HCMV jogs the ‘memory’ of NK cells in HBV

The treatment paradigm in hepatitis B virus (HBV) is rapidly transforming with many new viral and immune therapies entering the clinical trial pipeline. Although many of the immune targets focus on restoration of the exhausted virus-specific T cell response [1] an increasing body of work makes a case for NK-based immune therapies and vaccines [2]. Advances in NK cell biology have revealed that these ‘innate’ cells may also possess adaptive features. These cells ‘sitting-on-the-fence’ of innate-adaptive immunity, undergo significant changes in their repertoire during differentiation. A repeating stimulus such as cytomegalovirus (CMV) infection may result in the accumulation of highly differentiated NK cells, designated as ‘memory-like’. These ‘adaptive’ NK cells rapidly expand and generate an effective ‘recall’ responses. Features of memory-like NK cells have not been widely reported in chronic hepatotropic infections (HBV and HCV) where co-infection with human CMV (HCMV) is usually universal [3]. In this issue, Schuch, Zecher et al., describe for the first time the expansion of memory-like NK cells in patients with chronic HBV infection in conjunction with HCMV [4].

In the setting of HCMV, which has a broad impact on immune responses and induces an adaptive reconfiguration of the NK cell compartment; NK cell subsets predominantly express the terminal differentiation marker CD57, important in the control of HCMV. These cells also express the transmembrane protein DAP-12 coupled with NKG2C, a receptor of the C-type lectin family and the inhibitory CD85j [5]. This distinct memory-NK cell subset has a variable expression of cell surface receptors, reduced expression of signalling proteins and transcription factors compared to conventional NK cells. Epigenetic modifications are also known to drive the emergence and persistence of HCMV-adapted memory-NK cells [6]. There is now an increased focus on studying ‘antigen’-specific (memory) NK cell responses in humans, previously neglected despite their description in mice. In HIV memory-NK cells have been shown to demonstrate increased antibody-dependent cell cytotoxicity (ADCC) and a potential reduced capacity for T cell regulation, unlike conventional NK cells [7, 8].

In HBV infection, previous reports have described the expansion of functional NK cells following exposure to therapeutic Pegylated Interferon alpha (Peg-IFNα) with or without nucleos(t)ide analogue therapy [9, 10]. This expansion is however of proliferating functional CD56^bright NK cells, mediated via IL-15. This is of importance as the liver is enriched with CD56^bright NK cells and in a study where mice lacked T and B cells, CXCR6-postive liver NK cells were found to mediate hapten-specific hypersensitivity [11]. Prolonged expansion of NK cells with antigen-specific responses is disparate to the traditional view where NK cells are considered short-lived populations with rapid turnover and contraction following acute responses. However, it is now accepted that NK cells are diverse, possess both innate- and adaptive-like properties and even memory like-NK cell subsets may have differences in receptor, intracellular molecule and transcription factor expression. The in vitro exposure of NK cells to cytokine combination of IL-12/15/18 has shown the potential to generate memory like cells with increased effector functions [12]. The signalling pro-inflammatory cytokines appear to be pivotal to the long-term maintenance of NK cells with memory-like responses; IL-12/15 being especially important for CMV-specific, cytokine-induced and liver-resident memory NK cells. NK cell memory may be antigen-specific or antigen-independent. Antigen-independent memory like NK cells may expand following exposure to specific cytokines which can imprint long-lasting changes on their effector functions [13].

Schuch, Zecher et al, confirm that HCMV infection is common in HBV-infected patients and report the expansion of FcεRIγ- CD56^dim NK cells in HBV compared to healthy donors [4]. It has been reported that NK cells deficient for the adaptor protein FcεRIγ convincingly associate with HCMV [14]. In this article, FcεRIγ- CD56^dim NK cells have been designated as ‘memory-like’ NK cells as they share many of the characteristics of adaptive NK cells. The Fc chain associates with the transmembrane portion of CD16 to form the low-affinity IgG, FcγRIII also expressed on NK cells and thus they respond to CD16 stimulation and akin to NKG2C+ NK cells mount robust responses against HCMV-infected targets [15].. In line with this, the authors show augmented responses of memory-like FcεRIγ- CD56^dim NK cells to CD16 stimulation [4]. The expression of signalling molecules EAT2, SyK and transcriptional regulators PLZF, Helios was analysed. Together with the down-regulation of CD7, the expression of EAT2, SyK, PLZF and Helios was also reduced on FcεRIγ- CD56^dim memory-NK cells, with the noted expansion of CD57 in keeping with previous reports [15].

NK cell receptors exhibit varying expression in HBV. The activatory C-type lectin NKG2C was up-regulated on memory-like FcεRIγ- CD56^dim NK cells, with a compensatory down-regulation of the inhibitory NKG2A [10]. NKG2C and NKG2A are often mutually exclusive and thus the overexpression of NKG2C can shift the response to the non-classical MHC class I molecule, HLA-E from inhibitory, as seen with a predominance of NKG2A+/NKG2C- cells to an activating phenotype; NKG2C+/NKG2A−. Schuch, Zecher et al., also report reduced expression of the natural cytotoxity receptors (NCRs) NKp30 and NKp46 along with CD2 and Siglec-7 on memory-like FcεRIγ- CD56^dim NK cells, however the expression of these markers was also markedly different on bulk CD56^dim NK cells in HBV infected patients compared with healthy donors. Interestingly, the expression of CD2, NKG2A, NKG2C, NKp30, NKp46 and Siglec-7 on the bulk CD56^dim NK cell population correlated with their frequency on memory-like FcεRIγ- CD56^dim NK cells [4]. CD2 has been identified as a key co-stimulatory receptor, which like the Fc receptors, contributes to increased cytokine production in adaptive NK cells after synergising with CD16 [16].

NK cells have been shown to expand and proliferate in response to IL-15, which is also described for memory NK cells with features of ‘recall’ capacity [12]. The memory-like FcεRIγ- CD56^dim NK cells investigated here display a resting phenotype with the reduced proliferative capacity of FcεRIγ- CD56^dim NK cells vs. FcεRIγ+ CD56^dim NK cells, as previously described FcεRIγ deficient NK cells [14], ,especially in healthy donors, as HBV-infected patients did retain Ki67+ CD56^dim NK cells. This once again highlights the diverse phenotype of adaptive/memory NK cells. BCL2 analysis showed significant up-regulation on FcεRIγ- CD56^dim NK cells compared with the FcεRIγ+ subset [4]. Similarities in global epigenetic modifications have been described between adaptive NK cells and memory CD8 T cells [6] which drive the alteration in metabolic pathways and impact functional properties of NK cells [17]. Glucose uptake was reduced, yet polarised and functional mitochondria were increased on memory-NK cells; a pre-requisite for mitochondrial oxidative phosphorylation (OXPHOS). Interestingly, while memory-like and conventional NK cells differed in their metabolic potential, this was not influenced by HBV infection [4]. In contrast, metabolic requirements of CD8 T cells in HBV demonstrate mitochondrial dysfunction; exhausted antigen- specific T cells are dependent on abundant glucose supplies, unlike functional CMV-specific T cells which can utilise OXPHOS in the absence of glucose [18].

Importantly memory-like FcεRIγ- CD56^dim NK cells in HBV demonstrate a CD16-dependent increase in degranulation with increased IFNγ production. Intriguingly in response to stimulation with cytokine cocktails IL-12/15/18, IFNγ production was diminished from memory-like FcεRIγ- CD56^dim NK cells. Certain transcription factors, such as PLZF interact with several target genes, including IL12RBw, IL18RAP and KLRB1, which may account for the lack of responsiveness to cytokine stimulation [6]. DNA methylation is another feature of adaptive NK cells, silencing PLZF, SyK and EAT2 in line with the report by Schuch, Zecher et al,. Investigating DNA methylation of memory-like FcεRIγ- CD56^dim NK cells, Schuch, Zecher et al., analysed the expression FcεRIγ and Helios on NK cells. FcεRIγ+/Helios+ and FcεRIγ-/Helios+ make up the largest and smallest subsets of bulk the CD56^dim NK cells respectively, but importantly CD16-induced degranulation was increased on the FcεRIγ-/Helios- subset, compared to conventional FcεRIγ+/Helios+ NK cells. Hypermethylation of the FCERIG promoter in CD56^dim NK cells lacking FcεRIγ and CpG hypomethylation of the IFNG CNS1 and IFNG promoter region in FcεRIγ-/Helios- cells was demonstrated, confirming that FcεRIγ and Helios expression phenotypically, functionally and epigenetically regulate memory-like CD56^dim NK cells in HBV [4].

Schuch, Zecher et al,. in HBV, have prompted an avenue of previously neglected research into memory-NK cells in relation to HCMV. This is critical for a number of reasons and not least as the liver is a major site of HCMV infection [19] and the expansion of adaptive NK cells may serve to combat HCMV, whether this is also true for hepatotropic infections remains to be seen. As noted, in this article liver tissue-specific memory-like NK cells were not analysed and thus tissue-related HCMV activations cannot be excluded, but should be a line of further investigation along with the analysis of transcriptionally distinct tissue-resident NK cells [20]. In addition impact of age, gender and ethnicity of these cells in HBV is required along with the exploration of inhibitory markers such as PD-1 [21], ADCC persists in FcεRIγ- NK cells in HIV despite viral load reduction with long-term anti-retroviral therapy [22], whether such similarities are seen in HBV requires further investigation, along with the modulation of these cells in novel agents. Finally ADCC has importance for the design of NK cell therapy/vaccine type trials to promote HBs antibody responses and thus the further study of memory NK cells in HBV is urgently required.

Figure 1:

Diagram depicting the differences in FcεRIγ+ and FcεRIγ- CD56^dim mature NK cells in HBV. The latter are depicted as memory-NK cells with a diverse transcription profile along with epigenetic and metabolic alterations compared to conventional CD56^dim NK cells, following exposure to HCMV. In addition memory NK cells undergo increased cytokine responses and CD16-mediated ADCC.