Encephalitis Virus Persistence in California Birds: Experimental Infections in Mourning Doves (Zenaidura macroura)

Abstract

After-hatching and hatching year, mourning doves were infected by inoculation with either western equine encephalomyelitis (WEE) or St. Louis encephalitis (SLE) viruses; some birds in each group also were treated with the immunosuppressant cyclophosphamide before and during infection. Cyclophosphamide treatment significantly increased the WEE viremia but did not alter the antibody response. In contrast, cyclophosphamide-treated and -untreated doves did not develop a detectable SLE viremia but became antibody positive. Antibody peaked at 10 wk after inoculation for both viruses and remained detectable in most birds throughout the 26-wk study. When treated with cyclophosphamide the following spring, birds did not relapse and develop a detectable viremia. Previously infected birds were protected when challenged with conspecific virus (i.e., none produced a detectable viremia), but there was no anamnestic antibody response to reinfection. In agreement with our failure to detect relapses, all birds were negative for viral RNA when sera, spleen, lung, and kidney tissues were tested by reverse transcriptase-polymerase chain reaction after necropsy. Our results indicated that adult mourning doves were an incompetent host for SLE virus and probably do not serve as a suitable overwintering or dispersal host for either WEE and SLE viruses.

Our research recently has focused on the role of wild birds in the persistence of western equine encephalomyelitis (WEE, Togaviridae, Alphavirus) and St. Louis encephalitis (SLE, Flaviviridae, Flavivirus) viruses in California (Reeves 1990a). Initial studies on house finches in the order Passeriformes (Reisen et al. 2001) indicated that, similar to previous studies (Reeves 1990b), experimental WEE and SLE infections occasionally resulted in the long-term retention of virus (or viral RNA) in multiple organs at low titer. However, attempts to initiate relapse and infect mosquitoes the following spring were unsuccessful, even after immunosuppression. In this study, we attempted to establish chronic infections in mourning doves (Zenaidura macroura) in the order Columbiformes to determine if the results with house finches were generally applicable to California birds or if selected taxa may be more effective overwintering hosts for these endemic encephalitis viruses. We selected mourning doves because (1) they are abundant and are repeatedly serologically positive when WEE or SLE are active (Milby and Reeves 1990, Reisen et al. 2000c), (2) adults produce WEE viremias >2 log₁₀ PFU per 0.1 ml, the minimal titer necessary to infect Culex tarsalis Coquillett mosquitoes (Hardy and Reeves 1990a), (3) recent experiments documented that they occasionally establish chronic infections detectable in their peripheral blood stream (Reisen et al. 2003c), and (4) mourning doves move considerable distances and may exchange migration flyways (Kramer et al. 1997), thereby becoming important in virus dispersal as well as persistence. Studies with house finches indicated that cyclophosphamide treatment concurrent with virus inoculation enhanced viremia and perhaps the establishment of chronic infections (Reisen et al. 2003b). In this study, doves were treated with cyclophosphamide to enhance infection and again to stimulate the relapse of possible chronic infections.

Materials and Methods

Experimental procedures and assays followed those in previous studies (Kramer et al. 2002, Reisen et al. 2003b) and are described briefly below.

Birds

Hatching and after-hatching year doves were collected at an orchard near Bakersfield, Kern County, CA, during the week of 24 September 2001 and held in an outdoor aviary where they were given drinking water medicated with metronidiazole (400 mg/liter; Zenith Goldline Pharmaceuticals, Miami, FL), teramyacin (250 mg/liter; Pfizer Animal Products, Exton, PA), and piperazamide (7.5 g/liter; RX Veterinary Products, Grapevine, TX). Doves were fed mixed wild birdseed (millet, sorghum, and sunflower seeds).

Virus

The 1989 Kern217 strain of SLE and the 1983 Kern5547 strain of WEE were used throughout. Both viral strains isolated from Cx. tarsalis collected near Bakersfield were at Vero cell passage 1 or 2 at experimentation and produced elevated viremias in house finches (Reisen et al. 2000b). Doves were infected initially on 22 October 2001 by subcutaneous syringe inoculation with 3.3 WEE and 2.5 SLE log₁₀ plaque forming units (PFU) in the cervical region; subsequent challenge doses administered on 3 June 2002 were 2.0 and 2.3 log₁₀ PFU, respectively. Previous studies indicated that these doses fell within the range of virus delivered by mosquito bite and that WEE and SLE infection by mosquito bite and syringe inoculation produced comparable viremia and antibody responses in house finches (Reisen et al. 2000a).

Experimental Design

Before experimentation, all birds were bled by jugular puncture or from the wing vein to determine previous infection. For antibody determination, 0.1 ml blood was diluted into 0.9 ml physiological saline, clarified by centrifugation, and stored at −80°C. Birds were transferred to avian infection units, where two birds were maintained per cage. On days −5, −3, and 0 (22 October 2001, when all birds were infected with either WEE or SLE virus as described above), we inoculated eight birds per virus group in the breast muscle with cyclophosphamide (40 mg/kg); seven birds per virus group were infected concurrently but not treated with cyclophosphamide. Cyclophosphamide (Cytoxan; Mead Johnson, Princeton, NJ) doses were similar to those given previously to immunosuppress house finches (Reisen et al. 2003b) and chickens (Fulton et al. 1996), but were approximately two times higher than administered intraperitoneally to immunosuppress European starlings (Trust et al. 1994). An additional 12 birds were not inoculated or infected and were held as untreated controls. Birds infected with WEE and SLE were bled on days 1 and 2 or on days 2 and 3 after inoculation (PI), respectively, to detect viremia. Previous studies have shown that viremia typically peaks in mourning doves on these days (Reisen et al. 2003c). For viremia samples, 0.1 ml of whole blood was mixed immediately with 0.4 ml of virus diluent containing antibiotics and bovine serum albumin and frozen at −80°C. Four weeks later, when birds were no longer infectious, they were moved into a screened outdoor aviary. To monitor antibody levels, birds were bled on 19 November 2001 (4 wk PI), 7 January 2002 (10 wk PI), 4 March 2002 (18 wk PI), and 4 May 2002 (26 wk PI).

Surviving birds were divided into four treatment groups: VC, cyclophosphamide immunosuppressed, virus infected, and immunosuppressed the following spring; VV, virus infected without immunosuppression + virus reinfected the following spring (reinfection control); 0V, positive infection control, not previously infected or immunosuppressed and infected for the first time; and 00, negative control (untreated and uninfected). Birds in the VC groups were inoculated intramuscularly with cyclophosphamide (40 mg/kg body weight) on days −2, 0, and +2 before and after the day when positive controls were inoculated with virus on day 0 (3 June 2002, see above for challenge doses). Viremia was measured on days 1–5 PI and antibody on weeks 1–4 PI. On 2 July 2002, surviving doves were necropsied, and sera, and lung, spleen, and kidney tissues were screened for viral RNA by reverse transcriptase-polymerase chain reaction (RT-PCR).

Assays

Viral and serological assays were done under BL-3 conditions at the Center for Vectorborne Diseases Arbovirus Laboratory using methods described previously (Kramer et al. 2002, Reisen et al. 2003b). Blood samples for viremia determination were diluted 1:5 by volume in phosphate buffered saline (PBS) plus 20% fetal bovine serum (FBS) and antibiotics (100 U penicillin, 100 U streptomycin, and 200 U nystatin), frozen immediately on dry ice, and stored at −80°C. Samples were tested for virus by plaque assay on Vero cells (Reisen et al. 1993). Values listed as 0 had ≤2 PFU at a dilution of 1:5. Antibody levels were estimated by enzyme immunoassay (EIA) and/or plaque reduction neutralization test (PRNT) from 0.1 ml blood collected by jugular puncture and diluted into 0.9 ml saline (Chiles and Reisen 1998). To be positive for antibody by EIA, the ratio of the mean optical density of two antigen positive wells over the antigen negative well had to be >2.0. To be positive by PRNT, sera had to inhibit the formation of >80% of plaques at a dilution of ≥1:20.

Tissues taken at necropsy were tested for WEE and SLE RNA using RT-PCR. Approximately, 30 mg (3 by 3 by 3 mm) of tissue was triturated in Trizol, and RT-PCR was conducted using the OneStep RT-PCR Kit (Qiagen, Carlsbad, CA), according to manufacturer protocols. WEE-specific primers, 2471 forward and 3055 reverse (Molecular Biology Insights, Cascade, CO), produced a 584-bp fragment of the E1 region. For SLE, samples were tested by amplifying the E region of the viral genome with primers 1500v forward and 2315c reverse, which produced an 816-bp fragment (Kramer et al. 2002).

Results

One after-hatching year dove bled before experimentation on 15 October 2001 had detectable antibody, indicating that it previously had been infected with WEE. This bird had an EIA p/n ratio of 6.35 against WEE antigen and was the only bird treated with cyclophosphamide and inoculated with WEE that failed to produce a detectable viremia. On 4 May 2002, new control birds were added to the experiment and prebled. One control bird previously was infected with SLE and had a EIA p/n ratio of 8.72 and a PRNT titer of 1:40. A second SLE control bird was negative at prebleed (EIA p/n = 1.03), had no detectable viremia on days 1–5, and was negative for antibody at week 1 (EIA = 1.68), but seroconverted on week 2 (EIA p/n = 6.41, PRNT = 1:40). These data indicated bird-bird or fecal-oral transmission within cages. These three birds were deleted from further analysis in this experiment. Prebleed EIA p/n ratios for the remaining birds averaged 1.4 (range, 0.48–2.21), indicating that they had not been infected previously.

Effect of Cyclophosphamide on Infection

WEE viremias were compared by two-way model I analysis of variance (ANOVA) with cyclophosphamide treatment and days PI as main effects (Table 1). The mean viremia for birds inoculated with WEE and cyclophosphamide was significantly greater (F = 75.6, df = 1, 24, P < 0.001) than for birds inoculated only with WEE; there was no difference (P > 0.05) between mean viremias measured on days 1 and 2 PI, and the interaction effect was not significant (P > 0.05). These data indicated that cyclophosphamide treatment enhanced WEE viremia levels. In marked contrast, none of the doves inoculated with SLE produced a viremia detectable on days 2 and 3 PI, the time when SLE viremia generally peaks in mourning doves (Hardy and Reeves 1990b, Reisen et al. 2003c).

The effects of time after inoculation and cyclophosphamide treatment on antibody levels measured by EIA p/n ratios were analyzed by two-way ANOVA (Table 1). Overall mean EIA levels were not significantly different between cyclophosphamide-treated and -untreated groups (P > 0.05). Combined EIA values were significantly greatest at 10 wk PI and declined for the next 6 mo (F = 13.4, df = 3, 40; P < 0.001). A slightly different pattern was observed for PRNT titers (Table 1). Mean PRNT titers were marginally higher (F = 3.6, df = 1, 39; P = 0.067) among doves treated with cyclophosphamide than among doves not immunosuppressed. Combined PRNT titers also declined marginally over time (F = 3.70, df = 3, 39; P = 0.02), being lowest on weeks 18 and 26 after inoculation.

Cyclophosphamide did not significantly alter either the EIA or PRNT antibody response (P > 0.05) after SLE inoculation (Table 1). Combined EIA p/n ratios for antibodies against SLE were highest on week 10 PI (F = 45.2; df = 3, 37; P < 0.001); ratios for weeks 4 and 18 were not significantly different but increased on week 26 (Table 1). PRNT titers also were significantly highest on week 10 (F = 4.30; df = 3, 37; P = 0.011).

Reinfection

Previous infection prevented a detectable viremia in doves challenged with either WEE and SLE (Table 2). In contrast, all three WEE, but not SLE, positive control birds (OV group) produced a moderate but detectable viremia on days 1–3 PI. Treatment with cyclophosphamide did not result in a relapse, as measured by viremia in six and eight birds in the WEE and SLE VC groups, respectively (Table 2).

Inoculation of doves with WEE or SLE in the VV groups did not result in a detectable anamnestic response. In fact, both EIA and PRNT antibody levels against WEE and SLE in the VV groups remained similar to or significantly less than the antibody levels in the 0V control and unchallenged VC groups (Table 2). These data indicated that cyclophosphamide treatment did not prevent or eliminate humoral immunity. Serum specimens from some birds infected with SLE were compromised after EIAs were completed, precluding PRNT.

Necropsy

None of the 19 and 20 birds infected with WEE and SLE, respectively, and necropsied >4 wk after treatment or infection tested positive for viral RNA by RT-PCR. These totals included three and two birds that died during the winter >18 wk after infection, three and three positive virus controls, and five negative controls (including the two positive for SLE). Although cyclophosphamide treatment enhanced infection with WEE by producing elevated viremia titers, these enhanced infections did not result in chronic infections, that were detectable by testing lung, kidney, spleen, and blood tissues.

Discussion

Our results with mourning doves differed markedly from those obtained for house finches using similar protocols (Reisen et al. 2001, 2003a, b), even though both species were collected near Bakersfield and were infected with the same sympatric strains of WEE and SLE viruses. In house finches, PRNT, and to some extent EIA, antibody levels to both WEE and SLE declined after infection and were rarely detectable the following spring, whereas in the current studies, both WEE and SLE antibodies were detectable in mourning doves for >26 wk, even after cyclophosphamide treatment. Challenge of house finches with WEE and SLE resulted in a rapid and significant anamnestic response (Reisen et al. 2001, 2003a), whereas antibody levels in virus challenged mourning doves did not differ from new infections. Lack of a detectable anamnestic response may not be a general attribute of Columbiform immunology, because rock doves (Columbia livia) challenged with a second infection of both WEE and SLE produced a significant anamnestic response, measured by hemagglutination inhibition assays (Reisen et al. 1992). Immunity levels in all three species, as well as house sparrows (Passer domesticus) (McLean et al. 1983), were protective, and these species did not produce viremia after challenge with same viral strains.

Results of our experiments continue to question the role of avian chronic infection as the mechanism for encephalitis virus persistence and long-range dispersal. Mourning doves are markedly dispersive and frequently exchange flyways during migration (Kramer et al. 1997). Doves banded in Bakersfield, for example, have been captured over 1,000 km south in Mexico (McClure et al. 1962) and in Daytona Beach, FL (USGS bird band recovery report). However, similar to previous studies (Reeves 1990b), none of the 39 doves had a detectable chronic infection at necropsy. Negative results also were obtained after testing kidneys from doves shot in Coachella Valley after a summer of intense SLE transmission (Reisen et al. 2002). Even if infections were to persist, the immune response after relapse would seem adequate to prevent viremia of sufficient titer to infect mosquitoes.

Although frequently collected as SLE antibody–positive in nature (Gruwell et al. 2000, Reisen et al. 2000c), adult mourning doves rarely produced an SLE viremia of sufficient titer to infect Culex mosquitoes (Hardy and Reeves 1990a, b, Reisen et al. 2003c) and therefore should be classified as an incompetent host for SLE. However, like many other species of birds, nestling mourning doves produced elevated viremia titers sufficient to infect mosquitoes (unpublished data), and it is this age group that would seem important in viral amplification. In California, mourning doves have a long multi-brooded nesting season (our breeding colony produces offspring from March to November), perhaps assuring the presence of competent hosts throughout the mosquito season. In contrast to SLE, adult doves were a competent host for WEE and produced elevated viremias without associated mortality. Our preliminary studies have indicated that WEE infections of nestlings may result in death shortly after inoculation. Therefore, WEE transmission would seem to rely more on immunologically competent adult than incompetent immature birds as a reservoir host.

Acknowledgements

We thank B. Carroll and J. Dobson, Arbovirus Field Station, and S. Garcia, S. Clark, and T. Jenkins, Arbovirus Laboratory, for excellent technical support. This research was funded, in part, by National Institute of Allergy and Infectious Diseases Research Grants 1-R01-AI39483 and 1-R01-AI47855. Logistical support was provided by the Kern Mosquito and Vector Control District.

References Cited