Evidence of dysregulation of dendritic cells in primary HIV infection

Abstract

Myeloid and plasmacytoid dendritic cells (DCs) are important mediators of both innate and adaptive immunity against pathogens such as HIV. During the course of HIV infection, blood DC numbers fall substantially. In the present study, we sought to determine how early in HIV infection the reduction occurs and whether the remaining DC subsets maintain functional capacity. We find that both myeloid DC and plasmacytoid DC levels decline very early during acute HIV infection. Despite the initial reduction in numbers, those DCs that remain in circulation retain their function and are able to stimulate allogeneic T-cell responses, and up-regulate maturation markers plus produce cytokines/chemokines in response to stimulation with TLR7/8 agonists. Notably, DCs from HIV-infected subjects produced significantly higher levels of cytokines/chemokines in response to stimulation with TLR7/8 agonists than DCs from uninfected controls. Further examination of gene expression profiles indicated in vivo activation, either directly or indirectly, of DCs during HIV infection. Taken together, our data demonstrate that despite the reduction in circulating DC numbers, those that remain in the blood display hyperfunctionality and implicates a possible role for DCs in promoting chronic immune activation.

Figures

Figure 1

mDC and pDC frequencies are significantly reduced during HIV infection. (A) The percentages of mDCs and pDCs were calculated based on the following gating scheme. Total viable PBMCs were gated based on their forward and side scatter (left panel). After 4-color staining with anti-lineage (CD3, CD14, CD16, CD20, and CD56) and anti-HLA-DR antibodies, mDCs and pDCs were identified as lineage negative and HLA-DR positive cells (middle panel). Additionally, pDCs expressed CD123 (right top panel), while mDCs expressed CD11c (right bottom panel). The percentages of (B) mDCs and (C) pDCs within PBMCs in the following groups of subjects are shown: U, uninfected; HR, high risk; primary infection: Fiebig stages I-II (N = 11), III-IV (N = 3), and V-VI (N = 27); E, early established infection (N = 6); CU, chronic/untreated (N = 9); CT, chronic/treated (N = 9); LTNP, long-term nonprogressor (N = 8); and EC, elite controller (N = 6). Horizontal bars indicate the median values. Hollow dots represent outliers. The nonparametric Mann-Whitney U test was used to determine whether there were significant differences between subject groups in circulating DC frequencies and the results are summarized in the left panels. As indicated, DC frequencies in the HIV seronegative subjects (U plus HR groups) were significantly lower than those in all the HIV-infected subjects. The asterisks indicate P values: ***P < .001, **P < .01, and *P < .05; NS = not significant.

Figure 2

DC frequencies are strongly associated with plasma viral load. Correlation of plasma viral load (A-B) and CD4 counts (C-D) to mDC and pDC frequencies from primary HIV-infected subjects at Fiebig stages I-II (N = 11) (A,C) and III-VI (N = 30) (B,D). III-IV (N = 3) and V-VI (N = 27) were combined as III-IV contained too few N for statistical analysis. Associations between various parameters were evaluated using linear regression. Open diamonds correspond to mDCs and closed circles correspond to pDCs. Solid lines denote the relationship between mDC numbers and plasma viral load or CD4 counts. Dotted lines denote the relationship between pDC numbers and plasma viral load or CD4 counts. p denotes the P value of the slope parameter in the linear regression.

Figure 3

Longitudinal measurements of plasma viral load, CD4 cell, and DC numbers in HIV-infected subjects. HIV-infected subjects enrolled in the study were grouped as treated (N = 40, red open diamonds) and untreated (N = 13, blue solid circles) and were followed longitudinally for 52 weeks. Date of infection was estimated for all subjects using the Serologic Testing Algorithm for Recent HIV Seroconversion (STARHS or detuned ELISA) to establish whether patients with positive HIV antibody tests had recently acquired HIV. All monthly sample collection points were aligned accordingly. Left panels show measurements of (A) plasma viral load, (B) percent CD4, and (C) percent mDC, and (D) percent pDC frequencies obtained during the monthly visits. Right panels show the trend for each subject group (solid lines: red = ART treated and blue = untreated) as calculated using linear spline mixed effects models, and the dashed lines around the group average estimates represent the 95% confidence intervals. The horizontal gray dashed lines indicate the range of DC numbers observed in uninfected controls.

Figure 4

mDCs and pDCs from PHIV subjects stimulate allogeneic CD4+ and CD8+ T cells. mDCs and pDCs purified from HIV-seronegative subjects or people with primary HIV infection were cultured with PBMCs depleted of antigen-presenting cells from uninfected subjects. Proliferation was evaluated at day 5 by measurement of 3H-thymidine incorporation and was expressed as a SI. The SI was determined by dividing the proliferation induced in the presence of DCs by the proliferation observed in a culture with no DCs added. The mean T-cell proliferation induced by (A) mDCs and (B) pDCs is shown from uninfected subjects (N = 5, □) and subjects at Fiebig stages I-II (N = 1,), III-IV (N = 2, ■), and V-VI (N = 9,▧). The panels below show the correlation of (C-E) mDC-induced and (D-F) pDC-induced T-cell proliferation to (C-D) CD4 counts and (E-F) plasma viral load in all the HIV-infected subjects at Fiebig stages I-VI. Statistical analysis was performed using a Spearman correlation.

Figure 5

DCs up-regulate maturation markers in response to stimulation with TLR agonists. Purified mDCs and pDCs were stimulated with the following TLR7 agonists: 10μM R848 and 300 ng of p24CA equivalents AT-2 HIV for 16-18 hours. Cells were harvested and stained for expression of CCR7, CD40, and CD86. (A) Representative flow cytometric data from primary HIV-infected subject at Fiebig stages I-II. Light gray shading denotes staining of the unstimulated cells with isotype control antibodies. Dark gray shading denotes antibody staining of the unstimulated cells. Solid black lines denote antibody staining of the stimulated cells. Summary of (B) mDC and (C) pDC expression of CCR7, CD40, and CD86 after TLR stimulation. Each bar represents the mean staining (expressed as mean fluorescence intensity; MFI) of DCs from uninfected subjects, subjects at Fiebig stages I-II, subjects at Fiebig stages III-IV, and subjects at Fiebig stages V-VI. Statistical analysis was performed using the nonparametric Mann-Whitney test.

Figure 6

Levels of chemokines and cytokines produced by DCs in response to stimulation with TLR7 agonists. Purified mDCs and pDCs were stimulated with 10μM R848 or 300 ng of p24CA equivalents AT-2 HIV for 16-18 hours. Supernatants were collected, and cytokine and chemokine production was evaluated using a multiplex bead-based assay (Luminex) and an IFNα ELISA. (A) Summary of cytokines and chemokines secreted by mDCs in response to stimulation with 10μM R848. (B) Summary of cytokines and chemokines secreted by pDCs in response to stimulation with 10μM R848 or 300 ng of p24CA equivalents AT-2 HIV. Each bar represents the median analyte level (pg/mL) produced by DCs from uninfected subjects (N = 10, □), subjects at Fiebig stages I-II (N = 2,), subjects at Fiebig stages III-IV (N = 4, ■), and subjects at Fiebig stages V-VI (N = 10 for mDCs, 7 for pDCs,▧). The nonparametric Mann-Whitney U test was used to determine the statistical significance of differences in the levels of analyte production by DCs from primary HIV-infected subjects at Fiebig stages V-VI and uninfected controls. The asterisks indicate P values: ***P < .001, **P < .01, and *P < .05.

Figure 7

DCs display a transcriptional profile that is indicative of in vivo activation. (A-B) RNA was extracted from ex vivo purified mDCs and pDCs from uninfected and primary HIV-infected subjects at Fiebig stages V-VI and converted to cDNA. QPCR was conducted using primers for TLR7 (A) and IRF 7 (B). Expression levels were normalized using GAPDH. The mean relative gene expression levels in mDCs and pDCs from uninfected subjects (N = 5, □) and subjects with primary HIV infection (N = 5 subjects at Fiebig stages V-VI, ■) are shown. (C-F) RNA was extracted from ex vivo purified mDCs and pDCs and hybridized on Illumina Human HT-12 Expression Beadchips, which assays approximately 25 000 annotated genes with > 48 000 probes derived from the RefSeq (Build 36.2, Rel 22) and the UniGene (Build 99) databases. Comparisons of expression results were done in the Partek software using ANOVA. (C) PCA of complete gene expression profiles measured on each individual array. PCA was performed on log2 transformed and quantile normalized expression data. Each colored sphere (primary HIV-infected) and triangle (uninfected) indicates complete expression profiles of individual samples with similarity between data sets displayed as proximity in 3-dimensional space (red = mDC, blue = pDC). The lines connect samples from the same subject. (D) Venn diagram indicating overlap of probe sets with differential expression in mDCs and pDCs from primary HIV-infected subjects compared with uninfected subjects as calculated using ANOVA. (E-F) Heat map of probe sets associated with immune activation in mDCs (E) and pDCs (F). Each column represents different subjects: uninfected (N = 4, light gray) and primary HIV-infected (N = 5, dark gray). Genes that are down-regulated by up to 2-fold are blue, and genes that are up-regulated by up to 2-fold are red.