|Home | About | Journals | Submit | Contact Us | Français|
Sylvatic dengue viruses (DENV) are both evolutionarily and ecologically distinct from human DENV and are maintained in an enzootic transmission cycle. Evidence of sylvatic human infections from West Africa and Southeast Asia suggests that sylvatic DENV come into regular contact with humans. Thus, this potential of emergence into the human transmission cycle could limit the potential for eradicating this cycle with vaccines currently in late stages of development. We assessed the likelihood of sylvatic DENV-4 emergence in the face of natural immunity to current human strains and vaccination with two DENV-4 vaccine candidates. Our data indicate homotypic neutralization of sylvatic and human DENV-4 strains by human primary convalescent and vaccinee sera but limited heterotypic immunity. These results suggest that emergence of sylvatic strains into the human cycle would be limited by homotypic immunity mediated by virus neutralizing antibodies produced by natural infection or vaccination.
Dengue viruses (DENV) are arthropod-borne viruses (arboviruses) in the genus Flavivirus (family Flaviviridae) that utilize Aedes (Stegomyia) spp., primarily Ae. aegypti and to a lesser degree Ae. albopictus, as vectors for transmission in urban and peri-urban settings (urban transmission cycle). In Southeast Asia and West Africa, DENV are also transmitted in an enzootic cycle between non-human primates and arboreal Aedes spp. mosquitoes. There are four antigenically distinct but genetically related serotypes (DENV-1, -2, -3 and -4) within the dengue (DEN) antigenic complex (Calisher et al., 1989). Unlike most arboviruses, DENV are extremely restricted in their natural vertebrate host range, which most likely includes only primates. Currently, all four DENV serotypes can be found in nearly all urban and peri-urban tropical and subtropical environments where Ae. aegypti is present. By current estimates this distribution puts over half of the global human population at risk for infection. The impact of DENV infections on human health is enormous; over 200 million infections and 2 million cases of of dengue hemorrhagic fever (DHF) occur each year, with a case fatality rate of up to 5% (Kyle and Harris, 2008). Most profoundly, the majority of the DEN-associated disease in hyperendemic regions is borne by children (Clark et al., 2005; Mathers et al., 2007; Witayathawornwong, 2005), although recent evidence from Southeast Asia and Latin America suggests that adults are also at high risk (Fox et al., 2011; Guilarde et al., 2008; Hanafusa et al., 2008; Koh et al., 2008; Siqueira et al., 2005; Wichmann et al., 2004), particularly in urban areas that are transitioning or have already transitioned to hyperendemicity.
Phylogenetic (Chen and Vasilakis, 2011; Rico-Hesse, 1990; Twiddy et al., 2003; Vasilakis et al., 2008b; Wang et al., 2000) and ecological studies (Cordellier et al., 1983; Hervy et al., 1984; Monlun et al., 1992; Roche et al., 1983; Rudnick, 1965; Rudnick, 1978, 1984, 1986; Smith, 1956, 1958) indicate that the ancestral sylvatic DENV are both ecologically and evolutionarily independent from the current endemic DENV circulating within urban transmission cycles. However, data from West Africa and Southeast Asia suggest that sylvatic DENV come into regular contact with humans. For example, in West Africa the gallery forest-dwelling mosquito, Ae. furcifer, which is highly susceptible to sylvatic DENV infection (Diallo et al., 2005), disperses into villages and may be responsible for sylvatic DENV infection of humans (Diallo et al., 2003; Fagbami et al., 1977). In Southeast Asia, Ae. albopictus may transfer sylvatic DENV from the forest into human habitats (Rudnick, 1986). Sylvatic DENV infection can cause human disease in both rural peridomestic and urban settings as documented by spillover epidemics (Carey et al., 1971; Vasilakis et al., 2008c), and human infections in West Africa (Franco et al., 2011; Monlun et al., 1992; Robin et al., 1980; Saluzzo et al., 1986) and Southeast Asia (Cardosa et al., 2009). Clinical illness due to sylvatic DENV not only is indistinguishable from classic dengue fever (DF), but also has the potential to progress to severe disease (Cardosa et al., 2009; Franco et al., 2011). Because the only means to determine whether the DENV strain causing human infection is from the sylvatic or the urban transmission cycle is by sequencing the virus genome, human infection by sylvatic strains is often misclassified etiologically as dur to urban strains.
Collectively, these observations combined with: (i) experimental evidence in models of human (Vasilakis et al., 2007) and mosquito infection (Diallo et al., 2008; Diallo et al., 2005) (Hanley and Vasilakis, unpublished data) that adaptation is not required for urban transmission; and (ii) the continuing risk of DENV emergence from the sylvatic cycle resulting in human infection, which can manifest as symptomatic dengue disease, suggest that sylvatic dengue spillover may occur at a greater frequency than is currently recognized. A recent study showed broad neutralization by vaccinated monkey serum against a large panel of DENV-1–4, including a DENV-2 sylvatic strain (Barban et al., 2012). Our earlier study (Vasilakis et al., 2008a), evaluating the neutralization capacity of convalescent sera collected from dengue vaccine recipients and from DENV-2 and DENV-3 patients following primary infection, also suggested that the emergence of sylvatic DENV strains into the human transmission cycle would be limited by homotypic humoral immunity.
In this study, to assess the likelihood of current sylvatic DENV-4 re-emergence (emergence of extant sylvatic DENVs into the human transmission cycle) in the face of immunity induced by current urban-circulating DENV strains or vaccine candidates, we tested sera collected from DENV-4 vaccine recipients and convalescent-phase sera from DENV-infected patients against geographically and genetically diverse sylvatic and human DENV-4 strains (Table 1).
For vaccination to achieve effective immunity against DEN disease, the immune response must be effective against all four DENV serotypes and must be long-lived. Natural infection with DENV results in lifelong protection against reinfection with the same serotype (Halstead, 1974; Okuno et al., 1983; Papaevangelou and Halstead, 1977; Rosen, 1986; Sabin, 1952; Tadano et al., 1983) but only in transient cross-protection against heterotypic serotypes (Guirakhoo et al., 2006; Sabin, 1952). Homotypic immunity is mediated by neutralizing antibodies directed mainly against the viral envelope (E) protein (Crill and Roehrig, 2001; Kaufman et al., 1987; Roehrig et al., 1998) and other minor antigens located in the viral membrane [(M) (Kaufman et al., 1989) and nonstructural 1 (NS1) (Schlesinger et al., 1986) proteins] as well as by cell-mediated immunity. However, heterotypic immunity is associated with cross-reactive neutralizing antibodies to E and prM that decline rapidly after infection (Pengsaa et al., 2006; Sabin, 1952).
Previously, we assessed the likelihood of current sylvatic DENV re-emergence in the face of immunity to current vaccine candidates, including the DENV-4 candidate vaccine rDEN4Δ30, by evaluating the neutralizing capacity of sera from DENV4Δ30 vaccinees against reference sylvatic and human DENV-4 strains (Vasilakis et al., 2008a). Here, we assessed the potential of all available sylvatic DENV-4 strains to re-emerge in the face of immunity to an expanded set of human strains, as well as the more attenuated DENV-4 rDEN4Δ30-200,201 and rDEN4Δ30 vaccine candidates, by evaluating the neutralizing capacity of sera from human DENV vaccinees and convalescent patients from Puerto Rico, Mexico, Singapore, Sri Lanka, the West Indies, Honduras, and Thailand.
Sera collected at 42 days post-vaccination from 7 and 10 placebo-controls participating in the clinical trials evaluating rDEN4Δ30-200,201 and rDEN4Δ30, respectively, were unable to neutralize any DENV tested (PRNT60 < 20). Sera collected from all rDEN4Δ30-200,201 vaccinees were previously shown to neutralize the parent DENV-4 strain 814669 (McArthur et al., 2008) and were not retested in the current study. Overall, sera from 16/19 rDEN4Δ30-200,201 vaccinees neutralized some or all of the DENV-4 strains. Weak to modest (PRNT60 = 20 – 160) homotypic neutralization was exhibited by vaccinee sera (Table 2). Sera from 13/19 vaccinees neutralized all of the DENV-4 tested, while sera from three additional vaccinees neutralized all viruses tested except for the India G11337 (3 test sera 214.01.14, 214.01.16, 214.01.18) and the H241 strains (test serum 214.01.14). Weak to modest (PRNT60 = 20 – 160) homotypic neutralization responses were exhibited by sera collected at 42 days post vaccination (Table 2). The levels of homotypic neutralization further decreased 180 days post vaccination (PRNT60 = 20 – 40) (data not shown). Similarly, sera collected from 21/22 persons vaccinated with rDEN4Δ30 neutralized some (14) or all (7) of the DENV-4 tested, and several sera from other persons were unable to neutralize human or sylvatic DENV-4 strains (Table 3). The homotypic neutralization responses exhibited by sera collected 42 days after vaccination ranged from weak to strong (PRNT60 = 20 – >640) (Table 3) (sera collected 180 days post vaccination were not tested) and were collectively more potently neutralizing than sera from the rDEN4Δ30-200,201 vaccinees (mean log titers 1:40 and 1:28, respectively, P=0.0163), consistent with greater attenuation of the rDEN4Δ30-200,201 virus. Sera from all but two persons vaccinated with either rDEN4Δ30-200,201 or rDEN4Δ30 failed to neutralize heterotypic human DENV-1 – 3 or sylvatic DENV-2 (PRNT60 < 20) (Table 2 – 3). Serum from patient JHU02 exhibited a strong heterotypic response to the sylvatic DENV-2 DKD811 strain (Cardosa et al., 2009). Interestingly serum from one vaccinated subject (JHU 22) exhibited a modest to relatively strong heterotypic neutralizing response to DENV-1 and - 3 (PRNT60 = 40 – >640) and a robust neutralization titer (>1280) to DENV-2, suggesting a secondary antibody response (previous DENV-2 infection), or possibly an inapparent West Nile virus infection and ‘original antigenic sin’ (Halstead et al., 1983; Kuno et al., 1993). However, there were stringent eligibility criteria for this trial, including absence of prior exposure to flaviviruses (Durbin et al., 2006), and this volunteer was previously found to be seronegative by PRNT to DENV-1 – 4, yellow fever (YFV) and WNV and by hemagglutination-inhibition assay to Saint Louis encephalitis virus (SLEV), WNV, and Japanese encephalitis virus (JEV) prior to vaccination. Furthermore, it is very unlikely that this volunteer (JHU 22) would have been exposed to any flavivirus other than WNV because he/she was a Baltimore resident with no travel history to any DENV-endemic countries. We reported a similar observation previously (Vasilakis et al., 2008a). The broad heterotypic response after vaccination or previous, unknown exposure could be due to antigenic mismatch between the vaccine and DENV strains used for the PRNT (Lanata et al., 2012; Sabchareon et al., 2012).
The vaccinee sera offer a unique opportunity to explore the relationship between neutralization and DENV genetic diversity. Because the sequences are known for the vaccine strain and for the endemic and sylvatic strains used in the assays, it is possible to test the relationship between differences in neutralization titers of sera and genetic differences among the viruses used in the assays. Using published sequences for the rDEN4Δ30 vaccine strain and the endemic and sylvatic strains (Table 1), genetic p-distances between the amino acid sequences of the structural proteins were calculated. As expected, the structural proteins of rDEN4Δ30 were much more similar to those of the endemic DENV-4 strains than to that of the sylvatic E glycoprotein, with nearly 100% amino acid sequence identity between the vaccine strain and its parent 814669, followed by INH6412 (99.4%), Haiti73 (99.3%), IndiaG11337 (97.3%) and H241 (97.0%). The sylvatic strains’ structural proteins were all 95% identical to the vaccine strain. These distances were plotted against the log PRNT60 titers for each vaccine recipient and subjected to regression analysis. For the rDEN4Δ30 recipients, 18 of the 22 sera demonstrated a positive relationship between the genetic relatedness of the neutralized virus to the vaccine strain and the neutralization titer and, of these, 11 of the 18 were statistically significant (P<0.05). The only sera that failed to show a relationship were JHU 5 and JHU 11 (non-neutralizing sera), the weakly neutralizing sera JHU 2, and the potently neutralizing sera JHU 6.
Of note, the less potently neutralizing rDEN4Δ30-200, 201 vaccinee sera did not show a consistent relationship between genetic distance and neutralization titer. Of the 19 sera tested, only 11 showed a positive relationship between neutralization titer and genetic distance, and of these 11, only 2 had a statistically significant correlation (214.01.01 and 214.01.13). This likely is because of the overall lower antibody titers induced by these more attenuated vaccines. This explanation is supported by the observation that these vaccinees showed reduced homotypic neutralization at 180 days post vaccination.
We also assessed the ability of convalescent sera collected from primary DENV-4 pastients to neutralize sylvatic and endemic DENV-4, (Table 4), including (i) sera obtained 9–21 days (Puerto Rico and Singapore) or up to a year (Mexico) after onset (Table 4), as part of routine national surveillance programs, (ii) very late convalescent sera (defined as those collected >2 years after documented DENV infection) obtained as part of an ongoing dengue in traveler study (French West Indies and Honduras) and, (iii) anonymous sera from a Sri Lankan blood bank (length of convalescence unknown) and a reference serum collection (Thailand). Primary DENV infections were defined by the capture IgM:IgG ratio during the acute phase of disease (0–5 days after onset of symptoms) (Falconar et al., 2006) or by a monotypic neutralization profile on late convalescent sera. Because a limited volume of DENV-4 primary convalescent sera was available, we were unable to evaluate their neutralization capacity with a large collection of DENV-4 strains. Of the early convalescent Puerto Rico sera, all but one (PR-A02 with India G11337 strain), robustly neutralized of both endemic and sylvatic DENV-4 (PRNT60 40 – >1280) (Table 5). Sera obtained from Puerto Rico and Singapore exhibited weak to modest heterotypic neutralization of endemic DENV-1 – 3 or sylvatic DENV-1 and -2 (PRNT60 = 20 – 320), except serum PR-A04, which strongly neutralized the sylvatic DENV-1 strain (PRNT60 >1280). Several sera exhibited no heterotypic virus neutralization of endemic DENV-1 – 2 or sylvatic DENV-2 strains (PRNT60 < 20) (Table 5). Notably, both sera from Mexico weakly neutralized both endemic and sylvatic DENV-4 (PRNT60 20 – 40), but not endemic DENV-1 – 3 or sylvatic DENV-1 and -2 (PRNT60 < 20), whereas one serum exhibited a relatively strong virus neutralizing activity against sylvatic DENV-1 (PRNT60 = 160)(Table 5). Because these sera were collected several months after infection (Table 4), the lack of heterotypic neutralization is consistent with transient heterotypic cross-protection after infection. The relatively robust heterotypic responses observed in PR-A04 and MX-01 samples suggest that the patients had a previously undiagnosed, heterologous DENV infection. This concept was proposed by Halstead (Halstead et al., 1983) as a mechanism of DENV immune response involved in sequential infections. In such instances, the immune response to a secondary infection is dominated by the proliferation of cross-reacting memory cells induced by the primary infection, which may be of lower affinity for the secondary challenging antigen.
The late convalescent sera DT100 (West Indies, 4 years post acute infection) and DT102 (Honduras, 2 years post acute infection) showed that durable and broad homotypic neutralization can persist several years after infection. The blood donor sera DV002 and GS58, both from Sri Lanka, showed a less robust but equally broad neutralization against both endemic and sylvatic viruses. The DENV-4 reference sera, SS06/105, 302 and SS07/333 were interesting in that they were broadly neutralizing but had as much as a 16-fold difference in neutralization titers between endemic and sylvatic DENV-4. This suggests that, while natural infection induces broad intraserotypic neutralization, there can be significant variation within that serotype, underscoring the role of genotype-specific epitopes in antibody mediated protection.
Our results demonstrate the ability of convalescent sera collected after primary DENV-4 infection as well as from DENV-4-vaccinees to neutralize genetically diverse DENV-4. The observed protection is mainly due to homotypic immunity, and consistent with Sabin’s observations, we observed only a transient heterotypic immunity. Sera from human vaccinees collected at 42 and 180 days following vaccination with rDEN4Δ30-200,201 or at day 42 after rDEN4Δ30 vaccination exhibited limited to no heterotypic immunity. Furthermore, our rDEN4Δ30-200,201 data corroborate the reduced levels of neutralization titers observed previously, attributed to the greater attenuation of this candidate (McArthur et al., 2008).
Several investigators have suggested that amino acid differences in the lateral ridge of domain III of the E glycoprotein (Cockburn et al., 2012; Endy et al., 2004; Wahala et al., 2010; Zulueta et al., 2006) are responsible for variable intra-serotypic neutralization. For the viruses in this study, there are thirteen variable amino acid positions on the E glycoprotein that are conserved between the endemic and sylvatic strains. Five are in domain I, only one occurs in domain II, and the remaining eight residue differences are in domain III, of which only two occur along the lateral ridge (A329T and V382A). It is unlikely that these two polymorphisms alone would account for differences in neutralization. Moreover, recently published research suggests the flavivirus targets of human antibodies are likely to be non-EDIII or quaternary epitopes that potentially involve residues in multiple domains in variable neutralization (de Alwis et al., 2012).
As noted above, within-serotype neutralization varied broadly. For the rDEN4Δ30 vaccinee sera, neutralization titer was significantly correlated with genetic distances between the vaccine strain and the tested strains in half the sera tested. Additionally, the PRNT60 titers of recent convalescent sera against endemic and sylvatic strains did not vary significantly for most samples.
Collectively, our data suggest that re-emergence of sylvatic DENV-4 into the human transmission cycle may be limited by homotypic humoral immunity, induced either by vaccination or prior natural DENV-4 infection. The potential licensing of effective DENV vaccines raises the prospect of control or even eradication of the human transmission cycle, which relies solely on human-mosquito-human transmission. However, sylvatic transmission cycles are not amenable to control by human vaccination and epidemiological evidence suggests that sylvatic DENV come into regular contact with humans and can cause severe disease (Cardosa et al., 2009; Franco et al., 2011) or transient spill-over to urban settings (Carey et al., 1971; Vasilakis et al., 2008c). Furthermore, recent experimental evidence indicates a low to non-existent adaptive barrier for the emergence of sylvatic DENV into the human population (Vasilakis et al., 2007), implying that re-emergence is a clear and present danger. Therefore, even as efforts to control circulation of human dengue intensify in a manner analogous to historic efforts to control urban yellow fever, reduction and ultimate eradication of dengue from human populations solely due to vaccination campaigns could be, at best, short-lived. The urban yellow fever experience suggests that successful eradication of dengue will hinge on sustained vaccination of susceptible populations, including those who are at risk from the introduction of sylvatic DENV, with potently immunogenic and long-lasting vaccines, as well as control of the urban mosquito vectors.
Written consent to participate in the study was obtained from each subject enrolled in the dengue vaccine studies. All data were handled anonymously and confidentially. The studies were conducted at the Center for Immunization Research (CIR) at the John Hopkins Bloomberg School of Public Health under an investigational new drug application for both rDEN4Δ30-200,201 (BB-IND 12670) and rDEN4Δ30 (BB-IND 12977). The University of Texas Medical Branch Institutional Review Board reviewed the study as protocol UTMB-IRB#10-047. Human sera from Singapore were from the early dengue infection and outcome study (Low et al., 2006), and were collected following written informed consent and approved by the National Healthcare Group Domain Specific Review Board (DSRB 5B/05/013). Human sera from Puerto Rico, and Mexico, obtained from routine surveillance, and sera from US travelers were de-identified and approved for this research study under IRB exemption 4797 at the CDC, CI986-L62 Mexico and 08-0895 UNC, respectively.
Monolayer cultures Vero cells (obtained from American Type Culture Collection, Bethesda, MD) were grown at 37°C in Dulbecco’s minimal essential medium (DMEM) (4.5 g/L D-Glucose) with 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S). C6/36 mosquito cells (a generous gift from Ilya Frolov) were grown at 28°C in Dulbecco’s minimal essential medium (DMEM) (4.5 g/L D-Glucose) with 10% heat-inactivated FBS, 1% P/S and 1% tryptose phosphate broth (TPB).
DENV-1,-2, -3 and -4 isolates of low passage history (Table 1), were obtained from the World Reference Center for Emerging Viruses and Arboviruses, University of Texas Medical Branch. The Dominica strain, 814669, from which both tested DENV-4 vaccine candidates were derived, was provided by Dr. Anna Durbin. Viral isolates were passaged once in C6/36 cultures to obtain high titer stocks. Supernatants were clarified from cellular debris by low-spin centrifugation (630 × g, 20 min, 4°C) stabilized with the addition of 10X SPG (2.18 M Sucrose, 0.038M KH2PO4, 0.072M K2HPO4 and 0.054M L-glutamate), and stored at −;80°C.
The rDEN4Δ30 vaccine candidate was derived from the DENV-4 strain 814669 (Dominica 1981) by the removal of 30-nucleotides (nt) from the 3’ UTR of the genome (Durbin et al., 2001). The rDEN4Δ30-200,201 vaccine virus was derived from the rDEN4Δ30 vaccine virus by paired charge-to-alanine mutagenesis. (Hanley et al., 2004; McArthur et al., 2008). The rDEN4Δ30-200,201 virus was generated as described previously in Hanley etal. (Hanley et al., 2004).
Serum samples were collected as part of previously described DENV vaccine trials (Durbin et al., 2011; McArthur et al., 2008). Healthy adult male and non-pregnant female volunteers were recruited from metropolitan Baltimore, Maryland. Informed consent was obtained from each enrolled healthy volunteer, between 18 – 50 years of age, in accordance with the Code of Federal Regulations (CFR21, Part 50). Each enrolled volunteer met the following eligibility criteria: normal findings during physical examination; negative for antibodies to all DENV, yellow fever virus, West Nile virus, St. Louis encephalitis virus, Japanese encephalitis virus, human immunodeficiency virus, and hepatitis C virus; negative for hepatitis B surface antigen; normal values for complete blood count (CBC), and urinalysis. Additional safety-related exclusion criteria were also applied. Female volunteers were required to have a negative urine pregnancy test at least three days prior to vaccination and on the day of vaccination and to agree to use contraception or abstain from sexual intercourse for the duration of the study. Sera provided for this study represent collections at day 42 post vaccination as described in McArthur etal (McArthur et al., 2008).
Sera from convalescent patients after primary infection with DENV-4 obtained 9–21 days (Puerto Rico and Singapore) or up to a year (Mexico) after the onset of symptoms (Table 4), were obtained from routine surveillance, and were de-identified and approved for this research study under IRB exemption 4797 at the CDC. Late convalescent sera were collected as part of an ongoing dengue in traveler study under University of North Carolina (UNC) IRB approval 08-0895. Anonymous blood bank and reference sera were kindly provided by the lab of Aravinda de Silva, UNC School of Medicine, Department of Microbiology.
PRNTs were performed in 12-well, Vero-microplate-cell cultures, using a fixed virus inoculum [~800 focus forming units (FFU)] against varying serum dilutions (1:10--1:640). Serum samples were diluted in minimal essential medium (MEM), containing 2% fetal bovine serum. Virus was mixed with an equal volume of each serum dilution and the mixture was incubated 1hr at 37°C. Then, 250 µL of the serum-virus mixture was placed into Vero cultures and incubated 1hr at 37°C. A 1.5 mL volume of 4% methycellulose in OPTIMEM-I overlay was placed in each well and the plates were incubated at 37°C for 4–5 days. The plates were then fixed with 1:1 methanol:acetone and foci were stained immunologically and counted to determine the level of virus neutralization, as described previously (Vasilakis et al., 2008a; Vasilakis et al., 2007). The PRNT titers were scored as reciprocal of the highest dilution of serum that inhibited 60% of foci.
All statistical comparisons and regression analyses were calculated in JMP v7.0 (Cary, NC). Mean log titers were compared by one-way ANOVA. P<0.05 was considered significant in all comparisons. Correlations between PRNT60 titers and genetic P-distances were tested using a standard least squares model. P-distances between DENV-4 structural protein amino acid sequences were calculated with MEGA V5.0 (Tamura et al. 2011) default settings.
We thank Ilya Frolov for providing C6/36 cells. AP is supported by the National Institute of Allergy and Infectious Diseases Intramural Research Program, National Institutes of Health under contract HHSN272200900010C; SVM is supported by the James W. McLaughlin Fellowship Fund; SLR is supported by 2T32AI007536-11A1 postdoctoral training grant; EEO is supported by the Singapore National Research Foundation under its Translational Clinical Research Award administered by the National Medical Research Council; SW is supported for dengue research by NIH grant RO1 AI069145; WBM is supported by SERCEB/NIH U54 AI057157 SE-RP-004 and SE-CD-007 ; RT is supported by NIH contract HHSN272201000040I/HHSN27200004/D04; NV is supported by start-up funds provided by the Department of Pathology, UTMB.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
This work was presented in part at the 'Emerging and Re-emerging Vector-Borne and Zoonotic Viral Infectious Diseases in Southeast Asia', meeting in Hanoi, Vietnam, September 2010 and at the 58th Annual Meeting of the American Society for Tropical Medicine and Hygiene, Atlanta, November 2010.