Higher mucosal antibody concentrations in women with genital tract inflammation

Inflammatory cytokines augment humoral responses by stimulating antibody production and inducing class-switching. In women, genital inflammation (GI) significantly modifies HIV risk. However, the impact of GI on mucosal antibodies remains undefined. We investigated the impact of GI, pre-HIV infection, on antibody isotypes and IgG subclasses in the female genital tract. Immunoglobulin (Ig) isotypes, IgG subclasses and 48 cytokines were measured prior to HIV infection in cervicovaginal lavages (CVL) from 66 HIV seroconverters (cases) and 66 matched HIV-uninfected women (controls) enrolled in the CAPRISA 004 and 008 1% tenofovir gel trials. Pre-HIV infection, cases had significantly higher genital IgM (4.13; IQR, 4.04–4.19) compared to controls (4.06; IQR, 3.90–4.20; p = 0.042). More than one-quarter of cases (27%) had GI compared to just over one-tenth (12%) in controls. Significantly higher IgG1, IgG3, IgG4 and IgM (all p < 0.05) were found in women stratified for GI compared to women without. Adjusted linear mixed models showed several pro-inflammatory, chemotactic, growth factors, and adaptive cytokines significantly correlated with higher titers of IgM, IgA and IgG subclasses (p < 0.05). The strong and significant positive correlations between mucosal antibodies and markers of GI suggest that GI may impact mucosal antibody profiles. These findings require further investigation to establish a plausible biological link between the local inflammatory milieu and its consequence on these genital antibodies.


Scientific Reports
| (2021) 11:23514 | https://doi.org/10.1038/s41598-021-02954-0 www.nature.com/scientificreports/ attributed the presence of genital tract HIV-specific IgA as a correlate of protection in highly exposed persistently seronegative (HEPS) women [16][17][18] . In the RV144 trial vaccinees, circulating Env V1-V2 IgG correlated with lower HIV-1 risk through enhanced ADCC, ADCP and complement activation 19,20 . Whether these antibodies transduced from the circulation to the mucosal compartment to confer protection remains undefined. Collectively these findings underscore the importance of both locally and/or transduced antibody responses which may help to mitigate HIV infections. Inflammatory cytokines and chemokines influence B cell functions, including antibody production, isotype and IgG subclass switching 21,22 . Previously, cytokines were shown to influence effector functions of antibodies in the HIV-infected individuals 23 . Elevated inflammatory cytokines in the FGT were also associated with increased levels of protease expression leading to loss of antibody functions through proteolytic cleavage 7,24 . Our group showed significant associations between antibodies and pro-inflammatory cytokines in the semen of HIV-infected men 25 shedding light on the profile of mucosal antibodies in relation to local cytokines. Less well defined is the relationship between local cytokines and humoral immunity in the FGT. In this study we measured the antibody isotype and subclasses in the FGT of women from the CAPRISA 004 and CAPRISA 008 1% tenofovir gel trials 26,27 in relation to the mucosal cytokines that are used to gauge GI. Understanding the relationship between pre-existing GI and antibody isotypes and subclasses in the FGT are also important to the design of more effective prevention strategies.

Methods
Study participants and sample collection. This is a retrospective sub-study of N = 132 women from the CAPRISA 004 and CAPRISA 008 clinical trial studies. CAPRISA 004 was a randomized, double-blinded, placebo-controlled trial assessing the effectiveness of tenofovir gel to prevent HIV 27 . CAPRISA 008 was a two-arm, open-label, randomized controlled, non-inferiority implementation trial assessing whether tenofovir gel provision could be effectively integrated into family planning clinics 26 . All participants provided written informed consent for study participation and the sub-study was approved by the Biomedical Research Ethics Committee of University of KwaZulu-Natal (BE0207/17). All the experimental procedures are in accordance with the relevant ethical guidelines and regulations.
Cervico-vaginal lavage (CVL) samples were obtained at the pre-infection timepoint from women who subsequently became HIV-infected (cases, n = 66) and women who remained uninfected (controls, n = 66). Preinfection CVLs for each of the cases were the last HIV negative samples before testing HIV positive by PCR. The cases and controls were matched by time in study and tenofovir use. In controls, their timepoint was matched to the pre-infection timepoint for the cases. All the samples were chosen according to a 1:1 ratio of cases to controls. CVL samples were collected and stored from each woman according to the method described by Bebell et al 28 , CVL sample collection was deferred if women were menstruating. (IgG1, IgG2, IgG3 and IgG4) and isotypes IgA and IgM in CVL were quantified using the Bio-Plex Pro™ Human Isotyping Panel kit (Bio-Rad, USA) according to the manufacturer's instructions. The Bio-Plex 200 multiplex system (Bio-Rad, Hercules, CA) was used to determine the levels of immunoglobulins. Mean Fluorescent Intensity (MFI) were determined by 4-PL logistic regression using the Bioplex Manager 6.0 software (Bio-Rad, Hercules, CA). CVL samples were diluted at 1:10 in sterile PBS to ensure that the MFI for the IgG subclasses and isotypes were detected in the linear range of the standard curve.

IgG subclasses and Ig isotype quantification in CVL. IgG subclasses
Genital cytokine measurement by multiplex ELISA. Concentrations of 48 cytokines involved in different immunological functions were measured using Bio-Plex Pro™ Human Cytokine Group I 27-Plex Panel and Group II 21-Plex Panel (Bio-Rad, USA) and Single-plex Human Magnetic Luminex Assay (R&D systems, USA) in the CVLs. The cytokine panel included: interleukin (IL)-1β, IL-1Rα, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12p70, IL-12p40, IL-16, IL-18, IL-1α, IL-2RA, IL-3, IL-13, IL-15, IL-17, basic fibroblast growth factor (FGF), cutaneous T-cell attracting chemokine (CTACK), eotaxin, granulocyte colony-stimulating factor (G-CSF), granulocyte macrophage colony-stimulating factor (GM-CSF), growth regulated (GRO)-α, hepatocyte growth factor (HGF), interferon (IFN)-γ, IFN-α2, interferon gamma-induced protein (IP)-10, leukemia inhibitory factor (LIF), monocyte chemotactic protein (MCP)-1, MCP-3, macrophage colony-stimulating factor (M-CSF), monokine induced by gamma-Interferon (MIG), macrophage migration inhibitory factor (MIF), macrophage inflammatory protein (MIP)-1α, MIP-1β, nerve growth factor (NGF)-β, platelet derived growth factor (PDGF)-ββ, regulated upon activation normal T cell expressed and presumably secreted (RANTES), stem cell factor (SCF), stem Cell Growth Factor (SCGF)-β, stromal derived factor (SDF)-1α, tumor necrosis factor (TNF)-α, TNF-β, TNF-related apoptosis inducing ligand (TRAIL), and vascular endothelial growth factor (VEGF). The sensitivity of these kits ranged between 0.2 and 45.2 pg/ml for each of the 48 cytokines measured. Data was collected using Bio-Plex Manager software version 6.1, and a 5 PL regression formula was used to calculate sample concentrations from the standard curves. Cytokine levels below the lower limit of detection (LLOD) of the assay were reported as the mid-point between zero and the lowest concentration measured for that given cytokine. To control for inter-plate variability, CVL samples from the same participant were analyzed on separate plates. Baseline  www.nature.com/scientificreports/ mixed models with a random effect of the matched pairs were used to determine associations between cytokines and Ig isotypes and IgG subclasses. The models were adjusted for age, tenofovir use, HSV-2, the number of vaginal sex acts, contraceptives, male condom use and HIV infection status. GI was defined as having at least 5 of 9 elevated cytokines in the 75th upper quartile using previously published scoring criteria 2 . Mann-Whitney U test and unpaired t-test were performed to compare isotype and IgG subclasses between women with GI versus women without GI within cases and controls. Similar tests were also used for the stratified analyses based on tenofovir use within cases and controls to compare isotype and IgG subclasses (supplementary data). Cytokine concentrations were log transformed to ensure normality. P-values were adjusted for multiple comparisons using the Benjamini-Hochberg method. All statistical analysis were performed using GraphPad Prism Version 8.4.3 and SAS version 9.4 (SAS Institute Inc., Cary NC, USA).

Results
Study participants. In this case-control sub-study, a total of 132 mucosal specimens of women from CAPRISA 004 (N = 90) and CAPRISA 008 (N = 42) trials were analyzed. Of these, cases represent women (n = 66) at their pre-HIV infection timepoint who subsequently became HIV-infected. Controls represent women (n = 66) who remained HIV-uninfected and were matched to cases by time in study. No significant differences were found between cases and controls with respect to age, education, relationship status, condom use, lifetime number of partners, HSV-2 positivity at time of study enrolment ( Higher mucosal IgM in cases at pre-HIV infection timepoint. We investigated mucosal IgG subclasses and isotypes in the FGT to determine their differences in the women who become HIV-infected compared women who remained uninfected (Fig. 1) (Fig. 1A). IgG2, IgG3, IgG4 and IgA were similar between cases and controls ( Fig. 1B-D and F).

Mucosal IgG and IgM significantly increases with genital inflammation.
We next determined the effects of GI on antibodies by further stratifying women according to the presence (GI+) or absence (GI−) of GI irrespective of HIV infection. Among cases, women with GI had significantly higher IgG1 (p = 0.042), IgG3 (p = 0.0004), IgG4 (p = 0.0002) and IgM (p < 0.0001) compared to women without GI ( Fig. 2A and C-E). We found no significant difference for IgG2 and IgA among GI+ and GI− women ( Fig. 2B and F). IgG3, IgG4 and IgM remained significant even after multiple comparisons adjustment. Similarly, in controls, women with GI had significantly higher concentrations of IgG4 (p = 0.008), IgM (p = 0.008) ( Fig. 2D and E) and IgG3 trended higher (p = 0.054) (Fig. 2C) compared to women without GI, with IgG4 and IgM remaining significantly higher after multiple comparisons adjustment. This finding highlights the differential profile of antibodies in the presence of inflammation. In order to understand the effect of tenofovir in modifying genital antibodies we further stratified cases and controls by tenofovir use (tenofovir and placebo). Within the cases, and irrespective of the treatment arm, whether women used tenofovir or placebo, concentrations of IgG2 (p = 0.019), IgG3 (p = 0.002), IgG4 (p = 0.0001) and IgM (p = 0.004) were significantly higher in women with GI compared to women without GI ( Supplementary Fig. S1B-E). Similarly, among the controls, significantly higher IgG3 (p = 0.045) and IgM (p = 0.008) were observed in women with GI compared to women without GI ( Supplementary Fig. S2C and E).

Discussion
The majority of HIV infections in women are acquired during sexual intercourse, making it imperative to understand the immune microenvironment of the FGT for developing strategies to prevent infections. A significant effort has been made to understand if and how antibodies can confer protection in the FGT. Given the significant evidence of GI and increased HIV risk, very little is known about how GI can impact the profile and functions of antibodies in the genital tract. Here, we detected that in the presence of pre-existing GI, mucosal IgG subclasses and isotypes were significantly increased in both women who became HIV-infected and in those who remained uninfected. In this case-control analyses we found significantly higher IgM in cases compared to controls which may indicate that mucosal IgM can be an early marker of pathogen exposure. Circulating IgM are elicited during the early stage in response to any infection including HIV 29 . Moreover, HIV-specific IgM were functionally effective in neutralizing and reducing HIV infections in cervicovaginal tissue models 30 . Natural IgM can limit HIV infection by modulating inflammation and T cell activation 31,32 . In addition, IgM can directly bind to CD4+T cells and chemokine receptors CCR5, hindering HIV entry 33 . Together, these data highlight the varied functional capability of pentameric IgM to effectively crosslink and capture virus thereby reducing HIV infection. Upon class switching from IgM, IgG1 predominates during acute and chronic HIV infection 11 . We also found higher IgG1 in cases compared to controls. HIV-specific IgG1 have been associated with viral control and slower progression to HIV infection 34,35 . These data suggest that in the cases, at the pre-HIV infection stage, sexual exposure to HIV or other pathogens may have elicited increased mucosal IgG1 and IgM titres. However, whether these antibodies are functionally competent and HIV-specific IgG or IgM, remains undefined.
To determine genital antibody profiles in the presence of GI, women were stratified for the presence or absence of GI. IgM, IgG1, IgG3 and IgG4 were significantly higher in women with GI compared to women without GI. We previously reported higher titer of HIV-specific IgG and IgA in the genital tracts of women with breakthrough HIV infections who used tenofovir gel from the CAPRISA 004 trial 36 . However, in this study we found that irrespective of tenofovir use there was significantly increased IgG2, IgG3, IgM and IgG4 in the presence of GI. Together these data infer that tenofovir gel use was less likely than GI to modify genital antibodies. However, we cannot over interpret these data and the limited sample size in each sub-category precludes accurate conclusions. Both in vivo and in vitro studies suggested that IgM increases as a compensatory drive to resolve inflammation 32,37 . Therefore, increased mucosal IgM among women with GI may be in response to the inflammation. Studies have demonstrated cytokine-induced enhancement of IgG subclasses and isotypes 38,39 explaining  www.nature.com/scientificreports/ in part the increased antibodies with GI. Alternatively, compromised barrier integrity may also account for the increased antibody transudate in the presence of GI. Inflammation is strongly associated with the glycosylation of the fragment crystallizable (Fc) region of IgG subclasses particularly 40 . This reduces antibody affinities for Fcγ receptors decreasing functional activities 41,42 . Besides glycosylation, increased proteases have been associated with elevated inflammatory cytokines in the FGT 7 which can lead to non-specific proteolytic antibody degradation compromising the function of the antibodies 24 . Therefore, the impact of GI on antibody functions in the genital tract need to be investigated in future studies. Nevertheless, the findings suggest that inflammatory environment in the FGT may be responsible for maintaining specific antibody responses, but this needs further investigation in large cohort studies. Cytokines have been shown to influence Ig class switching and subclass synthesis 39,43 . In this study we demonstrated a link between genital inflammatory cytokines and mucosal antibodies. We found that prior to HIV infection, mucosal IgG subclasses and isotypes were positively and significantly associated with several inflammatory cytokines: TNF-α, IL-1α, IL-1β, IL-10 and IL-6; adaptive cytokines: IL-13, IL-4 and IFN-γ and chemotactic and growth factors: MIP-1α, MIP-1β, MCP-1, IL-8, CTACK and MIG after adjusting for potential confounders. Positive associations were observed between IL-13 and IgG4 and IgM. These data corroborate previous findings of IL-13 induced synthesis of IgM and IgG4 in human B cells in the presence of activated T cells 44 . Furthermore, IL-6 along with IFN-γ can inhibit IgG1 and enhance IgG2 production in stimulated PBMCs demonstrating the differential regulation of cytokines on human IgG subclass production 39 . The significantly direct association between IL-7 and IgG4 is also supported by a study showing IL-7 driven class switching to IgG4 45 . Collectively, these preliminary findings suggest that the local isotypes and IgG subclasses may be influenced, in part, by preexisting GI.
The strength of the study is the evaluation of rare pre-HIV infection mucosal samples from high-risk South African women in two clinical trials investigating mucosal Ig isotype/subclasses in relation to cytokines. Mucosal antibody profiling may be an additional predictor of pre-existing GI. However, these studies would indeed require large cohorts evaluating both mucosal cytokines and antibodies longitudinally and in parallel to establish their validity as an additional surrogate for GI. Understanding the association between GI and antibodies can provide further insight for developing effective HIV prevention strategies for at risk populations with underlying www.nature.com/scientificreports/ susceptibilities to GI. This study however has several limitations. We did not measure pathogen-specific antibodies, an important surrogate indicator of bacterial and viral pathogen exposure. All the participants were on some form of hormonal contraceptives, a mandatory inclusion criterion for trial participation. Therefore, we could not account for phase of the menstrual cycle, which is known to affect inflammatory cytokines and antibodies 46,47 . A previous study of women on DMPA showed significantly increased IgG1, IgG2 and IgA and IL-1α, IL-1β,  IL-2, MIP-1β, IP-10, IL-8 and TGF-β cytokines in the FGT 48 . These data affirm the local impact of exogenous hormones on immune responses in the FGT. In this study we adjusted for hormonal contraceptive use as one of the potential confounders in the linear mixed model. However, further studies are needed to understand the impact of endogenous and exogenous hormones on GI and mucosal antibodies. Another caveat to the present study is the lack of data for STIs and BV in the CAPRISA 004 trial which precluded the ability to establish significant associations of GI with STIs and BV. Previous studies have shown that both STIs and BV can induce inflammation in the genital tract impacting the mucosal immune responses 49,50 . Inflammatory cytokines in the FGT were similar in women with symptomatic and asymptomatic STIs 51 . In addition, women with asymptomatic STIs had significantly elevated cytokines compared to women with no STIs or BV 51 . Toll-like receptors (TLRs) recognize pathogens including STIs inducing specific T-cell and antibody response 52,53 . In addition, TLRs recognize pathogen associated molecular patterns (PAMPs) of bacterial STIs inducing NF-κB driven immune response resulting in inflammatory cytokine production 54,55 . Therefore, to exclude the bias of GI due to STIs and BV, these two factors should be adjusted for as confounders. Future studies would, therefore, require large sample sizes to account for the selection bias and confounding of GI secondary to STIs and BV. Despite these shortcomings, our data provides an enhanced dissection of how GI can affect the quantities of mucosal antibodies and, potentially impact their functions through the predominance of certain isotypes or subclasses over the other. Our findings suggest that GI can shape or reconfigure the mucosal Ig subclass and isotype signature. However, further investigation is required to verify a plausible link between the local inflammatory milieu and the effect on genital antibodies and their functions. These data may be important for future HIV vaccine efficacy studies.