Imbalanced gut microbiota fuels hepatocellular carcinoma development by shaping the hepatic inflammatory microenvironment

Abstract

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related deaths worldwide, and therapeutic options for advanced HCC are limited. Here, we observe that intestinal dysbiosis affects antitumor immune surveillance and drives liver disease progression towards cancer. Dysbiotic microbiota, as seen in Nlrp6^-/- mice, induces a Toll-like receptor 4 dependent expansion of hepatic monocytic myeloid-derived suppressor cells (mMDSC) and suppression of T-cell abundance. This phenotype is transmissible via fecal microbiota transfer and reversible upon antibiotic treatment, pointing to the high plasticity of the tumor microenvironment. While loss of Akkermansia muciniphila correlates with mMDSC abundance, its reintroduction restores intestinal barrier function and strongly reduces liver inflammation and fibrosis. Cirrhosis patients display increased bacterial abundance in hepatic tissue, which induces pronounced transcriptional changes, including activation of fibro-inflammatory pathways as well as circuits mediating cancer immunosuppression. This study demonstrates that gut microbiota closely shapes the hepatic inflammatory microenvironment opening approaches for cancer prevention and therapy.

Fig. 1. Loss NLRP6 augments liver disease progression in NEMO^∆hepa mice.

a Macroscopic appearance of NEMO^∆hepa and NEMO^∆hepa/Nlrp6^−/− livers at 52 weeks of age. b Quantification of liver tumors with a diameter larger than 1 cm in NEMO^∆hepa (n = 13) and NEMO^∆hepa/Nlrp6^−/− (n = 8) livers; unpaired two-tailed Student’s t test, 95% CI −2.239 to −0.107, p = 0.0328. c Serum ALT and GLDH levels of 52-week-old NEMO^∆hepa (n = 12), NEMO^∆hepa/Nlrp6^−/− (n = 8)) and respective controls (WT (n = 6), Nlrp6^−/− (n = 6)); one-way ANOVA with Tukey’s multiple comparisons test (ALT: WT vs. NEMO^∆hepa, 95% CI −708.8 to −141.2, p = 0.0018; WT vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −1023 to −410.1, p = <0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −519.1 to −0.93, p = 0.0489, GLDH: WT vs. NEMO^∆hepa, 95% CI −339.4 to −127.3, p < 0.0001; WT vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −494.2 to −265.1, p = <0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −243.1 to −49.49, p = 0.049). d Representative pictures of H&E-stained liver sections, immunohistochemical (IHC) staining of CD45 and Sirius red stained liver sections of 52-week-old WT, Nlrp6^−/−, NEMO^∆hepa, and NEMO^∆hepa/Nlrp6^−/− livers, representative of 2 independent experiments; Scale bar: 100 µm. e RT-qPCR analysis of pro-fibrotic mRNA expression (TGFβ, Col1a1) in whole liver of NEMO^∆hepa (n = 12), NEMO^∆hepa/Nlrp6^−/− (n = 11) and respective controls (WT (n = 6), Nlrp6^−/− (n = 8)); one-way ANOVA with Sidak’s multiple comparisons test (TGFb: WT vs. NEMO^∆hepa, 95% CI −0.661 to 0.8128, p = ns; Nlrp6^−/− vs. NEMO^∆hepa/Nlrp6^−/−,95% CI –1.418 to −0.0483, p = 0.0328; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −1.252 to −0.0219, p = 0.0405; Col1a1: WT vs. NEMO^∆hepa, 95% CI −13.08 to 10.38, p = ns; Nlrp6^−/− vs. NEMO^∆hepa/Nlrp6^−/−,95% CI –26.78 to −3.865, p = 0.006; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −24.53 to −3.950, p = 0.004). f Representative pictures of immunofluorescence (IF) stainings of CD11b and CD8 of NEMO^∆hepa, NEMO^∆hepa/Nlrp6^−/− and respective controls (WT, Nlrp6^−/−). Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. g RT-qPCR analysis of inflammatory mRNA expression (TNFα, IL-1β,Tlr4) in whole liver of NEMO^∆hepa (n = 12), NEMO^∆hepa/Nlrp6^−/− (n = 11) and respective controls (WT (n = 4), Nlrp6^−/− (n = 5)); one-way ANOVA with Sidak’s multiple comparisons test (Tnfa: WT vs. NEMO^∆hepa, 95% CI −2.37 to 0.6196, p = ns; Nlrp6^−/− vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI –4.04 to −0.806, p = 0.0015; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −2.505 to −0.006, p = 0.023; IL-1β: WT vs. NEMO^∆hepa, 95% CI −1.45 to 1.72, p = ns; Nlrp6^−/− vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI –2.90 to −0.1197, p = ns; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −2.601 to −0.1138, p = 0.047; Tlr4: WT vs. NEMO^∆hepa, 95% CI −5.62 to 4.156, p = ns; Nlrp6^−/− vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI –6.61 to −2.71, p = ns; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −8.989 to −1.323, p = 0.005). h Immunoblot analysis of liver protein extracts from 52-week-old mice of all indicated genotypes for JNK, p-JNK and GAPDH as loading control; representative of 2 experiment. i Histological quantification of Cleaved Caspase 3 positive cells per viewfield; NEMO^∆hepa (n = 17) and NEMO^∆hepa/Nlrp6^−/− (n = 24) viewfields one-way ANOVA with Tukey’s multiple comparisons test (NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −8.141 to −3.410, p < 0.0001). j Representative pictures of IF staining of KI67. Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. All Data are presented as the mean ± standard error of the mean (SEM). Experiments are considered significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***) and p < 0.0001 (****). Individual data points represent biological replicates unless otherwise stated. Source data are provided as a Source data file.

Fig. 2. Loss of NLRP6 results in intestinal dysbiosis and barrier impairment correlating with steatohepatitis activity and tumor burden.

a Cecal microbiota composition of 13-week-old mice (WT (n = 8), Nlrp6^−/− (n = 9), NEMO^∆hepa (n = 13), NEMO^∆hepa/Nlrp6^−/− (n = 11)) was analyzed using 16S rRNA gene amplicon sequencing. Bar chart based on permutational multivariate analysis of variance (ADONIS) presenting the percentage of variance of the gut microbiota explained by the factors “Genotype” (R² = 0.057, **p < 0.006), “Cage” (R² = 0.073, **p < 0.004), “Nemo” (R² = 0.079, **p < 0.002),” Line”. Genotype - all included genotypes; Cage – individual cages; Nemo – WT & Nlrp6^−/− mice vs. NEMO^∆hepa & NEMO^∆hepa/Nlrp6^−/−; Line – WT & NEMO^∆hepa vs. Nlrp6^−/− & NEMO^∆hepa/Nlrp6^−/− (R² = 0.16, ***p < 0.001). b Analysis for differential abundance of microbiota via DESeq analysis of 13 weeks old NEMO^∆hepa (n = 13) and NEMO^∆hepa/Nlrp6^−/− (n = 11) mice. c Linear discriminant analysis (LDA) of effect size (LEfSe) between NEMO^∆hepa (n = 13) and NEMO^∆hepa/Nlrp6^−/− (n = 11). d Representative pictures of IF stainings of ZO-1 in ileum and colon of NEMO^∆hepa, NEMO^∆hepa/Nlrp6^−/− and respective controls (WT, Nlrp6^−/−). Nuclei were counterstained with DAPI, representative of 2 independent experiments; Scale bar: 100 µm. e, f Immunoblot analysis of ileum and colon protein extracts from 52-week-old mice for Occludin and β-actin as loading control of indicated genotypes, representative of 2 experiments. g RT-qPCR analysis of inflammatory mRNA expression (IL-1β, Il-18, TNFα, Mcp1, Ccl5) in the ileum of NEMO^∆hepa (n = 12) and NEMO^∆hepa/Nlrp6^−/− (n = 12) mice, unpaired t-test (IL-1β: NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI 0.0235–2.421, p = 0.046; Il-18: NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI 0.0293–0.7369, p = 0.035; TNFα: NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −0.210 to 0.595, p = n.s.; Mcp1: NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI 0.097–1.932, p = 0.032; Ccl5: NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI 1.023–4.706, p = 0.0038. h In vivo intestinal permeability assessed by oral gavage of FITC-dextran. Rel. induction of FITC-dextran permeability in NEMO^∆hepa (n = 6) and NEMO^∆hepa/Nlrp6^−/− mice (n = 4), unpaired t-test: NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI 0.101–0.615, p = 0.012), pooled data from 2 independent experiments. i Strong correlation (Spearman) of intestinal barrier function evidenced by FITC-dextran permeability with tumor number in NEMO^∆hepa (blue dots) and NEMO^∆hepa/Nlrp6^−/− (red dots) mice (8 pairs), p = 0.0012 (two-sided), r = 0.96. All Data are presented as the mean ± standard error of the mean (SEM; experiments are considered significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****). Individual data points represent biological replicates unless otherwise stated. Source data are provided as a Source data file.

Fig. 3. mMDSC drive steatohepatitis activity in NEMO^∆hepa/Nlrp6^−/− mice.

a Representative pictures of H&E-stained liver sections and immunohistochemical staining of KI67 stained liver sections of 13-week-old WT, Nlrp6^−/−, NEMO^∆hepa, and NEMO^∆hepa/Nlrp6^−/− livers; Scale bar: 100 µm. b Histological quantification of Ki67⁺ cells per viewfield (WT n = 18, Nlrp6^−/− n = 12, NEMO^∆hepa n = 36, NEMO^∆hepa/Nlrp6^−/− n = 35, one way ANOVA, Tukeys’s multiple comparison test:Ki67: WT vs. NEMO^∆hepa, 95% CI −34 to −22, p < 0.0001; WT vs. NEMO^∆hepa/Nlrp6^−/−,95% CI –40 to −28, p < 0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −11.54 to −1.67, p = 0.0039). c Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels of 13-week-old NEMO^∆hepa (n = 13), NEMO^∆hepa/Nlrp6^−/− (n = 10) and respective controls (WT (n = 9), Nlrp6^−/− (n = 8), one-way ANOVA, bonferroni’s multiple comparison test: ALT: WT vs. NEMO^∆hepa, 95% CI −1.123 to −344.4, p < 0.0001; WT vs. NEMO^∆hepa/Nlrp6^−/−,95% CI –1700 to −875.6, p < 0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −931.8 to −177.0, p = 0.002; AST: WT vs. NEMO^∆hepa, 95% CI −764 to −210, p < 0.0001; WT vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI –1147 to −588, p < 0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −1.735 to −0.2338, p < 0.001). d Representative pictures and e quantification of immunohistochemical (IHC) staining of CD45 stained liver sections of 13-week-old mice (WT(n = 4), Nlrp6^−/− (n = 3), NEMO^∆hepa (n = 13), and NEMO^∆hepa/Nlrp6^−/− (n = 8), one way ANOVA, Tukey’s multiple comparisons test, WT vs. NEMO^∆hepa/Nlrp6^−/−,95% CI −2855 to −0.8097, p < 0.001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/− 95% CI −7.5 to −0.8, p = 0.007); Scale bar: 200 µm. f Flow cytometry analysis of monocytic myeloid-derived suppressor cells (mMDSC) isolated from whole liver of 13-week-old NEMO^∆hepa (n = 12), NEMO^∆hepa/Nlrp6^−/− (n = 11) and respective controls (WT (n = 7), Nlrp6^−/− (n = 8) one-way ANOVA, Sidak’s multiple comparison test: WT vs. NEMO^∆hepa, 95% CI −10.1 to −2.7, p = 0.0002; Nlrp6^−/− vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI –11.8 to −4.2, p < 0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI −7.5 to −0.8, p = 0.009). g The percentage of mMDSCs of CD45⁺ living cells strongly correlates with ALT levels in NEMO^∆hepa and NEMO^∆hepa/Nlrp6^-/- mice (Spearman, 23 pairs). h Flow cytometry analysis of CD3⁺CD8⁺ cytotoxic T cells isolated from whole liver of 13-week-old NEMO^∆hepa (n = 14), NEMO^∆hepa/Nlrp6^−/− (n = 10) and respective controls (WT (n = 9), Nlrp6^−/− (n = 9); one-way ANOVA: Sidak’s multiple comparison test: WT vs. NEMO^∆hepa, 95% CI −12.4 to −8.0, p < 0.0001; NEMO^∆hepa vs. NEMO^∆hepa/Nlrp6^−/−, 95% CI 5.3 to −9.5, p < 0.0001. i Percentage of mMDSCs of CD45⁺ living cells and CD8⁺ cytotoxic T-cells is inversely correlated (Spearman 18 pairs). j In vitro T-cell proliferation assay. T cells were co-cultured with granulocytic MDSCs or monocytic MDSCs, representative of 2 experiments. All Data are presented as the mean ± standard error of the mean (SEM) Experiments are considered significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****). Individual data points represent biological replicates unless otherwise stated. Source data are provided as a Source data file.

Fig. 4. Microbiota modulation reshapes the hepatic inflammatory microenvironment in NEMO^∆hepa livers.

a NEMO^∆hepa and NEMO^∆hepa/Nlrp6^−/− mice were treated with broad-spectrum antibiotics (ABx) from week 8 till week 13. Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels of NEMO^∆hepa (n = 12) and NEMO^∆hepa/Nlrp6^−/− with (n = 11) or without (n = 10) 5 weeks ABx treatment, control groups (–Abx) shared with Fig.3c, Fig.4e; one way ANOVA, Sidak’s multiple comparison test: ALT: NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/−, 95% CI −980.6 to −240.7, p = 0.0007; NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI −316.0 to 405.4, p = n.s.; NEMO^Δhepa/Nlrp6^−/− vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI 277.8 to 1033, p = 0.0004; AST: NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI 8.094 to 552.3, p = 0.047; NEMO^Δhepa/Nlrp6^−/− vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI −399.3 to 120.6, p = n.s. b Flow cytometry analysis of monocytic myeloid-derived suppressor cells (mMDSC) and CD3⁺CD8⁺ cytotoxic T-cell isolated from whole liver of 13-week-old NEMO^∆hepa (n = 12–14) and NEMO^∆hepa/Nlrp6^−/− with (n = 11) or without (n = 10–11) 5 wks Abx treatment control groups (–Abx) shared with Fig. 3d, f, one way ANOVA, Sidak’s multiple comparison test: CD11b: NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/−, 95% CI −7.624 to −1.307, p = 0.0036; NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI −0.9433 to 5.374, p = n.s. NEMO^Δhepa/Nlrp6^−/− vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI 3454 to 9908, p < 0.0001; CD3+ CD8+: NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/−, 95% CI − 3.946 to 10.84, p < 0.0001; NEMO^Δhepa vs. NEMO^Δhepa/NLRP6^−/−+ ABx, 95% CI −1.245 to 5.460, p = n.s.; NEMO^Δhepa/NLRP6^−/− vs. NEMO^Δhepa/Nlrp6^−/− + ABx, 95% CI −8.919 to −1.648, p = 0.003; control groups (–Abx) shared with Fig. 3f, h and (g), (h). c Representative pictures of immunofluorescence staining of CD11b of NEMO^∆hepa and NEMO^∆hepa/Nlrp6^−/− with and without Abx treatment. Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. d Analysis for differential abundance of microbiota via DESeq analysis of 13 weeks old NEMO^∆hepa (n = 13) and NEMO^∆hepa-FMT (n = 6) mice. NEMO^∆hepa group shared with Fig. 2. e Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels of NEMO^∆hepa mice treated with NEMO^∆hepa/Nlrp6^−/− (+FMT, n = 5) microbiota or without treatment (−FMT, n = 12), control group (–FMT) shared with Fig.3c, (a). f Representative pictures of immunofluorescence staining of CD11b of NEMO^∆hepa mice treated with +FMT or −FMT. Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. g Flow cytometry analysis of monocytic myeloid derived suppressor cells (mMDSC) isolated from whole liver of NEMO^∆hepa mice treated with +FMT (n = 5) or –FMT (n = 12) and NEMO^∆hepa/Nlrp6^−/− treated with +FMT (n = 4) or –FMT (n = 11), one way ANOVA, Holm–Sidak’s multiple comparison test: NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/−, p = 0.005; Nemo^Δhepa vs. Nemo^Δhepa +FMT, p = 0.040.; control group (–FMT) shared with Fig.4b, Fig. 3f. h Flow cytometry analysis of CD3⁺CD8⁺ cytotoxic T-cell isolated from whole liver of NEMO^∆hepa mice treated with +FMT (n = 5) or –FMT (n = 14) and NEMO^∆hepa/Nlrp6^−/− treated with +FMT (n = 4) or –FMT (n = 10), one-way ANOVA, Sidak’s multiple comparison test: NEMO^Δhepa vs. NEMO^Δhepa/Nlrp6^−/−, 95% CI 4.296 to 10.49, p < 0.0001; Nemo^Δhepa vs. Nemo^Δhepa + FMT, 95% CI 4.226 to 11.68, p < 0.0001; control group (–FMT) shared with (b), Fig. 3h. i Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels of NEMO^∆hepa/Tlr4^−/− mice treated with NEMO^∆hepa (n = 4) or NEMO^∆hepa/Nlrp6^−/− (n = 6) microbiota, unpaired t-test: AST: 95% CI −159.6 to 313.8, p = n.s., ALT: −711.9 to 713.5, p = n.s. j Flow cytometry analysis of mMDSC isolated from whole liver of NEMO^∆hepa/Tlr4^−/− mice treated with NEMO^∆hepa (n = 4) or NEMO^∆hepa/Nlrp6^−/− (n = 6) microbiota and respective controls (NEMO^∆hepa (n = 6) and NEMO^∆hepa/Tlr4^−/− mice (n = 8) without treatment), unpaired t-test: NEMO^Δhepa/Tlr4^−/− vs. NEMO^Δhepa 95% CI −7.038 to −0.09878, p = 0.045. All Data are presented as the mean ± standard error of the mean (SEM) and considered significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Individual data points represent biological replicates unless otherwise stated. Source data are provided as a Source data file.

Fig. 5. Akkermansia muciniphila supplementation ameliorates liver disease in NEMO^∆hepa mice.

a Inverse correlation (Spearman, 13 pairs) between mMDSC and abundance of A. muciniphila (AKK) in caeca of NEMO^∆hepa mice. b Experimental setup: 8-week-old NEMO^∆hepa mice were gavaged either with 2.1 × 10⁸  CFU AKK or sterile anaerobic PBS for 5 weeks. c Principle Coordinates Analysis (PCoA) plot of microbiota (cecal content, stool) of NEMO^∆hepa mice (n = 4) before and after AKK treatment. d Results of LDA effect size (LEfSe) of cecal microbiota of NEMO^∆hepa mice (n = 4) before and after AKK treatment at bacterial taxa level. e Representative pictures of immunofluorescence staining of MUC2 in colon of NEMO^∆hepa mice treated with PBS or AKK. Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. f Representative pictures of immunofluorescence staining of ZO-1 in ileum and colon of NEMO^∆hepa mice treated with PBS or AKK. Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. g Immunoblot analysis of colon protein extracts from NEMO^∆hepa mice either treated with PBS or AKK for 5 weeks for occludin and β-actin as loading control, representative of 2 experiments. h Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels of NEMO^∆hepa mice either treated with PBS (n = 6) or AKK (n = 8) for 5 weeks, unpaired t-test: ALT: NEMO^Δhepa AKK vs. NEMO^Δhepa PBS, 95% CI −835.1 to −312.4, p < 0.0001; AST: NEMO^Δhepa AKK vs. NEMO^Δhepa PBS, 95% CI −935.6 to −431.0, p < 0.0001. i Representative pictures of H&E-stained liver sections and immunofluorescence staining for CD11b in either PBS (n = 6) or AKK (n = 8) treated NEMO^∆hepa mice. Nuclei were counterstained with DAPI, representative of 2 experiments; Scale bar: 100 µm. j RT-qPCR analysis of inflammatory and fibrotic mRNA expression (TGFβ, Col1a1, Mcp-1, Tlr4) in whole liver of NEMO^∆hepa mice treated with PBS (n = 6) or AKK (n = 8) for 5 weeks, unpaired t-test, TGFβ: NEMO^Δhepa AKK (n = 8) vs. NEMO^Δhepa PBS (n = 6), 95% CI −0.4772 to −0.03676, p = 0.026; Col1a1: NEMO^Δhepa AKK (n = 8) vs. NEMO^Δhepa PBS (n = 6), 95% CI −1.023 to −0.1061, p = 0.019 Mcp-1: NEMO^Δhepa AKK (n = 7) vs. NEMO^Δhepa PBS (n = 6), 95% CI −0.5069 to −0.1172, p = 0.0048; Tlr4: NEMO^Δhepa AKK (n = 8) vs. NEMO^Δhepa PBS (n = 6), 95% CI −0.4238 to −0.07837, p = 0.008. All data are presented as the mean ± standard error of the mean (SEM) and considered significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), p < 0.001 (****), respectively. Individual data points represent biological replicates unless otherwise stated. Source data are provided as a Source data file.

Fig. 6. Hepatic bacterial 16s rDNA is increased in cirrhosis patients and shapes the hepatic transcriptomic landscape.

a Study outline: Snap frozen surgical liver tissue specimen were taken from 44 cirrhosis patients that underwent liver transplantation or 11 healthy controls. DNA and mRNA were isolated from the same tissue specimen and tissue region and subjected to 16 s rRNA gene amplicon sequencing or mRNA sequencing. b 16s rRNA gene copies/ng DNA determined by real time quantitative PCR in control (n = 12) and cirrhotic (n = 43) liver (Mann–Whitney-U-Test, p = 0.014). c Pathway activity based on mRNA-sequencing data and inferred by PROGENy computational pathway analysis in cirrhosis patients (n = 22) vs. healthy controls (n = 8). Correlation of 16S rRNA gene abundance with pathway activation (Spearman correlation, n = 30 pairs). d Computational Cell type enrichment analysis and correlation of calculated cell types with 16S rRNA gene abundance of cirrhotic patients (n = 22) and healthy controls (n = 8). e 16S rRNA gene abundance strongly correlates with the expression of immune checkpoint genes (Spearman correlation, n = 30 pairs, two-tailed). f Correlation of CTLA4 and g transcription factors involved in T-cell exhaustion (TOX, IRF4) with rRNA gene copies/ng genomic DNA (Spearman, n = 30 pairs, all 2-tailed, p < 0.0001). All Data are presented as the mean ± standard error of the mean (SEM) and considered significant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Source data are provided as a Source data file.