Neutralization activity in chronic HIV infection is characterized by a distinct programming of follicular helper CD4 T cells

Abstract

A subset of people living with HIV (PLWH) can produce broadly neutralizing antibodies (bNAbs) against HIV, but the lymph node (LN) dynamics that promote the generation of these antibodies are poorly understood. Here, we explored LN-associated histological, immunological, and virological mechanisms of bNAb generation in a cohort of anti-retroviral therapy (ART)-naïve PLWH. We found that participants who produce bNAbs, termed neutralizers, have a superior LN-associated B cell follicle architecture compared with PLWH who do not. The latter was associated with a significantly higher in situ prevalence of Bcl-6^hi follicular helper CD4 T cells (TFH), expressing a molecular program that favors their differentiation and stemness, and significantly reduced IL-10 follicular suppressor CD4 T cells. Furthermore, our data reveal possible molecular targets mediating TFH- B cell interactions in neutralizers. Together, we identify cellular and molecular mechanisms that contribute to the development of bNAbs in PLWH.

Keywords: HIV; TFH; broadly neutralizing antibodies; lymph node.

Fig. 1:. Cross-neutralization is associated with GCs displaying preserved features of maturation.

A) Diagram depicting the strategy employed for neutralization screening (left) as well as the distribution of cross-neutralization breadth in the population of study, defined as the percentage of the number of isolates recognized with an ID₅₀>40 out of 6 isolates (pie chart) and 20 isolates (bar graph) tested among study participants. B) Representative confocal images showing the distribution of CD20 (cyan) and Ki67 (magenta) in NNs and Ns and total B-cell follicle area quantification (CD20^hi/dim) in control tissues, NNs and Ns. Each circle represents a follicle, and data are for all tissues combined (***p<0.001). Original magnification x40 (NA 1.3), scale 500um. C) Box plots showing the number of total cells and GC/F ratios (areas) in NN and Ns as measured by quantitative imaging analysis and Histocytometry. Data for each follicle represent total counts of JOJO-1+ events (total cells) normalized to the total area (mm²) of follicle screened. Each circle represents an individual follicle, and data shown are for all follicles combined. p=0.4249, **p=0.00115. D) Representative confocal images showing the distribution of FDC (red) in B cell follicles (CD20^dim/hi areas, cyan) in NNs vs Ns (original magnification x40, scale 100um), and box plot summarizing the FDC areas (mm²) in NNs and Ns (pooled data) as calculated using data from quantitative imaging analysis and Histocytometry (*p=0.0352). Each circle represents and individual follicle, and data shown are for all follicles combined. p values were considered significant if p≤0.05.

Fig. 2:. Cross-neutralization is associated with an expansion of the LN CD4 T cell memory compartment and a distinct TFH cell phenotypic profiling.

A) Flow cytometry plots depicting the frequencies of memory CD4 T cell subsets in NNs and Ns within the CD3^hiCD4^hi gated populations. B) Bar graphs showing the relative frequencies of CD4 T cell subsets in Ns and NNs (left panel) and the positive association between the % of T_CM CD4 T cells and microneutralization ID₅₀ titers as assayed against a 20 HIV isolate panel. C) Cumulative flow cytometry plots showing the frequencies of receptors TIGIT, ICOS as well as CD57 epitope and CXCR3 within the PD-1^hi CXCR5^hi TFH cell compartment in NNs and Ns, and frequencies of combined CXCR3 CD57 expression. TIGIT **p=0.0075; ICOS, *p=0.0160; CD57 *p=0.0225, CXCR3 * p=0.0420 (Mann Whitney U test); NN CXCR5^loCD57^hi vs NN CXCR3^hi CD57^lo *p<0.0001 (ANOVA). D) Representative confocal images of GCs in NNs and N showing the distribution of TFH (CD4: green, PD-1^hi: red) within the Ki67 follicular areas (magenta) and pooled histocytometry quantitation of total GC-CD4, GC Bcl-6^hi Tfh and GC Bcl-6^hi Tfh/CD4 ratios in NNs (n=6, blue circles represent all follicles counted in those six individuals) vs Ns (n=6, red circles represent all follicles counted as for NNs). The numbers of cells per mm² of follicular area are shown. p values are *p= 0.0198 (ANOVA), **p= 0.003 and ****p<0.0001 (Mann-Whitney U test). Original magnification x40, scale bar = 50um. E) Regression plots showing the association between bulk TFH and Bcl-6^hi TFH cells (normalized cell counts adjusted for mm² of area screened) as quantified by confocal imaging analysis and histocytometry, and the breadth of neutralization expressed as the number of HIV isolates recognized out of the 20 isolates tested.

Fig.3:. TFH cells are characterized by a molecular profile favoring their differentiation and longevity in cross-neutralization participants.

UMAP projections of total LNMCs (A) and CD4 T cells (B), with color-coded cell populations and CD4 T cell subpopulations identified and their percentages in NNs vs Ns. C-D) Volcano plots and dot plots of differentially expressed genes (DEGs) and normalized enrichment scores (NES) of GSEA for selected signaling and downstream targets are shown. The DEGs were calculated using the MAST (Hurdle model based) test via the FindMarkers function using the Seurat package in R. The p-values were corrected using the Benjamini-Hochberg method and genes meeting an adjusted p< 0.05 and abs(log₂FC)>0.5 are highlighted on the volcano plot. Blue dots and red dots in volcano plots represent genes that are significantly downregulated or upregulated in neutralizers respectively. E) Representative confocal images showing FOXP3^hi CD4 T cells in extrafollicular and follicular areas in NNs vs N. FOXP3^hi CD4 T cells (white arrows) were defined by means of concurrent CD4 (green) and FoxP3 (red) expression within CD20 (cyan) and Ki67 (yellow) double positive areas. Original magnification 40x. Scale bars are 100um, except for the zoomed in images corresponding to the white dotted enclosures which are 20um. Box plots showing the normalized per imaged area numbers of FOXP3^hi CD4 T cells within B cell follicles, as quantified by Histocytometry (***p=0.00069, ANOVA). F) Box plots of CD25^hiFoxP3^hi and IL-10^hiCD25^hi FoxP3^hi CD4 T cells as measured by Histocytometry. Each blue and red circle represents a single follicle, and data presented are for all follicles measured in NNs (blue) and Ns (red) respectively except for extrafollicular and whole tissue data where each participant tissue is represented by a circle of a different color. ****p<0.0001; ****p<0.0001 (Mann-Whitney U test).

Fig. 4:. Higher viral evolution is associated with lower in situ prevalence of actively transcribed virus in cross-neutralization participants.

A) Plasma viral loads (VLs) in all study participants and association of VLs with neutralization breadth (number of isolates recognized out of the 20 isolates tested). B) Representative confocal images showing the distribution of viral RNA (yellow) on FDC networks (red) within B-cell follicles (CD20: cyan) and box plots summarizing the number of RNA+ cells and vRNA+ virions as measured by RNAscope in situ hybridization, Histocytometry (cells) and spot analysis (virions) in NNs and Ns. Images were captured at 40x magnification, with 1% zoom and scale is 100um and 15um (zoomed in detail). Each blue and red circle represents a single follicle, and data presented are for all follicles measured in NNs (blue) and Ns (red) tissues *p=0.0232 (Mann-Whitney U test); **p=0.0704 (ANOVA). Graphs showing the number of total RNA virions (counts) in individual NN and N participants are also shown. Each color represents a different participant tissue and individual circles the B-cell follicles in that tissue. C) Association of total RNA virion burden on FDCs and total FDC area (mm²) as measured by surface analysis. D) Entropy-H(x) plots showing the location of positions with residue mismatches on alignment to the reference Env HIV sequence, and the distribution of those mismatches over different HIV structural regions. Color shades denote distinct structural regions, and orange arcs the locations of gp120 variable loops.

Fig. 5:. Germinal center B cells express a molecular profile favoring the GC development in cross-neutralization LNs.

A) Representative confocal images showing the distribution of Bcl-6 (yellow) and Ki67 (magenta) within the GCs of NNs and Ns. Images were captured at 40x magnification, with 1% zoom and scale is 100um. Box plots and graphs show the pooled normalized numbers of CD20^hi and Bcl-6^hiKi67^hiCD20^hi B cells as measured by Histocytometry. Circles represent single follicles, and data are from all follicles measured in NNs (blue) and Ns (red) tissues (p=0.1692 and *p=0.0185, ANOVA). B) UMAP projections of B-cell populations in total LMNCs, color-coded by B cell category. C) Volcano plots and dot plots of differentially expressed genes (DEGs) and normalized enrichment scores (NES) of GSEA for select B-cell specific ligands and transcription factors are shown. The DEGs were calculated using the MAST (Hurdle model based) test via the FindMarkers function using the Seurat package in R. The p- values were corrected using the Benjamini-Hochberg method and genes meeting an adjusted p< 0.05 and abs(log₂FC)>0.5 are highlighted on the volcano plot. Dots in volcano plots represent genes that are significantly downregulated (blue) or upregulated (red) in neutralizers. D) Heatmap of pairwise TFH-GC B cell interactions as identified through ligand-target analysis. Darker colors denote interactions corresponding to an increased regulatory potential.