Non-linear multidimensional flow cytometry analyses delineate NK cell phenotypes in normal and HIV-infected chimpanzees

Definitive markers for chimpanzee NK cells

Abstract

Natural killer (NK) cells are primary immune effector cells with both innate and potentially adaptive functions against viral infections, but commonly become exhausted or dysfunctional during chronic diseases such as human immunodeficiency virus (HIV). Chimpanzees are the closest genetic relatives of humans and have been previously used in immunology, behavior and disease models. Due to their similarities to humans, a better understanding of chimpanzee immunology, particularly innate immune cells, can lend insight into the evolution of human immunology, as well as response to disease. However, the phenotype of NK cells has been poorly defined. In order to define NK cell phenotypes, we unbiasedly quantified NK cell markers among mononuclear cells in both naive and HIV-infected chimpanzees by flow cytometry. We identified NKG2D and NKp46 as the most dominant stable NK cells markers using multidimensional data reduction analyses. Other traditional NK cell markers such as CD8α, CD16 and perforin fluctuated during infection, while some such as CD56, NKG2A and NKp30 were generally unaltered by HIV infection, but did not delineate the full NK cell repertoire. Taken together, these data indicate that phenotypic dysregulation may not be pronounced during HIV infection of chimpanzees, but traditional NK cell phenotyping used for both humans and other non-human primate species may need to be revised to accurately identify chimpanzee NK cells.

Introduction

Natural killer (NK) cells are multifunctional innate immune cells that are among the first mobilized against viral infections and tumors (1–3). As a base definition, NK cells in human blood are subdivided by CD56 and CD16 expression with differing functional profiles—CD56^dimCD16^pos NK cells are typically cytotoxic, whereas CD56^bright NK cells are predominantly cytokine-producing (4). However, among these populations there are further specialized subsets which include tissue-specific NK cell subsets, cells capable of potent antibody-dependent cell-mediated cytotoxicity (ADCC) functions and even memory and memory-like responses (5, 6). The functional activity of these various NK cell responses is controlled by a balance of activating and inhibitory signals (7). Activating receptors include natural cytotoxicity receptors (NCRs) (8) such as NKp30, NKp44 and NKp46, activating killer immunoglobulin receptors (KIRs) and C-type lectin receptors NKG2D and NKG2C, and inhibitory receptors include inhibitory KIRs and NKG2A (9, 10).

NK cells play significant roles in the control of acute human immunodeficiency virus (HIV) infection via secretion of the anti-viral cytokine IFN-γ, CC-chemokines such as RANTES and macrophage inflammatory proteins 1α and β, and their cytotoxic functions (11–13). Strong NK cell responses early in infection are correlated with a lower viral set point (14). However, immune activation and inflammation such as that induced by chronic HIV infection, cause dramatic changes in NK cell phenotypes and functions (15) that include down-regulation of CD56 and NCRs in HIV-infected patients (14). Even after viral immune reconstitution of CD56⁺ mature NK cells in virally suppressed patients, sustained functional impairment was evidenced as decreased IFN-γ secretion by mature NK cells (16). In addition, reduced ADCC activity associated with down-regulated CD16 in cytotoxic NK subsets and expansion of dysfunctional CD56^neg cells are some of the NK cell exhaustion traits commonly described in HIV infection (16–20). In long-term non progressors and elite controllers, NK cells respond with similar ADCC activity as healthy individuals (21, 22). Anti-retroviral therapy partially restores NK cell numbers in both systemic and mucosal tissues (17, 23, 24) and improves ADCC functions of NK cells (25), indicating that NK cells are critical for HIV control, albeit the virus has evolved to evade and modulate NK cell activity.

Non-human primate models are commonly used to study human infections because of their genetic, physiological and immunological similarities to humans. However, NK cell frequencies vary significantly among primate species ranging up to 40% of blood lymphocytes in humans and Old World monkeys including rhesus, cynomolgus and pig-tailed macaques, sooty mangabeys and African green monkeys, to less than 5% in neotropical primates such as common marmosets and cotton-top tamarins (26–31). Variability in non-human primate NK cell phenotypes is also very common. For instance, CD56 is highly expressed on circulating NK cells from humans (where it is also used as a primary delineator) and apes, but is found infrequently on NK cells in neotropical and Old World monkeys (26, 30, 31). Similarly, delineating markers for NK cells in neotropical and Old World monkeys have largely been accepted as NKp46 and CD8α/NKG2A/C, respectively (26, 32). Generally, the functionality of NK cell populations remains conserved in humans and non-human primates despite these disparate phenotypes.

Historically, chimpanzees have played a significant role as animal models in the study of multiple human infections including HIV. For some infections such as Hepatitis B virus (HBV) and Hepatitis C virus (HCV), chimpanzees are the only fully susceptible non-human species and have provided valuable insights into the natural history and pathology induced by those agents (33, 34). Although HIV infection in chimpanzees leads to attenuated disease (35) and generally does not cause mortality, infection and vaccine studies in these animals initially helped identify viral immunity and screen vaccine candidates. However, ethical concerns and limited availability of these great apes have now abbreviated their use in biomedical research, but non-invasive or retrospective studies can still provide important insights into the evolution of the immune system. Previous studies (31, 36–38) identified NK cells in chimpanzees on the basis of CD56 and CD16 gating (used in humans) or CD8α and CD16 (used in Old World monkeys) and even more recently CD16 and CD94 on CD3^– cells. However, these definitions have often been marred by bias and inadvertently included non-NK cells or excluded minor NK cell subpopulations (39). Further clarification is needed to identify consistent NK cell phenotypes in chimpanzees in order to assess NK cell-mediated functions in health and disease. Therefore, in our study, we quantified NK cell markers in naive and HIV-1-infected chimpanzee peripheral blood mononuclear cells (PBMCs) with two goals: (i) to clarify an unbiased and concise NK cell phenotypic definition for the species and (ii) to identify differences in NK cell phenotypes between naive and chronically HIV-1-infected chimpanzees.

Methods

Ethics statement

Blood samples were collected as part of routine veterinary medical examinations, and all research and procedures were IACUC-approved and were carried out in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. Cryopreserved specimens that would have otherwise been discarded were used for this study and no new infections or procedures were initiated. Animals were housed at the Yerkes Primate National Research Center, Emory University reviewed and approved by Institutional Animal care and Use Committee of Emory Institute and at the Laboratory for Experimental Medicine and Surgery in Primates, New York University, the Coulston Foundation Alamogordo, New Mexico were reviewed and approved by the University of Alabama at Birmingham Institutional Animal Care and Use Committee.

Experimental samples

PBMCs from naive and chronically HIV-1-infected chimpanzees were used in this study. All samples were banked from previously approved research and would have otherwise been discarded. No new infections or animal studies were associated with this project.

Flow cytometric staining of PBMCs

Frozen PBMC samples were thawed using standard procedures and analyzed by flow cytometry. Approximately 1–2 million cells were first stained for Live/Dead cell discrimination by incubating with Aqua dye for 30 min at room temperature. This was followed by surface marker staining with a pool of antibodies against the following antigens: CD3 (clone SP34.2, APC-Cy7, BD Biosciences), CD8α (clone T8/7Pt-3F9, QD605, National Institutes of Health Nonhuman Primate Reagent Resource Program), CD14 (clone M5E2, BV650, BD Biosciences), CD16 (clone 3G8, Alexa 700, BD Biosciences), CD56 (clone NCAM1.2, PE-Cy7, BD Biosciences), CD117 (clone 104D2, BV711, BioLegend), CD127 (clone eBIORDR5, Pe-Cy 7, Life Technologies), HLA-DR (IMMU-157, ECD, Beckman Coulter), NKG2A (clone Z199, Pacific Blue, in-house conjugation), NKG2D (clone 1D11, APC, BD Biosciences), NKp30 (clone Z25, PE, Beckman Coulter) and NKp46 (clone BAB281, PE-Cy5, Beckman Coulter). Cells were then washed and permeabilized with Fix & Perm reagents (Invitrogen) and stained for perforin (clone Pf-344, FITC, MabTech) intracellularly. Acquisitions were performed on an LSR II (BD Biosciences) and analyzed by FlowJo v.10 (TreeStar).

Multidimensional data reduction analysis

Flow cytometric data were gated on live HLA-DR^–CD3^– cells (Fig. 1) in FlowJo v.10 and exported with compensated parameters into Cyt (40). t-dependent Stochastic Neighbor Embedding (t-SNE) analysis was performed on the transformed data for NKG2A, NKG2D, NKp30, NKp46, CD16, CD56 and perforin using Barnes-Hut SNE (bh-SNE) (41) approximations. This generated 2-D plots that clustered cells on the basis of marker expression profiles.

Fig. 1.

Characterization of NK markers in PBMCs. PBMCs from naive and HIV-1-infected chimpanzees were analyzed by flow cytometry. (A) Representative gating strategy for HLA-DR^–CD3^– cells in naive PBMCs, and (B) frequency of NK cell markers gated on HLA-DR^–CD3^– cells. Asterisk denotes significant difference between naive and infected at P value <0.05 by non-parametric Mann–Whitney U-test.

Statistical analysis

Statistical differences between naive and HIV-infected chimpanzees were analyzed by Mann–Whiney U-test using Prism software. Differences between the two groups were considered significant at P <0.05.

Results and discussion

Identification of an inclusive and accurate consensus NK cell phenotype has been lacking in the chimpanzee (39). To identify a highly specific marker(s) and thus prevent erroneous NK cell definition, we phenotyped basic NK cell receptors (NKRs) among live HLA-DR^–CD3^– cells from PBMCs of normal and HIV-infected chimpanzees (Fig. 1A and B). Although this population undoubtedly contains NK cells and various other lymphocyte populations such as innate lymphoid cells and monocytes (Supplementary Figure S1), an unbiased analysis was needed since many NK cell markers can be expressed on multiple cells types. NKG2D and NKp46 were the most expressed NKRs in naive animals with medians of 7 and 4% on total lymphocytes, respectively. Low to no expression of NKG2A and NKp30 was observed, similar to other reports (36, 38). CD8α, previously used as one of the primary markers for NK cell definition in chimpanzees (31), was expressed on up to 8% of the total lymphocytes, but was more variable and can be also expressed on dendritic cells and B cells in primates. Further analysis of the live HLA-DR^–CD3^– gating by t-SNE dimensionality reduction identified stable data patterns reconfirming NKG2D and NKp46 as the most highly expressed markers in both naive and HIV-infected animals (Fig. 2). Interestingly, CD56, a primary NK cell delineator in humans, did not fully overlap in expression with other NK cell markers, suggesting CD56 may not accurately identify all NK cells in chimpanzees. The cells expressing NKG2D and/or NKp46 clustered with other NK markers including CD8α, CD16, CD56 and perforin, indicating that this Boolean gating would be helpful in a base phenotype to delineate NK cells and eliminate contaminating cells. Indeed, functional analysis of naive NK cells gated by our definition showed elevated expression of CD107a, an important NK cell degranulation marker, when cross-linked with either CD16 or NKp46 (Supplementary Figure S2). Collectively, both phenotypic and functional data show that a HLA-DR⁻CD3⁻NKG2D⁺NKp46⁺ phenotype denotes NK cells in chimpanzees.

Fig. 2.

Multidimensional data analysis of HLA-DR^–CD3^– cells. Live HLA-DR⁻CD3⁻ gated cells from chimpanzee PBMCs PBMC were analyzed by t-SNE with bh-SNE to generate plots clustering cells with similar expression profiles. Relative expression of individual NK cell markers visualized over the bh-SNE plots as a colorimetric scale of blue (low expression) to red (high expression).

Unlike HIV infection in humans, NCR expression on peripheral NK cells remains generally more stable in HIV-infected chimpanzees (31, 36, 42). A study by Rutjens et al. showed that CD8⁺ NK cells expressed high levels of NCRs in naive animals and was maintained at high levels in chronic HIV-1-infected chimpanzees as well (31). Similarly, in our study, HIV-infected animals did not show changes in expression of NCRs and NKG2 family receptors (Fig. 1B). This is most likely due to the attenuated disease course and limited viremia in infected animals. Instead, the stable expression of dominant NCRs and NKG2 receptors can be used as a reliable NK cell-defining marker in chimpanzees, particularly in HIV infection. Unfortunately, no anti-KIR human antibodies were cross-reactive in our tested chimpanzees (data not shown). Overall, the NKG2D^+/– and/or NKp46^+/– definition of chimpanzee NK cells would allow better resolution of changes in other NK cell markers such as CD8α, CD16 and CD56 that are more likely modulated during infections.

Re-analysis of the same samples with our new gating strategy first identifying NK cells as HLA-DR^–CD3^–NKG2D^+/–NKp46^+/– cells (Fig. 3A) showed no significant differences in NK cell frequency between naive and HIV-infected PBMCs (data not shown). Down-regulated CD16 expression and nearly a 50% decrease in CD8α were evident with our NK cell definition (Fig. 3A) but not when gated on HLA-DR^–CD3^– (Fig. 1B), and is in accordance with other studies (20, 43) that have reported the loss of the CD56^dim CD16⁺ NK cell population in HIV-infected humans. In addition to phenotypic markers, we also characterized the intracellular expression of perforin, an important mediator and indicator of cytolytic function of NK cells. Elevated perforin expression was observed in HIV-1-infected animals in both HLA-DR^–CD3^– cells (Fig. 1B) and HLA-DR^–CD3^–NKG2D^+/–NKp46^+/– cells indicating that these are indeed activated cells with cytotoxic function (Fig. 3A). The perforin-expressing cell clusters predominantly associated with CD16-expressing cells when analyzed by t-SNE (Fig. 3B). A similar report of up-regulated perforin expression has been described in CD16⁺ NK cells of simian immunodeficiency virus-infected macaques (32). A small non-NK cell cluster of perforin-expressing cells was also observed and may represent an as yet undescribed cytolytic population (data not shown). In HIV-infected humans, reduced perforin expression correlated with a low cytotoxic capacity/functional anergy of NK cells (44, 45) and was restored in patients when treated with IFN-α (44). Further, it is tempting to speculate that the stable NCRs are capable of mediating NK cell functions in CD16 down-regulated NK cells since our cross-linking experiments showed that NKp46 cross-linking was capable of eliciting similar levels of degranulation as CD16 cross-linking albeit in naive PBMCs (Supplementary Methods and Figure S2). Thus, our results show that NK cells in HIV-1-infected chimpanzees are functional which may be related to the attenuated disease and lack of chronicity in chimpanzees compared to HIV-1-infected humans.

Fig. 3.

Suggested gating strategy for NK cell identification. (A) Representative gating strategy for delineation of NK cells by gating on both NKp46 and NKG2D positive cells from live HLA-DR^–CD3^– gated cells as in Fig. 1. Asterisk denotes significant differences between naive and infected chimpanzees at P value <0.05 by non-parametric Mann–Whitney U-test. (B) t-SNE with Barnes-Hut approximation (bh-SNE) analysis of NK cells gated on NKG2D and NKp46 markers in naive (pooled, n = 4) and HIV-infected chimpanzee (pooled, n = 5) PBMCs. Relative expressions of single markers were visualized over the bh-SNE plot as a colorimetric scale of blue (low expression) to red (high expression).

Overall, our data suggest that NKG2D and NKp46 are the major markers that can be useful for clear delineation of NK cells in chimpanzees. The unaltered NK cell samples indicate that while infection may activate NK cells, dysfunction is likely less pronounced than in humans. Thus, our study contributes to a better understanding of primate immunology, and also further clarifies limitations of this former disease animal model. Furthermore, these data highlight the importance of multidimensional, rather than single dimension analyses in identifying stable phenotypes for non-human primate lymphocytes.

Funding

This work was supported by National Institutes of Health (grants R01 AI120828 to R.K.R.). This project was also supported, in part, by the Yerkes National Primate Research Center (grant ORIP/OD P51OD011132).

Conflicts of interest statement: The authors declare no conflicts of interests.