Dengue is a mosquito-borne disease affecting over half of the world's population. It is over a century old disease in Malaysia [Reference Skae1]. All four dengue virus (DENV) serotypes caused dengue and the different DENV serotypes have co-circulated in many regions of the world. In the Klang Valley, Malaysia, the major outbreak cycles involving the same DENV serotype, the homotypic cycle, occurred at every 4–10 years interval [Reference Tan2, Reference Abubakar and Shafee3]. The major outbreak cycles involving the different serotypes, the heterotypic cycle, occurred in a sequential manner involving DENV-3, followed by DENV-1, and then DENV-2 (DENV-3/DENV-1/DENV-2) in that order [Reference Tan2]. So far, only three DENV-3/DENV-1/DENV-2 supra-serotype cycles have been recorded [Reference Tan2]. The DENV-4 maintained low circulation in the background of DENV-1, DENV-2 and DENV-3 [Reference Tan2]. DENV-3-homotypic cycles were recorded in 1986, 1992–1995 and 2002 [Reference Tan2]. The anticipated DENV-3 homotypic cycle, which could appear between 2006 and 2012, did not materialise [Reference Tan2]. Notably, during the 2002 DENV-3 homotypic cycle, DENV-3 was only implicated in 42% of the total serotyped cases compared to the 1992–1995 cycle, where DENV-3 contributed up to 90% of the typed DENV cases [Reference Tan2]. Although several DENV-3 genotypes have been found in Malaysia for the past-30 years [Reference Tan2, Reference Tan4, Reference Tan5], DENV-3 genotype II (DENV-3/II) has been the only genotype associated with the outbreak cycles [Reference Tan2]. Previous genetic analysis showed that these cycles were caused by two closely related groups within DENV-3/II [Reference Tan and AbuBakar6]. The different outbreak magnitude and duration of DENV-3/II-associated homotypic outbreak cycles in the same locality suggests possible underlying factor(s) contributing to the differences. Immune-driven selection of virus clade that can escape host immunity has been shown as one of the possible mechanisms for selecting the dominant circulating strains of many infectious agents [Reference Tan4, Reference Collins, Murphy and Weaver7] and shaping the epidemic pattern [Reference Adams8]. To understand if closely related strains within the same DENV-3 genotype would respond similarly to the existing DENV immunity in the local setting, we investigated the neutralisation capacities of homotypic and heterotypic immune sera against the representative DENV-3/II strains of 1992–1995 and 2002 DENV-3 outbreaks. The serum sample against DENV-4 infection was not available; hence, the heterotypic neutralisation by DENV4 serum was not assessed in the current study. Subsequently, we examined the genetic variations between the DENV-3/II strains that could influence the neutralisation capacity.

The representative DENV-3 strains responsible for 1992–1995 (DENV-3/II-E) and 2002 (DENV-3/II-F) outbreaks were prepared by serial propagation of the selected DENV-3 strains in C6/36 mosquito cells (Aedes albopictus cell line). The use of human serum samples in this study was approved by the Medical Ethics Committee of the University Malaya Medical Centre (MEC Ref No: 806.23 and 806.24). The dengue status of selected human serum samples was characterised using the anti-dengue IgM and IgG as previously described [Reference Tan4]. The IgM/IgG ratio of more than 1.4 was classified as the primary infection. The patients' sera with IgM/IgG ratio less than 1.2 or the detection of anti-dengue IgG concurrent with the isolation of DENV was characterised as a secondary infection [9]. The serum samples with IgM/IgG ratios between 1.2 and 1.4 were excluded from this study. The serum from donors tested negative for DENV antibodies (IgG and IgM) were used as a negative control. The selected serum samples were used in the focus reduction neutralisation test (FRNT) against the different DENV-3 strains as previously described [Reference Tan4]. The patients' sera representing the homotypic and heterotypic immunity in the local population used in the current study were listed in Table 1.

Table 1. List of patients' sera used in this study

Overall, the control sera (Negative sera) showed weak and non-reactive neutralisation responses against all DENV strains tested (FRNT₅₀<40; Table 2). All immune sera (homotypic and heterotypic) demonstrated specific neutralisation capacity (FRNT₅₀>40) against tested DENV-3 strains (DENV-3/E and DENV-3/F). The immune sera from patients with secondary infection neutralised DENV-3 strains effectively (FRNT₅₀⩾5120). The neutralisation titres of the homotypic immune sera (sDENV-3/I, sDENV-3/II-E, sDENV-3/II-F) against DENV-3/II-E and DENV-3/II-F ranged from 452.5 to 1280. The homotypic serum neutralisation capacity against DENV-3/II-E and DENV-3/II-F were similar with less than 2-fold of intergenotypic differences. For the analysis of heterotypic neutralisation of DENV-1 and DENV-2 immune sera against DENV-3 strains, results showed both DENV-1 immune sera (sDENV-1/I and sDENV-1/II) possess similar neutralisation capacities, with less than 2-fold of intergenotypic differences against DENV-3/II-E and DENV-3/F strains. By using DENV-1 homotypic neutralisation results (sDENV-1 neutralised DENV-1) as control, the sDENV-1/I neutralised DENV-3/II strains more efficiently than sDENV-1/II. The sDENV-1/I neutralised DENV-3/II-E (FRNT₅₀ = 640.0) and DENV-3/II-F (FRNT₅₀ = 806.3) were less than four-fold differences as compared to neutralisation against DENV-1 (FRNT₅₀ = 2031.9). The neutralisation capacity of the sDENV-1/II against DENV-3/II strains (FRNT₅₀ = 254.0–320.0), however, differed by at least four-fold to its own DENV-1 strain (FRNT₅₀ = 1280). For the heterotypic neutralisation of DENV-2 immune sera, the neutralisation efficiency of the sDENV-2/Cos against DENV-3/II-E (FRNT₅₀ = 100.8) was 5.04-fold lesser compared to DENV-3/II-F (FRNT₅₀ = 508.0). The DENV-3/II-F was neutralised by sDENV-2/Cos at a similar neutralisation capacity against its own serotype, DENV-2 (FRNT₅₀ = 452.5). The neutralisation capacities of sDENV-2/Asian against both DENV-3/II-E and DENV-3/II-F (FRNT₅₀ = 160–226.2) were four-fold less efficient than that against DENV-2 homotypic neutralisation (FRNT₅₀ = 905.1).

Table 2. Neutralisation capacity of DENV immune serum against DENV-3/II strains

DENV Envelope (E) protein is the main target protein of the neutralising antibodies. We examined the presence of naturally occurring amino acid variations encoded in DENV-3/II-E and DENV-3/II-F E protein. There were two amino acid variation sites, 132 H/Y and 479 V/A in E between the DENV-3/II-E and DENV-3/II-F (Fig. 1). The 132 H/Y is partially accessible residue located in domain I and 479 V/A in the transmembrane region. The solvent accessibility and the potential effects of the amino acid variation on protein stability were predicted using a Site-Directed Mutator (SDM) [Reference Worth, Preissner and Blundell10]. The genetic variation at residue E-132 was predicted to affect protein stability (Fig. 2).

Fig. 1. Genetic variations among DENV-3 genotype II within the envelope protein. The variation sites of DENV-3/II-E and DENV-3/II-F were highlighted in grey.

Fig. 2. Structural-based prediction of mutation-induced protein stability changes on DENV-3 ectodomain by SDM.

Overall, serum neutralisation results showed that both DENV-3/II-E (1992–1995 DENV-3 cycle) and DENV-3/II-F (2002 DENV-3 cycle) were sensitive to neutralisation by all homotypic (DENV-3) and heterotypic DENV-1 and DENV-2 sera. Our findings suggest that the broad cross-neutralisation of DENV-3/II strains by serum from both homotypic and heterotypic DENV serotypes could have mitigated transmission of DENV-3/II strains in DENV endemic areas such as Klang Valley [Reference Tan2, Reference Abubakar and Shafee3]. At any given time, the DENV transmission pattern at that location during the virus introduction could have been influenced by the dengue prevalence, the length of time since the last outbreak and the predominant DENV serotype responsible for the outbreaks. Many previous studies also suggested that the pre-existing population immune landscape and the circulating DENV genotypes act as the determinants that could shape the local virus transmission pattern [Reference Tan4, Reference Allicock11, Reference Santiago12]. Our study showed that different amplitude and outbreak cycles duration were caused by closely related DENV-3/II strains. The differential neutralisation of DENV-3/II genotypes by the cross-neutralising serum from patients with DENV-2 Cosmopolitan infection was observed. The DENV-3/II-F was effectively neutralised at a similar capacity against DENV-2, but not DENV-3/II-E. In Klang Valley, dengue outbreaks occurred in DENV-3/DENV-1/DENV-2 supra-serotype outbreak cycles pattern [Reference Tan2], where the local population have a high level of DENV-2 immunity but low DENV-1 immunity before the emergence of DENV-3 -associated outbreak cycles (Fig. 3). The DENV2 contributed to 54–86% and 57–63% of the dengue cases between 1990–1991 and 2000–2001 (2 years before the 1992 and 2002 outbreaks; Fig. 3). It was noted that the DENV-2/Cosmopolitan genotype was the dominant DENV serotype before the emergence of 1992–1995 and 2002 DENV-3 outbreaks cycles [Reference Chee HY13]. Hence, the high level of pre-existing DENV-2 immunity before the emergence of DENV-3 was mainly against the DENV-2/Cosmopolitan genotype. The susceptibility of DENV-3/II-F to DENV-2/Cosmopolitan serum could have mitigated its wider transmission, especially in DENV-2 endemic area as compared to DENV-3/II-E, resulting in a short and relatively low epidemic involving DENV-3/II-F reported in 2002. A similar observation was reported in our previous study where the DENV-2 immunity effectively neutralised DENV-3/I but not DENV-3/III [Reference Tan4]. Collectively, the neutralisation disparity of DENV-2 immune sera against different DENV-3 genotypes suggests that the DENV-2 heterotypic immunity could play a role in shaping the DENV-3 recurring outbreaks. The current study is a retrospective study, hence it completely relied on the type and quantity of samples stored in the repository. As only limited volume of serum samples were available, the current study, therefore, is limited to using pooled serum samples in the neutralisation assays

Fig. 3. Percentage of DENV-1, DENV-2 and DENV-3 isolated from University Malaya Medical Centre, Klang Valley.

.

Our study also revealed one amino acid difference at position 132 of E protein between DENV-3/E (Histidine) and DENV-3/II-F (Tyrosine). Both histidine and tyrosine are aromatic amino acids commonly found in the semi-buried area [Reference Betts and Russell14]. It was noted that the E-132 is a partially accessible residue located at domain I. It was previously identified as a critical residue interacting with the hinge region during conformational changes in fusion trimer formation [Reference Christian15]. Our results suggest that amino acid substitution at E-H132Y could lead to changes in amino acid solvent accessibility. Whether or not these biochemical differences involving histidine and tyrosine at E-132 would influence the contact of the DENV-2/Cosmopolitan antibodies against the DENV-3/II strains that may lead to the differential neutralisation between DENV-3/II-F and DENV-3/II-E remain to be investigated.

In summary, we described the disparity of DENV-2/Cos immune sera neutralisation capacities against DENV-3/II-E and DENV-3/II-F. This differential ability could have contributed to the different virus transmission efficiencies in the DENV-2 endemic areas, the Klang Valley pre-emergence of DENV-3, leading to different outbreak patterns. These findings could be useful to facilitate the understanding of the DENV spread pattern in a hyperendemic setting. Our findings also highlight the need to investigate further how the naturally occurring genetic variations within closely related DENV genotypes could impact the neutralisation profile and protective immunity.