Automated Technique for the Detection of Hemagglutinins to Sheep Erythrocytes

Transcription

1 APPLIED MICROBIOLOGY, May 1969, p Vol. 17, No. 5 Copyright 1969 American Society for Microbiology Printed in U.S.A. Automated Technique for the Detection of Hemagglutinins to Sheep Erythrocytes ROY CLEELAND AND E. GRUNBERG Department of Chemotherapy, Hoffmann-La Roche Inc., Nutley, New Jersey Received for publication 25 February 1969 The Technicon autoanalyzer system used for the detection of homologous human hemagglutinins has been modified for the detection of hemagglutinins to sheep red blood cells. Antibody quantitation was obtained by plotting the optical density () readings of peak heights of a serially diluted reference serum versus the serum dilutions. When the of the peak height of an unknown sample fell within the linear portion of the plot, a direct determination of the antibody titer was made. If the of the sample fell outside the linear portion, dilutions of the sample were carried out until a direct reading could be made. Assay by this method of at least 140 samples was possible within a day. Titers obtained with the autoanalyzer agreed very closely with those obtained by manual titration. Variation in humoral antibody production to red cells by randomly bred mice, differences in response of these animals to chemical agents, and the effect of proper timing of drug treatment all affect the ability of an agent to alter the immunological response. It is essential that numerous tests be carried out before establishing that a chemical agent actually exerts a consistent depressive or enhancing effect on humoral antibody production. Consequently, the time and labor involved in determining hemagglutinin titers by manual titrations, either by the macro or micro technique, greatly impede thorough testing of large numbers of chemical agents. The automated hemagglutination technique for measuring prenatal human antibody (R. E. Rosenfield and G. V. Haber, presented at the Technicon Symposium, New York, 1965) appeared to offer an answer to the problem. Our report describes the adaptation of this basic Technicon autoanalyzer system (Technicon Instruments Corp., Ardsley, N.Y.) for determination of hemagglutinin titers of serum from animals immunized with sheep red blood cells (SRBC). In addition, results on the use of the autoanalyzer in conjunction with agarose gel column chromatography for rapid, semiautomated separation of mouse IgM and IgG antibody are presented. SRBC. Cells were purchased (BBL) weekly as a washed, stabilized 10% cell suspension. These cells were used for periods of up to 1 week without further treatment. Immunization and bleeding. Animals were injected intraperitoneally (ip) with 0.2- or 0.5-ml amounts of a 1, 5, or 10% SRBC suspension prepared in 5% saline. Bleedings were done at 1, 2, and 3 weeks postimmunization. Animals were anesthetized with 30 to 50 mg of Nembutal (Abbott Laboratories, North Chicago, Ill.) per kg ip and were bled directly from the heart. Individual or pooled samples of blood were allowed to clot for 1 hr at 37 C and then for 18 to 24 hr at 4 C. Serum was removed by centrifugation and aspiration by employing a Pasteur pipette and, if not used immediately, was stored at -20 C. An alternative bleeding procedure was to cut the tips off the tails of mice and withdraw 5-jliter samples of blood in microliter pipettes which had been rinsed in 5% saline containing 100 units of heparin per ml. Samples from mice of the same group were then pooled in 2-ml autoanalyzer cups containing the heparinized saline to give a 1:10 dilution of the whole blood. The red cells were allowed to settle, and the supernatant plasma was sampled directly for antibody content. When plasma was to be stored for more than several days, it was aspirated from the cells by the use of a Pasteur pipette and stored at -20 C in capped vials. Final dilution of plasma was considered to be 1:20 to account for the red cell volume in the sample. Cells. The 10% SRBC suspension was diluted in MATERIALS AND METHS 0.9% saline containing sufficient Ficoll (mol wt Animals. Randomly bred Swiss albino mice (20 g) 400,000; Pharmacia Fine Chemicals, Inc., New York, were obtained from several commercial suppliers and N.Y.) to give a final concentration of 2.5% Ficoll. golden Syrian hamsters (60 to g) were purchased Ficoll solutions were prepared fresh daily. Final SRBC from Lakeview Farms. concentration was 6%, although concentrations of 5 726

2 VOL. 1 7, 1969 AUTOMATED DETECTION OF HEMAGGLUTININS 727 or 7% were sometimes required for certain cell lots. The concentration used was determined by obtaining a cell base line ranging between 0.7 and 0.9 optical density () units. Cells were kept in an ice bath and in suspension by use of a magnetic stirrer. Under these conditions, cell lysis as evidenced by drift in the cell base line was rarely observed for periods as long as 8 hr. Tween-saline. Saline (0.9%) containing 2 drops of Tween 20 per liter was employed for the initial reaction between SRBC and serum, since the addition of the Tween was necessary to lubricate the system and prevent irregular pulsing. Polyvinylpyrrolidone (PVP). PVP (K-90, mol wt 360,000) was prepared fresh daily as a 2% solution in 0.9% saline. Lysing agent. After removal of agglutinated cells, the residual SRBC were lysed with a 0.5% solution of Triton X-100 in distilled water or, in later tests, a 0.5% solution of 7-X detergent (Linbro Chemical Co., New Haven, Conn. in distilled water. hemagglutination tests. Standard macro tests were carried out by diluting the serum sample serially twofold in 0.5-ml saline blanks. A 0.5-ml volume of a 1%0 SRBC suspension prepared in saline was then added to all tubes. The tubes were shaken, incubated at room temperature for 1 to 2 hr, and read by the pattern method. The last tube to show positive hemagglutination was taken as the end point, and the titer was expressed as the reciprocal of the serum dilution giving this end point. Mercaptoethanol treatment of serum. Sera were divided into two samples, one treated with equal parts of phosphate-buffered saline (PBS; ph 7.2) and the other with equal parts of 0.2 M 2-mercaptoethanol (2-ME) made up in PBS. Mixtures were incubated for 2 hr at 26 C and then diluted for manual titration. Separation of IgM and IgG antibody. A K-15/90 column (1.5 by 90 cm) (Pharmacia Fine Chemicals, Inc.) was packed with agarose A-Sm gel (100 to 200 mesh) hydrated on M tris(hydroxymethyl) aminomethane (Tris)-ethylenediaminetetraacetic acid (EDTA) buffer (Bio Rad Laboratories, Richmond, Calif.). This gel separates proteins in a molecular weight range from 10,000 to 5,000,000. The column was packed according to the gel manufacturer's instructions to a height of approximately 70 cm. The column was then equilibrated with 0.1 M Tris buffer (ph 8.0) containing 5% NaCl. The void volume was determined with a Formalin-killed suspension of Proteus mirabilis. Serum (0.5 to 1 ml) was layered on top of the column and allowed to adsorb into the column, and 1 ml of 0.1 M Tris buffer was added to wash down residual serum. When the buffer wash was adsorbed, the column was filled with the buffer, and collection of fractions was started. Additional buffer was allowed to flow through the column by gravity from a reservoir. RESULTS Hemagglutination with the Technicon autoanalyzer. The automated analytical scheme (Fig. 1) is a modification from Technicon autoanalyzer methodology flow diagram N-57a. The red-cell suspension was pumped by the first proportioning pump continuously into a three-way fitting, into which samples to be analyzed were also pumped from the sampler. The serum-srbc mixture was then mixed at a second ml/min 2.0 FIG. 1. Flow diagram of autoanalvzer system for quantitative measurement of SRBC hemagglutination.

3 728 CLEELAND AND GRUNBERG APPL. MICROBIOL. 420 mp k Sample Sample I Serum dilutions Antibody units FIG. 2. Titration of hemagglutinins in mouse reference antiserum to SRBC including two unknown serum samples (I and II). three-way fitting with the PVP solution also being pumped continuously into the system, and this mixture was segmented at a PB-3 fitting with air. The segmented samples were mixed in a single four-turn coil, and the reaction was allowed to develop by passing the samples through four 10- min delay coils. Saline was added to the segmented samples emerging from the delay coils at a T fitting, and the resulting mixture was passed through an 1 1-turn mixing coil to break up nonspecific stacking of the red cells (Rouleau formation). The samples then were passed through a 2.5-turn settling coil, the agglutinated cells settling to the bottom of the coil. As the cells pass from the settling coil, a majority of the agglutinated cells were removed by gravity at a PT-6, three-way fitting. The remaining cells then were passed through a 1.5-turn settling coil, and more of the agglutinated cells were removed at a second three-way fitting (PT-5). Lysing agent was added to the residual cells, and lysis was completed by mixing in an 1-turn mixing coil. The resulting hemoglobin solution was passed through a second proportioning pump after being debubbled. The hemoglobin solution was again air-segmented, and additional lysing agent was added to dilute the hemoglobin in order to obtain a concentration which gives a satisfactory baseline value. Air-segmented samples were mixed in a 14-turn mixing coil, debubbled, and passed through a 15-mm tubular flow cell in the colorimeter. The at 420 nm was continuously recorded on a single-pen chart recorder. At the end of the day, the system was cleaned by passing a solution of 2.5% urea in 0.1 N NaOH through all lines for 15 to 20 min and rinsing for 30 to 50 min with deionized water. Use of standard reference serum for determination of antibody titers in defined units. Reference antisera were produced by immunizing groups of 40 mice or 40 hamsters, each with 0.5 ml of a 10% SRBC suspension ip. At 3 weeks postimmunization, the animals were bled from the heart; the serum was collected, pooled, and dispensed in 0.2-ml amounts into capped vials and stored at -20 C. Reference antiserum was then titrated in the following manner: 1:20, 1:40, 1:, 1:, 1:, 1:640, and 1:1,2 dilutions were made in 0.2 ml amounts of 0.9% saline in 2-ml autoanalyzer sample cups. The samples were placed, starting with the 1 :20 dilution, on the sampler and between each dilution a cup containing 1 ml of saline was inserted. The dilutions were then analyzed for antibody content (Fig. 2). The base line for this run was 6 units, and the last definite peak was at the 1: dilution, with a partial peak at the 1:640 dilution. The 1: dilution, therefore, was taken to represent 1 antibody unit; consequently, the 1: dilution would be 2 units, the 1: dilution 4 units, etc. The 1:640 dilution could therefore be considered to be 0.5 antibody unit. The of the peaks starting with the base line as zero was then plotted versus the antibody units (serum dilutions), and the best straight line was drawn through the plot (Fig. 3). All unknown samples which gave readings within the straight line portion of the graph could then be read directly in terms of antibody units, multiplied by whatever dilution factor was used for the unknown, and recorded as hemagglutinating units of antibody; e.g., two unknown sera (samples I and II) are shown at the end of the reference antiserum titration in Fig. 2. Sample I was diluted 1:20 and gave a peak value of. On the graph in Fig. 3, an value of represents 2.7 antibody units of reference serum. Since the dilution factor was 20, the total number of antibody units was 2.7 X 20 = 54. Sample II was also diluted 1:20. However, the value for the peak of this 420 ma 1.0 r Antibody units 1- Serum dilutions FIG. 3. Plot of antibody units versus at 420 nm

4 VOL. 17, 1969 AUTOMATED DETECTION OF HEMAGGLUTININS /hr CAM 420mD Without saline wash With saline wash between samples betwen samples 50/hr CAM 420m Without solins wash With soline wash 0.2 between samples between somples 0.3k I.O Reciprocal of Serum Dilutions I Reciprocal of Serum Dilutions 70/hr CAM 0m Without saline wash With 420m#v saline wash between samples between samples s Reciprocal of Serum Dilutions FIG. 4. Hemagglutination titrations of mouse antiserum to SRBC at different sampling rates with and without a saline wash between dilutions. sample was Since this value did not fall within the straight line portion of the graph, the sample would have had to be diluted an additional four times in saline and tested again. This process was repeated until the peak value fell within the straight line portion of the graph. Each day a serum standard was prepared as above, and a standard antibody titration was performed to correct for minor deviations in base line which occurred from day to day. If, as rarely happened, the base line shifted during the day, the standard was then plotted as antibody units versus A instead of direct readings. The values for unknown samples were then converted to A readings, and antibody units were determined as before. Effect of PVP and serum concentrations on hemagglutinaton. Addition of PVP was absolutely necessary for obtaining interaction between the red cells and antibody globulins, probably because it was necessary to neutralize the repellant charges on both reactants. For example, when the titration of the reference antiserum shown in Fig. 2 was carried out in the absence of PVP, only a small peak at the 1:20 dilution was observed, whereas in the presence of a 1% PVP concentration small peaks were seen at the 1:20, 1:40, and 1: dilutions. Increasing the PVP concentration to 3 or 4% (maximal concentration obtainable because of viscosity) resulted in an increased sensitivity over that observed with the 2% concentration as evidenced by both an increase in the height of each serum dilution peak and a definite peak at the 1:1,2 dilution. However, it was found that the plot of versus antibody units was linear over a greater range at the 2% PVP concentration. Since this meant that the antibody content of a greater number of serum samples could be directly determined without making additional dilutions, all further work was carried out at the 2% concentration of PVP. If undiluted normal serum or serum diluted 1:2 in saline was sampled, small inverted peaks extending downward toward an value of 2.0 resulted. The reason for the peak inversion was not known, but it did not appear to be due to free hemoglobin in the sample, since assay of normal serum free of obvious hemoglobin and normal serum containing appreciable amounts of hemaglobin gave similar peak inversions. Sera diluted 1:5 or higher did not show this inverted peak. In titrations of the standard reference antisera, the initial dilution was arbitrarily set at 1:20, since a dilution of 1:5 resulted in a peak height slightly lower than that at a 1:20 dilution, whereas a dilution of 1:10 gave a peak equal to the 1:20 dilution. No explanation for this phenomenon was readily available. Number of samples determined per hour. A serum standard was tested by the use of three

5 730 CLEELAND AND GRUNBERG APPL. MICROBIOL. TABLE 1. Comparison of hemagglutination titers of reference mouse or hamster antiserum as determined by manual titration and titration use of an autoanalyzer Mouseb Serum Hamsterb Hemagglutination titera 4 4 () 640 (2) 4 (2) 4 (230) a Reciprocal of last dilution showing hemagglutination manually or definite peak on the autoanalyzer. Values in parentheses indicate mean average. bfor each test, 40 animals were immunized with 0.5 ml of 10% SRBC suspension ip and bled at 3 weeks postimmunization, and sera from all animals in a group were pooled. sampling CAMs (40, 50, and 70 samples per hr with each CAM having a sampling to wash ratio of 2:1) with or without a saline wash cup between the samples (Fig. 4). The results show that, without a saline wash cup between samples, the residual serum carry-over markedly distorted the antibody peaks. In tests with unknown sera, this same distortion was observed especially if one serum sample having a high antibody content was followed by one with a low content. Use of the 70-samples-per-hr CAM did not permit sufficient sampling of the serum and greatly reduced the sensitivity of the test compared to the results using the 40- and 50-per-hr CAM. Since the 40- per-hr sampling CAM appeared to give slightly better sensitivity than the 50-per-hr CAM, all further tests were carried out by employing the 40-per-hr sample rate with a saline wash cup between each sample cup. Use of this sampling rate gave a final sample to wash ratio of 1 min of sampling (0.05 ml of serum) to 2 min of washing (0.1 ml of saline). Under these conditions in a normal day, approximately 140 samples could be analyzed. In order to help identify the peaks during this screening procedure, a standard marker of known antibody was placed in every 20th cup. TABLE 2. Comparison of the hemagglutination titers of individual mice determined by manual titration and by titration by the use of the autoanalyzera Mouse no.d < ( (58) (48) Weeks postimmunization <10 (107) (82) 640 <40 <20 40 c (162) (146) a For manual titrations, titers are expressed as the reciprocal of the last dilution showing hemagglutination. For autoanalyzer tests, the titer is expressed in antibody units determined by comparison with specific reference antiserum. Values in parentheses indicate mean averages. b Each mouse was immunized with 0.2 ml of a 1% SRBC suspension ip. c Mouse died; no serum was obtained. Comparison of the sensitivity of autoanalyzer titration with that of manual titration. To compare the sensitivity of the autoanalyzer titrations with those obtained by manual titration, standard reference sera obtained from different groups of mice or hamsters were diluted 1:20, 1:40, 1:, 1:, 1:, 1:, 1:4, 1:640, and 1:1,2 in 0.9% saline. The dilutions were then tested both manually and on the autoanalyzer with both titrations on a single pooled antiserum being done on the same day with the same lot of red blood cells; results shown in Table 1 are given as the reciprocals of the last dilution to show a definite peak on the autoanalyzer or positive hemagglutination by manual titration. From these results, it is evident that the titers obtained with the autoanalyzer closely agree but, on the average, tend to be slightly lower than those found by manual titration. Numerous other tests from other groups of mice always gave the same type of consistent results, the variation being what

6 VOL. 1 7, 1969 AUTOMATED DETECTION OF HEMAGGLUTININS 731 TABLE 3. Comparison of the hemagglutination titers ofpooled plasma from groups of mice bledfrom the tails at weekly intervals as determined by manual titrations and titrations by the use of the autoanalyzer- Immunizing dose of SRBCb 1% 5% (70) (120) (41) (98) Weeks postimmunization 40 (130) (140) (148) (153) 40 (90) () D.1 0 I1 2 3 Auto (78) (133) a See Table 2 for definition of titers. Values in parentheses indicate mean averages. b Each mouse was immunized with 0.2 ml ip. c Eight separate tests were performed on 10 mice per group. Each mouse was bled from the tail (5 gliter per mouse) and blood was pooled in heparinized saline to give a final plasma dilution of 1:20. would be expected in any standard serological test. A series of experiments was then set up to compare the serum titers of individual mice by both serological methods. Each of 45 mice was immunized with 0.2 ml of a 1 % SRBC suspension ip. Groups of 15 animals were individually bled from the heart at 1, 2, and 3 weeks postimmunization. Each serum was diluted 1:5 or 1:10 in 0.9% saline (in some cases because of the small volume collected the initial serum dilution was 1:20 to 1:40) and was tested directly on the autoanalyzer for determinations of antibody units. The same serum sample was also tested by manual titration and the titers obtained were compared (Table 2). Whereas appreciable individual variations occurred in the two methods, the mean average titers were very similar for both methods. Again, the manual results tended to be somewhat higher on the average than those obtained with the autoanalyzer. In all cases, when negative hemagglutination results were obtained by the autoanalyzer assay, the sera were also negative at the lowest dilution tested manually. In one instance, at 3 weeks, mouse no. 7 had a manual titer of less than 40 and an autoanalyzer titer of 122. This type of difference has occurred only rarely in numerous other tests, and as yet the reason for the discrepancy is not known. Individual variation among mice in antibody responses was detected by both methods equally well. Individual heart bleedings are time-consuming and, since they usually result in the death of the animal, preclude following the antibody response over a number of weeks. It was, therefore, of interest to compare the sensitivity of both methods in determining antibody content of pools of plasma obtained by tail bleeding at weekly intervals after immunization. Groups of 10 mice were immunized each with 0.2 ml of a 1 or 5 % SRBC suspension ip. At 1, 2 and 3 weeks postimmunization, the mice from each group were bled from the tails, the plasma was pooled, and the antibody titers were determined both manually and on the autoanalyzer. Eight experiments were performed, four with animals immunized with 1 % and four with 5% SRBC suspension (Table 3). At 1 week, the average titers obtained manually were higher than those determined by autoanalyzer assay. In experiments 3 and 4, the discrepancy in titer after 1 week was quite significant, the manual titer being five- to eightfold higher. This type of discrepancy can result when plasma or serum with a low antibody content is sufficiently diluted before testing so that the resulting of the peaks obtained by assay with the autoanalyzer is very close to the base line. (In this instance the initial dilutions were 1:20.) Reading these peaks becomes difficult and a slight error in reading can result in a substantial error in the actual antibody titer. After 2 weeks, the titers by both methods were quite comparable and remained so at the 3-week bleeding. Increasing the immunizing dose from 1 to 5% increased, as expected, the antibody response at the 1-week interval, maintained the response over the 3-week interval, and did not change the correlation in antibody titers between the two methods. Separation of IgM and IgG mouse antibody to SRBC. Mice were immunized ip with 0.2 ml of 5% SRBC suspension and bled from the heart at 5 days or 3 weeks postimmunization. Treatment of the pooled serum from the 5-day bleeding with 2-ME reduced the manual hemagglutination titer from 1 :40 to less than 1:5, whereas treatment of the 3-week serum with 2-ME did not change the hemagglutination titer from 1:. A 0.5 ml portion of the 2-ME-sensitive and 0.5 ml of the 2-ME-resistant serum were mixed and chromatographed on the agarose column using Tris-saline buffer. Fractions were collected directly onto the autoanalyzer sampler, and hemagglutinating activity was determined. The remainder of selected fractions was removed from

7 732 CLEELAND AND GRUNBERG APPL. MICRoBioL. A 420my *-4 Hemagglutination A 420mp X---X 2 mp 2-ME resistant antibody 2mp ME sensitive antibody FRACTION NO. 133 FIG. 5. Separation of mouse IgM and IgG antibody to SRBC by column chromatography by using agarose gel (results ofone test run). Hemagglutination measurements offractions were carried out by direct sampling on the autoanalyzer at 420 nm. at 2 nm was measured on a Beckman D U spectrophotometer the sample cup and diluted 1:10 with buffer, and the at 2 nm was measured on the DU spectrophotometer (Beckman Instruments, Inc., Fullerton, Calif.; Fig. 5). Fractions 46 through 54 gave both a hemagglutination reaction, the peak occurring at fraction 48, and a discernible protein reading with the peak again at fraction 48. A second hemagglutination reaction was observed between fractions 84 and 119 with the peak occurring at fraction 100. The protein readings rose from fraction to a peak at fraction 112. The protein readings returned to zero by fraction 133. When the 2-ME-sensitive (IgM) serum was tested, only one hemagglutination peak occurred at fraction 48. The 2-ME-resistant (IgG) serum, as expected, showed only a single broad peak similar to that shown in Fig. 5 for the second hemagglutination peak. Flow rate of the column was 45 ml per hr, and each fraction collected contained on an average 4 ml, which meant that the IgM antibody came off the column within approximately 1 hr. Total time for assay for hemagglutination was approximately 56 min, consequently the separation and assay for IgM antibody by the use of the combination of the agarose column and autoanalyzer required a total time of about 2 hr. In later tests, fractions were passed through the drop counter of a Fractomette 200 fraction collector (Buchler Instruments, Inc., Fort Lee, N.J.) directly into 2-ml cups on the autoanalyzer sampler after having been scanned at 2 nm by use of a Uviscan III (Buchler Instruments, Inc.) and recorded on a Recti recorder (Texas Instruments, Inc.). Thus, both the at 2 nm and hemagglutinating activity were simultaneously determined. Essentially the same results (Fig. 5) were obtained. DISCUSSION Use of the Technicon autoanalyzer system employing the hemagglutination manifold for the measurement of hemagglutinins to SRBC has been found to be both convenient and reproducible. Our data show that the antibody unit values obtained by autoanalyzer assay agree closely with but tend to be slightly less than those obtained by manual titration. Increasing the PVP concentration from 2 to 3% would probably make the autoanalyzer titers higher if complete agreement were necessary, but the increased sensitivity has the drawback that more dilutions of the serum samples would be required to obtain accurate antibody unit values. Modifications of the original system (R. E. Rosenfield and G. V. Haber, presented at the Technicon Symposium, New York, 1965) used for screening human prenatal antibody include lowering the red blood cell concentration from 30 to 6%, omission of proteolytic enzyme treatment of red cells, insertion of several additional delay coils to allow the reaction of cells and antibody to ripen, and the addition of a small amount of Tween 20 to saline in place of albumin. Use of this system has several major advantages over conventional manual titrations. A permanent record of each serum hemagglutination is obtained which, except at very weak antibody concentrations, is not based on the subjective reading of end points encountered in standard manual titrations. Once a particular range of anti-

8 VOL. 17, 1969 AUTOMATED DETECTION OF HEMAGGLUTININS 733 body titers for a given amount of antigenic stimulus is determined, with most of the test serum samples, only one or two dilutions need be employed to obtain final results. After the dilutions of sera have been made, only replacement of the sample trays and an occasional check on operation are required during the day. Cleaning the system at the end of the day and routine replacement of tubing and cleaning of glass parts have been the only normal maintenance needed in almost 2 years of continuous operation. In addition, washed stabilized SRBC can be used without further treatment and, under proper conditions, show little or no evidence of spontaneous lysis. Disadvantages encountered with the autoanalyzer are the inability to use undiluted serum or serum diluted 1:2 to obtain a titer and the difficulty in accurately determining whether a weak antibody peak is actually antibody or part of a slightly irregular base line.