Utility of the Focus Technologies West Nile Virus Immunoglobulin M Capture Enzyme-Linked Immunosorbent Assay for Testing Cerebrospinal Fluid

Focus Technologies has developed an immunoglobulin M (IgM) captureenzyme-linked immunosorbent assay (ELISA) kit that utilizesrecombinant West Nile virus (WNV) antigens to detect WNV IgMin serum. We evaluate here the utility of the kit for detectingWNV IgM in cerebrospinal fluid (CSF). The sensitivity was evaluatedby using 52 CSF specimens from the 2002 WNV season that werepositive in both the Public Health Service Laboratories WNVIgM ELISA and an in-house WNV IgM ELISA with native WNV antigen.The specificity was evaluated with two groups of specimens:(i) 73 CSF specimens submitted for in-house WNV IgM ELISA testingfrom February through April 2003 and yielding a negative WNVIgM result and (ii) 60 CSF specimens determined to be positivefor another virus by PCR testing. Using these 185 CSF specimensat a screening dilution of 1:2, the kit was determined to be100% sensitive and 100% specific. Endpoint titers were determinedfor 20 IgM-positive CSF specimens by testing serial twofolddilutions and ranged from 1:8 to 1:512. Index values (specimenabsorbance value/calibrator absorbance value) for the screeningdilution (1:2) showed no correlation with IgM titers, whereasindex values for higher dilutions showed significant correlationwith IgM titers. CSF screening dilutions of greater than 1:2are not recommended, however, due to the risk of obtaining false-negativeresults. These findings show that the Focus Technologies WNVIgM capture ELISA, when utilized as recommended, offers accuratequalitative detection of WNV IgM in CSF specimens.

INTRODUCTION

Top

Introduction

Laboratory assays designed to detect West Nile virus (WNV)-specificimmunoglobulin M (IgM) have proven to be very useful for thediagnosis of West Nile virus (WNV) infection (1, 2, 4, 5, 7,8, 10-13). The vast majority of WNV-infected patients have detectableWNV IgM in serum specimens collected at presentation (6, 9).Involvement of the central nervous system in the infectiousprocess is demonstrated by the detection of WNV-specific IgMin cerebrospinal fluid (CSF) (6, 9, 11, 12). Serum and CSF aretypically screened for WNV IgM by using an IgM capture enzyme-linkedimmunosorbent assay (ELISA). Due to the cross-reactivity observedamong flaviviruses in the Japanese encephalitis virus seroocomplex,ELISA-positive samples may be tested by using a confirmatoryassay, such as the plaque reduction neutralization test, toshow that the IgM detected is specific for WNV (6, 9, 10, 12).

Focus Technologies has developed a WNV IgM capture ELISA kitfor the detection of WNV IgM in serum. This kit, which utilizesrecombinant WNV premembrane and envelope proteins (3), demonstratesexcellent sensitivity and specificity compared to the assaysused during the 2002 WNV season in the United States (W. R.Hogrefe, R. J. Moore, and H. E. Prince, unpublished observations).The purpose of the studies presented here was to determine theperformance characteristics of the Focus Technologies WNV IgMcapture ELISA kit for detecting WNV IgM in CSF.

MATERIALS AND METHODS

Top

Materials and Methods

CSF specimen panels. The three panels used to assess assay sensitivity and specificityare described in Table 1. The sensitivity panel consisted offrozen (-70°C) aliquots of 52 CSF specimens giving a positiveresult in an in-house WNV IgM capture ELISA (utilizing nativeWNV antigens) performed by the Focus Technologies referencelaboratory Immunology section during the 2002 WNV season (10,11). This assay was performed by using CSF at a 1:5 dilutionas described previously (11). Duplicate aliquots of these specimenswere forwarded to PHSL (13 different State Public Health Laboratories)for further evaluation during 2002, and all 52 specimens werealso positive in the Public Health Service Laboratories (PHSL)WNV IgM capture ELISA (7, 11). Only five of these CSF specimenswere tested by the plaque reduction neutralization test, andall were positive for WNV-specific antibodies. The specificitywas evaluated by using two different panels. The first specificitypanel consisted of 73 CSF specimens submitted for testing bythe in-house WNV IgM capture ELISA (native antigen) during aperiod of low WNV endemicity (February through April 2003) andgiving a negative WNV IgM result. The second specificity panelconsisted of 60 CSF specimens positive for another virus byPCR testing performed by the Focus Technologies reference laboratoryMolecular Diagnostics section; samples were frozen (-70C) afterPCR testing until utilized for the present study.

View this table:
[in this window]
[in a new window]
TABLE 1. CSF specimens evaluated

WNV IgM capture ELISA kit. The Focus Technologies (Cypress, Calif.) WNV IgM capture ELISAkit utilizing recombinant antigens was used according to themanufacturer's instructions, with the exception that CSF wasdiluted 1:2. The kit contains a calibrator serum, which is includedin all runs; the reactivity of each kit control and unknownspecimen is expressed relative to the reactivity of the calibratorserum, thus allowing for the interassay comparison of results.Calibrator serum (in triplicate), negative control serum, andpositive control serum were diluted 1:101 in sample diluent(phosphate-buffered saline containing 0.1% bovine serum albuminand 0.1% Tween 20), and CSF specimens were diluted 1:2 in samplediluent. Diluted specimens (0.1 ml) were added to microtiterwells (Nunc PolySorp) coated with rabbit anti-human IgM (heavy-chainspecific). After an hour at room temperature, the wells werewashed three times with 0.25-ml volumes of wash buffer (phosphate-bufferedsaline plus 0.1% Tween 20); the wells then received 0.1 ml ofWNV recombinant antigen (lyophilized material reconstitutedwith sample diluent per the package insert) for 2 h at roomtemperature. After three washes as described above, 0.1 ml ofhorseradish peroxidase-conjugated anti-flavivirus monoclonalantibody (6B6C-1), supplied as a ready-to-use reagent, was addedto all wells. After incubation at room temperature for 30 minand three washes as described above, the wells received 0.1ml of enzyme substrate (tetramethylbenzidine), supplied as aready-to-use reagent, for 10 min; the reaction was then stoppedby the addition of 0.1 ml of 2 N sulfuric acid. The absorbanceat 450 nm was measured by using an ELISA reader. For each specimen,including serum controls and patient CSF specimens, the indexvalue was calculated by dividing the specimen absorbance valueby the mean absorbance value obtained for the calibrator serum.Index values of <0.90 were considered negative, values of0.90 to 1.10 were considered equivocal, and values of >1.10were considered positive for WNV IgM.

Titration of positive CSF specimens. To further characterize CSF specimens giving a positive resultwith the WNV IgM capture ELISA kit (recombinant antigen), serialtwofold dilutions of positive specimens with sufficient volume(n = 20) were tested as described in the previous section. Theendpoint titer was defined as the highest dilution giving anindex value of $>=$ 0.90.

Statistical analyses. The relationship between index values observed at various CSFdilutions and endpoint titers was evaluated by using linearregression anal-ysis. A correlation coefficient (R²) with aP value < 0.01 was considered statistically significant.

RESULTS

Top

Results

Sensitivity and specificity. The results for the sensitivity and specificity panels are shownin Fig. 1. All CSF specimens in the sensitivity panel exhibitedWNV IgM capture ELISA (recombinant antigen) index values of>1.10, whereas all CSF specimens in the two specificity panelsexhibited index values of <0.90. Thus, when used to testCSF specimens at a 1:2 dilution, the Focus Technologies WNVIgM capture ELISA kit was 100% sensitive and 100% specific.

View larger version (17K):
[in this window]
[in a new window]
FIG. 1. Distribution of index values for CSF specimens tested in the Focus Technologies WNV IgM Capture ELISA kit. Each point represents the result for one specimen.

Titration studies. WNV IgM-positive CSF specimens with sufficient volume were seriallydiluted and tested to determine the endpoint titer. Index valuesare presented as a function of dilutions in Fig. 2. For 16 of20 CSF specimens, the titer was measurable using dilutions between1:2 and 1:256; four specimens required retesting with dilutionsexpanded to cover the range 1:2 to 1:1,024. The distributionof endpoint titers was as follows: 1:8 (n = 1), 1:16 (n = 4),1:32 (n = 1), 1:64 (n = 5), 1:128 (n = 4), 1:256 (n = 2), and1:512 (n = 3).

View larger version (32K):
[in this window]
[in a new window]
FIG. 2. Titration curves for 20 WNV IgM-positive CSF specimens. Each curve represents the results for one CSF specimen.

Relationship of index values at various dilutions to endpoint titers. To determine whether index values could be used as a semiquantitativeindicator of endpoint titers, index values obtained for 1:2to 1:16 dilutions were plotted as a function of endpoint titersfor the 20 WNV IgM-positive CSF specimens tested by serial dilution(Fig. 3). No significant correlation was observed between endpointtiters and index values at a 1:2 dilution. In contrast, endpointtiters were significantly correlated with index values at 1:4,1:8, and 1:16 dilutions. Higher dilutions were associated withhigher correlation coefficients. However, the strong correlationbetween endpoint titers and index values at the 1:16 dilutionwas offset by a decrease in assay sensitivity; the specimenwith a titer of 1:8 exhibited a false-negative result at the1:16 dilution, and two of three specimens with titers of 1:16were equivocal.

View larger version (21K):
[in this window]
[in a new window]
FIG. 3. Relationship of index values at various dilutions to endpoint titers for 20 WNV IgM-positive CSF specimens. Correlation coefficient (R²) values at 1:4, 1:8, and 1:16 dilutions (B, C, and D, respectively) were statistically significant (P < 0.01), whereas the correlation coefficient for the 1:2 dilution (A) did not reach statistical significance.

DISCUSSION

Top

Discussion

The findings presented in this report clearly demonstrate thatthe Focus Technologies WNV IgM Capture ELISA kit, initiallydesigned for testing serum, accurately detects WNV IgM in CSF.Using a panel of well-characterized CSF specimens known to containWNV IgM, the kit exhibited 100% sensitivity. Likewise, the kitwas 100% specific; all 73 CSF specimens collected during a periodof low endemicity and negative for WNV IgM by using an in-houseELISA based on reactivity with native WNV antigens were alsoWNV IgM negative with the kit. Studies conducted with an additionalspecificity panel demonstrated that CSF specimens from patientswith infections caused by other viruses, including enterovirusesand herpesviruses, did not give false-positive results withthe WNV IgM kit.

Due to the unavailability of CSF specimens from patients withother flavivirus infections, we were not able to perform CSFcross-reactivity studies with the Focus Technologies IgM kit.However, studies using similar IgM capture assays to evaluateserum specimens have shown that cross-reactivity may be extensive,depending on the infecting flavivirus. Martin et al. (8) foundthat 63% of sera from patients infected with St. Louis encephalitisvirus gave a positive signal in a WNV IgM capture ELISA; incontrast, only 31% of patients with Japanese encephalitis virusinfection were reactive. These findings emphasize the need toperform additional testing for patients whose CSF and/or serumare positive in a flavivirus IgM capture ELISA.

Our choice of a screening dilution of 1:2 for CSF was a compromisebetween precedent and practicality. Although CSF specimens aretested undiluted in the PHSL WNV IgM ELISA (7, 8), our experiencein a reference laboratory setting demonstrates that the volumeof CSF submitted is often too low to permit testing of undilutedsample. Thus, some degree of dilution is required to facilitatetesting of all submitted CSF specimens. Further, subjectingCSF specimens to dilution serves to minimize any adverse matrixeffects that may characterize utilization of undiluted CSF.

Detailed characterization of WNV IgM detection in CSF at variousdilutions revealed that, at the screening dilution of 1:2, theWNV IgM capture ELISA kit functions as a qualitative assay.There was no significant correlation between the index observedat 1:2 and the endpoint titer determined by testing serial twofolddilutions. Although the results indicated that the assay exhibitedquantitative characteristics at higher dilutions, the routineutilization of higher dilutions for CSF testing is not recommended.This position is based mainly on the lack of data regardingthe endpoint titers of WNV IgM-positive CSF specimens exhibitingindex values in the range of 1.1 to 3.0 when tested at a 1:2dilution. Although three WNV IgM-positive CSF specimens exhibitedindex values in this range (Fig. 1), none had sufficient volumeto allow titration studies. Along these same lines, the populationof WNV IgM-positive CSF specimens used for these experimentswas biased toward strongly positive specimens, since they werepositive when tested at a 1:5 dilution in the 2002 WNV IgM ELISAutilizing native WNV antigen (11). Thus, it remains unclearwhat proportion of CSF specimens truly positive for WNV IgM,as assessed at a 1:2 dilution with the new WNV IgM ELISA kit,may give false-negative results at higher dilutions. Until thefindings of such studies are available, we recommend that theFocus Technologies WNV IgM capture ELISA kit be used as a qualitativeassay for the detection of WNV IgM in CSF specimens.

ACKNOWLEDGMENTS

We thank all PHSL staff members who generated and supplied PHSLWNV IgM ELISA results. Jane Filamor, Maryam Saber- Tehrani,and Susan Vogeli provided expert technical assistance.

FOOTNOTES

* Corresponding author. Mailing address: Focus Technologies, 5785 Corporate Ave., Cypress, CA 90630. Phone: (714) 503-2047. Fax: (714) 484-1296. E-mail: hprince{at}focustechnologies.com.

REFERENCES

Top