Abrupt and altered cell-type specific DNA methylation profiles in blood during acute HIV infection persists despite prompt initiation of ART

Abstract

HIV-1 disrupts the host epigenetic landscape with consequences for disease pathogenesis, viral persistence, and HIV-associated comorbidities. Here, we examined how soon after infection HIV-associated epigenetic changes may occur in blood and whether early initiation of antiretroviral therapy (ART) impacts epigenetic modifications. We profiled longitudinal genome-wide DNA methylation in monocytes and CD4+ T lymphocytes from 22 participants in the RV254/SEARCH010 acute HIV infection (AHI) cohort that diagnoses infection within weeks after estimated exposure and immediately initiates ART. We identified monocytes harbored 22,697 differentially methylated CpGs associated with AHI compared to 294 in CD4+ T lymphocytes. ART minimally restored less than 1% of these changes in monocytes and had no effect upon T cells. Monocyte DNA methylation patterns associated with viral load, CD4 count, CD4/CD8 ratio, and longitudinal clinical phenotypes. Our findings suggest HIV-1 rapidly embeds an epigenetic memory not mitigated by ART and support determining epigenetic signatures in precision HIV medicine. Trial Registration: NCT00782808 and NCT00796146.

Conflict of interest statement

I have read the journal’s policy and the authors of this manuscript have the following competing interests: LCN has served as an advisory board member for Abbvie, Cytodyn and ViiV for work unrelated to this project. All other authors declare no competing interests.

Figures

Fig 1. Monocyte cell type-specific differentially methylated loci associated with acute HIV infection.

(a) Diagram of experimental design. Created with BioRender.com (b) Manhattan plot of differentially methylated loci associated with AHI in monocytes displayed across chromosomes. P values transformed using -log10(P) (c) Volcano plot displaying HUGO gene symbols and related CpG ID (cg#) for top hypomethylated and hypermethylated sites. Difference in DNA methylation displayed as delta beta values plotted against P values transformed using -log10(P). NS: non-significant. (d-q) Plots demonstrating changes in DNA methylation in AHI Fiebig I (red), Fiebig II (blue) Fiebig III-V (purple) compared to HIV- (green) at single CpGs sites related to specific genes or intergenic regions of the genome displayed in bold italic above CpG ID (cg#). Significance determined comparing HIV- vs. AHI and using FDR adjusted P-values <0.05.

Fig 2. Associations of monocyte cell type-specific DNA methylation with viral load, CD4 count, and CD4/CD8 ratio.

Scatter plots showing the correlation between (a-e) plasma viral load, (f-g) pre-ART CD4 count, and (h-o) pre-ART CD4/CD8 ratio and DNA methylation levels at single CpGs sites related to specific annotated protein-coding genes displayed in bold italic above CpG ID (cg#) in monocyte cells during AHI at entry. P-values were calculated with the Spearman correlation test.

Fig 3. Cell-type independent differentially methylated loci associated with acute HIV infection.

(a) Diagram of experimental design displaying comparison utilized to identify DML in CD4+ T cells associated with AHI. Created with BioRender.com (b) Volcano plot displaying gene symbol and CpG ID (cg#) for top hypomethylated and hypermethylated sites associated with AHI in CD4+ T lymphocytes. Difference in DNA methylation displayed as delta beta values plotted against P values transformed using -log10(P). (c) Venn diagram displaying the number of overlapping DML and cell-type specific DML identified in CD4+ T cells and monocyte cells (d-s) Plots demonstrating changes in DNA methylation in AHI Fiebig I (red dots), Fiebig II (blue dots) Fiebig III-V (purple dots) participants compared to HIV- (green dots) at single CpGs sites related to annotated protein-coding genes displayed in bold italic above CpG ID (cg#) for CD4+ T lymphocytes and monocytes. Significance determined comparing HIV- vs. AHI and using FDR adjusted P-values <0.05.

Fig 4. Early initiation of ART during acute infection minimally impacts monocyte cell type-specific differentially methylated loci associated with acute HIV infection.

(a) Diagram of longitudinal experimental design utilized to identify DML-related to ART treatment in purified monocytes. Created with BioRender.com (b) Venn diagram displaying the overlap of DML between DNA methylation sites identified in monocytes related to AHI and DNA methylation sites identified in monocytes that changed before and after early initiation of ART during AHI. (c) Heatmap showing the unsupervised clustering of DML at annotated protein-coding genes related to interferon at pre-ART (aqua color) and post-ART (light blue) timepoints for AHI participants staged at Fiebig I (red color), II (blue color), or III-V (purple color). Dendrogram shown above. Colors in the heatmap indicate CpG methylation levels (blue to red: low to high methylation levels). (d-o) Plots displaying longitudinal DNA methylation levels of AHI participants at Pre-ART and Post-ART timepoints by Fiebig stage. Red dots represent individuals in Fiebig I. Significance determined comparing repeated measures comparison of Pre-ART vs. Post-ART and using FDR adjusted P-values <0.05.

Fig 5. Differentially methylated loci identified in monocyte cells that persist in acute and chronic HIV infection despite ART.

(a-l) Plots of durable differentially methylated loci in in AHI Fiebig I (red), Fiebig II (blue) Fiebig III-V (purple) at Pre-ART and Post-ART time points compared to HIV- (green) at single CpGs sites related to specific genes or intergenic regions of the genome displayed in bold italic above CpG ID (cg#). Validation of these same durable differentially methylated loci in monocytes displayed for chronic HIV infection at Pre-ART (aqua) and Post-ART (light blue) along with the Berlin patient (star). Significance determined comparing HIV- vs. AHI pre-ART, AHI post-ART, CHI pre-ART, and CHI post-ART and using FDR adjusted P-values

Fig 6. Early initiation of ART during acute infection impacts interferon-related gene programs in monocytes.

(a) Volcano plot showing gene symbol for top differentially expressed genes in monocytes comparing pre-ART and post-ART time points. Log fold change plotted against transformed -log10 P value. (b-m) Plots showing differentially expressed interferon-related genes in monocyte cells by Fiebig stage comparing pre-ART and post-ART timepoints. Fiebig I displayed in red, Fiebig II displayed in blue, and Fiebig III-V displayed in purple.

Fig 7. Monocyte gene expression associates with DNA methylation levels and clinical parameters during acute HIV infection.

(a-d) Correlation matrix plots showing the correlation coefficient for associations between gene expression levels, CD4 count, CD4/CD8 ratio, viral load, and DNA methylation levels at specific genomic loci. Significant correlations displayed as solid colored boxes (red displayed for negative correlations and blue for positive correlations). Correlation coefficient displayed.

Fig 8. DNA methylation related to the IRF7 gene in monocytes relates to CD4 fold change at 96 weeks following the initiation of ART during acute HIV infection.

(a-d) Line plots of longitudinal CD4 count in participants by Fiebig stage from week 0 (pre-ART) through post-ART time points out to week 96. (e) Correlation plot displaying relationship between IRF7 DNA methylation levels in monocyte cells assessed at baseline and CD4 fold change calculated from week 0 to week 96. Fiebig stage of participant displayed: Fiebig I displayed in red, Fiebig II displayed in blue, and Fiebig III-V displayed in purple.

Fig 9. DNA methylation features associated with AHI in monocytes and CD4+ T lymphocytes predict favorable clinical phenotypes and neurocognitive performance following ART.

(a-b) Feature importance plots of gradient boosting machine learning and logistic regression models with interaction features for predicting favorable clinical phenotype at week 96 post-ART utilizing cell-type specific DNA methylation measures and clinical parameters (estimated days of infection, viral load, CD4 count, CD8 count, and CD4/CD8 ratio) during viral establishment in AHI prior to ART initiation. (c) Receiver operating characteristic curve showing performance of classification models (area under curve AUC) in discovery AHI DNA methylation feature set. (e-f) Feature importance plots of gradient boosting machine learning and logistic regression models with interaction features for predicting neurocognitive performance trajectory at week 96 post-ART. (g) Receiver operating characteristic curve showing performance of classification models.