CN118421799A - A group of specific methylation diagnostic markers related to lung adenocarcinoma and their application - Google Patents

Abstract

The invention provides a group of specific methylation diagnosis markers related to lung adenocarcinoma and application thereof, and relates to genetic engineering, wherein the lung adenocarcinoma diagnosis markers are characterized in that: comprises at least one of cg24996482(NA)、cg07077277(HIST1H2AA)、cg05148722(HIST1H2BF)、cg00608169(NA)、cg15435730(TFAP2D)、cg21174344(NA)、cg15803869(NA)、cg23943136(GATA3)、cg03933026(NA)、cg27630311(TBX3)、cg08448665(C21orf88) and cg17900854 (C21 orf 88). Specific verification proves that the specific CpG sites of the 12 lung adenocarcinoma provided above are abnormally changed in early lung adenocarcinoma patients, the diagnostic effect of the 12 CpG sites is evaluated in TCGA lung adenocarcinoma data, the AUC range is between 0.652 and 0.911, the sensitivity is between 54.9 and 83.1 percent, and the specificity is between 81.2 and 100 percent; the comprehensive diagnosis effect of the 12 CpG sites is evaluated by constructing a random forest model based on TCGA lung adenocarcinoma data, the accuracy is 99.7%, and the accuracy is 83.9% and 97.9% respectively by verification in a GEO data set.

Description

A group of specific methylation diagnostic markers related to lung adenocarcinoma and their applications

Technical Field

The present invention relates to the technical field of biological diagnostic reagents, and in particular to a group of specific methylation diagnostic markers related to lung adenocarcinoma and applications thereof.

Background technique

According to the latest estimates by the International Cancer Agency, there will be about 2.2 million new cases of lung cancer in 2020, accounting for about 11.4% of new cancer cases each year; and the number of deaths will be about 1.8 million, accounting for 18%, and its incidence and mortality rate rank second and first respectively in the world. According to the "2016 China Cancer Statistics" released by the National Cancer Center, the incidence of lung cancer in my country has been increasing year by year in the past two decades; in 2016, the incidence and mortality of lung cancer in my country ranked first, with about 828,000 new cases and about 657,000 deaths. In addition, the incidence of lung cancer in my country varies greatly among different genders, age groups and regions. In recent years, the growth of patients has been mainly in women, people over 45 years old and urban areas. In addition to age, a number of high-risk factors related to lung cancer have been identified, such as smoking, air pollution, occupational exposure to silica, previous history of lung diseases (chronic obstructive pulmonary disease and pulmonary fibrosis, etc.) and family history of lung cancer, but there are many factors that affect the occurrence of lung cancer, and its pathogenesis remains to be explored. Lung adenocarcinoma is the most common subtype of lung cancer, accounting for about 60% of all lung cancers, and is showing a clear upward trend. It can be expected that the incidence of lung adenocarcinoma will continue to rise in the future, and it is one of the serious challenges threatening the health of Chinese residents.

Timely and accurate diagnosis is the prerequisite for the treatment of lung adenocarcinoma and one of the most effective ways to reduce the mortality of lung adenocarcinoma. Commonly used tissue biopsy methods include percutaneous lung puncture biopsy and bronchoscopy under the guidance of CT or color Doppler ultrasound. In addition to being invasive and having potential complications, tissue samples have certain tumor heterogeneity, which will affect the accuracy of diagnosis, and are difficult to perform in patients with early cancer or postoperative residual and cancer recurrence. At present, low-dose computed tomography (LDCT) of the chest is a recognized lung cancer screening method. The results of multiple randomized controlled studies have shown that LDCT screening can reduce the overall mortality of lung cancer by improving the detection of early lung cancer. However, due to the high sensitivity of LDCT, its false positive rate in clinical trials is as high as 23.3%. This suggests that there is a certain deviation between the comprehensive judgment of the benign and malignant nature of lung nodules based on imaging and the actual situation of the lesions, and its accuracy still needs to be improved. And due to concerns about false-positive imaging results and radiation exposure, the compliance of CT screening in high-risk populations still needs to be improved. Exfoliative cytology of sputum or detection of nucleic acid molecules (RNA, DNA, and microRNA) in epithelial cells of bronchoalveolar lavage fluid are simple and easy to perform and can be tested multiple times, but their sensitivity in early lung cancer is not high, and further research is needed to determine their clinical effectiveness and practicality.

In recent years, many studies have begun to explore the diagnostic effect of plasma free cfDNA methylation markers in lung cancer/non-small cell lung cancer/lung adenocarcinoma, suggesting the potential value of cfDNA methylation markers as markers for screening and early diagnosis of lung adenocarcinoma. Early studies focused on the diagnostic effect of the methylation level of a single gene in lung cancer cfDNA (such as SHOX2, SEPT9 and DCLK1, etc.), with a sensitivity ranging from 44.3% to 60%. Subsequently, more studies screened methylation markers based on whole genome methylation data. In 2017, Ooki A et al. screened 6 methylation markers (CDO1, HOXA9, AJAP1, PTGDR, UNCX and MARCH11) based on TCGA lung cancer tissue and paracancerous tissue 450K chip data. The sensitivity and specificity of the six markers in the combined diagnosis of early non-small cell lung cancer in cfDNA (N=85) were 72.1% and 71.4%, respectively. In 2017, Hulbert A et al. also identified 6 incompletely overlapping methylation markers (SOX17 and TAC1, etc.) based on TCGA lung cancer tissue and adjacent samples, with sensitivity and specificity ranging from 65-76% and 74-84% in non-small cell lung cancer cfDNA (N=210), respectively. In 2021, Liang N et al. also identified a set of pan-cancer markers based on TCGA tissue data, with a sensitivity of 52-81% in cfDNA of stage I-III lung cancer patients. We note that when determining methylation markers, the above studies only used lung cancer tissue/plasma and adjacent tissue/healthy control cfDNA to analyze and screen out markers with the greatest difference between the two groups, or pan-cancer markers and some known tumor suppressor genes for subsequent non-invasive diagnosis, without considering the tumor-specific characteristics of the markers. In addition, some studies have taken into account the specific methylation profiles of different cells or cancer types when exploring cancer markers. However, the sensitivity is limited in early lung cancer patients (21.9% (14.8-31.1%)). Currently, cfDNA methylation detection is mainly based on high-throughput sequencing and low-throughput detection such as whole-genome methylation detection and targeted sulfur sequencing. The above high-throughput sequencing has undoubtedly accelerated the development of markers, but more blood needs to be collected to meet the sample volume, which may reduce the compliance of participants. The complex technical equipment and high cost price limit its promotion in large sample populations.

Summary of the invention

The purpose of the present invention is to solve the problem in the prior art that high-throughput sequencing requires collecting more blood to meet the sample loading amount, which may reduce the compliance of the participants. The complex technical equipment and high cost price limit its promotion in large sample populations.

The purpose of the present invention is to solve the problem in the prior art of lacking a group of screening processes and preliminary verification of lung adenocarcinoma-specific methylation markers.

In order to achieve the above object, the present invention adopts the following technical solutions:

A diagnostic marker for lung adenocarcinoma, comprising at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88).

The present application also provides the use of CpG sites as lung adenocarcinoma tissue-specific methylation markers in the preparation of lung adenocarcinoma diagnostic products, wherein the CpG sites are at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88).

Preferably, the CpG sites are a combination of six CpG sites: cg07077277 (HIST1H2AA), cg08448665 (C21orf88), cg15435730 (TFAP2D), cg17900854 (C21orf88), cg27630311 (TBX3) and cg03933026 (NA).

Preferably, the lung adenocarcinoma diagnostic product is a reagent or a kit.

The present application also provides a lung adenocarcinoma diagnostic product, including at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88).

Preferably, other medically acceptable agents are also included.

Preferably, the product includes cg07077277 (HIST1H2AA), cg08448665 (C21orf88), cg15435730 (TFAP2D), cg17900854 (C21orf88), cg27630311 (TBX3) and cg03933026 (NA).

In summary, this application has proved through specific verification experiments that the 12 lung adenocarcinoma-specific CpG sites provided above show abnormal changes in early lung adenocarcinoma patients, and show lung adenocarcinoma-specific methylation changes compared with lung squamous cell carcinoma, large cell lung cancer or small cell lung cancer. The diagnostic effect of the 12 CpG sites was evaluated in the TCGA lung adenocarcinoma data, and the AUC range was between 0.652-0.911, the sensitivity was between 54.9-83.1%, and the specificity was between 81.2-100%; the random forest model was constructed based on the TCGA lung adenocarcinoma data to evaluate the comprehensive diagnostic effect of the 12 CpG sites, with an accuracy of 99.7%, and verified in the GEO data set, with accuracies of 83.9% and 97.9%, respectively. The above results prove that the CpG sites provided in this application have good and stable distinguishing performance in lung adenocarcinoma tissues in different data sets.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a principal component analysis of 30 samples in Example 1 of the present invention (A) and the differential methylation analysis results of 10 pairs of lung adenocarcinoma patients' cancer and adjacent tissues (B); differentially high methylation sites (C) and abnormally low methylation sites (D) after eliminating abnormal background noise of healthy human leukocyte DNA;

FIG2 is a correlation analysis (A) and a Venn diagram (B) of the methylation levels of internal lung adenocarcinoma differentially methylated CpG sites and TCGA lung adenocarcinoma differentially methylated CpG sites in one embodiment of the present invention;

FIG3 is a lung adenocarcinoma tissue-specific methylation marker screening process (A) and the methylation levels of candidate markers in lung adenocarcinoma (B), healthy human whole blood (C) and ten other types of tumor tissue samples (D) in one embodiment of the present invention;

FIG4 shows the methylation levels of 12 CpG sites in the present invention in lung adenocarcinoma patients and adjacent tissues at different TNM stages;

FIG5 shows the methylation levels of the lung adenocarcinoma tissue-specific methylated CpG sites in lung cancer tissue of the present invention. Note: LUAD: lung adenocarcinoma; LUSC: lung squamous cell carcinoma; LCC: large cell lung cancer; SCLC: small cell lung cancer; Normal: adjacent tissue to the cancer;

FIG6 is a comparison of the diagnostic effects of 12 CpG sites in the present invention in TCGA lung adenocarcinoma data;

FIG. 7 shows the principal component analysis of TCGA and GEO datasets (A) and the AUC values of different CpG site combination models in the three datasets (B) in the present invention.

Detailed ways

The present invention is further described in detail below in conjunction with specific embodiments.

Lung adenocarcinoma diagnostic markers, including at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88).

The invention relates to the use of CpG sites as lung adenocarcinoma tissue-specific methylation markers in the preparation of lung adenocarcinoma diagnostic products, wherein the CpG sites are at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88).

Preferably, the CpG sites are a combination of six CpG sites: cg07077277 (HIST1H2AA), cg08448665 (C21orf88), cg15435730 (TFAP2D), cg17900854 (C21orf88), cg27630311 (TBX3) and cg03933026 (NA).

The above contents are explained below in combination with specific verification experiments.

Example 1: Detection of DNA whole gene methylation in lung adenocarcinoma tissue:

A. Study subjects: The first phase included 10 lung adenocarcinoma cases and 10 healthy controls. Their baseline information is shown in Table 1.

Table 1 Characteristics of the subjects included in this study

Case group: newly diagnosed lung adenocarcinoma cases from Nantong First People's Hospital. Inclusion criteria: patients with pathologically confirmed primary lung adenocarcinoma who were untreated for the first time, excluding patients with a history of malignant tumors. The 10 lung cancer cases were all newly diagnosed lung adenocarcinoma cases in Nantong First People's Hospital, Jiangsu Province from October 2022 to November 2022. All cases were confirmed by pathological histology, excluding patients with a history of malignant tumors, metastasis to other organs, and chemotherapy or radiotherapy before blood collection.

Control group: healthy people who participated in routine physical examinations. The inclusion criteria for the control group were no history of any tumors and no nodules in lung imaging tests. 10 healthy controls were randomly selected from healthy people who received routine health examinations at Nantong First People's Hospital. All controls were healthy controls with no history of respiratory diseases or malignant tumors.

B. Biological specimen collection and methylation 850K chip detection

Fresh cancer tissues and matching adjacent tissues from patients with lung adenocarcinoma were collected and stored at -80°C. 10 mL of peripheral venous blood from healthy controls was collected using an Ardent Cell-Free DNA Blood Collection Tube, centrifuged at 1900 g/3000 rpm for 10 min at room temperature to separate plasma, leukocytes, and blood cells. The plasma was centrifuged again at 16,000 g for 10 min at 4°C, and the supernatant was transferred to another clean EP tube and stored at -80°C.

DNA from lung adenocarcinoma tissues and blood leukocytes of healthy controls were extracted and subjected to whole genome methylation 850K chip detection.

Specifically, the whole genome methylation detection (850K) was performed on the matched cancer and adjacent normal tissue DNA of 10 patients with early lung adenocarcinoma and the peripheral blood leukocyte DNA of 10 healthy subjects. First, after quality control, one cancer tissue sample with abnormal results was eliminated (Figure 1A). The differential methylation analysis was performed between the cancer tissue samples of 9 patients with early lung adenocarcinoma and 10 adjacent normal tissue samples. According to the FDR less than 0.05, a total of 57,659 CpG sites with abnormal high methylation and 95,768 CpG sites with abnormal low methylation in patients with early lung adenocarcinoma were screened ( Figure 1B); for abnormally high methylated CpG sites, CpG sites that showed a high methylation background in leukocyte DNA, i.e., 49,639 CpG sites with a median methylation level greater than 0.1, were eliminated, leaving 8,020 candidate high methylated CpG sites for lung adenocarcinoma (Figure 1C); for abnormally low methylated CpG sites, CpG sites that showed a low methylation background in leukocyte DNA, i.e., 77,988 CpG sites with a median methylation level less than 0.9, were eliminated, leaving 17,781 candidate low methylated CpG sites for lung adenocarcinoma (Figure 1D).

Example 2: Screening of lung adenocarcinoma-specific methylation markers

A differential analysis was performed on the TCGA lung adenocarcinoma whole genome methylation chip data, and the results of the previous lung adenocarcinoma whole genome 850K chip were jointly analyzed to retain the candidate CpG sites of lung adenocarcinoma with consistent differences and statistical significance. The whole genome methylation chip data of ten most common malignant tumor tissues, including breast cancer, squamous cell lung carcinoma, colorectal cancer, prostate cancer, gastric cancer, liver cancer, esophageal cancer, cervical cancer, thyroid cancer and bladder cancer, as well as paracancerous tissues and healthy human blood were further combined. For the above-mentioned lung adenocarcinoma candidate CpG sites, the sites with abnormal methylation background interference in the above-mentioned ten malignant tumor tissues and healthy human whole blood were removed to construct a lung adenocarcinoma tissue-specific DNA methylation map, and then obtain lung adenocarcinoma tissue-specific methylation markers.

Based on the above candidate methylated CpG sites of lung adenocarcinoma, we further combined the methylated 450K chip data of 11 types of cancer tumor tissues including lung adenocarcinoma in the TCGA and GEO databases (Table 2) and healthy human whole blood to perform pan-cancer analysis and construct a tissue-specific methylation map of lung adenocarcinoma.

Table 2 Demographic and clinical information of GEO samples

More stringent screening conditions (|β-cancer-β-cancer-para-cancer|>0.2 and FDR<0.05) were used to screen differentially methylated sites in 455 lung adenocarcinoma tissues and 32 adjacent adjacent tissues from TCGA. The methylation levels of differentially methylated CpG sites in the two datasets were highly consistent (Figure 2A). When crossed with the above-mentioned markers, a total of 3,255 highly methylated CpG sites and 837 low-methylated CpG sites were obtained (Figure 2B).

Finally, we further eliminated CpG sites that may have potential background interference in the whole blood DNA of 1,501 healthy people and CpG sites that had abnormal methylation changes in 4,633 cases of ten other types of cancer tissues, and finally screened out 12 CpG sites that were specifically methylated in lung adenocarcinoma tissues, which were annotated to 6 known genes (Table 3, Figure 3A). The 12 lung adenocarcinoma-specific methylation sites showed specific hypermethylation or hypomethylation in lung adenocarcinoma tissues, compared with adjacent tissues (Figure 3B), healthy peripheral blood leukocytes (Figure 3C) and ten other malignant tumor tissues (Figure 3D).

Table 3 Genomic information of 12 tissue-specific methylated CpG sites in lung adenocarcinoma

Example 3: Validation of lung adenocarcinoma tissue-specific methylation diagnostic markers:

In order to preliminarily verify the effect of the markers on distinguishing lung adenocarcinoma tissues from adjacent normal tissues, the TCGA lung adenocarcinoma 450K data were used as the training set, and three GEO lung adenocarcinoma tissue methylation 450K data sets (GSE39279, GSE52401, and GSE56044) were used for external validation. ROC analysis was performed for each CpG site to calculate the optimal cutoff value (at the maximum Youden index) and AUC value. The confusion matrix was generated according to the optimal cutoff value, and the diagnostic indicators such as sensitivity, specificity, and accuracy of each CpG site were calculated. A random forest model was established using all CpG sites to evaluate the comprehensive predictive performance of the selected markers. The constructed model was internally validated using 10-fold cross validation in the training set, and externally validated using the GEO data set. Indicators such as accuracy, sensitivity, specificity, and AUC were generated to evaluate the diagnostic efficacy of the model.

Specifically, in order to further explore the methylation levels of the 12 lung adenocarcinoma-specific CpG sites in patients with lung adenocarcinoma at different TNM stages, it can be seen that the methylation levels of the 12 CpG sites showed abnormal changes in early lung adenocarcinoma patients (Figure 4); in addition, compared with lung squamous cell carcinoma, large cell lung cancer or small cell lung cancer, lung adenocarcinoma-specific methylation changes were shown (Figure 5).

The diagnostic effect of the 12 CpG sites was evaluated in the TCGA lung adenocarcinoma data, with an AUC range of 0.652-0.911, a sensitivity of 54.9-83.1%, and a specificity of 81.2-100% ( FIG. 6 ).

A random forest model was constructed based on TCGA lung adenocarcinoma data to evaluate the comprehensive diagnostic effect of the 12 CpG sites, with an accuracy of 99.7%. When validated in the GEO dataset, the accuracies were 83.9% and 97.9%, respectively (Table 4).

Table 4 Construction of random forest confusion matrix for 12 CpG sites

Note: LUAD: lung adenocarcinoma

Subsequently, recursive feature elimination (REF) based on the random forest model (10-fold cross validation repeated 5 times) was used to construct models containing from 1 to 12 CpG sites, and the best model was determined based on the AUC value (maximum). It can be seen that when only six CpG sites, cg07077277 (HIST1H2AA), cg08448665 (C21orf88), cg15435730 (TFAP2D), cg17900854 (C21orf88), cg27630311 (TBX3) and cg03933026 (NA), were included in the model, the AUC of the joint model diagnosis reached 0.985 (Figure 7). When the model was applied to three GEO datasets, the AUCs of GSE39279&GSE52401 and GSE56044 were 0.995 and 0.994, respectively. The above results show that these CpG sites have good and stable distinguishing performance in lung adenocarcinoma tissues in different data sets.

In summary, this study preliminarily explored candidate diagnostic markers for lung adenocarcinoma methylation through whole-genome methylation chip detection, and then integrated and analyzed the whole-genome methylation chip data of large samples and multiple types of cancers, and found that there are methylation markers in lung adenocarcinoma tissue that are different from lung adenocarcinoma adjacent tissues, lung squamous cell carcinoma and other tumor tissues. Therefore, based on the theory of tumor tissue-specific methylation maps, assuming that there are fragments carrying this specific methylation pattern in the DNA fragments actively or passively released by lung adenocarcinoma cancer cells to the blood circulation system, plasma cfDNA detection is expected to achieve specific tracing of lung adenocarcinoma, thereby achieving specific auxiliary diagnosis.

Claims (7)

1. A diagnostic marker for lung adenocarcinoma, characterized by comprising at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88). 2. The use of CpG sites as lung adenocarcinoma tissue-specific methylation markers in the preparation of lung adenocarcinoma diagnostic products, characterized in that the CpG site is at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88). 3. Use of the CpG site according to claim 2 as a lung adenocarcinoma tissue-specific methylation marker in the preparation of a lung adenocarcinoma diagnostic product, characterized in that the CpG site is a combination of six CpG sites: cg07077277 (HIST1H2AA), cg08448665 (C21orf88), cg15435730 (TFAP2D), cg17900854 (C21orf88), cg27630311 (TBX3) and cg03933026 (NA). 4. Use of the CpG site according to claim 2 as a lung adenocarcinoma tissue-specific methylation marker in the preparation of a lung adenocarcinoma diagnostic product, characterized in that the lung adenocarcinoma diagnostic product is a reagent or a kit. 5. A lung adenocarcinoma diagnostic product, characterized by: comprising at least one of cg24996482 (NA), cg07077277 (HIST1H2AA), cg05148722 (HIST1H2BF), cg00608169 (NA), cg15435730 (TFAP2D), cg21174344 (NA), cg15803869 (NA), cg23943136 (GATA3), cg03933026 (NA), cg27630311 (TBX3), cg08448665 (C21orf88) and cg17900854 (C21orf88). 6. A lung adenocarcinoma diagnostic product according to claim 5, characterized in that it also includes other medically acceptable reagents. 7. A lung adenocarcinoma diagnostic product according to claim 5, characterized in that the product includes cg07077277 (HIST1H2AA), cg08448665 (C21orf88), cg15435730 (TFAP2D), cg17900854 (C21orf88), cg27630311 (TBX3) and cg03933026 (NA).