A Critical Role for the Mitochondrial Pyruvate Carrier in Hepatic Stellate Cell Activation

Abstract

Background & aims: Hepatic stellate cells (HSCs) are non-parenchymal cells of the liver that produce the extracellular matrix that forms fibrotic lesions in chronic liver disease, including metabolic dysfunction-associated steatohepatitis (MASH). The mitochondrial pyruvate carrier (MPC) catalyzes the transport of pyruvate from the cytosol into the mitochondrial matrix, which is a critical step in pyruvate metabolism. An MPC inhibitor has shown promise as a novel therapeutic for MASH and HSC activation, but a mechanistic understanding of the direct effects of MPC inhibition on HSC activation is lacking.

Methods: Stable lines of LX2 cells expressing short hairpin RNA against MPC2 were established and examined in a series of studies to assess HSC metabolism and activation. Mice with conditional, HSC-specific MPC2 deletion were generated and their phenotypes assessed in the context of diets that cause hepatic steatosis, injury, and early-stage fibrosis.

Results: Genetic suppression of MPC activity markedly decreased expression of markers of HSC activation in vitro. MPC knockdown reduced the abundance of several intermediates of the tricarboxylic acid cycle and attenuated HSC activation by suppressing hypoxia inducible factor-1α signaling. Supplementing alpha-ketoglutarate to replenish the tricarboxylic acid cycle intermediates was sufficient to overcome the effects of MPC inhibition on hypoxia inducible factor-1α and HSC activation. On high-fat diets, mice with HSC-specific MPC deletion exhibited reduced circulating transaminases, numbers of HSCs, and hepatic expression of markers of HSC activation and inflammation compared with wild-type mice.

Conclusions: These data suggest that MPC inhibition modulates HSC metabolism to attenuate activation and illuminate mechanisms by which MPC inhibitors could prove therapeutically beneficial for treating MASH.

Keywords: Collagen; Fibrosis; HIF1alpha; MASLD; TCA Cycle.

Graphical abstract

Figure 1

Genetic inhibition of the MPC reduces hepatic stellate cell activation in vitro. (A) A schematic for glucose metabolism and pyruvate entry into the mitochondrial TCA cycle is shown. IMM, inner mitochondrial membrane; OMM, outer mitochondrial membrane. Created with BioRender.com. (B) Stable human LX2 hepatic stellate cells were engineered to express shRNA against MPC2 (or Scr shRNA control). TGF-β-1 (5 ng/mL) was added to culture medium of some wells to activate HSC. Collagen or MPC proteins were detected by Western blotting. (C) Seahorse respirometry assays demonstrate that OCR and extracellular acidification rate (ECAR) are decreased by MPC2 shRNA. (D) Collagen isoform gene expression was decreased in activated LX2 cells expressing MPC2 shRNA. Data are expressed as mean ± SEM, relative to control cells expressing non-targeting shRNA. ∗P < .05 vs Scr shRNA cells of the same treatment group. (E) Col1a1, Col3a1, Acta2, and Timp1 gene expression of isolated HSCs from WT Mpc2^fl/fl mice and Lrat-Mpc2^-/- littermates that were cultured for 7 days (Day 7). Some cells were harvested after 1 day of culture (Day 1) for quiescent HSCs. Gene expression was measured by RT-qPCR, and data are expressed as mean ± SEM, relative to Day 1 HSC. ∗P < .01 vs WT day 7 cells as determined by 1-way ANOVA, followed by Tukey’s post hoc analysis. n = 3 technical replicates from a representative experiment.

Figure 2

Lrat-Mpc2^-/- mice are protected from MASH-inducing diet (HFC). At about 8 weeks of age, littermate male WT and Lrat-Mpc2-/- (knockout) mice were placed on either a LFD or an HFC diet for a period of 12 weeks. (A–B) Terminal body weight and changes in body weight relative to the baseline. (C) Liver weight, measured at sacrifice, relative to terminal body weight. (D–E) body composition, determined by EchoMRI, expressed as mean ± SEM (n = 7–11/group). (F) Liver weight, fat mass, and lean mass measured at sacrifice. (G) Plasma levels of circulating transaminases ALT and AST collected at sacrifice. (H, I) Analysis of triglycerides (TG), total cholesterol (TC), and NEFA from liver (H) and plasma (I), expressed as mean ± SEM (n = 7–11/group). (J) Representative liver sections with H&E staining and immunofluorescence staining for hepatic stellate cells (Desmin), macrophages (F4/80), and nuclei (DAPI). (K) Histologic scoring of H&E-stained liver sections assessing steatosis, macrosteatosis, lobular inflammation, and NAFLD activity score expressed as mean ± SEM (n = 6–7/group). (L) Quantification of desmin staining.

Figure 3

Lrat-MPC2^-/- induced cell death and protected from MASH-inducing diet (HFC). (A) Representative images indicating primary HSCs from knockout (KO) and WT mice stained using SYTOX green nucleotide stain with quantification. ∗∗P < .01. (B) Hepatic gene expression measured by RT-qPCR and expressed relative to WT LFD control group. All data expressed as mean ± SEM (n = 7–11/group). ∗P < .05; ∗∗P< .01 vs WT mice on the same diet as determined by 2-way ANOVA, followed by Sidak’s test for multiple comparisons.

Figure 4

MPC deletion in hepatic stellate cells protects mice from MASH-inducing diet (CDAA). At about 10 weeks of age, male (blue data points) and female (pink data points) Lrat-Mpc2^-/- knockout (KO) and WT littermates received a CDAA diet for a period of 10 weeks. (A) Body weight change in experimental period reported as percentage relative to initial body weight. (B) Plasma levels of ALT at sacrifice. (C) Liver weight measured at sacrifice normalized to final body weight. (D) Hepatic gene expression measured by RT-qPCR and expressed relative to WT chow control group. All data expressed as mean ± SEM (n = 3 [chow] or 10 [CDAA]/group). ∗P< .05; ∗∗P< .01; ∗∗∗P< .001 vs WT mice on the same diet as determined by 2-way ANOVA, followed by Sidak’s test for multiple comparisons. (E) Representative liver sections with H&E and Sirius red staining in knockout (KO) and WT mice fed a CDAA diet with quantification ∗∗∗∗P < .001.

Figure 5

Lrat-Mpc2^-/- mice are protected from MASH-inducing diets (HFC). (A–C) Volcano plots of differentially expressed genes with P< .05 comparing (A) HFC vs LFD in WT mice, (B) knockout (KO) vs WT mice on LFD, and (C) KO vs WT mice on HFC diet. DEGs with Log fold change (LogFC) less than −0.5, or greater than 0.5, were highlighted in either blue or red, respectively. (D–E) GSEA of perturbations in Hallmark gene set collections when comparing (D) HFC vs LFD in WT mice and (E) KO vs WT mice on HFC diet. The arrows highlight hypoxia signaling pathways. Differential expression analysis was then performed to analyze for differences between conditions. RNA-seq was performed on liver tissue from WT and Lrat-Mpc2-/- knockout (KO) mice on either a LFD or HFC diet (n = 5/group).

Figure 6

Lrat-Mpc2-/- mice are protected from MASH-inducing diets (CDAA). (A) Violin plots of number of genes detected per cell, number of UMIs per cell, and proportion of cell reads from mitochondrial genes. (B) Bar graph of annotated cell clusters expressed as percent of all cells in each respective genotype. (C) Uniform manifold approximation and projection (UMAP) of non-parenchymal cells from WT and knockout (KO) mice after CDAA diet-feeding for 10 to 12 weeks. (D) Dot plot of cell type specific markers for each cell cluster used for annotation. (E) Volcano plot of DEGs in the HSC cluster when comparing KO vs WT. Genes that had P < .05 were highlighted in either red (Log2FC >0.5) or blue (Log2FC <0.5) in KO vs WT. (F) Heatmap representation of highly altered Hallmark gene set pathways among all clusters when comparing KO vs WT using SCPA, expressed by the Qval (Q value statistic), which represents the size of distribution for change within a given pathway. cDC1 (conventional type 1 dendritic cell), cDC2 (conventional type 2 dendritic cell), CD4 TH (CD4 t helper), C-LAMs, CTL, γδT, hepatic stem (hepatic stem/progenitor cells), HSCs, ILC, KC, LAMs, Low q (cells with low RNA quality [high mitochondrial genes, low RNA count]), mig cDC (migratory conventional dendritic cell), NK, pDC (plasmacytoid dendritic cell), prolif cDCs (proliferating conventional dendritic cells), prolif T cells (proliferation T cells), trans monocytes (transitioning monocytes), T_EM.

Figure 7

Lrat-Mpc2^-/- mice have altered transcriptional responses in myeloid cells. (A) Schematic depicting interaction between macrophages and HSCs. In response to chronic steatotic liver disease, circulating monocytes enter the liver and differentiate into distinct populations of macrophages that localize in areas of lipid accumulation and fibrosis, forming hepatic crown-like structures, that are associated with increased stellate cells and fibrosis. (B) Uniform Manifold Approximation and Projection (UMAP) of the monocyte and macrophage subsets. (C) Bar graph of monocyte/macrophage subset clusters expressed as percent within each respective genotype. (D) UMAP representation of previously established markers that were used to annotate monocyte/macrophage clusters. (E) Representative immunofluorescence staining of liver sections for macrophages (F4/80) (Scale bar = 100 μm). (F–G) Volcano plot of DEGs in the C-LAM and LAM clusters, respectively, when comparing KO vs WT. Genes that had P < .05 were highlighted in either red (Log2FC >0.5) or blue (Log2FC <0.5) in knockout (KO) vs WT. (H) Violin plots of Spp1 expression in WT and KO monocyte/macrophage subsets.

Figure 8

MPC suppression markedly reduces several TCA cycle intermediates in LX2 cells treated with uniformly labeled ¹³C-glucose. (A) Schematic of glucose metabolism and the TCA cycle tracing the route of incorporation of ¹³C-glucose. (B–C) TGF-β-1 (5 ng/mL) was added into media and LX2 cells expressing either Scr control shRNA or shRNA against MPC2. Cells were then cultured overnight in the presence of uniformly labeled ¹³C-glucose. (B) Incorporation of ¹³C-glucose into glycolytic intermediates. (C) Incorporation of ¹³C-glucose into TCA cycle intermediates and amino acids; M2, M3, indicate 2 or 3 carbon atoms are derived from labeled glucose, respectively. (D) Relative abundance of intermediates in cells expressing MPC2 shRNA or Scr control shRNA. Data are expressed as mean ± SEM (n = 3) from a representative experiment. ∗P < .05 vs Scr shRNA as determined by unpaired Student t-test. Schematics created with BioRender.com

Figure 9

MPC suppression markedly reduces several TCA cycle intermediates in LX2 cells treated with uniformly labeled ¹³C-glutamine. (A) The schematic depicts glutaminolysis and the TCA cycle and traces the incorporation of ¹³C-glutamine. (B–C) Scr control or MPC2 shRNA expressing LX2 cells were treated with TGF-β-1 (5 ng/mL) and cultured overnight in the presence of uniformly-labeled ¹³C-glutamine. Enrichments of ¹³C-glutamine into TCA cycle intermediates and amino acids are shown. Data are expressed as mean ± SEM (n = 3) from a representative experiment. ∗P < .05 vs Scr shRNA as determined by unpaired Student t-test. (D) Western blots for collagen 3, collagen 1, SMA, HIF1α, and α-Tubulin in LX2 cells treated with or without TGF-β-1 (5 ng/mL), glutamine (Gln, 2 mM), and in the presence or absence of the glutaminase inhibitor BPTES (10 or 20 μM). (E) Western blot images for collagen 1, HIF1α, and α-Tubulin in LX2 cells treated with or without glutamine (Gln, 2 mM) or dm-αKG (5 mM) compared with TGF-β-1 (5 ng/mL) treated cells. (F–G) Protein abundance of the indicated proteins in LX2 cells treated with or without TGF-β-1 (5 ng/mL), glutamine (Gln, 2 mM), and (F) BPTES (10 μM) or (G) CB839. Cells in the last lane were supplemented with dm-αKG (5 mM). (H) Western blot images for collagen 1, SMA, HIF1α, and α-Tubulin in LX2 cells treated with or without TGF-β-1 (5 ng/mL) and HIF1α inhibitor VI (10 or 20 μM).

Figure 10

MPC deficiency leads to impaired HIF1α activation by a metabolic mechanism. (A) Western blot images for HIF1α, collagen 3, collagen 1, MPC2, MPC1, and α-Tubulin in LX2 cells expressing shRNA against MPC2. Cells treated with TGF-β-1 (5 ng/mL). Graph below depicts HIF1A mRNA expression in cells treated similarly. (B) Western blot images for collagen 1, HIF1α, and α-Tubulin in lysates from LX2 cells expressing Scr or MPC2 shRNA and treated with TGF-β-1 (5 ng/mL) with or without dm-αKG (5 mM). Graph below depicts mRNA expression of collagen genes in LX2 cells treated in the same way. (C) Relative intensity of TCA cycle intermediates in LX2 cells expressing Scr or MPC2 shRNA treated with TGF-β-1 (5 ng/mL) with or without dm-αKG (5 mM); Data are expressed as mean ± SEM (n = 3) from a representative experiment. ∗P < .0001 vs Scr shRNA as determined by 2-way ANOVA followed by Tukey’s post-hoc test. (D) Western blot images for HIF1α, collagen 1, collagen 3, and α-Tubulin in LX2 cells treated with or without TGF-β-1 (5 ng/mL), glutamine (Gln, 2 mM), dm-αKG (5 mM), or diethyl-succinate (1 or 5 mM).

Figure 11

MPC deficiency leads to impaired HIF1α activation by a metabolic mechanism. (A) The abundance of indicated proteins in LX2 cells treated with or without TGF-β-1 (5 ng/mL), glutamine (Gln, 2 mM), dm-αKG (5 mM), or D- or L-2-hydroxyglutarate (5 mM). (B) Protein abundance of collagen 1, collagen 3, and HIF1α in LX2 cells expressing Scr or MPC2 shRNA treated with TGF-β-1 (5 ng/mL) and with or without DMOG (1 mM), a cell-permeable HIF1α stabilizer. Quantified protein fold change in LX2 cells expressing Scr or shMPC2 treated in the same way. Data expressed as mean ± SEM (n = 3/group) and analyzed by 2-way ANOVA followed by uncorrected Fisher’s LSD test for multiple comparisons; ∗P < .05. (C) Selected HIF1α target genes from RNAseq data expressed as counts per million (CPM) and represented as mean ± SEM (n = 5/group). ∗P < .05 as determined by 2-way ANOVA, followed by Sidak’s test for multiple comparisons. RNA-seq was performed on liver tissue from WT and Lrat-Mpc2^-/- (KO) mice and placed on either an LFD or HFC diet (n = 5/group).