1. Introduction
Estimating the titer of arboviruses transmitted by mosquitoes during blood feeding on a host is critical to understand arbovirus transmission, especially by accurately simulating these natural infections in laboratory studies. There are several documented methods for collecting saliva and/or determining the efficiency and titer of virus transmitted from mosquito saliva to a host: (1) forced salivation (FS) of infectious mosquitoes into media contained in a capillary tube and testing the captured saliva for virus, (2) detecting virus in host tissues immediately after infectious mosquitoes blood feed or by later examining host infection or seroconversion rates, and (3) detecting virus in the remaining blood from artificial feeders fed upon by infectious mosquitoes [
1,
2,
3,
4,
5,
6]. It is also possible to detect arboviral nucleic acids transmitted into sugar solutions when wild or laboratory mosquitoes sugar feed on special collection devices. However, this does not estimate virus transmitted to a host and tends to only be a qualitative measure of transmission potential because the number of sugar-feeding mosquitoes and the frequency at which they sugar-feed is often unknown [
7]. The three former quantitative methods each have challenges and their success is dependent on mosquito species and virus. Therefore, it is essential to evaluate the effectiveness of these methods on different mosquitoes and virus combinations to determine the best laboratory practices for predicting viral transmission.
Forced salivation is often used as it can determine virus expectorated from a single mosquito and is not dependent on the mosquito blood feeding. Within this method, different media have been used within the capillary tube, the most common being either microscope immersion oil [
2,
4] or fetal bovine serum (FBS) [
8]. FBS was thought to be a better medium to use because it may aid in viral stabilization and preservation, however, placing the mosquito proboscis into FBS is more difficult due to the hydrophobic properties of the mosquito cuticle [
4,
9]. In this sense, immersion oil is an easier choice to work with. Results with
Aedes albopictus infected with Venezuelan equine encephalitis virus (VEEV) showed that there was no difference between virus titers when saliva was collected in either immersion oil or FBS using the FS technique [
4].
In addition to forced salivation, vertebrate hosts or artificial feeders can be used to estimate the amount of virus being transmitted from an infectious mosquito bite. However, mosquito feeding behaviors are inconsistent, especially in high-containment (BSL-3) laboratory settings because of rapid air exchange and personal protective equipment that limit body heat and odor cues, so there is no way to ensure any one mosquito, or recalcitrant species or strains, will take a blood meal [
10]. A study with Eastern equine encephalitis virus (EEEV)-infected
Aedes aegypti used mouse intracerebral 50% lethal doses to show that the amount of virus transmitted varied from being undetectable to 1.0 × 10
5 [
11]. An additional study with EEEV showed mosquitoes transmitted ~1.0 × 10
3 PFU as measured by FS collections with immersion oil [
12]. Comparing these two results suggested that the quantities of EEEV transmitted during blood feeding and FS collection were approximately equal. However, other studies with different virus-vector pairings have given disparate results. For example, a study that quantified the amount of West Nile virus (WNV) transmitted by
Culex tarsalis after blood feeding determined that virus transmitted was approximately 600-fold higher than virus transmitted during the FS technique [
13], but another study showed similar virus titers transmitted from
Culex pipiens quinquefasciatus by blood feeding and forced salivation [
14]. Taken together, these results show that there is variation in the amount of virus transmitted in saliva that may be dependent on the virus-vector pairing, and that detection methods vary widely in their accuracy and precision.
Here, we have attempted to quantify virus titers transmitted from
Aedes aegypti mosquitoes infected with Zika virus (ZIKV) and chikungunya virus (CHIKV). Our efforts examined variations on the FS technique and compared it to virus transmission during blood feeding on animals and artificial feeders, and also to re-ingested virus recovered from bloodmeals dissected out of the mosquitoes (
Figure 1). The results from this study add critical information to understanding the transmission of
Aedes aegypti-borne arboviruses, which are responsible for frequent human disease epidemics across the tropical and sub-tropical areas of the world.
2. Materials and Methods
2.1. Virus and Cells
African Green Monkey kidney cells (Vero; ATCC #CCL-81) were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (DMEM; Gibco Thermo Fisher, FBS; Hyclone, Logan, UT, USA), 2 mM L-glutamine (Gibco Thermo Fisher), 1.5 g/L sodium bicarbonate (Gibco Thermo Fisher), 100 U/mL penicillin (Gibco Thermo Fisher) and incubated at 37 °C in 5% CO2. ZIKV strain PRVABC59 (ZIKV-PR; GenBank:KU501215), originally isolated from a human traveler to Puerto Rico in 2015 with three rounds of amplification on Vero cells, was obtained from Dr. Aaron Brault (CDC, Ft. Collins, CO, USA). CHIIKV strain LR2006_OPY1 (GenBank: KT449801.1) was obtained from the University Texas Medical Branch and isolated from outbreak in Reunion Island in 2006 with three rounds of amplification on Vero cells.
2.2. Mosquito Infections
Aedes aegypti Poza Rica strain mosquitoes were fed an infectious artificial blood meal containing either CHIKV or ZIKV and held for 10–14 days before all subsequent experiments to ensure dissemination of virus to the salivary glands. Infectious bloodmeals were prepared with 1 mL fresh virus contained in the cell-culture supernatant of infected Vero cells and 1 mL of defibrinated calf blood. Back-titering of the bloodmeals ranged between 1 × 106–5 × 106 PFU/mL. Mosquitoes were sorted post blood feeding and were placed in cartons (Huhtamaki, paper food container 64oz) with an organdy cover and given water and a sugar source.
2.3. Mice Infection
A129 mice (interferon alpha/beta receptor -/-) 8–12 weeks old were obtained from breeding colony maintained at Colorado State University. Use of mice was approved by the Colorado State University Institutional Animal Care and Use Committee (protocol 15-6677 AA). All procedures were done in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. To infect mice by mosquito bite, Aedes aegypti Poza Rica strain mosquitoes were fed an infectious blood meal and held for 14–17 days. Mosquitoes were sorted post blood feeding and 10–20 blood-fed mosquitoes were place in cartons with an organdy cover and given water and a sugar source. To allow the mosquitos to feed on the mice, each mouse was anesthetized using 100 mg/kg ketamine/10 mg/kg xylazine (ketamine: Zetamine from VetOne, xylazine: XylaMed from VET ONE) and placed on the organdy cover of one carton for ~20 min.
2.4. Mosquito Sample Collections
Mosquitoes were immediately cold-anesthetized post-blood feeding and their saliva was collected by the FS method described previously [
1], briefly their legs and wings were removed and their proboscis was placed into a capillary tube containing either mineral oil or FBS + glycerol at a ratio of 1:1. After 20–30 min, mosquitoes were pulled off the capillary tube and the capillary tube contents were centrifuged into 150 µL of 2x DMEM and held at −80 °C. The bodies were place in a separate tube held at −80 °C to be homogenized in media for later testing. Infections of mosquito bodies and saliva were determined by plaque assay and qRT-PCR. Samples were titrated by Vero cell plaque assay, with a tragacanth gum overlay and staining at day 5 post-cell culture inoculation for ZIKV and day 2 post-cell culture inoculation for CHIKV.
2.5. Bloodmeal Dissections
Bloodmeal dissections were done immediately after individual mosquitoes underwent FS. Mosquitoes were dissected on the sides of glass wells partially filled with 200 µL of DMEM supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 1.5 g/L sodium bicarbonate, 100 U/mL penicillin. The midguts were dissected out and spilt open so that the blood meal contents could spill out into the media, and the torn midgut swished into media to extract the whole blood meal. Mock blood meal dissections were performed on non-blood fed mosquitoes exactly the same way but there was no blood meal that could spill out into the media. The media (plus blood meal) was then collected and placed into a tube for later testing, and the body plus torn midgut were placed into another tube and frozen at −80 °C to be homogenized in media for later testing.
2.6. RNA Extractions and qRT-PCR
Tubes containing mosquito bodies were homogenized and both saliva and bodies where centrifuged for 5 min at 14,000×
g. Bloodmeals were collected in 150 µL of DMEM supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 1.5 g/L sodium bicarbonate, 100 U/mL penicillin. RNA was extracted from all samples using the Mag-Bind Viral DNA/RNA 96 kit (Omega Bio-Tek) on the KingFisher Flex Magnetic Particle Processor (Thermo Fisher Scientific). RNA was eluted in 30 µL nuclease-free water. Progmeg GoTaq Probe 1-Step RT-qPCR System Kits were used on RNA extracted from saliva and bodies to quantify CHIKV and ZIKV RNA according to manufacturers’ instructions. Standard cycling condition were followed, one cycle at 45 °C for 15 min, one cycle at 95 °C for 2 min and 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Primers used for CHIKV were Forward (5′-CTTTGAAGTTTCCTTTCGGTGG-3′) and Reverse (5′-ACFFAAFFRAAACTGGTATGG-3′) and Probe-FAM (5′-TCTGCAGCGTCTTTATCCACGGG-3′). Primers used were ZIKV 1086 (5′-CCGCTGCCCAACACAAG-3′) and ZIKV 1162c (5′-CCACTAACGTTCTTTTGCAGACAT-3′). The probe used was ZIKV 1107-FAM (5′-AGCCTACCTTGACAAGCAGTCAGACACTCAA-3′) [
15]. Approximately 100 ng of RNA was added to each reaction. Standards were generated for each virus using a full-length viral RNA. RNA was quantified on a Qubit Fluorometer (ThermoFisher Scientific) and diluted to achieve serial 10-fold genome equivalent (GE) dilutions. The standard curve detection of 104–107 GE/reaction had a primer efficiency of 88.62% to 102% with an R
2 value of 0.971 to 0.997, a slope of −3629 to −3269, and y-intercept = 37.966 to 47.270.
2.7. Statistical Analyses
Results in figures were expressed as mean values (horizontal bars) with individual values showing the variance. The statistical details are noted in the figures and/or in the corresponding figure legends. Statistical significance was primarily determined using either Fisher’s exact test, unpaired Student’s
t-test or a one-way analysis of variance (ANOVA) with a Tukey’s multiple-comparison in GraphPad Prism. Correlation was determined by Spearman’s rank-order correlation in the GraphPad Prism (GraphPad Software,
http://www.graphpad.com/faq/viewfaq.cfm?faq=1362 (accessed on 29 March 2021), La Jolla, CA, USA).
4. Discussion
We used different experimental procedures and methods to estimate the titer of virus transmitted by
Aedes aegypti mosquitoes after being infected with either ZIKV or CHIKV. Different medias have been previously evaluated for FS, including immersion oil and a mixture of 1:1 FBS + glycerol, as well as blood and other medias [
2,
4,
6,
8,
16,
17,
18,
19]. Results from these studies showed that the effectiveness of the media depends on mosquito and virus species, but no other study compared these methods with
Ae. aegypti infected with CHIKV or ZIKV. We found no evidence that FBS + glycerol aided in viral stabilization and preservation as was previously suggested [
4,
9] and no difference in titers from ZIKV- or CHIKV-positive samples, however, infectious saliva collected using oil resulted in significantly more positive samples from individual mosquitoes. The use of oil with FS is also easier to perform because the mosquitoes are drawn into the capillary tube via the similar hydrophobic properties of the mosquito’s cuticle and the oil, which draws saliva out of the salivary glands into the oil. Additionally, proof of successful saliva capture can be observed, and its quantity estimated, because the hydrophobic oil and aqueous saliva do not mix [
20]. In the same paper, Sanchez-Vargas et al. estimated that individual
Aedes aegypti expectorated a mean of 6.8 nL using oil-based FS and observed no correlation of CHIKV titers with the saliva volume expectorated. Another potential benefit of oil-based FS is that the mosquitoes are unable to re-ingest their own saliva. It has been previously shown that anopheline mosquitoes will re-ingest many of the
Plasmodium sporozoites they deposit in the host when blood feeding [
21]. It follows that saliva re-ingestion could also influence the virus detection success in FS.
FS has been used as the standard method to determine transmission of mosquito-borne arboviruses [
1,
2,
4,
5], but we are not aware of any examination of forced salivation on mosquitoes that blood fed immediately prior. Given the large variances in virus transmission by any one mosquito (
Table 2 and
Table 3 and the
supplemental tables), we wanted to know if virus transmission estimates using FS were different if performed immediately after they imbibed a second blood meal relative to estimates using FS from sibling mosquitoes never given a second blood meal. If there were no differences, one would be able to estimate the amount of virus transmitted in laboratory experiments, or even natural experiments in the field, using FS immediately after one or more mosquitoes took a blood meal on a host. For example, one could capture indoor resting blood fed mosquitoes from the walls of a house, perform FS on them immediately post-capture, and reliably estimate the titer that they may have just transmitted to the people in the house whom they bit. We used groups of mosquitoes (1–10 mosquitoes/group) that were given three different blood feeding treatments (none, blood fed on a mouse, blood fed on an artificial feeder) and demonstrated that infectious virus and viral RNA titers determined from FS were not different between the treatments. Furthermore, the ratios of infectious virus to viral RNA quantified were not different between the treatment groups. Each treatment group showed ~100–10,000 times more viral RNA than infectious virus as has been reported in many other studies using FS on unfed mosquitoes alone [
2,
22].
To compare our post-blood feeding FS data with the quantity of virus transmitted during the blood feed, we analyzed outcomes of ZIKV-infected mosquitoes that blood fed on immunocompromised mice, and virus transmitted to artificial feeders after ZIKV- or CHIKV-infected mosquitoes fed on them. Mice became infected after mosquito feeding even when no infectious virus was detected in the mosquito saliva post feeding. Our limit of detection is 2 PFU and 10 genomes copies; however, we never detected anything under 10 PFU. Based on these results, we can assume these mice can become infected with less than 10 PFU transmitted by mosquitoes blood feeding on them. When examining the blood remaining in the artificial feeders, however, infectious virus was never recovered, and viral RNA was only recovered in four out of nine ZIKV groups and one out of nine CHIKV groups. This observation could be explained in at least two non-exclusive hypotheses. It may be that live virus was quickly inactivated and viral nucleic acid sequences were destroyed by proteases and nucleases in the artificial blood meal, making their detection difficult by plaque assay and qRT-PCR, respectively. However, this seems unlikely given that we rarely record a drop in virus titer of the original blood meal used to first infect the mosquitoes when it is ‘back-titered’ after sitting in the artificial feeder for ~30 min during the blood feed. Another possibility is that virus expelled with the saliva into the artificial blood meal may be immediately re-ingested through the suction force needed to bring blood into the food canal.
To address the latter hypothesis, we dissected out bloodmeals from the infected mosquito midguts after they were given different blood feeding treatments and then were processed with FS. For one of the blood feeding treatments, blood was pipetted up and down in the blood feeder during the time of feeding to determine if blood mixing might counteract the re-ingestion of a mosquito’s own expectorated virus during blood feeding. Compared to the control treatment (mock dissection of bloodmeals from empty midguts), significantly more virus and viral RNA was recovered from the bloodmeals of the two blood feeding treatment groups, and the mixing of the blood in the artificial feeder did not influence this. This indicated that mosquitoes re-ingest much of their expectorated virus while feeding on the artificial feeder and that poor detection of virus in the remaining blood from artificial feeding is likely due to re-ingestion. As each mosquito dissected of its bloodmeal was also processed via FS, we could determine the difference of virus titers between the bloodmeal and saliva to estimate the quantity of virus transmitted during blood feeding. The estimate was consistent between
Ae. aegypti transmitting either ZIKV or CHIKV and between quantities of infectious virus or viral RNA detected; between 50–100 times more virus is secreted during blood feeding than is detected in FS performed immediately after blood feeding, suggesting a large increase in virus transmission during blood feeding. One limitation of these data are that they are estimates of virus transmission by
Ae. aegypti derived from blood feeding on artificial feeders, which may not accurately reflect what occurs during blood feeding on live hosts, including transmission during probing but not blood feeding [
23]. However, blood feeding on live hosts results in diverse outcomes. Mosquitoes will capillary feed by either fully cannulating capillaries, or just pierce the capillary at a right angle with the tip of the labrum, or sometimes might only nick a capillary and perform ‘pool feeding’ on the blood that pools into the interstitial space of the dermis [
24]. Each of these methods are likely to result in differing quantities of saliva/virus deposited as well as being re-ingested back into the blood meal. As such, artificial feeders may be a more consistent blood source for this estimation. In natural blood feeding experiments, Secundino et al. determined that the ZIKV cDNA ranged from 2.0 × 10
2–2.1 × 10
10 when the mouse ear tissue was immediately removed and homogenized after being fed on by ZIKV-infected
Ae. aegypti mosquitoes [
25]. In our study, this quantity is more comparable to the RNA loads in bloodmeals rather than saliva, indicating that mosquitoes re-ingest much of their saliva during natural blood feeding.
Studies using different combination of viruses and mosquito species have evaluated the use of vertebrate hosts or artificial feeders to estimate the amount of virus being transmitted from an infectious mosquito and found varying results [
11,
12,
13,
14]. Our data allow for estimation of the amount of ZIKV or CHIKV from an infectious
Ae. aegypti mosquito by performing FS on it immediately post-blooding, quantifying infectious virus or viral RNA and then multiplying the titer determined by ~50–100. Quantifying infectious virus ensures measurement of true infectious units, but it is clearly of low sensitivity and so simultaneously quantifying viral RNA will give the best estimates of transmission dose. More experiments will be necessary to determine if the increase in virus transmission during blood feeding relative to FS we observed is because of (a) more saliva being released by
Ae. aegypti or (b) more virus being released from the salivary glands, or both. Similarly, results from this study should be replicated with other arbovirus vectors to determine if they are consistent across mosquito species. Overall, the methods developed here can be used as a better way to estimate the titer of arboviruses transmitted by blood feeding
Ae. aegypti and may be valuable for similar estimations with other mosquitoes.