Abstract
Background
Seminal human immunodeficiency virus (HIV) transmission from men to their partners remains the main driver of HIV epidemics worldwide. Semen is not merely a carrier of the virus, but also provides an immunological milieu that affects HIV transmission.
Methods
We collected blood and semen from people with HIV whose epidemiologically linked sexual partners either did or did not acquire HIV. Viral transmission was confirmed by phylogenetic linkage (HIV pol). We measured the concentration of 34 cytokines/chemokines by Luminex in the blood and semen of 21 source partners who transmitted HIV (transmitters) and 22 who did not transmit HIV (nontransmitters) to their sexual partners. Differences between cytokine profiles in transmitters versus nontransmitters were analyzed using the multivariate statistical technique of partial least square discriminant analysis.
Results
The cytokine profile in seminal fluid, but not in peripheral blood, was significantly different between men who have sex with men (MSM) who transmitted HIV and those who did not transmit HIV to their sexual partners (E = 19.77; P < .01). This difference persisted after excluding people with undetectable HIV RNA levels in nontransmitters.
Conclusions
Seminal cytokine profiles correlated with transmission or nontransmission of HIV from the infected MSM to their partners, independently from seminal viral load. Seminal cytokine spectra might be a contributing determinant of sexual HIV transmission, thus providing new directions for the development of strategies aimed at preventing HIV transmission.
Keywords: cytokine, semen, blood, HIV transmission
The seminal cytokine profile of people who transmitted human immunodeficiency virus (HIV) to partners was statistically different from that of people who did not transmit HIV, even after correction for viral load. The cytokine profiles in blood were not different.
Despite its low probability, human immunodeficiency virus (HIV) sexual transmission remains the main cause of new infections worldwide [1, 2]. HIV transmission and susceptibility are a multifactorial phenomenon associated with several established factors, including use of antoretroviral therapy (ART) host genetic variants or polymorphisms, the number of partners, male circumcision, condom use rates, and the type of receptive intercourse (anal vs vaginal). Semen and its constituents also influence the risk of HIV-1 sexual transmission/acquisition, as the probability of HIV-1 transmission increases with the seminal viral load [3]. However, semen is not only a carrier for HIV-1 but also provides an immunological milieu in which HIV-1 variants are selected. Therefore, it is crucial to understand the factors associated with increased seminal HIV replication and HIV transmission. In this context, we and others have reported that HIV-1 infection results in the reactivation of other seminal pathogens [4–8], as well as an aberrant production of cytokines/chemokines [5, 6, 9–15]. The profound disruption in the cytokine network is evident in blood and semen from the earliest stage of HIV infection, shortly after the first detection of systemic viremia. Cytokine network changes are maintained throughout the chronic phase of the infection and are not normalized despite ART and the suppression of plasma HIV-1-RNA [13]. We and others have suggested that the resulting altered blood and seminal milieu in HIV-1 infection may be a determinant of HIV-1 sexual transmission [5, 16–18]. In particular, changes in the levels of certain cytokines/chemokines in the blood or semen of people with HIV may facilitate HIV-1 infectiousness and onward transmission [17].
The importance of cytokine/chemokine profiles in determining the risk of HIV transmission has been shown in multiple cohorts of highly exposed, seronegative individuals who do not contract HIV despite repeated exposures to HIV-1. In particular, several studies have shown that women resistant to HIV have a distinct pattern of mucosal chemokine/cytokine expression that contributes to a lower level of immune activation, or “immune quiescence,” which is associated with a lower risk of HIV acquisition. Similarly, Fulcher et al [19] found that explants from highly exposed, seronegative men who have sex with men (MSM) produced fewer proinflammatory cytokines, compared with controls, following innate immune stimulation.
Here, we compared the cytokine/chemokine profiles in the blood and semen of source partners to investigate whether the profiles are associated with HIV transmission in the men’s recent sexual partners, who either became infected with a phylogenetically linked HIV strain (where the putative source is referred to as a transmitter) or did not acquire HIV (where the putative source partner is referred to as a nontransmitter).
METHODS
Study Participants
Participants with early HIV infection in this study were recruited from October 2002 to August 2011 through the San Diego Primary Infection Resource Consortium in a partner-pair recruitment study. Acute and early HIV infection were defined as described previously [20]. Participants subsequently recruited their most recent sexual partners for further evaluation and HIV testing.
Within each epidemiologically linked sexual-partner pair, we classified the source or recipient based on their HIV status and estimated date of infection (EDI). Putative sources were classified as the partner with the earlier EDI or the only partner with HIV in the pair. Recipients were classified either as HIV-negative partners or the partner with HIV and the later EDI. If the recipient partner did not have HIV, the source was classified as a nontransmitter. If the recipient partner was infected with HIV and both the source partner and recipient’s virus pol gens were similar (see below), the source was classified as a transmitter. We identified a total of 43 putative source partners and 48 recipient partners amongst 26 transmission pairs and 25 nontransmission pairs. The studies were conducted with appropriate written consent and approved by the University of California–San Diego Human Research Protections program.
Sample Collection and Processing
HIV transmitters were confirmed by phylogenetic linkage (Viroseq 2.0; Applied Biosystems) of pol sequences of HIV from each partner (genetic distance, ≤1%). For each source partner, semen was collected by masturbation and processed as previously described [21]. Blood was drawn for both the source partner and recipient partner, and each person’s HIV RNA level was quantified (Amplicor HIV Monitor Test; Roche Molecular Systems Inc.).
Multiplex Bead Array Assay for Cytokines/Chemokines Quantification
There were 34 cytokines/chemokines involved in different immunological functions that were measured in blood and semen from the 43 putative sources and in blood samples from the 48 recipients, as previously described [13] (see Supplementary Data). The National Institutes of Health laboratory that performed Luminex measurements is part of the Microbicide Quality Assurance Program [22].
Statistical Analysis
We performed 3 analyses, comparing cytokine profiles in the (1) semen of transmitters versus nontransmitters; (2) blood of transmitters versus nontransmitters; and (3) blood of recipients with and without HIV. For each analysis and within each cytokine type, cytokine values were log10-transformed, and then normalized by subtracting the mean log10 value and dividing the standard deviation of the log10 values. To avoid highly skewed predictors, we restricted each analysis to those cytokines in which at least 50% of the measurements were above the level of detection (Supplementary Table S1). Undetectable values were replaced by half the minimum level of detection.
We used a Fisher’s exact test and Mann-Whitney U test to compare demographics and clinical characteristics (viral load, CD4, etc) between transmitters versus nontransmitters and between recipients with HIV versus those without.
To visualize the separation of cytokines between groups and investigate the classification of groups based on cytokine profiles, we performed a partial least squares discriminant analysis (PLS-DA). PLS-DA reduces the data by creating composite, latent variables, in which each cytokine contributes a determined weight (loading). The optimal number of latent variables (n = 2) was determined by analyzing a marginal decrease in the classification error rate by the Mahalanobis distance. The loadings for each cytokine, along with the variance explained by each latent variable, were used to calculate a variable importance in projection (VIP) for each cytokine. If all cytokines are equally important, the VIP will be equal to 1 for all cytokines, leading us to classify cytokines with VIPs larger than 1 as “important.” We also computed the balanced error rate (BER) of predicted classification on our data to summarize the predictive power of cytokine profiles by group. We tested for the difference in the distribution of the cytokines between groups using the E-statistic for a 2-sample difference in the multivariate normal distribution [23]. Bootstrap replicates numbering 500 000 were used to generate P values for the E-statistic. Analyses were performed using R 3.5, the E-test was performed using the “energy” package, and the PLS-DA was performed using the “mixOmics” package.
RESULTS
Study Participants’ Demographics and Clinical Data
Transmitter Versus Nontransmitters (Source Partners)
Putative source partners (n = 43) were mostly MSM. The participants were non-Hispanic White (45.7%), Hispanic (25.7%), and other/multiracial (28.6%). The median age was 34 years old (range, 22–58), and 47% of participants were diagnosed during an acute or early HIV infection (diagnosed within 133 days from EDI). At the time of the blood and semen sample collection, participants had a median log10 plasma viral load in blood plasma of 4.5 copies/mL (IQR, 2.7–5.1) and a median CD4 count of 516 cells/μL (IQR, 340–714). At the time of specimen collection, 18% of source partners were on ART. We found a significant difference in viremia (detectable viral load in blood plasma) by transmission status (100% in transmitters and 68.2% in nontransmitters; P < .01). Transmitters also had more detectable EBV DNA in semen (55% vs 15%, P = .02). (Table 1).
Table 1.
Summary of Demographic and Clinical Data by Transmission Status
| Category | Nontransmitter | Transmitter | Total | P a |
|---|---|---|---|---|
| n = 22 | n = 21 | n = 43 | ||
| Gender, n (%) | ||||
| Male | 22 (100) | 21 (100) | 43 (100) | |
| Race/ethnicity, n (%) | … | … | … | .19 |
| White, non-Hispanic | 9 (40.9) | 7 (53.8) | 16 (45.7) | |
| Hispanic/Latino | 8 (36.4) | 1 (7.7) | 9 (25.7) | |
| Other/multiracial | 5 (22.7) | 5 (38.5) | 10 (28.6) | |
| Infection timeline, n (%) | … | … | … | .13 |
| AEH: EDI < 133 days | 13 (59.1) | 7 (33.3) | 20 (46.5) | |
| Chronic: EDI > 133 days | 9 (40.9) | 14 (66.7) | 23 (53.5) | |
| Age, median (range) | 34 (23–58) | 35 (22–48) | 34 (22–58) | .75 |
| Viremia status at sample, n (%) | … | … | … | <.01 |
| Viremic | 15 (68.2) | 21 (100.0) | 36 (83.7) | |
| Not viremic | 7 (31.8) | … | 7 (16.3) | |
| ART regimen, n (%) | … | … | … | .70 |
| On ART | 5 (22.7) | 3 (14.3) | 8 (18.6) | |
| Lab results, median (IQR) | ||||
| First viral load, log10 copies/106 | 5.2 (4.5–5.5) | 4.9 (4.4–5.5) | 5.0 (4.4–5.5) | .98 |
| Viral load at sample, log10 copies/106 | 3.8 (2.1–5.5) | 4.7 (4.1–5.0) | 4.5 (2.7–5.1) | .26 |
| First absolute CD4 count, cells/μL | 589 (396–681) | 469 (323–563) | 518 (382–626) | .06 |
| CD4 count at sample, cells/μL | 675 (366–746) | 471 (334–548) | 516 (340–714) | .09 |
| Risk behavior in month preceding enrollment | ||||
| Unprotected anal intercourse,b n (%) | 16 (76.2) | 14 (93.3) | 30 (83.3) | .37 |
| Median number of sexual partners (IQR) | 2 (2–6) | 1 (1–4) | 2 (1–4) | .11 |
| Presence of STIs in semen,c n (%) | ||||
| HIV RNA | 9 (45.0) | 15 (75.0) | 24 (60.0) | .11 |
| CMV DNA | 10 (50.0) | 16 (80.0) | 26 (65.0) | .10 |
| EBV DNA | 3 (15.0) | 11 (55.0) | 14 (35.0) | .02 |
| HSV-2 DNA | 0 (0) | 0 (0) | 0 (0) |
Abbreviation: AEH, acute or early HIV infection; ART, antiretroviral therapy; CMV, cytomegalovirus; EBV, Epstein-Barr virus; EDI, estimated date of infection; HIV, human immunodeficiency virus; HSV, herpes simplex virus; IQR, interquartile range; STI, sexually transmitted infection.
aCategorical variables were tested for difference using Fisher’s exact test and continuous variables were tested for difference using the Wilcoxon ranked-sum test.
bAssessed for 36 participants.
cAssessed for 40 participants.
People With Human Immunodeficiency Virus Versus Controls (Recipient Partners)
Recipient partners (n = 48) were mostly male (95.8%), with the remainder of participants being female (4.2%). The participants were non-Hispanic White (63%), Hispanic (13%), and other/multiracial (24%). Their median age was 34 years old (IQR, 19–55). Half (n = 24) of the recipient partners had HIV (all MSM). Among all recipients with HIV, 21 (87.5%) individuals were diagnosed during an acute or early HIV infection, and all 24 (100%) were ART naive and viremic at the time of the blood sample collection, with a median log10 viral load of 5.2 (IQR, 4.6–5.9). There was a significant difference in CD4 counts for recipients with and without HIV, with medians of 506 (IQR, 348–612) and 860 (IQR, 732–1070), respectively (P < .01). Recipients were not found to be different in other applicable categories (Table 2).
Table 2.
Summary of Demographic and Clinical Data by Human Immunodeficiency Virus Status
| Category | Not living with HIV | Living with HIV | Total | P a |
|---|---|---|---|---|
| n = 24 | n = 24 | n = 48 | ||
| Gender, n (%) | ||||
| Male | 22 (91.7) | 24 (100.0) | 46 (95.8) | .49 |
| Female | 2 (8.3) | … | 2 (4.2) | |
| Race/ethnicity,b n (%) | … | … | … | 1.00 |
| White, non-Hispanic | 14 (63.6) | 15 (62.5) | 29 (63.0) | |
| Hispanic/Latino | 3 (13.6) | 3 (12.5) | 6 (13.0) | |
| Other/multiracial | 5 (22.7) | 6 (25.0) | 11 (23.9) | |
| Infection timeline, n (%) | ||||
| AEH: EDI < 133 days | … | 21 (87.5) | … | |
| Chronic: EDI > 133 day | … | 3 (12.5) | … | |
| Age, median (range) | 36 (19–50) | 31 (19–55) | 34 (19–55) | .17 |
| Viremia status at sample, n (%) | ||||
| Viremic | … | 24 (100) | … | |
| Not viremic | … | … | … | |
| Lab results, median (IQR) | ||||
| Viral load at sample, log10 copies/106 | … | 5.2 (4.6–5.9) | … | |
| CD4 count at sample, cells/μL | 860 (732–1070) | 506 (348–612) | 664 (484–892) | <.01 |
Abbreviation: -, negative; +, positive; AEH, acute or early HIV infection; EDI, estimated date of infection; HIV, human immunodeficiency virus; IQR, interquartile range.
aCategorical variables were tested for difference using Fisher’s exact test and continuous variables were tested for difference using the Wilcoxon ranked-sum test.
bThere were 2 participants in the HIV- group with missing race/ethnicity data.
Visualizing Multivariate Cytokine Profiles in 2 Dimensions With Partial Least Squares Discriminant Analysis Projections
In semen, PLS projections showed a clear separation of cytokines between transmitters and nontransmitters (Figure 1A), which was supported by a statistically significant difference in the distribution of cytokines by transmission status (E = 19.77; P < .01). The PLS-DA procedure correctly classified all participants in the correct transmission group (BER = 0%; Supplementary Table S2A). The 2 latent variables from PLS-DA accounted for 49% of the variation in cytokine expression (LV1: 22% and LV2: 27%, respectively; Figure 1A). The most important cytokines in discriminating the 2 groups (in descending order of importance) were interferon (IFN)-γ, interleukin (IL)-13, macrophage colony-stimulating factor (M-CSF), IL-17, granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-4, IL-15, IL-16, IL-33, and eotaxin (Figure 2). Transmitters exhibited higher mean concentrations in IL-13, IL-15, and IL-33, and lower means in IFN-γ, IL-15, M-CSF, IL-17, GM-CSF, IL-4, IL-16, and eotaxin in semen (Supplementary Figure 1A).
Figure 1.
PLS-DA and VIP scores of cytokine expression profiles in T versus NT and in recipient partners who acquired HIV (living with HIV) or remained not living with HIV. PLS-DA projections in 2 LV with ellipses representing Hotelling’s 2-samples T2 with 95% confidence intervals. E-stat was used to test the difference in the distribution of scores. The PLS-DA scores are color coded by VIP scores. Presented are the cytokine profiles of (A) T versus NT in semen; (B) T versus NT in blood; and (C) recipient partners in blood without versus with HIV. Abbreviations: cal, calgranulin; cmvIL, cytomegalovirus encoded interleukin; E-stat, energy statistic; GRO, growth-regulated oncogene; HIV, human immunodeficiency virus; IFN, interferon; IL, interleukin; IP, interferon γ-induced protein; I-TAC, interferon–inducible T cell alpha chemoattractant; LV, latent variables; MCP, monocyte chemoattractant protein; MIG, monokine induced by gamma; MIP, macrophage inflammatory proteins; NT, nontransmitters; PLS-DA, partial least squares discriminant analysis; RANTES, regulated on activation, normal T cell expressed and secreted; T, transmitters; TNF, tumour necrosis factor; var, Variance; VIP, variable importance in projection.
Figure 2.
VIP scores from PLS-DA analysis. Cytokines with larger VIP scores are more important in the PLS-DA projections, and scores > 1 (dashed line) are considered important. Presented are stratified analyses of transmitters versus nontransmitters in blood (Tx: Blood) and in semen (Tx: Semen), and specimens in blood from participants with and without HIV (HIV: Blood). Abbreviations: cal, calgranulin; cmvIL, cytomegalovirus encoded interleukin; GM CSF, granulocyte-macrophage colony-stimulating factor; GRO, growth-regulated oncogene HIV, human immunodeficiency virus; IFN, interferon; IL, interleukin; IP, interferon γ-induced protein; I-TAC, interferon–inducible T cell alpha chemoattractant; MCP, monocyte chemoattractant protein; MIG, monokine induced by gamma; MIP, macrophage inflammatory proteins; PLS-DA, partial least squares discriminant analysis; RANTES, regulated on activation, normal T cell expressed and secreted; TGF, transforming growth factor; Tx, analysis of transmitters versus non transmitter; TNF, tumour necrosis factor; var, variance; VIP, variable importance in projection.
In blood, the 2 latent variables from PLS-DA accounted for 63% of the variation in cytokine expression (LV1: 51% and LV2: 12%, respectively), but the distributions of cytokines between transmitters and nontransmitters were not different (E = 7.03; P = .16), and the PLS projections of the cytokines showed considerable overlap between the groups, with misclassification of 21.0% (BER) of the participants (Figure 1B; Supplementary Table S2A; Supplementary Figure 1B).
Finally, we performed the same analysis on cytokines in the blood samples of recipient partners by HIV status (HIV+ versus HIV-). The distributions of cytokines by HIV status were significantly different (E = 11.67; P = .03), but there was some overlap in the PLS-DA projections (Figure 1C), as evidenced by the misclassification of 14.6% (BER) of participants (Supplementary Table S2B). The 2 latent variables from PLS-DA accounted for 67% of the variation in cytokine expression (LV1: 44% and LV2: 23%, respectively). IP-10, IL-18, and MIG were the most important cytokines in discriminating between the 2 groups (Figure 2), with HIV+ recipients having higher means in all 3 cytokines (Supplementary Figure 1C).
Sensitivity Analysis: Results Were Not Dependent on Luminex Runs or Viral Load
Technically, samples could not be run on the same plate. To ensure that the slight differences in the limits of detection between plates did not bias the results, the analysis was performed 4 times, replacing undetectable levels with each cytokine’s (1) minimum lower limit of detection; (2) maximum lower limit of detection; (3) mean lower limit of detection; and (4) lower limit with each batch. No appreciable differences were found in any of the 4 processes for imputed undetectable values, or by excluding those cytokines with the largest differences in the lower limit of detection (data not shown).
In contrast, the transmitter and nontransmitter groups differed significantly by viremia status at the time of sampling. To ensure this was not the driving source of the difference in the cytokine profiles, we performed a subset analysis, limiting the PLS-DA and E-test to only those participants classified as viremic at the time of sampling. No appreciable differences were found in PLS-DA projections, and there was no change in the significance of the difference by the E-test. Indeed, the distributions of cytokines between viremic transmitter and nontransmitter participants were still significantly different (E = 19.07; P < .01), and the projections still showed clear separation. The BER increased from 0% to ~2%, with 1 of the 36 participants being misclassified.
Finally, to investigate whether differences between the transmitter and nontransmitter groups were responsible for the difference found in the cytokine profile in semen, we repeated the PLS analysis and the calculation of the E-statistic on the set of cytokine concentrations, adjusted for the presence of HIV RNA, cytomegalovirus (CMV), or Epstein-Barr virus (EBV) in seminal plasma. Specifically, to obtain an adjusted concentration for each cytokine, we regressed the log-transformed, normalized level of each cytokine on each of (1) semen HIV RNA; (2) semen CMV DNA; and (3) semen EBV DNA. We then performed PLS-DA and calculated the E-statistic on each adjusted set. In all 3 instances, there was no appreciable difference in the E-statistic (semen HIV RNA, E = 17.6 (P < .01); semen CMV DNA, E = 16.8 [P < .01]; semen EBV DNA, E = 17.8 [P < .01]). Additionally, the separation in PLS projections was still evident after the cytokine concentrations were adjusted for semen HIV RNA and CMV DNA. We found considerable separation in those cytokines adjusted for the presence of EBV DNA; however, the difference was no longer evident (not shown).
DISCUSSION
The cytokine/chemokine network is part of the language used by the innate and adaptive immune system to orchestrate an effective immune response to infectious pathogens [24]. In the case of HIV infection, the immune system not only fails at protecting the host against HIV, but it also becomes dysfunctional, in part due to a dysregulation of the cytokine profiles [5, 6, 10, 16, 24]. HIV infection skews the cytokine/chemokine network towards a pro-inflammatory response, and different stages of HIV infection are characterized by a distinct cytokine/chemokine network, both in blood and semen. Such a dysregulation of inflammatory cytokine networks may have serious implications for HIV transmission [5, 6, 9, 11–15, 25]. In this study, we investigated the association between cytokines in the blood or semen of the putative source partner and HIV transmission. We measured the concentrations of 34 cytokines/chemokines in the blood and semen of 21 source partners who transmitted HIV and 22 source partners who did not transmit HIV to recipient partners.
Cytokines are known to signal in a complex hierarchical network, and most of them show pleiotropic, redundant, and synergetic actions [26]. Therefore, we analyzed the differences between cytokine profiles in transmitters versus nontransmitters, using the multivariate statistical technique of PLS-DA rather than any univariate analysis.
The PLS-DA analysis revealed that the cytokine profiles between transmitters and nontransmitters were statistically different in semen but not in blood. PLS-DA projections of the transmitter versus nontransmitter groups strongly overlapped in blood, suggesting that blood cytokines were not a good predictor of HIV transmission in our cohort of mostly MSM. There was a trend of increased concentrations of pro-inflammatory cytokines (RANTES, IL-18, IL-6, and GROα) in the blood of transmitters, compared to nontransmitters, but the overall difference was not found to be statically significant. A previous study in heterosexual discordant couples reported the association between IL-10 in the blood of the source partner and HIV transmission after excluding plasma HIV RNA, although it was found not significant in the primary multivariate models [27].
In contrast to blood, the cytokine profiles in semen were different between transmitters and nontransmitters. The difference was highly significant, with a BER of 0%.
Overall, this finding reflects the fact that semen is not just a vector for HIV, but a complex milieu that carries pro- and antiviral factors that may facilitate or inhibit HIV transmission [17, 28, 29]. The unique cytokine profile in the semen of the transmitter group suggests that seminal cytokines are likely to be an important determinant of HIV transmission.
In the analysis of a complex cytokine network, the PLS-DA statistical model allowed us to reduce a large number of variables (34 cytokines) to a smaller number of variables (latent variables 1 and 2) in order to better visualize the difference between seminal cytokines of the transmitter and nontransmitter groups. VIP extracted from the PLS-DA model revealed which cytokines accounted for most of the differences between the transmitter and the nontransmitter group. As expected, we did not find a single cytokine that was associated with a bigger risk of transmission. Rather, we found that there was an associated group of cytokines: namely, IFN-γ, IL-13, M-CSF, IL-17, and GM-CSF. All cytokines but IL-13 were found in higher concentrations in nontransmitters, compared to transmitters. IFN-γ and IL-13, the 2 cytokines with the highest VIP, belong to different functional helper T cells type 1 (Th1) groups: IFN-γ is the archetypal cytokine of the Th1 cells that supports cytotoxic T-cell responses, while IL-13 (together with IL-4) is secreted by helper T cells type 2 (Th2) cells, which counter-regulate Th1 responses and activate the humoral immunity. A Th1/Th2 imbalance was originally described as a critical step in the etiology of HIV infection. Th1-associated cytokines have protective effects against HIV infection, whereas a shift towards an augmented humoral Th2 response may be adverse and lead to the progression of HIV infection to acquired immunodeficiency syndrome [30, 31]. Also, natural HIV elite controllers are characterized by the persistence of highly differentiated Th1 cells with effector functions [32, 33].
IFN-γ is a major cytokine in controlling HIV infection. In vitro, IFN-γ has anti-HIV activity [34] and inhibits HIV transmission [35]. In vivo, IFN-γ enhances anti-HIV cytotoxic T lymphocytes activities in HIV+ individuals, contributing to the control of the viral load in long-term nonprogressors and elite controllers, or preventing HIV-1 infection in highly exposed, seronegative individuals [33]. Altogether, these observations are in agreement with the lower concentrations of IFN-γ found in transmitters versus nontransmitters.
Subsequently, we investigated the effect of a recent HIV infection on cytokine perturbation by comparing the cytokine profiles in the blood of recipient partners who became infected with HIV to those who remained uninfected. We found that HIV infection skewed the cytokine/chemokine network towards a pro-inflammatory response, with higher concentrations of IP-10, MIG, and IL-18 in the blood of HIV+ recipient partners, compared to those who remained uninfected. This result was in agreement with our earlier report in a study describing the cross-sectional association between cytokine levels and stages of HIV infection [13]. The similar findings reported in 2 different cohorts validate our analysis and assay methods.
Our study had a number of limitations, including different HIV-1 viral loads in the transmitter and nontransmitter groups at the time of sampling. To ensure that this difference in plasma viral loads was not the driving source of a difference in the cytokine profile between transmitters and nontransmitters, we repeated the PLS-DA analysis, excluding participants with undetectable viral load in the nontransmitter group. No appreciable differences in PLS-DA projections and no change in the significance of the difference by the E-test were found, confirming that seminal cytokine profiles were significantly different between transmitters and nontransmitters. We also repeated the analysis after adjusting for the presence of HIV RNA, CMV DNA, and EBV DNA in semen of the transmitter versus non transmitter, and found no appreciable difference in the E-statistic values.
Another limitation of our study may be its relatively small number of participants and the lack of available seminal samples from the recipients. With a small number of samples and a larger number of variables, PLS-DA projections tend to overemphasize the differences between 2 multivariate groups [26]. Also, the difference in sexual behaviors makes MSM a unique group because of the relatively higher frequency of sexual encounters and because receptive anal sex has a ∼10-time higher relative risk of HIV transmission than that for receptive vaginal sex [36].
Despite these limitations, this study provides the first correlations between seminal cytokines in the transmitting partner and HIV transmission. Identifying the seminal correlates of HIV transmission may provide new directions for the development of new effective strategies aimed at preventing HIV transmission.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments. The authors thank all the study participants and all the nurses at all the enrollment sites. They also thank Wendy Fitzgerald for her help in setting up the cytokine assay.
Financial support. This research was supported, in part, by the Intramural Research Program of the National Institute of Child Health and Human Development, National Institutes of Health (NIH); and by the Department of Veterans Affairs and the NIH (grant numbers AI036214, MH062512, MH107345, AI106039, AI134295, HD094646, AI027763, AI118422, MH113477, and 5T32AI007384).
Potential conflicts of interest. M. H. has received research funding from Gilead S. A. R. has received grant support from T32 AI 007384-28. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1. Royce RA, Seña A, Cates W Jr, Cohen MS. Sexual transmission of HIV. N Engl J Med 1997; 336:1072–8. [DOI] [PubMed] [Google Scholar]
- 2. Cohen MS. Preventing sexual transmission of HIV. Clin Infect Dis 2007; 45(Suppl 4):S287–92. [DOI] [PubMed] [Google Scholar]
- 3. Cohen MS, Smith MK, Muessig KE, Hallett TB, Powers KA, Kashuba AD. Antiretroviral treatment of HIV-1 prevents transmission of HIV-1: where do we go from here? Lancet 2013; 382:1515–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Sheth PM, Danesh A, Sheung A, et al. Disproportionately high semen shedding of HIV is associated with compartmentalized cytomegalovirus reactivation. J Infect Dis 2006; 193:45–8. [DOI] [PubMed] [Google Scholar]
- 5. Lisco A, Munawwar A, Introini A, et al. Semen of HIV-1-infected individuals: local shedding of herpesviruses and reprogrammed cytokine network. J Infect Dis 2012; 205:97–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Olivier AJ, Masson L, Ronacher K, et al. Distinct cytokine patterns in semen influence local HIV shedding and HIV target cell activation. J Infect Dis 2014; 209:1174–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Gianella S, Redd AD, Grabowski MK, et al. Vaginal cytomegalovirus shedding before and after initiation of antiretroviral therapy in Rakai, Uganda. J Infect Dis 2015; 212:899–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Gianella S, Scheffler K, Mehta SR, et al. Seminal shedding of CMV and HIV transmission among men who have sex with men. Int J Environ Res Public Health 2015; 12:7585–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Berlier W, Bourlet T, Lévy R, Lucht F, Pozzetto B, Delézay O. Amount of seminal IL-1beta positively correlates to HIV-1 load in the semen of infected patients. J Clin Virol 2006; 36:204–7. [DOI] [PubMed] [Google Scholar]
- 10. Katsikis PD, Mueller YM, Villinger F. The cytokine network of acute HIV infection: a promising target for vaccines and therapy to reduce viral set-point? PLOS Pathog 2011; 7:e1002055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Hoffman JC, Anton PA, Baldwin GC, et al. Seminal plasma HIV-1 RNA concentration is strongly associated with altered levels of seminal plasma interferon-γ, interleukin-17, and interleukin-5. AIDS Res Hum Retroviruses 2014; 30:1082–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kafka JK, Osborne BJ, Sheth PM, et al. Latent TGF-β1 is compartmentalized between blood and seminal plasma of HIV-positive men and its activation in semen is negatively correlated with viral load and immune activation. Am J Reprod Immunol 2015; 73:151–61. [DOI] [PubMed] [Google Scholar]
- 13. Vanpouille C, Introini A, Morris SR, et al. Distinct cytokine/chemokine network in semen and blood characterize different stages of HIV infection. AIDS 2016; 30:193–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Rametse CL, Adefuye AO, Olivier AJ, et al. Inflammatory cytokine profiles of semen influence cytokine responses of cervicovaginal epithelial cells. Front Immunol 2018; 9:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Houzet L, Perez-Losada M, Matusali G, et al. Seminal simian immunodeficiency virus in chronically infected cynomolgus macaques is dominated by virus originating from multiple genital organs. J Virol 2018; 92:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Anderson JA, Ping LH, Dibben O, et al. ; Center for Human Immunodeficiency Virus/Acquired Immunodeficiency Syndrome Vaccine Immunology HIV-1 populations in semen arise through multiple mechanisms. PLOS Pathog 2010; 6:e1001053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Doncel GF, Anderson S, Zalenskaya I. Role of semen in modulating the female genital tract microenvironment–implications for HIV transmission. Am J Reprod Immunol 2014; 71:564–74. [DOI] [PubMed] [Google Scholar]
- 18. Gianella S, Massanella M, Wertheim JO, Smith DM. The sordid affair between human herpesvirus and HIV. J Infect Dis 2015; 212:845–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Fulcher JA, Romas L, Hoffman JC, et al. Highly human immunodeficiency virus-exposed seronegative men have lower mucosal innate immune reactivity. AIDS Res Hum Retroviruses 2017; 33:788–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Martin TCS, Chaillon A, Graves SK, et al. Genetic network analysis to assess the risk of HIV transmission among MSM seeking partners on the Internet. Clin Infect Dis 2019:1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Gianella S, Morris SR, Vargas MV, et al. Role of seminal shedding of herpesviruses in HIV Type 1 transmission. J Infect Dis 2013; 207:257–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Fichorova RN, Richardson-Harman N, Alfano M, et al. Biological and technical variables affecting immunoassay recovery of cytokines from human serum and simulated vaginal fluid: a multicenter study. Anal Chem 2008; 80:4741–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Székely GJ, Rizzo ML. Energy statistics: a class of statistics based on distances. J Stat Plan Inference 2013; 143:1249–72. [Google Scholar]
- 24. Kedzierska K, Crowe SM. Cytokines and HIV-1: interactions and clinical implications. Antivir Chem Chemother 2001; 12:133–50. [DOI] [PubMed] [Google Scholar]
- 25. Introini A, Vanpouille C, Lisco A, Grivel JC, Margolis L. Interleukin-7 facilitates HIV-1 transmission to cervico-vaginal tissue ex vivo. PLOS Pathog 2013; 9:e1003148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Azizieh F, Dingle K, Raghupathy R, Johnson K, VanderPlas J, Ansari A. Multivariate analysis of cytokine profiles in pregnancy complications. Am J Reprod Immunol 2018; 79:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kahle EM, Bolton M, Hughes JP, et al. ; Partners in Prevention Herpes Simplex Virus/Human Immunodeficiency Virus Transmission Study Team Plasma cytokine levels and risk of HIV type 1 (HIV-1) transmission and acquisition: a nested case-control study among HIV-1-serodiscordant couples. J Infect Dis 2015; 211:1451–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Rametse CL, Olivier AJ, Masson L, et al. Role of semen in altering the balance between inflammation and tolerance in the female genital tract: does it contribute to HIV risk? Viral Immunol 2014; 27:200–6. [DOI] [PubMed] [Google Scholar]
- 29. Barreto-de-Souza V, Arakelyan A, Zicari S, Margolis L, Vanpouille C. Monocytes but not lymphocytes carrying HIV-1 on their surface transmit infection to human tissue ex vivo. J Virol 2016; 90:9833–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Clerici M, Shearer GM. A TH1–>TH2 switch is a critical step in the etiology of HIV infection. Immunol Today 1993; 14:107–11. [DOI] [PubMed] [Google Scholar]
- 31. Clerici M, Shearer GM. The Th1-Th2 hypothesis of HIV infection: new insights. Immunol Today 1994; 15:575–81. [DOI] [PubMed] [Google Scholar]
- 32. Vingert B, Benati D, Lambotte O, et al. HIV controllers maintain a population of highly efficient Th1 effector cells in contrast to patients treated in the long term. J Virol 2012; 86:10661–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Roff SR, Noon-Song EN, Yamamoto JK. The significance of interferon-γ in HIV-1 pathogenesis, therapy, and prophylaxis. Front Immunol 2014; 4:1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Rihn SJ, Foster TL, Busnadiego I, et al. The envelope gene of transmitted HIV-1 resists a late interferon gamma-induced block. J Virol 2017; 91:1–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Chehimi J, Luo Q, Azzoni L, et al. HIV-1 transmission and cytokine-induced expression of DC-SIGN in human monocyte-derived macrophages. J Leukoc Biol 2003; 74:757–63. [DOI] [PubMed] [Google Scholar]
- 36. Horton RE, McLaren PJ, Fowke K, Kimani J, Ball TB. Cohorts for the study of HIV‐1-exposed but uninfected individuals: benefits and limitations. J Infect Dis 2010; 202(Suppl 3):S377–81. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.