A Strain-Specific Antigen in Japanese Helicobacter pylori Recognized in Sera of Japanese Children

Abstract

An enzyme immuno assay (EIA) test based on Japanese strain-derived high-molecular-weight cell-associated proteins (JHM-CAP) was evaluated by comparing with a previously developed EIA test based on a U.S. strain-derived high-molecular-weight cell-associated proteins (HM-CAP). Serum samples of 131 Japanese asymptomatic children (mean age, 5.5 years; range, 0 to 21 years) were tested that include 43 positive and 88 negative children as judged by Helicobacter pylori stool antigen test (HpSA test). Both tests showed comparable and reliable specificities, but the sensitivity of JHM-CAP EIA, at 93.0%, was much higher than that of HM-CAP EIA, at 67.4%. More false-negative results of HM-CAP were obtained in children under 10 years of age. Immunoblot analysis revealed that the JHM-CAP but not the HM-CAP preparation had a 100-kDa antigen recognized by JHM-CAP positive sera. It was concluded that JHM-CAP EIA is highly accurate for the serodiagnosis of H. pylori infection in Japanese young children and that the high sensitivity of JHM-CAP EIA in contrast to HM-CAP EIA is due to the presence of a 100-kDa antigen in Japanese strains that may be recognized by the host immune system at an early stage of infection.

Helicobacter pylori infects half the population of the world, causes nonimmune gastritis and peptic ulcer diseases, and is associated with gastric cancer (27, 30, 31). It is generally believed that most H. pylori infections are acquired early in childhood (12, 25) and persist for decades with a variety of clinical outcomes, but the epidemiology, including the transmission route, is still obscure. In addition, association of H. pylori infections with epigastric clinical conditions such as iron deficiency anemia (2) and thrombocytopenic purpura (5) in children has been reported recently. Therefore, the demand for an accurate diagnosis of H. pylori infection in both symptomatic and asymptomatic children is increasing.

It has been reported that commercial enzyme immunoassay (EIA) tests may yield variable diagnostic performances in oriental populations in different regions, and the accuracy was greatly reduced when sera from developing countries and younger ages were evaluated (7, 9-11, 16, 28, 29). We reported similar results that serodiagnosis by HM-CAP EIA on Japanese asymptomatic children under 10 years of age had 51.4% sensitivity and 97.2% specificity in reference to H. pylori stool antigen test (HpSA test) (20). In contrast, a commercially available urine-based EIA test, URINELISA, is highly reliable with 92.8% sensitivity and 95.8% specificity in reference to the ¹³C-labeled urea breath test (UBT) and HpSA test on the same age group of a similar population (21). The difference in the sensitivity might be due to antigen preparation or strain variation; i.e., URINELISA antigen was made of sonic extracts of a Japanese isolate (21), whereas HM-CAP EIA was made of high-molecular-weight cell-associated proteins purified from a U.S. isolate (15).

Recently, a serum-based EIA test of Japanese strain-derived high-molecular-weight proteins (JHM-CAP) prepared by exactly the same procedure as for HM-CAP EIA has been developed. The diagnostic performances of JHM-CAP EIA and HM-CAP EIA on adult populations were reported using UBT as the gold standard. For the samples of Japanese populations, the performance of JHM-CAP EIA is slightly better than that of HM-CAP EIA (15, 18), although there was no difference for the samples of a U.S. population (15).

Antibodies in the sera of H. pylori-infected children could be analyzed by immunoblotting. Two commercial immunoblot tests, Helico Blot 2.0 and Helico Blot 2.1, have been developed that were proved to be suitable for the serodiagnosis of H. pylori infection in children (22, 24, 29). Both of these tests used antigens derived from western strains: Helico Blot 2.0 from NTCT 11916 and Helico Blot 2.1 from ATCC 49503 (23).

In the present study, we found that the diagnostic performance of JHM-CAP EIA is much better than that of HM-CAP EIA for Japanese asymptomatic children. We used the HpSA test as the gold standard, since the assay was proved to be accurate for Japanese children by a multicenter study (8). In addition, we have shown that both the sensitivity and the specificity of HpSA test were 100% in reference to UBT without any equivocal cases on diagnosis in a population of children similar to that of the present study (21). To elucidate the cause of such a remarkable difference, immunoblot analysis of JHM-CAP and HM-CAP antigens was performed. We found that the JHM-CAP preparation has an antigen reactive to the sera of Japanese asymptomatic children that is absent from the HM-CAP preparation.

MATERIALS AND METHODS

Study population and clinical samples.

The present study included 131 Japanese children (mean age, 5.5 years; range, 0 to 21 years; male, n = 69; female, n = 62) living in Wakayama Prefecture, a western part of Japan near Osaka. In a previous epidemiological study at Wakayama Rosai Hospital, 484 children were tested by HpSA, and 31 positive children were found (19). They were not treated because of the absence of upper-abdominal symptoms. In the present study, we reinvited 29 positive children from the previous study. None of the children had signs of gastrointestinal diseases or history of H. pylori eradication. Informed consent was obtained from all parents, and the research protocols were approved by the Institution Review Board of Wakayama Rosai Hospital. Serum and stool samples of the children were collected on the same day in most cases and within 1 month in some cases. These samples were stored at −20°C and used within guaranteed periods.

HpSA test.

H. pylori antigens in stool samples were determined by an EIA test (Premier Platinum HpSA; Meridian Diagnostics, Inc., Cincinnati, OH) according to the manufacturer's instruction. The test was based on immunoaffinity-purified polyclonal anti-H. pylori rabbit antibodies raised against a U.S. isolate. Diluted samples (ca. 50 mg of stool in 200 μl of buffer) and peroxidase-conjugated secondary polyclonal antibodies were added to microwells, followed by incubation for 1 h at room temperature. Plates were read spectrophotometrically at 450 nm with a reference at 630 nm using a plate reader. HpSA values of ≥0.120, <0.120 to ≥0.100, and <0.100 were judged to be positive, equivocal, and negative, respectively.

JHM-CAP EIA and HM-CAP EIA.

Serum immunoglobulin G antibodies to H. pylori were measured by using commercial EIA tests of JHM-CAP (Scimedx Corp., Denville, NJ) and HM-CAP (Scimedx Corp., Denville, NJ). The assay was carried out according to the manufacturer's instructions, and EIA values (EVs) of ≥2.3, <2.3 to ≥1.8, and <1.8 were judged to be positive, equivocal, and negative, respectively.

UBT.

UBT was performed after fasting for at least 4 h using an infrared spectrometer (UBiT-IR200; Otsuka Electronics Co., Hirakata, Japan). The dosage of [¹³C]urea was altered according to the age of children: <6 years old, 50 mg; 7 to 12 years old, 75 mg; >13 years old, 100 mg. An increase of more than 3.5‰ was considered positive (3).

Helico Blot 2.1 analysis.

Sera were assayed for anti-H. pylori antibodies by using a commercial immunoblot test, Helico Blot 2.1 (Genelabs Diagnostic, Singapore) according to the manufacturer's instructions. This serological test consists of a Western blot with a lysate of H. pylori strain ATCC 49503, together with a recombinant antigen designated as a current infection marker (23).

SDS-PAGE and immunoblot analysis.

The high-molecular-weight cell-associated proteins used for JHM-CAP EIA were mixtures of proteins purified from four Japanese strains isolated in Hokkaido University, whereas those for HM-CAP EIA were proteins purified from a U.S. strain (197SR-US) isolated at the Baylor College of Medicine (15) according to a published procedure (6). In brief, crude extracts solubilized with 1% n-octyl alcohol-containing phosphate-buffered saline were dialyzed against buffered saline and centrifuged to obtain soluble proteins. Then, high-molecular-weight proteins were separated by passing through an agarose A-5m column equilibrated with 0.05 M Tris-HCl buffer (pH 8.0), followed by ammonium-sulfate precipitation and dialysis to obtain JHM-CAP and HM-CAP antigens with protein concentrations of 7.2 and 11.2 mg/ml, respectively. Antigens were appropriately diluted and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with a 5 to 20% gradient gel. After electrophoresis, the proteins were transferred to a nitrocellulose membrane, and immunoblotting was carried out with 500- fold-diluted sera. Peroxidase-conjugated goat anti-human immunoglobulin G (American Qualex International, Inc.) diluted 10,000-fold by phosphate-buffered saline was used to visualize reactive antigens to be detected by enhanced chemiluminescence.

Statistical analysis.

Statistical analysis was performed with StatView software (SAS Institute, Cary, NC) to obtain a correlation coefficient for the EVs of JHM-CAP and HM-CAP. Variables were evaluated by using the Spearman rank correlation coefficient.

RESULTS

When H. pylori infection of 131 Japanese asymptomatic children was determined by HpSA test, 43 and 88 children were judged positive and negative, respectively, without any equivocal cases. The positive rate (32.8%) in the present study was higher than the average prevalence of 6 to 7% in this region of Japan, since we reinvited 29 positive children from a previous study (19).

The performances of JHM-CAP EIA and HM-CAP EIA on sera of the children were compared (Table 1). By using HpSA test as the gold standard, JHM-CAP EIA showed a higher sensitivity compared to HM-CAP EIA (overall, 93.0 to 67.4%). The higher sensitivity of JHM-CAP EIA than to HM-CAP EIA was more remarkable in younger ages of children (under 10 years of age, 90.9 to 57.6%; under 5 years of age, 100 to 57.1%). On the other hand, the specificities of the two EIAs were similar and higher than 93%. These results indicated that JHM-CAP EIA, but not HM-CAP EIA, is highly accurate and suitable for serodiagnosis of H. pylori infection in young Japanese children.

TABLE 1.

Performance of JHM-CAP and HM-CAP EIA on Japanese children in reference to HpSA test

Group and parameter	JHM-CAP EIA	HM-CAP EIA
Overall
% Sensitivitya	93.0 (40/43)	67.4 (29/43)
% Specificityb	95.4 (84/88)	93.2 (82/88)
% Accuracy	94.7	84.7
Under 10 years of age
% Sensitivity	90.9 (30/33)	57.6 (19/33)
% Specificity	97.3 (72/74)	94.6 (70/74)
% Accuracy	95.3	83.1
Under 5 years of age
% Sensitivity	100 (14/14)	57.1 (8/14)
% Specificity	98.0 (51/52)	94.0 (49/52)
% Accuracy	98.5	86.4

^aThe numbers in parentheses for sensitivity represent the numbers of positive children per number of HpSA-positive children.

^bThe numbers in parentheses for specificity represent the numbers of negative children per number of HpSA-negative children.

To analyze such a big difference between the two tests in the sensitivity, EVs of JHM-CAP EIA of individual children were plotted against those of HM-CAP EIA (Fig. 1). Although a good correlation was observed (correlation coefficient, r = 0.87, P < 0.001), the mean EV of HpSA-positive children of JHM-CAP EIA, 4.3, was higher than that of HM-CAP EIA, 3.3, and the difference was statistically significant (P < 0.001). Fourteen HpSA-positive children showing HM-CAP EIA- negative/equivocal results were exclusively under the age of 10 years.

FIG. 1.

Relationship of EVs between JHM-CAP and HM-CAP EIA of HpSA-positive (○) and -negative children (•). Dotted areas represent equivocal EVs, and the horizontal and vertical dotted lines represent the mean EVs of JHM-CAP and HM-CAP EIA, respectively, of 43 HpSA-positive children.

In Table 2 the results of 14 HpSA-positive but HM-CAP EIA-negative/equivocal cases are summarized. The infection status of most of these children was confirmed by UBT. Of 14 cases, 11 were JHM-CAP EIA positive, including 2 cases that had EVs of more than the average.

TABLE 2.

Cases of children showing false-negative or equivocal results with JHM-CAP and/or HM-CAP

Case	Age (yr)	Sexa	EVb		UBT (‰)
			JHM-CAP EIA	HM-CAP EIA
1	4	M	5.9	1.0	NDc
2	3	F	4.6	1.2	ND
3	9	F	4.2	1.8	54.1
4	3	F	3.7	1.7	40.0
5	2	F	3.6	1.6	50.5
6	3	M	3.5	1.7	21.7
7	8	F	3.4	2.1	28.4
8	5	F	3.0	1.7	31.7
9	8	F	3.0	1.1	29.7
10	2	F	2.8	1.5	ND
11	6	F	2.7	1.3	50.5
12	9	F	2.1	1.3	43.8
13	5	F	1.6	1.3	ND
14	5	M	1.5	1.7	28.0

^aM, male; F, female.

^bEVs of ≥2.3, >2.3 to ≥1.8, and <1.8 were judged to be positive, equivocal, and negative, respectively.

^cND, not determined.

The sera of case 3, case 7, and case 8 in Table 2 were further analyzed by a commercial immunoblot test, Helico Blot 2.1. The immunoblot bands of 116 kDa (CagA), 65 kDa (UreB), and 30 kDa (UreA), together with the current infection marker (23), were observed in all cases. Additional band of 89 kDa (VacA) was detected in case 3. Therefore, these cases were all positive by UBT, HpSA test, Helico Blot 2.1, and JHM-CAP EIA but were negative (case 8) or equivocal (case 3 and case 7) by HM-CAP EIA.

The antigens used to develop JHM-CAP EIA and HM-CAP EIA were purified by exactly the same procedure but from different strains. To determine whether the difference in sensitivity of JHM-CAP EIA and HM-CAP EIA is due to the strain difference, the antigen preparations utilized for developing each EIA test were analyzed by immunoblot. In Fig. 2A typical immunoblot patterns of JHM-CAP-positive/HM-CAP-negative sera (case 1 and case 8 in Table 2) are shown, together with JHM-CAP-positive/HM-CAP-positive (case 23) and JHM- CAP-negative/HM-CAP-negative sera (case 26). To our surprise, JHM-CAP-positive sera, irrespective of HM-CAP-positive or -negative reaction, showed a distinct a 100-kDa band with the JHM-CAP preparation that was absent from the HM-CAP immunoblot. In addition, the intensity of the 100-kDa band was strongest among bands detected in each JHM-CAP immunoblot. With case 23 serum, bands other than the 100-kDa band in JHM-CAP and HM-CAP lanes were similar withapproximate molecular sizes of 130, 65, 60, 30, and 25kDa. No band was observed with case 26 serum, the negative control. The immunoblot patterns of 14 cases listed in Table 2 were essentially the same as case 8 and case 1 showing the predominant 100-kDa band with JHM-CAP but not with HM-CAP preparation. SDS-PAGE of a large amount of JHM-CAP and HM-CAP preparations showed 65-, 60-, and 30-kDa protein bands representing possibly UreB, Hsp60, and UreA, respectively, without any 100-kDa band (Fig. 2B). It should be noted that the 100-kDa band detected by the JHM-CAP preparation was absent from the immunoblot with Helico Blot 2.1 that utilizes a western strain ATCC 49503 as the antigen (23).

FIG. 2.

(A) Immunoblotting of JHM-CAP-positive/HM-CAP-negative sera (cases 8 and 1), positive control (case 23), and negative control sera (case 26) with JHM-CAP antigen (JHM) and HM-CAP antigen preparations (HM). JHM-CAP (0.36 μg) and HM-CAP (0.56 μg) antigens were separated by SDS-PAGE, followed by immunoblotting. EVs of JHM-CAP and HM-CAP EIA of sample sera: case 8, 3.0 and 1.7; case 1, 5.9 and 1.0; case 23, 7.4 and 5.3; case 26, 0.6 and 0.3, respectively. M, molecular size markers (in kilodaltons). (B) SDS-PAGE of JHM-CAP (2.9 μg) and HM-CAP (2.24 μg) antigen preparations. M, molecular size markers (in kilodaltons).

Based on these results, we concluded that a strain-specific 100-kDa antigen is present in the JHM-CAP preparation that is highly reactive to sera of Japanese children tested. In addition, absence of the 100-kDa antigen in the HM-CAP preparation could explain false-negative results of HM-CAP EIA on JHM-CAP-positive/HpSA-positive sera of young Japanese children.

DISCUSSION

For diagnosis of H. pylori infection in children, an HpSA test based on detection of H. pylori antigens in stool samples with polyclonal antibodies is a reliable method comparable to UBT and is less expensive and easy to perform; thus, it is suitable for epidemiological studies (4, 8, 14, 17, 21). In the present study, we used HpSA test as the gold standard and carried out serodiagnosis of Japanese children without upper abdominal symptoms with a newly developed JHM-CAP EIA, and the results were compared to those with HM-CAP EIA. A remarkable sensitivity and accuracy was obtained with JHM-CAP EIA in children even younger than 10 years of age, indicating that JHM-CAP EIA is suitable for serodiagnosis of H. pylori infection in Japanese children. In addition, the assay is applicable for epidemiological studies, since these children were asymptomatic and may have lower immune responses than symptomatic children. Furthermore, the results in the present study suggested the importance of future studies on the performance of JHM-CAP and HM-CAP in pediatric populations of the United States and other countries.

The procedure for antigen preparations was exactly the same in both JHM-CAP EIA and HM-CAP EIA, suggesting that the difference in the performance could be attributed to the strain variation, JHM-CAP from Japanese strains and HM-CAP from a U.S. strain. The cells of strains were solubilized with 1% n-octyl alcohol and purified by gel filtration with an agarose A-5m column that has the exclusion limit of proteins with the molecular mass of 400 to 700 kDa (6). The major protein in the preparation was suggested to be urease consisting six each of UreA (30 kDa) and UreB (65 kDa) subunits and the associated heat shock protein Hsp60 (60 kDa) (13). Some other proteins, such as CagA (110 to 130 kDa) and VacA (80to 90 kDa) might also be included that could be variable in different populations and individuals (1). In fact, several weak bands, corresponding possibly to CagA, UreB, Hsp60, and UreA, were detected by both of the preparations, as well as in Helico Blot 2.1. Thus, the new 100-kDa band of the JHM-CAP preparation detected by sera of Japanese children is a strain-specific antigen.

When large amounts of antigens of JHM-CAP and HM-CAP were electrophoresed and stained by Coomassie blue, no band at 100 kDa could be seen, suggesting that the 100-kDa antigen might be a highly immunogenic protein or a cell membrane-associating complex molecule with a small amount of proteins. The possibility of VacA to be the 100-kDa antigen seems unlikely, since Helico Blot 2.1 showed a weak VacA band in only one among three samples tested, and no distinct band at around 100-kDa was detected. It remains to be determined whether the sera of U.S. children could recognize the 100-kDa antigen. Further studies on identification of its molecular properties are now in progress in our laboratory.

In our preliminary study, a Japanese adult serum could also recognize the 100-kDa antigen in the JHM-CAP preparation. However, the immunoblot pattern was different from that of children; the 100-kDa band was less prominent than those of UreB and UreA (unpublished results). These results suggested that the immunogenic potential of the 100-kDa antigen is higher in early life, whereas that of urease is low but increases with age. A number of protective antigens, including urease, catalase, CagA, and VacA, have been proposed as vaccine candidates. Although these have been shown to be effective in animal models, the efficacy in humans is not satisfactory (26). The 100-kDa antigen presented here might serve as a protective antigen and thus a new vaccine candidate in Japan as well as in East Asian countries where gastric cancer is prevailing.

In conclusion, an EIA test based on Japanese strain-derived antigens, JHM-CAP EIA, demonstrated remarkably better performance than an EIA test based on U.S. strain-derived antigens, HM-CAP EIA, on Japanese asymptomatic young children. The higher sensitivity of JHM-CAP EIA is due to the presence of a strain-specific 100-kDa antigen in Japanese strains that may be recognized by the host immune system at an early stage of infection.