Transcriptional Network Analysis Implicates Altered Hepatic Immune Function in NASH development and resolution - PubMed (original) (raw)

doi: 10.1038/s42255-019-0076-1. Epub 2019 Jun 14.

Luisa Vonghia 2 3, Denis A Mogilenko 1, An Verrijken 3 4, Olivier Molendi-Coste 1, Sébastien Fleury 1, Audrey Deprince 1, Artemii Nikitin 1, Eloïse Woitrain 1, Lucie Ducrocq-Geoffroy 1, Samuel Pic 1, Bruno Derudas 1, Hélène Dehondt 1, Céline Gheeraert 1, Luc Van Gaal 3 4, Ann Driessen 5, Philippe Lefebvre 1, Bart Staels 1, Sven Francque 2 3, David Dombrowicz 1

Affiliations

Transcriptional Network Analysis Implicates Altered Hepatic Immune Function in NASH development and resolution

Joel T Haas et al. Nat Metab. 2019 Jun.

Erratum in

Abstract

Progression of fatty liver to non-alcoholic steatohepatitis (NASH) is a rapidly growing health problem. Presence of inflammatory infiltrates in the liver and hepatocyte damage distinguish NASH from simple steatosis. However, the underlying molecular mechanisms involved in the development of NASH remain to be fully understood. Here we perform transcriptional and immune profiling of NASH patients before and after lifestyle intervention (LSI). Analysis of liver microarray data from a cohort of patients with histologically assessed NAFLD reveals a hepatic gene signature, which is associated with NASH and is sensitive to regression of NASH activity upon LSI independently of body weight loss. Enrichment analysis reveals the presence of immune-associated genes linked to inflammatory responses, antigen presentation and cytotoxic cells in the NASH-linked gene signature. In an independent cohort, NASH is also associated with alterations in blood immune cell populations, including conventional dendritic cells (cDC) type 1 and 2, and cytotoxic CD8 T cells. Lobular inflammation and ballooning are associated with the accumulation of CD8 T cells in the liver. Progression from simple steatosis to NASH in a mouse model of diet-driven NASH results in a comparable immune-related hepatic expression signature and the accumulation of intra-hepatic cDC and CD8 T cells. These results show that NASH, compared to normal liver or simple steatosis, is associated with a distinct hepatic immune-related gene signature, elevated hepatic CD8 T cells, and altered antigen-presenting and cytotoxic cells in blood. These findings expand our understanding of NASH and may identify potential targets for NASH therapy.

PubMed Disclaimer

Conflict of interest statement

Competing interests: BS and SF are consultants for Genfit S.A. SF and LV are consultants for Inventiva. All other authors have nothing to declare.

Figures

Figure 1

Figure 1

Identification of hepatic transcriptomic signature of NASH. a. WGCNA performed with 11784 transcripts in the liver in patients with or without histologically proven NASH (n=155 patients, see Supplementary Table 1). Clustering of co-expressed genes in 9 gene modules. b. Number of transcripts in each gene module. c. Overall transcriptional regulation of gene modules upon RYGB (n=21 patients), LSI in responders (n=10 patients), and LSI in non-responders (n=10 patients) at one year follow-up compared to baseline (see Supplementary Table 3 and 4). P-values are calculated by mean-rank gene set test using geneSetTest function as described in detail in methods. d. Volcano plots of average log2 fold-changes versus P-values (paired moderated t-test using limma package) of all transcripts (gray dots) and transcripts from gene module “blue” (red dots) in RYGB patients (n=21 patients), LSI responders (n=10 patients), and LSI non-responders at one year follow-up compared to baseline. e. Top hallmark pathways enriched in gene module “blue” (n=786 transcripts), calculated using GSEA software as described in detail in methods. f. Venn diagrams with transcripts in gene module “blue” down-regulated (P<0.05 by moderated paired t-test using limma package) in three groups of patients at follow-up versus baseline, immune-related genes are shown. g. Spearman correlations between NASH activity index and hepatic expression levels of genes in gene module “blue” in the 155 patients at baseline, top genes with maximal positive correlation coefficients are shown. h. CXCL9, CXCL10, and LYZ expression (by microarray) in NAFL (n=22 patients) and NASH (n=106 patients) patients at baseline. Data are presented as median with 1st and 3rd quartiles as the box edges. *P<0.05, **P<0.01 by unpaired two-sided Mann–Whitney U test.

Figure 2

Figure 2

Correlations between blood immune cell populations, disease activity in NASH, and genes in module “blue”. a. Hierarchical clustering of correlation coefficients in 38 patients (see Supplementary Table 5) between proportions of blood immune cell populations and histological liver parameters (Spearman's correlation), T2D-associated parameters (Pearson correlation) and systemic inflammation markers (Pearson correlation). Asterisks indicate P< 0.05 for the given correlation. b. Pearson correlations between selected immune cell populations in blood from a subset of 29 patients (see Supplementary Table 5) and hepatic expression levels of genes from module “blue”. NAS: NAFLD Activity Score. AI: Activity Index

Figure 3

Figure 3

A diet-induced NASH alters cDC and CD8 T cells and inflammation in the liver. Male C57BL/6J mice were fed conventional diet (CD) or NASH-diet (ND) during 24 weeks (see Supplementary Information). a. Representative flow cytometry plots of cDC in the liver: proportions of XCR1+ and CD172a+ of total cDC are shown (n=8 mice CD; n=6 mice ND). b. CD172a+ cDC2 cells as proportion of CD45+ cells (n=8 mice CD; n=6 mice ND). c. XCR1+ cDC1 cells as proportion of CD45+ cells (n=8 mice CD; n=6 mice ND). d. Ratio of cDC1/cDC2 cells (n=8 mice CD; n=6 mice ND). e. Representative flow cytometry plots of TCRβ+ T cells in the liver: proportions of CD4+ and CD8+ of total TCRβ+ T cells are shown. f. Proportion of CD8+ T cells of CD45+ immune cells in the liver (n=9 mice CD; n=20 mice ND). g. qPCR analysis of inflammatory gene expression in mouse livers. (n=9 mice CD; n=20 mice ND). Data are shown as mean ± SEM. Statistical significance of differences between groups are analyzed by unpaired two-sided t-test (*P < 0.05, **P < 0.01, ***P < 0.001, NS – not significant).

Figure 4

Figure 4

NASH and T2D alter activity of cytotoxic CD8 T cells. a. Proportion of IFNγ+ and TNFα+ CD8 T lymphocytes in blood from patients with/without NASH and/or T2D (see Supplementary Table 5). Groups: n=7 patients no NASH no T2D, n=6 NASH no T2D, n=7 patients no NASH T2D, n=7 patients NASH T2D. Data are shown as median with 1st and 3rd quartiles. Statistical significance of differences between groups were analyzed by unpaired two-way ANOVA for effects of NASH and T2D followed by Tukey’s post-hoc test. b. Representative flow cytometric plots and c. proportions of perforin, granzyme A and B expression in blood CD8 T lymphocytes from patients. Groups: n=7 patients no NASH no T2D, n=6 NASH no T2D, n=7 patients no NASH T2D, n=7 patients NASH T2D. Data are shown as median with 1st and 3rd quartiles. Statistical significance of differences between groups were analyzed by moderated paired t-test or two-way ANOVA (for effects of NASH and T2D) followed by Tukey's post-hoc test.

Figure 5

Figure 5

Hepatic CD8 T lymphocytes correlate with lobular inflammation, ballooning, and transcriptomic signature of NASH. a. Representative immunostaining for CD8 (red) with haematoxylin counterstaining on liver biopsies from patients with/without NASH and/or T2D. b. Quantification of CD8-positive cells per mm2. No-NASH No-T2D n = 10, NASH No-T2D n = 10, No-NASH T2D n = 7, NASH T2D n = 9. Data are shown as median with 1st and 3rd quartiles. c. Localization of CD8 T lymphocytes (red) near immune infiltrates, steatosis, and ballooned hepatocytes (indicated by arrows) in the liver from NASH patient. Scale bar is 50μm. d. Correlations between hepatic CD8 T lymphocyte number and histological features in the liver (n=36). e. Pearson correlations and −log10 P-values between hepatic CD8 T lymphocyte and expression levels of gene from module “blue” (n=29, Supplementary Table 5). Statistical significance of differences between groups were analysed by unpaired two-way ANOVA (for effects of NASH and T2D) followed by Tukey's post-hoc test (*P < 0.05).

References

    1. Haas JT, Francque S, Staels B. Pathophysiology and Mechanisms of Nonalcoholic Fatty Liver Disease. Annu Rev Physiol. 2016;78:181–205. doi: 10.1146/annurev-physiol-021115-105331. - DOI - PubMed
    1. Luyckx FH, Lefebvre PJ, Scheen AJ. Non-alcoholic steatohepatitis: association with obesity and insulin resistance, and influence of weight loss. Diabetes Metab. 2000;26:98–106. - PubMed
    1. Brunt EM. Pathology of nonalcoholic fatty liver disease. Nat Rev Gastroenterol Hepatol. 2010;7:195–203. doi: 10.1038/nrgastro.2010.21. - DOI - PubMed
    1. Brunt EM, et al. Nonalcoholic fatty liver disease (NAFLD) activity score and the histopathologic diagnosis in NAFLD: distinct clinicopathologic meanings. Hepatology. 2011;53:810–820. doi: 10.1002/hep.24127. - DOI - PMC - PubMed
    1. Bedossa P, et al. Histopathological algorithm and scoring system for evaluation of liver lesions in morbidly obese patients. Hepatology. 2012;56:1751–1759. doi: 10.1002/hep.25889. - DOI - PubMed

Publication types

MeSH terms

LinkOut - more resources