Development of a Fluorescence Polarization-Based Diagnostic Assay for Equine Infectious Anemia Virus

Abstract

The control of equine infectious anemia virus (EIAV) infections of horses has been over the past 20 years based primarily on the identification and elimination of seropositive horses, predominantly by a standardized agar gel immunodiffusion (AGID) assay in centralized reference laboratories. This screening for EIAV-seropositive horses has been to date hindered by the lack of a rapid diagnostic format that can be easily employed in the field. We describe here the development of a rapid solution-phase assay for the presence of serum antibodies to EIAV based on fluorescence polarization (FP) (patent pending). Peptides derived from antigenic regions of EIAV core and envelope proteins were initially screened for their utility as probes in an FP assay to select the best peptide antigen candidates. The FP assay was optimized to detect the presence of EIAV-specific antibodies by a change in the FP of a fluorescein-labeled immunoreactive peptide diagnostic antigen. The most sensitive and specific peptide probe was a peptide corresponding to the immunodominant region of the EIAV transmembrane protein, gp45. This probe was tested for its reactivity in the optimized FP assay with 151 AGID-positive horse sera and 106 AGID-negative serum samples. The results of these studies demonstrated that the FP assay reactivity correlated with reported AGID results in 106 of 106 negative serum samples (100% specificity) and in 135 of 151 positive serum samples (89.4% sensitivity). The FP assay was also found to have a very low background reactivity and to readily detect antibodies produced early in infection (≤3 weeks postinfection). The developed EIAV FP assay is rapid (5 to 20 min) and simple to perform and is equally suitable for use both in the field and in the diagnostic laboratory, thus providing the basis of an improved commercial diagnostic assay for EIAV infection of horses.

Equine infectious anemia virus (EIAV) is a lentivirus, genetically related to human immunodeficiency virus type 1, that infects horses worldwide (for a recent review, see reference 13). It causes a chronic disease characterized by a period of cyclic fevers and viremia, followed by clinical quiescence. The animals generally survive this disease but remain infected, becoming lifelong inapparent carriers; they appear to be healthy but in fact still harbor infectious virus that poses a threat of transmission to other horses. There are thousands of EIAV-positive horses in the United States; most of them reside in the hot zone, a group of 18 states along the Gulf of Mexico and in the Mississippi River valley (5). The disease is most frequent in this region due to the climate and associated prevalence of horseflies, the major vector of transmission of EIAV. To control the spread of this virus, horses are routinely tested for EIAV-specific serum antibodies before being allowed into shows, used for breeding, or crossing state lines. If a horse is found to be seropositive, it must be either euthanatized or quarantined with a 200-yard barrier for the rest of its life, depending on local governmental regulations. However, because testing is not yet mandatory for all horses, it is estimated that over 80% of U.S. equids have never been tested; the actual prevalence and potential reservoir of EIAV remain uncertain. Efforts are underway to encourage, and in some states mandate, testing of all equids to better control this disease and reduce the rate of infection (5).

EIAV-infected animals mount a vigorous immune response to the viral infection. This results in reduction of viremia during clinical quiescence to very low, often undetectable, levels. This immune response is characterized by high-titer antibodies directed to three major viral antigens: the envelope glycoproteins, surface unit (gp90) and transmembrane (gp45), and the capsid protein or core antigen (p26). Due to the presence of high levels of antibody and low levels of virus during most of the disease course, diagnostic assays have focused on detection of viral antibodies. Current U.S. Department of Agriculture-approved diagnostic assays for EIAV include agar gel immunodiffusion (AGID) (4), competitive enzyme-linked immunosorbent assay (C-ELISA), and synthetic-antigen ELISA. The first two assays detect antibodies to the major core protein p26, which has a well-conserved structure but is a relatively poor immunogen compared to the envelope proteins, gp90 and gp45. The synthetic-antigen ELISA detects antibodies to gp45 but reportedly may have a lower sensitivity than AGID (8). The major drawbacks of the AGID test are the length of time that it takes to test the samples and the technical difficulty in interpreting the results. ELISA-based tests can be completed in several hours, but in a recent study, the C-ELISA had a 2% false-positive rate, raising concerns about the unnecessary destruction of healthy horses (8).

A more comprehensive screening for EIAV-positive horses could be facilitated by the development of more rapid and simple diagnostic assays that can be readily used under field conditions. Fluorescence polarization (FP) has been used as a tool to monitor protein-protein, protein-peptide, and other intermolecular interactions (9). Most fluorophores, including fluorescein, emit light in the same direction in which it is absorbed. When a fluorophore is freely rotating in solution, the light is emitted in all directions as a result of the molecule's rotation during the lifetime of the fluorescence emission; it is nonpolarized. If, however, the fluorophore is part of a slowly rotating molecule (one that is large or in a viscous environment), the molecule does not rotate quickly with respect to the lifetime of the fluorescence, and the emission will occur in roughly the same direction as the absorption; it is therefore polarized. This property of fluorescence can therefore be used to distinguish small molecules (e.g., fluorescently labeled peptides) from large ones (e.g., peptide bound to antibody). Relatively recent advances in instrumentation have allowed the use of this phenomenon to develop rapid immunoassays for a large number of analytes including therapeutic drugs (10) and metabolites (18) as well as antibodies to infectious agents (12, 14, 15). These assays can be performed in a matter of minutes (versus hours or days for the other tests) and usually do not require extensive sample preparation. In addition, the materials required for the assay are relatively simple and highly stable, making this technique attractive for field use.

In light of the need for a more rapid assay that can be used in the field to detect EIAV-infected horses, we pursued FP as a medium on which to develop a new diagnostic test for serum antibodies to EIAV. We selected, synthesized, and evaluated several candidate peptides for their ability to detect the presence of antibodies to EIAV core and envelope proteins. This investigation led to the development of an FP-based assay which is rapid and possesses both high sensitivity and very high specificity for EIAV-specific antibodies in serum samples derived from experimentally and field-infected horses from diverse geographic locations in the United States.

MATERIALS AND METHODS

Horse sera.

Sera from EIAV-field-infected and uninfected horses were generous gifts from the Texas Animal Health Commission, the Missouri Department of Agriculture, and the University of Kentucky (Utah, Florida, and Oklahoma field-infected sera). Prior to use and after a freeze-thaw cycle, the sera were centrifuged at 12,000 × g for 2 min to pellet any precipitated protein.

Peptide design.

Candidate peptides were chosen based on previous studies that defined regions of broadly reactive antigenicity of the p26 capsid (3), gp45 transmembrane (2), and gp90 surface unit (1) proteins of EIAV. These peptides are summarized in Table 1. Sequences were based on the Prototype (cell-adapted Wyoming) strain of EIAV (17) and correspond to conserved regions of the envelope proteins (11, 16).

TABLE 1.

Summary of EIAV diagnostic peptides

Peptide name	Sequence	Sourcea	Labelb
R51	IGCIERTHVFCHTG	gp45 (env 534–547)	5CF, 6CF
R51G	GCIERTHVFCHTG	gp45 (env 535–547)	5CF, 6CF
R51C	CIERTHVFCHTG	gp45 (env 534–547)	5CF, 6CF
R51L	LIGCIERTHVFCHTG	gp45 (env 533–547)	5CF, 6CF
R32	KERQQVEETFNLI	gp45 (env 522–534)	5CF, 6CF
R32ER	ERQQVEETFNLI	gp45 (env 523–534)	FITC-6
R32R	RQQVEETFNLI	gp45 (env 524–534)	FITC-6
R32QQ	QQVEETFNLI	gp45 (env 525–534)	FITC-6
R32Q	QVEETFNLI	gp45 (env 526–534)	FITC-6
R32V	VEETFNLI	gp45 (env 527–534)	FITC-6
R32E	EETFNLI	gp45 (env 528–534)	FITC-6
R32/51	KERQQVEETFNLIIGCIERTHVFCHTG	gp45 (env 522–547)	5CF
Sam50	ADDWDNRHPLPNAPLVAPPQGPIPMT	p26 (170–201)	5CF
Sam50H	HPLPNAPLVAPPQGPIPMT	p26 (177–201)	5CF
Sam50A	APLVAPPQGPIPMT	p26 (182–201)	5CF
Sam51	VDCTSEEMNAFLDVVPGQAGQKQILLDAIDKI	p26 (202–227)	5CF
Peptide 12	LETWKLVKTSGVTPLPISSEANTGL	gp90 (env 408–432)	5CF
Pep12S	SGVTPLPISSEANTGL	gp90 (env 419–434)	FITC-6
Pep12P	PISSEANTGL	gp90 (env 425–434)	FITC-6
Peptide 1	YGGIPGGISTPITQQSEKSK	gp90 (env 1–20)	5CF

^aAll peptide sequences are based on the prototype strain of EIAV (17). 

^bFITC-6, fluorescein-6-isothiocyanate. 

Peptide synthesis and labeling.

Peptides were produced on a 0.2-mmol scale using a Millipore automated peptide synthesizer and standard 9-fluorenylmethoxy carbonyl chemistry as described previously (6). Peptides were labeled with 5- or 6-carboxyfluorescein (5CF or 6CF, respectively; Molecular Probes, Eugene, Oreg.) while still on the resin, thus placing the fluorophore on the N terminus of the peptide. The 9-fluorenylmethoxy carbonyl protecting group was removed from the N terminus of the peptide resin by 25% piperidine in dimethylformamide (DMF) followed by four washes with DMF. The fluorescent probe was dissolved in DMF to a concentration of 0.3 M, and this solution was mixed with 0.9 M N,N-diisopropyl ethylamine (DIPEA) and 0.6 M N-hydroxybenzotriazole–2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluroniumtetrafluoroborate (HOBT-TBTU) in a 5:4:2 ratio. The dye mixture was added to the resin and incubated overnight with shaking. Following four washes each with DMF and dichloromethane, the resin was dried under vacuum. The dye-conjugated peptides were cleaved from the resin using standard trifluoroacetic acid cleavage procedures followed by multiple ether extractions. Peptides were purified by reverse-phase high-performance liquid chromatography and analyzed by mass spectrometry to confirm that the desired product was obtained.

Antifluorescein capture ELISA.

To measure antibody binding to test peptides without regard to their suitability for FP, an antifluorescein capture ELISA was used. To each well of an Immulon 2 HB 96-well plate (Dynex, Chantilly, Va.) was added 50 μl of rabbit antifluorescein antibody (Molecular Probes), 3.5 μg/ml in 50 mM sodium bicarbonate, pH 9.6; the plates were sealed and incubated overnight. The wells were blocked with BLOTTO (5% nonfat dry milk, 5% normal bovine serum, 0.025% Tween 20 in phosphate-buffered saline [PBS]), 100 μl/well for 1 h at room temperature (RT). Peptide solutions (10 nM in BLOTTO) were added to the wells and incubated for 1 to 2 h, followed by washing the wells three times with 0.025% Tween 20 in PBS. The plates were then incubated with test horse sera, diluted 1:100 in BLOTTO, for 1 h at RT, washed as described above, and then incubated with anti-horse immunoglobulin G (IgG) (Fc)-horseradish peroxidase (United States Biochemical), diluted 1:10⁵ in BLOTTO, for 1 h at RT and washed. The substrate, TM Blue Soluble reagent (200 μl/well; Intergen, Milford, Mass.) was added and incubated for 20 min with shaking, and the reaction was stopped with the addition of 50 μl of 1 N H₂SO₄ per well for 5 min with shaking. Absorbance at 450 nm was measured on a Dynex MR5000 microplate reader. Because each peptide caused a slightly different background absorbance, control wells containing no horse serum were included for each peptide tested.

FP measurements.

The fluorescein-labeled peptides were evaluated for their suitability as probes for FP using an FPM-1 FP analyzer (Jolley Consulting and Research, Grayslake, Ill.) in batch mode with the following settings: gain of 80, heater off, and single read. Prior to use and after a freeze-thaw cycle, the sera were centrifuged at 12,000 × g for 2 min to pellet any precipitated protein. Serum was diluted 1:100 or 1:50 into 2 ml of buffer in 12- by 75-mm borosilicate glass tubes (VWR) and mixed well. After the blank was read, fluorescently labeled peptide was added to a final concentration of 1 to 2 nM (100K to 200K total intensity) and incubated for at least 15 min. The FP of the sample was measured and expressed as millipolarization units (mP). Some of the sera were very dark, presumably due to hemolysis. If such a serum sample had poor light transmission (lamp feedback, <0.63), a twofold further dilution was tested. Polarization data were output to a computer running the FPM-1 data collection software and then converted to an ASCII text file and imported into the Quattro Pro spreadsheet program (Corel, Ottawa, Ontario, Canada) for data analysis and graphing. Statistical analysis of the correlation between AGID and FP reactivity was performed using the SAS 6.12 program for logistic regression.

RESULTS

The initial panel of candidate probes comprised seven peptides, each derived from one of the three major proteins of EIAV: peptide 1 and peptide 12 from gp90 (surface unit); R51, R32, and R51/32 from gp45 (transmembrane); and Sam50 and Sam51 from p26 (capsid). The sequences of these peptides are listed in Table 1. In the case of the R51 and derivative peptides, two different fluorescent labels, 5CF and 6CF, were evaluated. The peptides chosen for initial evaluation correspond to conserved regions of the proteins that have been shown to react broadly with equine sera in an ELISA format (1–3).

Due to problems encountered with testing certain horse serum samples (see below), the initial evaluation of FP reactivity made use of purified IgG from a reference long-term, field-infected horse (Lady). Most of the peptides were found to be insensitive to the presence of 60 to 100 μg of Lady IgG per ml in PBS; however R51-5CF, derived from gp45, did undergo a twofold increase in FP in the presence of Lady IgG (Fig. 1). The other peptides in the panel exhibited only slight increases in polarization in the presence of Lady IgG. Based on these results, we used R51-5CF to explore the proper buffer conditions for interaction with antibodies in whole serum.

FIG. 1.

FP reactivity of candidate diagnostic peptides with anti-EIAV IgG. Purified IgG from a reference field-infected horse serum (100 μg/ml) was incubated with the candidate probe peptides (2 nM) in PBS for 20 min. Black bars, probe plus IgG; white bars, probe in buffer alone. All peptides were 5CF derivatives.

We observed that PBS (pH 7.4) with Tween 20, Triton X-100, or lithium dodecyl sulfate detergents often caused precipitation of serum proteins and resulted in low, and occasionally even negative, polarization values due to severe background intensities and low lamp feedback (a measure of transmittance of the light through the sample). For this technique to be successful, there cannot be light-scattering particles present in the sample. Several different buffer compositions and detergents were therefore tested for compatibility with horse serum samples. Surprisingly, when horse serum was diluted 1:50 or 1:100 into 20 to 50 mM sodium phosphate without NaCl, this problem was virtually eliminated. Low-salt conditions also obviated the need for a detergent in the buffer, although signal-to-noise ratios were slightly improved when 0.05% Tween 20 was added to the buffer (data not shown). Under the low-salt conditions, the polarization of peptide R51-5CF increased from 50 to over 200 mP with a 1:100 dilution of a strong positive EIAV antiserum from an experimental infection (pony 95). Thus, it was determined that the optimal buffer composition for the FP assay was 20 mM sodium phosphate, pH 6.8 to 7.0.

Once serum testing was enabled, we tested the panel of peptides with sera from both experimentally and field-infected horses. Although some reactivity was observed with peptide R32 and peptide 12 against pony 95, R51-5CF again was the only peptide from the original panel that was sensitive to serum from field-infected horses. These results reflected those obtained with purified IgG in the FP assay but contrasted with our ELISA results, in which these two peptides reacted very strongly with both pony 95 and Lady sera (data not shown). None of the peptides reacted with EIAV-negative horse serum in either the FP assay or ELISA.

Based on these data, we optimized the R51 peptide for maximum FP signal by exploring the effects of alterations in peptide length and fluorescein linkage. Because different fluorescein linkages can result in differences in sensitivity in the FP assay, we labeled the R51 peptide with 6CF to compare with the 5CF isomer. We also synthesized analogs of R51 possessing zero to three amino acid residues between the N-terminal cysteine and the fluorescein probe. We found that neither reducing nor increasing peptide length improved signal; however, changing from a 5CF to a 6CF label did significantly improve the signal of R51 with positive sera (220 mP for 5CF versus >300 mP for 6CF) without increasing background (Fig. 2). Thus, the R51-6CF probe was demonstrated to be the most sensitive in terms of serum reactivity in the FP assay.

FIG. 2.

Influence of peptide length and fluorescein linkage on FP reactivity of peptide R51. Peptides (approximately 2 nM) were incubated with a 1:100 dilution of serum in 50 mM sodium phosphate, pH 6.8, for 20 min. Peptides are shown in order of decreasing length. Black bars, strong positive PV infected (pony 95); white bars, uninfected (Petite).

In addition to R51, peptides R32 and Pep12 were engineered in an effort to improve their sensitivity in FP. These peptides showed strong and broad reactivity in the antifluorescein ELISA but did not exhibit an increase in FP upon mixing with purified antibodies from a field-infected animal (Fig. 1). A series of peptides of different lengths were synthesized and labeled at their N termini by fluorescein-6-isothiocyanate (Table 1). The complete R32 series was tested for reactivity to positive and negative sera. The most sensitive peptide was R32QQ, a 10-amino-acid peptide. The R32 peptides all reacted well with strongly positive experimentally infected animals (ponies 95 and 562) but did not react with serum from the field-infected horse (Lady). Likewise, neither of the Pep12 analogs displayed a large change in FP in the presence of Lady serum (data not shown). Therefore, we concluded that these peptides are sensitive only to experimentally infected horse sera and are not appropriate for a field diagnostic assay.

Focusing on our highly sensitive probe, R51-6CF, we next tested sera from both uninfected and field-infected horses. We first examined the specificity of the probe by testing 106 serum samples that were negative by AGID (Q. Muencks, personal communication). Tested at a 1:100 dilution, these serum samples had very low and somewhat uniform polarization values (73.6 ± 3.1 mP), indicating that specificity was very high for R51-6CF (Fig. 3). Out of all of the negative samples tested, only two initially reacted in the assay, and both of these had signs of contamination (observation of large solid matter in the tubes). Upon sterile filtration and retesting, these two samples gave consistently negative readings. Thus, provided that the samples were of reasonable quality (absence of light-scattering particles), the FP assay achieved a specificity of 100%. This represents a substantial improvement over the 2% false-positive rate reported for the enzyme immunoassay C-ELISA (8).

FIG. 3.

FP reactivity of R51-6CF with a panel of AGID-positive (151 samples) and -negative (106 samples) horse sera. All test horse sera were tested at a dilution of 1:100 using standard FP reaction conditions. The resultant data are presented as the number of serum samples (y axis) displaying particular levels of FP (x axis).

To determine the sensitivity of the FP assay, we tested a panel of 151 sera from Coggins test-positive field-infected equids at a 1:100 dilution. These sera were obtained from geographically distinct regions throughout the United States, including Texas, Utah, Missouri, Florida, and Oklahoma. These AGID-positive sera caused the polarization of R51-6CF to increase to an average value of 152.8 ± 51.4 mP, a clear difference from the average of the negative sera (Fig. 3). The probe reacted well with serum antibodies from diverse geographic regions, indicating that the peptide epitope is well conserved among various strains of EIAV. Whereas the polarization values for the AGID-negative sera were within a narrow range, the level of reactivity among AGID-positive sera varied extensively, ranging from 60 to 240 mP (Fig. 3). These results demonstrate a high level of FP reactivity with the panel of AGID-positive serum samples, in marked contrast to the lack of FP reactivity observed with AGID-negative serum samples. Thus, these data indicate a high degree of sensitivity and specificity of the FP assay.

To more rigorously define the FP assay's sensitivity and specificity, statistical analysis of the correlation between AGID and FP reactivity was examined by logistic regression. Hypothetical cutoff levels of FP values that effectively discriminated between infected and uninfected horses were determined by generating a receiver operating curve. This analysis revealed a high value of area under the receiver operating curve (c statistic = 0.098), indicating a high predictive probability of the model. A cutoff value of 81 mP resulted in 0 of 100 false positives and 16 of 151 false negatives in the EIAV FP assay. Lower cutoff values resulted in a sharp increase in the number of false positives with little reduction in the number of false negatives. Thus, a cutoff value of 81 mP was selected to provide the best correlation between FP and AGID data. This result is manifest in the histogram in Fig. 3, where the delineation between all positive values and a mix of positive and negative values is apparent at an FP value of 81 mP. Logistic regression analysis demonstrated a highly significant (P = 0.0001) correlation between FP and AGID reactivity, with a sensitivity of 89.4% and a specificity of 100% at the chosen cutoff value.

Discrepant serum samples were tested by both AGID (Carole Simard, unpublished data) and Western blotting to confirm their serological status. Of the samples that were nonreactive in FP but confirmed to be seropositive to EIAV, several exhibited very low reactivity to the envelope-derived peptides in the antifluorescein-capture ELISA and yet had good reactivity to the p26-derived Sam50 peptide. These data suggested that, for these serum samples, a shorter form of the Sam50 peptide might be more sensitive in the FP assay. We therefore synthesized two shortened analogs of Sam50: Sam50A, a 14-mer peptide, and Sam50H, a 19-mer peptide. When tested in the FP assay, neither of these analogs displayed a measurable interaction with the EIAV-positive sera (data not shown). This lack of reactivity may be due to the low levels of antibodies to this epitope and/or the problem that the peptide is still too long for the fluorophore to undergo a change in polarization upon antibody binding.

In addition to testing sera from various geographic areas, we examined the ability of R51-6CF to detect antibodies early in infection. Serum samples acquired weekly during an experimental infection of four ponies (7) were tested for the presence of EIAV-specific antibodies by FP. This assay detected antibody in a 1:100 dilution of serum at 3 weeks postinfection (Fig. 4), which is the same time at which antibody was first detected by concanavalin A capture ELISA (7). These data indicate that the FP assay is as sensitive as the most sensitive ELISA in detecting early antibody responses to EIAV infection. In addition, the test was as sensitive as or more sensitive than AGID in detecting anti-EIAV antibody: ponies 561, 562, and 564 were AGID positive on day 21, and pony 567 was not positive until day 23 (C. Issel and S. Cook, personal communication). The detection of early immune responses may be an advantage over AGID in that the immune response to the envelope protein tends to arise earlier and to higher levels than the antibodies to p26.

FIG. 4.

Reactivity of R51-6CF with sera (1:100 dilution) from early time points in an experimental infection. Four ponies (561, 562, 564, and 567) were experimentally infected with EIAV as described by Hammond et al. (7). Serum samples collected at weekly intervals were tested for reactivity with R51-6CF peptide.

DISCUSSION

Most diagnostic assays for lentiviruses are based on detection of antibodies to the capsid protein. Although p26 is the most abundant protein in the EIAV virion, the antibody titer to this protein is 10- to 100-fold lower than that to the two envelope proteins, gp90 and gp45 (7). However, the amino acid sequence of the capsid protein is highly conserved compared to the sequences of envelope proteins, particularly gp90, where great diversity can occur even within an infected animal (11). We evaluated peptides derived from all three of these proteins and found that R51, the peptide derived from gp45, had the best combination of high specificity and broad reactivity, as it was able to detect antibodies from horses infected with geographically diverse field strains of EIAV. The R51 peptide is based on a region that is immunodominant in lentiviruses and yet is well conserved. Although the amino acid sequences of envelope proteins of lentiviruses generally vary more than those of the capsid and other core proteins, antigenic variation was evidently not a large problem in our studies, as we have achieved nearly 90% sensitivity with a single envelope-based peptide antigen and the current serum panel. The sensitivity level achieved in this current study correlates well with the results of independent evaluations of the FP assay with larger panels of horse serum in which sensitivity levels of 94.5% (C. Issel and S. Cook, University of Kentucky, personal communication) and 95.4% (C. Simard and L. Delorme, Animal Diseases Research Institute of Agriculture Canada, personal communication) were achieved. Taken together, the three studies indicate an average sensitivity of about 93%. To date, these evaluations of the FP assay have demonstrated 100% specificity in that none of the AGID-negative serum samples tested have shown reactivity with the R51 peptide. Detailed analyses of these data are in preparation.

In these initial evaluations of the EIAV FP assay, we have used the standard AGID diagnostic assay to define EIAV-positive and -negative horse serum samples, as this assay is the U.S. Department of Agriculture reference diagnostic assay. Assuming the accuracy of these AGID determinations, the FP assay with the R51 peptide appears to have an average false-negative rate of about 7% in the three independent evaluations. The lack of serum reactivity by the few AGID-positive serum samples may be attributed to several different factors. First, it is possible that these particular horses did not produce antibodies to the R51 peptide region due to their particular immunogenetic idiotype that fails to recognize this particular antigenic determinant. Alternatively, it is possible that these few horses were infected with a strain of EIAV containing a variant R51 peptide sequence. There is to date no report of sequence variation in the R51 peptide region in the reported envelope sequences of various EIAV isolates, but this envelope sequence data set is rather limited in that most of the viral strains studies are lab adapted; there are no primary field isolate sequences. However, the possibility of a variant R51 peptide sequence in FP-negative–AGID-positive horses can be tested by characterizing viral envelope sequences from EIAV isolated from these FP-negative horses, and these experiments are in progress. Since our current panel of immune serum samples were obtained predominantly from slaughterhouses, we were not able to further assess the EIAV status of discordant serum samples by additional testing of the donor horse. However, studies are planned with other horse populations to examine in more detail the basis for the disagreement between the AGID and FP reactivities and to compare the FP directly to other commercial diagnostic assays for EIAV.

To address the few sera that were not positive in FP for anti-R51 antibodies, an additional peptide probe based on the EIAV p26 antigen may need to be developed. The R51 nonreactor ponies did show some reactivity to the p26-derived Sam50 in the peptide ELISA, indicating a potential complementary diagnostic peptide that can be optimized for FP reactivity as done with the R51 peptide. If successful, an FP assay based on a combination of p26 and envelope peptides may be developed to achieve even higher sensitivity than that observed with the single-peptide FP assay. Efforts to identify a more sensitive gp90-based peptide via peptide scanning (data not shown) did not produce any peptides with better seroreactivity than R51.

The data described here were acquired on an FPM-1 instrument, a single-read machine interfaced with a computer. However, it is important to note that FP instrumentation is available with a wide spectrum of sample capacities, ranging from robotic processing of large numbers of samples appropriate for centralized diagnostic laboratories to portable fluorescence polarimeters that can be used in the field to obtain instant results with small numbers of serum samples. This instrument versatility provides important advantages to EIAV testing in that the FP technology described here can be readily applied both to centralized diagnostic labs and to field use in detecting EIAV-infected horses. Thus, the EIAV FP assay represents an important advance over existing EIAV diagnostic assays.

Whereas the current commercial EIAV diagnostic assays have greatly contributed to the control of this equine disease, the availability of a more rapid assay that can be used both in reference labs and directly in the field can facilitate a more thorough screening of horses for this viral infection (10). In this current study, we have described the development of a novel EIAV diagnostic assay based on FP that is specific, sensitive, rapid, and equally applicable to use in centralized testing facilities and to use in field situations for testing small or large numbers of serum samples. Thus, we believe that the FP assay provides an important new diagnostic tool in the effort to identify EIAV-infected horses and to control and ultimately eradicate this equine disease.