Local persistence of novel regional variants of La Crosse virus in the Northeast United States

Background: La Crosse virus [LACV] (genus Orthobunyavirus, family Peribunyaviridae) is a mosquito-borne virus that causes pediatric encephalitis and accounts for 50-150 human cases annually in the USA. Human cases occur primarily in the Midwest and Appalachian regions whereas documented human cases occur very rarely in the northeastern USA. Methods: Following detection of a LACV isolate from a eld-collected mosquito in Connecticut during 2005, we evaluated the prevalence of LACV infection in local mosquito populations and genetically characterized virus isolates to determine whether the virus is maintained focally in this region. Results: During 2018, we detected LACV in multiple species of mosquitoes, including those not previously associated with the virus. We also evaluated the phylogenetic relationship of LACV strains isolated from 2005-2018 in Connecticut and found that they formed a genetically homogeneous clade that was most similar to strains from New York State. Conclusion: Our analysis argues for local isolation and long-term persistence of a genetically distinct lineage of LACV within this region. We highlight the need to determine more about the phenotypic behavior of these isolates, and whether this virus lineage poses a threat to public health.

not readily enter standard CDC light traps or gravid traps which are routinely used [3,9]. Additional sampling and testing of Ae. triseriatus and other locally abundant mosquito species are needed to accurately estimate the entomological risk of LACV in enzootic sites.
Here we report the results of mosquito monitoring activities detecting LACV in Connecticut. In 2018, we conducted two surveys that intensi ed the extent of mosquito sampling and testing in central and southwestern Connecticut and highlighted increased detection of LACV activity [10]. In addition, routine statewide surveillance in Connecticut, ongoing since 1997, detected four additional isolates of LACV since 2005. We present evidence for ongoing circulation of LACV in this region, and report on the entomological and viral phylogenetic data that supports local persistence of LACV in Connecticut.

Methods
Entomological data Mosquito Collections. Isolates of LACV analyzed in the present investigation were obtained from eldcollected mosquitoes that were procured from the following: a. Connecticut state has conducted annual monitoring of mosquitoes at 36 locations since 1997 and was increased to a total of 91 locations (8 counties) since 2001. Field surveillance was conducted using CDC light traps and gravid traps, and at some sites, BG Sentinel traps (Biogents, Germany), followed by mosquito identi cation and viral testing across the eight counties of Connecticut. Each site was sampled at least once every 10 days from June-October ( Figure 1). We consider here, data from the 2005 In addition, as a pilot study of vertical transmission at those sites, six oviposition cups (black plastic casino cups, lined with seed germination paper), were placed in the eld and checked weekly to collect container-breeding mosquito eggs.
Mosquitoes captured within each study were sorted from other insect fauna, identi ed morphologically to species level using a key [11], with a cold-chain maintained throughout.

Virus detection
Adult mosquitoes were identi ed and screened for arbovirus as follows. Pools of up to 50 mosquitoes (grouped by species, location, and capture date) are homogenized and inoculated on a Vero (African Green monkey kidney) cell line, for evidence of arboviral infection, as described Eastwood et al. 2019 [10]. Brie y, mosquitoes were homogenized in a vial with 1 mL PBS-G (phosphate-buffered saline containing 0.5% gelatin, 30% rabbit serum, 1× antibiotic/antimycotic) and a copper BB pellet, using a mixer-mill set for 4 min 25 cycles/second. Samples were then centrifuged for 5 min at 7,000 rpm at 4°C.
The supernatant (100 μl) was inoculated onto a con uent monolayer of Vero cells in 25-cm 2 culture asks, allowed to absorb for 5 min on a plate rocker, then provided with 4 mL of minimum essential media supplemented with 10% fetal bovine serum, 1× antibiotic/antimycotic. Flasks were incubated at 37°C with 5% CO2 and examined daily for cytopathic effect (CPE) for up to 7 d. Infected cell cultures showing CPE were harvested and stored at −80°C. To identify viruses, RNA was extracted from isolates using a QIAampViral RNA mini kit (Qiagen, Germantown, MD), eluted in a nal volume of 70 μL. A reverse transcription polymerase chain reaction (RT-PCR) was performed in a 25 μL reaction using a Titan One-Tube RT-PCR system (Roche Diagnostics, Indianapolis, IN) with generic orthobunyavirus primers [10,12]. Ampli cation products of the appropriate size, were puri ed as per Eastwood (2019), then commercially sequenced (Science Hill DNA Analysis Facility, Yale University, New Haven, CT).
Mosquito eggs collected using the oviposition cups, as part of the third collection study, were reared to adults in the laboratory of the CAES at 25°C with 16/8 h light/dark photoperiod. This F 1 generation was screened for virus as above, to test for evidence of LACV having been vertically transmitted.
Minimum infection rates [MIR] of LACV were determined parsed by mosquito species/site/year using the CDC-provided Excel Add-in tool for calculating bias-corrected maximum likelihood estimate pooled infection rates [13] .

Phylogenetic data
Nucleotide sequencing and genetic analyses Viral RNA was isolated from virus cultures using the QIamp viral RNA Kit (Qiagen, Valencia, CA). RT-PCR was performed using the Titan One-Tube RT-PCR System (Roche Diagnostics, Indianapolis, IN) with primers targeting each of the three genomic segments of LACV. Primer pairs BUNS+new/BUNS-new and M14C/M4510r were used to amplify the entire S and M segments as previously described (Armstrong and Andreadis 2006). In addition, a portion of L segment was ampli ed using primers LACL2fwd (GTAGTGTACTCCTATCTACAAAAC) and LACL1077rev (GTTGATATACCCTTTATGCTTTG) [6]. Ampli cation products were puri ed using the PCR puri cation kit (Qiagen, Valencia, CA) and sequenced at the DNA Analysis Facility on Science Hill (New Haven, CT) using an Applied Biosystems 3730xl 96capillary genetic analyzer (Foster City, CA). Overlapping sequence chromatograms were aligned and edited using ChromasPro (Technelysium Ltd. Tewantin, Australia) and virus sequences were deposited in GenBank (S1 Table).
Edited nucleotide sequences were compared to those available on GenBank using the Blastn search algorithm (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Multiple sequence alignments were generated by the ClustalW algorithm and phylogenetic relationships were evaluated by maximum-likelihood analysis in Mega 6.0. The optimal nucleotide substitution was selected by performing ML ts of 24 different models in Mega. Support for individual nodes was evaluated by performing 1000 bootstrap replicates. were identi ed at the Fair eld and Redding sites between 10 th July and 9 th October, during this intensive capture study. The isolates came from pools of Ae. triseriatus with the exception of one isolate from Aedes cinereus; Table 1c.

Results
The minimum infection rate (MIR) per 1,000 mosquitoes tested was calculated for each species infected by LACV (Table 2). MIRs ranged from 2.05-6.78/1,000 for Ae. triseriatus depending on the survey and year of sampling and were about 10-fold higher than those estimated for Ae. cinereus (0.24 and 0.33/1,000). LACV was also isolated from single pools of Aedes canadensis and Aedes trivittatus from the Hamden site in 2018 yielding MIRs of 1.18 and 1.09/1,000 respectively. The total number of mosquitoes collected and tested by species are given in S2 Table. In addition, we collected Aedes eggs via ovitraps to assess vertical transmission of LACV to mosquito offspring. A total of 1,204 Aedes japonicus, 13 Aedes albopictus, and 5,818 Ae. triseriatus (both male and female) were collected at Easton, Redding and Fair eld sites and reared to adults. None of these F 1 mosquitoes tested positive for LACV.

Phylogenetic data
To evaluate the phylogenetic relationships of the Connecticut strains of LACV, we analyzed the entire coding sequence of the M segment polyprotein gene (4,325 nt). Figure 2  However, we show that through intensive targeted trapping, further evidence of LACV activity in mosquitoes in the Northeast is produced. Following the original detection in 2005, no isolates of LACV were obtained in Connecticut for 10 further years. Corroborating data from a focused-surveillance studies with routine monitoring data in Connecticut, and isolates from previous years (2005, 2015 and 2016) enabled evidence of localized persistence of novel regional variants of LACV. Twelve of the fourteen new isolates were detected in 2018, when species-directed sampling efforts were increased to assess the persistence of this virus. To be noted, 2018 was a year when higher than usual numbers of mosquitoes, as well as arbovirus detections, were observed in general under the standard routine sampling in Connecticut (unpublished data, [14]), and this increased catch size may also have contributed to our detection of LACV.
Aedes triseriatus has been implicated as the main vector involved in the maintenance and transmission of LACV [15]. LACV isolates described here were detected between July and October from four different mosquito species but were most frequently detected in Ae. triseriatus, providing further evidence to support its role as the primary vector. Our detection of LACV in Ae. cinereus and Ae. trivittatus represents a novel nding in the USA. Aedes cinereus is frequently captured in Connecticut, and a source of several other arbovirus isolations as well, including Cache Valley, Eastern equine encephalitis, Jamestown Canyon, and West Nile viruses [11]. Nevertheless, detection of LACV in a particular mosquito species does not necessarily incriminate it as a competent vector for the virus, and further studies are needed to elucidate its respective role in LACV transmission in the region. Outside of the northeastern US, LACV has been isolated from a number of mammalian-biting mosquito species in addition to Ae. triseriatus, including Ae. albopictus, Ae. canadensis, Ae. japonicus, Ae. vexans and Psorophora howardii [6,16,17,18,19,20]. Aedes albopictus and Ae. japonicus are invasive mosquito species emerging in Connecticut [21,22]; however, no LACV isolates were found from either of these species in this study. Our detection of LACV in a pool of Ae. canadensis reinforces ndings in Ohio which showed this species to be a secondary vector of LACV, via both eld isolations of the virus and vector competence studies [23]; Aedes vexans is also thought to be an accessory vector for LACV in Virginia [24]. There is also experimental evidence that Culex spp. may play a limited role in LACV transmission dynamics [25].
Sciurids (chipmunks and squirrels) serve as the main vertebrate hosts for horizontal transmission and ampli cation of LACV, and a high rate of transovarial transmission is also reported in Ae. triseriatus in other regions of the USA [26,27]. The contribution of both vertical and horizontal transmission requires further study to better understand LACV dynamics in the northeastern USA. Our study here included limited testing for evidence of vertical transmission in F1 mosquitoes reared from eld-collected eggs, none of which tested positive. Finding no evidence of LACV in F 1 mosquitoes could be surprising, given the high rate of vertical transmission reported elsewhere in the USA. For example, minimal infection rates of LACV in mosquitoes collected as eggs ranged from 0.4 to 7.5/1,000 in study sites in West Virginia [28]. However, greatly increased numbers of eld samples are required, as well as controlled experiments in the laboratory, to fully assess the role of vertical transmission. Reduced levels of vertical transmission of virus in mosquitoes could be one reason why LACV is infrequently detected in the northeastern USA. Miller et al (1977) report that LACV remained infective to vertebrate hosts after eight transovarial passages in Ae. triseriatus with infection rates of up to 71% in the offspring of an infected female (25). From this, they estimated that LACV can persist four years or longer in the vector population in the absence of horizontal ampli cation in vertebrate hosts. Transovarial transmission (TOT) is ecologically signi cant in the persistence of many vector-borne pathogens and used frequently by members of Bunyavirales when classical horizontal transmission is not possible [29].
MIRs can be seen to be high in 2015 and 2016, re ecting that the small quantity of Ae. triseriatus captured also yielded LACV. Our focused study in 2017-18 aimed to improve the catch rate of this mosquito species to examine LACV in the region. Routine mosquito surveillance in Connecticut and throughout much of the northeast, largely focuses on the use of CDC light and gravid traps. Few adult Ae. triseriatus were captured during routine mosquito surveillance using CDC light traps, gravid traps, or standard deployment of BG sentinel traps, and it was not until our focused-LACV study 2017-18 using a wide variety of different traps, that we saw numbers of this species increase. We suggest BG-Sentinel baited not only with BG-lure, but additionally with CO 2 , may enable increased collections of adult Ae. triseriatus, and thus enhanced chance of detecting LACV if it is present in an area [3]; once commercially available, additional baits described Eastwood et al. (2020) are even more effective than BG-Lure [10]. Minimal eld infection rates reported elsewhere have ranged from 0.26 to 27 per 1,000 mosquitoes tested [15,28,30,31].
Variants of LACV identi ed during this study represent a distinct third lineage. In other geographic regions of the USA, where lineages one and two occur, the virus is associated with clinical symptoms of human disease. Conversely and to our knowledge, locally acquired clinical cases of La Crosse encephalitis are exceedingly rare in the northeastern US. Whether this is due to under-diagnosis of clinical illness to LACV (lack of case recognition), differences in virus strain virulence, a low prevalence of infection in mosquito vectors that effectively limits human exposure to biting activity in regions where the virus circulates, or limited vector competence by these mosquito species, remains to be determined. Clearly, given the presence of this arbovirus in mosquitoes in several areas of CT, there appears to exist an entomological risk. Furthermore, this risk occurs where there is currently no reported human disease, yet a sizeable human population which could come into contact with LACV-infected mosquitoes. Human serosurveys, in regions where the novel lineage of virus circulates, would be valuable to determining exposure rates and further illuminating epidemiology or possible differences in virulence.
Our ndings clearly warrant that further investigation be taken to assess the public health risk that LACV (lineage III) may pose to this region of the USA. There is a requirement for 1) vector competency studies with local populations of Ae. triseriatus and other identi ed mosquito species; 2) infection and virulence studies in an animal model; 3) assessment of the role of enzootic vertebrate reservoir hosts for this lineage of LACV, and 4) vertical transmission studies in Ae. triseriatus

Conclusions
In conclusion, there is a low-level prevalence of lineage-III LACV in Connecticut in unique local populations maintained independently of known areas of LACV with reported clinical cases, i.e. Midwest and Appalachia regions. We have identi ed two new mosquito species (Ae. cinereus and Ae. trivittatus) that acquire LACV infection and may be involved in virus transmission in addition to the natural vector Ae. triseriatus. Greater awareness is needed to assess and highlight the potential health risk of lineage-III strains of LACV in the northeastern USA. Authors' contributions -GE designed and performed eld collections of study (ii) and (iii), analyzed data, and wrote the original manuscript draft; JS coordinated mosquito identi cation and contributed to species analysis; JS and PA coordinated eld collections for CT State mosquito surveillance/study (i); MM performed virus screening of mosquito pools; PA and TA advised in the study designs, interpretation of data, and contributed to drafts of the manuscript. All authors read and approved the nal manuscript.

List Of Abbreviations
Authors information -GE is a vector-borne disease ecologist, currently based at Virginia Tech University, with a One Health approach to understanding the transmission cycles of vector-borne pathogens, and the potential for emergence of zoonotic and arthropod-borne diseases. She is particularly interested in arboviruses.