Distinct cytokine/chemokine network in semen and blood characterize different stages of HIV infection

Abstract

Objective: The cytokine/chemokine network is used by the innate and adaptive immune system to orchestrate effective immune responses. Here, we describe the cross-sectional association between cytokine levels and stage of HIV infection to gain novel insights into HIV-1 immunopathogenesis and identify novel therapeutic targets.

Design: Concentrations of 31 cytokine/chemokines were retrospectively measured in blood and seminal plasma collected from 252 individuals enrolled in four well characterized cohorts: HIV-uninfected, untreated HIV-infected in early phase of infection, untreated HIV-infected in late phase of infection, and HIV-infected on antiretroviral therapy with undetectable HIV RNA levels in blood (<50 copies/ml).

Methods: Cytokine/chemokine levels were measured by multiplex-bead array. Comparisons between groups were performed by Mann-Whitney U-test and P values were adjusted for multiple comparisons using the Benjamini-Hochberg method.

Results: Presence of HIV-infection skewed the cytokine/chemokine network towards a pro-inflammatory response in both blood and semen compared to HIV-uninfected controls. Such changes emerged within the first weeks of infection and were maintained thereafter: Among untreated HIV-infected individuals, none of the 31 measured cytokines were significantly different between early and later stages of infection. Suppression of plasma HIV RNA with ART did not result in normalization of the levels of pro-inflammatory cytokines in blood. In semen, several pro-inflammatory cytokines were even further upregulated in ART-treated compared with HIV-uninfected and HIV-untreated individuals.

Conclusion: A profound disruption in the cytokine/chemokine network is evident in blood and semen from the earliest stage of HIV infection shortly after the first detection of systemic viremia. These changes are maintained throughout the chronic phase of the infection and do not normalize despite ART and suppression of plasma HIV RNA.

Figure 1

Comparison of cytokine/chemokine median levels in blood plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 1

Comparison of cytokine/chemokine median levels in blood plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 1

Comparison of cytokine/chemokine median levels in blood plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 1

Comparison of cytokine/chemokine median levels in blood plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 2

Comparison of cytokine/chemokine median levels in seminal plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 2

Comparison of cytokine/chemokine median levels in seminal plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 2

Comparison of cytokine/chemokine median levels in seminal plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 2

Comparison of cytokine/chemokine median levels in seminal plasma from HIV-negative controls (group 1), early HIV-infected (< 90 days from estimated date of infection) antiretroviral therapy (ART)-naïve individuals (group 2), late HIV-infected ART-naïve individuals (group 3), and chronically HIV-infected individuals on ART with suppressed HIV RNA < 50 copies/ml (group 4). Represented are the cytokine/chemokine fold increase or decrease (ratio of medians) statistically significant between early HIV-1-infected (ART)-naïve individuals and HIV-1 uninfected controls (A), late HIV-infected ART-naïve individuals and HIV-1 uninfected controls (B), chronically HIV-infected individuals on ART and HIV-1 uninfected controls (C), and chronically HIV-infected individuals on ART or not (D). Comparisons between pairwise groups were done using the Mann Whitney U test. P values for pairwise comparisons were adjusted for multiple comparisons using the Benjamini-Hochberg method (resulting in a raw p<=0.0002 as a cutoff for statistical significance).

Figure 3

Cytokine spectra are different in blood and seminal plasma. Median of cytokine concentration ratios in seminal plasma and blood plasma (S:B) are shown for HIV-uninfected controls (group 1, white bars), HIV-infected individuals in early (group 2, light grey bars) or in a later stage (group 3, dark grey bars) of HIV infection and HIV-infected individuals under antiretroviral treatment (ART) (group 4, black bars). Ratios >1or <1 indicate enrichment of a cytokine in semen or blood plasma, respectively. Ratios for some of the cytokines in which concentration remained under the limit of detection either in blood or seminal plasma were not reported in Figure 11 (IL-17, IL-33, GM-CSF and IL-13).