Elephant in the room: natural killer cells don’t forget HIV either

Abstract

Purpose of review

Like elephants (and T cells), accumulating evidence suggest natural killer (NK) cells never forget. The description of adaptive or memory NK cells, which can be induced by HIV/SIV infections and vaccines and associated with protective effects in persons with HIV (PWH), has dramatically increased the interest in leveraging NK cells to prevent HIV infection or suppress HIV reservoirs. However, harnessing their full antiviral potential has been hindered by an incomplete understanding of mechanisms underlying adaptive NK cell development and infected cell recognition. Herein, we outline the main discoveries around the adaptive functions of NK cells, with a focus on their involvement in HIV infection.

Recent findings

NK cells with diverse adaptive capabilities, including antigen-specific memory, cytokine-induced and CMV-driven adaptive subsets, likely all play a role in HIV infection. Importantly, true antigen-specific memory NK cells have been identified that mediate recall responses against multiple infectious agents such as HIV, influenza, and SARS-CoV-2. The NKG2C receptor is pivotal for certain adaptive NK cell subsets, as it marks a population with enhanced antibody-dependent functions and has been described as the main receptor mediating antigen-specific responses via recognition of viral peptides presented by HLA-E.

Summary

Antiviral functions of adaptive/memory NK cells have tremendous, but as of yet, untapped potential to be harnessed for vaccine design, curative, or other therapeutic interventions against HIV.

INTRODUCTION

Natural killer (NK) cells are effector cells of the innate immune system that play critical roles in defense against a wide range of viral infections, including rapid elimination of HIV-infected cells without needing prior antigen sensitization [1–9]. Emerging evidence over the past 15 years has challenged the dogma that only B and T cells can exhibit adaptive immune functions, demonstrating that NK cells can also develop antigen-specific memory and recall responses [10–18,19^▪▪,20–22]. This review will summarize the current understanding of various adaptive and memory-like NK cell subsets that have been identified in both mouse models and humans, with a particular focus on their potential relevance and implications in the context of HIV infection.

The first evidence of NK cell memory was demonstrated in Rag2^−/− mice, which mounted hapten-specific NK cell-mediated contact hypersensitivity responses without functional B or T cells [14]. NK cell recall responses could be observed for at least 4 weeks and were antigen-specific as prior sensitization with the same hapten was required to elicit contact hypersensitivity. The same team subsequently reported that vaccines containing viral antigens, such as HIV gag/env, can also promote antigen-specific adaptive responses by hepatic NK cells and identified expression of the chemokine receptor CXCR6 as a requirement for the persistence of antigen-specific adaptive NK cells in the liver [15]. Importantly, this study and a follow-up report demonstrated that adoptive transfer of liver influenza-specific adaptive NK cells protects the recipient mouse against lethal influenza challenge [13,15]. Combined, these findings revealed that murine NK cells can mediate highly specific recall responses against a vast range of nonnative and vaccine antigens, which can be protective against subsequent infections.

Concomitantly with the description of murine CXCR6^pos adaptive NK cells, another subset of murine NK cells endowed with adaptive capabilities and targeting mouse cytomegalovirus (MCMV) was reported [23]. In this case, MCMV infection promoted the clonal expansion of NK cells expressing Ly49H, an activating germline-encoded NK cell receptor that evolved to recognize the MCMV-encoded protein m157. MCMV-experienced Ly49H^pos NK cells persist for months following viral infection and mediate rapid and robust secondary responses, conferring more efficient protective immunity against subsequent MCMV challenges.

Finally, murine memory-like NK cell responses can also be achieved by antigen-independent priming of NK cells, upon homeostatic proliferation [24], or following short exposure to inflammatory cytokines [25]. Upon infusion, interleukin (IL)-12/15/18-preactivated murine NK cells mediate long-lived memory-like responses, proliferate, self-renew, and, combined with irradiation or other therapeutics, can substantially reduce the growth of established tumors in animal models [26–29]. Overall, these initial findings in mouse models dispute the traditional view that only T and B cells can mediate adaptive immune responses and raise the possibility that adaptive NK cells also exist in humans. 

Box 1.

no caption available

HUMAN NK CELL ADAPTIVE FEATURES COME IN VARIOUS FLAVORS

In the decade following these significant discoveries in mouse models, various types of adaptive NK cells were identified in both humans and nonhuman primates, mirroring to some extent those reported in mice. As of this writing, adaptive or memory-like NK cell responses have been described against and/or elicited by a wide range of human pathogens encompassing viruses such as CMV, Epstein–Barr virus (EBV), varicella zoster virus (VZV), hepatitis A and B viruses (HAV and HBV), HIV, influenza, SARS-CoV-2, yellow fever virus, as well as mycobacterial antigens such as bacillus Calmette-Guerin (BCG) [17,18,19^▪▪,20,30–38,39^▪,40–44].

Cytokine-induced memory-like natural killer cells

Human NK cells can exhibit antigen-independent memory-like properties, including the capacity to mediate enhanced effector functions after a short preactivation with certain cytokine combinations that act in synergy, such as IL-12, IL-15, and IL-18, or upon overnight co-culture with tumor cell lines [45–47]. Re-stimulation of preactivated human NK cells using leukemia target cells, cytokines, or FcγRIIIa ligation is associated with increased responsiveness that can be retained for several weeks following their initial differentiation into cytokine-induced memory-like (CIML) NK cells [45,48–52]. Various fundamental processes have been identified as contributors to the potent effector functions of CIML NK cells, such as expression of the high-affinity IL-2 receptor αβγ (IL-2Rαβγ), demethylation of the conserved upstream noncoding enhancer region of the IFN-γ gene, recruitment of anergic unlicensed NK cells, enhanced antibody-mediated functions, release from KIR-mediated inhibition and signaling through the SEMA7A-Integrin-β1 axis [48,50,51,53]. The long-lived properties of CIML NK cells have tremendous potential to be exploited for cancer immunotherapy, and preliminary results from a first-in-human phase 1 clinical trial have shown that NK cells preactivated with IL-12, IL-15, and IL-18 exert robust responses against leukemia targets, leading to remission in a subset of acute myeloid leukemia (AML) patients [49]. Since then, other clinical studies reported beneficial effects of CIML NK cell infusions in patients with relapsed AML [54–57], and preclinical studies have been exploring CIML NK cell-based immunotherapies against other types of cancer, including melanoma, colorectal cancer, ovarian cancer, and hepatocellular carcinoma [58–62].

CMV-driven adaptive natural killer cells

Similarly to MCMV, human CMV (HCMV) and rhesus macaque CMV (rhCMV) both drive the expansion of long-lasting NK cells expressing high levels of the activating heterodimeric receptor CD94/NKG2C and often co-express the marker of NK cell terminal differentiation CD57 [43,44,63–66]. Accordingly, increased proportions of CD94/NKG2C^pos NK cells have been observed following HCMV reactivation in transplanted patients [44,63,67] and have been associated with a better prognosis in patients seropositive for HCMV receiving solid organ transplantation [44,68–70] as well as with viral control in a T–B+NK+ SCID patient experiencing acute primary HCMV infection [71]. In individuals previously exposed to HCMV, expanded NK cells expressing CD94/NKG2C largely overlap with adaptive NK cells deficient for the FcR intracellular gamma signaling chain (FcγRΔg NK cells). FcγRΔg NK cells express high levels of the Fc receptor CD16 and are endowed with extremely robust effector functions when granted specificity through antibody binding, and particularly enhanced expansion and IFN-y production in response to viral antibodies compared to conventional NK cells [41,42,72–74]. Functional specialization of FcγRΔg NK cells has been linked to a distinct epigenetic signature resulting in altered expression in key transcription factors and signaling adaptors [41,42].

Antigen-specific adaptive natural killer cells

Several reports convincingly support the existence of true antigen-specificity and recall responses by NK cells in humans, which have been described against herpesviruses HCMV and VZV, HAV, HBV, influenza, SARS-CoV-2, and HIV [17,18,19^▪▪,20,37,38,39^▪,40]. An initial elegant demonstration of NK cell-mediated recall responses was performed in adults infected with VZV several decades earlier, after injecting them intradermally with VZV-Skin Test Antigen (STA) and using a suction chamber to create a skin blister [17]. Analysis of blister fluids revealed VZV-STA challenge-dependent infiltration by NK cells expressing robust levels of the surrogate cytotoxicity marker CD107a and displaying phenotypic features consistent with tissue residency. CXCR6, the chemokine receptor associated with antigen-specific memory NK cells in mouse models, was also more frequently expressed on VZV-STA blister-fluid NK cells, indicating these adaptive NK cells may have been recruited from the liver.

A subsequent study indicated the presence of antigen-specific adaptive NK cells in cohorts of participants exposed to HBV antigens through either vaccination or infection [18]. This was shown through enhanced cytotoxic and proliferative NK cell responses against autologous monocyte-derived dendritic cells pulsed with HBV antigens in participants who had received HBV vaccination or with chronic HBV infection compared to unvaccinated healthy participants. HBV antigen-specific adaptive NK cells relied on activation through NKG2D to mediate enhanced responses upon re-stimulation, upregulated KLRG1 after challenge with HBV antigen-pulsed targets, and displayed transcriptional features reminiscent of memory T cells.

Further evidence pointing towards the liver as a reservoir for antigen-specific adaptive NK cells was provided by describing a subset of hepatic CD49^posCD16^neg NK cells with adaptive capabilities [20]. These NK cells display enhanced effector functions upon IL-2 and IL-15 stimulation compared to liver-derived CD49^negCD16^pos NK cells, linked to unique transcriptional and epigenetic profiles. Hepatic CD49^posCD16^neg NK cells also mediated robust antigen-specific cytotoxic responses directed against autologous target cells pulsed with HAV and/or HBV antigens, with recall responses being detected only in participants that had been previously exposed through vaccination or infection and with a magnitude similar to that of memory T cells. An additional property of CD49^posCD16^neg adaptive NK cells highlighted in this report is their augmented capacity to migrate based on in vitro assays, which may reflect their ability to egress the liver to reach sites of inflammation, including the skin upon contact hypersensitivity reactions.

While revealing the development of true antigen-specific adaptive NK cells upon exposure to viruses in humans and establishing their potential reservoir in the liver, these studies did not propose clear mechanisms by which NK cells recognize and distinguish different antigens. Partly filling this knowledge gap, more recent studies established that beyond marking adaptive FcγRΔg NK cells with enhanced functionality, CD94/NKG2C endows NK cells with the ability to directly mediate antigen-specific adaptive responses [19^▪▪,37,38,39^▪]. The activating CD94/NKG2C and its inhibitory counterpart CD94/NKG2A both interact with HLA-E bound to a peptide [75–78] and are usually not co-expressed on NK cells [64]. HLA-E mostly presents conserved nonameric peptides derived from leader sequences of other HLA class I (HLA-I) molecules but can also present pathogen-derived peptides [79–83], including peptides that selectively activate CD94/NKG2C^pos NK cells while mediating minimal inhibitory effects through NKG2A [84^▪▪]. In particular, variants of CMV UL-40-derived peptides that mimic canonical HLA-I-derived leader peptides are presented by HLA-E and finely tuned adaptive CD94/NKG2C^pos NK cell functions [37,38]. These observations provided the first evidence suggesting virus antigen-specific NK cell responses are mediated via the CD94/NKG2-HLA-E axis. Building on these findings, recent work showed that the major proportion of antigen-specific adaptive NK cells recognize HIV and other viruses, such as influenza and SARS-CoV-2, through a focused response to viral peptides presented by HLA-E to CD94/NKG2C/A, independently of prior exposure to HCMV [19^▪▪,39^▪]. So far, this is the only described mechanism underlying human antigen-specific adaptive NK cell responses, which have been reported against several pathogens.

ADAPTIVE NK CELLS IN HIV INFECTION

NK cells display robust functional potency against HIV. Initially, epidemiological studies defined particular killer-cell immunoglobulin-like receptor (KIR) genes expressed in conjunction with their HLA ligands associated with significantly slower HIV disease progression and lower viral set-point [85,86]. Functional studies subsequently demonstrated that upon HIV infection, NK cells expand in the peripheral blood prior to the development of virus-specific CD8^pos T cells, inhibit HIV replication in vitro, and mediate significant in vivo immune selection pressure on HIV replication in vivo[1–9]. The critical role of NK cells in virus control was further emphasized in experimental depletion studies in nonhuman primates (NHP) [87]. NK cells have also been associated with protection from infection in HIV-exposed seronegative individuals [3,4,88–91] and may be critical immune correlates for HIV remission in posttreatment viral controllers [92^▪,93]. Finally, durable control of HIV in the absence of antiretroviral treatment has recently been linked to an enrichment in NK cells with a unique transcriptional profile, enhanced effector functionality and memory-like attributes [94]. However, the study of adaptive NK cell subpopulations in persons with HIV (PWH) remains in its early stages.

Cytokine-induced memory-like natural killer cells in HIV infection

Adaptive properties of CIML NK cells are achieved by exposure to specific combinations of cytokines, including cytokines that are elevated in acute HIV infection. Exposure to inflammatory cytokines such as IL-12 and IL-15 was proposed to support the expansion of CD94^posCD56^hi NK cells that express high levels of the TCF7 transcription factor and display adaptive features in PWH [95] (Fig. 1). In particular, this NK cell subset had enhanced proliferative capacity, IFN-γ production, and degranulation in response to HIV-1-infected CD4^pos T cells. CIML NK cells are the only adaptive NK cells currently being evaluated in clinical trials to treat blood cancers [54–57]. A better understanding of the mechanisms behind CIML NK cell responses to HIV may lead to novel strategies to tailor CIML NK cell functions to efficiently target the HIV reservoir, possibly in combination with other approaches to boost Fc-mediated activity or antigen-specific responses.

FIGURE 1.

Phenotypic features of human adaptive NK cells in HIV infection. Schematic indicating the major cell surface molecules reported so far to mark and/or mediate functions by adaptive NK cell subpopulations in PWH. Some markers have been described on bona fide human antigen-specific NK cells in other viral infections or in humanized mouse models and are indicated with a question mark as their role in HIV-specific memory NK cell responses is still unclear. NK, natural killer. Created in BioRender. Kroll, K. (2025). https://BioRender.com/l35e424.

CMV-driven adaptive natural killer cells in HIV infection

Adaptive CD94/NKG2C^pos NK cells proliferate not only in response to HCMV reactivation or infection in patients receiving hematopoietic transplantation [44,63,96–98] but also upon de novo infection with different viruses including HIV, and upon HCMV reactivation in PWH [43,99,100]. HIV infection of people seropositive for HCMV results in a major and long-lasting reshaping of the NK cell repertoire [100,101], leading to the accumulation of mature CD57^posCD85j^posNKG2C^pos NK cells, which end up constituting a large portion of circulating NK cells (Fig. 1). Expanded CD94/NKG2C^pos NK cells include a subset that exhibits adaptive signatures of FcγRΔg NK cells and presents conserved effector functions in PWH [100]. Several reports strongly suggest that HCMV-associated adaptive NK cells improve the control of HIV infection. Higher frequencies of CD94/NKG2C^pos NK cells during primary HIV infection are linked to lower viral setpoints, and predict higher CD4^pos T cell counts and an overall better outcome in treated PWH [102,103]. In contrast, individuals with NKG2C gene deletions are more susceptible to HIV infection and, once infected, may have accelerated disease progression [104].

NK cells limit HIV infection, spread, and progression through various mechanisms, including antibody-dependent cellular cytotoxicity (ADCC), with the signaling downstream of the CD16 Fc receptor being one of the strongest activating triggers for NK cells [105–108]. Adaptive FcγRΔg NK cells are equipped with elevated capacity for ADCC, linked to the replacement of FcεRIγ with CD3ζ, an adaptor protein that delivers a more potent signal through CD16 [66]. HIV antibodies that elicit Fc-mediated NK cell ADCC functions have been associated with reduced risk of infection and HIV control [109–113]. Thus, strategies to optimize FcγR-mediated anti-HIV activity of FcγRΔg NK cells could significantly improve the efficacy of antibody-based approaches to prevent or treat HIV infection [114–119]. For instance, including IL-15 stimulation in therapeutic strategies targeting FcγRΔg NK cells may boost their functionality, as recent work showed that NK cell reprogramming with IL-15 partly restores metabolic requirements necessary for efficient anti-CD16 stimulation in HIV-1 infection [120].

Antigen-specific adaptive natural killer cells in HIV infection

The first clear evidence of bona fide antigen-specific NK cell memory in any primate species was demonstrated in non-human primate models upon HIV/SIV infection and vaccination [11] and suggested a role for the MHC-E/NKG2 axis in the observed recall responses. The ability of human NK cells to mount HIV antigen-specific memory responses was then shown in humanized mice immunized with HIV-encoded envelope protein [17]. Corroborating prior findings in regular mice, recall responses were mediated by HIV antigen-primed CXCR6^pos hepatic NK cells. Finally, investigations in PWH recently established that antigen-specific human NK cell memory develops upon exposure to HIV and indicated that antigen-specific adaptive NK cell responses are largely mediated via recognition of HLA-E–presenting viral peptides recognized by the activating CD94/NKG2C receptor [19^▪▪] (Fig. 1). In this study, the permanent acquisition of antigen-specificity by individual adaptive NK cells was validated using single-cell cloning. The central role of CD94/NKG2C in mediating HIV-specific responses was confirmed by knocking down NKG2C in HIV-specific NK cell clones, resulting in a decrease in the killing of targets pulsed with HLA-E-binding HIV Env peptides. Importantly, individual HLA-E-presented HIV peptides triggered robust and dominant NK cell responses in PWH in vivo and single NK cells that potently respond to these specific HIV-derived peptides could be clonally expanded. In contrast to prior studies, analyzed antigen-specific adaptive NK cells were not derived from the liver but from the peripheral blood, potentially explaining the lack of association between CXCR6 expression and antigen-specific adaptive responses by NK cells. However, the activation marker KLRG1 was upregulated on HIV-specific NK cell clones, in accordance with previous reports describing its expression on subsets of adaptive NK cells [18,32,121], and KLRG1 may therefore represent another useful marker to identify antigen-specific adaptive NK cells. An overlap between antigen-specific memory NK cells described in this study and the previously described CMV-driven adaptive CD94/NKG2C^pos NK cell subsets is possible. However, true antigen-specific adaptive NK cells were scarce in the peripheral blood, suggesting only a minor portion of all CD94/NKG2C^pos NK cells are endowed with antigen-specific features and can develop in people seronegative for HCMV. Moreover, antigen-specific adaptive NK cell responses likely do not solely depend on the NKG2C-HLA-E axis, and future studies will be required to determine complementary or independent pathways of viral recognition by antigen-specific adaptive NK cells. Interestingly, HIV antigen-specific adaptive NK cells appeared to be more abundant in HIV elite controllers compared to progressors [19^▪▪,122], suggesting they could play a key role in controlling viral replication.

Altogether, these reports convincingly point to HIV-specific adaptive NK cells representing a third arm of the immune system that could be targeted through vaccinations/immunotherapeutic interventions to prevent, control, or eradicate HIV and warrant further exploration.

CONCLUSION

In summary, comprehensive characterization of adaptive and memory NK cell subsets in PWH, as well as elucidation of the underlying molecular mechanisms driving their development and antigen-specific functions, will be critical to guide the engineering of future vaccinations and immunotherapeutic interventions. Harnessing the potential antiviral capacity of these specialized NK cell populations in vivo could lead to new strategies to prevent HIV infection or facilitate the elimination of persistent viral reservoirs.

Acknowledgements

We would like to thank Mr. Griffin Woolley for his helpful comments on the manuscript and Mr. Kyle Kroll for his help finalizing the figure.

Financial support and sponsorship

The authors were supported by National Institutes of Health (NIH) grants: R01AI158516 (to S.J.) and R01AI161010, P01AI162242 (to R.K.R.) as well as by the NIH-funded BEAT-HIV Martin Delaney Collaboratory to cure HIV-1 infection by Combination Immunotherapy UM1 AI164570.

Conflicts of interest

There are no conflicts of interest.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest

• ▪▪ of outstanding interest