Emerging and re-emerging arthropod-borne viruses (arboviruses) cause human and animal disease globally. Field and laboratory investigation of mosquito-borne arboviruses requires analysis of mosquito samples, either individually, in pools, or a body component, or secretion such as saliva. We assessed the applicability of mosquito excreta as a sample type that could be utilized during studies of Ross River and West Nile viruses, which could be applied to the study of other arboviruses.
Mosquitoes were fed separate blood meals spiked with Ross River virus and West Nile virus. Excreta was collected daily by swabbing the bottom of containers containing batches and individual mosquitoes at different time points. The samples were analyzed by real-time RT-PCR or cell culture enzyme immunoassay. Viral RNA in excreta from batches of mosquitoes was detected continuously from day 2 to day 15 post feeding. Viral RNA was detected in excreta from at least one individual mosquito at all timepoints, with 64% and 27% of samples positive for RRV and WNV, respectively. Excretion of viral RNA was correlated with viral dissemination in the mosquito. The proportion of positive excreta samples was higher than the proportion of positive saliva samples, suggesting that excreta offers an attractive sample for analysis and could be used as an indicator of potential transmission. Importantly, only low levels of infectious virus were detected by cell culture, suggesting a relatively low risk to personnel handling mosquito excreta.
Mosquito excreta is easily collected and provides a simple and efficient method for assessing viral dissemination, with applications ranging from vector competence experiments to complementing sugar-based arbovirus surveillance in the field, or potentially as a sample system for virus discovery.
Testing for the presence of arboviruses in mosquitoes used in laboratory experiments or surveillance usually involves collecting samples, from pools of hundreds of mosquitoes to the legs and wings of an individual mosquito and testing them by different methods. These methods can be labour intensive and costly and require sacrificing the mosquitoes. Arbovirus detection can be made from mosquito saliva; however, the amount of saliva mosquitoes expel is very small, making detection difficult. Here we demonstrate that mosquitoes excrete Ross River and West Nile viruses at levels sufficient to be detected by molecular assays as early as 2 days after they have fed on an infected blood meal. The amount of live (infectious) virus in excreta is low, suggesting that mosquito excreta poses a relatively low risk to people handling the samples. Mosquito excreta is easily collected in the laboratory and has a range of applications including experiments designed to incriminate mosquito species as vectors (i.e. vector competence experiments), arbovirus surveillance in the field, and discovery of previously unknown viruses.
Citation: Ramírez AL, Hall-Mendelin S, Doggett SL, Hewitson GR, McMahon JL, Ritchie SA, et al. (2018) Mosquito excreta: A sample type with many potential applications for the investigation of Ross River virus and West Nile virus ecology. PLoS Negl Trop Dis 12(8): e0006771. https://doi.org/10.1371/journal.pntd.0006771
Editor: Charles Apperson, North Carolina State University, UNITED STATES
Received: July 11, 2018; Accepted: August 20, 2018; Published: August 31, 2018
Copyright: © 2018 Ramírez et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper.
Funding: This work was supported by a National Health and Medical Research Council-funded ‘Improving Health Outcomes in the Tropical North: A multidisciplinary collaboration (HOT NORTH)’, grant (Number 1131932) and by an internal Forensic and Scientific Services research and development project grant (Number RSS17-031). ALR was supported by a doctoral scholarship granted by the Australian Institute of Tropical Health and Medicine, James Cook University QLD, Australia. SAR was supported, in part, by National Health and Medical Research Council research fellowship 1044698. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
It has been estimated that vector-borne diseases account for almost 20% of the global burden of infectious diseases, with more than 80% of the world’s population living in areas at risk . Mosquitoes are the most important vectors of arthropod-borne viruses (arboviruses) globally. In recent years, many arboviruses have emerged or re-emerged due to several factors. High viral mutation frequency, widespread urbanization, and changes in land use, together with globalization and the growth of air travel, facilitate vector population increase and dispersal, and enable rapid transit of viremic humans [2, 3, 4]. Since few vaccines and antiviral therapies are available, critical work to understand and prevent arbovirus outbreaks must be undertaken both in the laboratory, by performing vector competence experiments to incriminate candidate species, and in the field by undertaking studies of virus ecology, as well as routine surveillance to identify periods of elevated virus activity.
Vector competence refers to the ability of a mosquito or other hematophagous arthropod to acquire, replicate, and successfully transmit a pathogen . This is a key parameter to estimate vectorial capacity, namely the potential of a mosquito population to transmit an infectious agent to a susceptible host population . Vector competence is determined by intrinsic factors that regulate virus infection of the midgut, escape from the midgut into the hemocel and associated tissues (dissemination), and finally infection of the salivary glands . In the laboratory, vector competence is evaluated usually by feeding mosquitoes an infectious bloodmeal or allowing them to feed on an infected vertebrate. After a period of time, their ability to transmit the pathogen is evaluated. Several methods are used to assess transmission in the laboratory. Historically, transmission was evaluated by allowing mosquitoes to feed on susceptible vertebrate hosts (such as suckling mice) and then assessing infection (e.g. via clinical changes in the mice) [8, 9]. However, many arboviruses lack an appropriate model vertebrate host that will produce sufficient viremia or antibodies after exposure to be detected using standard laboratory assays . Additionally, not all laboratories have the required biological security to allow handling vertebrate hosts in the same space as mosquitoes. Transmission can also be assessed in vitro, by forcing mosquitoes to salivate into capillary tubes  and then testing the expectorate for virus by inoculation in cell culture or by molecular assays. This method is relatively simple and removes ethical and logistical issues with working with live vertebrates. However, it can be an insensitive system to demonstrate transmission for some arboviruses, such as dengue viruses (DENVs) and chikungunya (CHIKV) [12,13]. Although not ideal, an alternative to estimate transmission potential is to test mosquito legs, wings, and/or heads, and use dissemination as a proxy for transmission . This method fails to take into account possible salivary gland barriers to transmission [7, 15] and may overestimate the true transmission rate. The main limitation of in vitro methods is that since the mosquitoes must be sacrificed, they provide an end-point measurement preventing longitudinal measurements from the same individual.
In the field, routine arbovirus surveillance is carried out to detect elevated viral activity in order to implement disease control measures. Different strategies can be used for arbovirus surveillance  and one of the most widespread methods is the collection, identification, pooling and testing of wild mosquitoes by molecular assays or virus isolation. However, mosquito-based surveillance is time consuming and requires a continuous cold-chain to preserve virus viability for downstream processing. To overcome these limitations, a mosquito-free surveillance system based on the detection of arboviruses in saliva of infected mosquitoes has recently been developed [17, 18]. Saliva is collected on honey-baited nucleic acid preservation cards (Flinders Associate Technologies, FTA), which inactivate the virus and preserve viral RNA. Viral RNA is then eluted from the cards and detected using standard molecular assays. Importantly, the RNA preserved on the FTA cards serves as a template for nucleotide sequencing allowing strain identification and genotyping. This system has been successfully incorporated into routine surveillance programmes in Australia and is generally effective, as evidenced by numerous detections of arboviruses from multiple locations [19, 20, 21, 22]. Similar approaches using honey-baited cards or sugar-baited wicks have been evaluated in Florida  and California [24, 25]. Like any novel or emerging technology, there is always an opportunity to enhance the sugar-based arbovirus surveillance system. Since only a limited number of virions are passed during salivation [26, 27], the amount of virus on the FTA cards is generally of low concentration, indicating that the diagnostic assays are operating at their limits of detection . This may lead to false negatives or insufficient template for downstream nucleotide sequencing. Additionally, this method will only detect mosquitoes after the extrinsic incubation period (EIP) which can take up to 14 days for some arboviruses. Finally, infection rates and vector species identification cannot be determined from honey-baited cards .
An exciting new application involves the collection of a previously overlooked sample. It was recently demonstrated by Fontaine et al.  that DENV RNA can be detected in excreta from Aedes aegypti mosquitoes with a disseminated infection. Since collection of excreta does not require sacrificing the mosquito, it allows for “time-to-event” estimation of the time for dissemination, and consequently, an estimation of the EIP when used as a proxy for transmission potential, in individual mosquitoes. Detection of viral RNA in mosquito excreta can also be used to select mosquitoes based on extreme phenotypes (viral refractory or susceptible) for experiments exploring the genetic basis of a complex trait. Mosquito excreta can potentially be used to complement sugar-based surveillance. Indeed, it appears that viral RNA detection in excreta is more sensitive than detection in saliva (89% vs 33% for DENV) . Detection of arboviruses from excreta of infected mosquitoes could enable more sensitive detection of arboviruses than existing honey-baited FTA cards relying on collection of mosquito saliva alone.
The main objective of the current study was to determine whether mosquitoes excrete the Australian endemic arboviruses Ross River virus (RRV; family Togaviridae, genus Alphavirus,) and West Nile virus (Kunjin strain, WNVKUN; family Flaviviridae, genus Flavivirus) at levels sufficient to be detected by real-time reverse transcription polymerase chain reaction (RT-PCR) molecular assays. Building upon the Fontaine et al.  findings, we also determined if the association between virus dissemination and excretion extends to other arboviruses. Then, as a way to potentially enhance the sensitivity of the sugar-based surveillance system, we compared the detection of RRV and WNVKUN in mosquito excreta with virus detected in saliva via filter paper cards. Importantly, in the context of workplace health and safety regulations affiliated with arbovirus surveillance systems, we evaluated whether excreted virus was infectious.
Materials and methods
RRV was isolated from a pool of Verrallina carmenti collected from the Cairns suburb of Yorkeys Knob, Queensland, Australia in 2007 . The virus had been previously passaged three times in African green monkey kidney (Vero) cells (ATCC, CCL-81). WNVKUN was isolated from a pool of Culex annulirostris collected in the Gulf Plains region of Queensland, Australia in 2002 . The virus had been previously passaged twice in porcine-stable equine kidney (PSEK) cells  before a final passage in Aedes albopictus (C6/36) cells (ATCC, CRL-1660).
Aedes vigilax was selected based on its status as the coastal vector of RRV in Australia . Eggs from colonized Ae. vigilax were obtained from NSW Health Pathology-ICPMR, Westmead Hospital, Westmead, Australia. The colony was originally established at the Malaria Research Unit at Ingleburn in 1986 from material collected near Townsville, Queensland. Eggs were hatched in 2L of 33% seawater containing ~45 mg of brain-heart infusion powder. Larvae were reared at 26°C 12:12 L:D and fed fish flakes (Tropical Flakes, Aqua One®, Ingleburn, Australia). Pupae were placed in 150 mL containers inside a 30 x 30 x 30 cm insect rearing cage. Emerged adults were held at 26°C, 75% RH and 12:12 L:D, and maintained on 15% honey solution ad libitum.
Culex annulirostris was selected based on its status as the primary WNVKUN vector in Australia . Adult mosquitoes were collected in February 2017 using passive box traps  baited with CO2 (1kg dry ice) and operated for 14 h (1700–0700) in a mixed Melaleuca and mangrove swamp near Cairns, Australia (−16.826613°, 145.707065°). These field mosquitoes were transported to the laboratory where they were briefly anesthetized and female Cx. annulirostris were sorted and maintained on 15% honey solution ad libitum at 26°C, 75% RH and 12:12 L:D. Since there is no evidence that WNVKUN circulates in the Cairns region , it is unlikely that the mosquitoes had acquired the virus in the field.
Mosquitoes were starved for 24 h before oral infection with virus. Five to 7 day-old female Ae. vigilax were offered RRV diluted in washed defibrinated sheep blood (Institute of Medical and Veterinary Science, Adelaide, Australia) at 37°C using a Hemotek membrane feeding system (Discovery Workshops, Accrington, Lancashire, UK) with pig intestine as a membrane. Cx. annulirostris were exposed to WNVKUN diluted in washed defibrinated sheep blood via the hanging drop method . To determine the virus titer of the blood at the time of feeding and to assess if there was any reduction in titer, a 100 μL sample of the blood/virus mixture was taken before and after feeding, diluted in 900 μL of growth media (GM; Opti-MEM (Gibco, Invitrogen Corporation, Grand Island, NY) containing 3% foetal bovine serum (FBS; In Vitro Technologies, Australian origin), antibiotics and antimycotics), and stored at -80°C. After feeding, mosquitoes were briefly anesthetized with CO2 gas, and blood-engorged females sorted and placed in modified containers (see below) or in 900 mL containers covered with 100% polyester gauze (Spotlight Pty Ltd, Australia). All mosquitoes were maintained at 28°C, 75% RH and 12:12 L:D within an environmental growth cabinet for 15 days.
Collection of excreta from mosquito batches
For each virus, 20 batches of 5 mosquitoes were placed in modified 200 mL polypropylene containers for excreta collection. The gauze-covered containers had a false floor made of fiberglass insect screen that allowed excreta to pass through onto a parafilm M (Bemis NA, Neenah, WI) disc situated about 5 mm below the screen to avoid cross contamination. Mosquitoes were fed on cotton balls soaked in 15% honey dyed with blue food colouring to allow for excreta visualisation and were replaced daily. Excreta was collected daily from day 2 to day 15 post-exposure (PE) using a cotton swab (Livingstone International, Rosebery, Australia) moistened with GM + 3% FBS. Each swab was placed in a 2 mL tube containing 1 mL GM + 3% FBS and stored at -80°C. Parafilm discs were replaced daily to avoid cross contamination. Mosquito mortality was also recorded daily. To compare the sensitivity of detection of viral RNA in excreta with the sensitivity of detection in saliva expectorates, on day 14 PE, mosquitoes were allowed to feed on a 4 cm2 filter paper card (FP; low chamber filter paper, Bio-Rad Laboratories, California) soaked in 100% honey dyed with red food colouring. After 24 h, the FP cards were removed, placed in a 2mL tube containing 1 mL GM + 3% FBS and stored at -80°C.
Collection of excreta from individual mosquitoes
At three different timepoints (RRV: 7, 10, 14 days PE; WNVKUN: 6, 11, and 14 days PE), 20 individual mosquitoes were placed into 70 mL containers modified with the same design as described above. A 1 cm2 FP card soaked in 100% blue honey was offered as a sugar source. The mosquitoes were allowed to feed on the cards for 18–24 h, after which the excreta and the cards were collected as described above.
Assessment of infection, dissemination and transmission rates from mosquito cohorts
Because the mosquitoes used for the batches and individual analyses were derived from a cohort exposed to the same infectious blood meal, we assessed the infection, dissemination and transmission rates only from the experiments that used individual mosquitoes. Saliva was collected using the in vitro capillary tube method described by Aitken  from mosquitoes described above. Bodies and legs+wings were stored separately in a 2mL tube containing 1 mL GM + 3% FBS with a single 5 mm stainless steel bead to assess for infection and dissemination, respectively. Saliva expectorates were expelled into a 2mL tube containing 500 μL of GM + 3% FBS. All samples were stored at -80°C.
The blood/virus mixtures were titrated as 10-fold dilutions in 96-well microtiter plates containing confluent C6/36 cell monolayers. Bodies and legs+wings were homogenized using a QIAGEN Tissue Lyser II (Qiagen, Hilden, Germany) for 3 minutes at 26 hz and centrifuged briefly at 14,000 g. Mosquito homogenates (bodies, legs+wings) and saliva expectorates collected using capillary tubes were filtered using a 0.2 μm membrane filter (Pall Corporation, Ann Arbor, MI). Filtered mosquito homogenates were inoculated in duplicate and filtered saliva expectorates were inoculated in quadruplicate onto confluent C6/36 monolayers in 96-well microtiter plates. To assess the viability of virus in excreta, 50 excreta samples collected from mosquito batches (10 samples, 5 time points) were homogenized and filtered as described above, and inoculated as neat (not diluted) and as 10-fold dilutions onto confluent C6/36 monolayers in 96-well microtiter plates. Plates were incubated at 28°C for 7 days before being fixed in PBS/20% acetone with 0.2% BSA and stored at -20°C. Virus infection in cells was assessed using a cell culture enzyme immunoassay (CC-EIA) using monoclonal antibodies: B10 for RRV and 4G2 for WNVKUN  (provided by Roy Hall, University of Queensland, Australia).
Thawed excreta samples were homogenized in the Tissue Lyser II as describe above. Thawed FP cards were maintained on ice and briefly vortexed every 5 min for 20 min . Viral RNA was extracted from the excreta supernatant and eluted FP cards with a QIAxtractor (Qiagen, Hilden, Germany) using the QIAmp One-For-All nucleic acid kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Viral RNA was detected using real-time TaqMan RT-PCR assays specific for RRV  and WNV  in a Rotor-Gene 6000 real-time PCR cycler (Qiagen, Australia). With each run, positive controls included an extraction control (bovine viral diarrhoeal virus, BVDV) and a positive virus control extracted from a virus stock with known titer. Negative controls included at least one negative extraction control and a no-template control (molecular grade water). For each sample, the threshold cycle number (Ct) was determined; lower Ct values correspond to a greater amount of viral template. Any sample with a Ct value ≥40 was considered negative .
For all the samples titrated in the CC-EIA, 50% endpoints (tissue culture infectious dose50, TCID50) were calculated using the method of Reed-Muench  and expressed as TCID50/mL. The Mann-Whitney U test was used to determine if there was a difference between the Ct values observed for excreta samples from batches and individuals, and between excreta samples and saliva expectorates on FP cards. Fisher’s exact test was used to compare the difference in between detection of viral RNA in excreta and detection of virus by CC-EIA in legs+wings, as an indication of virus dissemination. Scatter plots, heat maps and all statistical analyses were performed using GraphPad Prism version 7.0c (GraphPad Software, La Jolla CA, www.graphpad.com).
Infection, dissemination and transmission rates in mosquito cohorts
For RRV with Ae. vigilax, the mean (± SD) virus titer at the time of feeding was 108.1±0.1TCID50/mL and the overall infection rate was 82% (Table 1). For WNVKUN with Cx. annulirostris, the mean (± SD) virus titer at the time of feeding was 107.3±0.3 TCID50/mL and the overall infection rate was 42% (Table 2). All Ae. vigilax with confirmed RRV midgut infection developed a disseminated infection. Transmission of RRV was first observed on day 8 PE when 9/19 mosquitoes expectorated the virus. Only 76% (19/25) of Cx. annulirostris with confirmed WNVKUN midgut infection developed a disseminated infection. Transmission of WNVKUN was first observed on day 12 PE when 3/20 mosquitoes expectorated the virus.
Detection of viral RNA in excreta from batches of mosquitoes
RRV and WNVKUN viral RNA was excreted every day from day 2 PE onward in both Ae. vigilax and Cx. annulirostris, respectively, at levels sufficient to be detected by real-time RT-PCR. With the exception of one batch of Ae. vigilax and one batch of Cx. annulirostris, viral RNA was detected in excreta from all the batches of mosquitoes on at least one day (Fig 1). For RRV positive samples, Ct values ranged from 24.6 to 38.8. For WNVKUN positive samples, Ct values ranged from 26.6 to 39.2.
(A) Detection of RRV RNA from Ae. vigilax excreta collected daily from day 2 to day 15 post exposure (PE) (B) Detection of WNVKUN RNA from Cx. annulirostris excreta collected daily from day 2 to day 15 post exposure (PE). Lower Ct values correspond to a greater amount of viral template; a blank square indicates that viral RNA was not detected. A skull indicates that the container was removed from the experiment due to mortality of all 5 mosquitoes. X = not tested.
Detection of viral RNA in excreta from individual mosquitoes
It was possible to detect RRV RNA in excreta from individual Ae. vigilax on all days tested PE (Fig 2). Sixty-four percent (35/55) of samples were positive, with Ct values ranging from 25.1 to 37.6. No significant difference (P>0.05) was observed between the median Ct values from excreta collected from batches of mosquitoes and from individual mosquitoes, with the exception of day 8 PE where the median Ct value for batches was higher (30.8 vs 27.5; P = 0.0001, S1 Fig).
Samples collected over 18–24 h from individual Ae. vigilax sampled at different timepoints post exposure (PE). Bars denote medians. P<0.05 (*), P<0.001 (**), P<0.0001(***). Each point represents an individual mosquito.
WNVKUN RNA was detected in excreta samples from individual Cx. annulirostris tested on all days PE (Fig 3). Twenty-seven percent (16/59) of samples were positive, with Ct values ranging from 28.9 to 39.2. No significant difference (P>0.05) was observed between the median Ct values from excreta collected from batches of mosquitoes and from individual mosquitoes (S2 Fig).
Samples collected over 18–24 h from individual Cx. annulirostris sampled at different timepoints post exposure. Bars denote medians. P<0.05 (*), P<0.001 (**), P<0.0001(***). Each point represents an individual mosquito. No mosquitoes expectorated virus onto filter paper cards on days 7 and 12 PE.
Association between disseminated infection and excretion of arboviruses
From 55 Ae. vigilax individuals tested, 45 (82%) mosquitoes had disseminated RRV infection. We detected RRV RNA in the excreta of 35 (78%) mosquitoes with a disseminated infection. None of the mosquitoes without a disseminated infection had positive excreta. From 59 Cx. annulirostris individuals tested, 19 (32%) had disseminated WNVKUN infection. Thirteen (68%) mosquitoes with a disseminated infection had excreta positive for WNVKUN RNA. Only 3 (8%) mosquitoes without disseminated infection had positive excreta. For both RRV and WNVKUN, there was a significant (P<0.0001) association between disseminated infection and excretion of viral RNA.
Comparison of detection of arboviruses in excreta and saliva
Saliva deposited on FP cards from batches of mosquitoes on day 15 PE was tested for viral RNA. For Ae. vigilax, the proportion of RRV positive excreta samples was higher than the proportion of RRV positive FP cards (89% (16/18) vs 22% (4/18); P<0.0001). For Cx. annulirostris, the proportion of WNVKUN positive excreta samples was higher than the proportion of WNVKUN positive FP cards (79% (15/19) vs 42% (8/19); P = 0.0448). For both viruses, no significant difference (P>0.05) was observed between the median Ct values obtained from positive excreta and saliva expectorates on FP cards (Fig 4)
(A) Detection of RRV RNA by real time RT-PCR in excreta and filter paper cards collected over 18–24 h from batches of 5 Ae. vigilax. (B) Detection of WNVKUN RNA by RT-PCR in excreta and filter paper cards collected over 18–24 h from batches of 5 Cx. annulirostris.
There was a significant difference (P < 0.05) between the proportions of RRV positive excreta and RRV positive FP cards obtained from individual Ae. vigilax at each time point (Table 3). With the exception of day 11 PE, where only one FP card was positive, median Ct values were significantly different between excreta and FP cards (day 8 PE: P<0.05; day 15 PE: P<0.01; overall: P<0.01; Fig 2).
For WNVKUN only 2 FP cards were positive on day 15 (Fig 3). With the exception of day 15 PE, there was a significant difference (P < 0.05) between the proportions of WNVKUN positive excreta and FP cards obtained from Cx. annulirostris at different time points (Table 3). There was no significant difference (P>0.05) between median Ct values obtained from excreta and FP samples (Fig 3).
Specificity and sensitivity of viral RNA detection in excreta and FP cards as a proxy for viral dissemination were calculated as described by . Mosquitoes with a confirmed disseminated infection (assessed by CC-EIA) and a positive RT-PCR result were considered true positives (TP) and those with a disseminated infection but a negative RT-PCR result were considered false negatives (FN). Mosquitoes without a disseminated infection and negative RT-PCR result were considered true negatives (TN) and those without a disseminated infection and positive RT-PCR result were considered false positives (FP). Using excreta as a proxy for viral dissemination, detection of RRV in excreta is highly specific (100%) and moderately sensitive (78%, 95%CI: 66–90). In contrast, detection of RRV in FP cards is highly specific (100%) but only slightly sensitive (18%, 95%CI: 7–29). For WNVKUN, detection in excreta also is highly specific (93%, 95%CI: 84–100) and moderately sensitive (68%, 95%CI: 48–90) while detection in FP cards is highly specific (100%) but slightly sensitive (11%, 95%CI: 0–24).
Viability of arboviruses in excreta
To evaluate whether the excreted virus was infectious, 50 samples collected from batches of mosquitoes from each experiment (10 batches from 5 time points, RRV: day 2, 3, 6, 9 and 13 PE; WNVKUN: day 2, 4, 6, 9 and 13 PE) were inoculated onto C6/36 cells and virus infection confirmed using the CC-EIA. Only 3 samples (6%) from different batches on different days had sufficient material to quantify the amount of RRV (day 2PE: 103.06 TCID50/mL; day 3PE: 101.30 TCID50/mL; day 9PE: 101.80TCID50/mL). Trace amounts of viable RRV were found on 8% (4/50) of the samples. In these samples CC-EIA indicated the presence of the virus in at least one well, but it was below the calculation cut-off value. Only one sample from day 9 PE showed trace amounts of viable WNVKUN (2%, 1/50).
Our results confirm that mosquitoes exposed to RRV or WNVKUN excrete viral RNA at levels sufficient to be detected by molecular assays. Our findings, together with previous observations on the excretion of DENV RNA by Ae. aegypti  support the hypothesis that the excretion of arboviruses by mosquitoes is a general phenomenon. Interestingly, even when the infection rate of WNVKUN in Cx. annulirostris (42%) was lower than the infection rate of RRV in Ae. vigilax (82%), we were able to detect viral RNA in excreta from batches of mosquitoes continually from day 2 to day 15 PE. This indicates that the detection of viral RNA in excreta is not a result of a high mosquito infection rate under laboratory conditions. Blood meal digestion times vary between mosquito species, but generally 72 hours after feeding it has finalized . Similar to the results of Fontaine et al., we observed brown excreta spots from digested blood meals in samples from day 2 and 3 PE, hence it is possible that viral RNA from those samples came directly from the blood meal. From day 4 onward, no dark excreta spots were visible, indicating that blood meal digestion was completed. The excreta from individual mosquitoes also provided sufficient material for detection of viral RNA at all timepoints indicating that the method is sensitive enough regardless of the volume of excreta collected. Indeed, we were able to detect viral RNA from containers with as little as one visible blue excreta spot.
We observed a correlation between viral dissemination and excretion of viral RNA. RRV RNA was not detected in excreta from any individual Ae. vigilax tested without a disseminated infection. Only 3 excreta samples from Cx. annulirostris without disseminated infection but with confirmed midgut infection were positive for WNVKUN RNA. However, it is important to note that viral dissemination was assessed by cell culture, which is less sensitive than RT-PCR  and may have failed to detect low titer disseminated infection. RRV disseminates quickly in Ae. vigilax; 2 days after ingesting an infectious bloodmeal  with transmission occurring from day 3–4 PE . Similarly, dissemination of WNVKUN in Cx. annulirostris is detectable as early as day 3, with initial transmission observed on day 5 and increasing from day 10 to day 14 PE . We detected RRV and WNVKUN RNA in 90% and 70% excreta samples from batches of Ae. vigilax and Cx. annulirostris, respectively, collected on day 4 PE, when viral dissemination has already occurred for both viruses. Our results from individuals and batches of mosquitoes support the idea that testing mosquito excreta could be used in vector competence experiments as an indicator of viral dissemination or as a proxy for virus transmission potential for arboviruses that do not have a suitable transmission model, such as the DENVs, without having to sacrifice the insects. A limitation of this method is that it is impossible to distinguish viral RNA resulting from blood meal digestion from that being excreted because of viral dissemination. In order to avoid false positives, excreta samples should be collected after blood meal digestion has finalized.
For both batches and individual mosquitoes (overall), the proportion of positive excreta samples was higher than the proportion of positive saliva samples, suggesting that excreta offers an attractive sample for analysis of mosquitoes with disseminated infection in the laboratory and potentially in the field. Although specificity of detection of viral RNA when used as a proxy for viral dissemination in both excreta and saliva is high, sensitivity is at least 4 times higher for excreta compared to saliva (RRV: 78% vs 18%: WNVKUN: 68% vs 11%). Indeed, for WNVKUN only 2 saliva samples were positive for viral RNA. These differences in sensitivity are expected, since detection of viral RNA in excreta and saliva result from different processes: dissemination and transmission. Not all mosquitoes with a disseminated infection transmit the virus, and the existence of a salivary gland infection barrier, where the virus is unable to enter or establish infection of the salivary glands prior to transmission has been documented. [7, 15]. In this experiment, only 44% and 42% of the mosquitoes with a disseminated infection transmitted RRV and WNVKUN, respectively, as measured by the capillary tube method. The median Ct values obtained from positive saliva expectorates were significantly higher than those from positive excreta samples obtained from individual mosquitoes. This is not surprising, since the volume of fluid excreted by mosquitoes is higher than what they expectorate (~1.5 μl  vs 4.7 nl ). This difference was not observed in batches of mosquitoes, possibly because there was more than one mosquito expectorating onto each filter paper card, potentially increasing the amount of viral RNA.
There is potential for mosquito excreta to be applied to enhance arbovirus surveillance. Honey-based surveillance provides a better estimate of transmission risk than testing pools of mosquitoes, since only transmitting mosquitoes will yield positive results [17, 46]. However, the proportion of mosquitoes in a population that survive the extrinsic incubation period can be low. Given that arboviruses can be detected in excreta as early as 2 days after the ingestion of an infectious blood meal, mosquito excreta could be used to obtain evidence of arbovirus circulation earlier. These results could be used to prompt intensive mosquito trapping for pooling and processing by traditional methods. Since mosquitoes expel only small quantities of saliva, the amount of virus on FTA cards is generally of low concentration which may lead to false negatives . In this study, we observed that detection of arboviruses in excreta is more sensitive than detection in saliva. Further experiments will be required to establish if large amounts of excreta from non-infected mosquitoes would reduce the ability to detect viral RNA from the excreta of a single mosquito and to evaluate its performance under field conditions. Additionally, a methodology would need to be developed to collect and preserve the viral RNA from excreta in light traps and passive mosquito traps [18, 35] in a way that is convenient for routine surveillance. Recently, a method was described to collect mosquito excreta for xenomonitoring of filarial parasites, malaria, and trypanosomes, using super hydrophobic cones to concentrate excreta either into tubes or FTA cards, enabling detection of parasite DNA from the samples . Finally, mosquito excreta could be used as an exploratory sample for virus discovery or metagenomic analysis by providing a template for next generation sequencing, greatly reducing associated costs (one sample vs several pools of mosquitoes per trap).
Only low or trace amounts of viable virus were found in excreta samples. It has been proposed that arbovirus virions in the midgut are inactivated by digestive proteases that affect the integrity of their envelope, rendering the virion non-infectious . The sample with the highest titer (RRV, 103.06 TCID50/mL) was obtained on day 2 PE and it is possible that this “higher” viral titer resulted from the digestion of the recently acquired infectious blood meal. It is unlikely that mosquito excreta has a role as an alternative route of transmission under field conditions. Firstly, arboviruses are labile in the environment; in fact, viability of arboviruses in infected mosquitoes decreases rapidly after their death in hot and humid conditions . Mosquito excreta also contains digestive enzymes  which could continue to inactivate remaining virions once they have been excreted. Secondly, arbovirus infection via aerosol has only been observed under circumstances of high virus concentration . Studies to test Japanese encephalitis virus (JEV) vaccines using Rhesus macaques exposed intranasally to JEV required at least 6.6 x 106 infectious units per animal to achieve infection [51, 52]. Our results obtained from batches of 5 mosquitoes with a high infection rate showed only low or trace amounts of viable virus. In the field, where only 1–2 mosquitoes out of thousands in a trap might be infected, the amount of viable virus in excreta would be even lower. Finally, it is well documented that mosquito saliva plays an important role in facilitating arbovirus transmission  and excreta lacks salivary proteins responsible for generating favourable replication conditions in the vertebrate host.
There are some factors that influence the outcome of experiments that rely on experimental infection of mosquitoes. A limitation of our study was the use of field collected Cx. annulirostris. It has been documented that the source of the vector population plays a role in the outcome of vector competence studies . Unknown factors such as age, previous exposure to other pathogens, temperature and vector microbiome can affect vector competence and the reproducibility of the experiment [55, 56]. Differences in blood meal titers could also influence rates of excreta detection. Midgut infection and escape barriers are dose dependent . Females exposed to higher viral doses tend to develop a disseminated infection quicker. In contrast, females ingesting lower viral doses have lower infection rates and take longer to amplify the virus . In our study, both mosquitoes were exposed to high viral titers, which could explain the early detection of viral RNA in excreta resulting from viral dissemination. While excreted viral RNA is detected earlier from mosquitoes exposed to higher titers, Fontaine et al. did not observe a difference in the amount of DENV RNA excreted between low and high titers. Further experiments will be required to determine if this applies to other arboviruses.
Important work to understand and prevent arbovirus outbreaks is undertaken in the laboratory and in the field analysing different mosquito samples. Mosquito excreta is an easily collected sample and provides a simple and efficient method for assessing virus dissemination in vector competence experiments. Although the use of mosquito excreta to enhance sugar-based arbovirus surveillance is still at experimental stage, our results suggest that excreta offers an attractive sample for analysis that could enable earlier and more sensitive detection of circulating arboviruses, and potentially be used for virus discovery.
S1 Fig. Detection of RRV RNA by real time RT-PCR in excreta from batches and individual mosquitoes.
Samples collected over 18–24 h from batches and individual Ae. vigilax sampled at different timepoints post exposure (PE). Bars denote medians. P<0.05 (*), P<0.001 (**), P<0.0001(***). Each point represents either a batch of 5 or an individual mosquito.
S2 Fig. Detection of WNVKUN RNA by real time RT-PCR in excreta from batches and individual mosquitoes.
Samples collected over 18–24 h from batches and individual Cx. annulirostris sampled at different timepoints post exposure (PE). Bars denote medians. P<0.05 (*), P<0.001 (**), P<0.0001(***). Each point represents either a batch of 5 or an individual mosquito.
The authors would like to thank Peter Burtonclay, Tanya Constantino and Bruce Harrower for their assistance and advice on maintaining cell cultures; Ian Mackay for his advice on analysis of molecular results; Frederick Moore, Amanda De Jong, Neelima Nair, Doris Genge, Jane Cameron, Sean Moody and Peter Moore for their technical assistance; Michael Townsend and Lili Usher-Chandler for the collection of adult mosquitoes; and Roy Hall for providing monoclonal antibodies.
- 1. World Health Organization. Global vector control response 2017–2030. Geneva: World Health Organization; 2017.
- 2. Liang G, Gao X, Gould EA. Factors responsible for the emergence of arboviruses; strategies, challenges and limitations for their control. Emerg Microbes Infect. 2015;4(3):e18 pmid:26038768
- 3. Hotez PJ. Global urbanization and the neglected tropical diseases. PLoS Negl Trop Dis. 2017;11(2):e0005308 pmid:28231246
- 4. Kilpatrick AM, Randolph SE. Drivers, dynamics, and control of emerging vector-borne zoonotic diseases. Lancet. 2012;380(9857):1946–55 pmid:23200503
- 5. Chamberlain RW, Sudia WD. Mechanism of transmission of viruses by mosquitoes. Annu Rev Entomol. 1961;6:371–90 pmid:13692218
- 6. Kramer LD. Complexity of virus-vector interactions. Curr Opin Virol. 2016;21:81–6 pmid:27580489
- 7. Hardy JL, Houk EJ, Kramer LD, Reeves WC. Intrinsic factors affecting vector competence of mosquitoes for arboviruses. Annu Rev Entomol. 1983;28:229–62 pmid:6131642
- 8. Hardy JL, Reeves WC. Experimental studies on infection in vectors. In: Reeves WC, editor. Epidemiology and control of mosquito-borne arboviruses in California, 1943–1987. Sacramento, CA: Califormia Mosquito Vector Control Association; 1990. p. 145–250.
- 9. van den Hurk AF, Nisbet DJ, Hall RA, Kay BH, MacKenzie JS, Ritchie SA. Vector competence of Australian mosquitoes (Diptera: Culicidae) for Japanese encephalitis virus. J Med Entomol. 2003;40(1):82–90 pmid:12597658
- 10. Gubler DJ, Rosen L. A simple technique for demonstrating transmission of dengue virus by mosquitoes without the use of vertebrate hosts. Am J Trop Med Hyg. 1976;25(1):146–50 pmid:3980
- 11. Aitken THG. An in vitro feeding technique for artificially demonstrating virus transmission by mosquitoes. Mosq News. 1977;37(1):130–3
- 12. Lequime S, Richard V, Cao-Lormeau VM, Lambrechts L. Full-genome dengue virus sequencing in mosquito saliva shows lack of convergent positive selection during transmission by Aedes aegypti. Virus Evol. 2017;3(2):vex031 pmid:29497564
- 13. Vazeille M, Mousson L, Martin E, Failloux AB. Orally co-Infected Aedes albopictus from La Reunion Island, Indian Ocean, can deliver both dengue and chikungunya infectious viral particles in their saliva. PLoS Negl Trop Dis. 2010;4(6):e706 pmid:20544013
- 14. Lambrechts L, Chevillon C, Albright RG, Thaisomboonsuk B, Richardson JH, Jarman RG, et al. Genetic specificity and potential for local adaptation between dengue viruses and mosquito vectors. BMC Evol Biol. 2009;9:160 pmid:19589156
- 15. Franz AW, Kantor AM, Passarelli AL, Clem RJ. Tissue barriers to arbovirus infection in mosquitoes. Viruses. 2015;7(7):3741–67 pmid:26184281
- 16. Ramirez AL, van den Hurk AF, Meyer DB, Ritchie SA. Searching for the proverbial needle in a haystack: advances in mosquito-borne arbovirus surveillance. Parasit Vectors. 2018;11(1):320 pmid:29843778
- 17. Hall-Mendelin S, Ritchie SA, Johansen CA, Zborowski P, Cortis G, Dandridge S, et al. Exploiting mosquito sugar feeding to detect mosquito-borne pathogens. Proc Natl Acad Sci U S A. 2010;107(25):11255–9 pmid:20534559
- 18. Johnson BJ, Kerlin T, Hall-Mendelin S, van den Hurk AF, Cortis G, Doggett SL, et al. Development and field evaluation of the sentinel mosquito arbovirus capture kit (SMACK). Parasit Vectors. 2015;8:509 pmid:26444264
- 19. Kurucz N, Wenham J, Hunt N, Melville L. Murray Valley encephalitis virus detection using honeybait cards in the Northern Territory in 2013. Mosq Bites. 2014;9(1)
- 20. Doggett S, Haniotis J, Clancy J, Webb CE, Toi C, Hueston L, et al: The New South Wales arbovirus surveillance & mosquito monitoring program. 2014–2015 annual report. Westmead, Australia: Department of Medical Entomology, ICPMR, Westmead Hospital; 2015.
- 21. Flies EJ, Toi C, Weinstein P, Doggett SL, Williams CR. Converting mosquito surveillance to arbovirus surveillance with honey-baited nucleic acid preservation cards. Vector Borne Zoonotic Dis. 2015;15(7):397–403 pmid:26186511
- 22. van den Hurk AF, Hall-Mendelin S, Townsend M, Kurucz N, Edwards J, Ehlers G, et al. Applications of a sugar-based surveillance system to track arboviruses in wild mosquito populations. Vector Borne Zoonotic Dis. 2014;14(1):66–73 pmid:24359415
- 23. Burkett-Cadena ND, Gibson J, Lauth M, Stenn T, Acevedo C, Xue RD, et al. Evaluation of the honey-card technique for detection of transmission of arboviruses in Florida and comparison with sentinel chicken seroconversion. J Med Entomol. 2016;53(6):1449–57 pmid:27330092
- 24. Lothrop HD, Wheeler SS, Fang Y, Reisen WK. Use of scented sugar bait stations to track mosquito-borne arbovirus transmission in California. J Med Entomol. 2012;49(6):1466–72 pmid:23270177
- 25. Steiner CD, Riemersma KK, Stuart JB, Singapuri A, Lothrop HD, Coffey LL. scented sugar baits enhance detection of St. Louis encephalitis and West Nile viruses in mosquitoes in suburban California. J Med Entomol. 2018:tjy064
- 26. Davis NC. Attempts to determine the amount of yellow fever virus injected by the bite of a single infected Stegomyia mosquito. Am J Trop Med Hyg. 1934;s1-14(4):343–54
- 27. Vanlandingham DL, Schneider BS, Klingler K, Fair J, Beasley D, Huang J, et al. Real-time reverse transcriptase-polymerase chain reaction quantification of West Nile virus transmitted by Culex pipiens quinquefasciatus. Am J Trop Med Hyg. 2004;71(1):120–3 pmid:15238700
- 28. van den Hurk AF, Hall-Mendelin S, Johansen CA, Warrilow D, Ritchie SA. Evolution of mosquito-based arbovirus surveillance systems in Australia. J Biomed Biotechnol. 2012;2012:325659 pmid:22505808
- 29. Fontaine A, Jiolle D, Moltini-Conclois I, Lequime S, Lambrechts L. Excretion of dengue virus RNA by Aedes aegypti allows non-destructive monitoring of viral dissemination in individual mosquitoes. Sci Rep. 2016;6:24885 pmid:27117953
- 30. Jansen CC, Prow NA, Webb CE, Hall RA, Pyke AT, Harrower BJ, et al. Arboviruses isolated from mosquitoes collected from urban and peri-urban areas of eastern Australia. J Am Mosq Control Assoc. 2009;25(3):272–8 pmid:19852216
- 31. van den Hurk AF, Nisbet DJ, Foley PN, Ritchie SA, Mackenzie JS, Beebe NW. Isolation of arboviruses from mosquitoes (Diptera: Culicidae) collected from the Gulf Plains region of northwest Queensland, Australia. J Med Entomol. 2002;39(5):786–92 pmid:12349863
- 32. Gorman BM, Leer JR, Filippich C, Goss PD, Doherty RL. Plaquing and neutralization of arboviruses in the PS-EK line of cells. Aust J Med Technol. 1975;6:65–70
- 33. Kay BH. Three modes of transmission of Ross River virus by Aedes vigilax (Skuse). Aust J Exp Biol Med Sci. 1982;60(3):339–44 pmid:6291499
- 34. Marshall I. Murray Valley and Kunjin Encephalitis. In: Monath TP, editor. The Arboviruses: Epidemiology and Ecology Vol 3. Boca Raton, Florida: CRC Press; 1988. p. 151–89
- 35. Ritchie SA, Cortis G, Paton C, Townsend M, Shroyer D, Zborowski P, et al. A simple non-powered passive trap for the collection of mosquitoes for arbovirus surveillance. J Med Entomol. 2013;50(1):185–94 pmid:23427669
- 36. Goddard LB, Roth AE, Reisen WK, Scott TW. Vector competence of California mosquitoes for West Nile virus. Emerg Infect Dis. 2002;8(12):1385–91 pmid:12498652
- 37. Broom AK, Hall RA, Johansen CA, Oliveira N, Howard MA, Lindsay MD, et al. Identification of Australian arboviruses in inoculated cell cultures using monoclonal antibodies in ELISA. Pathology. 1998;30(3):286–8 pmid:9770194
- 38. Hall RA, Prow NA, Pyke A. Molecular diagnostics for Ross River virus. In: Liu D, editor. Molecular Detection of Human Viral Pathogens. Boca Raton, FL: CRC Press; 2010. p. 349–59
- 39. Pyke AT, Smith IL, van den Hurk AF, Northill JA, Chuan TF, Westacott AJ, et al. Detection of Australasian Flavivirus encephalitic viruses using rapid fluorogenic TaqMan RT-PCR assays. J Virol Methods. 2004;117(2):161–7 pmid:15041213
- 40. Muench H, Reed LJ. A simple method of estimating fifty per cent endpoints. Am J Hyg. 1938;27(3):493–7
- 41. Gooding RH. Digestive processes of haematophagous insects. I. A literature review. Quaest Entomol. 1972;8:5–60
- 42. Hodinka RL. Point: is the era of viral culture over in the clinical microbiology laboratory? J Clin Microbiol. 2013;51(1):2–4 pmid:23052302
- 43. Kay BH, Jennings CD. Enhancement or modulation of the vector competence of Ochlerotatus vigilax (Diptera: Culicidae) for ross river virus by temperature. J Med Entomol. 2002;39(1):99–105 pmid:11931278
- 44. van den Hurk AF, Hall-Mendelin S, Webb CE, Tan CS, Frentiu FD, Prow NA, et al. Role of enhanced vector transmission of a new West Nile virus strain in an outbreak of equine disease in Australia in 2011. Parasit Vectors. 2014;7:586 pmid:25499981
- 45. Devine TL, Venard CE, Myser WC. Measurement of salivation by Aedes aegypti (L.) Feeding on a Living Host. J Insect Physiol. 1965;11:347–53 pmid:14330766
- 46. van den Hurk AF, Johnson PH, Hall-Mendelin S, Northill JA, Simmons RJ, Jansen CC, et al. Expectoration of Flaviviruses during sugar feeding by mosquitoes (Diptera: Culicidae). J Med Entomol. 2007;44(5):845–50 pmid:17915518
- 47. Cook DAN, Pilotte N, Minetti C, Williams SA, Reimer LJ. A superhydrophobic cone to facilitate the xenomonitoring of filarial parasites, malaria, and trypanosomes using mosquito excreta/feces. Gates Open Res. 2017;1:7 pmid:29377042
- 48. Johansen CA, Hall RA, van den Hurk AF, Ritchie SA, Mackenzie JS. Detection and stability of Japanese encephalitis virus RNA and virus viability in dead infected mosquitoes under different storage conditions. Am J Trop Med Hyg. 2002;67(6):656–61 pmid:12518858
- 49. Clements AN. Adult diuresis, excretion and defaecation. The Biology of Mosquitoes Vol 1. Oxfordshire, UK: CABI Publishing; 2000. p. 304–26
- 50. Kuno G. Dengue transmission without involvement of mosquito vector. Clin Infect Dis. 2005;40(5):774–5 pmid:15714437
- 51. Raengsakulrach B, Nisalak A, Gettayacamin M, Thirawuth V, Young GD, Myint KS, et al. An intranasal challenge model for testing Japanese encephalitis vaccines in rhesus monkeys. Am J Trop Med Hyg. 1999;60(3):329–37 pmid:10466957
- 52. Myint KS, Raengsakulrach B, Young GD, Gettayacamin M, Ferguson LM, Innis BL, et al. Production of lethal infection that resembles fatal human disease by intranasal inoculation of macaques with Japanese encephalitis virus. Am J Trop Med Hyg. 1999;60(3):338–42 pmid:10466958
- 53. Schneider BS, Higgs S. The enhancement of arbovirus transmission and disease by mosquito saliva is associated with modulation of the host immune response. Trans R Soc Trop Med Hyg. 2008;102(5):400–8 pmid:18342898
- 54. Jones RH, Foster NM. Heterogeneity of Culicoides variipennis field populations to oral infection with bluetongue virus. Am J Trop Med Hyg. 1978;27(1 Pt 1):178–83 pmid:204209
- 55. Wilson AJ, Harrup LE. Reproducibility and relevance in insect-arbovirus infection studies. Curr Opin Insect Sci. 2018;28:105–12
- 56. Richards SL, Lord CC, Pesko K, Tabachnick WJ. Environmental and biological factors influencing Culex pipiens quinquefasciatus Say (Diptera: Culicidae) vector competence for Saint Louis encephalitis virus. Am J Trop Med Hyg. 2009;81(2):264–72 pmid:19635881
- 57. Kramer LD, Hardy JL, Presser SB, Houk EJ. Dissemination barriers for western equine encephalomyelitis virus in Culex tarsalis infected after ingestion of low viral doses. Am J Trop Med Hyg. 1981;30(1):190–7 pmid:7212166
- 58. Mahmood F, Chiles RE, Fang Y, Green EN, Reisen WK. Effects of time after infection, mosquito genotype, and infectious viral dose on the dynamics of Culex tarsalis vector competence for western equine encephalomyelitis virus. J Am Mosq Control Assoc. 2006;22(2):272–81 pmid:17019773