Helicobacter pylori in the Indonesian Malay’s descendants might be imported from other ethnicities

Background Even though the incidence of H. pylori infection among Malays in the Malay Peninsula is low, we observed a high H. pylori prevalence in Sumatra, which is the main residence of Indonesian Malays. H. pylori prevalence among Indonesian Malay descendants was investigated. Results Using a combination of five tests, 232 recruited participants were tested for H- pylori and participants were considered positive if at least one test positive. The results showed that the overall H. pylori prevalence was 17.2%. Participants were then categorized into Malay (Aceh, Malay, and Minang), Java (Javanese and Sundanese), Nias, and Bataknese groups. The prevalence of H. pylori was very low among the Malay group (2.8%) and no H. pylori was observed among the Aceh. Similarly, no H. pylori was observed among the Java group. However, the prevalence of H. pylori was high among the Bataknese (52.2%) and moderate among the Nias (6.1%). Multilocus sequence typing showed that H. pylori in Indonesian Malays classified as hpEastAsia with a subpopulation of hspMaori, suggesting that the isolated H. pylori were not a specific Malays H. pylori. Conclusions Even though the ethnic groups live together as a community, we observed an extremely low H. pylori infection rate among Indonesian Malay descendants with no specific Indonesian Malay H. pylori. The results suggest that H. pylori was not originally among these groups and H. pylori was imported from other ethnic groups.

certain ethnic groups, despite having the same environmental exposure as other ethnic groups [3].
Malays, members of the Austronesian family, are an ethnic group who speak the Malayo-Polynesian language [4,5]. Malays predominantly inhabit the South-East Asia region, especially the Malay Peninsula, east coast of Sumatra, and the coast of Borneo. According to the "Taiwan" theory, the Malays originated from Taiwan and migrated to the Malay Peninsula through the Philippines and Borneo approximately 1,500 years ago, although they might have simultaneously traveled alongside people originating from Yunnan, China [6,7]. After reaching the Malay Peninsula, the Malays began to spread to Indonesia (predominantly Sumatra), several areas of Borneo, and the western tip of Java. Although Sumatra is predominantly occupied by Indonesian Malays, several ethnicities are classified as Proto-Melayu, including the Bataknese and Nias ethnic groups, which are considered older ancestors of the modern Indonesian Malays [8].
Other ethnic groups also reside in Indonesia, including the Javanese and Sundanese. The Javanese reside mainly on Java Island. Javanese and Sundanese have very similar cultures, languages, and cuisines. Importantly, the Sundanese reside almost exclusively in the western part of Java. The origin, history, and language of Peninsular Malays, Javanese, and Sundanese are very similar. In addition to the Melayu-Minang and Melayu-Bugis, who are sub-ethnic Malays in Malaysia, there is a sub-ethnic group named Melayu-Jawa, who have a close genetic relationship to the Indonesian population, including the Javanese [9]. These findings indicate that Javanese and Sundanese might have a common ancestral and cultural history with Peninsular Malays [4]. Therefore, Javanese and Sundanese are considered Indonesian Malay ethnic descendants.
To date, native inhabitants of several areas of Indonesia are still categorized as part of the Malay ethnic group, although they are divided into many sub-ethnic groups. This was reinforced by a population census conducted by the Dutch colonial government in 1930. The 1930 population census (volkstelling) used anthropological studies, language approaches, geography, history, and ethnography to determine ethnic groups; thus, many researchers use the 1930 census data by the Dutch government as a reference for tribal composition in Indonesia [10]. In the 1930 census, populations inhabiting Sumatra island, parts of Borneo, and the western tip of Java were still categorized as ethnic Malay. These ethnicities are still used with some extensions.
The H. pylori infection rate among Malays in Malaysia was only 19.6% and was significantly lower than the Chinese and Indian populations [11]. In contrast, the Javanese had a very low H. pylori infection rate of only 2.4% [12]. In addition, our previous data on the five largest islands in Indonesia showed a high H. pylori prevalence in Sumatra, the main residence of Indonesian Malays. Since there is a close relationship between Peninsular Malays and Indonesian Malays, we hypothesized that the H. pylori infection rate among Indonesian Malay descendants would be lower than the prevalence in Malaysia. Herein, we examined H. pylori prevalence among Indonesian Malay ethnic descendants.

Sample demographic characteristics
A total of 232 samples from 126 males and 106 females were analyzed. Participants were recruited from Banda Aceh (n = 38), Medan (n = 22), Dolok Sanggul (n = 47), Padang (n = 33), and Palembang (n = 38) on Sumatra island, Gunungsitoli (n = 32) on Nias Island, and Cimacan (n = 22) on Java Island. The mean age of the participants was 45.54 ± 14.64 and ranged from 17-83 years old. According to the ethnicity, 37 (15.9%) participants were Aceh, 67 (28.8%) participants were Bataknese, 4 (1.7%) participants were Javanese, 36 (15.5%) participants were Malay, 33 (14.2%) participants were Minang, 33 (14.2%) participants were Nias, and 22 (9.5%) participants were Sundanese. The ethnicity and history suggest that the Bataknese and Nias are older ethnicties than the current Malay ethnic group and are considered part of the Proto-Malay people [8]. Javanese are considered descendants of the Malays as part of the Austronesian expansion. Thus, we separated the dataset into an ethnic Malay group (Aceh, Malay, and Minang), an ethnic Java group (Javanese and Sundanese), ethnic Bataknese, and ethnic Nias. After separation into four groups, no significant differences in age and sex were detected between the groups (P = 0.322 and P = 0.321, respectively).

No H. pylori observed in the Java ethnic group
No H. pylori was detected in the Java ethnic group, similar to our observation among the Malay ethnic group, using any of the five different tests. The occurrence of H. pylori was 0/26 (0.00%) using RUT, IHC, histology, ELISA, or culture tests ( Table 3).

Origin of H. pylori in Indonesian Malays
Among all Indonesian Malay participants, we could only detect H. pylori in one participant and we could only isolate one sample. We performed a population genetic analysis of H. pylori from only one sample isolated from an Indonesian Malay originating from Padang city. Using no-admixture model of STRU CTU RE analysis, the H. pylori sample belonged to the hpEastAsia population (Fig. 1A). Subsequent analysis comparing only the hpEastAsia population showed that the sample belonged to the hspMaori subpopulation ( Fig. 1B). A phylogenetic tree analysis showed that this strain is located near Indonesian Minahasanese [13] and Melanesians hspMaori [14]. Table 1 Clinical outcome observations PUD, peptic ulcer disease; a , number of patients was only one, the SD value could not be calculated; b , this number was calculated based on at least one test positive consideration

Performance of each diagnostic test when H. pylori prevalence varied
Overall, we observed a good essential agreement (91.9%) between tests with a fair κ-coefficient value of 0.719 (P < 0.001). After we divided the participants into different ethnic groups, we found a high essential agreement of 97.5% and a fair κ-coefficient value (0.293, P < 0.001) among the Malay ethnic group. In addition, we found a good essential agreement of 72.3% and substantial κ-coefficient value 0.669 (P < 0.001) among Bataknese (Fig. 2).
Even though we found good inter-diagnostic method consistency, there was a major difference in invasiveness between the tests. We classified the diagnostic tests as invasive (RUT and culture) or non-invasive (ELISA). Then, we examined the performance of invasive versus non-invasive in at least one positive scenario for each group compared to histology/IHC as the positive group. The invasive test yielded a fair performance with sensitivity and specificity values of 78.9% and 92.5%, respectively, and overall accuracy of 91.4% (Table 5). The non-invasive test performed well, with an overall accuracy of 96.2%, a sensitivity of 83.3%, and a specificity of 97.4%. These results suggest that the non-invasive test was better for diagnostic testing in this population.   [19,20]. As for the pathogen virulence, the virulence may be low in indigenous Malay H. pylori, resulting in easier detachment from the gastric mucosa [21]. There are several H. pylori virulence factors related to colonization ability, including blood group antigen-binding adhesin, SabA, outer inflammatory protein, and H. pylori outer membrane protein Q [22]. The activity of these proteins may be associated with lower colonization in Indonesian Malay H. pylori. In addition, Malays add and consume "pegaga" (Centella asiatica) in their daily food and this consumption is significantly associated with lower H. pylori prevalence [23]. Centella asiatica may increase gastric mucus production [24], leading to lower H. pylori colonization. Consumption of "dadih", a traditional fermented buffalo milk, is common among Indonesian Malays. In addition, Indonesian Malays commonly consume Zanthoxylum acanthopodium, spicy flavored fruits [25]. Both Zanthoxylum acanthopodium and "dadih" have anti-microbial activity against several gram-positive and gram-negative pathogenic bacteria [26,27]. Therefore, consumption of these foods may prevent H. pylori infection.
We also found an extremely low prevalence of H. pylori among the Java ethnic group; in particular, no H. pylori was detected in the Sundanese. Sundanese is a specific ethnic group residing mostly in the western part  of Java. The Sudanese share a similar language, culture, and food with the Javanese, the most predominant ethnicity in Indonesia. This finding supports our previous results showing that the Javanese have a low prevalence of H. pylori [12]. Population genetics indicate that the Indonesian Malay H. pylori belongs to the hpEastAsia population and the hspMaori subpopulation of H. pylori. The Indonesian Malay H. pylori is highly similar to the H. pylori isolated from Manado [13] and Melanesians hspMaori [14] strains. These data suggest that the H. pylori originated from the Austronesian expansion in a recent infection, because the strain retains the genetic characteristics of its parental group. Since there was a similar ancestral history between Sundanese, Javanese, and Malays with a low H. pylori prevalence and no observation of specific Indonesian Malay H. pylori, we hypothesize that there is no separate H. pylori among the Indonesian Malay descendants. The currently reported H. pylori may have been imported following the intraracial spread instead of inter-racial spread, as explained by the "Racial Cohort" hypothesis [11,21]. We observed a high H. pylori prevalence among the Bataknese. The Bataknese descend from Austronesians who originated from Taiwan and the Philippines then migrated to North Sumatra via Java/Borneo. They settled mainly around the Great Toba Lake, which provided fresh water to support their agriculture activity [28]. Our observation confirmed a previous study showing a high prevalence of H. pylori in Bataknese [12,29]. Interestingly, we observed low H. pylori prevalence among the Nias ethnic group, which is believed to have the same roots as the Bataknese. Even Bataknese from Medan tended to have a higher prevalence of H, pylori than people from Dolok Sanggul, Bataknese still commonly use boiled water as their primary drinking water source [12]. The water source is associated with H. pylori infection. This hypothesis was supported by the high H. pylori infection rate in the Japanese, who have drunk well water since before World War II [30,31]. Our risk factor analysis among Indonesians from the five largest islands showed similar results [12]. This interesting distribution suggested that the main source of H. pylori among Bataknese might originate from Great Toba Lake.
We examined the H. pylori infection rate using five different diagnostic modalities. Even though we observed a different infection rate between the five tests, the intertest essential agreement value and substantial Cohen's Kappa value were good, suggesting valid results between tests and interchangeable usage. After dividing tests into invasive and non-invasive methods, we obtained better sensitivity and specificity for non-invasive methods compared with invasive tests, suggesting that the non-invasive methods were a better choice. Indeed, the ELISA is a favorable non-invasive method to detect H. pylori among clinicians. However, the ELISA has a bias toward current rather than past H. pylori infections; hence, infection should be confirmed with another diagnostic method. Other non-invasive methods, such as the stool antigen test (SAT) and 14 C Urea Breath Test (UBT) showed promising performances; however, only a polyclonal antibody with low sensitivity for SAT is available in Indonesia [32]. 14 C-UBT is widely available in Indonesia, but has not been locally validated yet.
Indonesia is widely known to have low H. pylori prevalence, but the number of dyspeptic patients is still high [12,33]. The results of this study demonstrate that older age and infection with H. pylori contribute to worsening gastric mucosal conditions. Interestingly, H. pylori infection also associated with the development and recurrence of Henoch-Schönlein Purpura (HSP) after gastrointestinal manifestations [34]. However, we observed inflamed gastric mucosa in the absence of H. pylori in a previous study [35]. This finding suggests that causes other than H. pylori are responsible for the gastro-duodenal diseases. Future studies elucidating non-H. pylori microbes that responsible for gastro-duodenal diseases are necessary.
There were several limitations to our study. First, the sample number is relatively small, which may lead to weak statistical inferences. In addition, we measured H. pylori infection, which is usually established during childhood. Our study may have examined the factors leading to the continuation of the infected state, rather than the development of infection. We did not obtain the data that explains the low prevalence in the culture and habit among our participants. Therefore, a study to elucidate the cultural and habitual activities related to low H. pylori among Malay is needed.

Conclusion
The prevalence of H. pylori infection among Malays is low, with no H. pylori in some ethnic groups, such as the Aceh. We also observed no H. pylori among the Javanese. Nevertheless, an exceptionally high H. pylori infection rate was observed among the Bataknese and a moderate infection rate was observed among the Nias. Even though the ethnic groups live together as a community, the H. pylori infection rate among Indonesian Malays descendant ethnic groups is extremely low, suggesting that no H. pylori originated among the Malays.

Sampling population and sample collection
We performed consecutive endoscopic surveys between January and February 2016 at 5 locations: 3 locations in Sumatra island, including Dolok Sanggul, Padang, and Palembang; 1 location in Nias Island, Gunungsitoli; and 1 location on Java Island, Cimacan. The mean age of the participants was 46.6 ± 14.5 and consisted of 96 males and 76 females. The ethnic groups were Bataknese (47), Javanese (2), Malay (36), Minang (32), Nias (33), and Sundanese (22). In addition, we analyzed 60 samples from our previous study [36]. These 60 samples originated from participants with a mean age of 41.6; 30 participants were male and 30 were female. The study inclusion criteria were patients who had a chronic dyspeptic symptom but already stopped proton pump inhibitor and antibiotic treatment outside of H. pylori eradication purposes for at least 2 weeks. The exclusion criteria were any history of H. pylori eradication therapy, partial/total gastrectomy, subjects with contraindication for endoscopic examination, and nonfasted patients. In addition, we collected patient demographics and history, including smoking and alcohol drinking habits by interview. Before taking a history and performing the upper endoscopic examination, we acquired written informed consent from all participants. Our current study protocol was approved by the ethics committees of Dr. Soetomo Teaching Hospital (Surabaya, Indonesia), Dr. Cipto Mangunkusumo Teaching Hospital (Jakarta, Indonesia), and Oita University Faculty of Medicine (Yufu, Japan).
We collected four gastric specimens for each patient during the endoscopic procedures. Three specimens were collected from the lesser curvature of the antrum, approximately 3 cm from the pyloric ring, which each of specimen was used for H. pylori culture, RUT, and histopathology examination, respectively. Another gastric specimen was taken from the greater curvature of the corpus, which was used only for histopathology examination. During the examination, the endoscopists also determined the gastric condition visually, including the presence of ulcers, inflammation of the gastric mucosa, and carcinoma. In addition, fasting serum was collected on the day of the endoscopy and stored at − 20 °C for ELISA.

Evaluation for H. pylori infection
We evaluated the H. pylori infection status by a combination of five different methods, including culture, histology, IHC, RUT, and serology.

Culture and rapid urease test
To culture H. pylori, one antral biopsy specimen was homogenized in saline and inoculated onto Mueller Hinton II Agar medium (Becton Dickinson, NJ, USA) supplemented with 7% horse blood without antibiotics. The plates were incubated for up to 10 days at 37 °C under microaerophilic conditions (10% O 2 , 5% CO 2 , and 85% N 2 ). H. pylori were identified based on colony morphology, Gram staining results, and positive reactions for oxidase, catalase, and urease. Isolated strains were stored at -80 °C in Brucella Broth (Difco, NJ, USA) containing 10% dimethyl sulfoxide and 10% horse serum. For the RUT examination, the gastric specimen collected from the antrum was directly inserted into the RUT slide (CLO test, Kimberly-Clark, USA).

Histology and immunohistochemistry examination
Biopsy specimens were stored in 10% buffered formalin then embedded in paraffin blocks. Serial sections were stained with hematoxylin and eosin and May-Giemsa stain. Stained samples were evaluated for H. pylori density using the updated Sydney system. The degree of bacterial density according to the updated Sydney system was: 0, normal; 1, mild; 2, moderate; and 3, marked [37]. Samples with bacterial density ≥ 1 were considered positive for H. pylori.
To increase the accuracy of detecting H. pylori, we also performed immunohistochemical staining. Briefly, tissue sections were incubated with anti-α-H. pylori antibody (DAKO, Glostrup, Denmark, product ID: B0471) overnight at 4 °C. After washing, the sections were incubated with biotinylated goat anti-rabbit IgG (Nichirei Co., Tokyo, Japan, product ID: 426011), followed by incubation with an avidin-conjugated horseradish peroxidase solution (Vectastain Elite ABC Kit; Vector Laboratories Inc., Burlingame, CA, USA). Peroxidase activity was detected using an H 2 O 2 /diaminobenzidine substrate solution [38]. The same experienced pathologist who analyzes for Myanmar, Vietnam, Bhutan, and the Dominican Republic evaluated all the specimens in this study to reduce the examiner bias [39][40][41][42][43].

Serology evaluation
We measured the H. pylori antibody titers with an ELISA kit (Eiken, Co. Ltd., Tokyo, Japan, product ID: 4,987,026,182,711). The manufacturer's recommended cut-off point for determining H. pylori infection was ≥ 10 U/mL. This cut-off point has been validated in the Indonesian population yielding a sensitivity and specificity of 66.7% and 97.2%, respectively [44].
Patients were considered to be negative for H. pylori infection when all five test results were negative, whereas patients with at least one positive test were considered positive for H. pylori infection.

Determination of population genetics
We performed the population genetic analysis using the multilocus sequence typing (MLST) approach. Seven housekeeping genes of H. pylori (atpA, efp, mutY, ppa, trpC, ureI, and yphC) were analyzed resulting in 3406 concatenated sequences. STRU CTU RE version 2.3.4 [45] with no-admixture model algorithm was used against 2544 available MLST data on the pubMLST (https:// pubml st. org/). We used the parameter K = 7, as this number of K has been identified as the generated population genetics of H. pylori [14,46,47]. For the determination of the H. pylori subpopulation, we picked the hpEastAsia population only and ran the subsequent analysis of the STRU CTU RE no-admixture model with K = 3, as previously described [14]. All of these analyses were carried out using Markov-Chain Monte Carlo of 1000,000 iterations with 100,000 burn-in. The phylogenetic tree was constructed using MEGA 7 [48] with Neighbor-Joining Tree [49] and Kimura-2 parameter substitution model [50], as these parameters were commonly used for phylogenetic analyses of H. pylori.

Statistical analysis
Discrete variables were tested using the chi-square test; continuous variables were tested using the Mann-Whitney U and t-tests. The Saphiro-Wilk test was carried out for testing data distribution. A multivariate logistic regression model was used to calculate the odds ratios (OR) of the clinical outcomes and H. pylori infection by age, sex, and other demographic factors. The OR and 95% confidence intervals (CI) were used to estimate the odds. We also evaluated the consistency of the results between diagnostic modalities using Cohen's Kappa test implemented in the irr package. A P-value of < 0.05 was accepted as statistically significant. All of these calculations were carried out on the R software version 4.0.3.