Host genetics

Apart from the influence of viral genetic determinants, the host's genetic background with varying polymorphisms might have important consequences for disease susceptibility (94). Improvements in high-throughput genotyping of genetic polymorphisms have permitted a genome-wide approach to the investigation of host genetic susceptibility. However, most studies have not attempted functional trials to try to link genetic association with any process in disease pathogenesis.
Indirect evidence of the host's genetic importance has been derived from Cuban dengue epidemics where a reduced risk for DHF/DSS was observed in those with an African ancestry compared to those with European ancestry (3). The Cuban observations coincide with the low susceptibility to DHF reported in African and Black Caribbean populations (86, 94). It is interesting that despite the circulation of DENV in 19 African countries, there are only sporadic reports of DHF cases (94).
In order to better understand these population differences, the polymorphic HLA genes have been among the most studied candidates for genetic associations with DHF/DSS. Several serological studies of HLA class I alleles have been performed in ethnically and geographically distinct populations, and positive correlations of various HLA class I alleles with susceptibility to DHF have been found. A significantly higher frequency of HLA class I alleles A*31 and B*15 have been found in Cuban individuals with symptomatic DENV infection compared to asymptomatic controls, who showed an elevated frequency of HLA II alleles DRB1*07 and DRB1*04. The DRB1*04 was also the most frequent allele associated with resistance to DHF in the Mexican Mestizo populations of the Americas. Since the Mexican Mestizo population and the Cuban population share the same Amerindian genetic background, it is possible that the identification of the same HLA class II allotype could explain the association to dengue disease protection.
A case-control study in ethnic Thai cases also reported the association of HLA class I alleles (A2, A*0207, B46, B51) with different clinical outcomes. A similar study in a Vietnamese population confirmed the association with polymorphism of the HLA class I loci and DHF suslceptibility. The same study also found that polymorphisms in the HLA-DRB1 allele are not associated with DHF susceptibility, highlighting the findings in the Amerindian populations.
The number of studies on polymorphisms within genes other than the HLA loci remains low. Variants of the vitamin D receptor and the FcγRIIA gene are associated with resistance to severe dengue. In addition, an allelic variant of the DC-SIGN1 coding gene CD209 is believed to protect against DHF (94).

Host health and age

An increased association between severe dengue and bronchial asthma, diabetes mellitus, peptic ulcers, and sickle cell anaemia has been observed (94, 95). However, the impact of dengue on chronic diseases and other pathogens needs to be further investigated.
Primary infections are supposed to cause mild disease in children, compared to secondary infections that tend to lead to severe dengue. In South-East Asia, DHF/DSS is predominantly an illness in children. The greater relative prevalence of DSS in children relative to adults is believed to be due to the intrinsically more permeable vascular endothelium in children (94). There is no clear consensus; studies conducted in South American countries have reported similar (84) as well as contradictory results indicating that adults are the most affected (94).

Autoimmune responses in dengue virus infection

Anti-DENV antibodies can cross-react to host proteins and endothelial cells, and this could enhance the endothelial dysfunction observed in DHF/DSS. Antibodies against the viral surface E protein cross-react with plasminogen and have been associated with bleeding in acute DENV infection, and anti-DENV NS1 antibodies cross-react with host proteins and endothelial cells (49, 94). In addition, immune activation markers (e.g. IL-6, IL-8, TNFα, IFNγ, and complement components 3A and 5A) together with altered platelet, DC, monocyte, and T-cell functions suggest that immune responses to various DENV components could contribute to autoimmune processes resulting in DHF/DSS (94).

Antibody-dependent enhancement

A secondary infection by a heterologous DENV serotype is an important risk factor for developing DHF/DSS. The explanation lies within the cross-reactive antibodies raised after a primary DENV infection (90, 92, 94). Serotype-specific antibodies confer life-long immunity to the homologous serotype, whereas cross-protection against heterologous serotypes last for 3–4 months. Beyond this time period of cross-protection, the preexisting antibodies of sub-neutralizing concentration will instead cross-react with the heterologous virus facilitating viral infection of FcγR-bearing cells. This phenomenon is known as ADE (94). The limited cross-protection between the four DENV serotypes has allowed them to coexist in the same or overlapping geographical areas. Thus, their antigenic uniqueness has implied an evolutionary advantage (96).
Low-affinity/Low-affinity, sub-neutralizing antibodies and DENV form virus-antibody immune complexes that bind to Fcγ-receptors on monocytes. The net result will be a larger number of infected cells compared to the primary infection when there were no cross-reactive antibodies present, or compared to earlier after the primary infection when antibody levels are high enough to achieve neutralization of the heterologous virus. Hence, the viral biomass will be larger during a secondary DENV infection compared to during a primary DENV infection.
 In vitro studies indicate that non-neutralizing antibodies against the viral prM protein can potentially mediate ADE. These anti-prM antibodies are in addition non-neutralizing even at high concentrations (97, 98). The proposed hypothesis for prM-mediated ADE is based on the fact that the viral prM protein needs to be cleaved to render the virus infectious. Hence, immature virus particles that would otherwise be non- or less-infectious are rendered infectious in combination with anti-prM antibodies that mediate ADE to infect new host cells (97).
The time interval between heterotypic DENV infections is another parameter influencing the magnitude of ADE; a longer interval between heterologous DENV infections causes higher DHF/DSS ratios (85). The differences in DENV genotype could influence the pathogenic consequences, but a contributing risk factor is the progressive loss of heterotypic neutralizing antibodies (99). The time effect of DENV-specific antibodies can be seen in DENV-immune mothers and their infants. Before the age of 3–4 months, the maternally derived DENV-specific antibodies confer protection against a DENV infection. However, primary infections in infants aged between 4 and 12 months of age run a higher risk of developing severe dengue due to maternally derived non-neutralizing antibodies. The risk of severe dengue decreases after the age of 1 year as the concentration of cross-reactive antibodies declines (94, 100, 101).
A higher viral burden elicits a greater host inflammatory response and increased plasma levels of proinflammatory cytokines. Secondary DENV infections and severe disease in DHF/DSS patients have elevated serum levels of IL-2, IL-6, IL-8, IL-10, IL-13, IL-18, IFNγ, TNFα, and MCP-1 (102- 108). Thus, an increased infected cell mass would stimulate T-cell and cytokine responses that are proportional to the antigenic stimulus. This hypothesis is consistent with the observations that a high initial viremia or high NS1 concentrations in blood during secondary infections are associated with DHF/DSS (46, 59, 109, 110).
Accumulating evidence questions whether ADE of infection alone is sufficient to explain DHF/DSS (59). Severe dengue with plasma leakage can occur in primary infection without ADE. In addition, by the time plasma leakage occurs, viral titers are several logs below peak levels, and there are patients with high viral titers that do not develop plasma leakage (59, 109, 111). Thus, increased viremia alone is not the direct cause of plasma leakage and other mechanisms are involved in the cytokine storm. Furthermore, ADE is not a useful correlate of disease risk (112, 113).

Dengue virus virulence

ADE has dominated as the explanatory model for severe dengue disease in secondary infections. However, evidence for ADE in humans is indirect and controversial results against ADE exist (112, 114). Many parts of the world have become hyperendemic, implying that all four serotypes of DENV co-circulate in the same country, with the consequence that secondary infections are common scenarios. Epidemiological data also indicate that not all secondary infections cause DHF/DSS, and that there are even cases of tertiary and quaternary DENV infections (115). Studies from Thailand report that 0.08%–0.8% of dengue hospitalizations may be caused by tertiary and quaternary DENV infections (67). In Cuba, 17.5% of the total DHF dengue cases were caused by third or fourth infections (116).
The four DENV serotypes 1–4 diverge at ∼30% across the polyprotein (117), but each serotype also consists of phylogenetically distinct ‘subtypes’ or ‘genotypes’, which have different geographical distributions  (94). The hypothesis that some DENV genotypes have greater virulence and epidemic potential than others was introduced during the 1970s around the same time that the ADE phenomenon was coined (118-120). However, in contrast to the ADE hypothesis, experimental evidence for increased virulence was for long absent and, therefore, primarily based on epidemiological observations. Recent work has shed light on this question and confirmed what Rosen, et al. proposed almost four decades ago (118, 119).
There have been specific geographic examples of the appearance of DENV genotypes correlating to DHF/DSS epidemics. The appearance of a South-East Asian DENV-2 strain in the Americas in 1981 resulted in the sudden emergence of DHF/DSS cases. It turned out that DENV-2 could be subdivided into a variety of genotypes, minimally Asian and American. The Asian genotype is more virulent and more likely to result in DHF/DSS than the American genotype even after a secondary infection (121, 122). Viruses to the South-East Asian DENV-2 lineage replicate to higher titers in human DCs than American genotype viruses. It was also seen that the South-East Asian genotype infects and disseminates to the head tissue of Ae. aegypti mosquitoes more rapidly and in a greater proportion compared to the American genotype viruses (121).
The emergence of group B subtype III DENV-3 strain in Sri Lanka in 1989 is another example of clade replacement correlating with an increase in DHF/DSS (88, 123, 124). As for the South-East Asian DENV-2 strain, the invasive DENV-3 strain replicated to higher levels in mosquitoes and disseminated to the head tissue more readily than the displaced, native DENV-3 strain from Sri Lanka (125). Both traits likely enhanced the capacity to spread and displace endemic strains.
Based on the examples given, one hypothetical mechanism for increased virulence suggests that highly pathogenic DENV strains have been selected for enhanced ability to replicate in key human targets, such as macrophages and DCs (94, 126). Thus, virulent DENVs would produce more viruses per cell, resulting in higher viremia and inflammatory response, than with a low pathogenic strain (94, 127).
The second hypothesis for increased virulence proposes that strains associated with severe DHF/DSS better escape neutralization by the presence of serotype cross-reactive antibodies in the semi-immune host compared to strains associated with DF (128). Enhancement of virus replication following heterologous infection may favor coexistence of multiple serotypes. If such enhancement also results in increased transmission, DENVs from different serotypes would benefit from prior and concurrent circulation of several serotypes in the same location (96).
It is still not known if the tendency of certain genotypes to cause severe disease results from greater intrinsic virulence, or if greater virulence is a result of enhanced infectivity in the presence of heterologous antibodies, or a combination of the two. Determining whether DENVs differ in virulence, as well as identifying the genetic basis of such differences, is of fundamental importance.