4.2 The Virus and its Effect on Clinical Outcome
4.2.2 Phylogeny of the 2005 Singaporean Dengue Outbreak
4.2.2.1 Isolated Serotype 1 and Serotype 3 Viruses show different phylogenetic Tree Structures suggesting evolutionary Differences
Our phylogenetic analysis revealed a high degree of similarity between strains of the same serotype but identified different tree structures in serotype 1 and serotype 3.
DENV-1 isolates did not clearly group into clades of reasonable virus numbers and the tree branches were represented by low evolutionary lengths (Figure 3.38; Page 170 and Figure 3.39; Page 171). On the other hand, DENV-3 isolates showed a clearer tree structure which was defined by longer branch lengths (Figure 3.40; Page 172 and Figure 3.41; Page 151). These two major differences possibly suggest evolutionary differences between DENV-1 and DENV-3 isolates from the 2005 Singaporean outbreak.
In this respect, it is interesting to note that a DENV-3 strain showing a high level of similarity to the EDEN DENV-3 isolates was already detected in 2003 and might represent a common ancestor of the EDEN DENV-3 viruses isolated during the major outbreak in 2005 (Ong Swee Hoe, personal communication). In addition, DENV-1 was predominant in 2004 as well as in 2005 and occurrence of dengue cases was increasing since 2004 and peaking in 2005 with a sharp decrease of dengue cases in 2006 (MOH, 2006). Hence, we can assume that DENV-3 was present in Singapore earlier than DENV-1 and that a newly arisen DENV-1 strain might have triggered the increase of dengue cases. Moreover, the fact that DENV-1 was predominant over DENV-3 might support the idea of a replicative advantage of DENV-1 which was additionally
supported by the observed significant differences in viral load. A possible explanation might lie in host immune recognition. Longer maintenance of DENV-3 in Singapore possibly resulted in a higher proportion of people who had already been exposed to serotype 3, whereas serotype 1 was newer for the host immune system, thus resulting in a higher proportion of susceptible individuals which would finally lead to a replicative advantage of DENV-1 over DENV-3. Such a phenomenon had also been proposed to explain the differences in transmission patterns of DENV-2 and DENV-4 in the Americas (Carrington et al., 2005).
Another study observed that an increase in clade diversity of one serotype leads to a decrease in prevalence of another serotype (Zhang et al., 2005). This might be another possible explanation for the observed differences in the tree structures of DENV-1 and DENV-3. Diversity of DENV-1 strains (0.184%) is higher than diversity of DENV-3 strains (0.118%) and the observation of four clear established clades in DENV-3 stay in contrast to many small subgroups all over distributed in the DENV-1 clade EDEN 1.4. We could suggest that the replicative advantage of DENV-1 led to a higher mutation frequency and thus to an increase in clade diversity whereas the four DENV- 3 clades were well established and showed no increase in clade diversity. This was further supported by the longer branch lengths observed in the DENV-3 phylogenetic tree.
Further phylogenetic analysis of EDEN DENV-3 revealed that the strains belonged to DENV-3 subtype 3. It was additionally found that the strains showed a higher level of similarity to a Sri Lankan strain which belonged to clade A of subtype 3 (Ong Swee Hoe, personal communication). This may underline the observation that DENV-3 was
associated to a lower degree of severity because earlier studies showed that subtype 3 clade A (pre-DHF) variants were associated with milder disease manifestations whereas the occurrence of subtype 3 clade B (post-DHF) strains directly correlated with the emergence of DHF in Sri Lanka (Messer et al., 2003; Messer et al., 2002).
The discussed potential reasons for the observed differences in the structures of the phylogenetic trees are hypothetical and more investigation is needed to elucidate the evolutionary mechanisms involved. Furthermore, we have to bear in mind that the mean nucleotide difference was almost negligibly small that an overall comparison of the isolated EDEN strains to geographically and temporally different isolates would naturally result in one EDEN cluster as observed in an overall DENV-3 tree including virus isolates collected throughout the globe (Ong Swee Hoe, personal communication).
4.2.2.2 Serotype 3 Clades show Differences with regard to viral Load and Platelet Count
Comparison of the observed EDEN clades within serotype 3 revealed differences in viral load. More specifically, the significant difference was between EDEN 3.1 (highest viral load) and EDEN 3.2 (lowest viral load) (Figure 3.42; Page 143). This difference is supported by a high bootstrap value and leads to the assumption of a real mutation affecting viral replication and thus disease severity. However, the identified mutation responsible for the clear split between these two clades was a synonymous mutation (Asparagine) in the E protein (EDEN 3.1: 2083C; EDEN 3.2: 2083T) The fact that we do not see any amino acid mutation leads to the assumption that the
observed difference occurred by chance due to only four viruses in clade EDEN 3.2.
Furthermore, it is unlikely that a synonymous substitution in the E protein would lead to differences in viral load or host immune responses because the amino acid sequence of the E protein is the main antigenic determinant during dengue virus infection. By way of example, it was reported that a unique amino acid mutation in the E protein of DENV-2 was responsible for the loss of a N-linked glycosylation site which led to a decreased infectivity (Ishak et al., 2001).