Publications

Invariant natural killer T (iNKT) cells are innate-like T cells that respond to lipid antigens presented by CD1d. These immunoregulatory cells have the capacity for rapid cytokine release after antigen recognition and are essential for the activation of multiple arms of the immune response. HIV-1 infection is associated with iNKT cell depletion in the peripheral blood; however, their role in the gastrointestinal-associated lymphoid tissue (GALT) is less well studied. Our results show that iNKT cells are found at a higher frequency in GALT compared with blood, particularly in HIV-1 elite controllers. The capacity of iNKT cells to produce interleukin-4 (IL-4) and IL-10 in the GALT was associated with less immune activation and lower markers of microbial translocation, whereas regulatory T cell frequency showed positive associations with immune activation. We hypothesized that the composition of the microbiota would influence iNKT cell frequency and function. We found positive associations between the abundance of several Bacteroides species and iNKT cell frequency and their capacity to produce IL-4 in the GALT but not in the blood. Overall, our results are consistent with the hypothesis that GALT iNKT cells, influenced by certain bacterial species, may have a key role in regulating immune activation in HIV-1 infection.

BACKGROUND

HIV-infected women risk sexual and perinatal HIV transmission during conception, pregnancy, childbirth, and breastfeeding. We compared HIV-1 RNA suppression and medication adherence across periconception, pregnancy, and postpartum periods, among women on antiretroviral therapy (ART) in Uganda.

METHODS

We analyzed data from women in a prospective cohort study, aged 18-49 years, enrolled at ART initiation and with ≥1 pregnancy between 2005 and 2011. Participants were seen quarterly. The primary exposure of interest was pregnancy period, including periconception (3 quarters before pregnancy), pregnancy, postpartum (6 months after pregnancy outcome), or nonpregnancy related. Regression models using generalized estimating equations compared the likelihood of HIV-1 RNA ≤400 copies per milliliter, <80% average adherence based on electronic pill caps (medication event monitoring system), and likelihood of 72-hour medication gaps across each period.

RESULTS

One hundred eleven women contributed 486 person-years of follow-up. Viral suppression was present at 89% of nonpregnancy, 97% of periconception, 93% of pregnancy, and 89% of postpartum visits, and was more likely during periconception (adjusted odds ratio, 2.15) compared with nonpregnant periods. Average ART adherence was 90% [interquartile range (IQR), 70%-98%], 93% (IQR, 82%-98%), 92% (IQR, 72%-98%), and 88% (IQR, 63%-97%) during nonpregnant, periconception, pregnant, and postpartum periods, respectively. Average adherence <80% was less likely during periconception (adjusted odds ratio, 0.68), and 72-hour gaps per 90 days were less frequent during periconception (adjusted relative risk, 0.72) and more frequent during postpartum (adjusted relative risk, 1.40).

CONCLUSIONS

Women with pregnancy were virologically suppressed at most visits, with an increased likelihood of suppression and high adherence during periconception follow-up. Increased frequency of 72-hour gaps suggests a need for increased adherence support during postpartum periods.