The role of natural killer cells in preventing HIV-1 acquisition and controlling disease progression.
In sub-Saharan Africa, women carry a disproportionate burden of the Human Immunodeficiency Virus Type 1 (HIV-1) pandemic. The high risk of HIV acquisition in these women and the variability in their disease progression is not fully understood. Natural Killer (NK) cells, which are innate immune antiviral lymphocytes, present systemically and at mucosal surfaces, may play a role in preventing HIV acquisition and/or altering disease progression, as they are key early mediators of the response to viral infections and are equipped to kill infected cells. The purpose of this study was to evaluate the role of NK cells in HIV-1 acquisition and following acquisition, in disease progression. The study participants were selected women who were participating in a randomized controlled trial assessing the effectiveness of 1% Tenofovir gel in preventing HIV-1 (CAPRISA 004 trial). The study design was a case-control study nested within the cohorts followed up in the CAPRISA 004 trial. In this trial, 889 sexually-active women aged 18-40 years were randomized to receive Tenofovir or placebo gel and prospectively followed. Assessment of HIV infection was performed monthly by rapid HIV-1 antibody tests, supplemented by HIV-1 RNA polymerase chain reaction (PCR), p24 Western blotting and/or ELISA. Samples obtained prior to the first positive rapid antibody test were retrospectively tested by HIV specific PCR to identify window period infections. The date of infection in this study was estimated as the midpoint between the last negative and first positive antibody test, or 14 days prior to the first HIV-1 RNA-PCR positive result. Multi-parametric flow cytometry techniques developed and validated in healthy blood donors were used to asses the bidirectional relationship between NK cells and HIV-1. To simulate in vivo interaction between NK cells and autologous HIV infected cells, an in vitro infection and coculture assay was used in addition to conventional assays of NK cell recognition of HLA-deficient cell lines. These were supplemented with measurement of plasma cytokines by Luminex and microbial products by ELISA. In this study, 44 cases who acquired HIV-1 were sampled prior to infection and 39 controls who remained HIV-1 negative despite high behavioural exposure at the timepoint when their preceding sexual activity was highest. To understand how HIV infection affected NK cells during early HIV-1 infection, the first sample obtained after acquisition was studied and compared to preinfection samples from the same participant. The case and control groups were broadly similar in the proportions using tenofovir gel, proportions infected with HSV-2 and number of sexual partners but tended to be marginally older than cases (27.6 vs 23.3 years). By design control women had higher sexual activity than cases (mean 11 vs. 5.7 sex acts per month). The frequency of IFN-γ secreting NK cells from women who acquired HIV infection were significantly lower than from women who remained uninfected in response to 721 cells-an EBV transformed B cell line (background-adjusted median 13.7% vs. 21.6%; p=0.03) and to autologous HIV infected T-cells (background-adjusted median 0.53% vs. 2.09%; p=0.007). NK cells from HIV acquirers displayed impaired proliferation but enhanced spontaneous degranulation compared with non-acquirers after co-culture with HIV uninfected or infected autologous T-cell blasts. Adjusting for age, gel arm, HSV-2 infection status and levels of NK cell activation, IFN-γ+ NK cell responses to autologous HIV infected cells were associated with reduced odds of HIV acquisition (OR 0.582; 95% CI 0.35-0.98; p=0.04). In addition, even in the absence of ex vivo stimulation, HIV acquirers had higher levels of generalised innate immune activation measured by systemic cytokine concentrations (TNF-α, IL2, IL-7 and IL12p40), peripheral blood platelet concentrations (p=0.038), and non-specific ex vivo NK cell activation (p<0.001). Generalised NK cell activation measured directly ex vivo without stimulation was associated with acquisition. Further, if innate immune activation was assembled as a principal component in an unsupervised fashion but taking into account all the measures made, it was significantly associated with HIV acquisition (OR adjusted for age, tenofovir gel use, and HSV-2 status for PC with innate immune factor loadings 11.27; 95% CI 1.84- 69.09; p=0.009). The causes of preinfection innate immune activation could not be established in this study but the degree of activation could not be explained by microbial translocation as both HIV acquirers and non-acquirers had similar levels of plasma lipopolysaccharide (LPS), soluble CD14 (sCD14) and intestinal fatty-acid binding protein (I-FABP). Similarly, both HIV acquirers and non-acquirers had similar NK cell and cytokine responses to Toll-like Receptor (TLR)-2, 3 or 7/8 agonists 11. During early HIV-infection, NK cells demonstrated significantly higher activation (p=0.03), expression of Killer-cell immunoglobulin-like Receptors (KIR) (p=0.006) and expression of chemokine receptor 7 (CCR7, p<0.0001) compared with prior to acquisition. Although NK cells had reduced cytolytic potential following HIV acquisition, antiviral IFN-γ secretion appeared to be preserved. NK cell responses were not different between tenofovir and placebo gel recipients, but women who acquired HIV whilst using tenofovir gel had higher gag-specific IFN-γ CD4+ T-cell responses during early infection. Overall, the findings suggest that the frequency of NK cells producing IFN-γ specifically after co-culture with HIV-1 infected target cells was associated with protection from HIV-1 acquisition but, generalised, non-specific activation of NK cells and other innate immune components enhanced HIV acquisition. Since neither microbial translocation nor TLR responsiveness were associated with pre-existing immune activation further studies will be required to identify the drivers of generalised innate immune activation. Methods to dampen generalised innate immune activation and/or augment specific NK cell antiviral responses in women at risk for HIV-1 may reduce HIV-1 acquisition. During primary HIV-1 infection, NK cells underwent impairment of cytolytic function but not IFN-γ secretory function; this may affect their ability to affect disease progression. Although Tenofovir gel did not alter innate immune responses in women with breakthrough infection, it preserved HIV-specific Tcell immune responses, the consequences of which need further exploration. Understanding how Tenofovir gel mediated preservation of adaptive immune responses may lead to interventions that will reinforce protective host responses. In conclusion, innate immune responses by NK cells have been shown to impact HIV acquisition; HIV-specific IFN-γ responses by NK cells were protective while generalised NK activation was detrimental. The causes of innate immune activation are not known but these effects were independent of the impact of Tenofovir gel. Future prevention strategies targeting mucosal transmission of HIV should assess their impact on NK cell responses, to avoid general innate immune activation and enhance their ability to protect against HIV acquisition.