Plasma proteomic signatures of liver steatosis and fibrosis in people living with HIV: a cross-sectional study

Abstract

Background: Insights into the mechanisms driving metabolic dysfunction-associated steatotic liver disease (MASLD) in people living with HIV (PLHIV) remain limited. Plasma proteomics holds promise for biomarker discovery and the elucidation of biological mechanisms.

Methods: We performed cross-sectional analyses on data from 1036 virally suppressed PLHIV using antiretroviral treatment (ART) from the Dutch multi-centre 2000HIV cohort. Participants underwent transient elastography to assess liver steatosis (controlled attenuation parameter (CAP) ≥263 dB/m) and -fibrosis (liver stiffness measurement (LSM) ≥7.0 kPa). Plasma protein concentrations (n = 2367) (Olink® Explore Panel) were compared between PLHIV with vs. without liver steatosis and PLHIV with vs. without fibrosis. Enriched pathways (using GO, KEGG and Reactome libraries) and correlations with clinical characteristics were assessed, and analyses were stratified by BMI category. In addition, concentrations of 242 proteins were compared between individuals ("controls") with and without liver steatosis (ratio of methylene:methylene and water >5.6% on magnetic resonance spectroscopy) from a separate cohort (300-OB), all having a BMI >26 kg/m².

Findings: Steatosis and fibrosis were associated with 67/2367 (2.2%) and 17/2367 (0.7%) differentially expressed proteins (DEP), respectively, enriched in mostly metabolic pathways. Immunoglobulin superfamily member 9 (IGSF9) was amongst the top DEP associated with both steatosis and fibrosis. Stratifying by BMI revealed 8/2367 DEP associated with steatosis in lean- and 12/2367 DEP in overweight/obese individuals, with two shared DEP (IGSF9 and GHR). Conversely, protein signatures of overweight/obese PLHIV (32/242 DEP) and overweight/obese HIV-uninfected individuals (32/242 DEP) exhibited substantial overlap with 16 shared DEP. Notably, DEP correlated with HIV characteristics in lean individuals but not in overweight/obese PLHIV.

Interpretation: Lean and overweight/obese PLHIV exhibit distinct proteomic signatures associated with liver steatosis, with the former being more strongly correlated with HIV-specific factors and ART. In addition, we identified a protein, IGSF9, strongly related to liver fibrosis and steatosis across BMI categories.

Funding: The 2000HIV study is funded by ViiV Healthcare.

Keywords: Fibrosis; HIV; MASLD; Proteomics; Steatosis.

Fig. 1

a and b Differentially expressed proteins in PLHIV with steatosis compared to PLHIV without steatosis. Fig. 1a shows a volcano plot with the results of the DE analysis for the discovery cohort. For readability of the figure, only validated proteins with a logFC < −0.25 or >0.25, or FDR-adjusted p-value <0.00001 in the discovery cohort are annotated. Fig. 1b shows the expression levels (in NPX) of the top 10 proteins by steatosis grade. p-values are derived from comparison by parametric ANOVA tests between each fibrosis grade with correction for multiple testing by FDR.

Fig. 2

a and b Differentially expressed proteins in PLHIV with fibrosis compared to PLHIV without fibrosis. a) Volcano plot showing the results of the DE analysis in the discovery cohort with validated proteins annotated. All validated proteins (i.e. FDR-adjusted p-value <0.05 in the discovery- and raw p-value <0.05 in the validation cohort) are annotated. b) Expression levels (in NPX) of the top 10 proteins (as defined by log fold change in discovery cohort) by steatosis grade. p-values are derived from comparison by parametric ANOVA tests between each fibrosis grade with correction for multiple testing by FDR.

Fig. 3

a and b Correlation between the proteins associated with steatosis (a) and fibrosis (b), with SNPs of interest, markers of inflammation, -immune activation, -cholesterol metabolism, HIV-characteristics, ART exposure, and cardiometabolic diseases. The colours show the direction and strength of the Spearman correlations, and asterixis refer to the FDR-adjusted p-value: ∗ = p-value 0.01–0.05; ∗∗ = p-value 0.001–0.01; ∗∗∗ = p-value <0.001.

Fig. 4

a–c Differentially expressed proteins in lean and overweight/obese PLHIV with steatosis compared to PLHIV without steatosis. a) Volcano plot showing the results of the DE analysis lean PLHIV, with only validated DEP (i.e. FDR-adjusted p-value <0.05 in the discovery- and raw p-value <0.05 in the validation cohort) annotated. b) Volcano plot showing the results of the DE analysis in overweight and obese PLHIV, with only validated DEP (i.e. FDR-adjusted p-value <0.05 in the discovery- and raw p-value <0.05 in the validation cohort) annotated. c and d) Correlation between the DEPs associated with steatosis identified in lean PLHIV (left) and overweight/obese PLHIV (right) and SNPs of interest, markers of inflammation, -immune activation, -cholesterol metabolism, HIV-characteristics, ART exposure, and cardiometabolic diseases. The colours show the direction and strength of the Spearman correlations, and asterixis refer to the significance level: ∗ = p-value 0.01–0.05; ∗∗ = p-value 0.001–0.01; ∗∗∗ = p-value <0.001. Abbreviations: D-drug = dideoxynucleoside analogue drugs (includes D4T, DDI, and DDC), D4T, stavudine; cART, combination antiretroviral treatment; NtRTI, nucleotide reverse transcriptase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; VL, HIV-1 RNA viral load; immunological non resp, immunological non-responder; IDV, indinavir; RAL, raltegravir, T2DM, type 2 diabetes mellitus; DDI, didanosine; DTG, dolutegravir; INSTI, integrase strand transfer inhibitor; DDC, zalcitabine; PI, protease inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; HBV, hepatitis B virus; HAV, hepatitis A virus; HCV, hepatitis C virus.

Fig. 5

a–c DEPs associated with steatosis identified in overweight/obese PLHIV and controls. a) Volcano plot showing the results for overweight/obese controls and b) Volcano plot showing the results for controls. All proteins with an FDR-adjusted p-value <0.05 are annotated. Figure c shows the differences and similarities of the proteomic signature of steatosis between PLHIV and controls.