDR < 0.05, the top five are shown (IRF8, n = 607; IRF1, n = 326; VDR, n = 161; STAT4, n = 896; SUZ12, n = 2812). To represent effect size, sets with fewer genes were given greater bar height than larger sets that yielded similar P-values. (B) Statistical significance of VDR-bound genes in clinical comparisons, as indicated. Primary data were from (14). P-values were calculated as in (A). (C) DNA-binding proteins implicated by analysis of upregulated genes annotated for Metabolism of Lipids and Lipoproteins (n = 348) in infected THP-1 cells. Gene sets were defined by the overlap between the pathway genes and genes annotated as bound by a transcriptional modulator (Supplemental Fig. 1). Each overlap set was tested for differential expression as in (A). Of 29 gene sets at FDR < 0.05, the top five are shown (VDR, n = 23; SOX2, n = 165; RUNX1, n = 126; ERG, n = 64; FOXA2, n = 93), plotted as in (A).

Figure 2. Effects of vitamin D on lipid droplet formation and M. tuberculosis growth in THP-1 cells

Differentiated THP-1 cells were infected with M. tuberculosis, treated with 100nM vitamin D (1,25α-dihydroxyvitamin D3), as indicated, and harvested at 24 and 48 hrs post-infection. Results shown are means from triplicate experiments (+/− standard error of the mean). (A) Cells were fixed and stained with LipidTOX Deep Red for detection of lipid droplets by flow cytometry. Data were calculated as MFI. (B) Infected cells were lysed and M. tuberculosis colony-forming units (CFU) were determined. The selected dose of vitamin D caused maximal inhibition of lipid droplets and induced significant expression of VDR target genes CAMP and CYP24A1 in uninfected and infected cells at 24 hrs post-infection (not shown).

Figure 3. Transcriptomic analyses of THP-1 cells infected with M. tuberculosis and treated with vitamin D

(A) 108 Gene Ontology gene sets containing the terms “lipid” or “cholesterol” were tested for increased and decreased expression. FDR values for these annotations (gray circles) were plotted for the comparison infected vs. uninfected cells (infection, I, x-axis) and for infected cells with/without vitamin D treatment (treatment, T, y-axis). Test results for increased (up arrows) and decreased (down arrows) gene expression were plotted. The gray area includes non-significant results for both comparisons (FDR ≥ 0.05), while the white area contains significant results (FDR < 0.05). (B) DNA-binding proteins implicated by downregulation of genes annotated for Lipid Binding upon vitamin D treatment of M. tuberculosis-infected THP-1 (upper panel). Shown are the top five gene sets at FDR < 0.05, plotted as in Figure 1. Gene set overlap was determined as in Fig. 1C, and the most significantly regulated genes are shown (p < 0.005) (lower panel).

Figure 4. Role of PPARγ

Differentiated THP-1 cells were infected with M. tuberculosis and treated with 100nM vitamin D, 1μM PPARγ agonist GW1929, and 100nM PPARγ agonist Rosiglitazone, alone or combined, as indicated. Results are means from triplicate experiments (+/− standard error of the mean). (A) Cells were infected at MOI 1:1, and total RNA was extracted at 24 hrs post-infection. Transcript measurements were obtained by qRT-PCR using gene-specific primers and molecular beacons. Data are normalized to the reference gene GAPDH. (B) Lipid droplets (LD) from cells infected as in panel A were detected by fluorescence staining and flow cytometry as in Figure 2B. MFI is shown. (C) M. tuberculosis CFU were determined in lysates from cells infected at MOI 1:30 at 4 days post-infection (MOI was reduced in this experiment to prolong infected THP-1 survival). For gene expression (A), lipid droplets (B), and CFU (C), the effect of vitamin D on response to infection and the further effect of PPARγ agonist were statistically significant (p < 0.05).