Laboratory Evaluation of the Rapid Analyte Measurement Platform Assay to Detect Dengue Virus in Mosquito Pools

In terms of human disease burden, dengue virus (DENV) is the most significant arboviral pathogen in the world (Bhatt et al. 2013). Any one of the 4 DENV serotypes can cause dengue fever, the symptoms of which include fever, myalgia, conjunctivitis, and joint pain, and in the most serious cases, DENV can cause severe dengue, a potentially fatal syndrome. The geographical spread of DENV's 2 primary vectors, Aedes aegypti (L.) and Ae. albopictus (Skuse), as well as the movement of people via urbanization and increased international travel have contributed to the global resurgence of DENV disease (Gubler 2011). Dengue virus is now thought to be present in more than 125 countries on 6 continents (Brady et al. 2012, Stanaway et al. 2016) with outbreaks occurring at greater frequencies and magnitudes across the globe (Murray et al. 2013). Concern that DENV disease may be introduced into the USA by infected travelers returning from DENV-endemic regions has been reinforced by recent local transmission events in Texas, Hawaii, the US Virgin Islands, and Florida (Effler et al. 2005, Clark 2008, Mohammed et al. 2010, Radke et al. 2012, Adams et al. 2019, Hasty et al. 2020).

While a vaccine of controversial efficacy is available in some countries, prevention and control strategies for DENV consist primarily of disease surveillance and vector control. It has been recommended that the testing of mosquito pools be included as a component of DENV surveillance to monitor active transmission in DENV-endemic areas (Eisen et al. 2009). In addition, mosquito testing has been utilized in DENV-endemic areas to incriminate vectors, assess and identify circulating or newly introduced DENV strains in the vector population, and determine precise locations where people are becoming infected (Chung and Pang 2002, Urdaneta et al. 2005, Yoon et al. 2012, Khan et al. 2016). Studies have reported the detection of positive mosquitoes as early as 6–8 wk prior to the onset of outbreaks, suggesting that mosquito-based DENV surveillance may serve as a predictive tool for human disease outside of the USA (Chow et al. 1998, Lourenco-de-Oliveira et al. 2002, Urdaneta et al. 2005).

Much of the mosquito pool testing is done via viral RNA detection using standard or real-time reverse transcriptase polymerase chain reaction (RT-PCR) assays (Chien et al. 2006, Gurukumar et al. 2009, Chen et al. 2010, Graham et al. 2011, Rohani et al. 2014, Cecílio et al. 2015, Dzul-Manzanilla et al. 2016, Khan et al. 2016). Some agencies have opted to incorporate antigen detection assays such as immunofluorescence assays (IFAs) or enzyme immunoassays (EIAs) in their testing programs (Angel and Joshi 2009, Vikram et al. 2015). Still others have used an antigen-detection assay to initially screen for DENV and an RT-PCR assay to characterize serotypes of presumptive positive pools (Dutta et al. 2015). While RT-PCR assays are sensitive and reliable, they require dedicated laboratory space, specially trained personnel, and equipment and reagents that are not commonly available in vector control agencies. Antigen detection assays such as IFAs or EIAs are often cheaper and easier to perform but are also carried out in specialized laboratory facilities by trained laboratorians. Often the laboratories where these assays are performed are in centralized locations, which may prove difficult for agencies in rural dengue-endemic areas to access. In addition, the lag time between collecting mosquitoes and receiving laboratory results can reduce the usefulness of mosquito-based surveillance.

For agencies that do not have access to fully equipped laboratories, simple assays are needed. In response to the need for fast, reliable, and technologically simple means of detecting DENV, commercial vendors have developed antigen detection assays in lateral-flow or dipstick formats that have proven suitable for mosquito pool testing (Tan et al. 2011, Muller et al. 2012, Voge et al. 2013, Wanja et al. 2014, Chao et al. 2015). These easy-to-use assays incorporate DENV-specific antibodies that bind to target antigens if present in the sample, producing a color change that is visible to the naked eye. By incorporating commercially available antigen detection assays into their surveillance programs, more agencies could test their mosquitoes with minimal laboratory equipment. These assays also capitalize on the relative stability of antigen, which is less susceptible to environmental inhibitors such as enzymes and to variable handling conditions in the field that may degrade viral RNA and infectious virus.

Recently, the manufacturer of the Rapid Analyte Measurement Platform (RAMP®) West Nile virus assay (Response Biomedical Corp., Burnaby, British Columbia, Canada) expanded their environmental product line to include a DENV RAMP antigen detection assay for the qualitative detection of all 4 serotypes of DENV in mosquito pools. This assay works similarly to the commercial antigen assays mentioned above in that antigens present in the sample bind to antibodies immobilized on a strip. However, while interpretation of dipstick assays relies on the visual observation of color development, the RAMP assay incorporates fluorescently labeled antibody complexes that are read by an instrument (RAMP Reader), thereby eliminating subjective interpretation of results. We evaluated the RAMP DENV assay for sensitivity against a panel of geographically distinct strains of each DENV serotype, and for specificity against nontarget arboviruses to ensure that it does not produce false positives.

Samples used to evaluate the sensitivity and specificity of the RAMP DENV assay were tested according to the manufacturer's instructions. All samples were homogenized in 1-ml RAMP buffer for 3 min and centrifuged for 3 min at 1700 × g. An aliquot of clarified supernatant was mixed with a fluorescently labeled DENV antibody conjugate supplied in the kit and applied to the test strip, which is housed in a cartridge. During the recommended 2-h incubation period at room temperature (RT), the mixture migrated along the strip, whereas antigen-bound particles were immobilized at the detection zone. Additional control particles continued to migrate and were immobilized at an internal control zone. After incubation, the cartridge was inserted into the RAMP reader, which measured the amount of fluorescence emitted by the particles bound at the detection and control zones and displayed the relative difference between the 2 zones as RAMP units. The cartridges were also retained for an additional ∼24 h at RT and read the following day to determine the effect of a longer incubation. Samples that returned results of ≥ 30 RAMP units were considered positive, according to the assay protocol. Samples that returned results of < 30 were considered negative (the lowest reading provided by the RAMP reader is “< 10”). A maximum pool size of 25 mosquitoes was used as recommended by the manufacturer.

The RAMP assay sensitivity limit was initially determined using representative DENV strains of each serotype. DENV-1 16007, DENV-2 New Guinea C, DENV-3 H-87, and DENV-4 1650 were diluted in RAMP buffer to create a range of virus titers and used to spike pools of 25 uninfected laboratory-reared Puerto Rico REXD strain Ae. aegypti. The samples were homogenized in RAMP buffer and tested with the RAMP DENV assay as described above (Table 1). Each virus stock was also diluted in bovine albumin-1 (BA-1) and tested by Vero cell plaque assay in 6-well tissue culture plates as described previously (Beaty et al. 1995) to estimate the titer of each spiked mosquito pool sample.

Once the sensitivity limit was determined for each representative serotype, virus stocks of the remaining DENV strains listed in Table 1 were prepared in RAMP buffer to produce titers slightly below and above the limit of detection and used to spike mosquito pools that were tested with the RAMP DENV assay as described above. To conserve supplies, titers well below and above the established limit of detection (< 3.0 and > 4.5 log₁₀ plaque-forming units [PFU]/ml, respectively) were not tested. While the limit of detection varied slightly among the strains (Table 1), overall, samples containing at least 4.3 log₁₀ PFU/ml were detectable by RAMP (produced a result ≥ 30 RAMP units) regardless of serotype or strain after incubating the cartridge for 2 h. Likewise, any sample containing less than 3.8 log₁₀ PFU/ml returned negative results. Incubating the cartridges for 24 h resulted in RAMP results that were slightly higher but did not alter the calculated sensitivity limit.

The 4 representative DENV strains listed above were also used to inoculate Ae. aegypti intrathoracically as described previously (Rosen and Gubler 1974). Mosquitoes were kept for 7–10 days at 28°C and 80–85% RH, after which the legs and head from each inoculated mosquito were removed, placed in a single tube, homogenized in BA-1, and tested using a real-time RT-PCR assay as described by Santiago et al. (2013) to determine infectivity. If the legs and head were positive, the corresponding body was added to a pool of 24 uninfected mosquitoes, homogenized in RAMP buffer, and tested with the RAMP assay as described above. The proportion of pools containing 1 positive mosquito that were detectable by the RAMP DENV assay after a 2-h incubation period were as follows: 83% (4 of 5) of DENV-1, 100% (5 of 5) of DENV-2, 50% (2 of 4) of DENV-3, and 100% (4 of 4) of DENV-4. After the 24-h incubation period, however, the RAMP DENV assay detected 100% of positive pools of all serotypes.

High-titered stock strains of nontarget arboviruses (Table 1) and pools of 25 uninfected Ae. aegypti (n = 10) were homogenized in RAMP buffer and tested with the RAMP DENV assay as described above. The RAMP DENV assay did not produce positive results for any nontarget virus tested, nor did it produce positives for negative control mosquito pools.

DENV is a biosafety level (BSL)-2 agent (U.S. Department of Health and Human Services, 2009). Since some agencies that will implement RAMP testing may not have the appropriate biosafety containment facilities that are required when working with DENV, we tested the ability of the RAMP buffer to inactivate virus using methods described previously (Burkhalter et al. 2016). Briefly, high-titered aliquots of 4 representative strains of each DENV serotype (DENV-1 16007, DENV-2 New Guinea C, DENV-3 H-87, and DENV-4 1650) were added to RAMP buffer and vortexed for 3 sec or 1 min (Table 2). An aliquot of the RAMP buffer samples from each incubation period was then serially diluted in cell culture medium BA-1 and tested for infectious virus using Vero cell culture assays as described above. While the RAMP buffer reduced the titer of all DENV serotypes after both incubation periods, it did not completely inactivate 1 serotype after the 1-min incubation period (Table 2). Therefore, we repeated the inactivation experiment using RAMP buffer that contained a 1% final concentration of Triton X-100 (Sigma-Aldrich, St Louis, MO). With the addition of this detergent, the buffer completely inactivated all 4 serotypes within the 3-sec incubation (Table 2).

The results of this laboratory evaluation indicate that the RAMP DENV assay can detect geographically distinct strains of all 4 DENV serotypes and can detect 1 positive mosquito in a pool of up to 25 mosquitoes. The 3 samples containing infected mosquitoes that produced negative RAMP results after the 2-h incubation period were below the limit of detection (14, 22, and 22 RAMP units) according to the manufacturer's recommended cutoff value of 30 RAMP units. After the 24-h incubation period, the RAMP results were slightly higher for all of the DENV samples, and these initially negative mosquito pool samples produced results just above 30 RAMP units (35, 58, and 49 RAMP units, respectively). All negative controls and nontarget viruses produced results of “< 10” after both of the 2- and 24-h incubation periods. Therefore, we recommend that all sample cartridges be read after 2 h for an initial screening, and any cartridge that produces a result ≥ 30 RAMP units may be considered positive. A cartridge returning a result between 11 and 29 RAMP units should be read after ∼24 h to detect samples that may benefit from a longer incubation and further development of a detectable signal. As this test detects all serotypes of DENV without differentiating them, serotyping would have to be done by other means (i.e., RT-PCR), if deemed necessary to do so. If an agency intends to perform the assay in a non-BSL-2 facility, we recommend processing the pools in RAMP buffer fortified with a 1% final concentration of Triton X-100 to inactivate the virus on contact, as described more fully in Burkhalter et al. (2016).

The authors thank Brandy Russell of the Diagnostic and Reference team, Arbovirus Disease Branch, CDC, Fort Collins, CO, for providing the viruses used in this evaluation.