Simian Immunodeficiency Virus Infection Induces Expansion of α4β7⁺ and Cytotoxic CD56⁺ NK Cells

Abstract

Herein we demonstrate that chronic simian immunodeficiency virus (SIV) infection induces significant upregulation of the gut-homing marker α4β7 on macaque NK cells, coupled with downregulation of the lymph node-trafficking marker, CCR7. Interestingly, in naïve animals, α4β7 expression was associated with increased NK cell activation and, on CD16⁺ NK cells, delineated a unique dual-function cytotoxic-CD107a⁺/gamma interferon (IFN-γ)-secreting population. However, while SIV infection increased CD107a expression on stimulated CD56⁺ NK cells, α4β7⁺ and α4β7⁻ NK cells were affected similarly. These findings suggest that SIV infection redirects NK cells away from the lymph nodes to the gut mucosae but alters NK cell function independent of trafficking repertoires.

Human peripheral blood contains two primary subsets of natural killer (NK) cells—a major CD16⁺ CD56^dim subset and a minor CD16⁻ CD56^bright subset. We have defined subsets of rhesus macaque (Macaca mulatta) NK cells and found that, similarly, macaque peripheral blood is dominated by a CD16⁺ CD56⁻ subset but contains two minor CD16⁻ CD56⁺ and CD16⁻ CD56⁻ subpopulations (34). As in humans, macaque CD16⁻ CD56⁺ NK cells are the primary producers of gamma interferon (IFN-γ) and express cell surface markers such as CCR7 and CD62L that mediate homing to lymph nodes, whereas CD16⁺ CD56⁻ NK cells do not express CCR7 or CD62L and primarily mediate cytolytic functions (7, 12, 30, 34). In both humans and macaques, the distribution of NK subsets in peripheral blood is distinct from that observed in lymph nodes and mucosal tissues, where NK cells are primarily CD56⁺ (9, 12, 30, 35).

NK cells are important for the control of multiple viral infections, and increasing evidence suggests that NK cells play a significant role in controlling human immunodeficiency virus (HIV) infection (3, 5, 13, 14, 19, 21, 22, 24, 33), as well as simian immunodeficiency virus (SIV) infection of rhesus macaques and sooty mangabeys (6, 16, 26). HIV and SIV primarily replicate in the gut mucosa (18), and although we and others have demonstrated the presence of NK cells in the gastrointestinal tracts of both humans and rhesus macaques (8, 9, 25, 30), the nature of these NK cells is still poorly understood. Interestingly, the integrin α4β7, which directs lymphocyte trafficking to the gut (4), has been shown to be expressed on peripheral blood NK cells in humans and rhesus macaques (11, 27). This finding suggests that the gut mucosa is a site of active NK cell trafficking.

Despite the importance of gut-associated lymphoid tissue in HIV/SIV pathogenesis, little is known about the effects of infection on NK cell homing to these tissues. In order to address this deficit, a total of 47 Indian rhesus macaques were studied, 27 of which were SIV naïve and 20 infected with either SIVmac239 (5) or SIVmac251 (15) for more than 150 days (mean duration of infection, 293 days). All animals were housed at the New England Primate Research Center and maintained in accordance with the guidelines of the Committee on Animals of the Harvard Medical School and the Guide for the Care and Use of Laboratory Animals (23a).

PBMC isolation and polychromatic flow cytometry staining were carried out using protocols described previously for our laboratory (29, 31); the antibodies used are listed in Table 1. NK cells were defined as CD3⁻ CD8α⁺ NKG2A⁺ (30, 34), and CD16 and CD56 expression were used to delineate three primary subsets: CD56⁻ CD16⁺ (CD16⁺), the dominant subset; CD56⁺ CD16⁻ (CD56⁺); and CD56⁻ CD16⁻ (double negative [DN]) (Fig. 1 A). The results of polychromatic flow cytometry analyses demonstrated that α4β7 was expressed at the highest levels on CD16⁺ NK cells and that, while expression on this subset was not altered during SIV infection, α4β7 was significantly upregulated on both CD56⁺ and DN NK cells in SIV-infected animals (Fig. 1B and C). Interestingly, CCR7, which is expressed only on the CD56⁺ and DN NK cell subsets in macaques (30, 34), was concomitantly downregulated on these subsets of NK cells during chronic SIV infection (Fig. 1B). The relationship between the two markers delineated a dichotomous expression pattern between naïve and SIV-infected macaques (Fig. 1D). This dramatic shift in CD56⁺ and DN NK cell trafficking repertoires is likely indicative of increased homing of these NK subsets to the gut coupled to decreased homing to lymph nodes. Also, as shown in Fig. 1E, the absolute numbers of both CD16⁺ and DN NK cells increased during chronic SIV infection, resulting in increased absolute numbers of gut-homing α4β7⁺ cells in both subsets. Interestingly, while the absolute numbers of all CD56⁺ NK cells tended to decrease during chronic SIV infection, the absolute numbers of the α4β7⁺ CD56⁺ NK cell subset increased slightly (Fig. 1E, middle panel), further suggesting that multiple subsets of α4β7⁺ NK cells increase during chronic SIV infection.

TABLE 1.

Antibodies used in polychromatic flow cytometry analyses

Antibody	Clone	Fluorochromec	Manufacturer
Anti-α4β7	A4B7	APC	NIH NPRRa
Anti-CCR7	150503	Alexa700b	R&D Systems (Minneapolis, MN)
Anti-CD3	SP34.2	APC-Cy7	BD Biosciences (La Jolla, CA)
Anti-CD8α	T8/7Pt-3F9	QDot 605	NIH NPRR
Anti-CD8α	SK1	APC-Cy7	BD Biosciences
Anti-CD16	3G8	Alexa700, PE, FITC	BD Biosciences
Anti-CD56	NCAM16.2	PE-Cy7	BD Biosciences
Anti-CD69	TP1.55.3	PE-Texas Red	Beckman Coulter (Fullerton, CA)
Anti-CD107a	H4A3	PerCP-Cy5.5	BD Biosciences
Anti-IFN-γ	B27	FITC	Invitrogen (Carlsbad, CA)
Anti-NKG2A	Z199	Pacific Blueb	Beckman Coulter

^aNIH Nonhuman Primate Reagent Resource.

^bIn-house custom conjugate.

^cAPC, allophycocyanin; FITC, fluorescein isothiocyanate; PE, phycoerythrin; PerCP, peridinin chlorophyll protein.

FIG. 1.

Comparison of α4β7 expression on NK cell subsets in naïve and SIV-infected macaques. (A) Macaque NK cell subsets were defined as CD3⁻ CD8α⁺ NKG2A⁺ (30, 34) and then further delineated into CD56⁺, CD16⁺, and DN subsets. (B) Representative flow cytometry plots of α4β7 and CCR7 expression on NK cell subsets in naïve and SIV-infected macaques. (C) Percentages of α4β7⁺ cells above the background level were compared between naïve and SIV-infected macaques for CD56⁺, CD16⁺, and DN NK subsets. (D) Relationships between α4β7 and CCR7 expression on CD56⁺ and DN NK cells in naïve and SIV-infected macaques. (E) Absolute numbers of total circulating NK cells were determined by using a bead-based flow cytometric assay as described previously (29, 30), and α4β7⁺ NK cell subset counts were extrapolated using these data combined with NK cell frequency data determined by polychromatic flow cytometry (panel A). Horizontal bars indicate median values for 20 to 27 animals. Student's t tests were used to compare naive and SIV-infected animal groups; P values of >0.05 are considered statistically significant.

Plasma viral loads were also determined for infected animals (range, 30 to 6,500,000 copy equivalents/ml), as described previously (10), but we found no correlation with either α4β7 or CCR7 expression (data not shown). However, even in infected animals with low levels of plasma viremia (i.e., <1,000 copies/ml), α4β7 expression was similar to that in animals with high viremia. This finding suggests that increased NK cell homing to the gut may occur even in instances of low-level viral replication.

We next examined whether α4β7⁺ NK cells were functionally different from their α4β7⁻ counterparts in either naïve or SIV-infected macaques. We analyzed IFN-γ production and CD107a degranulation, as a marker for cytotoxicity, in a dual-function-intracellular-cytokine-staining assay by stimulating NK cells with major histocompatibility complex (MHC)-devoid 721.221 cells using protocols optimized in our laboratory (15, 30). In response to stimulation, CD16⁺ NK cells upregulated CD107a, indicative of a more cytotoxic phenotype (Fig. 2B). However, we also found that, in many animals, a subset of CD16⁺ NK cells secreted IFN-γ; these were found almost exclusively among α4β7⁺ cells (Fig. 2A). Moreover, as indicated by the results of multifunction analysis (SPICE 4.2 software; Mario Roederer, NIH), IFN-γ-secreting CD16⁺ NK cells were not only α4β7⁺ but were mostly dual function, as indicated by their coexpression of CD107a (Fig. 2C), and this functional profile was present in both naïve and SIV-infected macaques. The dominant response of CD56⁺ NK cells to stimulation was IFN-γ secretion, and interestingly, α4β7⁺ CD56⁺ NK cells in naïve animals (although rare) secreted IFN-γ at statistically higher frequencies than their α4β7⁻ counterparts (P = 0.0015, Wilcoxon matched pairs test) (Fig. 2A). Furthermore, although CD56⁺ NK cells had low CD107a expression in naïve animals, this expression was significantly upregulated during chronic SIV infection (Fig. 2B). This expansion was most dramatic in monofunction CD107a⁺ degranulating cells but also occurred in dual-function IFN-γ-secreting cells (Fig. 2C). In infected animals, α4β7⁺ and α4β7⁻ CD56⁺ NK cells had virtually the same functional profiles, suggesting that the expansion of CD107a⁺ cells was SIV induced but occurred independently of gut-homing potential. DN NK cells were hyperresponsive to 721.221 cell stimulation, as manifested by high levels of CD107a expression and moderate levels of IFN-γ secretion (Fig. 2A and B). When the DN NK cells were examined for dual functionality, we observed that, like CD16⁺ NK cells, most of the IFN-γ-secreting cells expressed CD107a, indicative of a dual-function phenotype (Fig. 2C). Interestingly, however, α4β7⁺ and α4β7⁻ DN NK cells had virtually identical profiles in both naïve and SIV-infected macaques, with only a modest but not significant reduction in the frequency of dual-function cells. The fact that the DN NK subset expressed low levels of both CCR7 and α4β7 and had a high degree of both IFN-γ secretion and CD107a upregulation (even more so than the classical CD16⁺ effector population) suggests the possibility that the DN subset may be a less differentiated population than the other NK cell subsets. However, additional studies are necessary to better define the ontogeny of these macaque NK subsets and the in vivo function of the DN subset, especially with regard to potential cytotoxic function.

FIG. 2.

Function profiles of α4β7⁺ and α4β7⁻ NK cell subsets in naïve and SIV-infected macaques. Enriched NK cells were stimulated with 721.221 cells, and IFN-γ production (A) and CD107a expression (B) were measured on α4β7⁺ and α4β7⁻ NK cell subsets in naïve and SIV-infected macaques. The monofunction profile of each subset was determined by expressing each response as a proportion of the total cell subset. Horizontal bars indicate median values for 10 to 12 animals. Blue asterisks indicate statistically significant differences between α4β7⁺ and α4β7⁻ NK cell subsets in naïve animals and red asterisks indicate statistically significant differences between naïve and SIV-infected macaques using the Mann-Whitney U test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. (C) Multiparametric analyses were performed with SPICE 4.2 software (M. Roederer, NIH), and the pie charts represent the functional repertoires of all responding cells (nonresponsive cells are excluded for these analyses). Mean values for 10 to 12 animals are shown. Tables show the results of one-sided permutation tests comparing each of the pies as calculated by SPICE; P values of <0.05 are considered significant and are highlighted in yellow.

Interestingly, CD69 was expressed at the highest levels on CD16⁺ NK cells and was expressed at significantly higher levels on α4β7⁺ NK cells than on their α4β7⁻ counterparts (Fig. 3). These data, combined with the observation that CD69 is globally upregulated on NK cells during chronic SIV infection (30), suggest that α4β7 expression is closely associated with NK cell activation. This is consistent with previous observations in both humans and rhesus macaques showing that α4β7 is upregulated on NK cells with ex vivo interleukin-2 (IL-2) stimulation (27, 28) and that decreased CCR7 expression is associated with increased NK cell activation (17, 20).

FIG. 3.

Increased expression of the activation marker CD69 on α4β7⁺ NK cells and during chronic SIV infection. Percentages of CD69 expression above background staining were measured on α4β7⁺ and α4β7⁻ NK cell subsets in naïve and SIV-infected macaques. Horizontal bars indicate median values. Differences between α4β7⁺ and α4β7⁻ NK cell subsets were analyzed using a Wilcoxon matched-pairs test (black asterisks), and comparisons between naïve and SIV-infected macaques were performed using a Mann-Whitney U test (red asterisks). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Herein we demonstrate independent but overlapping features of macaque NK cell subsets: (i) NK cells in SIV-infected animals display changes in phenotypic markers that suggest a shift in trafficking from the lymph nodes to the gut mucosa; (ii) NK cell subsets can possess both cytotoxic and cytokine-secreting functions that can occur simultaneously—particularly notable with the identification of α4β7⁺ gut-homing dual-function CD16⁺ NK cells, a finding that challenges the conventional wisdom that CD16⁺ NK cells mediate only effector functions; and (iii) NK cell subsets have an inherent plasticity that allows the expansion of cytotoxic features during chronic SIV infection. Interestingly, however, our data suggest that these two phenomena occur independently. Perturbations in NK cell function have been documented both in HIV and SIV infections (1-3, 6, 13, 16, 24), and our findings of increased monofunction and dual-function CD107a⁺ degranulating CD56⁺ NK cells are consistent with these observations. Furthermore, because HIV/SIV replicate primarily in CD4⁺ T lymphocytes found in the gut mucosa (18), increased trafficking of NK cells to the gut could represent a physiologic mechanism of modulating innate immune responses to the dominant site of viral replication. Also, although the absolute increase in α4β7⁺ CD56⁺ and DN NK cells in SIV-infected animals is relatively small compared to the size of the dominant population of α4β7⁺ CD16⁺ NK cells, the fact that these CD16⁻ NK cells have a functional repertoire that is distinct from the repertoire of CD16⁺ NK cells suggests that the shift in NK cell trafficking may have consequences that are disproportionate to their frequencies. However, additional studies of mucosal tissues will be required to confirm the hypothesis that increased expression of α4β7 on NK cells from SIV-infected macaques enhances NK cell trafficking to the gut mucosa.

While the exact mechanisms responsible for increased numbers of circulating α4β7⁺ NK cells remain unknown, they could involve one or more of the following: (i) an overall shift in trafficking of preexisting α4β7⁺ NK cells to gut mucosa, resulting in increased numbers of α4β7⁺ NK cells in the blood; (ii) upregulation of α4β7 on previously α4β7⁻ differentiated NK cells by retinoic acid or dendritic cell imprinting as has been observed for T cells (23, 32); and/or (iii) increased expression of α4β7 as a result of imprinting during NK cell differentiation. Regardless of the mechanism, because gut-homing CD16⁺ NK cells had more dual-function cells than their α4β7⁻ counterparts and CD56⁺ NK cells had increased cytotoxicity coupled to increased α4β7 expression, the result would be greater numbers of monofunction cytotoxic or dual-function cells trafficking to the gut during chronic SIV infection. These data offer new insights into the role of innate immune responses in the control of mucosal SIV replication and raise the possibility that modulation of NK cells may affect future vaccine strategies and/or immunologic therapies for HIV/SIV infection.