Immunohistochemical Differentiation between Western and East Asian Types of CagA-Positive Helicobacter pylori in Gastric Biopsy Samples

Background Cag A-positive Helicobacter pylori isolated from human gastric mucosa is categorized as a Western or East Asian allele-type based on whether the cagA gene encodes an EPIYA-C or EPIYA-D motif. We aimed to differentiate between the 2 types of H. pylori by immunohistochemistry (IHC) using formalin-fixed paraffin-embedded (FFPE) gastric biopsy samples. Materials and Methods We developed 2 monoclonal antibodies (mAbs) that detect either the EPIYA-C or EPIYA-D motif of the H. pylori CagA protein by IHC using FFPE tissues. FFPE tissue sections from 30 Japanese and 39 Brazilian gastric biopsy samples with H. pylori infection confirmed by Giemsa staining (moderate/severe in the Sydney classification system) were examined by IHC with the novel mAbs followed by polymerase chain reaction (PCR) for EPIYA-C or EPIYA-D using DNA extracted from adjacent tissue sections. Results Differentiation among Western and East Asian types and CagA-negative H. pylori was successful in most (97%) samples by IHC with the novel mAbs and commercially available mAbs that react with a species-specific lipopolysaccharide or a common CagA motif of H. pylori. The detection status of EPIYA-C/D motifs by IHC with the novel mAbs was consistent with the PCR results in 61 (88%) of 69 samples: EPIYA-C(+)/D(−) in zero Japanese and 26 Brazilian samples, EPIYA-C(−)/D(+) in 26 Japanese and 1 Brazilian sample, and EPIYA-C(−)/D(−) in 1 Japanese and 7 Brazilian samples. The detection sensitivity and specificity of IHC with each novel mAb compared with the PCR results were, respectively, 84% and 97% for EPIYA-C, and 97% and 95% for EPIYA-D. Conclusions The novel mAbs specific to each EPIYA-C or EPIYA-D motif differentiated between Western and East Asian types of CagA-positive H. pylori by IHC using FFPE tissues. Applying these novel mAbs to large numbers of archived pathology samples will contribute to elucidating the association of these allele types with gastric cancer.


Introduction
Helicobacter pylori infection is critically involved in the development of gastric cancer. [1] Although H. pylori infect almost half of the population worldwide, the incidence of gastric cancer does not necessarily coincide with the rate of H. pylori infection in individual countries. [2] One possible reason for this is the existence of pathogenic gene polymorphisms in H. pylori and their geographic diferences [3].
Te most important virulence factor of H. pylori is cytotoxin-associated gene A (CagA), and the gene encoding the CagA protein is located within the cag pathogenicity island (cag PAI). In CagA-positive or -negative strains of H. pylori, the cag PAI is either present or absent, respectively. [4] EPIYA motifs, characterized by the amino acid sequence Glu-Pro-lle-Tyr-Ala (EPIYA), locate at the C-terminal region of CagA protein. Te 4 diferent types of EPIYA motifs, EPIYA-A, -B, -C, and -D, are determined by the surrounding amino acid sequence. Te EPIYA-A and EPIYA-B motifs are representative polymorphisms shared by almost all H. pylori isolates, followed by the EPIYA-C or EPIYA-D motifs. [5] H. pylori strains with the EPIYA-C motif are referred to as the Western type and those with the EPIYA-D motif are referred to as the East Asian type according to their regional prevalence. [6] Compared with the EPIYA-C motif, the EPIYA-D motif binds more strongly to host SHP-2, an ubiquitously expressed protein tyrosine phosphatase recognized as a proto-oncogenic phosphatase [5,7,8]. Clinical observations suggest that infection with H. pylori carrying East Asian type CagA is associated with severe gastric atrophy and gastric cancer [9,10].
Te CagA status and genotype are mainly determined by polymerase chain reaction (PCR) and sequencing of the EPIYA region using DNA extracted from H. pylori isolated from fresh gastric biopsy samples. To gain more information on the CagA status in clinical and histopathologic studies of gastric cancer, we aimed to diferentiate between the Western and East Asian types of CagA-positive H. pylori by immunohistochemistry (IHC) using formalinfxed parafn-embedded (FFPE) gastric biopsy samples. Toward this aim, we developed 2 monoclonal antibodies (mAbs) that detect either the EPIYA-C or EPIYA-D motif of the H. pylori CagA protein in FFPE tissue sections, and the sensitivity and specifcity of IHC with the novel mAbs were compared with the PCR detection results for EPIYA-C or EPIYA-D using DNA extracted from adjacent FFPE tissue sections.

Tissue Samples.
Te FFPE tissue blocks of gastric biopsy samples obtained from 30 Japanese and 39 Brazilian patients infected with H. pylori (confrmed by Giemsa staining as moderate/severe in the Sydney classifcation system [11]) were collected from the Tokyo Medical and Dental University Hospital pathology archives in Japan, and the Sao Paulo University Hospital and Santa Cruz Hospital in Brazil. All patients underwent an endoscopic examination for dyspepsia or gastric cancer screening during the years 2020 and 2021, which revealed no evidence of malignancy endoscopically or pathologically. At least 1 tissue with H. pylori confrmed by Giemsa staining was included in the Japanese biopsy samples from the gastric antrum and/or corpus (diferent number of tissues) and in the Brazilian biopsy samples from the gastric antrum (3 tissues) and corpus (2 tissues). Serial tissue sections were cut from each FFPE tissue block of the gastric biopsy samples and subjected to IHC and PCR analysis. Te study protocol received approval from the ethics committees of the Tokyo Medical and Dental University Hospital (reference M2018-244 and M2019-072) and the Sao Paulo University Hospital (reference HUUSP CEP76 CAAE 52744016.7.0000.0076).

mAb Production.
Novel mAbs specifc to either the EPIYA-C or EPIYA-D motif of H. pylori CagA protein were developed to determine the type (Western or East Asian) of H. pylori in FFPE gastric biopsy samples. Te mAbs were generated according to a published laboratory protocol [12] with some modifcations. BALB/c mice (CLEA Japan, Tokyo, Japan) were immunized with keyhole limpet hemocyanin-conjugated peptide synthesized by Eurofns Genomics (Ebersberg, Germany) from the whole sequence of EPIYA-C (N′-GFPLKRHDKVDDLSKVGRSVSPEPI YATIDDLGG-C′) or EPIYA-D (N′-AINRKIDRINKIA SAGKGVGGFSGAGRSASPEPIYATIDFDEANQAG-C′). Hybridoma cell lines producing antibodies for each peptide were evaluated by enzyme-linked immunosorbent assay (ELISA) with the keyhole limpet hemocyanin-free synthetic peptide. Hybridomas with positive results were screened by manual IHC as described previously, [13] using FFPE gastric biopsy samples that were preliminarily confrmed by PCR to be infected with Western or East Asian type H. pylori. Finally, we selected the hybridoma that produced an antibody generating a strong reaction specifc to each type of H. pylori on the gastric mucosa and cloned it by 2 rounds of limiting dilution. We implanted a single hybridoma clone in the intraperitoneal space in severe-combined immunodefciency mice (CLEA Japan). Ascites was collected 1 or 2 weeks later and used without purifcation as an undiluted antibody. Te antibodies obtained by immunization of either the EPIYA-C or EPIYA-D peptide antigen are referred to as EPIYA-C mAb (IgG2b,κ) or EPIYA-D mAb (IgG2b,κ), respectively, in the present study. Te Institutional Animal Care and Use Committee of Tokyo Medical and Dental University approved all of the animal experiments (approval number: A2021-097A and approval date: 01/04/2021).

H. pylori Samples.
Te specifcity of the novel EPIYA-C and EPIYA-D mAbs or the commercially available CagA mAb (sc-28368, Santa Cruz Biotechnology, CA, USA) specifc to a common CagA motif of H. pylori was examined by ELISA and Western blotting with a standard strain of Western type H. pylori (ATCC 43504), [14] an isogenic cagA-knockout H. pylori mutant (ATCC 43504 ΔcagA) [14], and a clinical isolate of East Asian type H. pylori (HpLN-4) [13]. Te H. pylori strains were cultured on sheep blood agar plates (BBL 252792, Tokyo, Japan) at 37°C in a jar with an AnaeroPack (Mitsubishi Gas Chemical Co., Inc., Tokyo, Japan) for 72 h. Whole cell lysates were prepared by sonicating the cultures on ice using an ultrasonic homogenizer (VP-5S ULTRA S homogenizer; Taitec, Koshigaya, Japan) at level 6 for 3 min with each bacterium in phosphate-bufered saline (PBS). Te whole cell lysate was centrifuged at 12,000 rpm for 10 min, and the supernatant was collected and used for ELISA and Western blotting. A BCA protein assay kit (23225; Pierce, Rockford, IL) was used to quantify the protein in the supernatant.

ELISA.
For the ELISA, fat-bottomed 96-well NUNCimmuno plates (Nalge Nunc International, Roskilde, Denmark) were coated with whole cell lysates of each of the 3 H. pylori strains (5 µg per well) in carbonate-bicarbonate bufer (pH 9.6) for 90 min at 37°C. Te mAbs were serially diluted in PBS containing 0.25% Tween-20 (T-PBS) and added to each well, and then, the plates were incubated for 90 min at 37°C. Tey were incubated further for 30 min with biotinylated rabbit anti-mouse immunoglobulins (E0354, DAKO, Glostrup, Denmark), followed by incubation for 30 min with horseradish peroxidase-conjugated streptavidin (P0397, DAKO), both at room temperature. Te plates were washed with T-PBS both before and after each step. After the reaction, citrate phosphate bufer (pH 5.4) containing 0.3% o-phenylenediamine dihydrochloride (MilliporeSigma Co.) and 0.012% H 2 O 2 was added to each well, and the plates were incubated in the dark for 15 min at room temperature. Te reaction was stopped by adding 25 µl of 2N HCl to each well. Te plates were read at 490 nm on an ELx 808 TM Absorbance Microplate Reader (BioTek, Tokyo, Japan).

Western
Blotting. Each whole cell lysate of the 3 H. pylori strains (12 μg per lane) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using 8% polyacrylamide gel. Te samples were electrophoretically transferred to polyvinylidene difuoride membranes using a Mini Trans-Blot cell (Bio-Rad, Tokyo, Japan). Block Ace (DS Pharma Biomedical Co. Ltd., Osaka, Japan) was used to block the membranes overnight at 4°C and the membranes were then incubated with diluted antibodies (EPIYA-C mAb; 1 : 500, EPIYA-D mAb; 1 : 500, and CagA mAb; 1 : 500) at room temperature for 90 min. Te membranes were incubated in the dark for 60 min with ECL Plex ™ goatα-mouse IgG-Cy3 (GE Healthcare, Buckinghamshire, UK) at a 1 : 1000 dilution ratio. Te membranes were washed in T-PBS before and after each of the last 2 steps. Finally, the membranes were washed in PBS before being dried in the dark for 1 h at 37°C. A Bio-Rad BPC-500, Molecular Imager (Quantity One-Analysis Software Version 4.6.9, Bio-Rad) was used to image the dried membranes.
2.6. IHC. All samples were examined by IHC with the novel EPIYA-C and EPIYA-D mAbs and the commercially available CagA mAb and H. pylori mAb (D369-3, MBL, Nagoya, Japan) specifc to H. pylori lipopolysaccharide (LPS). Serial tissue sections (4 μm thick) mounted on silanecoated slides (Muto Pure Chemicals Co. Ltd., Tokyo, Japan) were subjected to automated IHC with the Leica BOND-III (Leica Microsystems Inc., Tokyo, Japan) using a BOND Polymer Refne Detection kit (#DS9800, Leica Microsystems Inc.). Deparafnization, peroxidase inhibition, antigen retrieval, incubation with primary antibody at room temperature for 15 min, and counterstaining were performed according to the manufacturer's protocol. Te best protocol for antigen retrieval was determined, as shown in Table 1. Appropriate dilution in Primary Antibody Diluent (10-0001RUO, Sakura Finetek Japan Co., Ltd., Tokyo, Japan) was determined for the H. pylori mAb; 1 : 1000, CagA mAb; 1 : 100, EPIYA-C mAb; 1 : 2000, and EPIYA-D mAb; 1 : 400. Te IHC results were considered to be positive when at least 1 cluster of unequivocal dot-like signals was observed on the surface of the gastric mucosa in the biopsy sample.

PCR.
A tissue section (10 μm thick) from each biopsy sample was placed in a sterilized 1.5-ml centrifuge tube. Te DNA of each section was extracted with TaKaRa DEXPAT ™ (9091, Takara Shuzo, Shiga, Japan) and further purifed by Ethachinmate (318-01793, Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions. Real-time PCR was performed to amplify fragments of 16S ribosomal RNA, the cagA gene, and EPIYA-C and EPIYA-D, using primers and probes (supplementary Table 1) that were originally designed for droplet digital PCR for quantitative detection and genotyping of H. pylori from stool. [15,16] Real-time PCR mixtures contained 5 μl of template DNA, 100-nM concentrations of each primer, a 40-nM concentration of the probe, and 2× SensiFAST Probe Hi-ROX Mix (BIO-82020, BIOLINE, Cincinnati, USA) in a total volume of 50 μl. Amplifcation and detection were both performed using the ABI PRIZM 7900HT Sequence Detection System (Applied Biosystems, CA, USA) with the following program: 95°C for 5 min and then 50 cycles of 95°C for 15 s and 60°C for 1 min. Samples of serially diluted bacterial DNA were used as an internal standard to determine the amount of bacterial DNA in the sample used for PCR. Te amount of bacterial DNA was expressed in terms of the number of bacterial genomes using 1.25 × 10 10 Da per genome for the calculation [17]. Every PCR included negative controls without bacterial DNA. Positive samples were defned as those in which any bacterial genome was detected (supplementary Figure 1).

Statistical Analysis.
Associations between the detection status of EPIYA-C or EPIYA-D by IHC and PCR were evaluated using Fisher's exact test. Unweighted Cohen's kappa coefcient was used to assess variability in the detection status of the EPIYA C/D motif by IHC and PCR. In the analysis, p values < 0.05 were considered statistically significant. All analyses were carried out with the statistical package R (version 3.6.3; available from http://www.r-project.org).

Specifcity of the Novel mAbs.
Te novel EPIYA-C and EPIYA-D mAbs showed specifc reactivity to each corresponding peptide without any cross-reactivity between them. In the ELISA and the Western blotting analysis using the sonicated H. pylori whole cell lysate, the EPIYA-C mAb reacted with the Western type H. pylori (ATCC 43504) but not with the East Asian type H. pylori (HpLN-4) or the cagAknockout H. pylori (ATCC 43504 ΔcagA), whereas the EPIYA-D mAb reacted with the East Asian type H. pylori but EPIYA-C mAb EPIYA-D mAb CagA mAb   not with the Western type H. pylori or the cagA-knockout H. pylori (Figure 1). Te specifcity of the CagA mAb was unclear in the ELISA but was confrmed by Western blotting, showing a sharp band corresponding to the Western type or East Asian type CagA protein detected by the EPIYA-C or EPIYA-D mAb, respectively.  (Figure 3(c)). Te CagA status was undetermined in the 2 (3%) samples that were positive for the H. pylori and CagA mAbs but negative for the EPIYA-C or EPIYA-D mAbs. Te PCR detection status of both the EPIYA-C and EPIYA-D motifs was consistent with the IHC detection status in 61 (88%) samples (Table 2) (Table 3). Compared with the PCR results, the IHC detection sensitivity and specifcity were, respectively, 84% and 97% for EPIYA-C and 97% and 95% for EPIYA-D. Te 2 samples with undetermined results by IHC were both positive by PCR for cagA; 1 from a Japanese patient was negative by PCR for both EPIYA-C and EPIYA-D and the other from a Brazilian patient was negative by PCR for EPIYA-D, but positive by PCR for EPIYA-C.

Discussion
For the present study, we developed novel mAbs that can detect either the EPIYA-C or EPIYA-D motif of CagApositive H. pylori locating on the surface of FFPE gastric mucosa. Diferentiation among the Western type, East Asian type, and CagA-negative H. pylori was successful in most (97%) of the gastric biopsy samples by IHC with the novel mAbs and commercially available mAbs that react with a species-specifc LPS or a common CagA motif of H. pylori. Te detection sensitivity and specifcity of IHC with each novel mAb compared with the PCR results were, respectively, 84% and 97% for EPIYA-C, and 97% and 95% for EPIYA-D.
Validation of the H. pylori CagA typing by IHC was reported in 2020 by a research group from Oita University [18] who examined FFPE gastric biopsy samples by manual IHC with polyclonal antibodies (pAbs) raised against H. pylori (B0471, DAKO), CagA (b-300) (sc-25766, Santa Cruz, no longer available), or East Asian type CagA peptide (developed in their laboratory [19]). In their study, the pAb raised against East Asian type CagA peptide (N′-NAINR-KIDRINKIASAGKG-C′) had markedly low sensitivity (23.9%) compared with the cagA sequence data of isolated H. pylori and was considered inappropriate for determining the CagA status. In the present study, we obtained novel mAbs for determining the CagA status by selecting and cloning the best hybridoma cell lines that produced a specifc and strong reaction to the Western type or East Asian type of H. pylori on the gastric mucosa by IHC using FFPE tissue sections. Furthermore, an automated IHC method was used for the sample analysis to introduce the novel mAbs as a routine examination in pathology laboratories. Finally, unequivocal dot-like signals on the gastric mucosa with minimal background staining were obtained for Western or East Asian type H. pylori by the Leica system of IHC with the EPIYA-C or EPIYA-D mAbs, respectively.
Te commercially available H. pylori mAb (D369-3, MBL) that reacts with a species-specifc LPS was originally produced in our laboratory (clone: TMDU-D8) by immunizing mice with sonicated whole bacterial lysate of the Western type H. pylori followed by screening hybridoma cell lines by IHC using FFPE tissue sections of the H. pyloriinfected stomach resected from a Japanese gastric cancer patient. [13] Te H. pylori mAb was confrmed to be highly specifc for H. pylori species by ELISA and Western blotting, and the IHC detection sensitivities for H. pylori in the gastric mucosa and gastric lymph nodes were higher than those of the commercially available anti-H. pylori pAb (DAKO). [13] In the present study, the H. pylori mAb strongly reacted with H. pylori in all gastric biopsy samples from Japanese and Brazilian patients with H. pylori infection confrmed by Giemsa staining, indicating that H. pylori expresses high levels of species-specifc LPS within the cell wall outer membrane structure regardless of the CagA status. Tis antibody specifc to H. pylori LPS was useful for detection of both CagA-negative and CagA-positive H. pylori.
Te commercially available CagA mAb (Santa Cruz sc-28368) was raised against a common CagA motif comprising Canadian Journal of Gastroenterology and Hepatology amino acids 1-300 of CagA of H. pylori origin, according to the manufacturer's information. In the Western blotting analysis performed in the study, the CagA mAb reacted with both the Western type and East Asian type H. pylori, whereas the novel EPIYA-C or EPIYA-D mAb specifcally reacted with either the Western type or East Asian type H. pylori, respectively. Te detection status by IHC with the CagA mAb was almost consistent with the results by IHC with the novel EPIYA-C/D mAbs; IHC with the CagA mAb was positive in all 57 samples with either EPIYA-C or EPIYA-D detected, and negative in the 10 samples with both EPIYA-C and EPIYA-D undetected. According to the algorithm for determining the CagA status, we could diferentiate among Western type, East Asian type, and CagA-negative H. pylori in 67 (97%) of the 69 gastric biopsy samples. Te 3 Brazilian samples with EPIYA-D detected by IHC were all from patients of Japanese descent, and the 1 Japanese sample with EPIYA-C detected by IHC was from a foreign resident in Japan (private communications). Although it is not clear why the 2 (3%) samples were positive by the CagA mAb but negative by both the EPIYA-C/D mAbs, the possibility of a CagA-positive H. pylori without an EPIYA-C/D motif, such as a unique CagA type called ABB-type CagA, [18] cannot be ruled out, at least in the case of the Japanese sample with consistent results obtained by IHC and PCR.
To estimate the accuracy of H. pylori CagA typing by IHC with the novel mAbs, we performed PCR typing for the EPIYA-C/D motif using total tissue DNA extracted from FFPE gastric biopsy samples because H. pylori isolates from fresh biopsy samples were not obtained from these patients. While the detection status of the EPIYA-C/D motif was consistent between the results by IHC and PCR in many (88%) samples, we suspect that the inconsistent results in the 8 (12%) samples were likely due to the use of DNA extracted from FFPE tissue sections for the PCR. Formalin fxation usually causes DNA fragmentation, cytosine deamination, and cross-linking [20] and decreases PCR sensitivity. [21] Contamination by H. pylori carrying the prevalent type CagA into the tissue from other samples with H. pylori infection may occur during tissue processing for parafnembedding in identical chambers of the machine used in pathology laboratories. [22][23][24] [25][26][27] and Okinawa, Japan [10,28,29].
In conclusion, the novel mAbs specifc to each EPIYA-C or EPIYA-D motif could diferentiate between Western  (2), the EPIYA-C mAb (3), and the EPIYA-D mAb (4) are shown pairwise for each identical gastric biopsy sample with the Western type (a), East Asian type (b), or CagA-negative (c) H. pylori infection confrmed by PCR afterward. Te novel EPIYA-C and EPIYA-D mAbs showed a prominent contrast in the presence (a3/b4) or absence (a4/b3) of signals on the surface of gastric foveolar epithelium in the CagA-positive (a2/b2) samples with H. pylori infection (a1/b1). A sample with many H. pylori (c1) and minimum infammation was CagA negative (c2) with no reactivity by the novel EPIYA-C/D mAbs (c3/c4). Bars: 50 µm.  and East Asian types of CagA-positive H. pylori by automated IHC using FFPE tissue sections. Application of these novel mAbs to large numbers of samples accumulated in pathology archives will contribute to elucidating the potential associations of these H. pylori allele types with gastric cancer.