Zim CHIC: A cohort study of immune changes in the female genital tract associated with initiation and use of contraceptives

Problem Contraceptive hormones are systemically active, potent, and likely to invoke biological responses other than known fertility regulation impacts. We hypothesized that initiation of depot medroxyprogesterone acetate (DMPA) would increase genital HIV‐target‐cells and soluble immune mediators compared with baseline and initiation of other contraceptive methods. Method of Study We collected cervical cytobrushes and cervicovaginal fluid from healthy Zimbabwean women aged 18‐34 to assess immune cell populations, cytokines, and innate anti‐HIV activity at baseline and after 30, 90, and 180 days use of DMPA (n = 38), norethisterone enanthate (n = 41), medroxyprogesterone acetate/estradiol cypionate (n = 36), levonorgestrel implant (n = 43), etonogestrel implant (n = 47), or copper intrauterine device (Cu‐IUD) (n = 45). Cells were quantified by flow cytometry, cytokines were detected by multiplex assays, and innate anti‐HIV activity was assessed by in vitro HIV challenge. Results Compared to baseline, the number of cervical HIV target cells (#CD4 cells P < .04 and #CD11c cells P < .04), the concentration of the inflammatory cytokine IL‐1β (P < .01), and the innate in vitro anti‐HIV activity (P < .001) significantly decreased following DMPA initiation. In Cu‐IUD users, genital HIV target cells increased (#CD4 cells P < .001, #CD4CCR5 cells P = .02, #CD4CD69 cells P < .001, #CD8CD69 P = .01, and #CD11c cells P = .003) at day 30 and resolved by day 180. IFN‐γ (P < .001), IL‐1β (P < .001), IL‐6 (P < .001), IL‐8 (P < .001), IL‐10 (P < .01), and RANTES (P < .001) were also significantly increased at day 30. Minimal alterations were observed following initiation of subdermal implantable contraceptives. Conclusions This head‐to‐head study compared six contraceptives and found increased HIV target cells and cervical inflammation temporally associated with Cu‐IUD initiation. Use of hormonal contraception, including DMPA, did not increase cervical HIV target cells or inflammation. Clinical Trial Number: NCT02038335


| INTRODUC TI ON
Globally, approximately 60% of all reproductive-aged women use modern contraception, including ~256 million intrauterine device users and ~93 million injectable contraceptive users. 1,2 Observational studies evaluating HIV acquisition risk associated with contraceptive use suggest that use of depot medroxyprogesterone acetate (DMPA) may be associated with increased risk, 3 however, no substantial difference in HIV risk among DMPA, copper intrauterine device (Cu-IUD), and levonorgestrel implant (LNG-I) users was demonstrated in a recent randomized trial. 4 Increased susceptibility to HIV has been reported in women having vaginal dysbiosis [5][6][7][8][9] and increased genital inflammation. [10][11][12] Since immune cells are primary targets for HIV transmission, there is biologic plausibility for increased HIV acquisition risk associated with increased local inflammation, target cell density, dysbiosis, and decreased innate anti-HIV activity, any of which may be impacted by contraceptive use.
Contraceptives vary by composition (hormonal or non-hormonal and types of hormones), route of delivery (oral, transdermal, subdermal, injectable, vaginal, and intrauterine), and resultant drug pharmacokinetics and pharmacodynamics. This heterogeneity, compounded by the phenomenon of frequent contraceptive switching, has challenged careful study of biological changes in women.
We aimed to evaluate the impact of contraceptive initiation and use on genital tract immune cells, soluble mediators of inflammation, and innate anti-HIV activity, all of which could alter HIV susceptibility. We evaluated CD4 and CD8 T-cells, CD11c + antigen-presenting cells (APCs), as well as cellular expression of CCR5 (HIV co-receptor) and CD69 (activation marker), all implicated as necessary or important for acquisition of HIV in women. [13][14][15] We hypothesized that initiation and use of DMPA would increase genital tract HIV target cells and soluble immune mediators compared with baseline and to use of non-injectable contraceptive methods.

| ME THODS
We performed a parallel longitudinal cohort study (ClinicalTrials.gov number: NCT02038335) of healthy Zimbabwean women initiating contraception with an injectable (DMPA, Net-En, MPA/EC), an implant (LNG-I or ENG-I), or a Cu-IUD (copper T380A IUD). The study was designed to assess genital tract changes in HIV target cells, immunoregulatory cytokines, and innate anti-HIV activity compared to baseline with each woman serving as her own control after 30,90, and 180 days of contraceptive use relative to quantified serum progestogen concentrations. The primary objective was change from baseline in lower genital tract HIV target cells after 90-days of continuous contraceptive use. At each visit, we quantified serum progesterone, levonorgestrel (LNG), etonogestrel (ENG), norethindrone (NET), and medroxyprogesterone acetate (MPA) concentrations, covering the spectrum of regionally available progestins as previously described. 16,17 Participants having confirmed non-study hormone use at enrollment or follow-up were not included in this analysis.
We calculated a sample size of 35 women in each group would provide 80% power to detect a 20% change in HIV target cell populations, based on a paired samples t test and a mean CD3 + CD4 + CCR5 + cell recovery from cervical cytobrush at enrollment observed in a prior study. 18,19 To account for the use of non-parametric statistical methods and loss to follow-up, we planned to enroll 50 women per contraceptive group.  Screening included urine pregnancy testing, 2 rapid HIV screening tests, and collection of cervicovaginal fluid by swab to biologically confirm absence of genital tract infections as previously described. 16,20 Eligible participants then presented for enrollment during the follicular phase of menses (self-reported day [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. At enrollment and every follow-up visit, participants were free of vaginal bleeding and had refrained from vaginal and anal intercourse for ≥48 hours (self-report). Contraceptives were administered immediately following sample collection by a study clinician per standard practice at enrollment and as clinically indicated at follow-up visits, with strict adherence to the varied dosing schedules of three inject-

| Sample collection and processing
Peripheral blood mononuclear cells (PBMCs), endocervical cytobrush, neat vaginal fluid, and cervicovaginal lavage samples were   collected at enrollment and at 30, 90, and 180 days, placed on   ice and received by the Zimbabwean laboratory within 2 hours of collection. Blood and genital tract samples for hormonal and flow cytometric analyses were processed as previously described. 16,17 Peripheral blood mononuclear cells (PBMCs) were obtained by venipuncture using Vacutainer ® CPT tubes (BD Biosciences). Tubes were centrifuged at 1300 g for 30 minutes with the brake off. The cell layer was removed, washed twice with DPBS (Corning) by centrifugation at 400 g, resuspended in 2 mL of DPBS, and viable cells were counted using Trypan blue (Sigma-Aldrich) exclusion. 21 Endocervical specimens were obtained by inserting a cytobrush (Cooper Surgical, Trumbull, CT) into the cervical os, rotating 360°, and placing in 4 mL RPMI-1640 medium supplemented with 25 mmol/L HEPES, L-glutamine, and 10% fetal bovine serum (tRPMI). In the laboratory, mucus and cells were recovered by scraping the cytobrush on a polypropylene conical tube while rinsing with tRPMI until no visible material remained. The suspension was filtered through a 70 µm mesh filter (BD Biosciences) using a rubber-tipped plunger from a 5cc syringe while rinsing with tRPMI to push all material through the filter.
We centrifuged the sample at 400 g for 10 minutes, decanted the su-
We filtered the sample using Costar ® Spin-X Centrifuge Tube Filter (Corning), tested the eluent in duplicate for cytokines using magnetic bead kits (EMD Millipore) according to manufacturer instructions, and analyzed with MAGPIX and xPONENT ® 4.2 software (Luminex).

| Innate in vitro anti-HIV activity in vaginal fluid
We treated TZM-bl cells with CVL fluid or control buffer prior to challenge with HIV-1 Bal as previously described 23 to assess innate anti-HIV-1 activity in vitro. Results are reported as percent HIV suppression associated with CVL relative to the buffer-only control. seven hypotheses were tested for immune cell changes (number of cells and percent parent population of CD3 + CD4 + , CD3 + CD4 + CCR5 + , CD3 + CD4 + CD69 + , CD3 + CD8 + , CD3 + CD8 + CCR5 + , CD3 + CD8 + CD69 + , CD3 + CD11c + ) and seven hypotheses were tested for changes in soluble mediators (IFNg, IL-1b, IL6, IL8, IL10, RANTES, innate in vitro anti-HIV activity). To control the type I error for the analyses of multiple endpoints, the Holm-Bonferroni sequential correction 24 was used to adjust the P values presented in the tables. Outcome data are reported as medians with corresponding interquartile ranges and percent change from baseline. Evaluable participants were less likely than non-evaluable participants to select DMPA (15% vs 25%) and more likely to select ENG-I (19% vs 9%) and Cu-IUD (18% vs 11%) (P = .001). Evaluable participants were older (27 ± 4 vs 26 ± 4 years, P = .02) and otherwise did not differ in demographic or sexual behavioral features compared with non-evaluable participants.

| Participant enrollment and demographic characteristics
Among evaluable participants in self-selected contraceptive groups, women opting for DMPA had lower BMI, whereas women opting for Cu-IUD had less frequent intercourse and were less likely married/living with a partner (Table 1)

| Flow cytometry gating and baseline HIV target cell populations
Live single CD3 + T cells were identified from the lymphocyte popula-

| Impact of injectable contraceptives
After initiating DMPA, most HIV target cells evaluated did not change in number (#) or proportion (%) compared with baseline. There were fewer %CD3CD4 + cells 30 days after and fewer #CD3CD4 + cells 180 days in the cervix after DMPA initiation (P < .001 and P = .04 respectively). The #CD11c + APCs also decreased 180 days following DMPA initiation (P = .04) ( Figure 4A and Table 2

| Impact of contraceptive implants
After initiating LNG-I, there were no changes in cervical or systemic HIV target cell populations, cytokines and soluble mediators, or innate in vitro anti-HIV activity ( Figure 5A,B and Table 2).  Figure 5C and Table 2). There were no changes in any of the soluble mediators evaluated or the innate in vitro cervicovaginal anti-HIV activity through 180 days of use ( Figure 5D).

| Impact of contraceptive use on cervical CD8 cells and PBMCs
There were no changes in cervical CD8 cell populations with initiation and use of MPA/EC, LNG-I, or ENG-I (Table S1). Compared  (Table S2).

TA B L E 2 (Continued)
in cervicovaginal fluid at day 30 although the clinical significance of this finding is uncertain. 31 We similarly observed this effect in users  We found that analyzing data at time points of relatively high and low (nadir) progestin concentrations were more informative than time since initiation.
Our study has some limitations, including self-selection of con-

ACK N OWLED G M ENTS
We thank the Zim CHIC study participants and the dedicated research team at the Spilhaus Family Planning Centre and Clinical Research Site.

CO N FLI C T O F I NTE R E S T
The funder had no role in study design, data collection, data analy-

AUTH O R CO NTR I B UTI O N S
SLA designed the study protocol with assistance from LAM and SLH.

I M PLI C ATI O N S
Although the copper IUD has often been used as a non-hormonal comparator, we show that it is not immunologically inert, particularly in the first 30 days following insertion, which is an important framework for data interpretation. Future work focused on investigating biological mechanisms underlying the observed contraceptive-induced immune alterations may benefit future contraceptive and multipurpose product development.

D I SCL A I M ER
The protocol for this clinical study is publicly available at https:// magee womens.org/inves tigat or/sharo n-achil les-md-phd/. The deidentified data that support the findings of this study are available upon request from the corresponding author [SLA]. The data are not publicly available due to containing information that could compromise research participant privacy/consent.