Phenotypic and epigenetic profiles of circulating NK cells in spontaneous HIV-1 controllers

Abstract

Background: NK cells play a key role in eliminating HIV-infected cells, but it is unclear whether there are specific NK cell receptor signatures in spontaneous HIV controllers.

Methods: We conducted a cross-sectional analysis of circulating NK cell phenotypes in people living with HIV (PLHIV), divided into spontaneous HIV controllers (HIC), normal progressors on antiretroviral therapy (non-HIC), and first-degree HIV-negative family members. Using supervised and unsupervised flow cytometry, we assessed NK cell markers and receptors. We performed an epigenetic analysis of H3K4me3 chromatin enrichment in NK cells from both HIC and non-HIC and measured IFNγ, Perforin and CD107a expression in NK cells upon stimulation with IL-2/IL-15, K562 cells, and IFNα. Additionally, we conducted a genome-wide association study (GWAS) and quantitative trait locus (QTL) mapping using data from HIC and non-HIC part of the 2000HIV study.

Findings: HIV controllers had higher levels of CD56^bright NK cells and increased expression of NKp46, NKp30, and DNAM-1. The genetic association between protective MHC class I alleles and the NK cell receptor KIR2DL2/3 supports a genetic predisposition to HIV control. Unsupervised clustering identified an HIV-induced NK cell population, separate from CMV-induced NK cells. Epigenetic analysis revealed greater H3K4me3 marks in genes involved in immune response pathways, including IFNα, IL-15, and IL-2. The memory-like NK cell subpopulation was characterised by elevated expression of NKG2C and ILT2, with reduced KIR2DL2/3 in HIC. These memory NK cells were more responsive to stimulation with IFNα, resulting in increased production of IFNγ in HIC.

Interpretation: These results suggest that spontaneous HIV control is associated with an NK cell memory phenotype, shaped by HIV infection, epigenetic modifications, and genetic factors.

Funding: The authors are part of the 2000HIV study, which is supported by ViiV Healthcare.

Keywords: HIV; Innate immune memory; NK-cells; Normal progressors on ART; Spontaneous HIV-controllers.

Fig. 1

Frequency and receptor expression of NK cell subpopulations are influenced by HIV infection and differ in individuals who can spontaneously control HIV. (a) Workflow of whole-blood NK cell characterisation performed in the people living with HIV (PLHIV) from the 2000HIV-Discovery cohort and family members from the 2000HIV-TRAINED Substudy as a group of healthy controls (HC). (b) Percentages CD56^brightCD16⁻, CD56^brightCD16⁺, CD56^dimCD16⁺, CD56^dimCD16⁻, and CD56^lowCD16⁺ NK cell subpopulations assessed in whole blood of PLHIV (n = 1098) of the 2000HIV study and HC (n = 43) of the 2000HIV-TRAINED Substudy. (c) Percentages of CD45⁺CD3⁻CD56⁺ live cells expressing exhaustion markers PD1, CD38, the activation marker HLA-DR, and chemokine receptors CXCR5 and CXCR3 in whole blood of PLHIV and HC. (d) Percentages of CD56^brightCD16⁻, CD56^brightCD16⁺, CD56^dimCD16⁺, CD56^dimCD16⁻, and CD56^lowCD16⁺ NK cell subpopulations assessed in whole blood of HIV controllers (HIC) (n = 92) and non-controllers (non-HIC) (n = 1006) of the 2000HIV-Study. (e) Percentages of CD45⁺CD3⁻CD56⁺ live cells expressing exhaustion markers PD1, CD38, the activation marker HLA-DR, and chemokine receptors CXCR5 and CXCR3 in whole blood of HIC and non-HIC. Statistical analysis was performed using a linear regression model, corrected for sex, age, season of inclusion, time to lab, and COVID-19 vaccination. Data represented as median with interquartile range. Whiskers extend to 1·5 × IQR. Exact p-values are shown.

Fig. 2

SNPs in the MHC locus that are associated with HIV control downmodulate the gene expression and plasma levels of KIR2DL2/3 and its ligand HLA-C. (a) A genome-wide association study (GWAS) was performed on 67 HIC and 1179 non-HIC from the 2000HIV study. (b) Regional association plot of SNPs in the MHC locus of chromosome 6 that are significantly associated with HIV control. (c) eQTL analysis based on the identified top SNPs rs2853950 and rs4713462 regulating the gene expression of the KIR2DL2/3 and its ligand HLA-C. (d) pQTL analysis of the top SNP rs2853950 regulating the protein levels of KIR2DL2 and KIR2DL3. Data represented as median with interquartile range. Whiskers extend to 1·5 × IQR. Exact p-values are shown.

Fig. 3

NK cells of non-controllers on ART exhibit increased maturation towards CD56dim expressing CD57 compared to healthy controls, while HIC have increased cytotoxicity receptors on CD56bright NK cells. (a) CD56⁺ NK cells were obtained through MACS isolation from PLHIV (n = 23), including both HIC (n = 13) and non-HIC (n = 10), and from HC (n = 21) part of 2000HIV-TRAINED Substudy. Flowcytometry was performed to determine the expression of CD57, CD94, DNAM, ILT2, KIR2DL2/3, NKG2A, NKG2C, NKG2D, NKp30, and NKp40 on CD3⁻CD56⁺NK cells. (b) Heatmap representing percentage of CD56^brightCD16⁻, CD56^brightCD16⁺, CD56^dimCD16⁺, CD56^dimCD16⁻, and CD56^lowCD16⁺ NK cells subpopulations expressing CD57, CD94, DNAM, ILT2, KIR2DL2/3, NKG2A, NKG2C, NKG2D, NKp30, and NKp40 between PLHIV and HC (c) and HIC vs. non-HIC. Statistical analysis was performed using a linear regression model, corrected for sex and age. Data represented as estimates. Exact p-values are shown.

Fig. 4

Unsupervised analysis of NK cells with FlowSOM allows identification of differential metaclusters in PLHIV and healthy controls. (a) CD56⁺ NK cells were obtained through MACS isolation from PLHIV (n = 23), including both HIC (n = 13) and non-HIC (n = 10) and HC (n = 21) part of the 2000HIV-TRAINED substudy. Flow cytometry was performed to determine the expression of CD57, CD94, DNAM, ILT2, KIR2DL2/3, NKG2A, NKG2C, NKG2D, NKp30, and NKp40 on CD3⁻CD56⁺NK cells and unsupervised analysis with dimension reduction using UMAP and FlowSOM metaclustering was employed. (b) Representation of marker expression within NK cells on a UMAP scale (representative data of one individual with HIV). (c) Comparison of M1, M3 percentages relative to total NK cells (CD45⁺CD3⁻CD56⁺) between HC and PLHIV. (d) Expression of CD16, CD56, CD57, CD94, DNAM, ILT2, KIR2DL2/3, NKG2A, NKG2C, NKG2D, NKp30, and NKp40 in M1 and M3 as well as (e) M7 and M9 among HC or PLHIV. MFI depicted as log(10) values. Data are represented as median with interquartile range. Whiskers extend to 1·5 × IQR. Significance was tested using a Wilcoxon rank-sum test. Exact p-values are shown.

Fig. 5

CMV and HIV induce distinct memory NK cells with differential expression of KIR2DL2/3. (a) Overview of PLHIV and HC of the 2000HIV-TRAINED substudy who are CMV-seropositive (n = 23 PLHIV, n = 10 HC) or -seronegative (n = 11 HC). (b) Average percentage of CD3⁻CD56⁺ NK cells within metaclusters among CMV−HIV−, CMV+HIV−, and CMV+HIV+ individuals. (c) Percentage of CD3⁻CD56⁺ NK cells within M1 and M3 in healthy controls stratified as CMV− and CMV+ individuals. (d) UMAP representation of M1 and M3 from a CMV−HIV−, a CMV+HIV−, and a CMV+HIV+ individual, respectively (representative data of one individual of each condition). (e) Expression of CD16, CD56, CD57, CD94, DNAM, ILT2, KIR2DL2/3, NKG2A, NKG2C, NKG2D, NKp30, and NKp40 among M1 vs. M3 of CMV+PLHIV. MFI is depicted as log(10) values. Data are represented as median with interquartile range. Whiskers extend to 1·5 × IQR. Significance was tested using a Wilcoxon rank-sum test. Exact p-values are shown.

Fig. 6

NK cells from HIC show memory-like features with differential receptor expression of NKG2C, ILT2 and KIR2DL2/3 and are more responsive to IFNα stimulation. (a) Overview of HIC and non-HIC from who flowcytometry data CD3⁻CD56⁺ NK cell data were available. (b) Frequencies of CD3⁻CD56⁺ NK cells between HIC (n = 13) and non-HIC (n = 10) within M1 and M3. Heatmap comparing the MFI expression of NK cell markers within M1, M3, and M8 in HIC and non-HIC. (c) Expression of CD16, CD56, CD57, CD94, DNAM, ILT2, KIR2DL2/3, NKG2A, NKG2C, NKG2D, NKp30, and NKp40 among HIC vs. non-HIC in M3. MFI depicted as log(10) values. Boxes represent interquartile ranges (IQR) with medians; whiskers show 1·5 × IQR. (d) Heatmap showing the normalised H3K4me3 deposition at promoters of CD3⁻CD56⁺ sorted NK cells from HIC (n = 8) and non-HIC (n = 5). (e) Functional enrichment of up- and down-regulated regions in NK cells from HIC compared to non-HIC using the Hallmark database. (f) PBMCs of 4 HIC and 4 non-HIC were activated for 10 h with IL-2 and IL-15 or left unstimulated (RPMI) and afterwards co-cultured with either MHC-deficient K562 cell lines for 4 h or activated with IFNα. Flowcytometry was performed afterwards for intracellular (IFNγ and Perforin) and extracellular (CD107a) functional markers. Memory NK cells were gated for NKG2C+ILT2+KIR2− NK cells. (g) Expression of functional markers IFNγ, perforin and CD107a in memory NK cells between HIC and non-HIC. Data are represented as median with interquartile range. Whiskers extend to 1·5 × IQR. Significance was tested using a Wilcoxon rank-sum test. Exact p-values are shown.

Fig. 7

Overview–distinct genetic, phenotypic, and functional features of NK cells in HIC. (a) SNPs associated with HIV control are enriched near HLA-C and MICA/HLA-G loci on chromosome 6, with potential implications for reduced KIR2DL2/3 expression. (b) Phenotypic profiling of NK cells in HIC reveals an increase in Cd56bright cells and activating receptor expression (NKp30, NKp46, DNAM-1) (top). Memory-like NKG2C⁺ NK cells differ based on viral exposure: CMV-induced cells show increased KIR2DL2/3, whereas HIV-induced cells in HIC lack this upregulation (bottom). (c) NK cells in HIC exhibit epigenetic reprogramming (e.g., increased H3K4me3 at effector loci) and enhanced IFNα signalling, leading to a heightened IFNγ response upon stimulation, indicative of trained immunity. HIC, Spontaneous HIV-1 Controller; SNP, Single Nucleotide Polymorphism; CMV, Cytomegalovirus; IFN, Interferon; NKp30, Natural Cytotoxicity Receptor 3; NKp46, Natural Cytotoxicity Receptor 1; NKG2C, Natural Killer Group 2C receptor; ILT2, Immunoglobulin-like transcript 2; KIR2DL2/3, Killer-cell immunoglobulin-like receptor 2DL2/3; DNAM-1, DNAX accessory molecule-1; H3K4me3, Histone 3 lysine 4 trimethylation; STAT, Signal Transducer and Activator of Transcription. Figure created with Biorender.com.