Role of Nestling Mourning Doves and House Finches as Amplifying Hosts of St. Louis Encephalitis Virus

Abstract

Nestling mourning doves and house finches produced elevated viremias after inoculation with 2–3 log₁₀ plaque-forming units (PFU) of St Louis encephalitis (SLE) virus and infected 67 and 70% of Culex tarsalis Coquillett that engorged upon them, respectively. Mosquito infection rates as well as the quantity of virus produced after extrinsic incubation increased as a function of the quantity of virus ingested and peaked during days 3–5 postinoculation in mourning doves and days 2–4 in house finches. Only female Cx. tarsalis with body titers ≥4.6 log₁₀ PFU were capable of transmitting virus. Overall, 38% of females infected by feeding on mourning doves and 22% feeding on house finches were capable of transmission. The quantity of virus expectorated was variable, ranging from 0.8 to 3.4 log₁₀ PFU and was greatest during periods when avian viremias were elevated. Our data indicated that nestling mourning doves and house finches were competent hosts for SLE virus and that the quantity of virus ingested from a viremic avian host varies during the course of the infection and determines transmission rates by the mosquito vector.

Because of their abundance in nature and frequent infection detected during serosurveys, mourning doves, Zenaida macroura, have been considered as important hosts of St. Louis encephalitis virus (SLE, Flaviviridae, Flavivirus) in Florida (Day 2001) and elsewhere (McLean and Bowen 1980). In California, mourning doves frequently were fed upon by the primary vector species, Culex tarsalis Coquillett (Reisen and Reeves 1990) and were antibody positive during previous (Kern County, 7%, n = 145; Sacramento Valley, 36%, n = 87; Milby and Reeves 1990) and recent (Reisen et al. 2000c) serosurveys. However, adult birds failed to consistently produce viremias of sufficient titer to infect Cx. tarsalis after experimental infection (Hardy and Reeves 1990b; Reisen et al. 2003b,c), questioning their importance as amplifying hosts for SLE virus. However, many bird species such as chickens (Reisen et al. 1994) are incompetent hosts as adults but produce elevated viremias when infected shortly after hatching (Hardy and Reeves 1990b). Therefore, a major goal of our current research was to describe the viremia response of nestling mourning doves after inoculation with SLE virus. Adult house finches, Carpodacus mexicanus, also were infected frequently in California (Reisen et al. 2000c), but, in contrast to mourning doves, they are competent hosts for SLE virus as adults, developing viremia titers sufficient to infect susceptible Cx. tarsalis (Hardy and Reeves 1990b; Reisen et al. 2003b). However, the viremia response of house finch nestlings to SLE infection has not been described, even though this could be important in virus amplification (McLean and Bowen 1980).

There seems to be a quantitative relationship between the quantity of SLE virus ingested by Cx. tarsalis females and the resulting infection rate (Hardy et al. 1979, 1983; Hardy and Reeves 1990a). The amount of virus ingested not only influences the proportion of females infected but also the quantity of virus produced within their bodies during the extrinsic incubation period, and therefore their ability to transmit virus by bite. During the current study, we quantitatively assessed the vector competence of female Cx. tarsalis feeding on nestling mourning doves and house finches to provide a description of virus dynamics by using natural components of the amplification cycle.

Materials and Methods

Birds.

Nestling mourning doves and house finches were obtained as eggs or nestlings from breeding colonies maintained in screened aviaries at the Arbovirus Field Station. These colonies originated from adults collected in the Bakersfield area of Kern County, California. Parent birds had never been infected with SLE virus, because this virus has not been active in this region of California during the previous 5 yr (Steinlein et al. 2003). Eggs were hatched by parent birds or in an egg incubator and the nestlings aged for 1–8 d before infection. Nestlings were maintained in individual paper cups lined with paper within a temperature-regulated brooder, where they were fed a commercial diet (Kaytee Exact Hand-feeding Formula, Lafayette, CA) 6–8 times per day. In addition, each day nestling mourning doves were fed 0.1–0.2 ml of a solution of diluted "crop milk" collected by irrigating of the crop of a brooding female with distilled water. This solution may have helped to establish the natural gut flora in the nestlings as well as provide added nutrients. Birds were infected by subcutaneous inoculation in inguinal region with 50–100 μl of virus diluent (phosphate-buffered saline plus 20% fetal bovine serum, 100 U of penicillin, 100 U of streptomycin, and 200 U of nystatin) containing 1.5–3.9 log₁₀ plaque-forming units (PFU) per 100 μl of SLE virus. This titer of infectious virus was similar to that expectorated by Cx. tarsalis, and previous studies showed that the viremia response in house finches after inoculation did not differ from that elicited after infectious mosquito bite (Reisen et al. 2000a). Viremia in nestling doves was monitored by taking 50 μl of blood samples via jugular vein puncture. Blood was diluted 1:10 in virus diluent and then frozen at −80°C for later testing. Viremia response also was estimated indirectly for doves and finches from the average quantity of virus ingested by two to three female Cx. tarsalis that engorged on each bird each day and were immediately frozen at −80°C (see below). Samples were taken on days 1–5 postinoculation (PI), because most adult birds clear SLE virus infections by 4–5 d PI (Reisen et al. 2003b).

Virus.

The Kern217 strain of SLE virus was isolated from Cx. tarsalis collected in the Bakersfield area during 1989 and was at Vero cell passage 2 when used for experimentation. This strain elicited a consistent viremia response in house finches (Reisen et al. 2000b) and has been used previously in our avian infection experiments (Reisen et al. 2003b).

Mosquitoes.

The high virus producer (HVP) laboratory strain of Cx. tarsalis was used (Kramer et al. 1998). This strain was susceptible to SLE virus and has been used in previous infection experiments (Reisen et al. 1993). Immatures were reared following standard procedures (Mahmood 1999), maintained on 10% sucrose at 22–24°C for 3–8 d, and then starved overnight before being fed on nestling birds. Nestling birds were exposed to 4–15 females daily for 4–5 d PI. Two to three engorged mosquitoes were frozen at −80°C immediately after blood feeding to determine the amount of virus imbibed; the remaining females were maintained on 10% sucrose for 11–13 d at 26°C, after which transmission was attempted using the capillary tube method (Aitken 1977). Previous studies with the HVP strain showed that transmission rates were 100% in infected HVP females held for >10 d at 25°C (Reisen et al. 1993). For transmission, each mosquito was immobilized by chilling, the wings and legs removed, and the proboscis inserted into a capillary tube containing a 1:1 mixture of 20% fetal calf serum and 10% sucrose. After 15 min, the expectorate solution was expelled into 300 μl of virus diluent, and this solution and the mosquito body were frozen separately at −80°C for later testing.

Assays.

Individual mosquitoes and viremia blood samples were triturated or vortexed, respectively, in 1 ml of diluent, after which 100 μl was tested for virus by plaque assay by using Vero cell culture with a two-overlay system (Kramer et al. 2002). End point titers were estimated by interpolation from the numbers of plaques counted at 10× magnification after 7 d.

Results

Viremia titers in 1–5-d-old mourning doves remained elevated for the 5-d sampling period and were relatively consistent among birds during days 3–5 PI as evidenced by the small standard deviations (Fig. 1). Sampling was terminated on day 5 after individual birds died on days 4 and 5, perhaps due to infection or repeated handling and sampling. We did not include blood samples from 1–5-d-old house finches, because our preliminary attempts to take samples from such young birds resulted in unacceptable mortality.

Overall, 87 and 105 engorged females were tested immediately after blood feeding to determine the quantity of virus that mosquitoes ingested by feeding on nestling mourning doves and house finches, respectively (Table 1). Two mourning doves (129 and 131) and two house finches (148 and 162) produced low-titered viremias (Figs. 2 and 3). The quantity of virus ingested by mosquitoes feeding on mourning doves (Fig. 2) averaged ≈2 log₁₀ PFU less than estimated from jugular puncture samples (Fig. 1). This was expected volumetrically, because viremias were expressed as PFU per 100 μl in Fig. 1, whereas virus titers ingested by mosquitoes were based on a blood meal volume of ≈3 μl, i.e., ≈1.5 log₁₀ PFU <100 μl. Based on the amount of virus ingested by mosquitoes, the viremia patterns in mourning doves differed markedly from house finches (Figs. 2a and 3a). All three doves with a high viremia maintained this titer through day 5, indicating we did not describe the end of the viremia curve. In contrast, house finch viremia peaked on days 2–4 in six of seven birds; only bird 163 remained highly viremic on day 5.

Overall, 82 and 141 females that engorged on mourning doves and house finches, respectively, survived the 11–13 d extrinsic incubation period (Table 1). All mosquitoes feeding on dove 131 and tested immediately or after the 12 d incubation period were negative; therefore, data from this negative bird were deleted from further calculations. Overall, 67% (n = 72) of mosquitoes feeding on viremic mourning doves were infected, statistically similar (χ² = 0.07, P > 0.05) to 70% (n = 141) of mosquitoes feeding on house finches. However, the percentage of mosquitoes infected after feeding on house finches peaked on days 2 and 3 postinoculation (Fig. 3b), whereas infection was greatest on days 3–5 for mourning doves (Fig. 2b), agreeing with the quantities of virus ingested (Figs. 2a and 3a). When tested by two-way analysis of variance (ANOVA) (Hintze 1998) with bird species and days after inoculation as main effects, the mean virus titer in mosquitoes infected by feeding on mourning doves (mean = 5.1 log₁₀ PFU per mosquito) was significantly less (F = 5.17; df = 1, 136; P = 0.02) than mosquitoes infected by feeding on house finches (mean = 5.5 log₁₀ PFU per mosquito). Body titers in infected mosquitoes varied significantly (F = 6.75; df = 4, 136; P < 0.001) among the 5 d that the mosquitoes fed on birds after inoculation, being lowest on day 1 (4.7 log₁₀ PFU per mosquito) and highest on day 2 (5.9 log₁₀ PFU per mosquito). This pattern was consistent between species, because the interaction effect in the ANOVA was nonsignificant (P = 0.09). The mean amount of virus ingested from each bird on each day was well correlated (r = 0.58, df = 207, P < 0.01) with mosquito body virus titers after extrinsic incubation (Fig. 4). Although these factors were well correlated, the amount of virus within viremic blood meals was not entirely predictive of infection or the resulting body titer. For our plaque assay to be positive, samples had to produce ≥2 PFU after 5–7 d of incubation. This would be ≥7 PFU/μl of blood ingested per mosquito blood meal or based on dilutions, ≥3.8 log₁₀ PFU/ml of viremic blood. This viremia titer at the threshold of plaque assay detection was above the titer required to infect some mosquitoes, as indicated by the number of feeding episodes on birds where some mosquitoes were negative immediately after blood ingestion, whereas others were positive after incubation (Fig. 4). In addition, some mosquitoes that imbibed >3 log₁₀ PFU did not become infected.

Overall, 38% (n = 48) of Cx. tarsalis mosquitoes feeding on mourning doves expectorated virus, statistically comparable (χ² = 2.95, P = 0.09) with 22% (n = 98) of Cx. tarsalis infected by feeding on house finches (Table 1). The titer of virus in expectorate samples was correlated (r = 0.30, n = 146, P < 0.01) with the titer of virus in the bodies of infected females (Fig. 5). Females with body titers of <4.6 log₁₀ PFU failed to transmit virus, although a body titer ≥4.6 log₁₀ PFU per mosquito did not ensure transmission. The quantity of virus in expectorate samples from females that fed on mourning doves (mean = 1.7 log₁₀ PFU, n = 18) was not significantly (P > 0.05) different than the quantity of virus expectorated by females that fed on house finches (mean = 1.8 log₁₀ PFU, n = 22). The quantity of virus expectorated by females feeding on birds on days 4 (mean = 1.8 log₁₀ PFU, n = 5) and 5 (mean = 2.5 log₁₀ PFU, n = 6) was significantly greater (F = 3.29; df = 4, 30; P = 0.02) than expectorated by mosquitoes feeding on birds on days 1, 2, and 3, and this pattern was consistent between species because the interaction term in the ANOVA was not significant (P > 0.05) (Figs. 2c and 3c).

Discussion

Our data indicated that 2–10-d old mourning doves produce viremia titers >4 log₁₀ PFU per 0.1 ml for >5 d PI, readily infect Cx. tarsalis females that feed upon them during this period and therefore are effective amplifying hosts for SLE virus. This finding is epidemiologically important, because mourning doves are abundant throughout California as well as the rest of the United States and have from two to four broods per year (Gough et al. 1998). Doves in our colonies, for example, produced offspring from March through October, and this reproductive pattern, in combination with mild environmental temperatures (Reisen et al. 1993), could delineate the effective transmission season for SLE virus. Therefore, even though the viremia response in adult mourning doves is minimal (Reisen et al. 2003c), ample immatures should be available for continued virus amplification for most of the year. Our results agreed with previous reports of immature mourning dove experimental infections in Florida (Jennings 1969), but they differed markedly from our studies with pigeon, Columbia livia, squab <7 d of age (Reisen et al. 1992). In the latter study, two of five squab developed fleeting viremias <1 log₁₀ PFU/0.1 ml on days 3 and 5 PI that most likely were too low to infect mosquitoes. In the current study, 67% of Cx. tarsalis feeding on viremic doves became infected, and this value increased to 100% on day 5 (n = 10) when viremia estimates were >6 log₁₀ PFU/0.1 ml, after which sampling was terminated.

Adult house finches develop elevated viremias of 1–5-d duration after infection with SLE virus, and most Cx. tarsalis mosquitoes feeding on these birds during this period become infected. House finches also are widespread throughout the United States, abundant, and produce from one to two broods per year (Gough et al. 1998). The current study demonstrated that 70% of Cx. tarsalis females feeding on nestlings 2–12 d of age 1–5 d PI became infected and that this infection rate increased to >90% on days 2–3 PI when viremia, extrapolated for virus ingested by mosquitoes, peaked at >5 log₁₀ PFU/0.1 ml. Peak SLE viremia estimated here for nestlings was somewhat higher than reported by us for adult birds (Reisen et al. 2000b, 2003a,b), but similar to that observed for immunosuppressed adults (Reisen et al. 2003a). Impaired immunocompetence in adult birds may be similar to the immature immune system described here in that it enables viremia production.

SLE virus titers in female Cx. tarsalis after 11–13 d of incubation at 26°C ranged from 1.3 to 6.9 log₁₀ and were correlated with the amount of virus ingested in their blood meal. This was extremely variable, and females ingesting from 1.1 to 3.9 log₁₀ PFU of virus developed body titers >4 log₁₀ PFU. The resultant body titers were important epidemiologically, because only those females with body titers >4 log₁₀ PFU transmitted virus per os when evaluated by the in vitro capillary tube method. Similar to our previous reports (Reisen et al. 2000a), the quantity of virus expectorated averaged 1.5 log₁₀ PFU, but ranged from 0.8 to 3.4 log₁₀ PFU. Adult house finches inoculated with this range of infectious SLE virus produced a similar magnitude viremia response, but the timing of the peak viremia titer was altered by 1 d (Reisen et al. 2000a).

The overall percentage of infected females that transmitted virus was 27%, increasing to 31% if only females with body titers >4.6 log₁₀ PFU were included. This percentage transmission was considered to be low for the HVP colony; however, the body titers in the current study also were considerably lower than observed previously (Reisen et al. 1993). In the previous study, body titers for transmitting females ranged from 5.8 to 8.6 log₁₀ PFU and for nontransmitting females ranged from 1.3 to 7.8 log₁₀ PFU per mosquito after incubation at 26°C. Differences may have related to genetic changes within the HVP colony during the 10-yr intervening period. Elongating the time period or using warmer temperatures during incubation also may have increased body titers and therefore the percentage transmitting (Hardy and Reeves 1990a).

The current data clearly indicate that nestling mourning doves and house finches are competent hosts for SLE virus, supporting previous studies with other species (McLean and Bowen 1980). Interestingly, however, the phenology of SLE virus transmission in SE California does not support the role of nestlings as being critical in early season virus amplification (Reisen et al. 1995, 2002). Here, nesting activity begins in March for both species, but SLE virus typically is not detected until July, after the first and perhaps second broods of nestlings have fledged and left the nest. It may be that hot summer temperatures combined with continued reproductive activity by these and later breeding species such as Gambel’s quail are critical for amplification of virus to levels detected by our surveillance methods.

Acknowledgements

We thank V. M. Martinez, B. D. Carroll, and J. Dobson, Center for Vectorborne Diseases, for technical assistance. This research was funded, in part, by Research Grants R01-AI39483 and R01-AI47855 from the National Institute of Allergy and Infectious Diseases. Logistical support was provided by the Kern Mosquito and Vector Control District.

References Cited