Identifying the array of vectors that play a role in perpetuating West Nile virus (WNV) infection in endemic foci will help in controlling the disease. Aedes japonicus has the potential to be a vector in the wild of at least 3 kinds of encephalitis, including WNV. Aedes japonicus is a nonnative species in the USA that is temperature tolerant and a potential human biter. Detection of WNV in mosquito pools of this field-collected invasive species, combined with their ability to feed on humans, make this mosquito species a possible public health concern. In this study, we collected mosquito abundance data and tested them for WNV-positive mosquito samples from 3 counties in New York State. We found a significant association between the season and land demography and the likelihood of the virus in Ae. japonicus.
West Nile virus (WNV) is a mosquito-borne zoonotic disease. West Nile virus infections in people have been reported throughout the USA (CDC 2018, 2019a). According to the Centers for Disease Control and Prevention (CDC), 51,747 cases of WNV disease in people have been reported from 1999 to 2019. Of these, 25,264 (66%) were classified as neuroinvasive disease and 26,483 (34%) as nonneuroinvasive disease (CDC 2018, 2019b); this could be due to the underreporting of nonneuroinvasive WNV cases (CDC 2019b, McDonald et al. 2019).
Although Culex pipiens (L.) is identified as the most important vector, WNV has been detected in 66 species of mosquitoes in the USA (Blitvich 2008, CDC/ArboNet 2018). In 1998, Aedes japonicus (Theobald) was identified for the 1st time in North America, specifically in central New Jersey and eastern Long Island, New York (Peyton et al. 1999). This geographic region is the same location where WNV would arrive a year later. Now both WNV and this invasive mosquito species have become established in the USA (Armistead et al. 2008). In the USA, Ae. japonicus has been found in 33 states, including Hawaii, as of 2011 (Kampen and Werner 2014). In addition, Ae. japonicus is a competent vector of WNV (Sardelis and Turell 2001, Molaei et al. 2009). West Nile virus competence in Ae. japonicus has only been proven in the laboratory and has yet to be proven in the field.
In Tompkins County, gravid traps (BioQuip Products, Inc., Rancho Dominguez, CA) were used to capture ovipositing Ae. japonicus from July 10, 2008, through October 10, 2008, at 2 trap-days/wk at 3 sites. The CDC light traps baited with dry ice (CO2) and gravid traps were used to capture host-seeking and ovipositing Ae. japonicus in Nassau County. Traps (n = 16) were set from July 22, 2009, through September 30, 2009, at 3 trap-days/wk at 11 sites. The CDC light traps baited with dry ice, BG-Sentinel™ (BioQuip) traps, and gravid traps were used to capture host-seeking and ovipositing Ae. japonicus in Suffolk County. Traps (n = 16) were set from August 25, 2009, through September 29, 2009, at 2 trap-days/wk in August and 1 trap-day/wk in September at 5 sites for approximately 12 h from dusk to dawn. All gravid traps were baited with water. Batteries were used to facilitate the upward air draft to capture the mosquitoes in the netting.
A chill plate was used to process the mosquitoes (Andreadis et al. 2005). All Nassau and Suffolk County mosquitoes were sorted by their respective laboratory personnel and sent on dry ice to Cornell University. In order to prevent cross contamination, individual mosquitoes were dissected using single-use sterilized disposable no. 10 scalpels (BD Biosciences, Franklin Lakes, NJ) and forceps (BioQuip). Heads and thoraxes were removed by cutting between the 2nd and 3rd leg. Only the heads and thoraxes were tested following the procedure put in place in 2008 in our laboratory at Cornell University.
Mosquitoes were trapped in 3 WNV-endemic areas in New York: Tompkins, Nassau, and Suffolk counties. All the mosquito specimens were tested in pools. In Tompkins County, the mean number of specimens in each pool was 8 Ae. japonicus per trap-day, using the gravid trap. The range was 1–19 mosquitoes. In Nassau County, the mean number of specimens in each pool was 3 Ae. japonicus per trap-day using the CDC light trap and 2 per trap-day using the gravid trap. The range for the CDC light trap was 1–8 mosquitoes and gravid trap range was 1–6 mosquitoes. In Suffolk County, the mean number of specimens in each pool was 11 Ae. japonicus per trap-day using the gravid trap, 23 using the CDC light trap, and 58 using the BG-Sentinel trap. The range was 1–21 mosquitoes for the gravid trap, 2–99 mosquitoes for the CDC light trap, and 1–285 mosquitoes for the BG-Sentinel trap. Tompkins County traps were either in suburban developed areas with permission obtained from the owners or deciduous forest. Long Island locations were either suburban or open-space areas.
Disruption of the mosquito samples was done in sterile Eppendorf 1.5-ml conical tubes (Eppendorf North America, Hauppauge, NY), using sterile plastic pestles (Life Science Products, Inc., Chestertown, MD). Homogenization was accomplished through the use of Qiashredder according to manufacturer's instructions (Qiagen, Valencia, CA). Ribonucleic acid (RNA) isolation was achieved through the utilization of the RNeasy mini kit according to manufacturer's instructions (Qiagen). The RNA integrity was tested by a method of mosquito housekeeping gene detection of 18S rRNA developed in Hawaii (Hoffmann et al. 2004), using a primer pair 18S417/920c (Accession AB085210 GenBank). All samples tested were positive for 18S, ruling out RNA degradation in any of the samples. The testing was done on the same real time–polymerase chain reaction (RT-PCR) plates with their respective samples.
The TaqMan real-time and standard RT-PCR assays were run using previously established procedures with some cycling modifications to adhere to CDC standards of positive, negative, and equivocal results (Lanciotti et al. 2000, Shi et al. 2001). The ABI 7500 Fast RT-PCR system (Applied Biosystems™ by Life Technologies™, Carlsbad, CA) and AgPath-ID™ One-step RT-PCR kit (Ambion® Applied Biosystems by Life Technologies) were used with thermal cycling conditions of 1 cycle at 45°C for 10 min, 1 cycle at 95°C for 10 min, and then 36 cycles of 95°C for 15 sec, and 60°C for 45 sec. Known positive samples were run with all plates in order to show that the target DNA was successfully amplified. Standard RT-PCR: The Bio-Rad® iCycler® (Bio-Rad Laboratories, Hercules, CA) thermal cycling conditions consisted of 1 cycle at 50°C for 30 min, 1 cycle at 95°C for 15 min, and then 50 cycles of 95°C for 45 sec, 62°C for 45 sec, and 70°C for 1 min (Shi et al. 2001). This was the 2nd test of a 2-step process to verify positives.
The minimum infection rate (MIR) was calculated using the formula (number of positive pools/total specimens tested) × 1,000, with the data representing a single species collected over a single time period and from a single geographic area. The tool used for this was PooledInfRate, version 3.0, a Microsoft® Excel™ add-in (Biggerstaff 2006). This program also includes calculation of a 95% confidence interval (CI), which reflects, in part, the sample sizes used in the calculations. The data were stratified by county (see Table 1).
Estimates of the infection rate among mosquitoes are reported as interval estimates. In situations where no virus was detected, the 95% CI for zero events was calculated using the formula P ≈ 3/n (Hanley and Lippman-Hand 1983). The significance of the association between the putative factors (season of collection and geographic locality) was evaluated using the chi-square test in Statistix 9.0 (Analytical Software, Tallahassee, FL). The risk of detecting the virus in mosquitoes associated with a particular factor was quantified using the odds ratio as an approximate estimate of the relative risk.
There was a significant association between the time of year during mosquito season and the likelihood of the virus to be found in the Ae. japonicus. There was a greater likelihood (i.e., 10 times greater) of having a positive mosquito pool in the fall than in the summer (see Table 2, odds ratio = 0.1). Positive pools were also twice as likely to be found in a rural as in a suburban neighborhood. In the study of the invasive species Ae. japonicus on Long Island, none of the mosquito pools tested positive for WNV. The 95% CI for zero events in Nassau County was 1.7 (see Table 1). The 95% CI for zero events in Suffolk County was 0.2 (see Table 1). In Tompkins County, we trapped 193 female Ae. japonicus yielding a total of 24 pools that ranged from 1 to 19 mosquitoes per pool. Five pools tested positive for WNV. The MIR was 25.9 (see Table 1).
In this research, we collected mosquitoes from 3 counties of New York State and tested them to find WNV-positive mosquito pools. We found WNV-positive mosquito pools only in Tompkins County. No positive pools were found in the other 2 counties. A positive sample in our study was an indication of infection within the salivary gland. This denotes escape of the virus from the midgut where the mosquito can sequester the virus and not transmit it by bite. Results will be lower overall positive rates of infection, but positives indicated that the virus had disseminated within the mosquito. Therefore, there could have been the possibility of transmitting the virus by bite. The presence of WNV-positive Ae. japonicus in Tompkins County is a reason to test this species in that area. One confounding factor is that the testing in all 3 counties did not occur in the same year (2008 for Tompkins County versus 2009 for Nassau and Suffolk counties on Long Island) as it is known that mosquito population numbers and WNV activity may vary from year to year. There was low infectivity in 2009 in the New York metropolitan area which includes the 2 Long Island counties in this study (Roehrig 2013). Detection of WNV in pools of field-collected Ae. japonicus, combined with their ability to feed on humans, make these mosquitoes of some public health significance (Andreadis et al. 2001, Turell et al. 2001). Their daytime biting habits make this information worth disseminating to the public so that personal protection methods can be observed (CDC/EPA 2019).
The authors thank the USDA Hatch Act Formula grant number 478-6444-6650-444-000 for funding this research. We also acknowledge Barbara Barnett, Nassau County Department of Health, Hempstead, NY, and staff of the Suffolk County Department of Health Services for helping in mosquito collections.
Department of Population Medicine and Diagnostic Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853.
Suffolk County Department of Health Services, 360 Yaphank Avenue, Suite 2A, Yaphank, NY 11980.