La Crosse virus (LACV) is a leading cause of pediatric arboviral encephalitis in the U.S., particularly affecting children under 16 (Fagre et al. 2023). While LACV disease was initially recognized in Wisconsin, recent cases are concentrated in Ohio, North Carolina, West Virginia, and Tennessee (Day et al. 2023). In the southern Appalachian region, LACV exhibits persistent, localized clusters at county and zip code levels, with household-level risk suggesting the need for residential public health interventions (Byrd et al. 2018, Day et al. 2024).

The primary LACV vector is Aedes triseriatus (Say) with Ae. albopictus (Skuse) and Ae. japonicus (Theobald) as invasive secondary vectors (Westby et al. 2015). Aedes triseriatus is considered the main vector because of consistent findings of infected specimens near human cases, efficient laboratory transmission, transovarial transmission, and feeding behaviors that include humans and amplifying hosts (Pantuwatana et al. 1972, 1974: Watts et al. 1973, 1974). Vertical transmission within Ae. triseriatus populations likely facilitates overwintering and persistence of LACV. Because LACV disease risk is clearly associated with exposure to infected mosquitoes, effective methods to estimate biting rates, particularly for Ae. triseriatus, are necessary to better understand LACV exposure risk.

Host-seeking traps, baited with CO₂ or synthetic human odorants, capture mosquitoes actively seeking hosts and may serve as proxies for human exposure. However, standard traps like the Centers for Disease Control and Prevention (CDC) light trap (underrepresent diurnal container Aedes species. The BG-Sentinel 2 trap, designed to address this, effectively collects Ae. albopictus and Ae. aegypti, but its efficacy for Ae. triseriatus is less known (Maciel-de-Freitas et al. 2006, Camara et al. 2022). The BG-Pro a modified trap combining features of the CDC light trap and BG-Sentinel 2 is known to effectively trap Ae. albopictus (Degener et al. 2021). However, it has not been evaluated in the context of LACV vector surveillance, namely Ae. triseriatus, using a comparative approach in a LACV disease endemic area. This study aims to establish more effective methods for collecting host-seeking Ae. triseriatus, with secondary emphasis on invasive LACV vectors, by evaluating the BG-Sentinel 2 (with and without the BG human lure) and the BG-Pro (with the BG human lure), using the CDC light trap as a reference (Figs. 1a–d).

View Figure

• Download as Powerpoint
• Download Figure

Two 4 × 4 Latin square design studies were conducted in Jackson County, North Carolina, during summer 2023 (June 27–July 1 and August 25–29). Traps were placed in mixed hickory-oak forests near Western Carolina University, at sites with known LACV vector presence. Each site had 4 trap locations, approximately 50 m apart. Four trap types were tested: CDC Light Trap, BG-Pro, BG-Sentinel 2 with human lure, and BG-Sentinel 2 without lure. All traps were CO₂-baited with dry ice. Trap placement was randomized and rotated daily for 4 days. Traps were deployed between 1400–1500 h and collected between 0900–1010 h to align with mosquito activity and minimize loss of CO₂ sublimation.

Collected mosquito samples were frozen at −20°C, then identified and sorted by species, sex, location, date, and trap type (Harrison et al. 2016). Specimens were stored at –20°C until LACV testing. Mosquito pools (groups of ≤50 mosquitoes of the same species, collection site, collection date, and sex) in 2 ml microcentrifuge tubes with stainless steel beads were shipped on dry ice and stored at −80°C until further processing. To process the pools, 1 ml of (DMEM) Dulbecco's Modified Eagle Medium (supplemented with 5% Fetal Bovine Serum, 1% Penicillin-Streptomycin, 1 mg/ml Amphotericin B, and 100 mg/l Gentamicin) was added to each tube. Using a tissue lyser, the samples were homogenized at speed 20 for 3 min and immediately centrifuged at 12,000 RPM for 10 min. To extract ribonucleic acid (RNA) and test pools by quantitative reverse transcription-polymerase chain reaction (qRT-PCR,) 200 µl of each centrifuged sample was extracted using the MagMax Viral/Pathogen Nucleic Acid Isolation kit (Thermo Fisher, Valencia, CA, USA) in accordance with the manufacturer’s instructions. Molecular testing was performed in duplex via qRT-PCR, using the QuantiTect Virus +ROX Vial Kit (Qiagen, Redwood City, CA, USA). Forward, reverse, and probe primers were used in equal concentrations of 4 µM as dictated by the kit protocol. 5 µl of RNA from each sample was used. La Crosse virus primers are available upon request. The following cycling conditions were utilized: 50°C for 20 min, 95°C for 5 min; and 45 cycles of 95°C for 15 sec, 60°C for 45 sec. Virus isolation attempts were performed as follows. Mosquito pools were thawed from –80°C storage and centrifuged at 12,000 RPM for 10 min. 200 µl of each sample was filtered using 0.45 µm Multiscreen HTS) HA filter plates (MilliporeSigma, Burlington, MA, USA) with a Multiscreen HTS 96 format vacuum manifold (MilliporeSigma). Filtered samples were directly applied to 12 well monolayers of Vero cells. Plates were incubated at 37°C for 45 min, rocking every 15 min. Following adsorption, an additional 1 ml of complete DMEM was added to each well. Cells were monitored for cytopathic effects for 14 days. Samples demonstrating cytopathology were harvested, supplemented with 20% FBS, filtered with a 0.45 µm spin column (Corning, Corning, NY, USA), and stored at −80°C until subsequent genomic sequencing.

Data were analyzed using generalized linear mixed models (GLMMs) with a Poisson link function using the lme4 R package (Bates et al. 2015). Zero-inflated data were considered structural zeros (Blasco-Moreno et al. 2019). Odds ratios for each trap type were calculated relative to the CO₂-baited CDC trap. In any instance where the total number of individuals for a species collected in a trap type was equal to 0, we added a pseudocount equal to 1 which allowed for model estimation. Overdispersion was assessed, and negative-binomial GLMMs were used to account for any unequal mean/variance relationship. Non-parametric bootstraps (n = 250) estimated parameter distributions. Males were excluded from analyses as the focus was host-seeking female mosquitoes.

A total of 436 mosquitoes were collected over 128 trap days (142 in June, 294 in August) (Table 1a). Aedes albopictus was the most common LACV vector (n = 182, 41.7%), followed by Ae. triseriatus (n = 156, 35.7%) and Ae. japonicus (n = 22, 5.0%). These 3 species comprised 82.6% of total collections. The BG-Pro trap collected the highest average number of female Ae. triseriatus (2.53 per trap day), approximately three times more than the CDC light trap or BG-Sentinel 2 with lure (Table 1b). The odds of collecting Ae. triseriatus with the BG-Pro were 3.02 times higher (95% CI: 1.96–4.67) than with the CDC light trap (Fig. 1e). The BG-Sentinel 2 with lure collected the most female Ae. albopictus (2.5 per trap-day), about 1.5 times more than the BG-Pro and 4.46 times more than the CDC light trap (Table 1b). However, the odds of collecting Ae. albopictus with the BG-Sentinel 2 with lure (OR: 4.62, 95% CI: 2.76–7.74) were not significantly different from the BG-Pro (OR: 3.06, 95% CI: 1.78–5.24) (Fig. 1f). Only 22 Ae. japonicus were collected, limiting statistical analysis (Fig. 1g). No LACV was detected in any mosquito pools by qPCR or Vero cell culture.

Table 1a.Total number of mosquitoes collected by trap in Jackson County, NC (2023); June and August combined.

View Fullscreen

Table 1b.Summary statistics for LACV Vectors.

View Fullscreen

The BG-Pro trap demonstrated superior field effectiveness in capturing Ae. triseriatus compared to standard surveillance traps, collecting approximately three times more specimens than the CDC light trap or BG-Sentinel 2 with lure. The observed odds ratio of 3.02 (95% CI: 1.96–4.67) underscores the utility of the BG-Pro to sample the primary vector. Although the BG-Sentinel 2 with lure captured more Ae. albopictus, the difference was not statistically significant compared to the BG-Pro.

Our study's strengths include a robust 4 × 4 Latin square design, controlling for site-specific and temporal variations, and the use of GLMMs with appropriate adjustments for zero-inflated data. However, limitations include the relatively small sample size (n = 436), particularly for Ae. japonicus (n = 22), and the absence of LACV-positive mosquitoes, which limits conclusions about trap efficacy in monitoring infected populations. Notably, in summer 2024, the BG-Pro successfully collected LACV-positive Ae. triseriatus at the residence of a LACV-infected child (Byrd, unpublished data).

These findings suggest that the BG-Pro is a valuable tool for surveillance of host-seeking LACV vectors, especially Ae. triseriatus and Ae. albopictus. A comprehensive LACV surveillance strategy should incorporate multiple trap types to effectively monitor all three container-inhabiting vectors (Grim et al. 2007, Tamini et al. 2021, Sither et al. 2023). Further studies are warranted to enhance capture rates for these species and to validate the BG-Pro's utility in detecting LACV-infected mosquitoes (Day and Trout Fryxell 2025).

Madeleine Craig was supported, in part, through a summer internship funded by the CDC Cooperative Agreement Numbers 1U01CK000662: Southeastern Regional Center of Excellence in Vector-Borne Diseases: The Gateway Program. This study was also supported, in part, through Epidemiology and Laboratory Capacity for Prevention and Control of Emerging Infectious Diseases funding (NU50CK000530: CDC) awarded to the NCDHHS. We are grateful for the field assistance by Daygan Shouse, Nick Lundy, and Evan Joseph during the summer of 2023. We appreciate Ashley Evan (WCU Photography Services) for creating the photographs in Fig. 1. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the CDC or the North Carolina Department of Health and Human Services (NCDHHS).