Fig 1.
Experimental design used for gene array analysis.
(A) shows the total IgG expression levels in supernatants from autologous germinal center (GC)-Tfh and GC-B cell co-cultures from HIV negative (HIVneg) and HIV positive (HIVpos) lymph nodes (LNs). Supernatants were collected on day 5 of the co-culture and analyzed for total IgG levels by ELISA. Closed squares (n = 5) depict HIVneg and closed triangles (n = 5) depict HIVpos individuals. Results are shown in ng/ml and are represented as mean ± SEM. Data was analyzed with the unpaired two-tailed Student’s t-test using the Mann-Whitney test. Nominal p-values p<0.05 were considered statistically significant. ** p<0.01. (B) shows the co-culture/sorting strategy used for gene array analysis. At day 0, LN mononuclear cells from HIVneg and HIVpos individuals were labeled with CFSE, sorted into GC-Tfh and GC-B cells and co-cultured in the presence of Staphylococcal Enterotoxin B (SEB). At day 5, each cell population was re-sorted according to their CFSE expression (proliferating = CFSEneg or non-proliferating = CFSEpos) and prepared for gene array analysis. (C) Representative conventional flow cytometry plots showing hierarchal phenotype gating strategy of sorted CFSE-labeled LN GC-Tfh and GC-B cell populations from HIVneg or HIVpos individuals used in the co-culture assay. At day 0, sorted GC-Tfh cells were CFSE+ 7AAD- CD19- CD3+ CD4+ CD45RA- CXCR5hi. Sorted GC-B cells were CFSE+ 7AAD- CD3- CD19+ CD38int IgD- CD319-. (D) Representative flow cytometry plots showing hierarchal phenotype gating strategy of the re-sorted GC-Tfh and GC-B cells after 5 days of co-culture, based on their CFSE expression levels. 7AAD- CD3+ CD4+ GC-Tfh as well as 7AAD- CD19+ GC-B cells were sorted into CFSEneg and CFSEpos cells for gene array analysis. (E) shows the scheme used in the gene array assay. LNs from 11 donors (n = 5 HIVneg versus n = 6 HIVpos) were included in this experiment. Cells were labeled with CFSE, sorted, co-cultured and re-sorted for gene array analysis, as described above. Due to quality control, analyzed samples consisted of n = 8 HIVneg (5 CFSEneg and 3 CFSEpos) and n = 12 HIVpos (6 CFSEneg and 6 CFSEpos) cell populations. (F) shows the percent of proliferating (CFSEneg) GC-Tfh (open circles) and GC-B (closed circles) cells from HIVneg subjects (n = 5) and HIVpos patients (n = 6), 5 days after co-culture. Results are represented as mean ± SEM. Data was analyzed using one-way ANOVA followed by the Tukey multiple comparisons test. Nominal p-values p<0.05 were considered of statistical significance. ns indicates absence of statistical significance. Fig 1B and 1E were created with BioRender.com.
Fig 2.
Multi-dimensional Scaling Plot (MDS) reveals unique proliferation- and infection-related gene profile clustering in GC-Tfh cells from HIVpos versus HIVneg individuals.
We performed single-gene analysis of divided/activated versus non-proliferating LN GC-Tfh cells after 5 days of co-culture with autologous GC-B cells from HIVpos patients versus healthy subjects. Principle component (PC) analysis shown in the MDS plot along the first and second principle components, provides a visual representation of the similarity or distance between datasets from all samples. GC-Tfh cells were analyzed based on whether they are proliferating (CFSE negative CFSEneg: pink data points) or non-proliferating (CFSE positive CFSEpos: green data points), but also based on infection status (HIV negative HIVneg: circles or HIV positive HIVpos: triangles). MDS plot shows gene expression clustering between proliferating versus non-proliferating GC-Tfh cells as well as between proliferating HIVpos versus HIVneg GC-Tfh cells. Each symbol represents one sample. CFSEneg HIVneg: n = 4; CFSEneg HIVpos: n = 6; CFSEpos HIVneg: n = 3; CFSEpos HIVpos: n = 5).
Fig 3.
Differential gene expression characterizes proliferating GC-Tfh cells from LNs of HIV-infected and HIV-uninfected individuals.
We used single-gene analysis to determine the pattern of gene expression in proliferating LN GC-Tfh cells in co-culture with autologous GC-B cells from HIVpos patients versus HIVneg subjects. (A) Table summarizing differential gene expression analysis. Table shows the number of differentially expressed genes (DEGs) specific to proliferating (CFSEneg vs CFSEpos) HIVneg and HIVpos GC-Tfh cells after 5 days of co-culture with GC-B cells, selected for analysis based on nominal p-value p<0.05 or Benjamini Hochberg adjusted p-value (false discovery rate (FDR)) of <0.05. (B-C) Heatmaps of the top 100 DEGs in proliferating GC-Tfh cells from (B) HIVneg and (C) HIVpos individuals. We generated the heatmaps of the top 100 DEGs based on the highest absolute log2-fold change (logFC) in proliferating (CFSEneg vs CFSEpos) GC-Tfh cells from HIV-uninfected and HIV-infected subjects compared to non-dividing cells. Nominal and Benjamini Hochberg adjusted p-values (or FDR) of <0.05 were considered statistically significant. Upregulated genes are highlighted in purple and downregulated genes are highlighted in green.
Fig 4.
Pathway analysis reveals an alteration of immune-related pathway expression in proliferating HIVpos GC-Tfh cells.
Gene Set Variation Analysis (GSVA) was used to determine the biological characterization, statistical significance and differences in selected databases. We performed enrichment analysis using Igenuity Pathway Analysis (IPA) gene sets to determine the profile of GC-Tfh cell proliferation in the context of HIV or a non-HIV environment. (A-B) Checkerboard plots represent the top list of enriched pathways in the DEGs specific to proliferating GC-Tfh cells in the (A) HIVneg and (B) HIVpos contexts. Commonly enriched cell cycle signaling and regulation, DNA repair and metabolism pathways are significantly enriched in both settings, at p<0.05. The genes on the x axes in A and B are drivers of the specific pathway enrichments on the y axes. Genes shown in the heatmaps have the highest positive and negative log2-fold change (logFC), while pathways within the heatmaps are the top nominally significant based on GSVA enrichment analysis at p<0.05. Genes highlighted in purple are upregulated and genes highlighted in green are downregulated. (C) Table showing the commonly enriched pathways in the top pathway list between proliferating (CFSEneg vs CFSEpos) HIVneg and HIVpos GC-Tfh cells. Nominal p-values p<0.05 were considered significant. (D) Table showing a selection of enriched GC-Tfh-associated immunological pathways, altered in proliferating (CFSEneg vs CFSEpos) HIVpos versus HIVneg cells based on analysis by IPA. Statistical significance was considered with nominal p-values p<0.05.
Fig 5.
c-Maf signaling represents a key dysregulated pathway in proliferating HIVpos GC-Tfh cells.
(A) shows log2-fold change (logFC) in the expression of selected DEGs in different cell signaling and GC-Tfh-associated immunological pathways in proliferating (CFSEneg vs CFSEpos) HIVneg and HIVpos GC-Tfh cells. Among these, are genes coding for transcription factors (E2F2) and coregulators (HOPX), enzymes (ADA) and key GC-Tfh genes (ICOS, IRF4, BCL6, MAF and mediators of its signaling pathway IL6R, STAT3, BATF). (B) table showing log2-fold change (logFC) values and statistical significance between proliferating HIVpos versus HIVneg GC-Tfh cells for each selected DEG in (A). To assess statistical significance between the logFC of proliferating HIVpos versus HIVneg GC-Tfh cells, data in (A) was analyzed using a double contrast. CFSEneg values were baselined to CFSEpos values and an HIVpos versus HIVneg contrast was performed to compare significance between (CFSEneg vs CFSEpos) HIVpos and (CFSEneg vs CFSEpos) HIVneg cells. (C-D) Venn diagram analysis of proliferating (CFSEneg vs CFSEpos) HIVneg and HIVpos GC-Tfh cells. Diagrams show the numbers of unique and common statistically significant DEGs in the indicated GC-Tfh populations. (C) shows Venn diagram analysis of upregulated DEGs (represented in purple) and (D) shows Venn diagram analysis of downregulated DEGs (represented in green) in HIVneg and HIVpos cells. (B-D) Analysis was performed based on nominal p-value p<0.05.
Fig 6.
Specific ADA-1 blockade reduces c-Maf and BATF expression in healthy tonsillar GC-Tfh and pre-Tfh cells.
We sorted GC-Tfh, pre-Tfh and autologous GC-B cells from the tonsils of healthy individuals and co-cultured GC-Tfh:GC-B as well as pre-Tfh:GC-B cells in a 1:1 ratio in the presence of SEB (SEB) with or without 10uM specific ADA-1 inhibitor EHNA (EHNA). We harvested the cells one day after co-culture and performed intracellular flow cytometry staining for c-Maf and BATF expression in GC-Tfh and pre-Tfh cells. (A) shows the percent of c-Mafhi GC-Tfh and pre-Tfh cells in the total CD4+ T cell population (left) and c-Maf median fluorescence intensity (MDFI) in GC-Tfh and pre-Tfh cells, gated on the total CD4+ T cells (right) (n = 7–8). (B) shows the percent of BATFhi GC-Tfh and pre-Tfh cells in the total CD4+ T cell population (left) and BATF median fluorescence intensity (MDFI) in GC-Tfh and pre-Tfh cells, gated on the total CD4+ T cells (right) (n = 7–8). Negative controls are labeled as Fluorescence Minus One (FMO). Results are from 3 independent experiments and are represented as mean ± SEM. Data was analyzed with the two-tailed paired non-parametric Student’s t-test using the Wilcoxon matched-pairs signed rank test. Nominal p-values p<0.05 were considered of statistical significance. * p<0.05 and ** p<0.01.
Fig 7.
Specific ADA-1 blockade attenuates healthy tonsillar GC-Tfh cell function.
We sorted GC-Tfh, pre-Tfh and autologous GC-B cells from the tonsils of healthy individuals and co-cultured GC-Tfh:GC-B as well as pre-Tfh:GC-B cells in a 1:1 ratio in the presence of SEB (SEB) with or without 10uM of the specific ADA-1 inhibitor EHNA (EHNA). (A-B) We collected culture supernatants 5 days after co-culture and analyzed the levels of total IgG in ng/ml by ELISA. (A) Total IgG level in supernatant from GC-Tfh:GC-B cell co-cultures. (B) Total IgG level in supernatant from pre-Tfh:GC-B cell co-cultures. Results are from 2 independent experiments (n = 6) and are represented as mean as well as before and after treatment lines. Data was analyzed with the two-tailed paired non-parametric Student’s t-test using the Wilcoxon matched-pairs signed rank test. (C-L) We collected culture supernatants 5 days after co-culture and analyzed GC-Tfh and pre-Tfh cell cytokine and chemokine levels by the Luminex assay. Levels of (C) BLC, (D) IL-21, (E) IL-10, (F) IL-4, (G) IFN-γ, (H) IL-17A, (I) IL-2, (J) IP-10, (K) MIP-1α and (L) MIP-1ß are shown in pg/ml. Results are from 2 independent experiments (n = 6) and are represented as mean ± SEM. Data was analyzed with the two-tailed unpaired non-parametric Student’s t-test using the Mann-Whitney test. Nominal p-values p<0.05 were considered statistically significant. * p<0.05, ** p<0.01 and *** p<0.001.
Fig 8.
Co-culture supplementation with ADA-1 restores the defective pro-GC-Tfh IL-6 pathway in HIVpos GC-Tfh cells.
We sorted GC-Tfh and autologous GC-B cells from LNs of HIVpos individuals with chronic infection and co-cultured them in a 1:1 ratio in the presence of SEB (SEB) with or without 2.4uM ADA-1 (ADA-1). Cells were harvested 1 day after co-culture then GC-Tfh cells were re-sorted and total RNA extracted for RNA-seq analysis. We analyzed the data using Gene Set Variation Analysis (GSVA) by examining Molecular Signatures Databases (MSigDB) Canonical Pathways. Heatmaps illustrating the differential gene expression in the (A) IL-6, (B) IL-2 and (C) Th1/Th2 pathways in HIVpos GC-Tfh cells treated in co-culture with ADA-1 versus SEB are represented. RNA-seq analysis was performed on n = 3 SEB samples and n = 3 ADA-1 samples. Nominal p-values p<0.05 were considered statistically significant and are indicated for each pathway on the top left side. Where specified, statistical significance is shown for selected genes with *p<0.05. Absence of statistical significance for selected genes is represented by ns. Upregulated genes or pathways are highlighted in purple while downregulated genes or pathways are highlighted in green. Statistical analysis for all genes of the three pathways is shown in S4 Table.