CD4⁺ and CD8⁺ T‐cell immunity to Dengue – lessons for the study of Zika virus

Summary

Dengue virus (DENV) and Zika virus (ZIKV) are rapidly emerging mosquito‐borne flaviviruses that represent a public health concern. Understanding host protective immunity to these viruses is critical for the design of optimal vaccines. Over a decade of research has highlighted a significant contribution of the T‐cell response to both protection and/or disease enhancement during DENV infection, the latter being mainly associated with sub‐optimal cross‐reactive T‐cell responses during secondary infections. Phase IIb/III clinical trials of the first licensed tetravalent dengue vaccine highlight increased vaccine efficacy in dengue‐immune as opposed to dengue‐naive vaccinees, suggesting a possible immunoprotective role of pre‐existing DENV‐specific T cells that are boosted upon vaccination. No vaccine is available for ZIKV and little is known about the T‐cell response to this virus. ZIKV and DENV are closely related viruses with a sequence identity ranging from 44% and 56% for the structural proteins capsid and envelope to 68% for the more conserved non‐structural proteins NS3/NS5, which represent the main targets of the CD4⁺ and CD8⁺ T‐cell response to DENV, respectively. In this review we discuss our current knowledge of T‐cell immunity to DENV and what it can teach us for the study of ZIKV. The extent of T‐cell cross‐reactivity towards ZIKV of pre‐existing DENV‐specific memory T cells and its potential impact on protective immunity and/or immunopathology will also be discussed.

DENV and ZIKV: similarities and differences

Dengue virus (DENV) and Zika virus (ZIKV) belong to the Flavivirus genus of the Flaviviridae family of viruses along with other arthropod‐borne viruses that may have significant impact on human health such as Yellow fever virus (YFV), West Nile virus (WNV), Japanese encephalitis virus (JEV) and tick‐borne encephalitis virus (TBEV). No specific antiviral therapeutic is available for these viruses and treatments are supportive in nature. Protective vaccines are available for JEV, TBEV and YFV and a partially protective vaccine has recently been licensed for DENV.1 The live‐attenuated YFV vaccine, which is safe and extremely effective, was shown to elicit long‐lived neutralizing antibodies and a strong CD4⁺ and CD8⁺ T‐cell response,2, 3 components that we believe are key to a successful vaccine. However, the co‐circulation of DENV as four distinct serotypes (DENV 1–4) and the risk of immunopathology associated with sub‐optimal cross‐reactive B‐cell and T‐cell responses to heterologous serotypes represent critical factors for the development of a fully protective DENV vaccine.

Dengue virus, ZIKV and the other flaviviruses are enveloped viruses with a 10·7‐kb positive‐strand RNA genome encoding for a single polyprotein that is post‐translationally cleaved into three structural proteins (capsid, membrane, envelope) and seven non‐structural proteins (NS1, NS2a, NS2b, NS3, NS4a, NS4b, NS5). DENV 1–4 serotypes share approximately 70% amino acid identity whereas ZIKV displays an overall 43% homology with DENV (with up to 68% identity for the more conserved non‐structural proteins). Both DENV and ZIKV are principally transmitted by the bite of an infected Aedes aegypti/aldopictus mosquito but other minor routes of infection have been reported for ZIKV (sexual transmission, maternal transmission and through blood transfusions).4 Infection with DENV may be asymptomatic or it may cause a febrile illness (dengue fever) which is accompanied by severe headache, retro‐orbital pain, myalgia, arthralgia, gastrointestinal complications, liver inflammation and skin rashes. As the fever subsides, patients may develop more severe life‐threatening disease characterized by an increase in vascular permeability, plasma leakage and haemorrhagic manifestations, which may lead to hypovolaemic shock (dengue haemorrhagic fever and dengue shock syndrome, respectively). The factors responsible for the development of severe disease remain poorly defined and are largely associated with pre‐existing host immunity during secondary heterologous infections (cross‐reactive B‐cell and T‐cell responses).5, 6 The clinical features of ZIKV infection resemble – but are generally milder than – those caused by DENV and range from asymptomatic infection to a febrile illness characterized by headache, arthralgia, myalgia, maculopapular rash, conjunctivitis, vomiting and fatigue. However, severe neurological complications of ZIKV infection such as Guillain–Barré syndrome (GBS) in adults and congenital birth defects including macrocephaly in the developing fetus have emerged from recent epidemics, making ZIKV an emerging public health emergency. Clinical symptoms associated with ZIKV infection thus share common features with those developed upon infection with the mosquito‐borne encephalitic viruses (such as WNV and JEV) and with the viruses from the DENV group. Interestingly, phylogenetic analyses based on the amino acid sequences of the non‐structural protein NS5 result in the clustering of ZIKV with the encephalitic viruses, whereas analyses based on the amino acid sequence of the E protein cluster ZIKV with the DENV group, suggesting that ZIKV may have emerged as a recombinant virus between DENV and the encephalitic viruses.7 DENV was first isolated in 1943, has rapidly spread since the 1980s and is now endemic in over 100 tropical and sub‐tropical countries with a significant burden of disease in South‐East Asia, the Indian subcontinent and some areas of Latin America.8 ZIKV was first identified in 1947 in the Zika forest in Uganda and was later isolated in other African countries and in South‐East Asia.9, 10 Seroprevalence for ZIKV is high in populations throughout Africa and Asia but the exact areas of ZIKV exposure remain difficult to define because assays used in these studies also cross‐react with other endemic flaviviruses. The spread of ZIKV from Africa to the West Pacific and Americas became apparent during the ZIKV outbreaks on the Island of Yap in Micronesia in 2007, in French Polynesia in 2013/14 and in Brazil in 2015. The abundance of the Aedes aegypti vector and the availability of susceptible hosts have allowed ZIKV to spread rapidly from Brazil throughout Latin America and the Caribbean and further spread to other areas can be anticipated.

In summary, although DENV and ZIKV are closely related viruses that share structural and genetic characteristics, early clinical disease manifestations and the Aedes mosquito vector, ZIKV infection presents a unique profile of pathogenesis that is probably due to the distinct cellular tropism of ZIKV. In this review we discuss the main characteristics of the T‐cell response to DENV and how – based on the similarities and differences between DENV and ZIKV – this knowledge could provide us with insights into T‐cell immunity to ZIKV.

The ‘long and winding road’ of DENV T‐cell epitope discovery

A detailed analysis of the virus‐specific T‐cell response requires the previous knowledge of the immunogenic peptides of 9–10 or 12–15 amino acids in length that are recognized by CD8⁺ or CD4⁺ T cells in association with MHC class I or II molecules, respectively. In the last decade a large number of T‐cell epitopes has been identified for DENV, the majority of which are CD8⁺ T‐cell epitopes (1613 immunogenic peptides are currently listed in the Immune Epitope Database http://www.iedb.org/). T‐cell epitopes have been identified using a variety of methods ranging from traditional epitope mapping strategies with 15‐mer peptide libraries overlapping by 10 amino acids spanning the entire viral proteome11 or selected proteins,12 to more modern approaches based on predictive algorithms that can determine immunogenic epitopes associated with chosen HLA molecules.13 Traditional methods are laborious but provide a comprehensive and unbiased identification of both CD4⁺ and CD8⁺ T‐cell epitopes, whereas predicative algorithms represent a high throughput technology but may not be accurate in terms of identifying immunodominant epitopes, particularly for less characterized HLA molecules expressed by Asian populations.14 Collectively, these studies show that dengue infection elicits a broad dengue‐specific T‐cell response that peaks around days 8–10 from fever onset11, 12, 15 and targets all viral proteins with a preferential recognition of the non‐structural proteins NS3, NS4b and NS5.11, 12, 15 Dengue‐specific CD8⁺ T cells are present at higher frequencies compared with their CD4⁺ counterparts and preferentially target the non‐structural proteins NS3, NS4b and NS5, whereas CD4⁺ T cells are mainly directed towards the structural proteins capsid and envelope and the secreted protein NS1.11 A similar skewing of the CD4⁺ and CD8⁺ T‐cell response towards recognition of the structural or non‐structural proteins, respectively, was recently reported for JEV.16 Vaccination with the YFV 17D vaccine strain elicits a CD8⁺ T‐cell response that mainly recognizes NS3 and NS5 proteins, followed by the envelope protein,17 whereas the breadth of the CD8⁺ T‐cell response to WNV appears to be broader and to preferentially target capsid, membrane, envelope, NS3 and NS4b.18 Of note, a recent study showed that DENV 3 infections elicit a distinct pattern of immunodominance compared with the other DENV serotypes with CD8⁺ T cells targeting both structural and non‐structural DENV proteins and suggests that immune recognition may vary as a function of the infecting DENV serotype and between different flaviviruses.19

The identification of a large number of immunogenic epitopes with broad coverage of HLA types expressed worldwide provides us with the necessary tools for the characterization of virus‐specific T cells during natural dengue infection or after vaccination. However, choosing a T‐cell epitope restricted to an HLA molecule of interest (for example for peptide–HLA tetramer studies) can be a challenging task as for the majority of these epitopes there is scarce information on the frequency and magnitude of the T‐cell responses that they elicit. Studies using peptide–HLA tetramers are highly informative because they allow the direct ex vivo quantification and phenotypical/functional characterization of virus‐specific T cells. So far, for DENV these studies have focused mainly on CD8⁺ T cells specific for the HLA A*1101‐restricted NS3_133–142 epitope, which represents the best‐characterized T‐cell epitope for DENV.6, 15, 20, 21, 22 Other studies have characterized CD8⁺ T cells specific for epitopes localized in the NS1, NS5 or envelope protein (Table 1).22, 23, 24, 25 A comparison of the amino acid sequences of these DENV CD8⁺ T‐cell epitopes with the corresponding epitopes from ZIKV highlights the high sequence homology that exists between the two viruses and suggests that some of these CD8⁺ epitopes may also be applicable to ZIKV (Table 1). To our knowledge there are no published studies on dengue‐specific CD4⁺ T cells where peptide‐HLA tetramers were used.

Table 1.

CD8⁺ T‐cell epitopes from denque virus (DENV) that have been used for peptide‐HLA tetramer studies and corresponding peptide variants from other DENV serotypes or from Zika virus (ZIKV)

Epitope	HLA restriction	Sequence of peptides used for peptide‐HLA tetramer	Peptide variants from other DENV serotypes/ZKV	References
Env213‐221	A*02:01	DENV 1: FLDLPLPWT DENV 2: FLDLPLPWL DENV 3: FFDLPLPWT DENV 4: FFDLPLPWL	ZKV: FHDIPLPWH	Townsley et al. Immunology. 201423
NS126‐34	B*57:01	DENV 1‐4: HTWTEQYKF	ZKV: EAWRDRYKY	Townsley et al. Immunology. 201423
NS3133‐142	A*11:01	DENV 1: GTSGSPIVNR DENV 2: GTSGSPIVDR, GTSGSPIIDK DENV 3‐4: GTSGSPIINR	ZKV: GTSGSPILDK	Mongkolsapaya et al. Nat Med. 20036
				Duangchinda et al. PNAS 201015
				Friberg et al. Sci Rep. 201122
				Friberg et al. Immunol Cell Biol 201048
				Townsley et al. Immunology. 201423
				Rivino et al. Sci. Transl Med 201520
NS3222‐231	B*0702	DENV 2‐4: APTRVVAAEM	DENV 1: APTRVVASEM ZKV: APTRVVAAEM	Friberg et al. Sci Rep. 201122
NS3556‐564	A*24 (A*11)	DENV 1: QYSDRRWCF DENV 2: NYADRRWCF DENV 3: KYTDRKWCF DENV 4: SYKDREWCF	ZKV: TYTDRRWCF	Mongkolsapaya et al. J. Immunol. 200624
NS3542‐550	B*55:02	DENV 2: LPVWLAYRV	DENV 1: LPVWLSYKV DENV 3: LPVWLAHKV DENV 4: LPVWLSYKV ZKV: LPVWLAYQV	Chang et al. Eur. J. Immunol. 201325
NS5329‐337	B*55:02	DENV 2: KPWDIIPMV	DENV 1: KPWDVIPMV DENV 3: KPWDVVPMV DENV 4: KPWDVIPMV ZKV: KPWDVVTGV	Chang et al. Eur. J. Immunol. 201325

An alternative method for the detection of DENV‐specific T cells that does not require prior knowledge of the patient HLA types is to perform functional assays by stimulating with dengue peptide pools of overlapping peptides20 or pre‐selected ‘mega‐pools’ of predicted peptides.26 These assays allow us to visualize DENV‐specific T cells specific for a wide range of immunogenic peptides but they have the disadvantage of relying on a functional read‐out (e.g. production of interferon‐γ or tumour necrosis factor‐α) and hence fail to identify cells with poor functional capacity, which are in contrast readily detected with peptide‐HLA tetramers.6 It is important to be aware of the limitations of each method and ideally to consider the use of a combination of different methods.

In summary, a limited number of the defined CD8⁺ T‐cell epitopes from DENV have been used to directly characterize ex vivo dengue‐specific T cells during the course of infection. Some of these epitopes are highly conserved in ZIKV and may facilitate the study of T‐cell immunity to this virus. The lack of similar peptide–HLA studies for dengue‐specific CD4⁺ T cells highlights a substantial knowledge gap in the field.

Viral tissue tropism and impact on T‐cell immunity

There is robust evidence suggesting that upon entry of DENV into the host through the bite of an infected Aedes mosquito, the skin represents the initial site of virus replication and of host immune surveillance. Epidermal Langerhans cells and dermal dendritic cells of the skin were shown to support dengue replication ex vivo in humans and in vivo in immunocompromised mice.27 Furthermore, DENV could be detected systemically in the skin of non‐human primates following a primary or a secondary DENV infection,28 in the skin of a deceased dengue patient29 and within areas of skin rash in an individual receiving a live attenuated experimental dengue vaccine.30

Studies have shown that during T‐cell receptor activation T cells are imprinted with the expression of chemokine receptors and/or adhesion molecules that represent the ‘address‐code’ of that particular tissue, such that the activated T cells have the predisposition to home back to the site of their initial antigenic encounter.31, 32 For example, T‐cell priming by skin‐derived dendritic cells induces the expression of the cutaneous lymphocyte‐associated antigen (CLA), which confers T cells with the ability to migrate back to the skin. Accordingly, T cells present in peripheral blood that are specific for antigens acquired through the skin or the gut were shown to express skin‐homing or gut‐homing receptors, respectively.31 Dengue‐specific CD4⁺ and CD8⁺ T cells present in the blood of patients with acute dengue express high levels of the skin homing receptor CLA, consistent with the initial activation of these cells by skin‐derived dendritic cells in the skin‐draining lymph‐nodes. Expression of CLA correlates with the ability of dengue‐specific T cells to home to the skin as these cells could be detected at high frequencies in the skin of patients with acute dengue.20 Dengue‐specific CD8⁺ T cells were also shown to express inflammatory chemokine receptors that support T‐cell migration to peripheral inflamed tissues (CXCR3 and CCR5) and the putative liver‐homing receptor CXCR6, but not the gut‐homing receptor CCR9.20 In dengue‐immune individuals dengue‐specific CD4⁺ T cells with cytolytic activity were shown to express CX3CR1,33 a chemokine receptor that was previously identified on CD4⁺ T cells residing in the skin and lungs.34, 35

Based on these findings, we hypothesize that upon entry of DENV through the skin the virus infects local antigen‐presenting cells, which migrate to skin‐draining lymph‐nodes where they transport infectious viral particles and prime virus‐specific T cells. DENV‐specific T cells acquire expression of CLA and preferentially travel back to the skin (Fig. 1a). From the skin‐draining lymph nodes virus‐infected cells spread systemically to other lymph nodes (where they can activate more virus‐specific T cells) and through the bloodstream where the virus gains access to other tissues capable of sustaining its replication. Primed T cells expressing tissue‐specific homing receptors may also gain access to the corresponding tissues where they mediate viral clearance by targeting virus‐infected cells but where they may also contribute to tissue damage. DENV was shown to infect monocytes and macrophages and could be detected in lymph‐nodes28 and in the spleen but also in hepatoctyes and Kupffer cells (specialized macrophages) in the liver29 (Fig. 1a). In rare cases of dengue infection accompanied by neurological complications DENV could be detected in the brain.36 Interestingly, immunocompromised mouse models of DENV infection showed that clearance of DENV from the central nervous system was dependent on interferon‐γ production by CD8⁺ T cells and prevented a DENV‐induced paralysis of mice.37

Figure 1.

Schematic representation of T‐cell priming during dengue virus (DENV) and Zika virus (ZIKV) infection and acquisition of tissue‐homing receptors by virus‐specific T cells. (a) DENV enters the host through the bite of an infected Aedes mosquito, infects local antigen‐presenting cells, which then migrate to skin‐draining lymph nodes (LN) and prime virus‐specific T cells. Dengue‐infected, skin‐derived antigen‐presenting cells imprint expression of the skin‐homing receptor cutaneous lymphocyte‐associated antigen (CLA) on the activated T cells, such that CLA+ T cells preferentially return to the skin tissue (through the blood). During a secondary DENV infection virus‐specific T cells that may already be present in the skin can mount an immediate effector response. From the skin‐draining lymph nodes virus‐infected cells spread systemically to other lymph nodes, where they can activate more virus‐specific T cells, and through the bloodstream where the virus gains access to other tissues capable of sustaining its replication. Primed T cells expressing tissue‐specific homing receptors may also gain access to the corresponding infected tissue where they mediate viral clearance by targeting virus‐infected cells but may also contribute to tissue damage. For example, DENV was shown to infect cells in the liver and DENV‐specific T cells expressing the putative liver‐homing chemokine receptor CXCR6 have been identified in the blood of patients with acute dengue. (b) We speculate that during a ZIKV infection virus‐specific T cells are similarly primed in skin‐draining lymph nodes, up‐regulate CLA and subsequently migrate back to the skin. ZIKV spreads systemically through the bloodstream and may infect cells in other tissues, for example the virus displays tropism for neuronal cells of adults or neuronal progenitor cells in the brain of a developing fetus.

Recent studies have shown that ZIKV can infect and actively replicate in human Langerhans cells, and also in epidermal keratinocytes and skin fibroblasts,38 suggesting that the initial priming of ZIKV‐specific T cells may occur in a similar setting to that of DENV. However, subsequent tissue tropism of ZIKV appears, at least in some circumstances, to differ from that of DENV (Fig. 1b). ZIKV was detected in the brain of human fetuses and of infants born to ZIKV‐infected mothers, where it was associated with neuronal damage.39 In utero transmission of ZIKV was demonstrated in two different mouse models and resulted in ZIKV infection of neuronal progenitor cells and congenital abnormalities.40 The ability of ZIKV to infect neuronal cells has also emerged in adults affected by GBS, syndrome which appeared to be linked to a previous ZIKV infection during the epidemic in French Polynesia. GBS is an immune‐mediated neuropathy affecting the peripheral nervous system that has been reported to occur following bacterial and viral infections including DENV,41 cytomegalovirus and Epstein–Barr virus.42 Different forms of GBS feature the involvement of distinct immune cells with T cells being the main drivers of immunopathology in acute demyelinating GBS whereas antibodies play a predominant role in acute motor axonal neuropathy.42 A better understanding of the immune response during ZIKV infection is needed to clarify the possible involvement of T cells or of other immune cells in GBS immunopathology.

The concept of T‐cell imprinting of tissue‐homing properties was further evaluated in experimental vaccination studies whereby the route of immunization was shown to impact on the migratory capacity of the responding T cells and on the ability of antigen‐specific T cells to mediate protection upon re‐infection.43, 44 For example, vaccination in mice with live vaccinia virus was 100 000 times more effective in protecting against a subsequent challenge with vaccinia virus if delivered by skin scarification compared with subcutaneous, intradermal, and intramuscular vaccination.45 The higher efficacy of skin scarification was associated with the infection of local keratinocytes and the generation of a long‐lived population of tissue resident CD8⁺ T‐cell memory populations.44 Studies in mouse models of virus infection have shown that protection upon re‐infection is mediated exclusively by tissue‐resident T cells that are generated in the skin upon viral infection, are unable to exit the tissue and provide immediate protection upon viral re‐encounter.46 Further studies are needed to understand whether tissue‐resident T cells play similar roles and are maintained long‐term during infection in humans.

In conclusion, we highlight T‐cell migration as an important unappreciated aspect of T‐cell immunity that is closely related to the cellular and tissue tropism of the virus. Understanding the expression of tissue‐specific and adhesion molecules by virus‐specific T cells and the signals required for the regulation of these receptors will allow us to mimic or to alter the homing properties of antigen‐specific T cells in the context of vaccination or immunopathology.

Role of T cells during dengue infection and ‘original antigenic sin’

The role of T cells during dengue infection is still controversial with studies supporting either an immunoprotective or immunopathological role (reviewed in detail in ref. 47). Pioneering studies proposed that T cells have a detrimental role during secondary dengue infections in a process termed ‘original antigenic sin’. According to this theory, cross‐reactive T cells generated during the primary infection that recognize the secondary‐infecting DENV serotype with low affinity are poorly functional but prone to inducing immunopathology.6 As cross‐reactive memory T cells are present in increased numbers and have a low activation threshold, they may outcompete their naive counterparts that have high affinities for the secondary‐infecting serotype with an overall detrimental outcome for protective immunity.6 A number of studies report differential cytokine‐producing and functional capacities of DENV‐specific T cells upon stimulation with heterologous peptides compared with the natural peptide. These range from higher tumour necrosis factor‐α and lower interferon‐γ production,48 suboptimal degranulation but high cytokine producing‐capacity24 or similar cytotoxic capacity but impaired interferon‐γ production.49 Another study reports an altered pattern of cytokine secretion for some DENV epitopes but not for others.50 More recent studies confirmed that the preferential activation of cross‐reactive T cells during secondary DENV infection or upon sequential monovalent DENV vaccination resulted in the skewing of the T‐cell response towards conserved epitopes at the expense of serotype‐specific ones, but this process did not lead to a dysfunctional low‐avidity T‐cell response.19, 51 There is now accumulating evidence in both humans13, 52 and mice53, 54 that T cells play an important protective role during dengue infection. In particular, recent studies in humans showed that both CD4⁺ and CD8⁺ T cells restricted to HLA molecules that were previously reported to associate with protection from severe dengue display responses of higher magnitude and polyfunctionality compared with those restricted to HLA molecules associated with severe disease. These findings suggest that both the magnitude and the quality of the CD4⁺ and CD8⁺ T‐cell response are important factors in determining protection from severe disease.13, 55

Further support of a potential protective role of T cells emerges from phase II/III clinical trials of the first licensed tetravalent dengue vaccine constructed on the backbone of the YFV vaccine strain. In this vaccine formulation, the YFV prM and envelope proteins are replaced with those from DENV 1–4 while the non‐structural proteins and capsid, which are important targets of the T‐cell response, are derived from YFV. Preliminary evidence from a phase II trial has suggested that vaccination can boost pre‐existing CD8⁺ T cells specific for DENV NS3 proteins in flavivirus‐immune individuals whereas it is unable to induce these cells de novo in flavivirus‐naive individuals.56 In contrast, a robust CD8⁺ T‐cell response directed against the YF17D NS3 protein is observed in both naive and flavivirus‐immune individuals.56, 57 Interestingly, phase IIb/III clinical trials show that this vaccine protects against disease more effectively in flavivirus‐immune compared with flavivirus‐naive vaccinees,58 suggesting a possible contribution of dengue‐specific T cells to immune protection.

In conclusion, the preferential expansion of pre‐existing dengue‐specific T cells during secondary dengue infection has been extensively documented but the outcome of these events in terms of increased protective immunity or enhanced immunopathology is currently unclear. Further investigations are needed to strengthen the evidence emerged in recent studies that points towards a protective role of pre‐existing memory T cells during secondary infection/vaccination.

Cross‐reactive T‐cell responses between DENV and ZIKV

The above considerations are also relevant for ZIKV, which is spreading to areas of DENV hyper‐endemicity. Therefore, individuals exposed to ZIKV are likely to harbour pre‐existing DENV‐reactive memory T cells. Given the antigenic similarity between DENV and ZIKV and the reported cross‐reactivity to ZIKV of DENV envelope‐specific antibodies,59, 60 cross‐reactive recognition of ZIKV by pre‐existing dengue‐specific T cells is likely to occur. Sequence identity between DENV and ZIKV ranges from 44% and 56% for the structural proteins envelope and capsid, respectively, to 68% for the more conserved non‐structural proteins NS3 and NS5. The homology between immunogenic CD8⁺ T‐cell peptides from DENV and the corresponding ZIKV sequence is evident from Table 1. However, a recent study reported the lack of cross‐reactivity between envelope‐specific CD4⁺ T cells from DENV or ZIKV‐infected patients and the heterologous protein as well as between NS1 ZIKV‐specific CD4⁺ T cells and DENV NS1 protein.60 Studies are needed to further evaluate the extent and the implications of cross‐recognition of CD4⁺ and CD8⁺ T‐cell epitopes in DENV‐ and ZIKV‐immune individuals.

Studies on cross‐reactive T‐cell recognition (also known as ‘heterologous immunity’) in mouse models of lymphocytic choriomeningitis virus (LCMV) infection have demonstrated that, at least in mice, T‐cell cross‐reactivity can have opposing outcomes and can lead to either enhanced or diminished protective immunity and altered immunopathology.61 In LCMV‐immune mice cross‐reactive CD8⁺ T cells are impaired in recognition and clearance of a secondary‐infecting LCMV strain bearing mutations in T‐cell epitopes.62 In contrast, LCMV‐immune mice displayed stronger protective immunity towards infection with Pichinde virus (PV) and vaccinia virus, which was mediated by LCMV‐specific CD4⁺ and CD8⁺ T cells.63 In a different model, cross‐reactive LCMV‐specific T cells could protect against an otherwise lethal dose of vaccinia virus.63 Furthermore, pre‐existing immunity was shown to dramatically skew the hierarchies of the T‐cell response. For example, a normally sub‐dominant T‐cell epitope that is conserved between LCMV and PV became immunodominant in LCMV or PV‐immune mice infected with the heterologous virus.61 In patients with fulminant acute hepatitis C virus (HCV) infection the HCV‐specific T‐cell response is dominated by pre‐existing memory CD8⁺ T cells specific for influenza virus that cross‐react with an HCV epitope and expand vigorously, resulting in severe HCV‐induced liver pathology.64

Overall, these considerations highlight the complexity of the cross‐reactive T‐cell response, which in humans is further complicated by the individuals’ unique infection history and the highly polymorphic nature of HLA molecules. After over a decade of research we are still unable to reconcile the contradicting findings, which point to either a protective or a pathological role of cross‐reactive memory T cells during secondary dengue infection. Understanding the impact of heterologous immunity in the context of ZIKV infection represents a new and exciting challenge.

Concluding remarks

Given the high homology between DENV and ZIKV, the knowledge that we have gained on the T‐cell response to DENV may provide insights into the nature of T‐cell immunity to ZIKV. The geographical overlap of DENV and ZIKV infections highlights the need to understand the impact of pre‐existing immunity to DENV (acquired either through natural infection or vaccination) on protective immunity and/or immunopathology during ZIKV infection.

Disclosures

The authors declare that they have no financial or commercial conflict of interests.