DNA methylation marker composition for gastric cancer screening

Abstract

The invention provides the DNA methylation marker composition for screening the gastric cancer, the screened DNA methylation marker composition can be used for diagnosing the gastric cancer, and the sensitivity and the specificity of the DNA methylation marker composition are both greater than 88%, so that the biomarker composition has relatively high diagnosis efficiency on the gastric cancer.

Description

Technical Field

[0001] The present invention belongs to the field of biomedicine, and in particular relates to a DNA methylation marker composition for gastric cancer screening. Background Art

[0002] Gastric cancer is one of the most burdensome malignant tumors in the world. According to the cancer statistics report of the World Health Organization's International Agency for Research on Cancer (IARC), there were approximately 1.089 million new confirmed cases of gastric cancer and 769,000 deaths from gastric cancer in 2020, ranking sixth and third in global cancer incidence and mortality, respectively. The more obvious risk factors for gastric cancer currently include Helicobacter pylori infection, high salt intake, smoking and heavy drinking, and family history of gastric cancer. The occurrence of gastric cancer is a complex pathological process with multiple factors and steps. It is the result of the interaction of multiple factors such as environmental factors, Helicobacter pylori and host gene mutations (Parkin DM, Bary FI, Devesa SS. Cancer burden in the year 2000. The global picture. Eur J Cancer 2001; 37 (Suppl 8): S4–S66.). The survival time of gastric cancer patients is closely related to the early or late clinical diagnosis. Gastric cancer is the fifth most common cancer and the fourth leading cause of cancer death (Thrift AP, Wenker TN, El-Serag HB. Global burden of gastric cancer: epidemiological trends, risk factors, screening and prevention [J]. Nat Rev Clin Oncol, 2023, 20 (5): 338-49.). Most patients are already in the advanced stage when diagnosed, and the recurrence rate after surgical treatment is high and the prognosis is poor (Shitara K, Chin K, Yoshikawa T, et al. Phase II study of adjuvant chemotherapy of S-1 plus oxaliplatin for patients with stage III gastric cancer after D2 gastrectomy [J]. Gastric Cancer, 2017, 20 (1): 175-81.), which is an important reason why the incidence and mortality of gastric cancer are among the highest.

[0003] Gastric cancer screening can significantly improve the detection rate of early gastric cancer lesions in the population, improve patient prognosis, and significantly increase patient survival rate. At present, the main way to diagnose gastric cancer is still gastroscopy and biopsy (Hirata Y, Maeda S, Mitsuno Y, et al. Helicobacter pylori CagA protein activates serum response element-driven transcription independently of tyrosine phosphorylation. Gastroenterology. 2002; 123: 1962–1971.). In addition to traditional endoscopy, many new technologies have been invented, including fluorescence imaging, confocal microendoscopy, chromoendoscopy, magnified narrow-band imaging, etc. The use of these technologies has improved the accuracy of lesion detection (Goto T, Haruma K, Kitadai Y, et al. Enhanced expression of inducible nitric oxide synthase and nitro tyrosine in gastric mucosa of gastric cancer patients. Clin Cancer Res. 1999; 5: 1411–1415.), but as an invasive examination, gastroscopy has high requirements for equipment and technology, poor patient compliance and is difficult to promote on a large scale. The screening performance of markers such as Helicobacter pylori, pepsinogen, gastrin, CEA, CA19-9, and CA72-4 is not satisfactory.

[0004] DNA methylation can mediate various processes in cells, such as maintaining genome integrity, genomic imprinting, transcriptional regulation, and developmental processes (5KANWAL R, GUPTA S. Epigenetics and cancer [J]. Journal of applied physiology (Bethesda, Md: 1985), 2010, 109 (2): 598-605.). DNA methylation in eukaryotes mainly occurs on cytosine bases, where DNA methyltransferases (DNMTs) selectively add methyl groups to cytosine bases to form 5-methylcytosine (5mC), which regulates gene transcription by affecting promoters and non-coding DNA elements (WANG J, YANG J, LID, et al. Technologies for targeting DNA methylation modifications: Basic mechanism and potential application in cancer [J]. Biochimica et biophysica acta Reviews on cancer, 2021, 1875 (1): 188454.). This type of DNA repair can affect gene expression by changing DNA conformation, chromatin structure, DNA stability and DNA-protein interaction without changing the gene sequence (MILLER J L, GRANT P A. The role of DNA methylation and histone modifications in transcriptional regulation in humans [J]. Sub-cellular biochemistry, 2013, 61: 289-317.). In addition, many studies have shown that DNA methylation markers have the advantages of high stability, good accuracy and independence from internal reference control in tumor tissues (PFEIFER G P. Defining Driver DNA Methylation Changes in Human Cancer [J]. International journal of molecular sciences, 2018, 19 (4).).More and more studies have shown that DNA methylation plays a very important role in regulating tumors (KOCHA, JOOSTENSC, FENGZ, et al. Analysis of DNA methylation in cancer: location revisited [J]. Nature reviews Clinical oncology, 2018, 15 (7): 459-66.). DNA methylation, as an epigenetic modification, participates in the regulation of gene expression. Normal methylation status plays an important role in cell growth, development, proliferation, differentiation, etc., while abnormal methylation status can cause abnormalities in cell cycle, DNA damage repair, gene transcription, etc., leading to the occurrence of tumors. The application of DNA methylation in cancer screening is a hot topic in current research, and the screening performance for gastric cancer needs to be improved. Summary of the invention

[0005] The present invention screens out a DNA methylation marker composition for gastric cancer screening, which has a strong diagnostic efficacy for diagnosing gastric cancer. Based on this, the present invention is completed.

[0006] In a first aspect, the present invention provides a DNA methylation marker composition for gastric cancer screening, wherein the biomarker composition comprises 200 DNA methylation regions in the following table:

[0007]

[0008]

[0009]

[0010]

[0011]

[0012]

[0013] Furthermore, the biomarkers are derived from whole blood and plasma of patients.

[0014] Furthermore, the DNA methylation levels of relevant regions in the samples were detected, and the malignancy degree of each methylation region of cfDNA was calculated by the DQDA method, and the screening score of each sample was calculated. The calculation formula of the screening score was:

[0015]

[0016] Where p is the screening score; x is the malignant ratio of the i-th region; a is the i-th region coefficient; b is the cutoff value, and the cutoff value b of the screening model is -0.4131;

[0017] When the screening score is ≥0.3886, it is judged as a positive gastric cancer screening; when the screening score is <0.3886, it is judged as a negative gastric cancer screening.

[0018] Furthermore, the gastric cancer includes adenocarcinoma, squamous cell carcinoma, signet ring cell carcinoma, etc.

[0019] In a second aspect, the present invention provides an application of a biomarker composition in the preparation of a reagent for diagnosing gastric cancer, wherein the biomarker composition comprises the 200 DNA methylation regions described in the first aspect of the present invention, and the reagent is a reagent that can detect the methylation levels of the 200 DNA methylation regions in a patient's biological sample.

[0020] Furthermore, the biomarkers are derived from whole blood and plasma of patients.

[0021] Furthermore, the DNA methylation level in the sample was detected, and the malignancy degree of each methylation region of cfDNA was calculated by the DQDA method, and the screening score of each sample was calculated. The calculation formula of the screening score was:

[0022]

[0023] Among them, p is the screening score; x is the malignancy of the i-th region; a is the i-th region coefficient; b is the cutoff value, and the cutoff value b of the screening model is -0.4131;

[0024] When the screening score is ≥0.3886, it is judged as a positive gastric cancer screening; when the screening score is <0.3886, it is judged as a negative gastric cancer screening.

[0025] Beneficial Effects

[0026] The DNA methylation marker combination screened by the present invention can be used to diagnose gastric cancer, and its sensitivity and specificity are both greater than 88%, indicating that the biomarker combination of the present invention has a high diagnostic efficacy for gastric cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is the technical roadmap of the present invention;

[0028] Figure 2 The box plot and ROC curve of the gastric cancer screening model in the training set;

[0029] Figure 3 The box plot and ROC curve of the gastric cancer screening model in the validation set;

[0030] Figure 4 The box plot and ROC curve of the gastric cancer screening model in the independent validation set;

[0031] Figure 5 The bar graph shows the performance comparison between gastric cancer screening model and tumor markers. DETAILED DESCRIPTION

[0032] The specific embodiments of the present invention are further described below. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. In addition, the technical features involved in the embodiments described below can be combined with each other as long as they do not conflict with each other.

[0033] The experimental methods in the following examples are conventional methods unless otherwise specified, and the experimental materials used in the following examples are commercially available unless otherwise specified.

[0034] The "DQDA" described in the present invention is Diagonal Quadratic Discriminant Analysis. QDA is a statistical method for classification, which assumes that the data of each category obeys a multivariate normal distribution with a different covariance matrix. In QDA, each category has its own covariance matrix, which makes the model more flexible to adapt to data sets of different shapes or sizes. The DQDA classifier is a modification of QDA in which the non-diagonal elements of the mixed sample covariance matrix are set to zero.

[0035] Example 1 Screening of gastric cancer-related DNA methylation markers from public databases

[0036] DNA methylation data of 383 gastric cancer tissue samples (mainly foreign population) were downloaded from the public database TCGA (The Cancer Genome Atlas) (Illumina Infinium Human Methylation 450BeadChip), and DNA methylation data of 61 gastric normal tissue samples were downloaded from the GEO (Gene Expression Omnibus) database (Illumina Infinium Human Methylation 450BeadChip, GSE103186). The R package ChAMP was used to perform differential analysis on the DNA methylation data, and DNA methylation sites with significant differences between gastric cancer tissues and normal tissues (p < 0.05, |Δβ| ≥ 0.2) were screened, and the DNA methylation sites located on chromosomes X and Y were removed. Finally, 43,882 differential DNA methylation sites were obtained.

[0037] Example 2 Screening of gastric cancer-related DNA methylation markers from self-tested gastric cancer tissue samples

[0038] Cancer tissue samples and paired adjacent tissue samples from 43 Chinese gastric cancer patients were collected. DNA methylation was detected using the Roche NimbleGenSeqCap Epi method (Roche methylation probe is approximately 80 Mb, containing 5.5 million DNA methylation sites, including all DNA methylation sites of the Illumina 450K chip). Difference analysis was performed using the Wilcoxon test to screen DNA methylation sites with significant differences between cancer tissues and adjacent tissues (p < 0.05, |Δβ| ≥ 0.2), and DNA methylation sites on chromosomes X and Y were removed. Finally, 73,999 differential DNA methylation sites were obtained, which were then combined with 43,882 differential DNA methylation sites screened from the public database and merged into 35,295 DNA methylation regions according to the rule that the distance between adjacent sites was < 100 bp.

[0039] Example 3 Screening of DNA methylation markers for gastric cancer screening

[0040] 1. Obtaining plasma samples

[0041] Plasma samples of subjects were derived from 74 healthy controls and 53 gastric cancer patients (including 24 patients in stage I-II and 29 patients in stage III-IV).

[0042] Healthy controls: aged 30-80 years old; subjects are able to undergo gastroscopy and other related examinations; subjects fully understand the nature of this study, related examination methods and the risks that may need to be borne by participating in this study, and can understand and voluntarily sign the informed consent; endoscopy and / or pathology are negative, no other cancers, and patients with gastric cancer and precancerous lesions found in previous gastroscopy are excluded.

[0043] Gastric cancer patients: aged 30-80 years old; subjects are able to undergo gastroscopy and other related examinations; subjects fully understand the nature of this study, related examination methods and the risks that may need to be borne by participating in this study, and can understand and voluntarily sign the informed consent; patients with different stages of gastric cancer (I, II, III, IV) are required to be diagnosed by gastroscopy and / or pathological examination, without concomitant malignant tumors of other systems or organs, and have not been treated with surgery, radiotherapy or chemotherapy before blood collection.

[0044] Exclusion criteria: small blood sample volume, plasma less than 4 mL, unable to perform DNA methylation testing; incomplete blood sample information (including sample collection time, sample number, subject age, gender, medical history, imaging results, gastroscopy and pathological results, etc.); blood sample quality control failure, including samples that do not meet the requirements for sample collection and processing, repeated freezing and thawing, hemolysis, microbial contamination, use of reagents with quality problems (including expired, contaminated, improperly stored, etc.) for sample testing, other human errors during the test process, etc.; pregnant women, breastfeeding women or women of childbearing age who are preparing to conceive; any situation that the research physician considers unsuitable for inclusion.

[0045] 2. DNA methylation detection using plasma samples and calculation of screening scores

[0046] DNA methylation was detected using the Twist Human Methylome Panel (about 123Mb, covering 84.2% of CpG island sites in the human genome), and the malignancy of each methylated region of cfDNA (cfDNA malignant ratio) was calculated by the DQDA (diagonal quadratic discriminant analysis) method. The Wilcoxon test was used to analyze the differences between 74 healthy controls and 53 gastric cancer patients. According to the criteria of p<0.05 and malignant ratio diff≥0.01, 2676 DNA methylation regions with high malignancy in gastric cancer were screened from the 35295 DNA methylation regions screened by tissues, and the top 200 DNA methylation regions were sorted from small to large p values ​​as markers (see Table 1).

[0047] Table 1 Information of 200 DNA methylation regions

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054] Note: Reference genome hg19

[0055] The logistic regression machine learning algorithm was used for modeling, and the model hyperparameters were determined using ten-fold cross validation to calculate the screening score for each sample. The standard logistic regression formula, that is, multiple linear regression and sigmoid transformation, was used to obtain the screening score. The screening score calculation formula is:

[0056]

[0057] Among them, p is the screening score; x is the malignancy level of the i-th area; a is the i-th area coefficient; b is the cutoff value, and the cutoff value b of the screening model is -0.4131.

[0058] The threshold value of 0.3886 corresponding to the maximum Youden index was taken as the threshold value of the screening model. If the screening score was ≥0.3886, it was judged as a positive gastric cancer screening; if the screening score was <0.3886, it was judged as a negative gastric cancer screening.

[0059] 3. Experimental results

[0060] The screening results were compared with the gold standard (i.e., gastroscopy combined with pathological biopsy results). The sensitivity and specificity of the gastric cancer screening model in the training set were 96.23% and 90.54%, respectively, and the AUC was 0.987. The sensitivity of stage I-II and stage III-IV were 91.67% and 100%, respectively, and the AUC were 0.978 and 0.994, respectively (see Figure 2 , Table 2).

[0061] Table 2 Performance statistics of gastric cancer screening model in training set, validation set and independent validation set

[0062]

[0063]

[0064] Example 4 Validation of DNA methylation markers for gastric cancer screening

[0065] 1. Obtaining plasma samples

[0066] Plasma samples of subjects: from 25 healthy controls and 17 gastric cancer patients (7 patients in stage I-II and 10 patients in stage III-IV). The inclusion criteria and grouping of subjects were the same as those in Example 3.

[0067] 2. DNA methylation detection using plasma samples and calculation of screening scores

[0068] The plasma samples were tested for DNA methylation using the Twist Human Methylome Panel (about 123 Mb, covering 84.2% of CpG island sites in the human genome), and the malignancy of each methylated region of cfDNA (cfDNA malignant ratio) was calculated using the DQDA (diagonal quadratic discriminant analysis) method. The screening scores of 25 healthy controls and 17 gastric cancer patients in the validation set were calculated using the logistic regression model determined in Example 3.

[0069] 3. Experimental results

[0070] By comparing the screening score with the threshold of 0.3886, the negative / positive screening results were obtained, and then the screening results were compared with the gold standard. The sensitivity and specificity of the gastric cancer screening model in the validation set were determined to be 94.12% and 84%, respectively, and the AUC was 0.875. The sensitivity of stage I-II and stage III-IV was 85.71% and 100%, respectively, and the AUC was 0.794 and 0.932, respectively (see Figure 3 , Table 2).

[0071] Example 5 Independent Validation Set Plasma Samples Validation Model Performance

[0072] 1. Obtaining plasma samples

[0073] Plasma samples of subjects: from 25 healthy controls and 18 gastric cancer patients (8 patients in stage I-II and 10 patients in stage III-IV). The inclusion criteria and grouping of subjects were the same as those in Example 3.

[0074] 2. DNA methylation detection using plasma samples and calculation of screening scores

[0075] The plasma samples were tested for DNA methylation using the Twist Human Methylome Panel (about 123 Mb, covering 84.2% of CpG island sites in the human genome), and the malignancy of each methylated region of cfDNA (cfDNA malignant ratio) was calculated using the DQDA (diagonal quadratic discriminant analysis) method. The screening scores of 25 healthy controls and 18 gastric cancer patients in the external validation set were calculated using the logistic regression model determined in Example 3.

[0076] 3. Experimental results

[0077] By comparing the screening score with the threshold of 0.3886, the negative / positive screening results were obtained, and then the screening results were compared with the gold standard. The sensitivity and specificity of the gastric cancer screening model in the independent validation set were determined to be 88.89% and 88%, respectively, and the AUC was 0.913. The sensitivity of stage I-II and stage III-IV were 75% and 100%, and the AUC were 0.85 and 0.964, respectively (see Figure 4 , Table 2).

[0078] The results verified the diagnostic efficacy of the DNA methylation marker combination for gastric cancer screening determined by the present invention and confirmed its feasibility in actual clinical diagnosis.

[0079] Example 6 Performance comparison of gastric cancer screening model and tumor markers

[0080] 48 healthy controls and 64 gastric cancer patients (including 31 patients in stage I-II and 33 patients in stage III-IV) with tumor markers (CEA / CA199 / CA50 / CA125 / CA242 / CA724) and methylation test results in the training set of Example 3, the validation set of Example 4, and the independent validation set of Example 5 were selected for performance comparison. Among them, if any one of the tumor markers CEA / CA199 / CA50 / CA125 / CA242 / CA724 is detected positive, the patient is judged as positive; when all the tumor markers CEA / CA199 / CA50 / CA125 / CA242 / CA724 are detected negative, the patient is judged as negative. By comparing with the gold standard, the sensitivity and specificity of the tumor markers are determined to be 55% and 85%, wherein the sensitivity of stage I-II and stage III-IV is 35% and 72%, respectively. The sensitivity and specificity of the gastric cancer screening model were 98% and 88%, with the sensitivities of stage I-II and stage III-IV being 96% and 100%, respectively (see Table 3 and Figure 5 ). The results showed that the sensitivity of the gastric cancer screening model was significantly better than that of tumor markers, especially for early gastric cancer.

[0081] Table 3 Gastric cancer screening model and tumor marker performance statistics

[0082]

Claims

1. A DNA methylation marker composition for gastric cancer screening, the biomarker composition comprising 200 DNA methylation regions in the following table:

2. The composition according to claim 1, wherein the biomarkers are derived from whole blood or plasma of a patient.

3. The composition according to claim 1, detecting the DNA methylation level of the relevant region in the sample, calculating the malignancy degree of each methylated region of cfDNA by the DQDA method, and calculating the screening score of each sample, wherein the calculation formula of the screening score is: in, p is the screening score; x is the malignant ratio of the i-th region; a is the i-th region coefficient; b is the cutoff value, and the cutoff value b of the screening model is -0.4131; When the screening score is ≥0.3886, it is judged as a positive gastric cancer screening; when the screening score is <0.3886, it is judged as a negative gastric cancer screening.

4. Use of a biomarker composition in the preparation of a reagent for diagnosing gastric cancer, the biomarker composition comprising the 200 DNA methylation regions described in the first aspect of the present invention, and the reagent being a reagent that can detect the methylation levels of the 200 DNA methylation regions in a patient's biological sample.

5. The use according to claim 4, wherein the biomarker is derived from the patient's whole blood or plasma.

6. The use according to claim 4, detecting the DNA methylation level in the sample, calculating the malignancy degree of each methylation region of cfDNA by the DQDA method, and calculating the screening score of each sample, wherein the calculation formula of the screening score is: in, p is the screening score; x is the malignancy of the i-th region; a is the i-th region coefficient; b is the cutoff value, and the cutoff value b of the screening model is -0.4131; When the screening score is ≥0.3886, it is judged as a positive gastric cancer screening; when the screening score is <0.3886, it is judged as a negative gastric cancer screening.

Images