Abstract Chronic gastritis (CG) and osteoporosis (OP) are common and occult diseases in the elderly and the relationship of these two diseases have been increasingly exposed. We aimed to explore the clinical characteristics and shared mechanisms of CG patients combined with OP. In the cross-sectional study, all participants were selected from BEYOND study. The CG patients were included and classified into two groups, namely OP group and non-OP group. Univariable and multivariable logistic regression methods were used to evaluate the influencing factors. Furthermore, CG and OP-related genes were obtained from Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) were identified using the GEO2R tool and the Venny platform. Protein–protein interaction information was obtained by inputting the intersection targets into the STRING database. The PPI network was constructed by Cytoscape v3.6.0 software again, and the key genes were screened out according to the degree value. Gene function enrichment of DEGs was performed by Webgestalt online tool. One hundred and thirty CG patients were finally included in this study. Univariate correlation analysis showed that age, gender, BMI and coffee were the potential influencing factors for the comorbidity (P < 0.05). Multivariate Logistic regression model found that smoking history, serum PTH and serum β-CTX were positively correlated with OP in CG patients, while serum P1NP and eating fruit had an negative relationship with OP in CG patients. In studies of the shared mechanisms, a total of 76 intersection genes were identified between CG and OP, including CD163, CD14, CCR1, CYBB, CXCL10, SIGLEC1, LILRB2, IGSF6, MS4A6A and CCL8 as the core genes. The biological processes closely related to the occurrence and development of CG and OP mainly involved Ferroptosis, Toll-like receptor signaling pathway, Legionellosis and Chemokine signaling pathway. Our study firstly identified the possible associated factors with OP in the patients with CG, and mined the core genes and related pathways that could be used as biomarkers or potential therapeutic targets to reveal the shared mechanisms. Subject terms: Endocrinology, Gastroenterology, Pathogenesis, Risk factors Introduction Chronic gastritis (CG) and osteoporosis (OP) are common and occult diseases in the elderly. According to the data of recent epidemiological studies, there has been a decrease in the proportion of Helicobacter pylori-associated gastritis with an increase in the contribution of other etiological factors^[44]1. The international survey in 2003–2012 found that the prevalence of CG based on endoscopic diagnosis was close to 90%^[45]2. Meanwhile, OP is a systemic skeletal disease affecting affect up to 50% of postmenopausal women and 20% of men older than 50 years, with high health and economic burden worldwide^[46]3. With the progress of OP, osteocalcin will continuously lose and the patients will experience pain and spinal deformation. Moreover, severe osteoporotic fractures may occur; some patients have decreased muscle volume and strength, and are prone to falls, leading to an increased risk of fractures and a decline in quality of life. In recent years, the relationship between CG and OP has received more attention and has been increasingly exposed^[47]4. Many patients with CG will experience a significant reduction in bone density throughout the body. Some hormone substances secreted by secretory cells of gastric mucosa can regulate bone calcium, which are closely related to the occurrence of OP^[48]5. The occurrence of CG is mainly due to Helicobacter pylori infection, diet, lifestyle, and so on^[49]6–[50]9. The risk factors of OP are mainly concentrated in age, diet, and lifestyle, and OP could be caused by some diseases such as gastrointestinal diseases, endocrine diseases, etc.^[51]10,[52]11. It is a pity that there are few researches to explore the relationship of these two diseases and the related risk factors. The shared mechanisms of CG and OP are not very clear at present, which would further hamper our investigation of the comorbidity of these two diseases. In any gastrointestinal disease, the response of low circulating leptin to weight loss may be an important factor in reducing bone mass^[53]4. It is also found that intestinal microorganisms are closely related to the regulation of intestinal permeability and bone health^[54]12. Consequently, the identified core genes may become a new research focus, and the obtained molecular mechanisms and signaling pathways contribute to the study of the relationship between CG and OP^[55]13. In this study, we implemented the cross-sectional study in 10 communities in Beijing, China^[56]14. We aimed to analyze the clinical characteristics and associated factors for patients with CG complicated with OP. And we enriched the signaling pathways common to both by mining the shared novel genes of CG and OP, revealing their shared mechanisms. To the best of our knowledge, this might be the first study to analyze the clinical characteristics and explore the shared gene signatures between CG and OP through a cross-sectional study and using a systems biology approach. Materials and methods Cross-sectional studies Study design All participants were selected from 1540 community residents who completed past medical history inquiries and physical examinations in Chaoyang and Fengtai districts of Beijing, China from November 2017 to July 2018^[57]14. In the survey, we collected the general information of 1540 community residents, current diseases, bone mineral density, diet, drug use, biological indicators and other relevant data information. The trial was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained for all material from each participant. This study had been registered on the Chinese Clinical Trial Registry already. (Website: [58]http://www.chictr.org.cn) (Registration Number: ChiCTR-SOC-17013090). Diagnostic criteria CG patients were determined by clinical diagnosis reports and patient self-report, regardless of the different types of CG^[59]15. The diagnostic criteria of OP patients referred to the recommendation of the World Health Organization (WHO), namely, taking into account the T value of bone mineral density: T value > − 2.5 SD was non-OP; T value ≤ − 2.5 SD was OP^[60]16 (Fig. [61]1). Figure 1. [62]Figure 1 [63]Open in a new tab Crowd screening flow chart. Inclusion and exclusion criteria The inclusion criteria were as follows: (1) aged from 45 to 80 years; (2) the subjects lived locally lasting for more than 5 years; (3) have clinical diagnosis report and patient self-report diagnosed as CG; (4) the population with complete clinical and laboratory information including BMD, bone metabolic markers. The exclusion criteria included: (1) the participants who had missing important information; (2) having digestive tract tumors and diseases that affect bone metabolism or calcium absorption, such as diabetes and thyroid diseases, hematological diseases, leukemia, myeloma, systemic lupus erythematosus and kidney diseases; (3) Patients who received drugs or treatments that may affect the study within the first three months of the study, such as glucocorticoids and immunosuppressants. Information acquisition The study focuses on information collection in three aspects: population characteristics (age, gender, BMI, drinking and smoking), serum biochemical markers (β-CTX, PTH, ALP, P1NP), diet types (fruits, milk, yogurt and coffee). Bone mineral density Dual-energy X-ray absorptiometry device (Hologic, WI, USA) was used to assess the value of BMD (g/cm^2). The anteroposterior L1-L4 and left proximal femur including the femoral neck and the total hip BMD were detected, and the T and Z values of each site were recorded. According to the WHO criteria, the OP and non-OP populations were identified based on bone densitometry. Bone metabolism index detection Fasting blood samples of the participants were collected from 8 a.m. to 9 a.m. in the sitting position. The measurements were conducted through an automated electrochemiluminescence immunoassay system (Roche, Cobas E601, Germany). Additionally, the blood samples were centrifuged to get serum and stored at minus 80 centigrade. As a professional third-party testing organization, Guangzhou KingMed Diagnostics Limited Liability Company was responsible for collecting and testing blood samples. Data collection All the data, including demographics, clinical characteristics, potential influencing factors, laboratory test and BMD results, were checked by the two independent researchers. Studies of shared mechanisms Access to disease genes Microarrays related to CG and OP were acquired from the GEO database (update time: April 10, 2022, [64]https://www.ncbi.nlm.nih.gov/gene). The keywords “chronic gastritis”, “osteoporosis”, and only those genes from “Homo sapiens” were used as the research targets, analyzed, and discussed in this paper. The gene expression profiling data of CG gastric epithelial tissues were obtained under accession number [65]GSE60427, containing 15 patients with CG and 7 healthy controls, whose level was [66]GPL17077; Gene expression profiling data of peripheral blood of OP, number [67]GSE56116, containing 10 OP patients and 3 healthy controls, whose platform was used as [68]GPL4133. Analysis and selection of DEGs GEO2R ([69]http://www.ncbi.nlm.nih.gov/geo/geo2r/) is an intelligent online analysis tool that can integrate and analyze two datasets under the same experimental conditions, or split any GEO data^[70]17. In this study, GEO2R tool was used to analyze the genes between CG and OP. The genes that adjusted the P < 0.05 and |log2FC|(Fold Change) > 1 were considered as DEGs. After obtaining the differentially expressed genes, online Venn analysis tool ([71]http://bioinformatics.psb.ugent.be/webtools/Venn/) was used to obtain DEGs intersection and specific DEGs shared by CG and OP. Protein–protein interaction analysis To identify the relationships among the intersection targets, we used the STRING database ([72]https://string-db.org/). According to the comprehensive analysis of the topological parameters “closeness,” “betweenness,” and “degree”^[73]18,[74]19. Subsequently, 10 core genes were further screened out by the cytoHubba plug-in of Cytoscape v3.6.0 software^[75]20–[76]22. GO (gene ontology) functional enrichment and KEGG signaling pathways analysis Gene Ontology (GO) is a commonly used bioinformatics tool, which can provide relevant information according to the defined characteristics, including comprehensive information on gene function of a single genome product. GO enrichment analysis can explain gene functions from three aspects: molecular function (MF), biological process (BP) and cellular component (CC). Kyoto Encyclopedia of Genes and Genomes (KEGG) is a database that systematically analyzes the metabolic pathways and functions of gene products in cells. In this study, GO and KEGG analyses were performed using WebGestalt ([77]http://www.webgestalt.org/)^[78]23,[79]24. Statistical analysis The classification variable was expressed as frequency and percentage (%). The χ^2 test or Fisher exact test or Bonferroni’s method was used to compare the categorical variables between the two groups. Student’s t test or Mann–Whitney U test was used to analyze continuous variables. In addition, we conducted an in-depth analysis of the clinical characteristics of the OP and non-OP groups. Univariate and multivariate logistic regression models were used to explore the risk factors of CG complicated with OP. All the analysis was carried out with IBM SPSS Statistics 23.0 software. Bidirectional alpha less than 0.05 was considered statistically significant. Ethics approval The study was approved by the Ethics Committee of Wangjing Hospital, Chinese Academy of Chinese Medical Sciences (NO. WJEC-KT-2017-020-P001) and registered on the platform of China Clinical Trial Registry (NO. ChiCTR-SOC-17013090). Consent to participate Informed consent was obtained from all individual participants included in the study. Results Cross-sectional studies Participants characteristics A total of 130 CG patients were included in this study. According to whether CG patients combined with OP, the subjects were divided into OP group (48 cases) and non-OP group (82 cases). Compared with the non-OP group, the patients of OP group were older, including 5 males (10.4%) and 43 females (89.6%) in the OP group, 26 males (31.7%) and 56 females (68.3%) in the non-OP group, and the BMI of patients in OP group was lower than that of the non-OP group. There were significant differences in age (P = 0.04), gender (P = 0.01), and BMI (P = 0.03) between OP and non-OP patients. In addition, there was no statistically significant difference in drinking and smoking history (Table [80]1, Fig. [81]2). Table 1. Descriptive characteristics of the study population. Characteristics Total (n = 130) CG + OP (n = 48) CG + non-OP (n = 82) t/2 P Age 65.11 ± 7.96 66.94 ± 7.73 64.04 ± 7.94 2.05 0.04 Gender Male 31 (23.85%) 5 (10.4%) 26 (31.7%) 7.56 0.01 Female 99 (76.15%) 43 (89.6%) 56 (68.3%) BMI 24.20 ± 2.97 23.46 ± 2.69 24.63 ± 3.05  − 2.27 0.03 Drinking history Non-drinking 114 (87.69%) 42 (87.50%) 72 (87.80%) 0.36 0.84 Drinking 12 (1.54%) 4 (8.30%) 8 (9.80%) Abandoned drinking 4 (3.08%) 2 (4.20%) 2 (2.40%) Smoking history Nonsmoking 84 (64.62%) 32 (66.70%) 52 (63.40%) 3.81 0.28 Active smoking 14 (10.77%) 2 (4.20%) 12 (14.60%) Passive smoking 20 (15.39%) 9 (18.80%) 11 (13.40%) Smoking has been stopped 12 (9.23%) 5 (10.40%) 7 (8.50%) [82]Open in a new tab Data are x ± S, n (%), or n/N (%). p values were calculated by the Mann–Whitney U test, χ^2 test, or Fisher’s exact test, as appropriate. Figure 2. [83]Figure 2 [84]Open in a new tab Descriptive characteristics of the study population. Comparison of biochemical markers There was no significant difference in β-CTX, PTH, ALP, and P1NP between OP and non-OP patients (Table [85]2). Table 2. Comparison of biochemical markers between two groups. Group β-CTX (ng/mL) PTH (pmol/L) ALP (U/L) P1NP (ng/mL) CG + OP 0.32 ± 0.13 3.50 ± 1.26 78.65 ± 19.38 53.21 ± 21.36 CG + non-OP 0.31 ± 0.13 3.24 ± 2.14 84.15 ± 21.55 58.43 ± 25.13 t 0.65 1.24  − 1.86  − 1.59 P 0.52 0.22 0.07 0.12 [86]Open in a new tab Comparison of diet types In terms of diet types, coffee had a significant difference between OP and non-OP patients, while fruits, milk or yogurt had no significant difference between OP and non-OP patients (Table [87]3, Fig. [88]3). Table 3. Comparison of diet types between groups. Group CG + OP CG + non-OP F P No eating (drinking) Occasional Often Daily Frequent No eating (drinking) Occasional Often Daily Frequent Fruits 3 (6.3%) 6 (12.5%) 10 (20.8%) 16 (33.3%) 13 (27.1%) 1 (1.2%) 6 (7.3%) 16 (19.5%) 34 (41.5%) 25 (30.5%) 2.51 0.12 Milk or yogurt 0 (0.0%) 1 (2.1%) 2 (4.2%) 12 (25.0%) 33 (68.8%) 1 (1.2%) 1 (1.2%) 4 (4.9%) 27 (32.9%) 49 (59.8%) 0.20 0.66 Coffee 12 (25.0%) 26 (54.2%) 9 (18.8%) 0 (0.0%) 1 (2.1%) 13 (15.9%) 47 (57.3%) 17 (20.7%) 3 (3.7%) 2 (2.4%) 4.69 0.03 [89]Open in a new tab Figure 3. [90]Figure 3 [91]Open in a new tab CG + OP and CG + non-OP groups diet types. Multivariable logistic regression analysis for influencing factors Age, gender, and BMI were taken as covariates for classification, and their correlations were taken as independent variables for Logistic regression equation analysis. Resulting that age, gender, BMI, serum PTH, serum β-CTX, serum P1NP and fruit were all related factors for CG combined with OP (Table [92]4, Fig. [93]4). Ultimately, smoking history, serum PTH and serum β-CTX were positively correlated with OP in CG patients, while serum P1NP and eating fruit had an obviously negative relationship with OP in CG patients. Table 4. Multivariable logistic regression analysis. Factor β Wald P OR (95% CI) Age 10.360 0.016 Age (1)  − 22.207 0.000 0.999 0.000 (0.000 ~) Age (2)  − 2.597 10.254 0.001 0.075 (0.015 ~ 0.365) Age (3)  − 1.389 4.624 0.032 0.249 (0.070 ~ 0.884) Gender (1)  − 4.109 13.817 0.000 0.016 (0.002 ~ 0.143) BMI 9.959 0.019 BMI (1) 1.121 0.358 0.549 3.068 (0.078 ~ 120.384) BMI (2) 3.113 8.354 0.004 22.484 (2.724 ~ 185.599) BMI (3) 1.745 2.936 0.087 5.725 (0.778 ~ 42.126) Drinking history 0.840 1.255 0.263 2.316 (0.533 ~ 10.058) Smoking history 0.590 4.413 0.036 1.804 (1.040 ~ 3.127) Serum β-CTX 6.334 7.272 0.007 563.635 (5.644 ~ 56,288.907) Serum PTH 0.596 6.781 0.009 1.816 (0.159 ~ 2.844) Serum ALP 0.022 2.366 0.124 1.022 (0.994 ~ 1.051) Serum P1NP  − 0.031 3.889 0.049 0.970 (0.941 ~ 1.000) Fruits  − 0.655 6.276 0.012 0.519 (0.311 ~ 0.867) Milk or yogurt  − 0.011 0.002 0.960 0.989 (0.642 ~ 1.523) Coffee  − 0.576 0.754 0.385 0.562 (0.153 ~ 2.065) [94]Open in a new tab Figure 4. [95]Figure 4 [96]Open in a new tab Forest plot of factors associated with OP in CG patients. Studies of comorbid mechanisms Disease genes of OP and CG P < 0.05 and |log2FC|(Fold Change) > 1 were as criteria for defining differential genes. Through GEO2R analysis, we found 1803 CG DEGs and 691 OP DEGs. Then, the DEGs of CG and OP were analyzed via the Venn platform, and a total of 75 DEGs were obtained. These are presented in Fig. [97]5A and Table [98]5. Figure 5. [99]Figure 5 [100]Figure 5 [101]Open in a new tab Venn diagram of CG-OP intersection targets (A), DAG (B), PPI network and topology analysis (C,D), GO functional enrichment analysis (E), KEGG enrichment analysis of signaling pathways (F,G). Table 5. Common gene of CG and OP. Number Gene Number Gene Number Gene 1 ANO3 27 ICAM4 53 PTAFR 2 ANPEP 28 IGSF6 54 RANBP17 3 ATP8B3 29 IL19 55 RASSF4 4 C2 30 KCNK12 56 RNF17 5 CADM3 31 KCNQ1 57 S100A12 6 CCL8 32 KCTD15 58 SCG5 7 CCR1 33 KIAA2022 59 SCML4 8 CD14 34 LILRA4 60 SCNN1B 9 CD163 35 LILRB2 61 SELE 10 CEBPE 36 LILRB3 62 SELL 11 CXCL10 37 LIPG 63 SIGLEC1 12 CXCL12 38 LOC100505622 64 SIRPB2 13 CXorf21 39 LPL 65 SLC2A6 14 CYBB 40 MPEG1 66 SLC37A2 15 ANO3 41 MRAS 67 SLC6A12 16 CYP1B1-AS1 42 MS4A6A 68 SNX29 17 DRD3 43 MS4A7 69 SOD2 18 DUOX2 44 NAIP 70 TAC1 19 ESR2 45 NCF1 71 TCF4 20 F5 46 NPCDR1 72 TFEC 21 FCN1 47 OLFM1 73 TLR6 22 FSCN1 48 OLIG1 74 TMEM176B 23 HDAC9 49 P2RY13 75 TNFAIP6 24 HGF 50 PKD2L1 25 HLA-DOA 51 PLEKHA2 26 HLA-DRB5 52 PPP1R1C [102]Open in a new tab PPI network construction and core gene screening A total of 10 core genes were obtained by a 12-parameter correlation analysis of 75 intersection genes with cytoHubba. Having thus obtained the PPI data, we imported these data into Cytoscape to plot the PPI network. These are presented in Fig. [103]5B–D and Table [104]6. Table 6. Core gene of CG and OP. Number Gene Number Gene 1 CD163 6 SIGLEC1 2 CD14 7 LILRB2 3 CCR1 8 IGSF6 4 CYBB 9 MS4A6A 5 CXCL10 10 CCL8 [105]Open in a new tab GO enrichment analysis and KEGG pathways enrichment analysis The results of GO enrichment analysis consist of biological processes (BP), cell components (CC), and molecular functions (MF). To further study the biological function of key targets, an enrichment analysis of KEGG pathway was carried out. Meanwhile, the top 10 signaling pathways of key targets, including Ferroptosis, Toll-like receptor signaling pathway, Pathogenic Escherichia coli infection, Legionellosis, Chemokine signaling pathway, Cytosolic DNA-sensing pathway, Acute myeloid leukemia, RIG-I-like receptor signaling pathway, Phagosome, and Cytokine-cytokine receptor interaction (Fig. [106]5E–G). Discussion CG and OP are highly prevalent diseases around the world. A previous study showed that OP could be caused by gastrointestinal diseases, endocrine diseases, etc.^[107]10,[108]11. With the progressive increase in the prevalence of CG and OP, it is necessary to explore the clinical characteristics and mechanism in these 2 diseases and discover the associated factors and early potential targets to prevent disease development^[109]25,[110]26. Helicobacter pylori (Hp) infection seemed to be one of the most important factors that may increase the risk of CG patients complicated with OP. A Japanese research of 230 male patients aged 50–60 years found that HP infection increased the risk of low bone mass by 1.83 times and atrophic gastritis increased the risk of low bone mass by 2.22 times. Therefore, HP infection and atrophic gastritis were considered as high-risk factors for OP^[111]27. However, the relationship between HP infection and OP is also controversial. Through a cross-sectional study of 85 female patients, Brazilian researchers found that HP-associated gastritis and autoimmune gastritis were not risk factors for abnormal bone mass^[112]28. Iranian researchers included 967 elderly people over 60 years old. The results also showed that HP infection was not significantly associated with bone mineral density change^[113]29. The prevalence of CG increases with age^[114]30. This is mainly related to the increase in Hp infection rate with age, and atrophy, intestinal metaplasia and “aging” are also related to a certain extent. In our study, CG patients with OP were older than those without OP. In addition, the high risk of elderly patients with OP may be attributed to their overall poor health, bone loss accelerated and an increase in the number of complications. In addition, our study found that there were gender differences in patients with CG complicated with OP. Marked gender differences in the disease have been confirmed in other observational studies reporting the total incidence of autoimmune gastritis in the population is 2%, in which the incidence of elderly women is as high as 4% to 5%, and there is no race or region specificity^[115]31. Poor lifestyle is a commonly associated factor for the occurrence of CG patients complicated with OP, including dietary preferences and drink