Use of Immunoglobulin G Avidity Assays for Differentiation of Primary from Previous Infections with West Nile Virus

Since its introduction in 1999, West Nile virus (WNV) infectionshave spread rapidly across the North American continent. Diagnosisof acute WNV infection by detection of WNV-specific immunoglobulinM (IgM) is complicated by the persistence of detectable IgMfor more than 1 year in some patients. IgG antibody aviditytesting was assessed as a supplemental assay in the diagnosisof current infections. Three groups of serum samples were assayedin parallel by two different IgG avidity test systems (indirectimmunofluorescence test [IIFT] and prototype enzyme-linked immunosorbentassay [ELISA]; EUROIMMUN, Luebeck, Germany). Group I (40 serataken between 2 and 9 days after the onset of influenza-likesymptoms) and group II (40 sera taken between 10 and 43 daysafter onset) were acute and convalescent specimens from patientswith a positive anti-WNV IgM test (ELISA; Focus Diagnostics,Cypress, CA). Group III consisted of 43 patient sera collectedbetween 6 and 12 months after infection. IgG antibodies specificfor WNV were detected in 38% (ELISA) and 50% (IIFT) of groupI sera, in 90% (ELISA and IIFT) of group II sera, and in 100%(ELISA and IIFT) of group III sera. Low-avidity IgG antibodieswere demonstrated in 86% (ELISA) and 95% (IIFT) of IgG-positivepatient samples taken between 2 and 43 days after the onsetof symptoms (groups I and II). High-avidity IgG antibodies weredetected in 100% of group III sera obtained 6 months or moreafter the onset of symptoms (ELISA and IIFT). IgG avidity testsfor WNV infections are rapid and simple to perform. The determinationof IgG avidity provides additional diagnostic certainty in differentiatingbetween recently acquired and previous infections with WNV.

INTRODUCTION

Top

Introduction

West Nile virus (WNV) has spread rapidly across North Americasince it was first detected in New York in 1999 (6, 8, 12, 14,15). Although WNV was first detected in Canada in 2001, thefirst cases of human illness were not identified until 2002(5). In 2003, large outbreaks of WNV-associated illness occurredin the Midwestern United States (2) and the prairie provincesof Canada (http://www.phac-aspc.gc.ca/wnv-vwn/pdf_sr-rs_2004/surveillance_table_042904_hm.pdf).As a result, a significant proportion of the population of Saskatchewanseroconverted during the 2003 mosquito season. In the southwesternregion of the province, approximately 10% of the populationwere seropositive when surveyed in the spring of 2004 (7), mostof whom had no history of WNV infection.

The diagnosis of WNV infection has relied heavily upon the detectionof immunoglobulin M (IgM) antibodies in blood (or in the cerebrospinalfluid in cases of neurological infection). Unfortunately, IgMantibodies are persistent for more than 1 year in some patients(17). This finding has significant implications for the differentiationof recent or current infection with WNV from persistent seropositivityderived from the previous WNV season. The ability to identifyimmune responses persisting from previous years may be importantfor focusing public health efforts to control acute outbreaks.It is often difficult to obtain convalescent samples from patientsafter mild illness; thus, it may not be possible to confirmrecent infection by detection of seroconversion using a hemagglutination(HA) inhibition (HI) assay and a plaque reduction neutralizationtest (PRNT). We investigated the use of IgG avidity to excludehistorical infection as a cause of IgM reactivity.

MATERIALS AND METHODS

Top

Materials and Methods

Serum specimens. Three groups of specimens, representing recent and previousinfections, were studied. These included 80 sera from patientswhose symptoms were consistent with a diagnosis of acute WNVinfection, comprised of group I (40 sera taken between 2 and9 days after the onset of influenza-like symptoms) and groupII (40 sera taken between 10 and 43 days after onset). GroupIII contained 43 sera collected 6 to 12 months after exposurefrom patients known to have been infected with WNV in the summerof 2003.

WNV IgM ELISA. IgM antibodies against WNV were detected using an IgM captureenzyme-linked immunosorbent assay (ELISA) (Focus Diagnostics,Cypress, CA), as described by the manufacturer. Reactive serawere retested following a background subtraction procedure (9)to eliminate false-positive IgM results.

WNV IgG ELISA and HI assay. Serosurvey samples were screened for the presence of flavivirusIgG antibody using a monoclonal antibody-based capture enzyme-linkedimmunosorbent assay as previously described (10). The HI assaywas performed on selected serum samples essentially as describedpreviously (3). In brief, dilutions of acetone-treated serumwere mixed with 8 HA units of suckling mouse WNV antigen andincubated at 4°C overnight. Goose erythrocytes were thenadded to the mixture, and the solution was incubated for anadditional hour at room temperature. The HI titer was determinedas the highest dilution of serum that caused complete inhibitionof erythrocyte agglutination by 8 HA units of viral antigen.

Plaque reduction neutralization test. ELISA/HI-reactive samples were confirmed for the presence ofWNV-specific antibody by carrying out a PRNT. The standard PRNTwas performed essentially as described previously (1). In brief,mixtures of virus (200 PFU) and dilutions of human sera wereincubated at 37°C for 1 h and added to six-well plates containingVero cell monolayers. After a 37°C incubation of 1 h, anutrient agar overlay was added, and the plates were placedin a CO₂ incubator for approximately 3 days, after which a secondoverlay containing neutral red as a vital stain was added. Plateswere then checked periodically over the next few days for plaqueformation. The highest serum dilution with a plaque reductionof at least 90% was defined as the titration end point.

IgG avidity IIFT. Sera were tested using an indirect immunofluorescence test (IIFT)(Euroimmun AG, Luebeck, Germany). Briefly, 25 µl of serumdiluted 1:10 in buffer was added to each of two reaction fieldson a BIOCHIP slide. Each field contained two BIOCHIPS, one withWNV-infected cells and the second with uninfected cells. Afterincubation at room temperature for 30 min, cells were exposedto either 6 M urea solution or to phosphate-buffered saline-Tweenfor 10 min. After washing with phosphate-buffered saline-Tween,cells were incubated with fluorescein-labeled anti-human globulinfor 30 min and then viewed at x20 magnification using a 488-nmexcitation filter and a 520-nm blocking filter. In a specificpositive reaction, the cytoplasm of infected cells demonstratedfine to coarse granular fluorescence and inclusion bodies insome cells. The presence of low-avidity IgG antibodies was inferredby the absence or significant reduction of fluorescence of cellstreated with urea compared with the buffer-treated controls.

IgG avidity ELISA. Sera were tested using a prototype ELISA (Euroimmun AG, Luebeck,Germany). Briefly, 100 µl of serum diluted 1:100 in bufferwas added to each of two wells coated with WNV. After incubationfor 60 min at 37°C, wells were exposed to either 6 M ureasolution or to phosphate buffer for 10 min. After washing, wellswere incubated with peroxidase-labeled anti-human IgG for 60min at 37°C, substrate was added, and the reactions werestopped after 15 min by the addition of 100 µl stop solutionper well. The reactions were read immediately at a wavelengthof 450 nm using a reference wavelength of 650 nm. A relativeavidity index (RAI) was calculated for each specimen and wasexpressed as the percentage of reactivity remaining in the urea-treatedwell. Samples with RAIs of <40% were regarded as having lowavidity, and those with RAIs of >60% were regarded as havinghigh avidity.

RESULTS

Top

Results

IgM antibodies were detected in all 80 specimens from acutelyill patients with WNV infection, and the 43 patients with previousinfections were diagnosed by detection of IgM and IgG antibodiesand of specific neutralizing antibodies by PRNT. IgG antibodieswere detected by IIFT in 20 samples (50%) from acutely ill patients,rising to 90% of convalescent samples (Table 1). The IgG ELISAwas slightly less sensitive than the IIFT, and only 15 (38%)of the acute specimens were IgG positive by ELISA. In the convalescentsamples, both methods detected IgG antibodies in 36 patients(90%). In the specimens collected 6 months or more after theonset of symptoms, IgG seroprevalence was 100% using both IIFTand ELISA methods.

View this table:
[in this window]
[in a new window]
TABLE 1. Detection of IgG antibodies and IgG avidity in patients with recent or previous W NV infection

Among the IgG-positive acute and convalescent samples, the greatmajority were of low avidity (Table 1). The number of low-aviditysamples was highest in the convalescent samples, reflectingthe higher prevalence of IgG in this group. In addition, thenumber of low-avidity samples was greater using the IIFT thanthe ELISA because of the higher sensitivity of the IIFT.

When tested using the IgG ELISA, there was very little differencein the proportion of IgG-positive samples showing low avidity,but the IgG ELISA appeared to be less sensitive than the IIFTfor detection of IgG (Table 1). No samples from previously infectedsubjects showed low avidity (Fig. 1), although a few specimensfrom recently infected patients exhibited high avidity indices.

View larger version (8K):
[in this window]
[in a new window]
FIG. 1. Difference in relative avidity index between recent and previous infection detected by ELISA (left) and difference in intensity of fluorescence in IIFT (reduction of fluorescence of cells treated with urea compared with the buffer-treated cells) (right).

DISCUSSION

Top

Discussion

IgM antibodies to WNV have been shown to persist for at leasta year and for up to 500 days in some patients (17). It is notknown whether persistence for longer periods occurs and, ifso, if it occurs in sufficient proportions of cases and at titershigh enough to be clinically confusing. Since IgM persistencemay be problematic when current cases of WNV infection are identifiedusing serological assays such as the IgM ELISA, other approachesmay need to be considered. Documentation of seroconversion usingacute- and convalescent-phase sera has been the "gold standard"for demonstrating a current infection; however, convalescentsamples are not always available, and rises in antibody titermay be missed if acute samples are collected too late afterexposure. Several approaches have been used to differentiaterecent from previous flavivirus infection. Detection of IgAantibodies in patients with recent, but not previous, WNV infectionhas also been reported (16); further studies are needed to determinethe duration of IgA persistence and its diagnostic value inhelping to characterize the time frame of exposure.

IgG avidity has been employed to distinguish primary and secondarydengue virus infections (4, 13) with sensitivity similar tothose of the assays described in this study. In the early stagesof the immune response, IgG antibodies are produced againsta wide range of antigenic epitopes, producing an IgG responseof relatively low specificity, which results in low IgG avidity.Thus, the detection of low-avidity IgG after a recent febrileillness, in conjunction with a positive IgM ELISA, is presumptiveevidence of an acute or recent WNV infection.

In contrast, as the immune response matures and B-cell clonalselection occurs, the resulting IgG antibodies are more specific,and the IgG avidity increases. In this study, all sera tested6 months or more after exposure to WNV contained high-avidityIgG antibodies. Thus, our results indicate that individualsexposed to WNV during the preceding mosquito season will inalmost all cases demonstrate high-avidity IgG antibodies iftested the following spring.

Antibodies to flaviviruses are cross-reactive, but cross-reactivityis somewhat less pronounced in immunofluorescence assays thanin ELISA formats (11). Moreover, in many cases, differentiationis possible using a flavivirus BIOCHIP mosaic (dengue, St. Louisencephalitis, tick-borne encephalitis, and yellow fever viruses)with the determination of antibody profiles. A very small proportionof specimens from recent infections showed high IgG avidity(Fig. 1). It is not known whether cross-reacting flavivirusantibodies will result in a high-avidity result, leading toa recent WNV infection being misclassified as a previous infection,but this possibility requires further study.

The Euroimmun prototype IgG ELISA appeared to be less sensitivethan the IIFT for detection of IgG early after the onset ofsymptoms. However, there was minimal difference between ELISAand IIFT methods for avidity determination. An ELISA using amore purified antigen, in order to increase the sensitivityof IgG detection and reduce cross-reactivity, is in development.

The value of avidity determination will be great in years followinglarge outbreaks of WNV infection, as occurred in Saskatchewanin 2003. Weather conditions in Western Canada during the summerof 2004 limited vector development and activity, resulting invery few cases of WNV infection. Thus, most of the positiveIgM tests recorded in 2004 were reflective of previous infections.Epidemiological investigation of these cases consumed much timeand effort, which might have been saved if the capacity to testfor the presence of high-avidity IgG antibodies had existed.

ACKNOWLEDGMENTS

We thank Sina Erdmann and Britta Arp for excellent technicalsupport.

FOOTNOTES

* Corresponding author. Mailing address: 3211 Albert Street, Regina, Saskatchewan S4S 5W6, Canada. Phone: (306) 787-3135. Fax: (306) 787-1525. E-mail: plevett{at}health.gov.sk.ca.

REFERENCES

Top