Serodiagnosis of HeZicobacter p y Zori-Associated Gastritis with a Monoclonal Antibody Competitive Enzyme-Linked Immunosorbent Assay R. NEGRINI, I. ZANELLA, A. SAVIO. C. POIESI, R. VERARDI, S. GHIELMI, A . ALBERTINI, 0. SANGALETTI, M. LAZZARONI & G. BIANCHI PORRO Institute of Chemistry, School of Medicine, University of Brescia, and 'S. Orsola' Fatebenefratelli Hospital. Brescia, and Dept. of Gastroenterology, L. Sacco Hospital, Milan, Italy
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
Negrini R , Zanella I, Savio A, Poiesi C, Verardi R, Ghielmi S, Albertini A , Sangaletti 0 , Lazzaroni M, Bianchi Porro G. Serodiagnosis of Helicobacterpylori-associated gastritis with a monoclonal antibody competitive enzyme-linked immunosorbent assay. Scand J Gastroenterol 1992;27:599-605, Forty-nine monoclonal antibodies against Helicobacier pylori were screened to investigate their capacity to be used in enzyme-linked immunosorbent assay (ELISA) competitive systems for the serodiagnosis of Helicobacterpylori infection. On the basis of the inhibition pattern showed by the sera of five infected patients. the antibodies were subdivided into five groups. The immunoblotting analysis showed that thc antibodies recognized a total of nine different antigenic determinants. In a study of the reaction of the antibodies with 12 isolates of H . pylori a total of 9 antigenic profiles were identified. Two monoclonal antibodies, HpN44 and HpN45, which recognized a 64-kD protein, were inhibited by all 5 positive sera. Antibody HpN45 was labeled with horseradish peroxidase, and the competitive ELISA was compared with an ordinary indirect ELISA in a study of 102 patients undergoing gastroscopy. Seventy-three patients proved to he infected by H . pylori according to urease or histologic tests. The sensitivity and specificity were 90.4% and 89.6%, respectively, for the indirect ELISA and 100% and 89.6% for the HpN45 competitive assay. The three patients who were 'false seropositive' with both serologic tests had atrophic gastritis. The high diagnostic performance and simplicity of the HpN45 monoclonal competitive ELISA make it suitable for routine serodiagnosis of H . pylori infection.
Key words: Enzyme-linked immunosorbent assay; Helicobacier pylori; monoclonal antibodies: serology Riccardo Negrini, M . D..Lahoratorio di Biotecnologie, Spedali Civili di Bre.wia. P.le Spedali Civili, 1 . 25123 Bre.wia, Ilaly
Using indirect enzyme-linked immunosorbent assay (ELISA) tests, many authors have demonstrated a high degree of correlation between the presence of serum antibodies against Helicobucter pylori and gastric infection by this bacterium. However, these tests showed a different grade of sensitivity and specificity (1-9), due to a large extent to variations in a) the antigenic preparation used (4, 9), b) the different dilutions of serum, c) the different values of the set cut-off point, and d) the method used as a 'gold standard' to decide whether a patient has H. pylori in the stomach. The major problem with regard to specificity seems to be cross-reactions due to the presence of H. pylori antigens shared with other bacteria, such as Cumpylobucter jejuni (2,lO-12). The use of monoclonal antibodies for the detection of specific antibodies against H. pylori by indirect ELISA has already been reported (13). In a previous paper we described the use of monoclonal antibodies for the specific immunohistochemical detection of H . pylori (14). The present study was undertaken to evaluate the possible capacity of monoclonal antibodies to be used in competitive ELISA systems for the detection of human antibodies against single H. pylori antigens. One of
these tests was selected and compared with the conventional indirect approach.
PATIENTS A N D METHODS Putients
One-hundred and two patients consecutively undergoing gastroscopy were examined. In each patient biopsy specimens were obtained from the antral mucosa for the histologic and microbiologic examination. Blood samples were collected for the detection of antibodies against H . pylori. Histology
The biopsy specimens were immediately formalin-fixed and paraffin-embedded; the diagnosis of gastritis was made by examining sections stained with haematoxylin and eosin. The histologic changes of the antrum (chronic inflammation, atrophy. activity, intestinal metaplasia. and density of H . pylori) were divided into four grades, in accordance with the Sidney system for the classification of gastritis. The presence of H . pylori was detected by the examination of sections stained with Giemsa.
600
R . Negrini et d.
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
Urease test The presence of bacterial urease in the biopsy specimen was detected by means of a commercial rapid test (Quick Campy1 Test-Brocades, Milan, Italy). Bacteria The 12 isolates of H . pylori used in this study were kindly provided by Dr. F. Carozzi (Biolife-Milano, Italy) and were derived from stored bacteria obtained from human gastric biopsy specimens. The bacteria were identified on the basis of the morphologic features of the colonies and from Gram stain and positive reaction to urease, catalase, and oxidase. The isolates were subcultured on 6% horse blood agar at 37°C in a microaerobic atmosphere. After 5-8 days the bacteria were harvested and washed in phosphate-buffered saline (PBS). The resulting pellet was suspended in PBS at a conccntration of IO’ceIIs/mI and frozen at -20°C. A pool of all the 12 isolates of H. pylori, at a final concentration of 10’ cells/ml in PBS-0.5% Nonidet P40, was ultrasonicated and used as antigen for serologic and iminunoblotting studies. Isolates of C. jejuni, Escherickia coli, Vihrio parahaernoliticus, Klebsiella pneumoniae, Proteus rettgeri, and Pseudoinonas aeruginosa were kindly provided by Dr. P. Maini (USL.31, Ferrara).
Indirect ELlSA The polyclonal serological IgG activity against H . pylori was measured as follows: 100 PI of serum diluted 1 :200 in PBS-l% (BSA)-0.05% Tween 20 were added in duplicate to antigen-coated microwells. Serial dilutions (from 1 : 200 to 1:12,800) of a pool of the sera from 73 infected patients were included in triplicate for the construction of a standard curve. After 60min at room temperature plates were washed, and 100 PI of peroxidase-conjugated anti-human IgG immunoglobulins (DAKO Immunoglobulins, Copenhagen) at a dilution of 1 :3000 were added to each well. After further washing, color was developed with H202/ophenylendiamine, and absorbance read at 492 nm. Production and screening of rnonoclonul anlihodies Monoclonal antibodies CB-4, CB-10, CB-26, and CB-14 were obtained in a previous fusion experiment (14). None of these antibodies cross-react with other Gram-negative bacteria, whereas antibodies CB-4, CB-10, and CB-14 recognize an antigenic determinant expressed also o n gastric mucosa epithelial cells. The new monoclonal antibodies were produced iis follows: Balb/c mice were immunized by subcutaneous administration o f lox sonicated H . pylori organisms at 2-week intervals for a total of three injections. Three days after the last injection the spleen cells were fused with
PU OD 492 nrn 2.0 1.8
>2000
1
**
. . .
::
2000
.*
1.6 1.4 1000
1.2
** **
1.o
.
0.8
500
0 .
400
0.6
250
0.4
125
0.2
0
0
0
2
4
6
8
10
+
-
PU
A
B
Fig. 1 . Indirect enzyme-linked iinmunosorhent assay. 1A. Standard curve. 1B. Serum IgG antibody cnncentriition in 102 endobcopcd patienth whose antral gastric mucosa was positive (+) or negative ( - ) for the presence of‘ Helicohacler pvlori. PU - polyclonul u n i b .
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
Serology of H. pylori with Monoclonal Antibodies
P3-NSl/lAg4.1 murine myeloma cells by the protocol of Galfrk (15). The hybridoma supernatants were assayed for specific antibody production by ELISA, using sonicatecoated microwells as solid phase and peroxidase-conjugated rabbit anti-mouse (DAKO-Immunoglobulins, Copenhagen) as second antibody. Positive hybridomas producing antibodies giving an optical density greater than 0.7 were expanded in 24-well tissue culture plates. To select monoclonal antibodies reacting with antigenic determinants immunogenic also for humans, the positive hybridomas were further screened by a competitive ELISA. A panel of five sera from infected patients and one serum from an uninfected patient was used for this experiment. In brief, SO pi of supernatant of the positive hybridomas were added in duplicate to the antigen-coated wells together with 50 kl of each serum or PBS as control. After 1 h at room temperature the wells were washed and incubated with peroxidase-conjugated rabbit anti-mouse immunoglobulins as above. The color reaction was developed, and the absorbances were read at 492 nm.
601
Bound antibody was then detected with peroxidase antimouse conjugate. The substrate reaction and reading of the test were performed as described above.
Conjugation of antibody HpN45 with horseradish peroxidase The chosen antibody, HpN45, an I g G l immunoglobulin, was purified from ascitic fluid by protein A affinity chromatography and conjugated to horseradish peroxidase (HRP) (Type VI, Sigma Chemical Co.) by the method of Nakane & Kawaoi (16). The conjugate (antibody concentration, 4 mg/ml) was stored at -20°C in PBS-50% glycerin (v/v). Competitive ELISA with labeled monoclonal antibody HpN45 The competition between antibody HpN45 and human antibodies for the H . pylori-specific antigen was measured as follows: 50 pi of each undiluted serum sample were added to sonicate-coated wells. Serial dilutions (from 1: 1 to 1: 128) of the pool containing the 73 sera from the infected patients were included for the construction of the standard curve. Soon afterwards, 50 kl of conjugated antibody diluted 1: 100 in PBS-l% BSA were added to all the wells. After 1 h of incubation at room temperature the plates were washed, and the color reaction was performed as described above. The specimens showing an out-of-range optical density were diluted and re-assayed.
Immunoblotting The H. pylori sonicate was diluted in sodium dodecyl sulfate buffer, heated for 5 min at 100°C, run on a 10% polyacrylamide gel, and transferred to nitrocellulose. After being blocked with 3% BSA, the sheets were incubated with the hybridoma supernatants or positive and negative sera diluted 1 : 500 in PBS with 3% BSA for 90 min at 37°C. After being washed with PBS-0.05% Tween 20 the strips were incubated for 1 h at 37°C with peroxidase-conjugated sheep anti-mouse IgG or anti-human IgG (Amersham). After further washing steps, antibody binding was detected with diaminobenzidine as chromogen.
RESULTS
Reactiiiity of the monoclonal antibodies with H. pylori isolates and other Gram-negative bacteria Each of the 12 bacterial isolates was suspended in distilled water at a concentration of 10’ cells/ml. Each suspension was then distributed into microtiter wells (100 pl/well) and allowed to dry overnight at 37°C. The wells were washed three times in PBS and incubated with hybridoma supernatants for 1 h at room temperature.
H. pylori and gastritis Among the 102 patients examined, H . pylori was detected in 7 by urease test alone, in 4 by histology alone, and in 62 by both of the methods. All 73 infected patients showed antral chronic gastritis histologically. Of the 29 patients who were not infected, 26 had normal antral mucosa and 3 had chronic gastritis with intestinal metaplasia and glandular atrophy.
Statistical analysis The significance of histologic and antibody assay results were determined by chi-square with Yates correction.
Table 1. Patterns of inhibition of the binding of 49 monoclonal antibodies to Helicobacrer pylon antigens by 6 different human sera Group la lb
2 3 4
5
Monoclonal antibody CB-26,HpN 1-16 CB-10,14, HpN 17-31 HPN 32-33 CB-4, HpN 34-40 HpN 41-43 HpN 44-45
* OD 492 nm. 0.C0.2 = -; 0.2-0.5 PBS = phosphate-buffered saline.
= + ; 0.5-1.0
PBS
s1
s2
s3
s4
s5
S6
++” ++ ++ ++ ++ ++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ + +++ +++ +++ ++ ++ +++ +++ +++ ++ ++ ++ ++ ++ ++ +++ +++ = ++; > 1000 = +++; S1 = uninfected patient serum; S2-S6 = infected patient sera;
R . Negrini et ul.
602
kD
-
110
200 11697 66
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
45 31
-
21 14
-
-
84-
47
-
33-
a b c d e f g h i
I
Fig. 2. Reactivity by immunoblotting of monoclonal and polyclonal antibodies against Helicobucter pylori: (a) antibodies HpN I , HpN2, HpN3. HpN17, HpN21, HpN41, klpN42. HpN43; (h) CB-10. CB14, HpN22, HpN23, HpN24; (c) HpN25, HpN26, HpN27; (d) HpN16; (e) HpN32, HpN33; (ICB-4. ) HpN34, HpN36; ( 8 ) HpN35; (h) HpN44, HpN45; (i) H . pylori mouse immune serum; and (I)
normal mouse serum.
Indirect ELISA The standard curve was constructed by using the dilutions of the calibration serum pool that gave optical density ranging betwccn 0.2 and 2.0 (Fig. 1). The highest point of the curve, corresponding to a dilution of 1 :400, was arbitrarily considered as 10 polyclonal units (PU). Among the 73 infected patients 32 had a titer between 250 and 2000 PU, and 41 were above 2000PU (mean, 4000PU). In the 26 patients with normal histology the titer ranged from 80 to 360 PU (mean, 220 PU); of the 3 patients with negative H . pylori detection but with gastritis and intestinal metaplasia 1 had ii titer of 860 and 2 above 2000 P U . Monoclonul antibodies
After initial screening a total of 45 antibodies giving a reading above 0.7 optical density (OD) were selected. On the basis of the pattern of inhibition activity shown by the five human positive sera (Table 1) these new antibodies and the previous antibodies CB-26, CB-4, CB-10, and CB-14 were subdivided into five groups. In the first group there were 34 antibodies whose reactivity was not affected by any sera. Some of these antibodies, studied by immunoblotting, were shown to react with at least four different antigenic determinants: antibody CB-26 was previously shown to react with a 80-kD H. pytori antigen (14). Antibodies HpN1,2.3,17,21 reacted with an antigenic band of 28 kD (Fig. 2, lane a), antibodies CB-10, CB-14, and HpN22-24 with a group of antigenic bands between 33 and 50 kD (Fig. 2, lane b), antibodies HpN25,26,27 with bands between
16 24
HpNSlS2 S 3 S S 5 S 6
45 Fig. 3. Reactivity by immunoblotting of the monoclonal antibody HpN 45, a negative human serum (Sl), and five infected parient sera (S2-S6).
18 and 40 kD (Fig. 2. lane c), and antibody IipNl6 was unreactive (Fig. 2, lane d). The two antibodies of group two (HpN32 and HpN33) were inhibited only by serum 5 (Table I). They reacted strongly with a band of 19kD and, 'very weakly, with a band of 39 kD (Fig. 2, lane e). The immunoblotting analysis of the six human sera confirmed that an immunoresponse against 19 kD antigen was evident only in serum 5 (Fig. 3). Antibodies of group 3 were weakly inhibited by sera 2 and 3. Of these, antibodies CB-4, HpN34, and HpN36 recognized a group of antigens of 30-50 k D (Fig. 2, lane f), and antibody HpN35 a group of antigens between 33 and 43 k D (Fig. 2, lane g). Antibodies of group 4, lIpN41.42,43. were inhibited only by serum 3. They reacted with a 28-kD antigen, like ;Intibodies €lpN1,2,3,17,21.The reactivity of serum 3 with this antigen was also evident by immunoblotting (Fig. 3). The antibodies of group 5 , €IpN44 and 45, were completely inhibited by all the five human sera. They reacted with a 64kD antigen, and this reactivity was evident by immunoblotting in all the live positive sera (Fig. 3). After testing 24 monoclonal antibodies with 12 isolates of H. pylori (Table II), we observed that 19-, 28-, 64-, and 80kD antigens were detected in all isolates; the other antigens were expressed in various percentages. When the entire antigenic profile was considered, the 12 isolates could be subdivided into 9 groups (I-IX, Table 11). None of the antibodies showed a positive reaction when tested with other bacteria (Table 11).
Serology of' H. pylori wirh Monoclortal Ariribodirs I
I
I
I
I
I
+
+
+++ + + + I
I
-
I
+
1+
+ +
+
:++11++;++'>
+
+ +
+
i
++ + + +
+ + ++-
+
I + + + + -
+ +
4
+
t
DISCUSSION
++
+
+y
++++ I
t
+
+++
+
+++
:++11++;+:>
+ + + + + + +;++ ++l:l;++:> +;' + +++ + +
,
+ + + + , + , I + ' =
++
Competitive ELISA with antibody HpN45 The standard curve and the ranges of competitive antibody serum titers in each patient in whom H. pylori was or was not detected are shown in Fig. 4. The level of inhibition of monoclonal antibody HpN45 binding present in the last point of the curve, corresponding to a dilution of the pooled positive sera of 1 : 4, was assumed as 1600 monoclonal units (MU). As a consequence, the mean level of inhibition activity of infected patients was 6400 MU. Of the infected patients 29 had levels of antibodies between 1500 and 4000 MU and 44 above 4000 MU. The mean titer of the noninfected patients was 75 MU. Of the three patients negative for H. pylori infection but with gastritis and intestinal metaplasia one had a titer of 2200 and two above 4000MU. Neither competitive ELISA nor indirect ELISA showed significant statistical correlation between antibody level and grade of activity o r chronic inflammation of the gastric mucosa ( p > 0.05).
++++?
+:Is
8
603
I
::+:> +; + +
I
+;+
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
I
+
+
+
+
+
+
These results, like those of previous studies (11, 17, lX), demonstrated the existence of differences in the antigenic composition between H . pytori isolates. Using a panel of 49 monoclonal antibodies which recognized at least 9 different epitopes (Table II), we were able to subdivide the 12 examined isolates into 9 subgroups, on the basis of their antigenic profile. Four antigenic determinants, expressed on 19-, 28-, 64-, and 80-kD proteins were present in all isolates, whereas the others existed in various percentages. The antigenic variability among H . pylori strains is without doubt responsible for the heterogeneity of the human immunoresponse against H. pylori demonstrated by immunoblotting (6, 19-21). On the other hand, another reason for this heterogeneity seems to be the varied capacity between the patients in immunoresponding to certain antigens. One of these antigens is the 19-kD protein recognized by antibodies HpN 32 and HpN33. A second antigen is a 28-kD protein recognized by antibodies HpN41,42,43. Both these antigenic bands proved to be immunogenic in only one of five sera both by immunoblotting and competitive ELISA. However, we must take into account that at a screening level it was not possible to standardize an optimal condition of sensitivity for every competitive assay; therefore the true immunogenicity grade of the H . pylori antigens will only be determined by this technique in future experiments. The immunoblotting reactivity of the human and murine immune sera showed that other monoclonal antibodies need to be produced t o obtain a complete panel of antibodies against all H . pylori antigens. Using these reagents it will be possible to subdivide this bacteria into serotypes and to ascertain whether there could be any relationships between the antigenic characteristics of the strains and pathogenic activity for the gastroduodenal mucosa. Recently, we demonstrated that, owing to the existence of at least three
604
R . Negrini et ul.
OD 492 nm
. .:
4000'
.*
3500-
3000
-
0
. *. . .
. *
2500-
.
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
2000-
.
1500-
0.4
0.2 0
--
1000-
--
500-
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 O
0
N
O
O
u
O
0
5
O
0
O
80
O
0
,
O
0
O
~
0
~
MU
~
$
-~
+
A Fig. 4. Cornpetitivc enzyme-linked imrnunosorhent assay with monoclonal antibody HpN 45. 4A. Standard curve. 4B. Serum ccirnpetitive antibody concentration in 102 endoscoped patients whose antral gastric mucosa was positive (+) or ncgative ( - ) for the presence o f Helicobucier pvlori. MU = HpN45 monoclonal units.
epitopes that are common for H . pylori and gastric mucosa, H . pylori infection produces antibodies that cross-react with autologous gastric epithelial cells (22). Two of these epitopes, recognized by antibodies CB-4 and CB-10, are shared between H. pylori antigens of molecular weights between 30 and 50 kD and mucosubstances synthesized by gastric epithelial cells. The two determinants are expressed independently by 9 of 12 bacterial isolates. Preliminary results with regard to the study of the immunohistochemical reactivity of some of our new monoclonal antibodies with human and murine gastric mucosa indicated the existence of two other antigenic determinants shared by H . pylori and gastric epithelium. One is recognized by antibody HpN16 and is expressed by only 2 of the 12 H . pylori isolates; the other is recognized by antibody HpN35 and is expressed by 8 isolates (Table 11). These data suggested that differences between strains in the capacity to induce a host autoimmune response exist, implying possible variations in their pathogenicity. After screening by competitive ELISA we identified two antibodies, HpN44 and HpN45, whose binding was inhibited by all five infected patient sera. These antibodies react with an antigen of about 64kD that could correspond to the highly immunogenic protein of about 63 kD described by
Newell (12) and von Wullfen (20). Antibody HpN45 was conjugated with peroxidase, and a one-step Competitive ELISA was developed for serologic study in a group of patients subjected to gastroscopy. In comparison with a classical indirect ELISA, the competitive test showed a greater ability to discriminate between positive and negative sera, probably due to the use of undiluted serum. In contrast. the indirect ELTSA, as reported by other authors (1-9). showed a critical range of values in which both infected and uninfected patients appear. This negative effect is probably due to the use of diluted sera, necessary to minimize the non-specific adsorption of human immunoglobulins by the solid phase. Another advantage of this competitive test is that the specificity of antibody HpN45 for H . pylori eliminates the problem of cross-reactions with C . jejuni (2, 10, 11,12), which may cause misinterpretation of an indirect ELISA. Three patients who were apparently negative for the infection had chronic gastritis with atrophy and intestinal nietaplasia and a high antibody titer in both serologic tests. We believe that they were in fact infected by H . pylori and that the infection persisted in residual areas of non-atrophic mucosa which were hidden from the site of the biopsy. [f we
Serology of H. pylori with Monoclonal Antibodies
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
consider these three patients as being- negative, the indirect ELISA showed a SeLsitivity of 90.4% and a specificity of 89.6%,and the monoclonal competitive ELISA a sensitivity of 100% and specificity of 89.6%. If we consider these patients as being positive, the specificity of both tests was 100%. In conclusion, the competitive ELISA improved the value of serology as a non-invasive tool for the diagnosis of H . pylori infection. Moreover, as a result of its practicality, this test may be considered a very effective method to screen a large number of sera for epidemiologic investigations.
REFERENCES
I . Newell DG, Johnston BJ, Ali MH. Reed PI. An enzyme-linked immunosorbent assay for the serodiagnosis of Campylobactcr pylori-associated gastritis. Scand J Gastrocnterol 1988;23 Suppl 141:53-7. 2. Goodwin CS, Blincow E, Peterson G , et al. Enzyme-linked iniinunosorbcnt assay for Campylobacter pyloridis: correlation with prcscnce of C. pyloridis in t h e gastric mucosa. J Infcct Dis 1987; 155:488-04. 3. Pcrcz-Perez GI, Dworking BM, Chodos JE, Blascr MJ. Campylobacter pylori antibodies in humans. Ann Intern Med l988;l00:l1-7. 4. Bolton FJ, Hutchinson DN. Evaluation of three Campylobacter pylori antigen prcparations for scrccning sera from patients undergoing endoscopy. J Clin Pathol 1989;42:723-6. 5. Lolfcld RJLF, Stobbcringh E, Flcndrig JA. van Sprccuwel JP, Arcnds JW. Diagnostic value o l an immunoassay t o detect anti Campylobacter pylori antibodies in non-ulcer dyspepsia. Lancet 19x0; 1 : 1 182-5. 6. von Wulffen H, Heesemann J . Buetzow GH, et al. Detection of Cainpylobactcr pylori in paticnts with antrum gastritis and pcptic ulcer by culture. complement fixation test and immunoblot. J Clin Microbiol 1986;24:71620. 7. Booth L. Holdstock G, MacBride H , et al. Clinical importance of Campylobacter pylori and associated serum IgG and IgA antibody rcsponscs in patients undergoing upper gastrointestinal endoscopy. J Clin Pathol 1986;39:215-9. 8. Evans DJ Jr. Evans DG, Graham DY, Klein PD. A scnsitive Received 24 May 1991 Accepted 28 January 1992
605
and specific serologic test for detection of Campylobacter pylori .. . . infection. Gastroenterology 1989%: 1004-8. 9. Hirschl AM, Pletschette M, Hirschl MH, Berger J, Stanke G, Ratter ML, of differentantiecn in an evaluation of the immune response to &mpylobacter pylori. Eur C1in MicTobiol 1988;7:57*5. 10. Rathbone BJ, Wyatt JI, Worsley BW, et al. Systemic and local antibody responses to gastric Campylobacter pyloridis in nonulcer dyspepsia. Gut 1986;27:642-7. 11. Perez-Perez GI, Blaser MJ. Conservation and diversity of Campylobacter pylori major antigens. Infect lmmun 1987;55:1 2 5 6 63, 12. Newell DG. Identification of the outer membrane proteins of Campylobacter pyloridis and antigenic cross-reactivity between C. pyloridis and C. jejuni. J Gen Microbiol 1987;133:163-70. 13. Sugiyama T, Imai K, Yoshida H , et al. A novel cnzyme immunoassay for serodiagnosis of Helicobacter pylori infection. Gastroenterology 1991;101:77-83. 14. Negrini R, Lisato L, Cavazzini L, et al. Monoclonal antibodies for specific immunoperoxidase detection of Campylobacter pylori. Gastroenterology 1989;96:414-20. 15. Galfrk G. Preparation of monoclonal antibodies: strategies and procedures. Methods Enzymol 1987;73:1-46. 16. Nakane PK, Kawaoi A . Peroxidasc-labclcd antibody; a ncw method of conjugation. J Histochem Cytochem 1974;22:108491. 17. Burnie JP, Lee W, Dent JC, McNulty CAM. Immunoblot fingerprinting of Campylobacter pylori. J Med Microbiol 1988;27:153-9. 18. Danielsson D, Blomberg B, JBrnerot G , Kosunen TU. Heterogeneity of Campylobacter pylori as demonstrated by coagglutination testing with rabbit antibodies. Scand J Gastroentcrol 1988;23 Suppl 142:58-63. 19. Jones DM, Eldridge J , Fox AJ, Sethi P, Whorwell PJ. Antibody to the gastric Campylobacter-like organism (Campylobacter pyloridis). Clinical correlations and distribution in the normal populations. J Med Microbiol 1986;22:57-62. 20. von Wulffen H, Grote HJ, Gatermann S, Loning T, Berger B, Buhl C. Immunoblot analysis of immune response to Campylobacter pylori and its clinical associations. J Clin Pathol 1988;4 I :653-9. 21. Kaldor J , Tee W, Nicolacopolous C, Demirtzoglou K, Noonan D, Dwyer B. Immunoblot confirmation of immune response to Campylobacter pyloridis in patients with duodenal ulcers. Med J Aust 1986;145:133-5. 22. Negrini R, Lisato L, Zanella I, et al. Helicobacter pylori infection induces antibodies cross-reacting with human gastric mucosa. Gastroenterology 1991;1O1:437-45.
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
Negrini R , Zanella I, Savio A, Poiesi C, Verardi R, Ghielmi S, Albertini A , Sangaletti 0 , Lazzaroni M, Bianchi Porro G. Serodiagnosis of Helicobacterpylori-associated gastritis with a monoclonal antibody competitive enzyme-linked immunosorbent assay. Scand J Gastroenterol 1992;27:599-605, Forty-nine monoclonal antibodies against Helicobacier pylori were screened to investigate their capacity to be used in enzyme-linked immunosorbent assay (ELISA) competitive systems for the serodiagnosis of Helicobacterpylori infection. On the basis of the inhibition pattern showed by the sera of five infected patients. the antibodies were subdivided into five groups. The immunoblotting analysis showed that thc antibodies recognized a total of nine different antigenic determinants. In a study of the reaction of the antibodies with 12 isolates of H . pylori a total of 9 antigenic profiles were identified. Two monoclonal antibodies, HpN44 and HpN45, which recognized a 64-kD protein, were inhibited by all 5 positive sera. Antibody HpN45 was labeled with horseradish peroxidase, and the competitive ELISA was compared with an ordinary indirect ELISA in a study of 102 patients undergoing gastroscopy. Seventy-three patients proved to he infected by H . pylori according to urease or histologic tests. The sensitivity and specificity were 90.4% and 89.6%, respectively, for the indirect ELISA and 100% and 89.6% for the HpN45 competitive assay. The three patients who were 'false seropositive' with both serologic tests had atrophic gastritis. The high diagnostic performance and simplicity of the HpN45 monoclonal competitive ELISA make it suitable for routine serodiagnosis of H . pylori infection.
Key words: Enzyme-linked immunosorbent assay; Helicobacier pylori; monoclonal antibodies: serology Riccardo Negrini, M . D..Lahoratorio di Biotecnologie, Spedali Civili di Bre.wia. P.le Spedali Civili, 1 . 25123 Bre.wia, Ilaly
Using indirect enzyme-linked immunosorbent assay (ELISA) tests, many authors have demonstrated a high degree of correlation between the presence of serum antibodies against Helicobucter pylori and gastric infection by this bacterium. However, these tests showed a different grade of sensitivity and specificity (1-9), due to a large extent to variations in a) the antigenic preparation used (4, 9), b) the different dilutions of serum, c) the different values of the set cut-off point, and d) the method used as a 'gold standard' to decide whether a patient has H. pylori in the stomach. The major problem with regard to specificity seems to be cross-reactions due to the presence of H. pylori antigens shared with other bacteria, such as Cumpylobucter jejuni (2,lO-12). The use of monoclonal antibodies for the detection of specific antibodies against H. pylori by indirect ELISA has already been reported (13). In a previous paper we described the use of monoclonal antibodies for the specific immunohistochemical detection of H . pylori (14). The present study was undertaken to evaluate the possible capacity of monoclonal antibodies to be used in competitive ELISA systems for the detection of human antibodies against single H. pylori antigens. One of
these tests was selected and compared with the conventional indirect approach.
PATIENTS A N D METHODS Putients
One-hundred and two patients consecutively undergoing gastroscopy were examined. In each patient biopsy specimens were obtained from the antral mucosa for the histologic and microbiologic examination. Blood samples were collected for the detection of antibodies against H . pylori. Histology
The biopsy specimens were immediately formalin-fixed and paraffin-embedded; the diagnosis of gastritis was made by examining sections stained with haematoxylin and eosin. The histologic changes of the antrum (chronic inflammation, atrophy. activity, intestinal metaplasia. and density of H . pylori) were divided into four grades, in accordance with the Sidney system for the classification of gastritis. The presence of H . pylori was detected by the examination of sections stained with Giemsa.
600
R . Negrini et d.
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
Urease test The presence of bacterial urease in the biopsy specimen was detected by means of a commercial rapid test (Quick Campy1 Test-Brocades, Milan, Italy). Bacteria The 12 isolates of H . pylori used in this study were kindly provided by Dr. F. Carozzi (Biolife-Milano, Italy) and were derived from stored bacteria obtained from human gastric biopsy specimens. The bacteria were identified on the basis of the morphologic features of the colonies and from Gram stain and positive reaction to urease, catalase, and oxidase. The isolates were subcultured on 6% horse blood agar at 37°C in a microaerobic atmosphere. After 5-8 days the bacteria were harvested and washed in phosphate-buffered saline (PBS). The resulting pellet was suspended in PBS at a conccntration of IO’ceIIs/mI and frozen at -20°C. A pool of all the 12 isolates of H. pylori, at a final concentration of 10’ cells/ml in PBS-0.5% Nonidet P40, was ultrasonicated and used as antigen for serologic and iminunoblotting studies. Isolates of C. jejuni, Escherickia coli, Vihrio parahaernoliticus, Klebsiella pneumoniae, Proteus rettgeri, and Pseudoinonas aeruginosa were kindly provided by Dr. P. Maini (USL.31, Ferrara).
Indirect ELlSA The polyclonal serological IgG activity against H . pylori was measured as follows: 100 PI of serum diluted 1 :200 in PBS-l% (BSA)-0.05% Tween 20 were added in duplicate to antigen-coated microwells. Serial dilutions (from 1 : 200 to 1:12,800) of a pool of the sera from 73 infected patients were included in triplicate for the construction of a standard curve. After 60min at room temperature plates were washed, and 100 PI of peroxidase-conjugated anti-human IgG immunoglobulins (DAKO Immunoglobulins, Copenhagen) at a dilution of 1 :3000 were added to each well. After further washing, color was developed with H202/ophenylendiamine, and absorbance read at 492 nm. Production and screening of rnonoclonul anlihodies Monoclonal antibodies CB-4, CB-10, CB-26, and CB-14 were obtained in a previous fusion experiment (14). None of these antibodies cross-react with other Gram-negative bacteria, whereas antibodies CB-4, CB-10, and CB-14 recognize an antigenic determinant expressed also o n gastric mucosa epithelial cells. The new monoclonal antibodies were produced iis follows: Balb/c mice were immunized by subcutaneous administration o f lox sonicated H . pylori organisms at 2-week intervals for a total of three injections. Three days after the last injection the spleen cells were fused with
PU OD 492 nrn 2.0 1.8
>2000
1
**
. . .
::
2000
.*
1.6 1.4 1000
1.2
** **
1.o
.
0.8
500
0 .
400
0.6
250
0.4
125
0.2
0
0
0
2
4
6
8
10
+
-
PU
A
B
Fig. 1 . Indirect enzyme-linked iinmunosorhent assay. 1A. Standard curve. 1B. Serum IgG antibody cnncentriition in 102 endobcopcd patienth whose antral gastric mucosa was positive (+) or negative ( - ) for the presence of‘ Helicohacler pvlori. PU - polyclonul u n i b .
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
Serology of H. pylori with Monoclonal Antibodies
P3-NSl/lAg4.1 murine myeloma cells by the protocol of Galfrk (15). The hybridoma supernatants were assayed for specific antibody production by ELISA, using sonicatecoated microwells as solid phase and peroxidase-conjugated rabbit anti-mouse (DAKO-Immunoglobulins, Copenhagen) as second antibody. Positive hybridomas producing antibodies giving an optical density greater than 0.7 were expanded in 24-well tissue culture plates. To select monoclonal antibodies reacting with antigenic determinants immunogenic also for humans, the positive hybridomas were further screened by a competitive ELISA. A panel of five sera from infected patients and one serum from an uninfected patient was used for this experiment. In brief, SO pi of supernatant of the positive hybridomas were added in duplicate to the antigen-coated wells together with 50 kl of each serum or PBS as control. After 1 h at room temperature the wells were washed and incubated with peroxidase-conjugated rabbit anti-mouse immunoglobulins as above. The color reaction was developed, and the absorbances were read at 492 nm.
601
Bound antibody was then detected with peroxidase antimouse conjugate. The substrate reaction and reading of the test were performed as described above.
Conjugation of antibody HpN45 with horseradish peroxidase The chosen antibody, HpN45, an I g G l immunoglobulin, was purified from ascitic fluid by protein A affinity chromatography and conjugated to horseradish peroxidase (HRP) (Type VI, Sigma Chemical Co.) by the method of Nakane & Kawaoi (16). The conjugate (antibody concentration, 4 mg/ml) was stored at -20°C in PBS-50% glycerin (v/v). Competitive ELISA with labeled monoclonal antibody HpN45 The competition between antibody HpN45 and human antibodies for the H . pylori-specific antigen was measured as follows: 50 pi of each undiluted serum sample were added to sonicate-coated wells. Serial dilutions (from 1: 1 to 1: 128) of the pool containing the 73 sera from the infected patients were included for the construction of the standard curve. Soon afterwards, 50 kl of conjugated antibody diluted 1: 100 in PBS-l% BSA were added to all the wells. After 1 h of incubation at room temperature the plates were washed, and the color reaction was performed as described above. The specimens showing an out-of-range optical density were diluted and re-assayed.
Immunoblotting The H. pylori sonicate was diluted in sodium dodecyl sulfate buffer, heated for 5 min at 100°C, run on a 10% polyacrylamide gel, and transferred to nitrocellulose. After being blocked with 3% BSA, the sheets were incubated with the hybridoma supernatants or positive and negative sera diluted 1 : 500 in PBS with 3% BSA for 90 min at 37°C. After being washed with PBS-0.05% Tween 20 the strips were incubated for 1 h at 37°C with peroxidase-conjugated sheep anti-mouse IgG or anti-human IgG (Amersham). After further washing steps, antibody binding was detected with diaminobenzidine as chromogen.
RESULTS
Reactiiiity of the monoclonal antibodies with H. pylori isolates and other Gram-negative bacteria Each of the 12 bacterial isolates was suspended in distilled water at a concentration of 10’ cells/ml. Each suspension was then distributed into microtiter wells (100 pl/well) and allowed to dry overnight at 37°C. The wells were washed three times in PBS and incubated with hybridoma supernatants for 1 h at room temperature.
H. pylori and gastritis Among the 102 patients examined, H . pylori was detected in 7 by urease test alone, in 4 by histology alone, and in 62 by both of the methods. All 73 infected patients showed antral chronic gastritis histologically. Of the 29 patients who were not infected, 26 had normal antral mucosa and 3 had chronic gastritis with intestinal metaplasia and glandular atrophy.
Statistical analysis The significance of histologic and antibody assay results were determined by chi-square with Yates correction.
Table 1. Patterns of inhibition of the binding of 49 monoclonal antibodies to Helicobacrer pylon antigens by 6 different human sera Group la lb
2 3 4
5
Monoclonal antibody CB-26,HpN 1-16 CB-10,14, HpN 17-31 HPN 32-33 CB-4, HpN 34-40 HpN 41-43 HpN 44-45
* OD 492 nm. 0.C0.2 = -; 0.2-0.5 PBS = phosphate-buffered saline.
= + ; 0.5-1.0
PBS
s1
s2
s3
s4
s5
S6
++” ++ ++ ++ ++ ++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ + +++ +++ +++ ++ ++ +++ +++ +++ ++ ++ ++ ++ ++ ++ +++ +++ = ++; > 1000 = +++; S1 = uninfected patient serum; S2-S6 = infected patient sera;
R . Negrini et ul.
602
kD
-
110
200 11697 66
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
45 31
-
21 14
-
-
84-
47
-
33-
a b c d e f g h i
I
Fig. 2. Reactivity by immunoblotting of monoclonal and polyclonal antibodies against Helicobucter pylori: (a) antibodies HpN I , HpN2, HpN3. HpN17, HpN21, HpN41, klpN42. HpN43; (h) CB-10. CB14, HpN22, HpN23, HpN24; (c) HpN25, HpN26, HpN27; (d) HpN16; (e) HpN32, HpN33; (ICB-4. ) HpN34, HpN36; ( 8 ) HpN35; (h) HpN44, HpN45; (i) H . pylori mouse immune serum; and (I)
normal mouse serum.
Indirect ELISA The standard curve was constructed by using the dilutions of the calibration serum pool that gave optical density ranging betwccn 0.2 and 2.0 (Fig. 1). The highest point of the curve, corresponding to a dilution of 1 :400, was arbitrarily considered as 10 polyclonal units (PU). Among the 73 infected patients 32 had a titer between 250 and 2000 PU, and 41 were above 2000PU (mean, 4000PU). In the 26 patients with normal histology the titer ranged from 80 to 360 PU (mean, 220 PU); of the 3 patients with negative H . pylori detection but with gastritis and intestinal metaplasia 1 had ii titer of 860 and 2 above 2000 P U . Monoclonul antibodies
After initial screening a total of 45 antibodies giving a reading above 0.7 optical density (OD) were selected. On the basis of the pattern of inhibition activity shown by the five human positive sera (Table 1) these new antibodies and the previous antibodies CB-26, CB-4, CB-10, and CB-14 were subdivided into five groups. In the first group there were 34 antibodies whose reactivity was not affected by any sera. Some of these antibodies, studied by immunoblotting, were shown to react with at least four different antigenic determinants: antibody CB-26 was previously shown to react with a 80-kD H. pytori antigen (14). Antibodies HpN1,2.3,17,21 reacted with an antigenic band of 28 kD (Fig. 2, lane a), antibodies CB-10, CB-14, and HpN22-24 with a group of antigenic bands between 33 and 50 kD (Fig. 2, lane b), antibodies HpN25,26,27 with bands between
16 24
HpNSlS2 S 3 S S 5 S 6
45 Fig. 3. Reactivity by immunoblotting of the monoclonal antibody HpN 45, a negative human serum (Sl), and five infected parient sera (S2-S6).
18 and 40 kD (Fig. 2. lane c), and antibody IipNl6 was unreactive (Fig. 2, lane d). The two antibodies of group two (HpN32 and HpN33) were inhibited only by serum 5 (Table I). They reacted strongly with a band of 19kD and, 'very weakly, with a band of 39 kD (Fig. 2, lane e). The immunoblotting analysis of the six human sera confirmed that an immunoresponse against 19 kD antigen was evident only in serum 5 (Fig. 3). Antibodies of group 3 were weakly inhibited by sera 2 and 3. Of these, antibodies CB-4, HpN34, and HpN36 recognized a group of antigens of 30-50 k D (Fig. 2, lane f), and antibody HpN35 a group of antigens between 33 and 43 k D (Fig. 2, lane g). Antibodies of group 4, lIpN41.42,43. were inhibited only by serum 3. They reacted with a 28-kD antigen, like ;Intibodies €lpN1,2,3,17,21.The reactivity of serum 3 with this antigen was also evident by immunoblotting (Fig. 3). The antibodies of group 5 , €IpN44 and 45, were completely inhibited by all the five human sera. They reacted with a 64kD antigen, and this reactivity was evident by immunoblotting in all the live positive sera (Fig. 3). After testing 24 monoclonal antibodies with 12 isolates of H. pylori (Table II), we observed that 19-, 28-, 64-, and 80kD antigens were detected in all isolates; the other antigens were expressed in various percentages. When the entire antigenic profile was considered, the 12 isolates could be subdivided into 9 groups (I-IX, Table 11). None of the antibodies showed a positive reaction when tested with other bacteria (Table 11).
Serology of' H. pylori wirh Monoclortal Ariribodirs I
I
I
I
I
I
+
+
+++ + + + I
I
-
I
+
1+
+ +
+
:++11++;++'>
+
+ +
+
i
++ + + +
+ + ++-
+
I + + + + -
+ +
4
+
t
DISCUSSION
++
+
+y
++++ I
t
+
+++
+
+++
:++11++;+:>
+ + + + + + +;++ ++l:l;++:> +;' + +++ + +
,
+ + + + , + , I + ' =
++
Competitive ELISA with antibody HpN45 The standard curve and the ranges of competitive antibody serum titers in each patient in whom H. pylori was or was not detected are shown in Fig. 4. The level of inhibition of monoclonal antibody HpN45 binding present in the last point of the curve, corresponding to a dilution of the pooled positive sera of 1 : 4, was assumed as 1600 monoclonal units (MU). As a consequence, the mean level of inhibition activity of infected patients was 6400 MU. Of the infected patients 29 had levels of antibodies between 1500 and 4000 MU and 44 above 4000 MU. The mean titer of the noninfected patients was 75 MU. Of the three patients negative for H. pylori infection but with gastritis and intestinal metaplasia one had a titer of 2200 and two above 4000MU. Neither competitive ELISA nor indirect ELISA showed significant statistical correlation between antibody level and grade of activity o r chronic inflammation of the gastric mucosa ( p > 0.05).
++++?
+:Is
8
603
I
::+:> +; + +
I
+;+
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
I
+
+
+
+
+
+
These results, like those of previous studies (11, 17, lX), demonstrated the existence of differences in the antigenic composition between H . pytori isolates. Using a panel of 49 monoclonal antibodies which recognized at least 9 different epitopes (Table II), we were able to subdivide the 12 examined isolates into 9 subgroups, on the basis of their antigenic profile. Four antigenic determinants, expressed on 19-, 28-, 64-, and 80-kD proteins were present in all isolates, whereas the others existed in various percentages. The antigenic variability among H . pylori strains is without doubt responsible for the heterogeneity of the human immunoresponse against H. pylori demonstrated by immunoblotting (6, 19-21). On the other hand, another reason for this heterogeneity seems to be the varied capacity between the patients in immunoresponding to certain antigens. One of these antigens is the 19-kD protein recognized by antibodies HpN 32 and HpN33. A second antigen is a 28-kD protein recognized by antibodies HpN41,42,43. Both these antigenic bands proved to be immunogenic in only one of five sera both by immunoblotting and competitive ELISA. However, we must take into account that at a screening level it was not possible to standardize an optimal condition of sensitivity for every competitive assay; therefore the true immunogenicity grade of the H . pylori antigens will only be determined by this technique in future experiments. The immunoblotting reactivity of the human and murine immune sera showed that other monoclonal antibodies need to be produced t o obtain a complete panel of antibodies against all H . pylori antigens. Using these reagents it will be possible to subdivide this bacteria into serotypes and to ascertain whether there could be any relationships between the antigenic characteristics of the strains and pathogenic activity for the gastroduodenal mucosa. Recently, we demonstrated that, owing to the existence of at least three
604
R . Negrini et ul.
OD 492 nm
. .:
4000'
.*
3500-
3000
-
0
. *. . .
. *
2500-
.
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
2000-
.
1500-
0.4
0.2 0
--
1000-
--
500-
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 O
0
N
O
O
u
O
0
5
O
0
O
80
O
0
,
O
0
O
~
0
~
MU
~
$
-~
+
A Fig. 4. Cornpetitivc enzyme-linked imrnunosorhent assay with monoclonal antibody HpN 45. 4A. Standard curve. 4B. Serum ccirnpetitive antibody concentration in 102 endoscoped patients whose antral gastric mucosa was positive (+) or ncgative ( - ) for the presence o f Helicobucier pvlori. MU = HpN45 monoclonal units.
epitopes that are common for H . pylori and gastric mucosa, H . pylori infection produces antibodies that cross-react with autologous gastric epithelial cells (22). Two of these epitopes, recognized by antibodies CB-4 and CB-10, are shared between H. pylori antigens of molecular weights between 30 and 50 kD and mucosubstances synthesized by gastric epithelial cells. The two determinants are expressed independently by 9 of 12 bacterial isolates. Preliminary results with regard to the study of the immunohistochemical reactivity of some of our new monoclonal antibodies with human and murine gastric mucosa indicated the existence of two other antigenic determinants shared by H . pylori and gastric epithelium. One is recognized by antibody HpN16 and is expressed by only 2 of the 12 H . pylori isolates; the other is recognized by antibody HpN35 and is expressed by 8 isolates (Table 11). These data suggested that differences between strains in the capacity to induce a host autoimmune response exist, implying possible variations in their pathogenicity. After screening by competitive ELISA we identified two antibodies, HpN44 and HpN45, whose binding was inhibited by all five infected patient sera. These antibodies react with an antigen of about 64kD that could correspond to the highly immunogenic protein of about 63 kD described by
Newell (12) and von Wullfen (20). Antibody HpN45 was conjugated with peroxidase, and a one-step Competitive ELISA was developed for serologic study in a group of patients subjected to gastroscopy. In comparison with a classical indirect ELISA, the competitive test showed a greater ability to discriminate between positive and negative sera, probably due to the use of undiluted serum. In contrast. the indirect ELTSA, as reported by other authors (1-9). showed a critical range of values in which both infected and uninfected patients appear. This negative effect is probably due to the use of diluted sera, necessary to minimize the non-specific adsorption of human immunoglobulins by the solid phase. Another advantage of this competitive test is that the specificity of antibody HpN45 for H . pylori eliminates the problem of cross-reactions with C . jejuni (2, 10, 11,12), which may cause misinterpretation of an indirect ELISA. Three patients who were apparently negative for the infection had chronic gastritis with atrophy and intestinal nietaplasia and a high antibody titer in both serologic tests. We believe that they were in fact infected by H . pylori and that the infection persisted in residual areas of non-atrophic mucosa which were hidden from the site of the biopsy. [f we
Serology of H. pylori with Monoclonal Antibodies
Scand J Gastroenterol Downloaded from informahealthcare.com by Queen's University on 01/03/15 For personal use only.
consider these three patients as being- negative, the indirect ELISA showed a SeLsitivity of 90.4% and a specificity of 89.6%,and the monoclonal competitive ELISA a sensitivity of 100% and specificity of 89.6%. If we consider these patients as being positive, the specificity of both tests was 100%. In conclusion, the competitive ELISA improved the value of serology as a non-invasive tool for the diagnosis of H . pylori infection. Moreover, as a result of its practicality, this test may be considered a very effective method to screen a large number of sera for epidemiologic investigations.
REFERENCES
I . Newell DG, Johnston BJ, Ali MH. Reed PI. An enzyme-linked immunosorbent assay for the serodiagnosis of Campylobactcr pylori-associated gastritis. Scand J Gastrocnterol 1988;23 Suppl 141:53-7. 2. Goodwin CS, Blincow E, Peterson G , et al. Enzyme-linked iniinunosorbcnt assay for Campylobacter pyloridis: correlation with prcscnce of C. pyloridis in t h e gastric mucosa. J Infcct Dis 1987; 155:488-04. 3. Pcrcz-Perez GI, Dworking BM, Chodos JE, Blascr MJ. Campylobacter pylori antibodies in humans. Ann Intern Med l988;l00:l1-7. 4. Bolton FJ, Hutchinson DN. Evaluation of three Campylobacter pylori antigen prcparations for scrccning sera from patients undergoing endoscopy. J Clin Pathol 1989;42:723-6. 5. Lolfcld RJLF, Stobbcringh E, Flcndrig JA. van Sprccuwel JP, Arcnds JW. Diagnostic value o l an immunoassay t o detect anti Campylobacter pylori antibodies in non-ulcer dyspepsia. Lancet 19x0; 1 : 1 182-5. 6. von Wulffen H, Heesemann J . Buetzow GH, et al. Detection of Cainpylobactcr pylori in paticnts with antrum gastritis and pcptic ulcer by culture. complement fixation test and immunoblot. J Clin Microbiol 1986;24:71620. 7. Booth L. Holdstock G, MacBride H , et al. Clinical importance of Campylobacter pylori and associated serum IgG and IgA antibody rcsponscs in patients undergoing upper gastrointestinal endoscopy. J Clin Pathol 1986;39:215-9. 8. Evans DJ Jr. Evans DG, Graham DY, Klein PD. A scnsitive Received 24 May 1991 Accepted 28 January 1992
605
and specific serologic test for detection of Campylobacter pylori .. . . infection. Gastroenterology 1989%: 1004-8. 9. Hirschl AM, Pletschette M, Hirschl MH, Berger J, Stanke G, Ratter ML, of differentantiecn in an evaluation of the immune response to &mpylobacter pylori. Eur C1in MicTobiol 1988;7:57*5. 10. Rathbone BJ, Wyatt JI, Worsley BW, et al. Systemic and local antibody responses to gastric Campylobacter pyloridis in nonulcer dyspepsia. Gut 1986;27:642-7. 11. Perez-Perez GI, Blaser MJ. Conservation and diversity of Campylobacter pylori major antigens. Infect lmmun 1987;55:1 2 5 6 63, 12. Newell DG. Identification of the outer membrane proteins of Campylobacter pyloridis and antigenic cross-reactivity between C. pyloridis and C. jejuni. J Gen Microbiol 1987;133:163-70. 13. Sugiyama T, Imai K, Yoshida H , et al. A novel cnzyme immunoassay for serodiagnosis of Helicobacter pylori infection. Gastroenterology 1991;101:77-83. 14. Negrini R, Lisato L, Cavazzini L, et al. Monoclonal antibodies for specific immunoperoxidase detection of Campylobacter pylori. Gastroenterology 1989;96:414-20. 15. Galfrk G. Preparation of monoclonal antibodies: strategies and procedures. Methods Enzymol 1987;73:1-46. 16. Nakane PK, Kawaoi A . Peroxidasc-labclcd antibody; a ncw method of conjugation. J Histochem Cytochem 1974;22:108491. 17. Burnie JP, Lee W, Dent JC, McNulty CAM. Immunoblot fingerprinting of Campylobacter pylori. J Med Microbiol 1988;27:153-9. 18. Danielsson D, Blomberg B, JBrnerot G , Kosunen TU. Heterogeneity of Campylobacter pylori as demonstrated by coagglutination testing with rabbit antibodies. Scand J Gastroentcrol 1988;23 Suppl 142:58-63. 19. Jones DM, Eldridge J , Fox AJ, Sethi P, Whorwell PJ. Antibody to the gastric Campylobacter-like organism (Campylobacter pyloridis). Clinical correlations and distribution in the normal populations. J Med Microbiol 1986;22:57-62. 20. von Wulffen H, Grote HJ, Gatermann S, Loning T, Berger B, Buhl C. Immunoblot analysis of immune response to Campylobacter pylori and its clinical associations. J Clin Pathol 1988;4 I :653-9. 21. Kaldor J , Tee W, Nicolacopolous C, Demirtzoglou K, Noonan D, Dwyer B. Immunoblot confirmation of immune response to Campylobacter pyloridis in patients with duodenal ulcers. Med J Aust 1986;145:133-5. 22. Negrini R, Lisato L, Zanella I, et al. Helicobacter pylori infection induces antibodies cross-reacting with human gastric mucosa. Gastroenterology 1991;1O1:437-45.
