Abstract
Dengue is the most prevalent arboviral disease of humans. The host and virus variables associated with dengue virus (DENV) transmission from symptomatic dengue cases (n = 208) to Aedes aegypti mosquitoes during 407 independent exposure events was defined. The 50% mosquito infectious dose for each of DENV-1–4 ranged from 6.29 to 7.52 log10 RNA copies/mL of plasma. Increasing day of illness, declining viremia, and rising antibody titers were independently associated with reduced risk of DENV transmission. High early DENV plasma viremia levels in patients were a marker of the duration of human infectiousness, and blood meals containing high concentrations of DENV were positively associated with the prevalence of infectious mosquitoes 14 d after blood feeding. Ambulatory dengue cases had lower viremia levels compared with hospitalized dengue cases but nonetheless at levels predicted to be infectious to mosquitoes. These data define serotype-specific viremia levels that vaccines or drugs must inhibit to prevent DENV transmission.
Keywords: infectious disease, virology, entomology
Dengue is globally the most important mosquito-borne viral disease of humans, with a global burden of ∼100 million cases per annum (1, 2). Aedes aegypti mosquitoes are the primary mosquito vectors of dengue viruses (DENV), of which there are four virus types (DENV-1–4). Multiple factors influence the likelihood of infection and dissemination of DENV in Ae. aegypti and include the amplitude of daily temperature fluctuations (3), mean temperature (4), and the genotype of mosquito and virus (5), among others (6). The extrinsic incubation period (EIP), a critical determinant of vector competence (7, 8), is widely accepted to be 7–14 d for DENV in Ae. aegypti, although a recent modeling analysis of historical DENV transmission data has suggested a wider range of 2–15 d at 30 °C (9). A major caveat to many of these observations is that they stem from laboratory experiments with artificially generated virus-spiked blood meals and often in-bred colony mosquitoes.
The temporal and virological variables associated with the transmission of DENV from a naturally infected human to a biting Ae. aegypti mosquito are poorly understood. Natural history studies of experimental DENV infection of small cohorts of human volunteers in the 1920s by Siler et al. (10, 11), likely using DENV-4 (12), and subsequent studies by Simmons et al. (13), likely using DENV-1 (12), suggested that the window of time before the onset of clinical symptoms that DENV-1 or DENV-4 could be transmitted to Ae. aegypti mosquitoes was 6–18 h or 2 d, respectively (14). After fever onset, the duration of infectiousness was 4–5 d for DENV-1 and up to 2 d for DENV-4, with an EIP in the mosquito of 10 d or more. Consistent with this, mosquito-biting studies by Gubler et al. in the 1960s (15–18) collectively estimated that dengue cases were infectious for 4–5 d after illness onset (range, 2–12 d). The human viremia level required to infect Ae. aegypti mosquitoes is unknown, and therefore it is uncertain what percentage of symptomatic (or asymptomatic) dengue cases are infectious to mosquitoes or for how long they are infectious. It is also unknown whether the human viremia level at the time a mosquito takes a blood meal influences the EIP and therefore the probability of onward DENV transmission in the lifespan of the mosquito.
A better definition of the variables associated with human-to-mosquito DENV transmission can inform the development of successful dengue vaccines or antiviral drugs by identifying the level to which viremia resulting from a natural DENV infection should be suppressed. Similarly, analysis of immune responses that reduce human infectiousness during acute dengue might help in understanding immunity. To these ends, the aim of the present study was to define host and viral parameters that shape transmission of DENV from naturally infected humans to Ae. aegypti mosquitoes.
Results
Human to Mosquito Transmission: Duration of Human Infectiousness and the 50% Mosquito Infectious Dose.
Between April and December 2011, 210 adult dengue cases with ≤72 h of fever were enrolled and experimentally exposed to field-derived Ae. aegypti mosquitoes on two randomly allocated days during their first 4 d in the study. The patient enrollment flowchart, with mosquito exposure and engorgement frequencies, is shown in Fig. S1. The final evaluable cohort comprised 208 DENV viremic patients, from which there were 407 independent mosquito exposure events. Experimental exposure to Ae. aegypti was well tolerated; no patient had a severe adverse event, and only one patient had a moderate, transient local hypersensitivity response that led the study physician to withdraw them from the study. All other hypersensitivity responses at the site of mosquito exposure were clinically unremarkable. The baseline characteristics of the patients at the time of enrollment are shown in Table 1. DENV-1 and DENV-2 were responsible for 38% and 40% of the viremic dengue cases, respectively. Phylogenetic analyses of envelope (E) gene sequences from these viruses identified the DENV-1 as genotype 1 viruses and the DENV-2 as predominantly from the Asian 1 lineage, with a small minority of Cosmopolitan genotype viruses (Fig. S2). All blood-fed mosquitoes were incubated for 12 d in conditions reflecting the mean temperature and relative humidity inside residential houses in Ho Chi Minh City (HCMC) during the rainy, high DENV transmission season (Fig. S3).
Table 1.
Characteristics of the 208 dengue cases that were exposed to Ae. aegypti on 407 independent occasions
| Characteristic | DENV-1 (n = 78) | DENV-2 (n = 83) | DENV-3 (n = 25) | DENV-4 (n = 22) | All patients (n = 208) |
| Age, y | 21 (19–27) | 23 (20–30) | 24 (19–29) | 23 (20–31) | 23 (20–28) |
| Sex | |||||
| Female | 48 (62) | 48 (58) | 13 (52) | 11 (50) | 120 (58) |
| Male | 30 (38) | 35 (42) | 12 (48) | 11 (50) | 88 (42) |
| Day of illness at enrollment | |||||
| 1 | 0 | 0 | 1 (4) | 0 | 1 (<1) |
| 2 | 18 (23) | 21 (25) | 7 (28) | 3 (14) | 49 (24) |
| 3 | 48 (62) | 53 (64) | 8 (32) | 10 (45) | 119 (57) |
| 4 | 12 (15) | 9 (11) | 9 (36) | 9 (41) | 39 (19) |
| Serology | |||||
| Primary | 25 (32) | 11 (13) | 4 (16) | 2 (9) | 42 (20) |
| Secondary | 42 (54) | 56 (67) | 13 (52) | 19 (86) | 130 (63) |
| Indeterminate | 11 (14) | 16 (19) | 8 (32) | 1 (5) | 36 (17) |
| Enrolment viremia, log10-copies/mL | 8.35 (7.44–9.02) | 7.77 (6.85–8.61) | 8.55 (7.57–9.21) | 6.58 (6.17–7.84) | 7.99 (6.93–8.78) |
| Clinical classification | |||||
| Dengue | 63 (80) | 58 (70) | 18 (72) | 15 (68) | 154 (74) |
| Dengue with warning signs | 15 (20) | 24 (29) | 7 (28) | 7 (32) | 53 (25) |
| Severe dengue | 0 | 1 (1) | 0 | 0 | 1 (<1) |
| Transferred to ICU | 1 (1)(a) | 2 (2)(a),(b) | 0 | 0 | 3 (1) |
Clinical reasons for transfer to intensive care unit (ICU): (a) narrow pulse pressure; (b) hypovolemic shock. Values are n (%) or median (interquartile range).
The likelihood of successful human to A. aegypti transmission of DENV, detected by RT-PCR of individual mosquito abdomens, was coincident with the kinetics of plasma viremia and overall corresponded to the patient’s febrile period, with few transmission events occurring after the clinically observed day of defervescence (Fig. 1). Using day of illness (DOI) as a reference, human to mosquito transmission seldom occurred beyond 5 d after illness onset (Fig. S4). The 50% mosquito infectious dose (MID50) for each serotype was estimated by comparison of the DENV plasma viremia level (copies/mL) at the time of each exposure with the proportion of DENV-infected A. aegypti in the corresponding cohort of blood-fed mosquitoes (Fig. 2). The estimated log10 MID50 [±95% confidence interval (CI)] was similar for DENV-1 (6.51 ± 0.33 log10-copies/mL) and DENV-2 (6.29 ± 0.23 log10-copies/mL) and ∼10 times higher for DENV-3 (7.52 ± 0.59 log10-copies/mL) and DENV-4 (7.49 ± 0.63 log10-copies/mL). Pairwise comparisons of whether the log10-patient-viremia–transmission associations differed by serotype indicated that DENV-1 and DENV-2 (P = 0.16), as well as DENV-3 and DENV-4 (P = 0.93), showed similar associations but that DENV-1 and DENV-2 show highly significantly different associations from DENV3 and DENV4 (all pairwise P < 0.0001) (Fig. 2). These data define the duration of human infectiousness from illness onset and the 50% infectious dose for Ae. aegypti for specific genotypes of DENV-1–4 via the natural route of infection.
Fig. 1.
Profile of DENV transmission from 208 hospitalized dengue cases to Ae. aegypti mosquitoes by fever day and virus serotype. (Left) Patients’ longitudinal plasma viremia profiles according to fever day (gray lines) with a LOESS scatterplot smoother (blue lines) and stratified by virus serotype. The day of defervescence is assigned fever day = 0 for all patients. (Right) Scatterplots of the proportion of mosquitoes with DENV-infected abdomens 12 d after blood-feeding by fever day at the time of mosquito exposure. The graphs also show the estimated associations (blue lines) with pointwise 95% confidence intervals (gray-shaded areas) based on marginal logistic regression.
Fig. 2.
Dose–response scatterplots of plasma viremia vs. the proportion of mosquitoes with DENV-infected abdomens after feeding on 208 hospitalized dengue patients. Each dot point shows the proportion of all mosquitoes that took a blood meal during an exposure event and had DENV-infected abdomens 12 d later vs. the corresponding plasma viremia in the patient at the time of mosquito exposure, stratified by serotype. The graphs show the estimated associations (blue lines) with pointwise 95% confidence intervals (gray shaded areas) based on marginal logistic regression. Serotype-specific associations between the odds of abdomen infection and viremia were DENV1: OR 3.93 (95% CI 2.89–5.33; P < 0.001), DENV2: OR 5.48 (95% CI 3.81–7.90; P < 0.001), DENV3: OR 2.61 (95% CI 1.48–4.58; P < 0.001), and DENV4: OR 2.30 (95% CI 1.65–3.21; P < 0.001) per +1 log10-copies/mL increase in viremia.
Human Virological and Immunological Variables Associated with Successful Human to Ae. aegypti Transmission.
In univariate analyses (Table S1), increases in the patient’s tympanic temperature or plasma viremia level were associated with a higher probability of DENV transmission. Other summary measures of viremia, such as the average slope of the viremia curve and the area under the curve, were not associated with DENV transmission. Conversely, increases in DOI, IgM, and IgG titers to recombinant E protein ectodomain or recombinant E domain III were associated with reduced likelihood of transmission. In multivariable regression analysis only data from DENV-1 and DENV-2 infections were considered (161 patients, 313 exposure events); DENV-3 and -4 were excluded because they had much higher MID50 values and there were fewer cases infected with these virus types. A higher viremia level at the time of exposure was independently associated with a greater likelihood of DENV-1 or DENV-2 transmission (Table 2). Conversely, higher DENV-reactive IgM levels, or higher IgG titers to serotype-matched recombinant E domain III, were independently associated with a reduced likelihood of transmission (Table 2). Each additional DOI was also independently associated with a reduction in transmission risk (Table 2), suggesting factors other than those measured here also contributed to reduced risk of DENV transmission.
Table 2.
Covariates and their association with successful human-to-mosquito abdomen infection by DENV-1 or DENV-2 in multivariable regression analysis
| Variable | OR (95% CI) | P |
| Day of illness (+1 d) | 0.60 (0.46–0.80) | 0.0005 |
| Viremia (+1 log10-copies/mL) | 2.97 (2.29–3.86) | <0.0001 |
| Serotype DENV-2 (vs. DENV-1) | 0.80 (0.44–1.44) | 0.46 |
| Serology | ||
| Secondary (vs. primary) | 0.94 (0.52–1.70) | 0.83 |
| Indeterminate (vs. primary) | 1.86 (0.78–4.41) | 0.16 |
| Antirecombinant DIII IgG titer (twofold increase) | 0.88 (0.78–0.99) | 0.03 |
| IgM units (twofold increase) | 0.66 (0.53–0.82) | 0.0002 |
Early Plasma Viremia Levels Are Associated with the Duration of Infectiousness to Ae. aegypti.
Because the plasma viremia level at the time of mosquito exposure was an important determinant of DENV transmission (Table 2), the extent to which the patient’s first recorded viremia level determined the likelihood of transmission at a later time point was examined. Of the 407 exposure events, 276 exposures took place ≥1 d after the enrollment day. After adjusting for serotype, DOI at enrollment and DOI at exposure, each +1 log10-copies/mL higher plasma viremia level at the time of enrollment was significantly associated with an overall greater probability of human to mosquito transmission on a subsequent day; overall odds ratio (OR) 1.94 (95% CI 1.50–2.51; P < 0.0001). This effect diminished with time since enrollment; for exposures 1 d after enrollment only, OR 2.45 (95% CI 1.73–3.48; P < 0.0001), for exposures 2 d after enrollment only, OR 1.82 (95% CI 1.34–2.48; P = 0.0001), and for exposures 3 d after enrollment only, OR 0.94 (95% CI 0.39–2.25; P = 0.88). These data provide evidence that dengue cases with high early DENV viremia levels have relatively longer durations of infectiousness to Ae. aegypti.
DENV Plasma Viremia Levels and the Phenotype of DENV-Infected Ae. aegypti.
The outcome for Ae. aegypti that imbibe blood meals with high DENV viremia levels was examined in further detail. First, a significant positive correlation was observed between plasma viremia at the time of Ae. aegypti feeding and the concentration of DENV RNA in infected mosquito abdomen tissues measured 12 d later (Fig. S5), indicating that highly viremic blood meals led to more severe infections of the mosquito body. To explore whether highly viremic blood meals also influenced the prevalence of mosquitoes having infectious saliva, a laboratory assay was refined, and an additional 72 dengue cases were prospectively enrolled and exposed to Ae. aegypti on 113 occasions. The baseline features of this patient cohort are described in Table S2. A significant positive association (OR 2.02; 95% CI 1.53–2.67; P < 0.0001) was observed between the DENV plasma viremia level at the time of mosquito exposure and the proportion of blood-fed Ae. aegypti that had saliva containing infectious virus 14 d after blood-feeding (Fig. 3); there was no clear evidence that the strength of this association differed by serotype (interaction test: P = 0.13). Collectively, these data underscore the importance of high DENV plasma viremia levels; they serve as a marker of the duration of infectiousness and also the likelihood of a blood-fed mosquito becoming infectious within 2 wk of blood-feeding.
Fig. 3.
Dose–response scatterplots of plasma viremia vs. the proportion of mosquitoes with infectious saliva after feeding on 72 hospitalized dengue patients. Each dot point shows the proportion of all mosquitoes that took a blood meal during an exposure event and had saliva containing infectious DENV 14 d after blood-feeding vs. the corresponding patient’s plasma viremia at the time of exposure, stratified by serotype. Each dot represents one mosquito exposure, and the size of the dot is proportional to the number of mosquitoes assessed (ranging from 2 to 17). Serotype-specific associations between the odds of saliva positivity and viremia were DENV1: OR 1.47 (95% CI 0.96–2.711; P = 0.08), DENV2: OR 2.53 (95% CI 1.75–3.65; P < 0.0001), DENV3: OR 1.69 (95% CI 1.09–2.60; P = 0.02) and DENV4: OR 2.77 (95% CI 1.68–4.55; P < 0.0001) per +1 log10 copies/mL increase in viremia.
Viremia Characteristics in Hospitalized vs. Ambulatory Dengue Cases.
The preceding viremia–transmission data were from 208 hospitalized patients, and it was uncertain whether ambulatory dengue cases, a major clinical subgroup, would have similar viremia profiles. Plasma viremia levels were therefore measured in a cohort of 262 dengue cases consecutively enrolled at an outpatient clinic at the Hospital for Tropical Diseases (HTD) (cohort details in Methods), of which 189 cases were managed according to the prevailing standard of care as ambulatory patients. Absolute viremia levels at the time of enrollment and the proportion of patients with viremia levels exceeding the MID50 were significantly lower in ambulatory cases with DENV-1, -2, or -3 infections compared with hospitalized cases (Table 3). Nevertheless, among ambulatory patients 79% of DENV-1, 75% of DENV-2, and 32% of DENV-4 cases had viremia levels at the time of enrollment that exceeded the relevant MID50 values (Table 3). The duration of viremia was similar between ambulatory and hospitalized dengue cases (Fig. S6). Thus, a substantial proportion of symptomatic ambulatory dengue cases have viremia characteristics that strongly suggest they are infectious to Ae. aegypti for 1–3 d after they first seek health care.
Table 3.
Summary of enrollment viremia values of hospitalized study dengue patients (n = 208) and community-based ambulatory dengue patients (n = 189)
| Virology | Hospitalized dengue patients (n = 208) | Ambulatory dengue patients (n = 189) | Comparison (P value)* |
| Serotype: DENV-1 | |||
| Viremia, median (IQR) log10-copies/mL | 8.35 (7.44–9.02) | 7.69 (6.69–8.51) | 0.01 |
| Viremia > serotype MID50, x/n (%) | 67/78 (86) | 41/52 (79) | |
| Serotype: DENV-2 | |||
| Viremia, median (IQR) log10-copies/mL | 7.77 (6.85–8.61) | 7.25 (6.41–8.11) | 0.01 |
| Viremia > serotype MID50, x/n (%) | 74/83 (89) | 39/52 (75) | |
| Serotype: DENV-3 | |||
| Viremia, median (IQR) log10-copies/mL | 8.55 (7.57–9.21) | 6.03 (5.24–6.73) | <0.0001 |
| Viremia > serotype MID50, x/n (%) | 19/25 (76) | 2/23 (9) | |
| Serotype: DENV-4 | |||
| Viremia, median (IQR) log10-copies/mL | 6.58 (6.17–7.84) | 7.05 (5.88–7.58) | 0.07 |
| Viremia > serotype MID50, x/n (%) | 8/22 (36) | 20/62 (32) |
Comparison of log10-viremia between hospitalized and community patients based on linear regression with adjustment for day of illness at enrollment. IQR, interquartile range.
Discussion
Human-to-mosquito transmission of arboviruses is a critical but poorly understood step in their epidemiological cycle. Here, the MID50 was defined with high precision for DENV-1 and -2 and, to a lesser extent, for DENV-3 and -4. Strikingly, there was twofold importance to patients with high early plasma viremia levels: such cases were infectious for longer, and Ae. aegypti that fed at times corresponding to high plasma viremia levels were more likely to have infectious saliva 14 d later. These findings were generalizable to ambulatory dengue patients being managed at the primary health care level given the similarities in viremia characteristics.
The MID50 for DENV-1 and DENV-2 was lower than for DENV-3 and DENV-4. This may reflect relatively lower infectiousness of DENV-3 and DENV-4 for Ae. aegypti mosquitoes in this setting and/or a higher ratio of noninfectious to infectious virus particles in plasma. Additionally, a much higher proportion of ambulatory patients with DENV-1 or DENV-2 infections had viremia levels that exceeded their respective MID50 values when they first presented for health care (Table 3). These characteristics suggest that in the confines of the study setting and time period DENV-1 or -2 should be more prevalent in the community in comparison with DENV-3 or -4. Accordingly, virus surveillance in southern Vietnam indicates DENV-1 and -2 have accounted for a large majority of the dengue case burden over the last decade (19), although we recognize that multiple other factors determine the prevalence of DENV serotypes in endemic settings.
The results of this work address knowledge gaps in the understanding of DENV transmission dynamics and thus support mathematical models of dengue epidemiology that can themselves inform how to optimally use interventions such as dengue vaccines (20). The serotype-specific MID50 viremia estimates defined in this study are summary measures, and other factors such as day of illness and serological response independently influenced the probability of DENV transmission. Further studies are needed to dissect out the relative roles and specificities of the anti-DENV IgM and IgG antibody populations that appear during acute infection and that presumably play a pivotal role in reducing the infectiousness of viremic humans to mosquitoes. Nonetheless, the dose (viremia) response (mosquito infection) curves described here provide a helpful reference point for vaccines and antiviral drug interventions. In clinical trials of therapeutic antiviral dengue drugs (21), the MID50 or similar estimates (e.g., MID10) could be used as trial endpoints on the basis that they represent a threshold that a drug must rapidly accelerate viremia below to attenuate the duration of infectiousness. Similarly, prophylactic antiviral drugs or vaccines must prevent natural infection from attaining viremia levels that are infectious to Ae. aegypti, and the data provided here help define these viremia thresholds.
The probability of a DENV viremic human contributing to transmission depends on whether, and for how long, the individual is infectious to Ae. aegypti and the resulting EIP in those blood-fed mosquitoes. Our findings suggest that dengue cases with high early viremia levels have a longer window of infectiousness relative to patients with lower viremia levels and thus may “over-contribute” to the population of DENV-infected mosquitoes in a given setting. Additionally, we observed a positive relationship between the magnitude of plasma viremia at the time of mosquito blood-feeding and the prevalence of mosquitoes with infectious saliva 14 d later. This dose-dependency between plasma DENV RNA concentrations and the prevalence of infectious saliva 14 d after blood-feeding were consistent with the finding that DENV RNA concentrations in Ae. aegypti abdomen tissues were positively correlated with the patient’s plasma viremia level. Taken together, these data suggest that when Ae. aegypti acquires a blood meal from a highly viremic human host that there is relatively faster accumulation and dissemination of DENV in the mosquito body, and critically, a higher probability that infectious virus will be present in saliva 14 d after blood feeding. The importance of viremic blood meals that result in mosquitoes becoming infectious within 14 d is underscored by the ∼3- to 4-wk median lifespan of Ae. aegypti (22) and the inevitability that a proportion of mosquitoes that feed on a DENV viremic human host at a random time in the insect’s lifespan will die before becoming infectious.
In the setting of this study, a majority of symptomatic ambulatory dengue patients presenting to an outpatient clinic with 3 d or less of illness history had DENV viremia levels above the MID50. These cases remain in the community for the duration of their illness and are not captured by public health dengue case surveillance systems (that typically count hospitalized cases only) but are clearly potential sources of onward DENV transmission. Notwithstanding the strong evidence that dengue cases are infectious to mosquitoes before the onset of symptoms (10, 11, 13, 14), public health efforts should include preventing ambulatory index cases from being sources of further DENV transmission.
This study used low-passage, field-derived Ae. aegypti for randomized exposure to a large cohort of well-characterized dengue cases. We investigated transmission dynamics in NS1-positive, hospitalized dengue patients, yet the profiles of viremia in these cases were similar to ambulatory dengue cases being managed through a primary health care facility, suggesting the overall conclusions regarding transmission probably extend to ambulatory patients as well. There are several caveats to our study. Although sampled from the same district of HCMC, the two cohorts of field-derived Ae. aegypti used in this study were unlikely to be genetically identical populations; whether this contributed materially to the experimental variance is unknown. Additionally, the data described here may be specific to the combination of Ae. aegypti from HCMC, Vietnam, and to the viral genotypes that were prevalent during the study period. We did not explore whether ambulatory cases with mild febrile viremic infections, but who never presented to health care, or individuals with asymptomatic DENV infections, can transmit to Ae. aegypti. However, we speculate such cases will be much less likely to transmit DENV because they are likely to have lower and shorter viremias than symptomatic dengue cases. Human movement is a factor in the spatial dissemination of DENV (23), and it is plausible that ambulatory cases with mild febrile viremic infections might be more mobile and have a higher frequency of contact with Ae. aegypti than cases who are clinically more unwell; more research is needed in this area.
The challenge of developing an efficacious dengue vaccine (24, 25) underscores why continued research into DENV infection, transmission, and immunity remains important. Further insights into the transmission dynamics between viremic humans and Ae. aegypti should provide opportunities for a range of interventions aimed at reducing the force of infection in dengue endemic countries.
Methods
Patient Cohorts.
The study site was the HTD in HCMC, Vietnam. Two sequential cohorts of dengue patients were enrolled for investigation of human-to-mosquito transmission. The inclusion criteria were (i) adult patients (≥15 y of age) with ≤72 h of fever in whom there was a clinical suspicion of dengue, (ii) a positive NS1 rapid test result, and (iii) written informed consent. Exclusion criteria were (i) pregnancy (determined by clinical examination or urine dipstick test for β-human chorionic gonadotropin), (ii) patients in intensive care units, (iii) intellectual disability, (iv) a history of severe hypersensitivity reactions to mosquito bites, or (v) severe dermatological conditions. The primary cohort of 213 dengue patients was enrolled between April and December 2011. The follow-up cohort of 72 dengue patients was enrolled between July and October 2012. In both cohorts, serial venous blood samples were collected for 5 consecutive days from the day of enrollment for serological and virological investigations. An independent cohort of predominantly ambulatory dengue cases was enrolled in the International Research Consortium on Dengue Risk Assessment, Management and Surveillance study (ClinicalTrials.gov #NCT01550016) in the outpatients department of HTD between October 2011 and November 2012. The inclusion criteria were (i) presentation to the outpatient clinic with fever or history of fever for ≤72 h, (ii) clinical symptoms consistent with possible dengue (i.e., suspected dengue and/or undifferentiated fever in a patient from a dengue endemic area, (iii) age ≥5 y, and (iv) written informed consent. Exclusion criteria were (i) localizing features suggesting an alternative diagnosis or (ii) judgment by a physician that the patient is unlikely to attend daily follow-up. Case severity was classified according to 2009 World Health Organization classification criteria (26). Demographic and clinical information were recorded prospectively in a standard case record form. All patients provided written informed consent to participate in the study. The study protocols relevant to this work were reviewed and approved by the Scientific and Ethical committee of the HTD (CS/NĐ/09/24, CS/NĐ/12/15, and CS/ND/11/08) and the Oxford Tropical Research Ethical Committee (OxTREC 20-09, OxTREC 29-12, and OxTREC 40-11).
DENV E Gene Sequencing and Phylogenetic Inference.
Multiple overlapping amplimers spanning the DENV E gene were generated by RT-PCR directly from RNA extracted from plasma samples (primers available upon request), and sequencing was performed by standard Sanger sequencing methods. To place these sequences in a global context, E gene sequences representing the global genetic diversity of DENV-1 and DENV-2 were collated from GenBank and manually aligned using Se-AL (http://tree.bio.ed.ac.uk/software/seal/). Publicly available sequences were then subsampled to represent the global diversity of DENV-1 (n = 106) and DENV-2 (n = 173), focusing on the viral genotypes circulating in our study population. Serotype-specific phylogenies were inferred for the subsampled alignments using the maximum likelihood method available in RAxML (27), using the GTRGAMMA model with 500 bootstrap replicates and visualized using FigTree (http://tree.bio.ed.ac.uk/software/figtree/).
Ae. aegypti for Human to Mosquito Transmission Studies.
All of the mosquitoes that were fed on dengue patients were from two independent F3 populations, each derived from pooled larval collections sampled from three discrete locations in District 8 of HCMC. Repeated sampling from the same geographic area was designed to minimize to a level that was practical the genetic differences between each F3 population. Briefly, field-caught Ae. aegypti larvae were pooled and fed commercial dry fish and dog food. Adults (F1 generation) were kept in cages containing males and females in an environmental chamber with 12:12 light:dark hours, 27 °C, and 70% relative humidity. F1 females were provided blood meals by direct feeding on afebrile healthy volunteers for multiple gonotrophic cycles over a period of 45 d. Sucrose (15%) was provided ad libitum. Eggs from F1 females were hatched and reared and the subsequent F2 females provided with human blood meals for multiple gonotrophic cycles as described above. When female F2 mosquitoes were 12 d of age they were killed and pooled into groups of 10 mosquitoes. Each pool was homogenized and tested, along with appropriate controls, by RT-PCR to confirm the absence of DENV, Japanese encephalitis virus, and chikungunya virus. Eggs collected and stored from F2 females were the source of F3 females (3–7 d old) that were used for direct feeding on dengue patients.
Experimental Exposure of Patients to Ae. aegypti Mosquitoes.
After obtaining informed consent, each patient was assigned a schedule of two exposures to Ae. aegypti mosquitoes on two different study days during the first 4 d of enrollment. The schedule of exposures was randomly generated and was blinded to the participant and investigators until the time point that the study number was assigned to the patient. The randomized schedule for exposures was weighted (29% to each of day 1, 2, and 3 and 13% on day 4). At each allotted time point, the patient’s forearm was exposed to 25–40 female 3- to 7-d-old Ae. aegypti mosquitoes contained in a mesh-covered 350 mL plastic cup that was held against the patient’s forearm for 5 min. Mosquitoes were then cold anesthetized at 4 °C for 45 s and engorged mosquitoes transferred to 500 mL plastic cups and maintained in an environmental chamber with 12:12 light:dark hours, 27 °C, and 70% relative humidity for 12 d (cohort 1) or 14 d (cohort 2); the incubation time was determined by available human resources. The number of dead mosquitoes was recorded daily.
Clinical Adverse Events.
Postexposure severe adverse events were defined as any event that was clinically significant (i.e., requiring a clinical intervention, prolonged hospitalization, or admission to intensive care unit) that was possibly, probably, or definitely related to experimental exposure to Ae. aegypti mosquitoes.
Detection of DENV in Mosquito Tissues and Saliva.
All laboratory assays of insects were performed by technicians blinded to the clinical and virological details of the patients. Mosquitoes were killed by cold exposure. The abdomen was dissected from the rest of the mosquito body and suspended in 0.5 mL of mosquito diluent [2% (v/v) heat-inactivated FCS in RPMI1640, antibiotics, and antimycotics]. Individual mosquito abdomens were homogenized with 1 mm zirconia/silica beads for 15 min at 30 Hz using a TissueLyser II (Qiagen). Mosquitoes were scored as being DENV-infected using a quantitative, internally controlled RT-PCR assay (28) on homogenized tissue and the results expressed as copies per tissue.
Infectious virus in the saliva of individual mosquitoes was detected by placing the proboscis of dewinged and delegged mosquitoes into the end of the micropipette tip containing 6 μL of filtered saliva medium (a 1:1 solution of 15% (v/v) sucrose in normal saline and inactivated FCS) for 30 min at room temperature. After 30 min the 6 μL of saliva medium was ejected and then drawn into a pointed glass capillary tube (tip diameter <0.3 μm). The volume of saliva medium derived from one mosquito was then injected into the thorax of four to six Ae. aegypti mosquitoes (4–7 d old, ∼1 μL injected per mosquito) and the injected mosquitoes maintained for 7 d in an environmental chamber with 12:12 light:dark hours, 27 °C, and 70% relative humidity. After 7 d the injected mosquitoes for each saliva sample were killed and the bodies pooled, homogenized, and tested by quantitative RT-PCR for DENV infection, with saliva samples scored as positive or negative depending on this result.
Virological and Immunological Investigations.
A venous blood sample was drawn within 30 min of every mosquito exposure. Measurements of DENV-reactive IgM in these plasma specimens were performed using an IgM antibody capture ELISA (Panbio) according to the manufacturer’s instructions. Measurements of IgG antibodies to DENV recombinant E (recE) proteins (Hawaii Biotech) or recE DIII (gift from Jane Cardosa Universiti Malaysia, Sarawak, Malaysia) in plasma specimens matched to the time points of mosquito exposure were performed by indirect ELISA. End point titers were defined as the reciprocal of the dilution that yielded an optical density of 0.3 after subtracting the absorbance obtained for the uncoated wells.
Dengue Diagnostics.
Classification of primary and secondary dengue was based on serology results (IgM and IgG antibody capture ELISAs; Panbio) on study day 5, with IgM/IgG ratios >1.8 interpreted as primary infection and <1.2 interpreted as secondary infection. IgM/IgG ratios <1.8 but >1.2 were called indeterminate. DENV plasma viremia levels were measured by a validated, quantitative RT-PCR assay that has been described previously (28) and used in recent clinical trials (29, 30). In the validation process the RT-PCR assay was calibrated against infectious virus (grown in insect cells), and the ratio between genome copies per mL and plaque-forming units per mL was 214:1 for DENV-1, 73:1 for DENV-2, 436:1 for DENV-3, and 101:1 for DENV-4.
Statistical Analysis.
The proportion of infected mosquitoes at each exposure was modeled as depending on patient- and exposure-specific covariates based on marginal logistic regression models. Robust SEs were used throughout to account for potential within-patient correlation due to multiple exposures as well as overdispersion. MID50s (plasma viremia levels corresponding to a 50% probability of human-mosquito transmission) were derived on the basis of the logistic regression coefficients, and corresponding 95% confidence intervals were calculated on the basis of the delta rule.
Supplementary Material
Acknowledgments
We thank the patients and their families for their participation in this study; the staff of the Hospital for Tropical Diseases; and Huynh Le Phuong Huy, Huynh Le Phuong Thuy, Nguyen Thi Giang, Vu Tuyet Nhu, and Vu Ty Hang for providing outstanding technical assistance. This work was supported by a Wellcome Trust Senior Fellowship award (Grant 084368/Z/07/Z to C.P.S.).
Footnotes
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1303395110/-/DCSupplemental.
References
- 1.Simmons CP, Farrar JJ, Nguyen V, Wills B. Dengue. N Engl J Med. 2012;366(15):1423–1432. doi: 10.1056/NEJMra1110265. [DOI] [PubMed] [Google Scholar]
- 2.Brady OJ, et al. Refining the global spatial limits of dengue virus transmission by evidence-based consensus. PLoS Negl Trop Dis. 2012;6(8):e1760. doi: 10.1371/journal.pntd.0001760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lambrechts L, et al. Impact of daily temperature fluctuations on dengue virus transmission by Aedes aegypti. Proc Natl Acad Sci USA. 2011;108(18):7460–7465. doi: 10.1073/pnas.1101377108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Watts DM, Burke DS, Harrison BA, Whitmire RE, Nisalak A. Effect of temperature on the vector efficiency of Aedes aegypti for dengue 2 virus. Am J Trop Med Hyg. 1987;36(1):143–152. doi: 10.4269/ajtmh.1987.36.143. [DOI] [PubMed] [Google Scholar]
- 5.Lambrechts L, et al. Genetic specificity and potential for local adaptation between dengue viruses and mosquito vectors. BMC Evol Biol. 2009;9:160. doi: 10.1186/1471-2148-9-160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kramer LD, Ebel GD. Dynamics of flavivirus infection in mosquitoes. Adv Virus Res. 2003;60:187–232. doi: 10.1016/s0065-3527(03)60006-0. [DOI] [PubMed] [Google Scholar]
- 7.Smith DL, et al. Ross, Macdonald, and a theory for the dynamics and control of mosquito-transmitted pathogens. PLoS Pathog. 2012;8(4):e1002588. doi: 10.1371/journal.ppat.1002588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Christofferson RC, Mores CN. Estimating the magnitude and direction of altered arbovirus transmission due to viral phenotype. PLoS ONE. 2011;6(1):e16298. doi: 10.1371/journal.pone.0016298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chan M, Johansson MA. The incubation periods of dengue viruses. PLoS ONE. 2012;7(11):e50972. doi: 10.1371/journal.pone.0050972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Siler J. Transmission of dengue fever by mosquitoes. Proc Soc Exp Biol Med. 1925;23:197–201. [Google Scholar]
- 11.Siler J, Hall M, Hitchens A. Dengue: Its history, epidemiology, mechanism of transmission, etiology, clinical manifestations, immunity and prevention. Philipp J Sci. 1926;29:1–302. [Google Scholar]
- 12.Halstead SB. Etiologies of the experimental dengues of Siler and Simmons. Am J Trop Med Hyg. 1974;23(5):974–982. doi: 10.4269/ajtmh.1974.23.974. [DOI] [PubMed] [Google Scholar]
- 13.Simmons J, St. John J, Reynolds F. Experimental studies of dengue. Philipp J Sci. 1931;44:1–247. [Google Scholar]
- 14.Nishiura H, Halstead SB. Natural history of dengue virus (DENV)-1 and DENV-4 infections: Reanalysis of classic studies. J Infect Dis. 2007;195(7):1007–1013. doi: 10.1086/511825. [DOI] [PubMed] [Google Scholar]
- 15.Gubler DJ, Reed D, Rosen L, Hitchcock JR., Jr Epidemiologic, clinical, and virologic observations on dengue in the Kingdom of Tonga. Am J Trop Med Hyg. 1978;27(3):581–589. doi: 10.4269/ajtmh.1978.27.581. [DOI] [PubMed] [Google Scholar]
- 16.Gubler DJ, Suharyono W, Lubis I, Eram S, Sulianti Saroso J. Epidemic dengue hemorrhagic fever in rural Indonesia. I. Virological and epidemiological studies. Am J Trop Med Hyg. 1979;28(4):701–710. [PubMed] [Google Scholar]
- 17.Gubler DJ, Suharyono W, Lubis I, Eram S, Gunarso S. Epidemic dengue 3 in central Java, associated with low viremia in man. Am J Trop Med Hyg. 1981;30(5):1094–1099. doi: 10.4269/ajtmh.1981.30.1094. [DOI] [PubMed] [Google Scholar]
- 18.Gubler DJ, Suharyono W, Tan R, Abidin M, Sie A. Viraemia in patients with naturally acquired dengue infection. Bull World Health Organ. 1981;59(4):623–630. [PMC free article] [PubMed] [Google Scholar]
- 19.Vu TT, et al. Emergence of the Asian 1 genotype of dengue virus serotype 2 in viet nam: In vivo fitness advantage and lineage replacement in South-East Asia. PLoS Negl Trop Dis. 2010;4(7):e757. doi: 10.1371/journal.pntd.0000757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Beatty M, et al. WHO-VMI Dengue Vaccine Modeling Group Assessing the potential of a candidate dengue vaccine with mathematical modeling. PLoS Negl Trop Dis. 2012;6(3):e1450. doi: 10.1371/journal.pntd.0001450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Simmons CP, et al. Therapeutics for dengue: recommendations for design and conduct of early-phase clinical trials. PLoS Negl Trop Dis. 2012;6(9):e1752. doi: 10.1371/journal.pntd.0001752. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Harrington LC, et al. Age-dependent survival of the dengue vector Aedes aegypti (Diptera: Culicidae) demonstrated by simultaneous release-recapture of different age cohorts. J Med Entomol. 2008;45(2):307–313. doi: 10.1603/0022-2585(2008)45[307:asotdv]2.0.co;2. [DOI] [PubMed] [Google Scholar]
- 23.Stoddard ST, et al. House-to-house human movement drives dengue virus transmission. Proc Natl Acad Sci USA. 2013;110(3):994–999. doi: 10.1073/pnas.1213349110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sabchareon A, et al. Protective efficacy of the recombinant, live-attenuated, CYD tetravalent dengue vaccine in Thai schoolchildren: A randomised, controlled phase 2b trial. Lancet. 2012;380(9853):1559–1567. doi: 10.1016/S0140-6736(12)61428-7. [DOI] [PubMed] [Google Scholar]
- 25.Halstead SB. Dengue vaccine development: A 75% solution? Lancet. 2012;380(9853):1535–1536. doi: 10.1016/S0140-6736(12)61510-4. [DOI] [PubMed] [Google Scholar]
- 26.Wold Health Organization . Dengue: Guidelines for Treatment, Prevention and Control. Geneva: WHO; 2009. p. 146. [PubMed] [Google Scholar]
- 27.Stamatakis A, Ludwig T, Meier H. RAxML-III: A fast program for maximum likelihood-based inference of large phylogenetic trees. Bioinformatics. 2005;21(4):456–463. doi: 10.1093/bioinformatics/bti191. [DOI] [PubMed] [Google Scholar]
- 28.Hue KD, et al. Validation of an internally controlled one-step real-time multiplex RT-PCR assay for the detection and quantitation of dengue virus RNA in plasma. J Virol Methods. 2011;177(2):168–173. doi: 10.1016/j.jviromet.2011.08.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tam DT, et al. Effects of short-course oral corticosteroid therapy in early dengue infection in Vietnamese patients: A randomized, placebo-controlled trial. Clin Infect Dis. 2012;55(9):1216–1224. doi: 10.1093/cid/cis655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Nguyen NM, et al. A randomized, double-blind placebo controlled trial of balapiravir, a polymerase inhibitor, in adult dengue patients. J Infect Dis. 2013;207(9):1442–1450. doi: 10.1093/infdis/jis470. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.