West Nile and St. Louis Encephalitis Virus Antibody Seroconversion, Prevalence, and Persistence in Naturally Infected Pig-Tailed Macaques (Macaca nemestrina)

Abstract

Pig-tailed macaques (Macaca nemestrina) naturally infected with West Nile virus were monitored from 1999 to 2005 to determine virus-specific antibody seroconversion, prevalence, and persistence. Antibodies persisted for up to 36 months, as detected by epitope-blocking enzyme-linked immunosorbent and hemagglutination inhibition assays. Exposure to cocirculating St. Louis encephalitis virus was evaluated by Western blotting and immunofluorescence assays.

The seroepidemiology of wild macaque populations demonstrates a natural seroconversion to flaviviruses including West Nile virus (WNV) and Japanese encephalitis virus (10, 15). Across species, there is a wide variation in WNV antibody persistence. Immunoglobulin M antibodies may persist less than 3 weeks in fowl, a sentinel species, and up to 500 days in humans, a dead-end host (11, 19, 24). Domestic pigs have remained seropositive (hemagglutination inhibition assay [HAI] antibodies) and immune to both WNV and Japanese encephalitis virus for more than 3 years (7, 9). In some cases, antibody persistence in the form of a recurrent, but intermittent, appearance of specific plaque reduction neutralization test (PRNT) antibodies after a single exposure to virus has been attributed to a host immune reactivation of persistent virus (13, 21). Neutralizing antibody persistence for more than 5 months due to a single experimental inoculation of macaques (intracerebral or subcutaneous) without recurrence has been documented (16, 17).

West Nile virus emerged in Louisiana during the spring of 2002, with endemic establishment and year-round regional virus activity reported during the winter months of 2003 to 2004 (12, 23). WNV and St. Louis encephalitis virus (SLEV) share a close phylogenetic relationship and elicit cross-reactive antibodies (5). Both viruses are maintained in similar transmission cycles with mosquito vectors (mainly Culex spp.), avian reservoirs, and amplification hosts. St. Louis encephalitis virus has been endemically established in the southern United States for many decades. Just prior to the emergence of WNV in Louisiana, a regional outbreak of SLEV took place (fall of 2001); SLEV and WNV now cocirculate within the region. Serological distinction between flaviviruses is hampered by the extensive homologies of viral structural proteins. West Nile virus infection exhibits low pathogenicity compared to many other members of the flavivirus family while producing broad-spectrum immunity (14). Sequential infections with one or more flaviviruses elicit strong cross-reactive anamnestic responses, which may confer immunity (22).

The Washington National Primate Research Center (WaNPRC) and the Tulane National Primate Research Center (TNPRC) house animals in the same outdoor breeding cohorts. The WNV seroprevalence in TNPRC animals was shown to be greater than 30% (18). We therefore aimed to assess the WNV exposure level among WaNPRC animals through serologic techniques (enzyme-linked immunosorbent assay [ELISA], immunofluorescence assays [IFAs], PRNT, and HAI). Since repeated exposures, persistent antibody response, and exposure to cocirculating flaviviruses can complicate the interpretation of serologic results, we also aimed to establish methods to discern between WNV and SLEV infections.

Of the cross-reactive viral structural proteins, the envelope (E) protein is the most immunogenic (25). The E protein is highly conserved among flaviviruses and elicits an antibody response with relatively low specificity. Cross-reactivity is confounding to assays that depend to a great extent on antibodies to the E protein, for example, HAI. This may be overcome, at least partially, by employing assays based on nonstructural (NS) proteins, which elicit more virus-specific immune responses.

A colony serosurvey was performed using a WNV epitope-blocking ELISA validated in multiple avian and mammalian species including the pig-tailed macaque (Macaca nemestrina) (3, 4, 8). ELISAs were performed using monoclonal antibody 3.112G (Chemicon International, Inc., Temecula, CA), which is specific for the NS1 protein of Kunjin virus, a subtype of WNV (7, 20). Immunoassays were performed with banked (−70°C and −20°C) plasma samples collected from 1999 to 2005 for viral screening from approximately 700 WaNPRC pig-tailed macaques housed at the TNPRC in Covington, LA. Negative serum samples were obtained from WaNPRC animals born and housed indoors in Seattle, Washington. Western blot analysis was performed using infected Vero cell lysates (ChimeraVax SLEV and WNV; Acambis Int., Cambridge, MA) as previously described (2). Immunofluorescence assays (PanBio, Inc., Columbia, MD) were performed according to the manufacturer's specifications. Viral screening was performed under a general husbandry protocol approved by the University of Washington Institutional Animal Care and Use Committee. Both the WaNPRC and the TNPRC are AAALAC-accredited facilities. Hemagglutination inhibition assays were performed at the University of Texas Medical Branch using a previously published protocol (6). Previously obtained PRNT (Arthropod-Borne Infectious Disease Laboratory, Colorado State University) and HAI (University of Texas Medical Branch) data were used for correlation analysis (8).

Plasma samples collected at 6-month intervals from 1999 to 2005 were tested by ELISA. Samples from 2002 and 2003 were tested by HAI. No WNV-specific antibodies were detected in samples from 1999, 2000, or 2001. WNV antibodies were demonstrated in colony animals from 2002 to 2005.

The distribution of seroconversion and maintained titers among seropositive animals is summarized in Fig. 1. Seroconversion rates are a close approximation, as a small number of animals leave and enter the WaNPRC colony each year (7% average animal turnover from 2002 to 2004). Additionally, a small percentage of samples received were unsuitable for testing (0.5% in 2002, 1.1% in 2003, and 0.6% in 2004).

FIG. 1.

Distribution of seroconversion and maintained titers among seropositive animals surveyed from 2002 to 2005. The black bars represent the percentage of seroconversions; the dashed bars represent the percentage of maintained titers. In 2002, 40 previously naive animals developed WNV titers. In 2003, 69 animals were seropositive for WNV: 50 animals (72%) seroconverted, and 19 (28%) maintained ELISA titers from 2002. In 2004, 60 animals were positive for WNV: 25 animals (42%) seroconverted, and 35 (58%) maintained antibody titers (2002 or 2003). In 2005, 11% of the animals tested demonstrated WNV seroconversion by ELISA.

This serosurvey demonstrated that under natural conditions of repeated environmental exposure, WNV-specific antibodies may persist considerably longer in M. nemestrina than previously reported (16, 17). In the 2002 to 2005 colony serosurvey, 92 animals had NS1 antibody titers that persisted for 6 to 12 consecutive months. Antibody persistence of 18 months' duration was demonstrated in 13 animals, while persistence of 24 months' duration was demonstrated in 4 animals, persistence of 30 months' duration was demonstrated in 16 animals, and persistence of 36 months' duration was demonstrated in 14 animals. Titers that are maintained for more than 1 year affect the accuracy of seroprevalence estimates, as new cases must be differentiated from maintained titers to document accurate viral transmission (1). Over the course of 36 months, antibody titers slowly declined in some animals (Fig. 2A) and fluctuated in others (Fig. 2B to D). Fluctuations in ELISA titers may be related to intermittent environmental WNV exposure or to anamnestic responses to infection with SLEV. The Louisiana State Department of Health monitors arboviral activity (Arbonet) and reported SLEV-positive birds within the region of the TNPRC in 2001, with continued SLEV activity (avian, mosquito, and human) through 2005 (http://arbonet.caeph.tulane.edu/). Western blotting was performed to distinguish between repeated infections with WNV (Fig. 3) and coinfection with SLEV (Fig. 4).

FIG.2.

ELISA results from 10 representative WNV-seropositive animals illustrating four possible reasons for persistent titers. (A) Viral exposure prior to the fall of 2002 with a gradual decline in antibody titers over the following 36 months. (B) Viral exposure prior to the fall of 2002 and a second exposure in 2004, evidenced by a rise in ELISA titer. (C) Viral exposure prior to the fall of 2002 and a second exposure in 2005, evidenced by a rise in ELISA titer. (D) Initial viral exposure during the winter of 2002 to 2003 with multiple subsequent exposures in 2003 and 2005.

FIG. 3.

Western blot illustrating multiple exposures to flavivirus. Four separate columns of WNV Vero cell lysates (reducing gel) reacted with four serum samples collected from one animal at 6- and 12-month intervals. Two distinct flaviviral protein bands are present on Western blots: NS5 protein (90 to 100 kDa) and a merging band of the E protein (50 kDa) and the NS1 protein (45 kDa). The E protein elicits the strongest antibody response. No reaction was seen on blots of noninfected Vero cells (not shown). The first column shows Precision Plus Protein Standards (Bio-Rad) run on a 10% Tris-HCl gel. Column 2 (t = 0) represents flavivirus reactivity due to a primary exposure. Column 3 (t = 6 months) represents a recent, second flaviviral exposure. Column 4 (t = 18 months) demonstrates decreased signal strength, correlating with falling ELISA titers. Column 5 (t = 30 months) represents a third exposure to a flavivirus.

FIG. 4.

Two Western blots illustrating the distinction between WNV and SLEV infection. Each blot consists of two columns; in the left column, serum proteins are reacting to WNV-infected cell lysates (reducing gel), while in the right column, serum proteins are reacting to SLEV-infected Vero cell lysates (reducing gel). The first column shows Precision Plus Protein Standards (Bio-Rad) run on a 10% Tris-HCl gel. Blot 1, serum from an animal infected with SLEV. The E protein signal is much stronger in the SLEV (right) column than in the WNV (left) column. Blot 2, serum from an animal infected with WNV. The E protein signal is present in the WNV (left) column and absent in the SLEV (right) column.

Plaque reduction assays are the “gold standard” for definitive viral diagnosis; however, these assays require a high level of biosecurity, which may not be readily available. When PRNTs are run in parallel for cocirculating, cross-reactive viruses, the specificity of each assay is increased. In lieu of PRNTs, a Western blot assay may be used for comparison of responses to two or more viruses. Using this approach, Western blots for WNV and SLEV were compared with WNV PRNT titers, ELISA titers, and both SLEV and WNV IFAs. West Nile virus antibody titers measured by ELISA correlated with PRNT titers with 93% sensitivity and 100% specificity (8). Western blot banding patterns were found to correlate with the NS1 ELISA (88% concordance), PRNT (76% concordance), and IFA (66% concordance) data, demonstrating the ability of Western blots to distinguish between WNV and SLEV (Table 1).

TABLE 1.

SLEV and WNV reactivities^a

Animalb	WNV PRNT titer	% NS1 MAb inhibitionc	WNV HAI titer	IFA scored		Western blot banding patterne
				WNV	SLEV
1	0	0	Not available	0	0	No bands
2	20	4.7	1:80	4	3	WNV = SLEV
3	40	0	0	1	1	WNV = SLEV
4	40	11.8	0	2	3	WNV = SLEV
5	80	18.5	0	2	3	WNV < SLEV
6	160	5.1	1:40	3	1	WNV = SLEV
7	160	18.7	1:20	3	0	WNV > SLEV
8	160	40.5	1:80	5	1	WNV = SLEV
9	160	46.0	1:20	3	3	WNV > SLEV
10	160	76.0	1:20	3	0	WNV = SLEV
11	320	22.9	1:40	2	0	WNV > SLEV
12	320	23.0	1:40	2	1	WNV > SLEV
13	320	50	Not available	3	0	Not available
14	640	32.7	1:160	5	4	WNV ≫ SLEV
15	640	62.5	1:320	2	3	WNV > SLEV
16	>1,280	56.2	1:160	4	1	WNV > SLEV
17	>1,280	68.9	1:40	5	2	WNV > SLEV
18	>1,280	92.5	1:1,280	4	3	WNV > SLEV

^aIFA and HAI detect cross-reactive antibodies (WNV and SLEV). NS1 ELISA, Western blottng, and PRNT are specific for WNV.

^bCohort chosen to represent a range of HAI and NS1 ELISA titers.

^cMAb, monoclonal antibody.

^dIFA score, 0 (no fluroescence) to 5 (strong florescence).

^eWestern blot banding pattern is a measurement of the relative signal strength of the E protein (50 kDa).

In 2002, 63 animals were positive by HAI and negative by ELISA, suggesting exposure to a cocirculating flavivirus. In no case did an HAI-negative animal test positive for WNV by ELISA. Discordant samples were evaluated by Western blot assay. Four of the 63 animals positive by HAI and negative by ELISA (in 2002) were found to have strong SLEV-specific reactivity; the remaining animals (59/63 animals) demonstrated an SLEV band intensity equal to or less than that for WNV.

When developing flaviviral monitoring programs for macaques in zoos and primate centers with outdoor facilities, complications in data interpretation from persistent antibody titers must be considered. Persistent titers may be related to continued flaviviral environmental exposure and/or the presence of cocirculating flaviviruses. We have demonstrated that macaques can maintain WNV seropositivity by ELISA for more than 36 months with continual environmental exposure. Western blots demonstrated repeated flavivirus exposure and discerned between WNV and SLEV exposures.