Keywords

18.1 Introduction

NK cells are large granular lymphocytes that are a component of the innate immune system. NK cells kill target cells spontaneously without the requirement of priming or the expression of major histocompatibility antigens (MHC) on the target cell. NK cells lack the receptor diversity generated by DNA rearrangement found in prototype members of the adaptive immune response, T and B cells. However, similar to T cells, NK cells are able to distinguish healthy cells from virally infected cells. NK cells produce abundant antiviral cytokines, release cytotoxic granules such as granzyme B and trigger the FAS/TRAIL pathway to induce apoptosis of infected cells [35]. In order to mount a response, NK cells have to integrate signals received from many inhibitory and activating receptors to sense the environment and respond. Regulation of NK cell activity is achieved by a balance between these inhibitory and activating signals although signals received by inhibitory receptors tend to be dominant [21]. Many of the initial insights on NK cell function came from studying chronic viral infections in mice such as LCMV and MCMV [7]. The clearest demonstration that NK cells are important to humans stem from the findings that patients with NK cell deficiencies develop severe, recurrent viral infections [6, 54]. Our knowledge of select receptors on human NK cells modulated by viral infection and viral peptides that can modulate NK cell function has significantly advanced in the last decade in the context of human viral infections such as HMCV, HIV and HCV [7, 19]. Some of the findings seen in these human chronic viral infections are beginning to be observed for viral infections such as Influenza and Dengue thus opening up new areas of research that will enhance our understanding of how these innate effector cells can control acute viral infections [1, 4, 57, 58, 72]

18.2 Early and Current Observations on NK Cell Responses to Dengue Infections Using Samples from Clinical Cohorts

Human NK cells were shown to lyse dengue virus (DENV)-infected cells to a greater degree compared to uninfected cells over 30 years ago [38]. Antibodies to DENV further augmented the lysis of DENV-infected cells by NK cells. The traditional phenotype used to define human NK cells is based on the absence of CD3 (to exclude T cells) and relative expression of CD56 (neural cell adhesion molecule) and CD16 (low-affinity receptor for the Fc portion of immunoglobulin G). It is now appreciated that CD56 “bright” cells can produce cytokines abundantly but are poorly cytolytic while CD56 “dim” cells (which represent 90% of circulating NK cells) produce moderate cytokines when stimulated but are strongly cytolytic [60]. Using PBMC samples obtained from children in Thailand undergoing severe dengue infection nearly 20 years ago, Green et al. observed increased frequencies of activated NK cells (CD56+ CD69+ cells) in PBMC from patients with severe dengue disease compared to patients with milder disease suggesting that that NK cells are actively involved in the defense against dengue virus [26]. Homchampa et al. found evidence of NK cell cytotoxicity in fresh cells from children with acute dengue and the level of cytotoxicity was related to disease severity [31]. At the time, many of the activating and inhibitory receptors present on NK cells were not identified and the technologies to assess markers on individual NK cells using multiparametric flow cytometry and CyTOF were not available.

In a study published in 2008, frequencies of CD56+ CD69+ NK cells were found to be more activated in hospitalized Vietnamese infants with more severe dengue disease [15]. Recently Petitdemange et al. evaluated NK cell activation in patients from Gabon undergoing acute dengue infection [59]. They examined frequencies of subsets of NK cells and included a number of NK cell receptors NKp30, NKG2A, NKp44 and CD161 which was a significant advance from the early studies where only CD56 expression was used to identify NK cells. Compared to healthy donors, NK cells from patients who had DENV-2 infections expressed significantly less NKp30, NKG2A, and CD161, whereas expression of ILT-2 was increased at the early acute and convalescent time point. Interestingly, frequencies of NKp44+ NK cells were elevated at the early acute time point. Principal component analysis highlighted an association between NKp44 and DENV-2 infections. NKp44 has previously been shown to directly interact with the envelope protein of DENV and WNV and this subset of NK cells and its ligand in target cells is worthy of study in other cohorts [30]. While there have been only a few studies that have looked at NK cell responses in clinical cohorts, the data indicate that subsets of NK cells are involved in the immune responses to natural dengue infection thus warranting further study. NK cell responses in individuals receiving live attenuated tetravalent dengue vaccines have not been examined in any detail. Whether different subsets of NK cells respond to vaccination with live attenuated dengue vaccines versus subunit vaccines and inactivated vaccine antigens are currently unknown. Furthermore, whether the magnitude of NK cell responses is different between vaccination and natural infection has not been explored.

18.3 Activating Receptors on NK Cells

NK cells have a number of receptors that can activate NK cell reactivity (activating receptors) or dampen reactivity (inhibitory receptors). Activating receptors expressed on human NK cells include the C-type lectin receptors NKG2D and NKG2C/E, natural cytotoxicity receptors (NCR) NKp44, NKp30, NKp46, and CD16 (FC-γ-RIII), the low-affinity Fc receptor which mediates antibody dependent cytotoxicity (ADCC) (Fig. 18.1 and Table 18.1). Of the many activating receptors on NK cells and ligands on target cells, there is evidence for interaction of DENV with NKp44, the low-affinity FcR receptor CD16 and ligands for the NKG2D ligand MICB (Fig. 18.1).

Fig. 18.1
figure 1

Model of NK cell interaction with virus-infected cells. NK cells can interact with DENV infected target cells through direct and indirect mechanisms. To date, only NKp44 has been shown to directly interact with the E protein of DENV. NK cells can interact with DENV infected APCs by responding to cytokines, interacting with yet to be identified ligands on virus-infected cells and upregulating inhibitory or activating receptors on its surface. This may lead to degranulation and abundant cytokine secretion by NK cells and lysis of virus-infected cells. Alternatively, dominant interaction of inhibitory receptors on NK cells with viral proteins and peptides could dampen responses and represent an immune evasion strategy by DENV. NK cells can also be directed to DENV infected APCs through engagement of CD16 with the Fc receptor of DENV-specific Abs and target cells are lysed through ADCC

Full size image
Table 18.1 NK cell receptors and ligands
Full size table

In a large genome-wide association study (GWAS) performed in 2011, a single nucleotide polymorphism in MICB (ligand for NKG2D) was found to be a susceptibility locus for severe dengue disease dengue shock syndrome in over 2000 Vietnamese children [37]. In a subsequent study by the same group, MICB was also associated with the less severe form of dengue DF in both children and adults in Vietnam [77]. This genetic susceptibility of a SNP in MICB was confirmed in another study in the Thai population. In Cuban individuals (Guzman), a strong association of DF and DHF with MICA 008 and MICB 008 compared to asymptomatic individuals [24]. Interestingly soluble MICB levels were increased in the sera between pre-infection and acute illness among infants with symptomatic primary DENV infections in the Philippines and suggest an immune evasion strategy in dengue viral infection [43]. The data suggest that ligands for the activating receptor NKG2D, MICB may be involved in the immune response against DENV.

A direct interaction has been shown between the NK-activating receptor NKp44 and the DENV and WNV envelope protein [30]. Recombinant NKp44 directly bound purified DENV and WNV envelope proteins. The results suggest that NKp44 expressing NK cells are able to interact directly with the E protein on flavivirus-infected cells.

NK cells express the FcγRIIIA receptor (CD16) that binds the constant (Fc) domain of IgG antibodies. CD16 engagement is a strong activator of NK cell function and allows antigen-specific recruitment of NK responses. Very early on, NK cell-mediated lysis of DENV-2-infected Raji cells was significantly higher when these cells were pre-incubated with DENV immune serum through antibody dependent cell-mediated cytotoxicity (ADCC) [38]. More recently efforts have been made to understand in a prospective cohort whether ADCC activity in sera obtained pre-illness in a prospective cohort correlated with subsequent dengue disease severity and viremia levels. A higher level of ADCC activity measured in immune sera before secondary DENV-3 but not secondary DENV-2 infection was associated with lower subsequent viremia levels [41]. These results suggest a protective role for ADCC antibodies in sera obtained from in some children with secondary infection. In vivo, characteristics of both the Abs and NK cells in each individual are likely to impact the function of these Abs. Future studies should compare the contribution of functionally distinct Abs (those that mediate neutralization versus Abs that mediate ADCC) in the sera of individuals who subsequently develop clinical versus subclinical disease.

In order to understand soluble and membrane found factors important for activating NK cells, Lim et al. used DENV-infected monocyte-derived dendritic cells in co-culture with autologous NK cells from healthy donors [44]. The results suggest that a combination of cytokines type I IFNs, TNF-☐ secreted by DCs and cell surface receptor–ligand interactions all play some part triggering the anti-dengue response of primary human NK cells while each of these in isolation has a low to minimal response on the activation of human NK cells.

18.4 Inhibitory Receptors on NK Cells

18.4.1 KIRs

A major set of inhibitory receptors important for NK cell licensing is the killer cell immunoglobulin-like receptors (KIRs). The KIR receptors are expressed stochastically, leading to the generation of subpopulations of NK cells within each individual [50]. While only some KIR ligands are known, all are thought to be class I MHC molecules (Table 18.1). Epidemiological studies of immunological events, including responses to infectious diseases, autoimmune diseases, complications with pregnancy, and tumor responses, have implicated KIR/MHC interactions in protective or pathological roles [5, 34, 32, 39, 53, 63, 68].

In contrast to the T-cell receptor, which recognizes a given viral peptide-MHC complex, KIRs typically recognize subsets of MHC class I molecules with common amino acid motifs in their α1 domains ([13, 33]). However, several studies have shown that inhibitory KIRs are also sensitive to the specific peptides bound by the HLA molecules [12]. In particular, residues 7 and 8 of endogenously derived peptides bound to MHC class I can promote or abrogate binding of KIR. During viral infection, the peptide repertoire changes and a number of viral peptides are presented on MHC molecules for recognition by cytotoxic T lymphocytes. Changes in the endogenous repertoire of peptides can perturb or continue to maintain inhibition of NK cells (Fig. 18.2) depending on how strongly the viral peptide is able to bind and replace the endogenous peptide. Several viral peptides (HIV, HCV, SIV) have been shown to modulate KIR-MHC interactions thus providing a mechanism for the virus to modulate the interaction of the inhibitory KIR with its MHC ligand thus preventing the activation of subsets of NK cells [22, 47]. How changes in the peptide repertoire following viral infection alter NK cell recognition in vivo will depend on the relative amounts of viral peptides being presented, the nature of the peptides produced by infection and the frequency of NK cells expressing the relevant KIR.

Fig. 18.2
figure 2

A model of NK-cell inhibition during DENV infection. Specific amino acid changes in the peptides presented by major histocompatibility complex (MHC) class I can impact KIR recognition of peptide/MHC complexes. We hypothesize that the DENV NS1 peptide generated during infection presented by HLA B57 alters the affinity of the inhibitory KIR3DL1 receptor expressed on NK cells for its putative ligand, resulting in the inhibition of NK cells early during viral infection while MHC class I is upregulated. Following the clearance of viremia, HLA-B57 levels return to normal which decreases the inhibitory signal received by KIR3DL1+ NK cells. This release of inhibition allows KIR3DL1+ NK cells to be activated at fever day 0, coincident with the activation of CD8+ T cells

Full size image

18.5 KIR Peptide-MHC Interactions During Dengue Infection

During a study of HLA B57 restricted T cells in acute dengue infection, Townsley et al. found that a T cell tetramer for a CD8+ T cell epitope in the NS1 protein of DENV bound CD56dim NK cells [72]. Using depletion studies and KIR-transfected cell lines, the HLA B57 restricted dengue NS1 tetramer was found to bind the inhibitory receptor KIR3DL1 on NK cells clearly demonstrating an interaction with an inhibitory KIR. Using clinical samples from a long-standing cohort in Thailand the authors further found late activation of NK cells in HLA B57+ individuals. The peak of NK cell activation occurred at the critical time point in people with severe dengue disease which supported the hypothesis that the HLA B57 restricted NS1 peptide can modulate the function of KIR3DL1+ NK cells. The data strongly suggest that subpopulations of NK cells may contribute to disease pathogenesis and enabled the authors to speculate how the NS1 peptide may modulate function during acute dengue infection. At steady state (Fig. 18.1a), the interaction of an HLA B57 presenting endogenous peptides on an antigen presenting cell (APC) with KIR3DL1 maintains an inhibitory signal in NK cells thus preventing lysis of healthy target APCs. During many viral infections, HLA molecules are downregulated which allows NK cells to get activated and lyse virus-infected APCs (Fig. 18.1b). However it is well known that in DENV infection HLA class I molecules are upregulated [79]. We hypothesize that presentation of the NS1 peptide on HLA B57 during the viremic phase in infected individuals maintains the inhibitory interaction with KIR3DL1 (since the tetramer has been shown to bind KIR3DL1) thus delaying the activation of this important subset of NK cells (Fig. 18.1c, d). As viremia decreases, Class I expression decreases and NK cells are able to sense the change in Class I expression, triggering their activation during the critical phase of illness (around fever day 0) (Fig. 18.1d). Our clinical data support our hypothesis but we are yet to show functional alteration of NK cell responses by the NS1 peptide.

We are currently using NK cell sensitive TAP-impaired and TAP-blocked target cells (721.221) transfected with HLA alleles in functional assays. Since self-peptide is able to maintain inhibition, the use of TAP blocked and TAP impaired cell lines HLA transfected cell lines are essential to evaluate whether viral peptides added exogenously such as the DENV NS1 peptide can modulate the function of KIR3DL1+ NK cells. We anticipate that additional KIR/HLA interactions occur which are likely to be modulated by viral peptides during acute dengue.

Petitdemange et al. performed the genotypic analysis of KIR in the combination with their cognate HLA-class I ligands in 73 CHIKV and 55 DENV2 adult cases, compared with 54 healthy individuals [57]. Petitdemange et al. found no evidence of a role for KIR genotypes in patients infected with DENV-2. In contrast, Beltrame et al. detected an association between certain KIR genes and their cognate HLA ligands in the context of infection with DENV-3 in Southern Brazil [3, 57]. Differences in population origin and the infecting DENV serotype may explain these disparate results. Petitdemange et al. in a subsequent paper assessed NK cell activation by multiparametric flow cytometry in patients with DENV-2, CHIKV or in a few subjects with co-infections [59]. They found different signatures of NK cell responses between the two infections. NK cells in CHIKV infection were activated early and expressed a terminal differentiation pattern with prolonged persistence of NKG2C + CD57+ cells which the authors speculate may contribute to the chronic arthralgia seen in CHIKV infection. DENV-2 infections were associated with an increase in KIR2DL1+ NK cells which recognize HLA-C2. Together, these studies suggest that KIR-MHC interactions are likely to be important during acute dengue infection.

18.6 LILRB

LILRB is an inhibitory receptor present on monocytes, dendritic cells, and NK cells. It interacts with a wide range of MHC Class I molecules and maintains a negative feedback loop to prevent autoimmunity [70]. THE UL-18 protein of HCMV has been shown to bind with higher affinity to LILRB and protect against NK cell recognition in the context of HCMV infection. Recently using dengue-specific antibodies at neutralizing and sub-neutralizing concentrations, mechanisms of antibody-dependent enhancement (ADE) were further characterized in a resistant (THP-1R) and susceptible (THP-1S) subclone of THP-1 cells [14]. In this context, DENV was shown to interact with the inhibitory receptor LILRB. This inhibitory interaction blocked FcγRII signaling and dampened the expression of IFN stimulated genes and enhances DENV replication.

HLA-E is a non-classical MHC molecule that interacts with both activating (CD94C, D, and E) and inhibitory receptors (CD94A). Interaction of HLA-E with the inhibitory receptor NKG2A is of higher affinity compared to the interaction with most known HLA-E/peptide complexes than those transmitting activating signals. A related flavivirus, Japanese Encephalitis virus has been shown to upregulate HLA-E but no work has been published yet for DENV [69].

18.7 The Role of NK Cells in Modulating Adaptive Immune Responses

NK cells have also been implicated in shaping the adaptive response to viral infections in a number of ways including promoting maturation or elimination of DCs, perforin-dependent elimination of CD8+ T cells, and cytokine production [64]. Waggoner et al. used the model of lymphocytic choriomeningitis virus (LCMV), to show that NK cells can regulate CD4 T-cell-mediated support for the antiviral CD8 T cells [74, 76]. NK cells have also been shown to be important for long-term CD4+ T cell memory and subsequent antibody responses [18]. The data suggest that NK cells continue to participate in immune modulation well after initial infection when NK cells are traditionally thought to be active. In a previous study of CD8+ T cells by Townsley et al. frequencies of the HLA-B57-restricted epitope, were assessed over the course of acute DENV infection [71]. Given the highly conserved nature of this epitope, we predicted that PBMC from donors with secondary dengue infection would have significantly higher frequencies of B57-NS126–34 CD8+ T cells compared to PBMC from donors with primary dengue infection. While we detected tetramer-positive T cells in all subjects tested, the frequencies in subjects with secondary infections were not higher than in subjects with primary infections, with one exception. We speculated that an unidentified factor may dampen activation of CD8+ T cells directed at this epitope but had not yet identified that the NS1 peptide presented on HLA-B57 could bind KIR3DL1 an inhibitory receptor on NK cells. Our new findings suggest that NK cells could shape CD8+ T cell responses but given the lack of an authentic animal model that mimics human dengue infection it will be challenging to provide definite proof that inhibitory NK cells can modulate adaptive responses at the epitope level in vivo.

The varying combinations of inhibitory and activating receptors on NK cells and the number of unknown ligands make it difficult to assess changes in absolute frequencies of subsets of NK cells between subjects with mild or severe dengue illness. Furthermore, for meaningful comparisons to be made samples must be collected at multiple points and compared in subjects with mild and severe dengue disease rather than compare responses in all subjects with dengue disease and responses in healthy subjects. Since the hallmark of dengue hemorrhagic fever DHF (severe disease) is plasma leakage, if NK cells are hypothesized to contribute to DHF, then changes in the frequency or function of subsets of NK cells during or prior to defervescence the critical phase of illness must be demonstrated. In vitro assays will need to be performed in order to understand which receptors and ligands interact with virus-infected cells and are important for cytokine production and cytolytic activity in subsets of NK cells following DENV infection as little information exists to date. Basic biological studies of NK cells and multiple studies using clinical samples in chronic viral infections suggest that subsets of NK cells are an important contributor to the overall immune response. We have a partial understanding of how NK cells can regulate dengue immune responses but these findings point to multiple avenues whereby NK cells may control DENV infection and shape adaptive immune responses.

18.8 T-Cell Responses to Dengue Vaccination

T cell immune responses to natural dengue infection will be reviewed by Rivino et al. in Chap. 15. With the advent of a number of dengue tetravalent vaccine trials, T cell responses are being evaluated in vaccinated individuals [66]. These studies indicate that CD4 and CD8 T cells are directed against a number of epitopes across the entire DENV genome. In support of smaller studies performed in individuals receiving monovalent DENV vaccines, the data indicate that non-structural proteins are dominant targets of T cell responses following tetravalent vaccination as well [51]. There are significant differences in the composition of dengue antigens included in each vaccine which are likely to impact the immune response elicited [66]. Prior flavivirus immunity in many volunteers who received dengue vaccines in endemic countries will also influence the magnitude and quality of T cell immune responses. Here, published work to characterize CD4+ and CD8+ T-cell responses to dengue virus antigens and peptides will be reviewed.

The Sanofi Pasteur vaccine is a chimeric Yellow fever virus that expresses the structural components (prM and E) of each of the four serotypes of dengue [65, 66]. In three independent phase 1 clinical trials conducted in Australia and in the USA, Dengue 1–4 and Yellow Fever 17D (YF 17D)-specific CD4 and CD8 cellular responses induced by tetravalent chimeric dengue vaccines (CYD) were analyzed in flavivirus-naive or immune individuals [27]. Significant YF 17D NS3- specific CD8 responses and dengue serotype-specific T helper responses were detected. These responses were dominated by serotype 4 in naive individuals, but a booster vaccination broadened serotype-specific responses. A similar, broader response was seen after primary tetravalent immunization in subjects with pre-existing dengue 1 or 2 immunity caused by prior monovalent live- attenuated dengue vaccination. There was an absence of a Th2 response, and a more dominant response to IFN-☐ compared to TNF-α, for both CD4 and CD8 responses.

CD4 and CD8 T cell responses were assessed before and 28 days after a first and third injection of Sanofi Pasteur live attenuated CYD-TDV and 1 year after the third injection in 80 adolescents and adults enrolled in a phase II trial in Singapore [29]. CD4/IFN-γ and TNF-α responses specific to dengue NS3 were detected before vaccination. Vaccination induced YF-17D-NS3-specific CD8 IFN-γ responses. Using a multiplex analysis of peripheral blood mononuclear cells (PBMC) supernatants after restimulation with each the CYD vaccine viruses, higher levels of IFN-γ were detected compared with TNF-α, and low-level IL-13 levels.

The Takeda vaccine is a live-attenuated tetravalent dengue vaccine (TDV) candidate that consists of an attenuated DENV-2 strain (TDV-2), and 3 chimeric viruses containing the premembrane (prM) and envelope (E) genes of DENV-1, -3, and -4 expressed in the context of the TDV-2 genome (TDV-1, TDV-3, TDV-4, respectively). CD8+ T-cell responses in flavivirus-naive human volunteers vaccinated with 2 doses of the Takeda vaccine 90 days apart via two routes (subcutaneous or intradermal) were assessed using overlapping peptide pools against the NSl, NS3 and NS5 proteins [16]. The TDV vaccine-elicited CD8+ T cell TNF-☐ and IFN-☐ responses against the nonstructural NS l, NS3, and NS5 proteins of DENV-2. Responses were highest on day 90 after the first dose and were still detectable on 180 days after the second dose. In addition, CD8+ T cells were multifunctional, producing 2 cytokines simultaneously, and cross-reactive to NS proteins of the other 3 DENV serotypes.

For the NIH vaccine, attenuation was achieved by deleting a portion of the 31 untranslated region (UTR) of the dengue genome. Longitudinal specimens were collected in flavivirus-naive volunteers and tested for T cell responses against inactivated DENV antigen preparations made in Vero cells following low dose vaccination with a live attenuated DENV-1 candidate vaccine (DEN1L130) from the NIH. CD4+ T cells were found to secrete IFN-γ, TNF-α and IL-2, 3 weeks following exposure to DENV-1 and were detected upto 6 weeks. T cells produced 2 cytokines simultaneously [45]. There was little cross-reactivity in T cell responses.

Weiskopf et al analyzed CD8+ T cell responses in PBMC from healthy naïve volunteers in Vermont who received either monovalent or tetravalent live attenuated DENV vaccine from the NIH [75]. Using peptide pools (9- and 10-mer peptides that were predicted for their binding affinity to 27 major histocompatibility complex (MHC) class I molecules to monovalent and tetravalent DENV vaccines from the NIH. They found broad responses (IFN-☐ and TNF-☐ responses) to both structural and non-structural proteins were elicited in response to a monovalent vaccine (up to 47 months post-vaccination) while responses to a tetravalent vaccine (up to 12 months post-vaccination) were directed against the non-structural proteins.

Immune responses in vaccinated individuals are continuing to be monitored. How durable T cell responses to any of the attenuated vaccines in clinical trials is of interest as waning immunity to dengue is proposed to increase the risk of severe illness.

Questions

  • Are distinct subsets of NK cells elevated in natural DENV infection and do they differ between mild and severe disease, primary or secondary dengue infection?

  • Can viral peptides modulate NK cell function by interacting with inhibitory and activating KIRs?

  • Can distinct subsets of NK cells be induced following vaccination and how do they differ from natural infection?

  • Do preexisting dengue antibodies modulate NK cell responses?

  • Are there new ligands on target cells that are important for flavivirus infections?