- Email: [email protected]
A spectrophotometric method for evaluating a latex agglutination assay of Salmonella typhi lipopolysaccharide
Journal of Immunological Methods, 115 (1988) 269-274
269
Elsevier JIM 04993
A spectrophotometric method for evaluating a latex agglutination assay of Salmonella typhi lipopolysaccharide Pak-Leong Lim and Wai-Fun Choy Department of Microbiology, University of Hong Kong~ Pokfulam Road, Hong Kong, Hong Kong (Received 8 February 1988, revised received 20 June 1988, accepted 4 August 1988)
A spectrophotometer set at a wavelength of 400 nm was used to read reaction mixtures containing a 0.01% suspension of antibody-sensitized latex particles and different amounts of soluble antigen, after incubation in tubes for 30 min. The antigen used was Salmonella typhi lipopolysaccharide, and the antibody was an O-9-specific monoclonal antibody. Agglutination was indicated by a fall in the turbidimetric measurement compared to control, unagglutinated latex. It was observed that agglutination increased with increasing concentrations of antigen to a maximum at 0.25-1.0 /~g/ml, after which, the amount of agglutination decreased at a similar rate as far as 1 m g / m l of antigen, when no agglutination occurred at all. Thus, a bi-symmetrical curve was obtained, suggestive of a precipitation reaction. At the 'equivalence point', the turbidity of the reaction mixture compared to control fell from OD4o0 1.1 to OD40o 0.3. A similar inhibition of reaction at 'antigen excess' was observed visually in the reaction mixtures. Conventional slide tests performed in parallel could also be inhibited, but at higher antigen levels. Inhibition could be achieved with a commercially available latex used for the detection of Neisseria meningitidis antigen. Key words: Latex agglutination; Turbidimetry; Monoclonal antibody, 0-9; Salmonella typhi lipopolysaccharide; Antigen detection
Introduction
Latex particles sensitized with antibodies or antigens are widely used in slide agglutination tests for the detection of antigens or antibodies from biological fluids (Thomas, 1986). Indeed, tests employing antibody-latex conjugates (many of which are commercially available) have become increasingly popular for the diagnosis of a variety of clinical conditions, including bacterial meningitis (Marcon et al., 1984), streptococcal group A throat infection (Graham et al., 1986), and rota-
Correspondence to: P.L. Lim, Department of Microbiology, University of Hong Kong, Pokfulam Road, Hong Kong, Hong Kong.
viral gastroenteritis (Sambourg et al., 1985). The popularity of this technique is due to its low cost, simplicity (a one-step reaction) and speed (less than 15 min). It is thus well suited for clinical use. However, the main drawbacks of the test are: (1) its sensitivity is less than desirable in comparison with enzyme-linked i m m u n o s o r b e n t assays (ELISA), and (2) interpretation of its end-point can be subjective. As a result, several attempts have been made to improve these features of the test using instrumentation to read the results. Turbidimetry (Dezelic et al., 1971), light scattering (Blume and Greenberg, 1975) and particle counting (Cambiaso et al., 1977) have all yielded results in terms of sensitivity and objectivity. More recently, Miotti et al. (1986) used antibody-latex
0022-1759/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
270 conjugates instead of enzyme labels to develop an antigen-capture assay on microplates, analogous to the traditional solid-phase ELISA. Results were read either visually or using a microplate reader, and high sensitivities were obtained. However, none of these modifications have gained much popularity because they make the test more cumbersome, than of the traditional method. We recently modified the latex agglutination technique by performing the test in tubes rather than on slides. This simple modification retained some of the basic features of the slide method, but increased the sensitivity of the test at least four fold, and the results could be read more easily (Lim and Fok, 1987). The minimal reaction time used was 1 h. In the present study, we show that this time can be shortened, and the results can be objectively read using a spectrophotometer, which further improves the sensitivity of the test.
Materials and methods
Reagents Latex particles (diameter, 0.797 ttm, Sigma Chemical Co., Poole, England) sensitized with Salmonella 0 - 9 monoclonal immunoglobulin M (IgM) antibodies (Lim, 1986) were prepared as described previously (Lim and Fok, 1987). The Neisseria meningitidis B/Eseherichia coli K1 latex agglutination kit (Wellcogen) was obtained from Wellcome Diagnostics, Dartford, England. Salmonella typhi lipopolysaccharide (LPS) was purchased from Difco Laboratories, Detroit, MI. N o r m a l urine was obtained and pooled from five healthy laboratory workers in our Department.
Tube latex agglutination test As described previously (Lim and Fok, 1987), 1 ml (or 0.1 ml in some cases) 0.1 M glycine-0.9% sodium chloride, buffer ( p H 8.2) or urine, was delivered into a screw-capped glass vial (Flow Laboratories, Uxbridge, Middlesex, England) and spiked with a known amount of S. typhi LPS. Sensitized latex suspension was then added to a final concentration of 0.00125%-0.04%. The mixture was incubated on a mixer (Coulter Electronic, Luton, Bedfordshire, England) for 30 min at room temperature. The mixer (cat. no. C10) consisted of
a series of horizontal cylindrical rollers (radius, 17 mm; length, 32 cm) and the tube placed flat between two of these for rotation about the horizontal axis at a speed of 50 rpm. The results were read immediately by naked eye. Positive reactions were scored on a decrease in turbidity of the suspension compared with that of the control and the presence of agglutinated material at the bottom of the tube; these were graded ( + , + , + + , + + + ) according to the prominence of these features. Following this, the top half (0.5 ml) of the reaction mixture was transferred to a 10 m m quartz cuvette (Pye Unicam, Cambridge, England) for the determination of absorbance in a spectrophotometer (DMS 90, Varian Techtron Pty, Mulgrave, Australia).
Sfide agglutination test This was performed as described previously (Lim and Fok, 1987) using 30 /zl test sample and 10 txl of a 1% latex suspension. The mixture, placed on a reaction card (Wellcome), was incubated for 5 min on a flat-bed Clinical Rotator (Model 2995-G10, Arthur H. T h o m a s Co., Philadelphia, PA) at 120 rpm. Results were then read and scored on the degree of clumping.
Results
The absorption spectrum of the 0 - 9 latex particles, as determined using two separate batches of latex, showed an absorption m a x i m u m at about 400 nm. This wavelength was consequently chosen for use in subsequent studies. At this wavelength, the absorption of the latex suspension increased proportionally with its particle concentration (Fig. 1). Thus, for a 0.01% suspension (the concentration normally used in our visual test), the OD4oo was about 1.1. When this concentration (final) of latex was used with a range of concentrations (0-1 m g / m l ) of specific antigen (S. typhi LPS) in tube agglutination assays, the results obtained from reading the reaction mixtures in a spectrophotometer set at a wavelength of 400 nm, or by visual inspection, are shown in Fig. 2. There was general agreement between the two methods of examination, although the spectrophotometric method, which detected at least 4 n g / m l of anti-
271
gen, was 2-4-fold more sensitive than visual reading. This, when compared with our conventional slide agglutination method, was eight-fold more sensitive. In the spectrophotometric method, the degree of agglutination increased with increasing concentrations of antigen used, up to 250 n g / m l . Maximum agglutination was observed in the range between 250 n g / m l and 1 /~g/ml of antigen. At concentrations higher than this, agglutination decreased with increasing amounts of antigen and no agglutination occurred at 1 m g / m l or more of antigen. Parallel results were obtained with visual inspection of the reaction mixtures, although the point of maximal agglutination was less distinct but seemed biased to slightly higher concentrations of antigen (up to 16 # g / m l ) . Inhibition of agglutination by high concentrations of antigen was also seen in the conventional slide assay (Fig. 2) but total inhibition of agglutination required a much higher concentration of antigen (32 mg/ml). Similar inhibitory reactions at antigen excess were also seen using another latex system (Table I). In this commercially obtained system, the antigen used was derived from N. meningitidis, while the latex particles were sensitized with an IgM monoclonal antibody raised against the organism. Note, however, that the test was not performed as recommended by the manufacturer and also that the
2 0-
1 8
1 6
I tl
1 2
o 0.8
0.6
0.4
0.2
0-
I
I
I
0. 00125
1
I
0.005
latex
0.02
c o n c e n t r a t i o n (%)
Fig. 1. Absorbance of different concentrations of the 0 - 9
antibody-latex conjugate in buffer at 400 nm wavelength.
Slide
-
+
+
4.1-
-tfl-
Tube
ND
ND
ND
ND
ND
~
~
--
-I-
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
-fit-
+1-
+
--
+
+
-I¢-
-if-
~
~
-tit-
-fit-
~
~
4-H-
-It-
4+-
+
+
4-
1.0
0.8 0
o
0.6
a O 0.4
0.2
I
I
I
I
I
I
A
I
I
I
I
I
I
I
32
16
4
1
250
63
16
4
1
250
63
16
4
0
mg
per
ml
JJg per ml Antigen
ng per
ml
concentration
S. typhi LPS in buffer. Results of tube tests done in triplicate based on visual examination (average) or on spectrophotometric measurement (mean _+ SD) are shown. Also displayed are the results of slide agglutination tests performed in parallel. - , no agglutination; 4- to + + + , increasing agglutination; N D = not done. Fig. 2. Agglutination of the 0 - 9 antibody-latex conjugate (0.01%) by a range of concentrations of
272 _+
+
q-t-
I 250
I 63
I 16
4+
-IF
4+
+t-
4+
4+
+
+
3.0
2.6
2.2 O
£3 O
1.8
1.4
1.0 000
I 4
I 1
I 250
I 63
I 16
#g per ml
I 4
I 0
ng per ml Antigen c o n c e n t r a t i o n
Fig. 3. Effect of using increased concentration of the 0-9 antibody-latex conjugate (0.04%) on its agglutination by S. typhi LPS in buffer. Results of visual examination of tube reactions are also shown. Other legends as in Fig. 2.
concentration of antigen used in the kit as a positive control was not within the inhibitory concentrations observed in our test. When the S. typhi titration was run using a four-fold more concentrated suspension of the 0 - 9 latex particles, the U-shaped profile was again seen based on the spectrophotometric readings (Fig. 3), and the point of maximal agglutination _+
+
4+
+1-
was not significantly different. There was a high background of unagglutinated latex in all tubes causing a loss of sensitivity particularly in the visual readings. The possibility of measuring the agglutination reaction in urine instead of buffer was examined. Pooled normal urine samples spiked with different amounts of S. typhi LPS, were incubated with +H-
J-H-
+H-
+H-
4+
4+
4"t-
+
+
I 63
I 16
I 4
1.0
0.8
O
E3 O
0.6
0.4 ==
0.2 I 1000
I 250
I 63
I 16
I 4
I 1
#g per ml
I 250
I 0
ng per ml Antigen
concentration
Fig. 4. Effect of urine on the agglutination of the O-9 antibody conjugate (0.01%) by S. typhi LPS. Conditions of reaction as in Fig. 2, except that undiluted normal human urine spiked with antigen was used instead of buffer. Results of visual examination of tube reactions are also shown.
273 TABLE I D E T E C T I O N OF N. M E N 1 N G I T I D I S A N T I G E N IN B U F F E R BY A G G L U T I N A T I O N W I T H ANTIBODYSENSITIZED LATEX PARTICLES: EFFECT OF ANTIGEN CONCENTRATION Antigen dilution a
Agglutination results b Tube method
Slide method c
1/2
-
-
1/4
+
1/8 1/16 1/32 1/64 1/128 1/256 1/512 1/1024 1/2048 1/4096 1/8192 1/16384 1/32768 No antigen
+ + + + + + + + + + + -
+
+ + + + + + +
+ + + +
++ ++ ++ ++ 4ND ND ND ND ND ND -
a Of stock solution made by dissolving contents in vial (as supplied) with 0.2 ml distilled water (instead of 1.0 ml as recommended by manufacturer). b From visual examination. --, no agglutination; + to + + + , increasing agglutination; N D = not done. c As described in text, using 30 /tl antigen and 10 ~1 latex suspension. (Method recommended by manufacturer uses 40 /~1 antigen and 25/.tl latex.)
0.01% 0-9 latex suspension. The results were similar to those obtained using buffer (Fig. 4).
Discussion We have previously described a tube modification of the latex agglutination method which showed improved sensitivity and readability (Lim and Fok, 1987). Since using this method, agglutination of the latex particles was associated with a decrease in turbidity of the latex suspension, it occurred to us that the results of the test could perhaps be read by transmission spectrophotometry. The present investigation confirmed that this was possible. The reaction time used was relatively short (30 min) in comparison with the photometric methods previously described, though longer than the con-
ventional slide test. Although the reaction mixtures were transferred to cuvettes to be read separately in a spectrophotometer in this study (and this is an inconvenience), it would be possible in future to design suitable reaction tubes for direct use in specially designed spectrophotometers capable of reading small volumes of fluid, say 0.1 ml. The microplate readers currently in use for ELISA systems can be adapted for use with the latex particles. We have experimented with the MR710 (Dynatech, Alexandria, VA) and the Titertek Uniskan (Flow Laboratories, North Ryde, NSW) readers for this purpose, using microplates to contain the test suspensions, and found them to be moderately satisfactory (unpublished observations). However, in our experience, the microplates are not ideal reaction vessels. Suitable conditions for reading the latex agglutination results by spectrophotometry were found. The wavelength chosen was 400 nm, which is the absorption maximum of the 0 - 9 latex conjugate, as well as of the unconjugated latex particles (data not shown). A latex concentration of 0.01% was found to be suitable for this purpose, and is the concentration found to be ideal for the visual tube agglutination test (Lim and Fok, 1987). A surprising finding from the turbidimetric measurements was the U-shaped profile of the reaction (Figs. 2-4). Although inhibition of reaction at antigen excess is seen in many immunoassay systems and commonly referred to as a prozone phenomenon, the effect seen in our case was rather pronounced and spread over many antigen dilutions. The plots obtained were bi-symmetrical, analogous to precipitation reactions where the points of maximal agglutination are equivalence points. The reaction may be considered a precipitation since soluble and not particulate antigen is involved. Inhibition at 'antigen excess' is not a peculiarity of the turbidimetric measurement used since it was also observed in both the tube and slide methods of latex agglutination, though the effects were less pronounced. Nor is it a peculiarity of our latex system, since a modified commercially available system which used another IgM monoclonal antibody to detect N. meningitidis antigen, also showed it (Table I). On the other hand, we have not observed a prozone effect with whole S.
274
typhi bacteria at concentrations as high a s 1012 o r g a n i s m s / m l using the 0 - 9 latex conjugate in a slide test. This is the first time an inhibition of latex agglutination has been documented. The clinical significance of this is unknown, since, particularly for the slide test, high concentrations ( m g / m l ) of antigen are required to inhibit the reaction. However, since latex agglutination is widely used, e.g., for diagnosing meningitis, and since clinical samples are sometimes concentrated before testing, the observation is nevertheless noteworthy. The phenomenon unfortunately invalidates any attempt at using spectrophotometric measurement to quantitate latex agglutination, unless serial dilutions of the test sample are made to locate the 'equivalence point' of the assay.This is a preliminary report confirming the potential use of instrumental reading for latex agglutination. Other antigen detection systems employing both monoclonal and polyclonal reagent antibodies, as well as other types of latex particles, will have to be examined to determine the generality of the application. With this improvisation in our system, the test becomes more sensitive than the visual tube method, and is at least eight-fold more sensitive than the conventional slide test when used under optimal conditions (Fig. 2). This, admittedly, is at the expense of rapidity and simplicity of the conventional slide method. Whether it is also at the expense of specificity as a result of the improved sensitivity, remains a possibility. Nevertheless, unlike the visual methods in which the end-point of the assay is often ill-defined, the modified test gives objective results. Furthermore, it may be possible to introduce automation to latex agglutination assays, which would be particularly useful for screening large numbers of samples at a time. We have shown in this report that urine presents no problems when used as diluent, including the instrumental reading of the results (Fig. 4). This source of material, which is convenient to obtain, has proven to be immensely useful for the diagnosis of various infectious diseases (Coonrod, 1983). However, a range of clinical specimens will have to be used for a proper evaluation of the reliability of our method, since these can behave quite differently from spiked urine samples, especially when these involve the use of stored and pooled urine, as in our case.
Acknowledgements We thank Otis K.H. Ko for skilled technical assistance and Candy Leung for typing the manuscript.
References Blume, P. and Greenberg, L.J. (1975) Application of differential light scattering to the latex agglutination assay for rheumatoid factor. Clin. Chem. 21, 1234. Cambiaso, C.L., Leek, A.E., De Steenwinkel, F., Billen, J. and Masson, P.L. (1977) Particle counting i m m u n o a s s a y (PACIA). I. A general method for the determination of antibodies, antigens, and haptens. J. Immunol. Methods 18, 33. Coonrod, J.D. (1983) Urine as an antigen reservoir for diagnosis of infectious diseases. Am. J. Med. 75 (suppl. 1B), 85. Dezelic, G., Dezelic, N., Muic, N. and Pende, B. (1971) Latex particle agglutination in the immunochemical system hum a n serum albumin-anti-human serum albumin rabbit serum. Eur. J. Biochem. 20, 553. Graham, Jr., L., Meier, F.A., Centor, R.M., Garner, B.K. and Dalton, H.P. (1986) Effect of medium and cultivation conditions on comparisons between latex agglutination and culture detection of group A streptococci. J. Clin. Microbiol. 24, 644. Lim, P.L. (1986) Diagnostic uses of monoclonal antibodies to Salmonella. In: A.J. Macario and E.C. de Macario (Eds.), Monoclonal Antibodies Against Bacteria, Vol. 3. Academic Press, New York, p. 29. Lira, P.L. and Fok, Y.P. (1987) Detection of group D salmonellae in blood culture broth and of soluble antigen by tube agglutination using an 0 - 9 monoclonal antibody latex conjugate. J. Clin. Microbiol. 25, 1165. Marcon, M.J., Hamoudi, A.C. and Cannon, H.J. (1984) Comparative laboratory evaluation of three antigen detection methods for diagnosis of Haemophilus influenzae type b disease. J. Clin. Microbiol. 19, 333. Miotti, P.G., Viscidi, R.P., Eiden, J., Cerny, E. and Yolken, R.H. (1986) Centrifugation-augmented solid-phase immunoassay (CASPIA) for the rapid diagnosis of infectious diseases. J. Infect. Dis. 154, 301. Sambourg, M., Goudeau, A., Courant, C., Pinon, G. and Denis, F. (1985) Direct appraisal of latex agglutination testing, a convenient alternative to enzyme i m m u n o a s s a y for the detection of rotavirus in childhood gastroenteritis, by comparison of two enzyme immunoassays and two latex tests. J. Clin. Microbiol. 21,622. Thomas, M. (1986) Agglutination methods for rapid analysis. Nature 320, 289.
269
Elsevier JIM 04993
A spectrophotometric method for evaluating a latex agglutination assay of Salmonella typhi lipopolysaccharide Pak-Leong Lim and Wai-Fun Choy Department of Microbiology, University of Hong Kong~ Pokfulam Road, Hong Kong, Hong Kong (Received 8 February 1988, revised received 20 June 1988, accepted 4 August 1988)
A spectrophotometer set at a wavelength of 400 nm was used to read reaction mixtures containing a 0.01% suspension of antibody-sensitized latex particles and different amounts of soluble antigen, after incubation in tubes for 30 min. The antigen used was Salmonella typhi lipopolysaccharide, and the antibody was an O-9-specific monoclonal antibody. Agglutination was indicated by a fall in the turbidimetric measurement compared to control, unagglutinated latex. It was observed that agglutination increased with increasing concentrations of antigen to a maximum at 0.25-1.0 /~g/ml, after which, the amount of agglutination decreased at a similar rate as far as 1 m g / m l of antigen, when no agglutination occurred at all. Thus, a bi-symmetrical curve was obtained, suggestive of a precipitation reaction. At the 'equivalence point', the turbidity of the reaction mixture compared to control fell from OD4o0 1.1 to OD40o 0.3. A similar inhibition of reaction at 'antigen excess' was observed visually in the reaction mixtures. Conventional slide tests performed in parallel could also be inhibited, but at higher antigen levels. Inhibition could be achieved with a commercially available latex used for the detection of Neisseria meningitidis antigen. Key words: Latex agglutination; Turbidimetry; Monoclonal antibody, 0-9; Salmonella typhi lipopolysaccharide; Antigen detection
Introduction
Latex particles sensitized with antibodies or antigens are widely used in slide agglutination tests for the detection of antigens or antibodies from biological fluids (Thomas, 1986). Indeed, tests employing antibody-latex conjugates (many of which are commercially available) have become increasingly popular for the diagnosis of a variety of clinical conditions, including bacterial meningitis (Marcon et al., 1984), streptococcal group A throat infection (Graham et al., 1986), and rota-
Correspondence to: P.L. Lim, Department of Microbiology, University of Hong Kong, Pokfulam Road, Hong Kong, Hong Kong.
viral gastroenteritis (Sambourg et al., 1985). The popularity of this technique is due to its low cost, simplicity (a one-step reaction) and speed (less than 15 min). It is thus well suited for clinical use. However, the main drawbacks of the test are: (1) its sensitivity is less than desirable in comparison with enzyme-linked i m m u n o s o r b e n t assays (ELISA), and (2) interpretation of its end-point can be subjective. As a result, several attempts have been made to improve these features of the test using instrumentation to read the results. Turbidimetry (Dezelic et al., 1971), light scattering (Blume and Greenberg, 1975) and particle counting (Cambiaso et al., 1977) have all yielded results in terms of sensitivity and objectivity. More recently, Miotti et al. (1986) used antibody-latex
0022-1759/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
270 conjugates instead of enzyme labels to develop an antigen-capture assay on microplates, analogous to the traditional solid-phase ELISA. Results were read either visually or using a microplate reader, and high sensitivities were obtained. However, none of these modifications have gained much popularity because they make the test more cumbersome, than of the traditional method. We recently modified the latex agglutination technique by performing the test in tubes rather than on slides. This simple modification retained some of the basic features of the slide method, but increased the sensitivity of the test at least four fold, and the results could be read more easily (Lim and Fok, 1987). The minimal reaction time used was 1 h. In the present study, we show that this time can be shortened, and the results can be objectively read using a spectrophotometer, which further improves the sensitivity of the test.
Materials and methods
Reagents Latex particles (diameter, 0.797 ttm, Sigma Chemical Co., Poole, England) sensitized with Salmonella 0 - 9 monoclonal immunoglobulin M (IgM) antibodies (Lim, 1986) were prepared as described previously (Lim and Fok, 1987). The Neisseria meningitidis B/Eseherichia coli K1 latex agglutination kit (Wellcogen) was obtained from Wellcome Diagnostics, Dartford, England. Salmonella typhi lipopolysaccharide (LPS) was purchased from Difco Laboratories, Detroit, MI. N o r m a l urine was obtained and pooled from five healthy laboratory workers in our Department.
Tube latex agglutination test As described previously (Lim and Fok, 1987), 1 ml (or 0.1 ml in some cases) 0.1 M glycine-0.9% sodium chloride, buffer ( p H 8.2) or urine, was delivered into a screw-capped glass vial (Flow Laboratories, Uxbridge, Middlesex, England) and spiked with a known amount of S. typhi LPS. Sensitized latex suspension was then added to a final concentration of 0.00125%-0.04%. The mixture was incubated on a mixer (Coulter Electronic, Luton, Bedfordshire, England) for 30 min at room temperature. The mixer (cat. no. C10) consisted of
a series of horizontal cylindrical rollers (radius, 17 mm; length, 32 cm) and the tube placed flat between two of these for rotation about the horizontal axis at a speed of 50 rpm. The results were read immediately by naked eye. Positive reactions were scored on a decrease in turbidity of the suspension compared with that of the control and the presence of agglutinated material at the bottom of the tube; these were graded ( + , + , + + , + + + ) according to the prominence of these features. Following this, the top half (0.5 ml) of the reaction mixture was transferred to a 10 m m quartz cuvette (Pye Unicam, Cambridge, England) for the determination of absorbance in a spectrophotometer (DMS 90, Varian Techtron Pty, Mulgrave, Australia).
Sfide agglutination test This was performed as described previously (Lim and Fok, 1987) using 30 /zl test sample and 10 txl of a 1% latex suspension. The mixture, placed on a reaction card (Wellcome), was incubated for 5 min on a flat-bed Clinical Rotator (Model 2995-G10, Arthur H. T h o m a s Co., Philadelphia, PA) at 120 rpm. Results were then read and scored on the degree of clumping.
Results
The absorption spectrum of the 0 - 9 latex particles, as determined using two separate batches of latex, showed an absorption m a x i m u m at about 400 nm. This wavelength was consequently chosen for use in subsequent studies. At this wavelength, the absorption of the latex suspension increased proportionally with its particle concentration (Fig. 1). Thus, for a 0.01% suspension (the concentration normally used in our visual test), the OD4oo was about 1.1. When this concentration (final) of latex was used with a range of concentrations (0-1 m g / m l ) of specific antigen (S. typhi LPS) in tube agglutination assays, the results obtained from reading the reaction mixtures in a spectrophotometer set at a wavelength of 400 nm, or by visual inspection, are shown in Fig. 2. There was general agreement between the two methods of examination, although the spectrophotometric method, which detected at least 4 n g / m l of anti-
271
gen, was 2-4-fold more sensitive than visual reading. This, when compared with our conventional slide agglutination method, was eight-fold more sensitive. In the spectrophotometric method, the degree of agglutination increased with increasing concentrations of antigen used, up to 250 n g / m l . Maximum agglutination was observed in the range between 250 n g / m l and 1 /~g/ml of antigen. At concentrations higher than this, agglutination decreased with increasing amounts of antigen and no agglutination occurred at 1 m g / m l or more of antigen. Parallel results were obtained with visual inspection of the reaction mixtures, although the point of maximal agglutination was less distinct but seemed biased to slightly higher concentrations of antigen (up to 16 # g / m l ) . Inhibition of agglutination by high concentrations of antigen was also seen in the conventional slide assay (Fig. 2) but total inhibition of agglutination required a much higher concentration of antigen (32 mg/ml). Similar inhibitory reactions at antigen excess were also seen using another latex system (Table I). In this commercially obtained system, the antigen used was derived from N. meningitidis, while the latex particles were sensitized with an IgM monoclonal antibody raised against the organism. Note, however, that the test was not performed as recommended by the manufacturer and also that the
2 0-
1 8
1 6
I tl
1 2
o 0.8
0.6
0.4
0.2
0-
I
I
I
0. 00125
1
I
0.005
latex
0.02
c o n c e n t r a t i o n (%)
Fig. 1. Absorbance of different concentrations of the 0 - 9
antibody-latex conjugate in buffer at 400 nm wavelength.
Slide
-
+
+
4.1-
-tfl-
Tube
ND
ND
ND
ND
ND
~
~
--
-I-
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
-fit-
+1-
+
--
+
+
-I¢-
-if-
~
~
-tit-
-fit-
~
~
4-H-
-It-
4+-
+
+
4-
1.0
0.8 0
o
0.6
a O 0.4
0.2
I
I
I
I
I
I
A
I
I
I
I
I
I
I
32
16
4
1
250
63
16
4
1
250
63
16
4
0
mg
per
ml
JJg per ml Antigen
ng per
ml
concentration
S. typhi LPS in buffer. Results of tube tests done in triplicate based on visual examination (average) or on spectrophotometric measurement (mean _+ SD) are shown. Also displayed are the results of slide agglutination tests performed in parallel. - , no agglutination; 4- to + + + , increasing agglutination; N D = not done. Fig. 2. Agglutination of the 0 - 9 antibody-latex conjugate (0.01%) by a range of concentrations of
272 _+
+
q-t-
I 250
I 63
I 16
4+
-IF
4+
+t-
4+
4+
+
+
3.0
2.6
2.2 O
£3 O
1.8
1.4
1.0 000
I 4
I 1
I 250
I 63
I 16
#g per ml
I 4
I 0
ng per ml Antigen c o n c e n t r a t i o n
Fig. 3. Effect of using increased concentration of the 0-9 antibody-latex conjugate (0.04%) on its agglutination by S. typhi LPS in buffer. Results of visual examination of tube reactions are also shown. Other legends as in Fig. 2.
concentration of antigen used in the kit as a positive control was not within the inhibitory concentrations observed in our test. When the S. typhi titration was run using a four-fold more concentrated suspension of the 0 - 9 latex particles, the U-shaped profile was again seen based on the spectrophotometric readings (Fig. 3), and the point of maximal agglutination _+
+
4+
+1-
was not significantly different. There was a high background of unagglutinated latex in all tubes causing a loss of sensitivity particularly in the visual readings. The possibility of measuring the agglutination reaction in urine instead of buffer was examined. Pooled normal urine samples spiked with different amounts of S. typhi LPS, were incubated with +H-
J-H-
+H-
+H-
4+
4+
4"t-
+
+
I 63
I 16
I 4
1.0
0.8
O
E3 O
0.6
0.4 ==
0.2 I 1000
I 250
I 63
I 16
I 4
I 1
#g per ml
I 250
I 0
ng per ml Antigen
concentration
Fig. 4. Effect of urine on the agglutination of the O-9 antibody conjugate (0.01%) by S. typhi LPS. Conditions of reaction as in Fig. 2, except that undiluted normal human urine spiked with antigen was used instead of buffer. Results of visual examination of tube reactions are also shown.
273 TABLE I D E T E C T I O N OF N. M E N 1 N G I T I D I S A N T I G E N IN B U F F E R BY A G G L U T I N A T I O N W I T H ANTIBODYSENSITIZED LATEX PARTICLES: EFFECT OF ANTIGEN CONCENTRATION Antigen dilution a
Agglutination results b Tube method
Slide method c
1/2
-
-
1/4
+
1/8 1/16 1/32 1/64 1/128 1/256 1/512 1/1024 1/2048 1/4096 1/8192 1/16384 1/32768 No antigen
+ + + + + + + + + + + -
+
+ + + + + + +
+ + + +
++ ++ ++ ++ 4ND ND ND ND ND ND -
a Of stock solution made by dissolving contents in vial (as supplied) with 0.2 ml distilled water (instead of 1.0 ml as recommended by manufacturer). b From visual examination. --, no agglutination; + to + + + , increasing agglutination; N D = not done. c As described in text, using 30 /tl antigen and 10 ~1 latex suspension. (Method recommended by manufacturer uses 40 /~1 antigen and 25/.tl latex.)
0.01% 0-9 latex suspension. The results were similar to those obtained using buffer (Fig. 4).
Discussion We have previously described a tube modification of the latex agglutination method which showed improved sensitivity and readability (Lim and Fok, 1987). Since using this method, agglutination of the latex particles was associated with a decrease in turbidity of the latex suspension, it occurred to us that the results of the test could perhaps be read by transmission spectrophotometry. The present investigation confirmed that this was possible. The reaction time used was relatively short (30 min) in comparison with the photometric methods previously described, though longer than the con-
ventional slide test. Although the reaction mixtures were transferred to cuvettes to be read separately in a spectrophotometer in this study (and this is an inconvenience), it would be possible in future to design suitable reaction tubes for direct use in specially designed spectrophotometers capable of reading small volumes of fluid, say 0.1 ml. The microplate readers currently in use for ELISA systems can be adapted for use with the latex particles. We have experimented with the MR710 (Dynatech, Alexandria, VA) and the Titertek Uniskan (Flow Laboratories, North Ryde, NSW) readers for this purpose, using microplates to contain the test suspensions, and found them to be moderately satisfactory (unpublished observations). However, in our experience, the microplates are not ideal reaction vessels. Suitable conditions for reading the latex agglutination results by spectrophotometry were found. The wavelength chosen was 400 nm, which is the absorption maximum of the 0 - 9 latex conjugate, as well as of the unconjugated latex particles (data not shown). A latex concentration of 0.01% was found to be suitable for this purpose, and is the concentration found to be ideal for the visual tube agglutination test (Lim and Fok, 1987). A surprising finding from the turbidimetric measurements was the U-shaped profile of the reaction (Figs. 2-4). Although inhibition of reaction at antigen excess is seen in many immunoassay systems and commonly referred to as a prozone phenomenon, the effect seen in our case was rather pronounced and spread over many antigen dilutions. The plots obtained were bi-symmetrical, analogous to precipitation reactions where the points of maximal agglutination are equivalence points. The reaction may be considered a precipitation since soluble and not particulate antigen is involved. Inhibition at 'antigen excess' is not a peculiarity of the turbidimetric measurement used since it was also observed in both the tube and slide methods of latex agglutination, though the effects were less pronounced. Nor is it a peculiarity of our latex system, since a modified commercially available system which used another IgM monoclonal antibody to detect N. meningitidis antigen, also showed it (Table I). On the other hand, we have not observed a prozone effect with whole S.
274
typhi bacteria at concentrations as high a s 1012 o r g a n i s m s / m l using the 0 - 9 latex conjugate in a slide test. This is the first time an inhibition of latex agglutination has been documented. The clinical significance of this is unknown, since, particularly for the slide test, high concentrations ( m g / m l ) of antigen are required to inhibit the reaction. However, since latex agglutination is widely used, e.g., for diagnosing meningitis, and since clinical samples are sometimes concentrated before testing, the observation is nevertheless noteworthy. The phenomenon unfortunately invalidates any attempt at using spectrophotometric measurement to quantitate latex agglutination, unless serial dilutions of the test sample are made to locate the 'equivalence point' of the assay.This is a preliminary report confirming the potential use of instrumental reading for latex agglutination. Other antigen detection systems employing both monoclonal and polyclonal reagent antibodies, as well as other types of latex particles, will have to be examined to determine the generality of the application. With this improvisation in our system, the test becomes more sensitive than the visual tube method, and is at least eight-fold more sensitive than the conventional slide test when used under optimal conditions (Fig. 2). This, admittedly, is at the expense of rapidity and simplicity of the conventional slide method. Whether it is also at the expense of specificity as a result of the improved sensitivity, remains a possibility. Nevertheless, unlike the visual methods in which the end-point of the assay is often ill-defined, the modified test gives objective results. Furthermore, it may be possible to introduce automation to latex agglutination assays, which would be particularly useful for screening large numbers of samples at a time. We have shown in this report that urine presents no problems when used as diluent, including the instrumental reading of the results (Fig. 4). This source of material, which is convenient to obtain, has proven to be immensely useful for the diagnosis of various infectious diseases (Coonrod, 1983). However, a range of clinical specimens will have to be used for a proper evaluation of the reliability of our method, since these can behave quite differently from spiked urine samples, especially when these involve the use of stored and pooled urine, as in our case.
Acknowledgements We thank Otis K.H. Ko for skilled technical assistance and Candy Leung for typing the manuscript.
References Blume, P. and Greenberg, L.J. (1975) Application of differential light scattering to the latex agglutination assay for rheumatoid factor. Clin. Chem. 21, 1234. Cambiaso, C.L., Leek, A.E., De Steenwinkel, F., Billen, J. and Masson, P.L. (1977) Particle counting i m m u n o a s s a y (PACIA). I. A general method for the determination of antibodies, antigens, and haptens. J. Immunol. Methods 18, 33. Coonrod, J.D. (1983) Urine as an antigen reservoir for diagnosis of infectious diseases. Am. J. Med. 75 (suppl. 1B), 85. Dezelic, G., Dezelic, N., Muic, N. and Pende, B. (1971) Latex particle agglutination in the immunochemical system hum a n serum albumin-anti-human serum albumin rabbit serum. Eur. J. Biochem. 20, 553. Graham, Jr., L., Meier, F.A., Centor, R.M., Garner, B.K. and Dalton, H.P. (1986) Effect of medium and cultivation conditions on comparisons between latex agglutination and culture detection of group A streptococci. J. Clin. Microbiol. 24, 644. Lim, P.L. (1986) Diagnostic uses of monoclonal antibodies to Salmonella. In: A.J. Macario and E.C. de Macario (Eds.), Monoclonal Antibodies Against Bacteria, Vol. 3. Academic Press, New York, p. 29. Lira, P.L. and Fok, Y.P. (1987) Detection of group D salmonellae in blood culture broth and of soluble antigen by tube agglutination using an 0 - 9 monoclonal antibody latex conjugate. J. Clin. Microbiol. 25, 1165. Marcon, M.J., Hamoudi, A.C. and Cannon, H.J. (1984) Comparative laboratory evaluation of three antigen detection methods for diagnosis of Haemophilus influenzae type b disease. J. Clin. Microbiol. 19, 333. Miotti, P.G., Viscidi, R.P., Eiden, J., Cerny, E. and Yolken, R.H. (1986) Centrifugation-augmented solid-phase immunoassay (CASPIA) for the rapid diagnosis of infectious diseases. J. Infect. Dis. 154, 301. Sambourg, M., Goudeau, A., Courant, C., Pinon, G. and Denis, F. (1985) Direct appraisal of latex agglutination testing, a convenient alternative to enzyme i m m u n o a s s a y for the detection of rotavirus in childhood gastroenteritis, by comparison of two enzyme immunoassays and two latex tests. J. Clin. Microbiol. 21,622. Thomas, M. (1986) Agglutination methods for rapid analysis. Nature 320, 289.