Changes in the Vα7.2+ CD161++ MAIT cell compartment in early pregnancy are associated with preterm birth in HIV‐positive women

Human immunodeficiency virus (HIV) infection is associated with an increased risk of adverse pregnancy outcomes, including preterm birth (PTB), despite viral suppression with antiretroviral therapy. Mucosal‐associated invariant T (MAIT) cells are an immune cell subset involved in antimicrobial immunity at mucosal surfaces. MAIT cells have been found at the maternal‐foetal interface, and MAIT cells are typically depleted early in HIV infection. We aimed to investigate changes in MAIT cells in relation to maternal HIV/ART status and PTB.


| INTRODUC TI ON
Evidence suggests that 5%-12% of all births worldwide occur prematurely. 1 Complications of preterm birth (PTB) are the primary cause of death among children under 5 years of age, 2 and lead to significant short-term and long-term morbidity. 1 In 2014, UNAIDS reported that 91% of HIV-positive (HIV+) pregnant women resided in sub-Saharan Africa. 3 Antiretroviral therapy (ART)-naive maternal HIV infection in this region is associated with an increased risk of PTB, small for gestational age newborns, low birth weight (LBW) and stillbirth. 4 Although ART improves maternal health and reduces the risk of mother-to-child transmission of HIV, 5 its use in pregnant women in low-and middle-income countries has been implicated in adverse pregnancy outcomes, including PTB. 6 Preconception ART initiation has been reported to be associated with worse perinatal outcomes, including PTB, than antenatal ART initiation. 7 Both HIV infection and pregnancy induce significant immunological changes. HIV infection is characterized by chronic immune activation 8 and depletion of immune cell subsets 9 -most notably CD4 + T cells-resulting in immunodeficiency. Placentation, pregnancy maintenance and the timely onset of labour depend on changes at the maternal-foetal interface involving various adaptive and innate immune cell subsets. 10

Mucosal-associated invariant T (MAIT) cells are important for
anti-bacterial immunity at mucosal interfaces. 11 They express a semi-invariant αβ T-cell receptor comprising an invariant TCR α chain and a restricted range of Vβ chains (Vα7.2Jα33 paired with Vβ2 or Vβ13). 12 They can be activated by vitamin B metabolites of bacteria and yeast presented on the MR1 protein, 12 or by innate cytokines. 11 MAIT cells display an intrinsic effector-memory phenotype and secrete pro-inflammatory cytokines and cytolytic products. 11 Early MAIT cell depletion, irreversible with ART, occurs in HIV+ adults independently of markers of disease severity (eg CD4 + T cell count or viral load). [13][14][15][16][17][18][19] Some suggest that MAIT cells are depleted due to excess activation and subsequent apoptosis 13,15 ; alternative theories include the downregulation of the MAIT cell marker CD161 15,17,20 and PD-1-mediated inhibition of MAIT cell proliferation. 17 In HIV infection, MAIT cells have defective cytolytic function, 21 and reduced IFN-γ and TNF-α expression in response to activating ligands. 18 The role of MAIT cells in pregnancy is incompletely understood. They accumulate in the placental intervillous space in uncomplicated pregnancies 22 and are depleted in the peripheral circulation in early-onset pre-eclamptic patients compared with healthy controls. 23 Given that maternal HIV infection and ART are associated with PTB, and MAIT cells are irreversibly depleted in adults with HIV infection, we investigated circulating MAIT cell frequencies in relation to maternal HIV/ART status and PTB.

| Patients
Blood samples were obtained from patients participating in a prospective cohort study at Chris Hani Baragwanath Academic Hospital, Soweto, South Africa. 24 Patient inclusion criteria were as follows: black South African, living in Soweto, aged 18 or above, spontaneous conception, singleton pregnancy. Exclusion criteria were as follows: multiple pregnancy, body mass index (BMI) >35 kg/m 2 or an intellectual or physical disability.
All patients had a first trimester dating ultrasound scan. HIV testing was routinely offered to those not known to be HIV+. Patient data from medical records, antenatal cards and/or interviews were collected on around 200 items, including socioeconomic characteristics; and smoking, alcohol, drug, medical, gynaecological and obstetric history. HIV disease stage, ART initiation timing, use of ART and ART regimens were documented (with consent) for HIV+ women. Perinatal outcomes were recorded at delivery, with birthweight accurately measured by dedicated research midwives within 12 hours of birth.

| Exposure definitions
HIV+ women were categorized as preconception ART initiators if they had started ART before the date of their last menstrual period (LMP) and as antenatal ART initiators if they had started ART after the LMP date. Two patients in the antenatal ART initiation group had samples collected in the first trimester before the ART initiation date, the remaining patients after the ART initiation date.

| Outcome definitions
PTB was defined as any birth before 37 weeks' gestation, spontaneous preterm labour (Sp-PTL) as spontaneous onset of labour before 37 weeks' gestation, LBW newborns as newborns weighing under 2500g and small for gestational age (SGA) newborns as newborns under the 10th centile of the INTERGROWTH-21st Newborn Standard birth-weight-for-gestational-age/sex. 25 Oxford on dry ice, where they were stored in liquid nitrogen.

| Flow cytometry
South African patient samples were chosen for analysis based on the pregnancy outcome, rather than to be representative of HIV+ or HIV− pregnant women. Healthy HIV-negative (HIV−) non-pregnant control PBMCs were isolated from leukocyte cones (NHS Blood Services, UK) and used to optimize flow staining; these samples are anonymized, so data on age and gender are not available. Frozen PBMCs were thawed in a water bath at 37°C. Each vial of thawed cells (~2.5 × 10 6 cells) was added to 50 μL of DNAse I solution and suspended in warm R10 medium (37°C). To identify live cells, cells were stained with the Zombie Aqua

| Statistical analysis
Patient characteristics at baseline were analysed for normality and compared using the Mann-Whitney U test or unpaired t test for continuous variables, and Fisher's exact test or chi-squared test for categorical variables. Global changes during pregnancy were assessed by comparing values at each trimester with those at every other trimester using the Kruskal-Wallis test. Mann-Whitney U tests were used to compare median cell frequencies/proportions between two groups. For paired samples, a Friedman test followed by a Dunn's multiple comparisons test was used.

| Ethical approval
Written informed consent was obtained from all study participants upon en-

| Patient characteristics
Characteristics of HIV+ and HIV− patients were largely comparable (Table 1). Maternal age was higher among HIV+ women compared to HIV− women. Maternal age was not associated with overall MAIT cell levels, nor with frequencies of MAIT cell subsets in trimester 1, 2 or 3 except for a weak positive correlation seen between maternal age and CD8 + MAIT cell proportions in trimester 1 (Spearman coefficient = 0.295, P = .037). However, when analysing HIV+ and HIV− patients separately, this correlation was not seen (data not shown). There was a significant difference in maternal education, but, importantly, there were no significant differences in parity, obstetric history, pre-pregnancy BMI, smoking status or alcohol intake.

| MAIT cell frequencies remain stable throughout pregnancy
In accordance with previous literature, we defined blood MAIT cells as T cells co-expressing the Vα7.2 TCRα chain and high levels of CD161 26,27 ; the majority of Vα7.2+CD161++ cells stained with an MR1 tetramer loaded with the 5-OP-RU ligand ( Figure 1A). Overall, for all patients, the median MAIT cell frequency is between 0.73% and 1.06% of CD3 + lymphocytes during pregnancy. There was no

| Overall proportions of MAIT cell subsets remain stable throughout pregnancy, but show notable inter-patient variation within trimesters
There was remarkable variation in the proportions of MAIT cell sub- No significant changes in the proportions of MAIT cell subsets during pregnancy, at delivery or postnatally were found overall or in HIV+ women. However, in HIV− women, the proportion of CD8 + MAITs increased during pregnancy and was significantly higher in trimester 3 than in trimester 1 (P < .05, Figure 1J), with a reciprocal decrease in the proportion of DN MAITs (P < .01, Figure 1M); TA B L E 1 Characteristics of HIV-positive (HIV+) and HIV-negative (HIV-) patients

[21] 8 [18]
Spontaneous preterm labour (Sp-PTL) 13 [28] 12 [27] .711 (Continues) the proportion of CD4 + MAITs in HIV− women showed a decreasing trend ( Figure 1G). Paired analysis including only patients with samples for all time points showed a similar pattern, although among HIV− women the increase in the proportion of CD8 + MAITs was not significant, whereas the decreases in the proportions of CD4 + and DN MAITs were significant ( Figure S1). There were no changes in frequencies of CD4 + , CD8 + or DN CD3 + T cells during pregnancy among all women and the HIV+ and HIV− subgroups ( Figure S2).

| HIV-positive women have a higher proportion of CD8 + MAIT cells in early pregnancy compared to HIV-negative women
There are no significant differences in overall MAIT cell frequencies between HIV+ and HIV− women in trimester 1, 2 or 3 ( Figure 2A).
However, HIV+ women have a higher proportion of CD8 + and lower proportion of CD4 + MAIT cells in trimesters 1 and 2 than HIV− women ( Figure 2B,C). The proportion of DN MAIT cells is also lower in HIV+ women than HIV− women in trimester 1 ( Figure 2D). Overall, there is a shift towards CD8 + MAIT cells in early pregnancy in HIV infection.

| Women with preterm birth have a higher proportion of CD8 + MAIT cells in early pregnancy than women with term birth
There are no differences in total MAIT cell frequencies between women with PTB and term birth (TB) in trimester 1, 2 or 3 overall, or within HIV+ and HIV− subgroups ( Figure 4A,E,I). However, in trimester 1, there is a higher proportion of CD8 + MAIT cells in PTB than in TB women overall and within HIV+ and HIV− subgroups ( Figure 4C,G,K), and a lower proportion of CD4 + MAIT cells in PTB women overall ( Figure 4B). Therefore, within HIV+ women, who already have higher CD8 + MAIT proportions in early pregnancy compared to HIV− women, there is a shift towards even higher CD8 + MAIT cell proportions in those with PTB.

| Women with spontaneous preterm labour have a higher proportion of CD8 + MAIT cells compared to spontaneous term labour (Sp-TL)
As Sp-PTL leads to the majority of PTBs, 28

| D ISCUSS I ON
In early pregnancy, the CD8 + MAIT cell proportion is higher (and Elective preterm Caesarean section 8 [17] 9 [20] Small for gestational age (SGA) 19 [40] 11 [24] .221 TA B L E 1 (Continued) We demonstrate no difference in overall MAIT cell frequencies in peripheral blood between HIV+ and HIV− patients during pregnancy, in contrast to previous findings in non-pregnant populations. [13][14][15][16][17][18][19] Given that peripheral MAIT cell frequencies in our study (median between 0.73% and 1.06% of CD3 + lymphocytes) are lower   in early pregnancy may serve as predictive biomarkers for women at risk of delivering preterm.

ACK N OWLED G M ENTS
We thank all the study participants whose blood samples and data contributed to this study. We thank the doctors, nurses and other staff in the Department of Obstetrics and Gynaecology at Chris Hani Baragwanath Academic Hospital who facilitated the study.
We thank Dr C. Wedi for training the study nurses and laboratory technician and assisting with coordination of sample collection and processing at the study site. We thank the sonographers who performed the ultrasound scans and the study nurses who collected the data and blood samples. We thank Ritchie Mata for processing and storing the blood samples.