Vol. 10, No.3
JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1979, p. 279-284 0095-1137/79/09-0279/06$02.00/0
Demonstration of Antibodies Against Yersinia enterocolitica Lipopolysaccharide in Human Sera by Enzyme-Linked Immunosorbent Assay MARIANNE GRIPENBERG,* ANTTI NISSINEN,t EINO VAISANEN,t AND EWERT LINDER
Department of Bacteriology and Immunology, University of Helsinki, SF-00290 Helsinki 29, Finland Received for publication 5 June 1979
Antibodies against Yersinia enterocolitica serotype 0:3 lipopolysaccharide sera from patients with Yersinia infection were studied by using an enzyme-linked immunosorbent assay. Of the sera with significant bacterial agglutination titers against Y. enterocolitica type 0:3, 86% contained anti-lipopolysaccharide antibodies of the immunoglobulin G class. With the sera of some patients, we demonstrated increasing anti-lipopolysaccharide antibody levels of immunoglobulin G class in spite of decreasing bacterial agglutination titers. The assay was specific for lipopolysaccharide from Y. enterocolitica type 0:3, and in inhibition experiments lipopolysaccharide could be detected in amounts of :0.5 ,lg/
present in
ml.
Infections caused by Yersinia enterocolitica serotypes 0:3 and 0:9 are associated with a strong antibody response. Isolation of bacteria from feces is not always achieved, and consequently the diagnosis often depends on serology (1), the antibodies being demonstrated by bacterial agglutination (21). The antibody titers usually rise a few days after the onset of clinical symptoms and decline within 2 to 6 months, but sometimes high titers are seen for up to 2 years
serological diagnosis of salmonella infections (7). We have used the ELISA to study the occurrence of antibodies of immunoglobulin G (IgG) class against Y. enterocolitica LPS in the sera of patients with Yersinia infection.
(1).
The symptoms caused by Yersinia infections are mainly abdominal and usually are not serious. Occasionally, however, the infection is associated with extraintestinal symptoms such as "reactive" arthritis, carditis, iritis, and erythema nodosum (2, 3). In such complicated cases of yersiniosis, the bacterial agglutination titers tend to be high (1, 2). A further evaluation of the immune response in Yersinia infection could thus help toward the understanding of the pathogenesis of the complications. The enzyme-linked immunosorbent assay (ELISA) (8, 9) offers a sensitive and specific method for assaying various class-specific antibodies (18). Experimentally induced antibodies against bacterial lipopolysaccharide (LPS) can be detected by ELISA (6), and the method has been used for the demonstration of antibodies against LPS from Brucella abortus and Y. enterocolitica type 0:9 in rabbit antisera (5). ELISA has also found clinical application in the t Present address: Department of General Microbiology, University of Helsinki, SF-00100 Helsinki 10, Finland.
279
MATERIALS AND METHODS Bacterial antigens. Y. enterocolitica serotype 0: 3 and Y. enterocolitica serotype 0:9 were cultivated on blood agar plates for 48 h at 22°C (21). For agglutination tests, bacteria were harvested and washed in saline and used in a suspension giving a Klett turbidity value of 90 at 620 nm (Klett-Summerson photoelectric colorimeter, Klett Mfg. Co., Inc., New York, N.Y.) corresponding to a concentration of approximately 4 x 10'° bacteria per ml. For the isolation of LPS, Y. enterocolitica type 0:3 were harvested in distilled water, washed, and killed by adding Formalin to a final concentration of 1%. LPS was extracted by the hot phenol-water method (20). The purity of the LPS preparation was checked by measuring the ribonucleic acid (RNA) (20) and protein (14) contents. The LPS preparation prepared in this way and used in the subsequent studies contained 15.6% RNA and no demonstrable protein. LPS from Escherichia coli and S. typhi (E. coli 0111:B4 LPS and S. typhi 0901 LPS) were purchased from Difco Laboratories, Detroit, Mich. A suspension
of heat-killed B. abortus was obtained from the Public Health Laboratory, Helsinki, Finland. Bacterial agglutination. Bacterial agglutination tests were performed on standard microtiter trays by using serial twofold dilutions of sera in 25 pd of saline and equal volumes of the bacterial suspension. Agglutination was read after incubation overnight at room temperature, and a titer (defined as the reciprocal of the highest serum dilution showing a definite agglutination) above 160 was regarded as positive.
280
J. CLIN. MICROBIOL.
GRIPENBERG ET AL.
Sera. A total of 94 sera, each showing a bacterial agglutination titer 2160 against Y. enterocolitica type 0:3, were used in this study. They were obtained from 68 patients, including 17 males and 51 females (mean age, 33; range, 3 to 75 years), with clinically suspected yersiniosis. One serum sample was obtained from each of 53 patients, whereas serial serum samples were collected from 15 patients. In addition, three sera with agglutinating antibodies against Y. enterocolitica type 0:9 were studied. A pool of nine sera with a bacterial agglutination titer of 1,280 was prepared and used as a positive reference serum. Sera from healthy blood donors were used as controls, and a negative reference pool was prepared from these sera. A further control group consisted of 22 sera from patients with systemic lupus erythematosus (kindly provided by B. Skrifvars IV, Department of Medicine, University Central Hospital, Helsinki, Finland). Ig fractions of sera. IgG and IgM fractions of five individual patient sera were separated by using diethylaminoethyl-celiulose (Pharmacia Fine Chemicals AB, Uppsala, Sweden) (19). The IgG and IgM concentrations in whole sera and in the concentrated Ig fractions were determined by single radial immunodiffusion on Tri-Partigen immunodiffusion plates (Behringwerke AG, Marburg/Lahn, W. Germany). The sera and the separated fractions were then tested in dilutions contaSing the same Ig concentrations. Reduction of IgM antibodies was performed by mixing 100 pd of 0.2 M 2-mercaptoethanol with 100 y1 of sera diluted 1:100 and incubating the mixture overnight at room temperature. Anti-Y. enterocolitica LPS ELISA. The ELISA was performed essentially as described earlier (5, 10). Disposable 1-ml polystyrene microcuvettes (Finnpipette, Labsystems, Helsinki, Finland) were incubated with 200Ml of LPS solution (0.1 to 10 Mg of LPS/ml of 0.05 M carbonate buffer, pH 9.6) for 3 h at 37°C. After washing with phosphate-buffered saline (pH 7.2), containing 0.05% Tween 20 (PBS-T), the cuvettes were incubated for 5 h at room temperature with 175 p1 of sera diluted in PBS-T. The cuvettes were then washed again, and 150 Ml of a sheep anti-human IgG-alkaline phosphatase conjugate (10) diluted in PBS-T was added to the cuvettes and left at room temperature overnight. The cuvettes were then washed as described above and rinsed again with distilled water before the addition of the substrate solution (1 mg of Sigma 104 phosphatase substrate; Sigma Chemical Co., St. Louis, Mo., per ml of a 1 M diethanolamine buffer, pH 9.7, containing 0.5 mM MgCl2). The enzymatic reaction was allowed to proceed for 60 min at 37°C, after which the reaction was stopped by 250 ,ll of 0.16 M NaOH. The absorbance was measured at a wavelength of 404.7 nm vertically through the cuvettes by using an automated nine-channel photometer (Finnpipette analyzer FP-9, Labsystems, Helsinki, Finland). PBS-T in LPS-coated cuvettes and sera in noncoated cuvettes were run as technical controls. Inhibition studies. Inhibition studies with LPS were performed by mixing equal volumes of sera diluted 1:50 and solutions containing LPS (0.001 to 100 Mlg/ml). After incubation overnight at 4°C, the sera were tested in the ELISA in a final dilution of 1:500.
Absorptions with whole bacteria were performed by using equal volumes of sera diluted 1:500 and dilutions of the bacterial suspensions. After overnight incubation and centrifugation, the supernatants were tested.
RESULTS
Initial studies using the pooled positive anti-Yersinia reference serum. To find the optimal LPS concentration, the pooled positive reference serum was tested in the ELISA after coating the cuvettes with various amounts of LPS. The amount of antibody bound, as reflected by the shape of the titration curve, was dependent on the antigen concentration used (Fig. 1). Based on these experiments, a concentration of 1 jig of LPS per ml was chosen for coating in the subsequent studies. Specificity of the assay. Complete inhibition of antibody binding to Y. enterocolitica type 0:3 LPS-coated cuvettes was seen after absorption of sera with Y. enterocolitica type 0: 3 bacteria, whereas no effect was seen after absorption with Y. enterocolitica type 0:9 or B. abortus (Table 1). Likewise, LPS extracted from Y. enterocolitica type 0:3 inhibited the antibody activity, whereas the addition of up to 0.1 mg of E. coli LPS or S. typhi LPS to the sera did not interfere with the results of the assay (Table 2). 2.0
E
1.0
o.5
-of4
-3
SERUM DILUTION FIG. 1. The pooled positive anti- Y. enterocolitica type 0:3 reference serum titrated in the anti-LPS ELISA by using the following concentrations of antigen in the coating process (micrograms of LPS extracted from Y. enterocolitica type 0:3 per milliliter of carbonate buffer): 5 (E); 1 (E); 0.5 (-); and 0.1
(o).
ANTI-Y. ENTEROCOLITICA LPS ELISA
VOL. 10, 1979
The inhibition experiments were also used to quantitate free LPS. The sensitivity of the assay was found to be greatest when using an initial serum dilution of 1:100. In this way, 0.5 ,ug of Y. enterocolitica type 0:3 LPS per ml still caused a significant decrease of the absorbance in the ELISA (Fig. 2). Testing of individual sera in the anti-LPS ELISA. Representative titration curves of individual sera from patients with Yersinia infection are shown in Fig. 3. The slopes of the individual titration curves vary, and a remarkable prozone effect was seen with many sera. To study this phenomenon, the effect of IgM antibodies on the assay was evaluated. When the isolated IgG fractions of five individual sera were assayed, no prozone was seen, and the slopes of the titration curves ran almost in parallel (Fig. 4). The bacterial agglutination titers of the IgG fractions were 3 to 5 titer steps below the titers of the IgM fractions which corresponded to the titers of the unfractionated sera. A total of 20 sera were treated with 2-mercaptoethanol; this reduction did not affect the results obtained in the ELISA but resulted in a significant decrease in the TABLE 1. Specificity of the assay for antibodies against Y. enterocolitica type 0:3
281
|
E
0.5
0.1
1.0
10
100
IPS ug9 mi
FIG. 2. Quantitation of Y. enterocolitica type 0:3 LPS by absorption of the pooled positive anti-Y. enterocolitica type 0:3 reference serum with Y. enterocolitica serotype 0:3 LPS. Symbols: O, pooled positive reference serum; 0, pooled negative reference serum, final serum dilution 1:1,0X0.
Absorbance with the
following amt of bac-
bacterial agglutination titers (P < 0.01 by the Wilcoxon signed-rank test). 2 x 1010 I x 109 To minimize the day-to-day variation in ab0.360 Y. enterocolitica type 0:3 0.020 sorbance values (10 to 20%), results were exY. enterocolitica type 0:9 1.120 1.130 pressed in relation to the pooled positive referB. abortus 1.060 1.280 ence serum run simultaneously. A dilution of 1: a Equal volumes of the pooled positive anti-Y. en- 5,000 was chosen as standard (Fig. 1 and 3). terocolitica type 0:3 serum (diluted 1:500) and the When the absorbance value of the pooled referbacterial suspensions were mixed and incubated over- ence serum was expressed as 100 ELISA units, night at 4°C. After centrifugation, the supernatants the individual sera investigated gave results of were tested in the ELISA. Absorbance values were relative binding ranging from 3 to 220 U. The 59 obtained at the final serum dilution of 1:1,000. sera from blood donors and the 22 systemic lupus erythematous sera tested gave a mean TABLE 2. Specificity of the assay for antibodies value of 3 ± 1 standard deviation. Values below against LPS extracted from Y. enterocolitica 3 equal the technical controls. The three sera type 0:3 with agglutinating antibodies against Y. enterAbsorbance with the folocolitica type 0:9 gave results comparable to lowing amt of LPS added those for the normal controls. Source of LPS (mg/ml)a: Correlation between bacterial agglutiOrganism
teria added/ml':
0.1
0.01
0.001
0.820 1.680 1.890 Y. enterocolitica type 0:3 E. coli 2.090 1.930 1.950 S. typhi 2.040 1.930 1.980 a Equal volumes of the pooled positive anti-Y. enterocolitica type 0:3 serum (diluted 1:50) and the LPS solutions were mixed, incubated overnight at 4°C, and then tested in the ELISA. Values were obtained at a final serum dilution of 1:500.
nation titers and results obtained in the anti-LPS ELISA. Antibodies of IgG class against LPS extracted from Y. enterocolitica type 0:3 were detected in 86% of the sera from the patients with yersiniosis. Figure 5 shows the anti-LPS ELISA units plotted against the corresponding bacterial agglutination titers. Although considerable divergence of results from the two assays was evident with many individual
282
J. CLIN. MICROBIOL.
GRIPENBERG ET AL
was obtained 2 to 4 weeks after the onset of clinical symptoms. Three characteristic types of results were obtained (Fig. 6). In some cases, the curves for the agglutinating antibodies and the ones for the IgG anti-LPS antibodies run in parallel (Fig. 6a). In some cases, there was an initial increase in the IgG anti-LPS antibodies when the agglutination titer already was decreasing (Fig. 6b and c), and in some cases this increase in anti-LPS antibodies of IgG class was very late (Fig. 6d, e, and f).
E
.-W
DISCUSSION Although antibody affinity affects the results / ,{ / / of ELISA (4), the method can be regarded as quantitative. Being sensitive, specific, and easy to perform, ELISA offers many advantages com_5 lo-5 104 pared with other serological methods. We have 103 lo-2 SIlN IIUTIN developed an ELISA for antibodies against Y. FIG. 3. Representative titration curves obtained in enterocolitica type 0:3 LPS and used it to quanthe anti-LPS IELISA, using cuvettes coated with 1pg titate antibodies of IgG class in patient sera. of Y. enterocc Vitica type 0:3 LPS per ml. Symbols: The assay was specific for Y. enterocolitica 0, pooledposii tive reference serum; U, pooled negative type 0:3 LPS, because the inhibition experi-
... reference seru[m;*, *, 0 three individual sera with' bacterial agglIutination titers against Y. enterocolit- ments performed with Y. enterocolitica type 0: 9 or with LPS from E. coli and S. typhi did not ica type 0:3 o) f0and and 1,280,
fa640,32ti,
entely respectively, Y
affect the results. Since the LPS preparation used for coating the cuvettes had some RNA contamination, we further tested the specificity
2.0 F
100 I 0
1.0
r
9
-a I
.a
I10 5
10 4
SIUM IUTION FIG. 4. Three individual sera
If3
io-2
(LI 0, A) and the corresponding IgG fractions (E 0, A) tested in the ELISA, with the anti-IgG conjugate diluted 1:500. patient sera, overall a statistically significant correlation between the two methods was found to exist (correlation coefficient, 0.5; P < 0.001). With 15 patients, serial serum samples taken at different intervals were studied. The first sample
rr-
160
320
.
a
640 1280 2560 5128 1020 20450
BACTERIAL AGGLUTINATION TITER FIG. 5. Comparison between bacterial agglutination titers and anti-LPS results. The ELISA units are calculated as percentage of the absorbance value obtained with the pooled positive reference serum in the same dilution run on the same occasion.
ANTI-Y. ENTEROCOLITICA LPS ELISA
VOL. 10, 1979
283
C'
b 20480
100
20460
lii
10240 5120 _ a
10240 5120
X
2560 _ , a
b-.
1290
1280
640
640
320 I 1
I
I
I
5
3
320
I I
7
9
..I
a
1
3
TIME (WEEKS)
TIME (WEEKS)
a.
I
.
5
7
aI
I
9
TIME ( WEEKS )
--
r
.
e 20480
l-u
1
0240
10240
lIII
5120
5120
c 2560. _ _
_"
1260
1280
640
640
II*
I
I
I
320 II 1
..
3
5
7
TIME (WEEKS)
9
-
II
I
3
320
5
TIME (WEEKS)
7
9
1
3
FIG. 6. Bacterial agglutination titers (-) and ELISA units (0) in serially studied individual patients. See text.
by assaying a group of sera from patients with systemic lupus erythematosus which are known to contain anti-RNA antibodies (17). These sera gave results similar to those of the normal controls, indicating that the RNA contamination did not affect the results. The titration curves of some sera in this study exhibited a prozone. In an assay such as ELISA, antibodies of a different class present in the same serum sample may compete for the antigenic determinants (11). Such competition, apparently prinicipally by IgM antibodies, appeared to be the probable cause of the prozone effect in this study, because no prozone was seen when isolated IgG fractions of individual sera were studied (Fig. 4). Competition between antibodies of different classes, as well as differences in avidity, could also explain the variations in the shape of the titration curves of whole sera.
5
7
9
TIME ( WEEKS ) serum
samples from
The serum dilution of 1:5,000 chosen for expression of the results as units was in each case well beyond a possible prozone and should not cause any false-negative results. The development of an assay specific for antibodies of IgG class against Y. enterocolitica LPS was motivated by our interest in the extraintestinal "reactive" complications occasionally seen in patients with yersiniosis (2, 3). The pathogenesis of these conditions has not been clarified, nor is it known whether the prolonged presence of bacteria or bacterial products is a prerequisite for the development of these complications. It is generally assumed that immunological mechanisms in combination with host factors are of importance (12, 15). Bacterial LPS is a potent mediator of diverse biological effects (16), but its possible role, if any, in the pathogenesis of the complications in yersiniosis is not
284
GRIPENBERG ET ALC J. CLIN. MICROBIOL.
known. In this study, some patients exhibited elevated IgG anti-LPS values for long periods of time. The rise in anti-Y. enterocolitica LPS antibodies of IgG class despite a decline of bacterial agglutination titers observed in some cases with joint symptoms was interesting and encourages further studies in patients with reactive conditions. Our ELISA was primarily designed to detect antibodies against LPS, but the inhibition studies indicate that the assay may be modified for the quantitation of LPS. In terms of sensitivity, such an inhibition assay cannot compete with, for instance, the Limulus amoebocyte lysate test (13) which can detect picogram quantities of LPS (22). A modification of the anti-LPS ELISA offers, however, a method for detecting specific LPS in concentrations down to 0.5 ,ug/ml and may be useful for some purposes. ACKNOWLEDGMENTS This work was supported by grants from the Sigrid Juselius Foundation and Finska Lakaresallskapet. The technical assistance of Pipsa Kaipainen is gratefully acknowledged. We thank Pirjo H. Makela for reviewing the manuscript. 1. 2. 3. 4.
5.
6.
7.
LITERATURE CITED Ahvonen, P. 1972. Human yersiniosis in Finland. I. Bacteriology and serology Ann. Clin. Res. 4:30-38. Ahvonen, P. 1972. Human yersiniosis in Finland. II. Clinical features. Ann. Clin. Res. 4:39-48. Arvastson, B., K. Damgaard, and S. Winblad. 1971. Clinical symptoms of infection with Yersinia enterocolitica. Scand. J. Infect. Dis. 3:37-40. Butler, J. E., T. L. Feldbush, P. L. MeGivern, and N. Stewart. 1978. The enzyme-linked immunosorbent assay (ELISA): a measure of antibody concentration or affinity? Immunochemistry 15:131-136. Carlsson, H. E., B. Hurvell, and A. A. Lindberg. 1976. Enzyme-linked immunosorbent assay (ELISA) for titration of antibodies against Brucella abortus and Yersinia enterocolitica. Acta Pathol. Microbiol. Scand. Sect. C 84:168-176. Carlsson, H. E., A. A. Lindberg, and S. Hammarstrom. 1972. Titration of antibodies to Salmonella 0 antigens by enzyme-linked immunosorbent assay. Infect. Immun. 6:703-708. Carlsson, H. E., A. A. Lindberg, S. Hammarstrom, and A. Ljunggren. 1975. Quantitation of Salmonella
8.
9.
10.
11.
12. 13.
14.
15. 16. 17. 18.
19.
20. 21. 22.
0-antibodies in human sera by enzyme-linked immunosorbent assay (ELISA). Int. Arch. Allergy Appl. Immunol. 48:485-494. Engvall, E., K. Joneson, and P. Perlmann. 1971. Enzyme-linked immunosorbent assay. II. Quantitative assay of protein antigen, immunoglobulin G, by means of enzyme-labelled antigen and antibody coated tubes (BBA 35998). Biochim. Biophys. Acta 251:427-434. Engvall, E., and P. Perlmann. 1972. Enzyme-linked immunosorbent assay, ELISA. III. Quantitation of specific antibodies by enzyme-labeled anti-immunoglobulin in antigen-coated tubes. J. Immunol. 109:129-135. Gripenberg, M., E. LUnder, P. Kurki, and E. Engvall. 1978. A solid phase enzyme-linked immunosorbent assay (ELISA) for the demonstration of antibodies against denatured, single-stranded DNA in patient sera. Scand. J. Immunol. 7:151-157. Holmgren, J., and A.-M. Svennerholm. 1973. Enzymelinked immunosorbent assays for cholera serology. Infect. Immun. 7:759-763. Laitinen, O., M. Leirisalo, and G. Skylv. 1977. Relation between HLA-B27 and clinical features in patients with yersinia arthritis. Arthritis Rheum. 20:1121-1124. Levin, J., T. E. Poore, N. P. Zauber, and R. S. Oser. 1970. Detection of endotoxin in the blood of patients with sepsis due to gram-negative bacteria. N. Engl. J. Med. 283:1313-1316. Lowry, 0. H., N. J. Rosebrough, A. L Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. McDevitt, H. O., and W. F. Bodmer. 1974. HL-A, immune response genes, and disease. Lancet i:1269-1275. Neter, E. 1969. Endotoxins and the immune response. Curr. Top. Microbiol. Immunol. 47:82-124. Schur, P. H., and M. Monroe. 1969. Antibodies to ribonucleic acid in systemic lupus erythematosus. Proc. Natl. Acad. Sci. U.S.A. 63:1108-1112. Schuurs, A. H. W. M., and B. K. van Weemen. 1977. Enzyme-immunoassay. Clin. Chim. Acta 81:1-40. Sober, H. A., F. J. Gutter, M. M. Wyckoff, and F. A. Peterson. 1956. Chromatography of proteins. II. Fractionation of serum proteins on anion exchange cellulose. J. Am. Chem. Soc. 78:756-763. Westphal, O., 0. Luderitz, and F. Bister. 1952. Uber die Extraction von Bakterien mit Phenol/Wasser. Z. Naturforsch. 7b:148-155. Winblad, S., B. Nilehn, and N. H. Sternby. 1966. Yersinia enterocolitica (Pasteurella X) in human enteric infections. Br. Med. J. 2:1363-1366. Yin, E. T., C. Galanos, S. Kinsky, R. A. Bradshaw, S. Wessler, 0. Luderitz, and M. E. Sarmiento. 1972. Picogram-sensitive assay for endotoxin: gelation of Limulus polyphemus blood cell lysate induced by purified lipopolysaccharides and lipid A from gram-negative bacteria. Biochim. Biophys. Acta 261:284-289.
JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1979, p. 279-284 0095-1137/79/09-0279/06$02.00/0
Demonstration of Antibodies Against Yersinia enterocolitica Lipopolysaccharide in Human Sera by Enzyme-Linked Immunosorbent Assay MARIANNE GRIPENBERG,* ANTTI NISSINEN,t EINO VAISANEN,t AND EWERT LINDER
Department of Bacteriology and Immunology, University of Helsinki, SF-00290 Helsinki 29, Finland Received for publication 5 June 1979
Antibodies against Yersinia enterocolitica serotype 0:3 lipopolysaccharide sera from patients with Yersinia infection were studied by using an enzyme-linked immunosorbent assay. Of the sera with significant bacterial agglutination titers against Y. enterocolitica type 0:3, 86% contained anti-lipopolysaccharide antibodies of the immunoglobulin G class. With the sera of some patients, we demonstrated increasing anti-lipopolysaccharide antibody levels of immunoglobulin G class in spite of decreasing bacterial agglutination titers. The assay was specific for lipopolysaccharide from Y. enterocolitica type 0:3, and in inhibition experiments lipopolysaccharide could be detected in amounts of :0.5 ,lg/
present in
ml.
Infections caused by Yersinia enterocolitica serotypes 0:3 and 0:9 are associated with a strong antibody response. Isolation of bacteria from feces is not always achieved, and consequently the diagnosis often depends on serology (1), the antibodies being demonstrated by bacterial agglutination (21). The antibody titers usually rise a few days after the onset of clinical symptoms and decline within 2 to 6 months, but sometimes high titers are seen for up to 2 years
serological diagnosis of salmonella infections (7). We have used the ELISA to study the occurrence of antibodies of immunoglobulin G (IgG) class against Y. enterocolitica LPS in the sera of patients with Yersinia infection.
(1).
The symptoms caused by Yersinia infections are mainly abdominal and usually are not serious. Occasionally, however, the infection is associated with extraintestinal symptoms such as "reactive" arthritis, carditis, iritis, and erythema nodosum (2, 3). In such complicated cases of yersiniosis, the bacterial agglutination titers tend to be high (1, 2). A further evaluation of the immune response in Yersinia infection could thus help toward the understanding of the pathogenesis of the complications. The enzyme-linked immunosorbent assay (ELISA) (8, 9) offers a sensitive and specific method for assaying various class-specific antibodies (18). Experimentally induced antibodies against bacterial lipopolysaccharide (LPS) can be detected by ELISA (6), and the method has been used for the demonstration of antibodies against LPS from Brucella abortus and Y. enterocolitica type 0:9 in rabbit antisera (5). ELISA has also found clinical application in the t Present address: Department of General Microbiology, University of Helsinki, SF-00100 Helsinki 10, Finland.
279
MATERIALS AND METHODS Bacterial antigens. Y. enterocolitica serotype 0: 3 and Y. enterocolitica serotype 0:9 were cultivated on blood agar plates for 48 h at 22°C (21). For agglutination tests, bacteria were harvested and washed in saline and used in a suspension giving a Klett turbidity value of 90 at 620 nm (Klett-Summerson photoelectric colorimeter, Klett Mfg. Co., Inc., New York, N.Y.) corresponding to a concentration of approximately 4 x 10'° bacteria per ml. For the isolation of LPS, Y. enterocolitica type 0:3 were harvested in distilled water, washed, and killed by adding Formalin to a final concentration of 1%. LPS was extracted by the hot phenol-water method (20). The purity of the LPS preparation was checked by measuring the ribonucleic acid (RNA) (20) and protein (14) contents. The LPS preparation prepared in this way and used in the subsequent studies contained 15.6% RNA and no demonstrable protein. LPS from Escherichia coli and S. typhi (E. coli 0111:B4 LPS and S. typhi 0901 LPS) were purchased from Difco Laboratories, Detroit, Mich. A suspension
of heat-killed B. abortus was obtained from the Public Health Laboratory, Helsinki, Finland. Bacterial agglutination. Bacterial agglutination tests were performed on standard microtiter trays by using serial twofold dilutions of sera in 25 pd of saline and equal volumes of the bacterial suspension. Agglutination was read after incubation overnight at room temperature, and a titer (defined as the reciprocal of the highest serum dilution showing a definite agglutination) above 160 was regarded as positive.
280
J. CLIN. MICROBIOL.
GRIPENBERG ET AL.
Sera. A total of 94 sera, each showing a bacterial agglutination titer 2160 against Y. enterocolitica type 0:3, were used in this study. They were obtained from 68 patients, including 17 males and 51 females (mean age, 33; range, 3 to 75 years), with clinically suspected yersiniosis. One serum sample was obtained from each of 53 patients, whereas serial serum samples were collected from 15 patients. In addition, three sera with agglutinating antibodies against Y. enterocolitica type 0:9 were studied. A pool of nine sera with a bacterial agglutination titer of 1,280 was prepared and used as a positive reference serum. Sera from healthy blood donors were used as controls, and a negative reference pool was prepared from these sera. A further control group consisted of 22 sera from patients with systemic lupus erythematosus (kindly provided by B. Skrifvars IV, Department of Medicine, University Central Hospital, Helsinki, Finland). Ig fractions of sera. IgG and IgM fractions of five individual patient sera were separated by using diethylaminoethyl-celiulose (Pharmacia Fine Chemicals AB, Uppsala, Sweden) (19). The IgG and IgM concentrations in whole sera and in the concentrated Ig fractions were determined by single radial immunodiffusion on Tri-Partigen immunodiffusion plates (Behringwerke AG, Marburg/Lahn, W. Germany). The sera and the separated fractions were then tested in dilutions contaSing the same Ig concentrations. Reduction of IgM antibodies was performed by mixing 100 pd of 0.2 M 2-mercaptoethanol with 100 y1 of sera diluted 1:100 and incubating the mixture overnight at room temperature. Anti-Y. enterocolitica LPS ELISA. The ELISA was performed essentially as described earlier (5, 10). Disposable 1-ml polystyrene microcuvettes (Finnpipette, Labsystems, Helsinki, Finland) were incubated with 200Ml of LPS solution (0.1 to 10 Mg of LPS/ml of 0.05 M carbonate buffer, pH 9.6) for 3 h at 37°C. After washing with phosphate-buffered saline (pH 7.2), containing 0.05% Tween 20 (PBS-T), the cuvettes were incubated for 5 h at room temperature with 175 p1 of sera diluted in PBS-T. The cuvettes were then washed again, and 150 Ml of a sheep anti-human IgG-alkaline phosphatase conjugate (10) diluted in PBS-T was added to the cuvettes and left at room temperature overnight. The cuvettes were then washed as described above and rinsed again with distilled water before the addition of the substrate solution (1 mg of Sigma 104 phosphatase substrate; Sigma Chemical Co., St. Louis, Mo., per ml of a 1 M diethanolamine buffer, pH 9.7, containing 0.5 mM MgCl2). The enzymatic reaction was allowed to proceed for 60 min at 37°C, after which the reaction was stopped by 250 ,ll of 0.16 M NaOH. The absorbance was measured at a wavelength of 404.7 nm vertically through the cuvettes by using an automated nine-channel photometer (Finnpipette analyzer FP-9, Labsystems, Helsinki, Finland). PBS-T in LPS-coated cuvettes and sera in noncoated cuvettes were run as technical controls. Inhibition studies. Inhibition studies with LPS were performed by mixing equal volumes of sera diluted 1:50 and solutions containing LPS (0.001 to 100 Mlg/ml). After incubation overnight at 4°C, the sera were tested in the ELISA in a final dilution of 1:500.
Absorptions with whole bacteria were performed by using equal volumes of sera diluted 1:500 and dilutions of the bacterial suspensions. After overnight incubation and centrifugation, the supernatants were tested.
RESULTS
Initial studies using the pooled positive anti-Yersinia reference serum. To find the optimal LPS concentration, the pooled positive reference serum was tested in the ELISA after coating the cuvettes with various amounts of LPS. The amount of antibody bound, as reflected by the shape of the titration curve, was dependent on the antigen concentration used (Fig. 1). Based on these experiments, a concentration of 1 jig of LPS per ml was chosen for coating in the subsequent studies. Specificity of the assay. Complete inhibition of antibody binding to Y. enterocolitica type 0:3 LPS-coated cuvettes was seen after absorption of sera with Y. enterocolitica type 0: 3 bacteria, whereas no effect was seen after absorption with Y. enterocolitica type 0:9 or B. abortus (Table 1). Likewise, LPS extracted from Y. enterocolitica type 0:3 inhibited the antibody activity, whereas the addition of up to 0.1 mg of E. coli LPS or S. typhi LPS to the sera did not interfere with the results of the assay (Table 2). 2.0
E
1.0
o.5
-of4
-3
SERUM DILUTION FIG. 1. The pooled positive anti- Y. enterocolitica type 0:3 reference serum titrated in the anti-LPS ELISA by using the following concentrations of antigen in the coating process (micrograms of LPS extracted from Y. enterocolitica type 0:3 per milliliter of carbonate buffer): 5 (E); 1 (E); 0.5 (-); and 0.1
(o).
ANTI-Y. ENTEROCOLITICA LPS ELISA
VOL. 10, 1979
The inhibition experiments were also used to quantitate free LPS. The sensitivity of the assay was found to be greatest when using an initial serum dilution of 1:100. In this way, 0.5 ,ug of Y. enterocolitica type 0:3 LPS per ml still caused a significant decrease of the absorbance in the ELISA (Fig. 2). Testing of individual sera in the anti-LPS ELISA. Representative titration curves of individual sera from patients with Yersinia infection are shown in Fig. 3. The slopes of the individual titration curves vary, and a remarkable prozone effect was seen with many sera. To study this phenomenon, the effect of IgM antibodies on the assay was evaluated. When the isolated IgG fractions of five individual sera were assayed, no prozone was seen, and the slopes of the titration curves ran almost in parallel (Fig. 4). The bacterial agglutination titers of the IgG fractions were 3 to 5 titer steps below the titers of the IgM fractions which corresponded to the titers of the unfractionated sera. A total of 20 sera were treated with 2-mercaptoethanol; this reduction did not affect the results obtained in the ELISA but resulted in a significant decrease in the TABLE 1. Specificity of the assay for antibodies against Y. enterocolitica type 0:3
281
|
E
0.5
0.1
1.0
10
100
IPS ug9 mi
FIG. 2. Quantitation of Y. enterocolitica type 0:3 LPS by absorption of the pooled positive anti-Y. enterocolitica type 0:3 reference serum with Y. enterocolitica serotype 0:3 LPS. Symbols: O, pooled positive reference serum; 0, pooled negative reference serum, final serum dilution 1:1,0X0.
Absorbance with the
following amt of bac-
bacterial agglutination titers (P < 0.01 by the Wilcoxon signed-rank test). 2 x 1010 I x 109 To minimize the day-to-day variation in ab0.360 Y. enterocolitica type 0:3 0.020 sorbance values (10 to 20%), results were exY. enterocolitica type 0:9 1.120 1.130 pressed in relation to the pooled positive referB. abortus 1.060 1.280 ence serum run simultaneously. A dilution of 1: a Equal volumes of the pooled positive anti-Y. en- 5,000 was chosen as standard (Fig. 1 and 3). terocolitica type 0:3 serum (diluted 1:500) and the When the absorbance value of the pooled referbacterial suspensions were mixed and incubated over- ence serum was expressed as 100 ELISA units, night at 4°C. After centrifugation, the supernatants the individual sera investigated gave results of were tested in the ELISA. Absorbance values were relative binding ranging from 3 to 220 U. The 59 obtained at the final serum dilution of 1:1,000. sera from blood donors and the 22 systemic lupus erythematous sera tested gave a mean TABLE 2. Specificity of the assay for antibodies value of 3 ± 1 standard deviation. Values below against LPS extracted from Y. enterocolitica 3 equal the technical controls. The three sera type 0:3 with agglutinating antibodies against Y. enterAbsorbance with the folocolitica type 0:9 gave results comparable to lowing amt of LPS added those for the normal controls. Source of LPS (mg/ml)a: Correlation between bacterial agglutiOrganism
teria added/ml':
0.1
0.01
0.001
0.820 1.680 1.890 Y. enterocolitica type 0:3 E. coli 2.090 1.930 1.950 S. typhi 2.040 1.930 1.980 a Equal volumes of the pooled positive anti-Y. enterocolitica type 0:3 serum (diluted 1:50) and the LPS solutions were mixed, incubated overnight at 4°C, and then tested in the ELISA. Values were obtained at a final serum dilution of 1:500.
nation titers and results obtained in the anti-LPS ELISA. Antibodies of IgG class against LPS extracted from Y. enterocolitica type 0:3 were detected in 86% of the sera from the patients with yersiniosis. Figure 5 shows the anti-LPS ELISA units plotted against the corresponding bacterial agglutination titers. Although considerable divergence of results from the two assays was evident with many individual
282
J. CLIN. MICROBIOL.
GRIPENBERG ET AL
was obtained 2 to 4 weeks after the onset of clinical symptoms. Three characteristic types of results were obtained (Fig. 6). In some cases, the curves for the agglutinating antibodies and the ones for the IgG anti-LPS antibodies run in parallel (Fig. 6a). In some cases, there was an initial increase in the IgG anti-LPS antibodies when the agglutination titer already was decreasing (Fig. 6b and c), and in some cases this increase in anti-LPS antibodies of IgG class was very late (Fig. 6d, e, and f).
E
.-W
DISCUSSION Although antibody affinity affects the results / ,{ / / of ELISA (4), the method can be regarded as quantitative. Being sensitive, specific, and easy to perform, ELISA offers many advantages com_5 lo-5 104 pared with other serological methods. We have 103 lo-2 SIlN IIUTIN developed an ELISA for antibodies against Y. FIG. 3. Representative titration curves obtained in enterocolitica type 0:3 LPS and used it to quanthe anti-LPS IELISA, using cuvettes coated with 1pg titate antibodies of IgG class in patient sera. of Y. enterocc Vitica type 0:3 LPS per ml. Symbols: The assay was specific for Y. enterocolitica 0, pooledposii tive reference serum; U, pooled negative type 0:3 LPS, because the inhibition experi-
... reference seru[m;*, *, 0 three individual sera with' bacterial agglIutination titers against Y. enterocolit- ments performed with Y. enterocolitica type 0: 9 or with LPS from E. coli and S. typhi did not ica type 0:3 o) f0and and 1,280,
fa640,32ti,
entely respectively, Y
affect the results. Since the LPS preparation used for coating the cuvettes had some RNA contamination, we further tested the specificity
2.0 F
100 I 0
1.0
r
9
-a I
.a
I10 5
10 4
SIUM IUTION FIG. 4. Three individual sera
If3
io-2
(LI 0, A) and the corresponding IgG fractions (E 0, A) tested in the ELISA, with the anti-IgG conjugate diluted 1:500. patient sera, overall a statistically significant correlation between the two methods was found to exist (correlation coefficient, 0.5; P < 0.001). With 15 patients, serial serum samples taken at different intervals were studied. The first sample
rr-
160
320
.
a
640 1280 2560 5128 1020 20450
BACTERIAL AGGLUTINATION TITER FIG. 5. Comparison between bacterial agglutination titers and anti-LPS results. The ELISA units are calculated as percentage of the absorbance value obtained with the pooled positive reference serum in the same dilution run on the same occasion.
ANTI-Y. ENTEROCOLITICA LPS ELISA
VOL. 10, 1979
283
C'
b 20480
100
20460
lii
10240 5120 _ a
10240 5120
X
2560 _ , a
b-.
1290
1280
640
640
320 I 1
I
I
I
5
3
320
I I
7
9
..I
a
1
3
TIME (WEEKS)
TIME (WEEKS)
a.
I
.
5
7
aI
I
9
TIME ( WEEKS )
--
r
.
e 20480
l-u
1
0240
10240
lIII
5120
5120
c 2560. _ _
_"
1260
1280
640
640
II*
I
I
I
320 II 1
..
3
5
7
TIME (WEEKS)
9
-
II
I
3
320
5
TIME (WEEKS)
7
9
1
3
FIG. 6. Bacterial agglutination titers (-) and ELISA units (0) in serially studied individual patients. See text.
by assaying a group of sera from patients with systemic lupus erythematosus which are known to contain anti-RNA antibodies (17). These sera gave results similar to those of the normal controls, indicating that the RNA contamination did not affect the results. The titration curves of some sera in this study exhibited a prozone. In an assay such as ELISA, antibodies of a different class present in the same serum sample may compete for the antigenic determinants (11). Such competition, apparently prinicipally by IgM antibodies, appeared to be the probable cause of the prozone effect in this study, because no prozone was seen when isolated IgG fractions of individual sera were studied (Fig. 4). Competition between antibodies of different classes, as well as differences in avidity, could also explain the variations in the shape of the titration curves of whole sera.
5
7
9
TIME ( WEEKS ) serum
samples from
The serum dilution of 1:5,000 chosen for expression of the results as units was in each case well beyond a possible prozone and should not cause any false-negative results. The development of an assay specific for antibodies of IgG class against Y. enterocolitica LPS was motivated by our interest in the extraintestinal "reactive" complications occasionally seen in patients with yersiniosis (2, 3). The pathogenesis of these conditions has not been clarified, nor is it known whether the prolonged presence of bacteria or bacterial products is a prerequisite for the development of these complications. It is generally assumed that immunological mechanisms in combination with host factors are of importance (12, 15). Bacterial LPS is a potent mediator of diverse biological effects (16), but its possible role, if any, in the pathogenesis of the complications in yersiniosis is not
284
GRIPENBERG ET ALC J. CLIN. MICROBIOL.
known. In this study, some patients exhibited elevated IgG anti-LPS values for long periods of time. The rise in anti-Y. enterocolitica LPS antibodies of IgG class despite a decline of bacterial agglutination titers observed in some cases with joint symptoms was interesting and encourages further studies in patients with reactive conditions. Our ELISA was primarily designed to detect antibodies against LPS, but the inhibition studies indicate that the assay may be modified for the quantitation of LPS. In terms of sensitivity, such an inhibition assay cannot compete with, for instance, the Limulus amoebocyte lysate test (13) which can detect picogram quantities of LPS (22). A modification of the anti-LPS ELISA offers, however, a method for detecting specific LPS in concentrations down to 0.5 ,ug/ml and may be useful for some purposes. ACKNOWLEDGMENTS This work was supported by grants from the Sigrid Juselius Foundation and Finska Lakaresallskapet. The technical assistance of Pipsa Kaipainen is gratefully acknowledged. We thank Pirjo H. Makela for reviewing the manuscript. 1. 2. 3. 4.
5.
6.
7.
LITERATURE CITED Ahvonen, P. 1972. Human yersiniosis in Finland. I. Bacteriology and serology Ann. Clin. Res. 4:30-38. Ahvonen, P. 1972. Human yersiniosis in Finland. II. Clinical features. Ann. Clin. Res. 4:39-48. Arvastson, B., K. Damgaard, and S. Winblad. 1971. Clinical symptoms of infection with Yersinia enterocolitica. Scand. J. Infect. Dis. 3:37-40. Butler, J. E., T. L. Feldbush, P. L. MeGivern, and N. Stewart. 1978. The enzyme-linked immunosorbent assay (ELISA): a measure of antibody concentration or affinity? Immunochemistry 15:131-136. Carlsson, H. E., B. Hurvell, and A. A. Lindberg. 1976. Enzyme-linked immunosorbent assay (ELISA) for titration of antibodies against Brucella abortus and Yersinia enterocolitica. Acta Pathol. Microbiol. Scand. Sect. C 84:168-176. Carlsson, H. E., A. A. Lindberg, and S. Hammarstrom. 1972. Titration of antibodies to Salmonella 0 antigens by enzyme-linked immunosorbent assay. Infect. Immun. 6:703-708. Carlsson, H. E., A. A. Lindberg, S. Hammarstrom, and A. Ljunggren. 1975. Quantitation of Salmonella
8.
9.
10.
11.
12. 13.
14.
15. 16. 17. 18.
19.
20. 21. 22.
0-antibodies in human sera by enzyme-linked immunosorbent assay (ELISA). Int. Arch. Allergy Appl. Immunol. 48:485-494. Engvall, E., K. Joneson, and P. Perlmann. 1971. Enzyme-linked immunosorbent assay. II. Quantitative assay of protein antigen, immunoglobulin G, by means of enzyme-labelled antigen and antibody coated tubes (BBA 35998). Biochim. Biophys. Acta 251:427-434. Engvall, E., and P. Perlmann. 1972. Enzyme-linked immunosorbent assay, ELISA. III. Quantitation of specific antibodies by enzyme-labeled anti-immunoglobulin in antigen-coated tubes. J. Immunol. 109:129-135. Gripenberg, M., E. LUnder, P. Kurki, and E. Engvall. 1978. A solid phase enzyme-linked immunosorbent assay (ELISA) for the demonstration of antibodies against denatured, single-stranded DNA in patient sera. Scand. J. Immunol. 7:151-157. Holmgren, J., and A.-M. Svennerholm. 1973. Enzymelinked immunosorbent assays for cholera serology. Infect. Immun. 7:759-763. Laitinen, O., M. Leirisalo, and G. Skylv. 1977. Relation between HLA-B27 and clinical features in patients with yersinia arthritis. Arthritis Rheum. 20:1121-1124. Levin, J., T. E. Poore, N. P. Zauber, and R. S. Oser. 1970. Detection of endotoxin in the blood of patients with sepsis due to gram-negative bacteria. N. Engl. J. Med. 283:1313-1316. Lowry, 0. H., N. J. Rosebrough, A. L Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. McDevitt, H. O., and W. F. Bodmer. 1974. HL-A, immune response genes, and disease. Lancet i:1269-1275. Neter, E. 1969. Endotoxins and the immune response. Curr. Top. Microbiol. Immunol. 47:82-124. Schur, P. H., and M. Monroe. 1969. Antibodies to ribonucleic acid in systemic lupus erythematosus. Proc. Natl. Acad. Sci. U.S.A. 63:1108-1112. Schuurs, A. H. W. M., and B. K. van Weemen. 1977. Enzyme-immunoassay. Clin. Chim. Acta 81:1-40. Sober, H. A., F. J. Gutter, M. M. Wyckoff, and F. A. Peterson. 1956. Chromatography of proteins. II. Fractionation of serum proteins on anion exchange cellulose. J. Am. Chem. Soc. 78:756-763. Westphal, O., 0. Luderitz, and F. Bister. 1952. Uber die Extraction von Bakterien mit Phenol/Wasser. Z. Naturforsch. 7b:148-155. Winblad, S., B. Nilehn, and N. H. Sternby. 1966. Yersinia enterocolitica (Pasteurella X) in human enteric infections. Br. Med. J. 2:1363-1366. Yin, E. T., C. Galanos, S. Kinsky, R. A. Bradshaw, S. Wessler, 0. Luderitz, and M. E. Sarmiento. 1972. Picogram-sensitive assay for endotoxin: gelation of Limulus polyphemus blood cell lysate induced by purified lipopolysaccharides and lipid A from gram-negative bacteria. Biochim. Biophys. Acta 261:284-289.
