Early and nonreversible decrease of CD161++ /MAIT cells in HIV infection

Abstract

HIV infection is associated with immune dysfunction, perturbation of immune-cell subsets and opportunistic infections. CD161++ CD8+ T cells are a tissue-infiltrating population that produce IL17A, IL22, IFN, and TNFα, cytokines important in mucosal immunity. In adults they dominantly express the semi-invariant TCR Vα7.2, the canonical feature of mucosal associated invariant T (MAIT) cells and have been recently implicated in host defense against pathogens. We analyzed the frequency and function of CD161++ /MAIT cells in peripheral blood and tissue from patients with early stage or chronic-stage HIV infection. We show that the CD161++ /MAIT cell population is significantly decreased in early HIV infection and fails to recover despite otherwise successful treatment. We provide evidence that CD161++ /MAIT cells are not preferentially infected but may be depleted through diverse mechanisms including accumulation in tissues and activation-induced cell death. This loss may impact mucosal defense and could be important in susceptibility to specific opportunistic infections in HIV.

Figures

Figure 1

MR1-dependent activation of CD161++CD8+ T cells by E coli. Healthy isolated PBMCs were exposed to E coli overnight and analyzed for CD69 expression and IFNγ and IL17A production by CD8+ T cells as described in “Methods.” (A) There was a significant increase in the cell surface expression of CD69 on the CD161++ population in response to E coli exposure compared with the mock-infected control and to the CD161+ and CD161−E coli–exposed populations (n = 12). (B-C) A greater proportion of the E coli–exposed CD161++ population produced IFNγ or IL17A than the CD161− population (n = 12). Note data points lie on the x-axis (values of 0% were arbitrarily ascribed a value of 0.01% or 0.001% so as to appear on the log scale). (D) Increasing concentrations of anti-MR1 blocking antibody reduced the level of activation of the CD161++CD8+ T cells compared with the isotype antibody (n = 6). “Activation” on the y-axis is measured as a ratio of the geometric mean fluorescent intensity of CD69 on the antibody-treated E coli–exposed CD161++CD8+ T cells compared with that of the E coli–exposed CD161++CD8+ T cells. Plots are gated on CD8+ T cells.

Figure 2

Loss of CD161++/MAIT cells in HIV infection. Isolated PBMCs from healthy controls (n = 23) and patients with early (n = 35) or chronic stage HIV infection (n = 13) were stained for CD161 expression on CD8+ T cells. (A) There was a lower frequency of CD161++ cells in early and chronic HIV infection compared with the healthy control cohort, with no differences seen in the CD161+ populations. Results are displayed as a proportion of the CD8+ T cell population. (B) Representative FACS plots showing differences in the frequency of CD161++CD8+ T cells. Plots are gated on CD3+ cells. (C-D) PBMCs from the healthy control and chronic HIV infection cohorts were stimulated with PMA/ionomycin as described in “Methods.” The CD8+ T cells were analyzed for production of IFNγ, IL17A, and IL22. The gating strategy is shown in supplemental Figure 9a. (C) Representative FACS plots of PMA and ionomycin stimulated PBMCs in HIV+ patients and healthy subjects. Plots are gated on CD8+ T cells. (D) There were lower frequencies of IL17A producing cells in the HIV+ patients but no difference in the frequency of IFNγ or IL22 producing cells. Note data points lie on the x-axis (values of 0% were arbitrarily ascribed a value of 0.001% so as to appear on the log scale).

Figure 3

CD161++/MAIT cell analysis in the colon in HIV. Sections of colon from HIV-infected patients (n = 12) or controls (n = 12) were stained for CD3, CD8, and MDR-1. CD161++/MAIT cells were defined as CD3+CD8+MDR1++. (A) Representative images from control (left) and HIV-infected tissue (right) are shown. Arrows indicate CD3+CD8+MDR1++ cells. Scale bars, 50 μm. (B) The proportion of CD3+CD8+ cells expressing high (++) levels of MDR1 was reduced in HIV infection. (C) The proportion of CD3+ cells expressing CD8 was increased in HIV infection. (D) No significant difference was seen in the number of CD3+CD8+MDR1++ cells per mm2. (E) The number of CD3+CD8+ cells per mm2 was increased in HIV infection.

Figure 4

CD161++/MAIT cells fail to recover with HAART. Patients with chronic-stage HIV infection were followed for 2 years of HAART, with samples taken before the start of treatment (T0), then at 1 year (T1) and 2 years (T2) into treatment (n = 29). PBMCs were stained for markers of interest including CD161, TCR Vα7.2 and CCR6, and a random selection were stimulated with PMA and ionomycin to analyze the production of IFNγ, IL17A and IL22 (n = 13). PBMCs from healthy controls were stained for surface markers of interest including CD161 (n = 23), TCR Vα7.2 (n = 15), and CCR6 (n = 15) or were stimulated with PMA and ionomycin to analyze the production of IFNγ, IL17A and IL22 (n = 23). (A) All HIV+ patients showed complete suppression of viral load and recovery of CD4+ T cells over the course of treatment. (B) There were no changes in the frequency of CD161++CD8+ T cells or CD161+CD8+ T cells during treatment. (C) CD161++Vα7.2+CD8+ T cells and IL17A-producing CD8+ T cells failed to recover over the course of treatment. Note data points in plots (B) and (C) lie on the x-axis (values of 0% were arbitrarily ascribed a value of 0.001% or 0.0001% so as to appear on the log scale). (D) There was a lower frequency of CCR6+CD8+ T cells in the HIV+ cohort at all treatment time points compared with healthy controls, and this population of cells failed to recover over the course of treatment. There was no difference in the frequency of CCR6+CD4+ T cells between healthy controls and the HIV+ cohort at any of the treatment time points, and no change in the frequency of these cells over the course of treatment.

Figure 5

CD161++/MAIT cells are not preferentially infected in vitro. PBMCs from healthy subjects (n = 9) were activated with PHA, cultured in rhIL2 and rhIL7, and infected with a CCR5-tropic virus (JR-CSF) or CXCR4 tropic virus (MN) at an MOI of 10. FACS analysis was performed on days 6 and 9 after infection as described in “Methods.” The gating strategy is shown in supplemental Figure 9B through D. (A) CD8+ T cells were infected at a low frequency by both viruses, with the greatest frequency of p24+ cells observed on day 6 after infection with JR-CSF. (B) Representative FACS plots of CD3+ lymphocytes from day 6 after infection with JR-CSF. (C) When gating on the total p24+CD8+ T cells, the majority of the infected CD8+ T cells were found in the non-MAIT CD161−CD8+ T-cell population on both day 6 and day 9 after infection with JR-CSF virus. (D) Representative FACS plots of p24 staining of CD8+ T cells mock infected or infected with JR-CSF. (E) Comparison of the frequency of infection within the CD161++/MAIT, and non-MAIT CD161+ and CD161−CD8+ T-cell populations. CD161++/MAIT cells were infected, but not more frequently than either the non-MAIT CD161− or CD161+CD8+ T-cell populations on either day 6 or day 9 after infection with JR-CSF virus. (F) There was no significant difference in the survival of the cells in the HIV infected cultures compared with the uninfected cultures. Cell survival ratio was calculated by normalizing the frequency of CD161++/MAIT cells in the infected culture to that of the uninfected culture.

Figure 6

E coli exposure triggers apoptosis and destruction of CD161++/MAIT cells in vitro. (A-B) Sections of colon from HIV-infected patients (n = 12) or controls (n = 12) were stained for lipopolysaccharide (LPS). (A) Representative images from control (left) and HIV-infected tissue (right) are shown. Scale bars, 50 μm. (B) Significantly more LPS+ cells were seen in the lamina propria in HIV infection. The number of LPS+ cells in the lamina propria was determined by manual counting as outlined in “Methods.” (C-G) PBMCs from healthy subjects (n = 12) were activated with PFA-fixed E coli at a bacteria per cell (BpC) ratio of 1, 10, or 100 or mock-treated and were analyzed after 20 hours incubation. (C) E coli–exposed CD161++CD8+ cells (BpC of 10) had higher frequencies of activated caspase-3–positive cells versus mock-treated cells or E coli–exposed CD161− and CD161+ cell populations. (D) Representative FACS plots, gated on CD8+ T cells, comparing activated caspase-3 expression in mock-treated and E coli–exposed PBMC cultures. (E) At all BpC of E coli, there was a reduction in survival of the CD161++CD8+ T-cell population. No significant differences were noted in the CD161+ or the CD161−CD8+ T-cell populations. Cell survival was determined by normalizing the frequency of the cell population of interest in the E coli–exposed culture to that in the mock-treated culture. A value of 1 is equivalent to 100% survival. (F) Anti-MR1–blocking antibody reduced the frequency of activated caspase-3–positive CD161++CD8+ T cells compared with the isotype control. Blocking of E coli–induced apoptosis was determined by normalizing the frequency of activated caspase 3+ CD161++C8+ T cells treated as indicated to that of the culture exposed to E coli alone. (G) Anti-MR1–blocking antibody increased the survival of CD161++CD8+ T cells in the E coli–exposed culture. Cell survival was determined by normalizing the frequency of CD161++CD8+ T cells in the antibody treated, E coli–exposed culture to that of the mock-treated culture.