JOURNAL OF CLINICAL MICROBIOLOGY, May 1992, p. 1294-1296 Vol. 30, No. 5 0095-1137/92/051294-03$02.00/0 Copyright C 1992, American Society for Microbiology Application of Quantitative Immunofluorescence to Clinical Serology: Antibody Levels of Treponema pallidum GRACE L. PICCIOLO* AND DAVID S. KAPLAN Center for Devices and Radiological Health, Food and Drug Administration, 12200 Wilkins Avenue, Rockville, Maryland 20852 Received 27 March 1991/Accepted 20 January 1992

A previously reported method of quantitative immunofluorescence, employing a calibrated photometric system and chemically stabilized fluorescence intensity, was used to replace the subjective, visual method of endpoint determination with a quantitative, calibrated measurement of antibodies to Treponema paUlidum in serum. The results of the quantitative immunofluorescence method showed a 90% correlation with the subjective determinations of the visual method.

A quantitative immunofluorescence (QIF) method to de- measured by using an acridine orange filter set, with a 400- to termine immunofluorescence with a quantitative, calibrated 440-nm excitation and LP 470 emission filters. photometric intensity of the fluorescent reaction product has Reducing agent. Dithioerythritol (DTE) (Sigma Chemical been reported previously by us (4). This method used uranyl Co., St. Louis, Mo.) was prepared as stock solution contain- glass slides in the calibration and standardization of the ing 0.3 M reducing agent in 0.5 M Tris buffer, pH 8.2 (Sigma microscope-photometer voltage measurements. To stabilize Chemical Co.). DTE was diluted 1:9 with standard buffered the fluorescence emission, dithioerythritol (DTE) was incor- glycerol mounting medium (Clinical Sciences, Inc., Whip- porated into the buffered glycerol mounting medium of the pany, N.J.) (4, 8). prepared microscope slides, resulting in a slowed, linear, Instrument calibration. Use of a uranyl glass slide (Corn- and predictable rate of decay (8). Results from the applica- ing Glass, Inc., Corning, N.Y.) calibrated the microscope- tion of a QIF method to determine levels of antibody to photometer to 100%, compensating for the aging effects of Treponema pallidum in serum are presented in this report, the excitation lamp, its focusing, and other variables (5). The along with a preliminary validation with clinical specimens, sample fluorescence intensities were corrected on the basis which were tested with the fluorescent treponemal antibody- of statistical regression parameters (3). absorption (FTA-ABS) test kit. The Clinical Sciences, Inc. FTA-ABS test kit was used. A The first application of the QIF method was to determine total of 42 human serum samples with various reactivities in antibodies to Toxoplasma gondii in serum (4), showing a 94% correlation with the visual method used for 62 clinical TABLE 1. Comparison of visual readings obtained for specimens. Kaufman and Nester (6, 9), who measured whole fluorescently labeled treponemes in the fields of treponemes, reported the only other attempt to use presence or absence of DTE quantitative microfluorometric techniques to correlate mea- surements of labeled treponemes with subjective estimates Serum sample no. or Fluorescence reading" of fluorescence for a test for T. pallidum antibodies in serum. control reading' Without DTE With DTE Our work represents a system that can be developed for quality control of reagents by commercial manufacturers and Serum sample other laboratories to improve the reproducibility of stan- 1 -+/N N 2 + + dardizing reagents. In the present study we have relied upon 3 + + the correlations between fluorescence reading and clinical 4 ++ disease state, previously defined by the Centers for Disease 5 +/1+ 1+ Control, by using commercially available FTA-ABS test 6 +/1+ 1+ kits. 7 1+ 1++ (This work was submitted by D. S. Kaplan in partial 8 1+ 1+ fulfillment of the requirements for a Ph.D. from the Depart- 9 1+ 2+ ment of Microbiology, George Washington University, 10 2+ 2,-,3+ Washington, D.C.) 11 2++ 3+ Instrumentation. The 12 3+ 4+ instrumentation consisted of a mi- 13 3+ 4+ croprocessor-controlled Zeiss photomicroscope with a pho- 14 3++ 4+ tometer and scanning stage and a 100-W HBO excitation 15 4+ 4+ lamp (4, 5, 8). Fluorescence filters for fluorescein isothiocyanate pro- Controls duced a 450- to 490-nm excitation and an emission wave- 1+ 1+ 1+, -, 1++ length of 520 nm or longer. The uranyl glass standard was 4+ 4+ 4+ a Fifteen serum samples were mounted in buffered glycerol with and without 0.033 M DTE. Controls were diluted as specified in the manufactur- er's kit directions. bN, negative; 1+, minimally reactive. See the legend to Fig. 1 for other * Corresponding author. definitions. 1294 VOL. 30, 1992 NOTES 1295

0.7 f Q7 r

0.6 F 0.6 I-

05 F 05 I

0.4 H 0.4 -

._ 03 F c 03 F- .0 0C 8) E

0.2 F- Q2 F L-

0 0)0 I 0.1 1 0.1 H .

.,. 0 0 I 9 0

-0.1 -0.1 I

-0.2 -0.2

0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 N ± 1+ 2+ 3+ 4+ N ± 1+ 2+ 3+ 4+ Visual Intensity Visual Intensity FIG. 2. Correlation between the mean corrected intensity and FIG. 1. Correlation between the mean corrected intensity and the visual reading for labeled T. pallidum cells measured in the the visual reading for labeled T. pallidum cells measured in the absence of DTE. Visual intensities: 100, 4+; 87, 3++; 75, 3+; 62, presence of DTE. Visual intensities: 100, 4+; 87, 3++; 75, 3+; 62, 2++; 50, 2+; 37, 1++; 25, 1+; 6 to 12, ±; 0, negative (N). 2++; 50, 2+; 37, 1++; 25, 1+; 6 to 12, +; 0, negative (N). the FTA-ABS tests were provided by the Washington Hos- Calculation of mean corrected intensity. The intensities pital Center. The specimens were mounted in 0.03 M DTE- were corrected on the basis of the linear regression param- glycerol, except that 15 of these specimens were also eters of the standard calibration curve. The mean intensity mounted in glycerol alone. for the 10 cells reacted with the negative control was Localization of samples. The cells were located under subtracted from the intensity of each of the cells reacted with dark-field microscopy in order not to fade the specimen sera which tested positive for T. pallidum antibodies. For before quantitative measurements. For dark-field micros- each patient's serum, the intensity of the 10 cells was copy, an Apo 40x/1.00 phase 3 oil objective with an adjust- averaged. The visually determined results were related to a able iris diaphragm was used with a 1.2/1.4 oil dark-field linear scale ranging from 0 to 100. condensor. Localization for dark-field microscopy was per- Comparison ofvisual readings in the presence of DTE. As is formed with the 0.6 aperture, but the aperture was opened to shown in Table 1, the addition of DTE to the mounting 1.0 for quantitative fluorescence measurements. The FTA- medium did not significantly change the visual interpretation ABS tests were read visually by observer 1 by using a Leitz of the fluorescence intensity, as was also previously shown epifluorescence microscope with similar optics. with the toxoplasmosis IF kit (4), with which the DTE Alignment of cells under measuring aperture. The cells significantly altered the visual intensity only 6 to 11% of the were aligned under the 0.63-mm photomultiplier tube mea- time. However, the DTE did appear to enhance the fluores- suring diaphragm. The cells did not always fit completely cence, since the intensities in the presence of DTE tended to inside the measuring aperture, but the same sensing area was be higher than the intensities in the absence of DTE. The always measured. For each sample (well), including control presence of DTE allows the measurement of the initial, wells, an instantaneous intensity reading was taken for each unfaded fluorescence intensity (4, 8). of 10 cells by using the manual mode of the Zeiss PMI Since a 1 + intensity reading may not actually indicate the program. presence of antibodies to T. pallidum, as the test is clinically Serum dilutions. All sera (including control sera) were interpreted (1, 7, 10), but a 2+ reading usually does, a diluted as specified in the kit directions. change from a 1+ to a 2+ intensity could result in a 1296 NOTES J. CLIN. MICROBIOL.

TABLE 2. Correlation of quantitative intensities obtained for TABLE 3. Sensitivity of and false-positive rate for the QIF fluorescently labeled treponemes with the visual method compared with the visual methoda interpretation of fluorescence intensity Visual method QIF results results Positive Negative Total Groupa Reactivity intensityQuantitativerange' No readingVisual 1 Nonreactive -0.07-0.015 15 NC/+ Positive 16 1 17 2 Borderline 0.02-0.07 9 1+ Negative 3 22 25 3 Reactive 0.08-0.16 18 2+ Total 19 23 42 0.24-0.30 3+ a Sensitivity: (positive observations by QIF method)/(total sera positive by 0.33-0.58 4+ visual method) x 100 = 89%. False-positive rate: (number of sera negative by visual method that tested positive by QIF method)/(total sera positive by QIF a The intensities were divided into three groups on the basis of the correlation method) x 100 = 16%. of the photometrically determined intensity values with the subjective, visual readings used in the serological diagnosis of antibodies to T. pallidum. b The quantitative intensity range represents the average intensity for each group plus or minus two standard deviations. The QIF technique shows a 90% agreement with the visual c N, negative. See the legend to Fig. 1 for other definitions. method of determining levels of antibody to T. pallidum. By comparison with the visual method, Table 3 lists the results of the QIF method for 42 specimens, showing an overall false-positive reading. However, since in only one case sensitivity of 89% and a false-positive rate of 16%. (11.1%) did the DTE raise the visual reading from a 1+ or The QIF method sets the stage as a prototype method for below to a 2+, the intensities in the presence of DTE standardizing both the IF reagents and the IF test. The probably reflect the actual antibody concentration in the diagnostic significance of the IF procedure could be im- serum. However, since a clinical history was not available, proved by an accurate, reproducible method for monitoring this change in intensity may not be significant. the fluorescent reaction product, although visual readings Quantitative measurements of fluorescein isothiocyanate- have done fairly well in the past. labeled treponemes. Figure 1 shows a scatter plot of the mean corrected intensity versus the visual intensity for labeled We express our sincere appreciation to Kris Aldrich of the treponemes measured in the presence of DTE. A similar Washington Hospital Center for assistance in preparing the FTA- plot, for cells measured in the absence of DTE, is shown in ABS test slides and guidance in the interpretation of the FTA-ABS Fig. 2. The 0.90 correlation coefficient (Fig. 1) found by test results. linear regression analysis (2) between the visual and the REFERENCES determined values showed that the photometrically intensity 1. Fitzgerald, T. J. 1985. Treponema, p. 485-489. In E. H. Len- visual intensity is directly correlated with the quantitative nette, A. Balows, W. J. Hausler, Jr., and H. J. Shadomy (ed.), intensity from the 1+ up to the 4+ intensity. Manual of clinical microbiology, 4th ed. American Society for Correlation of quantitative intensity values with visual Microbiology, Washington, D.C. readings. The intensities were divided into three groups 2. Hald, A. 1967. Statistical theory with engineering applications. (Table 2) on the basis of the correlation of the photometri- John Wiley & Sons, Inc., New York. cally determined intensity values with the subjective, visual 3. Kaplan, D. S. 1985. Application of quantitative microfluorimetry readings. These groups represent the following intensities: to clinical serology. Ph.D. dissertation. George Washington (i) nonreactive (no measurable concentration of antibodies to University, Washington, D.C. 4. Kaplan, D. S., and G. L. Picciolo. 1989. Application of T. pallidum); (ii) borderline, including 1+ control (the serum quanti- tative immunofluorescence to clinical serology: antibody levels was then retested for antibodies to T. pallidum); and (iii) of Toxoplasma gondii. J. Clin. Microbiol. 27:2008-2013. reactive. 5. Kaplan, D. S., and G. L. Picciolo. 1989. Characterization of Use of QIF to prepare 1+ control sera. The intensity of the instrumentation and calibrators for quantitative microfluorom- 1+ standard is currently calibrated by the reagent manufac- etry for immunofluorescence tests. J. Clin. Microbiol. 27:442- turer by visually comparing dilutions of the control serum to 447. the intensity of the Centers for Disease Control 1 + reference 6. Kaufman, G. I., and J. F. Nester. 1972. Automation of the standard until an equivalent intensity is found (1, 7, 10). The fluorescent treponemal antibody-absorption test. J. Histochem. 1 + control sera may differ among lots from the same Cytochem. 20:255-260. manufacturer and especially between various manufactur- 7. Larsen, S. A., C. E. Farshy, B. J. Pender, M. R. Adams, D. E. Pettit, and E. A. Hambie. 1986. Staining intensities in the ers' kits. The QIF prepare method offers the opportunity to fluorescent treponemal antibody-absorption (FTA-ABS) test: the control sera by employing a quantitative, calibrated association with the diagnosis of syphilis. Sex. Transm. Dis. photometric system and, if validated by testing 1+ controls 13(4):221-227. from a number of manufacturers, could eliminate the neces- 8. Picciolo, G. L., and D. S. Kaplan. 1984. Reduction of fading of sity to match specific lots of control sera. fluorescent reaction product for microphotometric quantitation. Quantitatively determined intensities. The quantitatively Adv. Appl. Microbiol. 30:197-234. determined intensity values obtained for the labeled 9. Ploem, J. S. 1982. Automated methods in immunofluorescence treponemes showed that, although these cells emit a very studies, p. 73-94. In G. Wick, K. N. Trail, and K. Schauenstein weak fluorescence signal, the QIF method distinguishes (ed.), Immunofluorescence technology: selected theoretical and even weak signals from the electronic noise. The high clinical aspects. Elsevier Biomedical Press, New York. correlation 10. United States Department of Health, Education and Welfare. between the visually and quantitatively deter- 1969. Manual of tests for syphilis. Public Health Service publi- mined intensities validates the quantitative use of this tech- cation no. 411. National Communicable Disease Center, Vene- nique for the FTA-ABS test. real Disease Program, Atlanta.