A novel DNA methylation marker to identify lymph node metastasis of colorectal cancer

Lymph node metastasis (LNM) of colorectal cancer (CRC) is an important factor for both prognosis and treatment. Given the deficiencies of conventional tests, we aim to discover novel DNA methylation markers to efficiently identify LNM status of CRC. In this study, genome-wide methylation sequencing was performed in a cohort (n=30) using fresh CRC tissue to discover differentially methylated markers. These markers were subsequently validated with fluorescence quantitative PCR in a cohort (n=221), and the optimal marker was compared to conventional diagnostic methods. Meanwhile, immunohistochemistry was used to verify the effectiveness of the antibody corresponding to this marker in a cohort (n=56). LBX2 achieved an AUC of 0.87, specificity of 87.3%, sensitivity of 75.7%, and accuracy of 81.9%, which outperformed conventional methods including imaging (CT, PET-CT) with an AUC of 0.52, CA199 with an AUC of 0.58, CEA with an AUC of 0.56. LBX2 was also superior to clinicopathological indicators including the depth of tumor invasion and lymphatic invasion with an AUC of 0.61and 0.63 respectively. Moreover, the AUC of LBX2 antibody was 0.84, which was also better than these conventional methods. In conclusion, A novel methylation marker LBX2 could be used as a simple, cost-effective, and reliable diagnostic method for LNM of CRC.

Introduction

Colorectal cancer (CRC) is the third most common cancer in the world. Until now, the incidence and mortality rate have increased to the third and the second among all cancers. However, lymph node metastasis (LNM) is the main cause of the increasing mortality in CRC (1). According to National Comprehensive Cancer Network (NCCN) guidelines on the treatment of CRC, surgical operation is still the preferred treatment for CRC, meanwhile, lymph node dissection is recommended whenever there is an opportunity to remove the tumor (2). Although lymph node dissection could reduce the recurrence of CRC, the patients without LNM could not benefit from lymph node dissection, but it could bring many complications such as postoperative intestinal adhesion, intestinal obstruction, lymphatic leakage, sexual dysfunction, and postoperative bleeding, which lead to excessive medical treatment (3).

Currently, the clinical diagnosis of LNM of CRC mainly relies on imaging including computed tomography (CT) and positron emission tomography-computed tomography (PET-CT), or clinicopathological characteristics including depth of tumor invasion, ulceration, lymphatic vascular invasion, etc. (4, 5). In addition, some clinical serological indicators such as carcinoembryonic antigen (CEA) and carbohydrate antigen 199 (CA199) could also be used as a basis for LNM of CRC (6). However, the accuracy and reliability of these methods is not ideal. This may be the primary reason why the NCCN guidelines on surgical treatment of CRC recommend lymph node dissection, despite its potential for postoperative complications.

DNA methylation is one of the important epigenetic modifications. It has been proved that abnormal DNA methylation is related to cancer. During tumorigenesis, changes in DNA methylation patterns may be easily detected, thus tumor-related methylation markers have more accurate and direct effects on cancer diagnosis (7). So far, many studies on DNA methylation of CRC are based on early diagnosis and prognosis. In terms of early diagnosis, methylation of the promoter of RASSF1A (8), methylation of the CpG of Caveolin-1 (9), hypomethylation of transcription suppressor HES1 (10), hypomethylation of histone lysine methyltransferase encoding gene SMYD3 (11) are proved to be associated with CRC. In terms of prognosis assessment, methylation of the promoter of CDX2 is an independent indicator of prognosis of CRC (12), and methylation of the promoter of RAI2 is a poor indicator of prognosis of CRC (13).

Our previous work has demonstrated that the methylation markers of KCNJ12, VAV3-AS1, and EVC could be used as the basis for stage and stratification of CRC, with an area under curve (AUC) of 0.87, sensitivity of 83.0%, and specificity of 71.2% (14). Currently, it is common to take CRC tissue under colonoscopy for preoperative diagnosis in clinical practice. By obtaining CRC tissue samples, this study aims to identify LNM status of CRC by discovering novel DNA methylation markers, which could be used for the formulation of clinical treatment plans and prognosis evaluation of CRC.

Methods

Study design and patient recruitment

In this study, a three-phase strategy was designed (Figure 1), which included a marker discovery cohort (n = 30, fresh frozen (FF) tissue samples) and a marker validation cohort (n =221, FF and formalin-fixed paraffin-embedded (FFPE) tissue samples). The proportion of tumor in all tissue samples was more than 60%, which was obtained by two qualified pathologists on observing paraffin sections with high-power microscopy. Genome-wide methylation sequencing was performed on 30 FF tissue samples from the marker discovery cohort to identify LNM-specific methylation markers. These methylation markers were verified by fluorescence real-time quantitative PCR (qPCR) from the marker validation cohort (n=221). The optimal methylation markers were selected to compare with imaging (CT and PET-CT), serological indicators (CEA and CA199) and clinicopathological characteristics in a validation cohort. All CRC patients were recruited from Zhujiang Hospital, Southern Medical University. CRC samples (FF, n=76; FFPE, n=182) were derived from January 2017 to March 2020. Samples with less than twelve lymph nodes (15) and failed DNA quality control (n=37) were excluded from the study. Tissue sample of CRC patients was tumor surgical specimens before radiotherapy or chemotherapy. And these samples corresponding to pathological reports and LNM status were confirmed by at least two gastrointestinal pathologists. The clinicopathological characteristics containing age, gender, depth of invasion (t-stage of TMN), tumor size, lymphatic vessel invasion (LVI), blood vessel invasion (BVI), neural invasion (NI) and ulceration were shown in Table 1.

FIGURE 1

Figure 1 Schematic workflow of the study design.

TABLE 1

Table 1 Characteristics of CRC patients in the validation cohorts.

Discovery of methylation markers

30 FF tissue samples of CRC (19 LNM+, 11 LNM-)were collected to identify differential methylation markers. Next, we independently constructed a genome-wide methylation library using TruSeq^® Methyl Capture EPIC Library Prep Kit (Illumina, USA, Catalog No. Fc-151-1002). After EPIC library was quality-assured with Agilent High-Sensitivity DNA Kit (Agilent, USA, Catalog No. 5067-4626), high-throughput sequencing was performed on Illumina X-TEN platform.

DNA extraction, bisulfite treatment

Genomic DNA was extracted from FF tissue samples and FFPE tissue samples with AllPrep DNA/RNA Mini Kit (Qiagen, Germany, Catalog No. 80204) and AllPrep DNA/RNA FFPE Kit (Qiagen, Germany, Catalog No. 80234). Subsequently, the extracted DNA was quantified by the qubit dsDNA Customs Assay Facility (Thermal Fisher Science, USA, Catalog No.Q32851). The quality controlled criteria of CRC samples required that DNA content was more than 100 nanograms and the main band of agarose gel electrophoresis exceeded 500 bps. 50 nanograms of genomic DNA was taken from each sample, and EZ-96-DNA Methylation Direct MagPrep Kit (Zymo Research, USA, Catalog No. D5044) was used for bisulfite treatment of DNA.

Methylation analysis by fluorescence qPCR

The primer and probe sequences of the selected genes were designed through the biological software Beacon Designer V8.14. Fluorescence qPCR was used for methylation analysis in a validation cohort (n=221,103 LNM+ and 118 LNM-) (16). qPCR methylation analysis was performed on the Quant Studio 3 Real-Time PCR System (Thermo Fisher, USA). Based on our previous study (14), ACTB was selected as an internal reference gene. △Ct value obtained by fluorescence qPCR was used to indicate the methylation level of the target gene (△Ct= Ct value of the target gene - Ct value of the reference gene). If the Ct value is not present, the Ct value was set to 40.

Immunohistochemical analysis

Immunohistochemistry (IHC) was subsequently performed on the optimal genes validated by fluorescence qPCR. A total of 56 CRC paraffin sections (28 LNM+, 28 LNM-) were colllected for immunohistochemical analysis. First, the 2μm thick paraffin sections were roasted at 65°C for 1 hour, dewaxed with xylene and 100% ethanol, repaired with citrate buffer solution (PH 6.0) for 3 minutes under high pressure, and incubated with 3% H202 for 10 minutes. Next, the paraffin sections were sealed with goat serum for 30 minutes, and the primary antibody LBX2 (Bioss, Beijing, China) was diluted at 1:100 and incubated in a metal bath at 37°C for 1 hour. After washing with phosphate buffered saline PBS (PH 7.6), enzyme-labeled sheep anti-mouse/rabbit IgG polymer (Second Antibody, GeneTech, Shanghai, China) was selected to incubate at 37°C for 30 minutes. Peroxidase activity was cultured with 3, 3-diaminobenzidine hydrochloride (DAB) in sterile H2O2 solution for 2 minutes. Finally, nuclear re-staining was performed with Mayer hematoxylin solution. All the slices were independently examined by two observers. The positive composite score was used in this study, which was the staining intensity multiply the percentage of positive cells. The staining intensity is classified into four levels. No staining was rated 0 point, light yellow was rated 1 point, pale brown was rated 2 points, brown was rated 3 points (Figures 3A-D). In addition, the percentage of positive cells was evaluated as 0 points for 0 ~ 5%, 1 point for 6% ~ 25%, 2 points for 26% ~ 50%, 3 points for 51% ~ 75% and 4 points for >75%.

Comparison of DNA methylation markers and clinicopathological features, imaging, serological indicators

Eight clinicopathological variables were included in univariate analysis to explore their correlation with LNM. Variables with P value less than 0.05 were included in a multivariate analysis. Stepwise regression was used to assess 95% confidence interval (CI) of odds ratio (OR) values to identify independent predictors. DNA methylation marker was compared to the selected clinicopathological indicators, imaging and serological indicators (CEA and CA199) by the area under the receiver operating characteristic (ROC) curve including the specificity, sensitivity, accuracy, positive predictive value (PPV), and negative predictive value (NPV).

Statistical analysis

R package DSS(2.0.16) of ComplexHeatmap and Corrplot were used for unsupervised hierarchical clustering and correlation analysis, pROC (1.16.1) was used for ROC, AUC and AUC confidence interval calculations, ggplot2 (3.2.1) and RColorBrewer (1.1.2) were used for visualization of figures. Differences between 2 groups were analyzed with the unpaired Student’s t test (2-tailed tests), and 1-way ANOVA followed by Dunnett’s multiple comparisons tests when more than 2 groups were compared. Pearson’s χ2 test was used to analyze the clinical variables on sensitivity and specificity. Univariate and multivariate logistic regressions were used to evaluate the clinicopathological variables. Comparison of AUC values were conducted by Hanley and McNeil tests or DeLong test, when appropriate. The AUC values, sensitivity, specificity, and accuracy of methylation maker LBX2, clinicopathological features, serologic tumor markers and image in detecting LNM of CRC were used for comparison. A p value < 0.05 on two sides of all hypothesis tests were considered statistically significant.

Results

Genome−wide screening of DNA methylation markers to detect LNM in CRC tissue samples

A schematic workflow of the study design is shown in Figure 1. To identify LNM-specific DNA methylation markers in CRC, we first performed genome-wide methylation analysis containing 3.34 million CpG sites on fresh tissue samples from LNM+ group (n=19) and LNM- group (n=11). A total of 734 CpG sites with differential methylation were found (p<0.001, β value difference ≥0.15). The unsupervised heretical clustering showed that LNM+ and LNM- were clearly distinguished by majority of specific DNA methylation markers (Figure 2A). Based on these methylation sites, we further analyzed the differential methylation region (DMR) status and screened out twelve markers with DMR status. There were three hypomethylation markers including LBX2, SS18L1, CYTH2 in LNM + group, meanwhile, there were nine hypermethylation markers including ACHE、RPS15、APC2、BAHCC1、LEFTY1、RTN4RL2、KCNQ1、STMN3、LINC01072 in LNM+ group.

FIGURE 2

Figure 2 The discovery of DNA methylation markers to detect LNM in CRC tissue. (A) In the discovery cohort, an unsupervised hierarchical cluster analysis was based on differential methylation between LNM+(n=19) and LNM-(n=11), with each column representing a patient and each row representing a CpG marker. (B) Methylation levels of exon region of LBX2 in LNM- and LNM+ groups in CRC (C-E). The methylation level distribution of LBX2, STMN3, and SS18L1 between LNM+(n=19) and LNM-(n=11) was represented by the β value from genome-wide methylation sequencing in the discovery cohort. (F) ROC curve of three methylation markers. The effectiveness of three DNA methylation methods was evaluated by comparing AUC values. (G) The accuracy, NPV, PPV, sensitivity and specificity of three methylation marker were compared respectively. *p < 0.05, **p < 0.01 and ***p < 0.001. NS, not statistically significant.

Our primary goal was to develop a simple methylation-specific qPCR to detect LNM status (16). These twelve markers were further verified by qPCR in a cohort (n=65) (Supplemental material: Figure S2). However, nine markers were excluded due to low AUC value and inconsistent methylated patterns. Only three markers including LBX2, STMN3, SS18L1 showed higher AUC values and consistent methylated patterns in sequencing and methylation-specific qPCR analysis, and significantly differentiated LNM+ from LNM- in the same samples (Figures 2C–E). In addition, we found exons of LBX2 included significant methylation differences between LNM+ and LNM-(Figure 2B). To sum up, these results indicated that these three methylation markers and qPCR-based analysis were reliable and could be used for large-scale cohort analysis.

LBX2 had the best test performance in three target methylation makers

Through the biological software Beacon Designer V8.14, the information about three target genes (LBX2, STMN3, SS18L1) was input to set appropriate conditions. The primer and probe sequence were shown in Table 2. Based on validation of 221 tissue samples, LBX2 achieved an AUC of 0.87 (95%CI 0.82-0.92, p<0.001), specificity of 87.3%, sensitivity of 75.7%, accuracy of 81.9%. STMN3 achieved an AUC of 0.54 (95%CI 0.46-0.61, p=0.30), specificity of 85.6%, sensitivity of 26.2%, and accuracy of 57.9%. SS18L1 achieved an AUC of 0.56 (95%CI 0.48-0.63, p=0.15), specificity of 81.4%, sensitivity of 38.8%, and accuracy of 61.5%. The comparison of these three methylation markers was shown in Figure 2F, G. Obviously, LBX2 had higher efficiency in accuracy, specificity, sensitivity and AUC compared to other two methylation markers.

TABLE 2

Table 2 Designed primer and probe sequences of target genes.

LBX2 antibody could identify LNM status by IHC

IHC of LBX2 antibody was performed on CRC tissue section. These sections were amplified by 400 times, the staining of cancer cells was mainly observed. LNM- tissue sections (Figure 3E) showed light staining of cancer cells. However, LNM+ tissue section (Figure 3F) showed deep staining of cancer cells, especially in the nucleus where there was dark brown, presenting strong positive. In addition, the immunohistochemical score of LNM+ group was significantly higher than that of LNM- group (p<0.001) (Figure 3G). Comparing IHC with qPCR in a cohort (LNM+, n=28 and LNM-,n=28), the AUC of IHC was 0.84 (95%CI 0.74-0.94, p<0.001), but qPCR achieved an AUC of 0.93 (95%CI 0.89-0.97, p< 0.001) (Figure 3H). Moreover, the specificity, sensitivity and accuracy of the two methods were compared. The specificity, sensitivity and accuracy of immunohistochemical method were 92.9%, 64.3% and 78.6%, respectively. However, the qPCR method achieved a specificity of 100%, sensitivity of 82.1% and accuracy of 91.1% (Figure 3I). Obviously, both methods had good discrimination efficiency, despite the qPCR method was better than IHC.

FIGURE 3

Figure 3 The performance of antibody LBX2 to detect LNM in CRC tissue. (A–D) Immunohistochemical staining depth corresponding to the score. Stained samples (1:100 diluted concentration & 100X Magnification) were divided into four grades. (A 0=non-staining B 1=light yellow, C 2=pale brown D 3=brown). (E) Faint yellow stained samples (1:100 diluted concentration & 400X Magnification) was considered as LNM- CRC. (F) Brown staining samples (1:100 diluted concentration & 400X Magnification) was considered as LNM+ CRC. (G) The level of LBX2 antibody was compared between LNM+(n=28) and LNM-(n=28). (H) ROC curve of two methods (IHC&. qPCR). The effectiveness of these two methods was evaluated by comparing AUC values. (I) The accuracy, NPV, PPV, sensitivity and specificity of two methods (IHC&. qPCR) were compared respectively. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p<0.0001.

LBX2 showed stable table discriminative efficacy in different classification

DNA methylation marker LBX2 could identify LNM of CRC well in both male and female populations (P < 0.001) (Figure 4A). Similarly, LBX2 had good discrimination effect in the group under 55 years old or the group over 55 years old (P < 0.001) (Figure 4B). LBX2 could identify LNM of CRC well at T3 and T4 stages (P < 0.001) and T2 stage (P < 0.05). However, the strength of evidence was weak due to the small sample size at T1 stage (Figure 4C). LBX2 also could identify LNM of CRC in tumor diameter less than 5cm and tumor diameter more than 5cm (P < 0.001) (Figure 4D). In different clinicopathological groups, LBX2 could identify LNM of CRC well in both ulcerated and non-ulcerated groups, both lymphatic and non- lymphatic invasion groups, both vascular and non-vascular invasion groups, and both neural and non-neural invasion groups (P < 0.001) (Figures 4E-H). In conclusion, DNA methylation marker LBX2 had stable performance in different groups of each factor.

FIGURE 4

Figure 4 LBX2 has a good discriminative effect in different classification. (A) Performance of LBX2 in the male and female groups. (B) Performance of LBX2 in age less than 55 years and age more than 55 years groups. (C) Performance of LBX2 in t₁, t₂, t₃, t₄-staging. (D) Performance of LBX2 in tumor size less than 5cm and tumor size more than 5cm groups. (E) Performance of LBX2 in ulcerative and non- ulcerative groups. (F) Performance of LBX2 in lymph-vessel invasion and non-lymph-vessel invasion groups. (G) Performance of LBX2 in vascular invasion and non- vascular invasion groups. (H) Performance of LBX2 in nerve invasion and non-nerve invasion. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.

LBX2 was superior to clinicopathological features in distinguishing LNM

The relation between clinicopathological features and LNM status was further analyzed. Clinicopathological features included gender, age, depth of tumor invasion (t-stage of TMN), tumor size (demarcated by 5cm), ulcerative, LVI, BVI, and NI. In univariate analysis, there were four factors associated with LNM, including t-stage (OR 1.78, 95% CI 1.23-2.63, p<0.01), LVI (OR 4.52,95%CI 2.33-9.63,p<0.001), BVI (OR 3.47,95%CI 1.84-6.80, p<0.001), NI (OR 2.33,95%CI 1.34-4.09, p<0.01). (Table 3); Taking these four factors into account in multifactorial analysis, only LVI (OR 6.41,95%CI 1.31-47.66, p<0.05) and t-stage (OR 1.71,95%CI 1.15-2.61, p<0.01) were related to LNM. Therefore, among these eight clinicopathological features, LVI and t-stage were closely associated with LNM. Next, compared to LVI and t-stage, the LBX2 achieved an AUC of 0.87 (95%CI 0.82-0.92, p<0.001), accuracy of 81.9%, sensitivity of 75.7%, and specificity of 87.3%. Turning to other two clinicopathological features, the LVI achieved an AUC of 0.63 (95%CI 0.57-0.69, p<0.001), accuracy of 64.7%, sensitivity of 37.9%, and specificity of 88.1%. while the t-stage achieved an AUC of 0.61(95%CI 0.54-0.67, p<0.01), accuracy of 59.3%, sensitivity of 49.55%, and specificity of 67.8% were 59.3%, 49.5%, and 67.8%. Compared to LBX2, LBX2 was clearly superior to clinicopathological features (Figures 5A, B).

TABLE 3

Table 3 Univariate and multivariate logistic regression analyses associated with LNM.

FIGURE 5

Figure 5 The comparison of LBX2 and traditional indicators (clinical pathology, CA199, CEA, and image) to detect LNM of CRC. (A) ROC curve of three indicators (LBX2, LVI, and T). The effectiveness of three indicators was evaluated by comparing AUC values. (B) The accuracy, NPV, PPV, sensitivity and specificity of these three indicators were compared respectively. (C) ROC curve of four indicators (LBX2, CA199, CEA and image). The effectiveness of four indicators was evaluated by comparing AUC values. (D) The accuracy, NPV, PPV, sensitivity and specificity of these four indicators were compared respectively. *p < 0.05, **p < 0.01 and ***p < 0.001. NS: not statistically significant.

LBX2 was superior to CEA, CA199, imaging in distinguishing LNM

The relation between CA199, CEA, imaging and LNM status was analyzed. The AUC of CA199 was 0.58 (95%CI 0.51-0.66), with the specificity of 49.1%, the sensitivity of 70.6%, and the accuracy of 59.3%. Moreover, the AUC of CEA was 0.56 (95%CI 0.48-0.64), with the specificity of 43.4%, the sensitivity of 71.8% and the accuracy of 56.9%. In addition, the AUC of imaging (CT and PET-CT) was 0.52 (95%CI 0.45-0.59), with the specificity of 46.2%, the sensitivity of 58.2%, and the accuracy of 51.8%. Compared to these three conventional methods, the AUC of LBX2 was 0.87 (95%CI 0.82-0.92, p<0.001),with the specificity of 87.3%, the sensitivity of 75.7%, and the accuracy of 81.9%, which was better than these current clinical examination (Figures 5C, D).

Discussion

DNA methylation profiles may represent relatively stable long-term programming of the genome and underlying cellular functions, which is a reliable method of the diagnosis of cancer occurrence and progression (17). Therefore, in this study, genome-wide methylation sequencing on CRC tissues (n=30) was performed and three LNM related specific methylation markers were selected. These three methylation markers were further validated by a large retrospective cohort of 221 tissue samples. We found that a qPCR-based methylated marker LBX2 had the best discriminative performance for the diagnosis of LNM, which was superior to traditional clinicopathological features, as well as imaging, CEA, and CA199. LBX2 also had stable discriminative efficacy in different groups including age, sex, tumor size, depth of tumor invasion and clinicopathological feature. At the same time, the antibody corresponding to LBX2 also showed good performance in differentiating LNM of CRC in immunohistochemical validation. In addition, a more comprehensive approach was used to analyze CRC-associated LNM differential methylation sites, which covered more than 3.34 million CpG sites, accounting for 97.3% of the CpG islands in the genome. To date, few studies have used such a wide range of genome-wide methylation strategies to discover methylation markers for the diagnosis of LNM in CRC.

In previous similar studies, Tsuyoshi Ozawa et al. used microRNA sequencing data from the Cancer Genome Atlas (TCGA) and analyzed a five microRNAs model (MIR32, MIR181B, MIR193B, MIR195, and MIR411) that could distinguish LNM in t1-t2 CRC. This verified model achieved an AUC of 0.74 (18), lower than that of the single DNA methylation marker LBX2 (AUC 0.87) in this study. Moreover, Ailin Qu et al. found a four microRNAs model (Mir-122-5p, Mir-146B-5p, Mir-186-5p and Mir-193a-5p) related to LNM status of CRC from high-throughput sequencing data of CRC tissues (n=20), and it showed that the AUC of the four microRNAs model was 0.88 through a verification cohort (n=198) (19), which was similar to the detection efficiency of DNA methylation sites in our experiment. However, compared with RNA markers, DNA methylation markers are more stable as diagnostic biomarkers and relatively stable clinical specimens, which are not easily degraded. Therefore, it is easier to be applied to clinical practice.

Interestingly, LBX2 could be used to identify the LNM status of CRC not only in qPCR verification, but also in immunohistochemical verification, which indicated that LBX2 played a significant role in the differentiation of both molecular level and protein level. LBX2 is located on chromosome 2, which starts at 74725882 and terminates at 74726332, with a total length of 451bp and containing 34 CG. In the LNM- and LNM+ groups, there were significant differences in the methylated levels of LBX2 in exon region (Figure 2B). Due to the low methylated level of LBX2 in the LNM+ group, it would be overexpressed in the process of protein translation. On the contrary, LBX2 in the LNM- group has a high methylated level, which results in low expression in the process of protein translation. This view is well explained by our immunohistochemical verification. Subsequently, compared qPCR with IHC in 56 CRC samples, it is clearly that qPCR had better performance (AUC 0.93 vs. 0.84). Apparently, qPCR method had better differentiated efficiency because of more sophisticated level quantification from the qPCR instrument, while IHC was manually assessed and divided into only four grades according to the depth of staining. Because IHC examination is cheaper and easier to generalize, both methods could be applied flexibly for clinical practice.

Currently, there are few studies on gene LBX2 related to CRC. Some researchers have found that LBX2-AS1 could promote cell proliferation, migration and invasion through Mir-4766-5P mediated CXCL5 upregulation in gastric cancer (20). In addition, it has been reported that knockout of LBX2-AS1 in hepatoma cells could reduce its proliferation (21). Moreover, it has been proved that Zinc-finger E-box binding homeobox 1 (ZEB1) could induce upregulation of LBX2-AS1 to enhance the stability of ZEB1 and ZEB2, which could promote the migration and mesenchymal transformation of esophageal squamous cell carcinoma (22). The potential biological pathways of LBX2 upregulation remain to be proved further. It is widely accepted that LBX2 may be involved in the positive regulation of Wnt signaling pathway, which is active in the nucleus. Meanwhile, Wnt signaling pathway may be a complex protein action network, whose function not only participates normal physiological processes and embryonic development, but also induces cancer (23). Wnt signaling pathway mainly occurs in intestinal epithelial cells. Under normal conditions, colonic epithelial cells could bind secretory Frizzled related proteins (SFRP) to inhibit Wnt signaling. Once SFRP is silenced under epigenetic regulation, the Wnt signaling pathway would be activated and other molecules in the signaling pathway may mutate, which promotes cell proliferation and inactivation of cells into differentiation and results in the occurrence and invasion of tumors (24).

Although there are a few clinical methods to identify LNM of CRC, the discrimination efficiency of these methods is generally limited. In this study, it had been found that the AUC of imaging, CEA, and CA199 were only 0.52, 0.56 and 0.58, respectively. In addition, the AUC of LVI and the depth of tumor invasion in clinicopathology was 0.63 and 0.61. This may be the main reason that the CRC surgical treatment guidelines suggest we should remove intact tumor with lymph node dissection (2). In fact, the incidence of LNM in many CRC patients, especially those in t1-t2 stage, is only 16% (18). Therefore, excessive medical treatment frequently exists on many CRC patients. Since LBX2 achieved an AUC of 0.87, which is significantly superior to the current clinical diagnostic methods, meanwhile, DNA samples are more stable than RNA samples. Therefore, DNA methylation marker LBX2 is easy to be transformed into clinical application and it has the opportunity to become a novel clinical indicator for the identification of LMN of CRC.

Turning to clinical application of LBX2 in the future, the LNM status of CRC could be determined by immunohistochemical analysis or qPCR analysis of biopsy tissue obtained by colonoscopy. In addition, because CRC tumor cells are easily shed into stool and blood, we could also extract DNA of stool and ctDNA of blood and detect LNM of CRC by LBX2 probe. This makes it possible to identify LNM of CRC early by minimally invasive or noninvasive methods.

Conclusion

In conclusion, a novel DNA methylation marker LBX2 could be used as a simple, cost-effective, easy-to-implement, and reliable diagnostic method for LNM of CRC compared to traditional methods, it holds the potential to provide a better clinical diagnosis for the precise treatment of CRC.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

Ethics statement

The studies involving human participants were reviewed and approved by Ethics Committee of Zhujiang Hospital, Southern Medical University; Southern Medical University. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Author contributions

J-BF and CC conceived, designed, and directed the study. YY designed the experiments and developed the methodology. YY and WX completed experimental work. YY, JW, LX, and XL performed the analyses and interpretation of data. YY, ZL, SC, and QP acquired the patient samples and information. YY and J-BF wrote and critically reviewed the manuscript. All authors reviewed and approved the final manuscript. All authors contributed to the article and approved the submitted version.

Funding

This study was supported by Science and Technology Planning Project of Guangdong Province, China(Grant NO.2017B020226005), the 2020 Guangzhou Development Zone International Science and Technology Cooperation Project(NO:2020GH15), Scheme of Guangzhou Economic and Technological Development District for Leading Talents in Innovation and Entrepreneurship (Grant NO.2017-L152), Scheme of Guangzhou for Leading Talents in Innovation and Entrepreneurship(Grant NO.2016007), and Scheme of Guangzhou for Leading Team in Innovation(Grant NO.201909010010).

Acknowledgments

I thank the staff of the J-BF’s Laboratory in Southern Medical University and the AnchorDx R&D team (AnchorDx Medical Co., Ltd) for their excellent technical assistance. I also thank CC (Department of General Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou, China) for providing valuable clinical samples.

Conflict of interests

The authors J-BF, JW, XL, LX are current employees of AnchorDx Medical Co., Ltd or AnchorDx, Inc.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc.2022.1000823/full#supplementary-material

Abbreviations

LNM, Lymph node metastasis; CRC, colorectal cancer; IHC, Immunohistochemistry; NCCN, National Comprehensive Cancer Network; CT, Computed tomography; PET-CT, Positron emission tomography-computed tomography; CEA, Carcinoembryonic antigen; CA199, Carbohydrate antigen 199; FF, Fresh frozen; FFPE, Paraffin-embedded; qPCR, Quantitative PCR; LVI, Lymphatic vessel invasion; BVI, Blood vessel invasion;NI, Neural invasion; CI, Confidence interval; OR, Odds ratio; ROC, Receiver operating characteristic (ROC); PPV, Positive predictive value; NPV, Negative predictive value; DMR, Differential methylation region.

References

1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin (2021) 71(3):209–49. doi: 10.3322/caac.21660

PubMed Abstract | CrossRef Full Text | Google Scholar

2. Benson AB, Venook AP, Al-Hawary MM, Arain MA, Chen YJ, Ciombor KK, et al. Colon cancer, version 2.2021, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw (2021) 19(3):329–59. doi: 10.6004/jnccn.2021.0012

PubMed Abstract | CrossRef Full Text | Google Scholar

3. Nakamura T, Watanabe M. Lateral lymph node dissection for lower rectal cancer. World J Surg (2013) 37(8):1808–13. doi: 10.1007/s00268-013-2072-z

PubMed Abstract | CrossRef Full Text | Google Scholar

4. Kwak JY, Kim JS, Kim HJ, Ha HK, Yu CS, Kim JC. Diagnostic value of FDG-PET/CT for lymph node metastasis of colorectal cancer. World J Surg (2012) 36(8):1898–905. doi: 10.1007/s00268-012-1575-3

PubMed Abstract | CrossRef Full Text | Google Scholar

5. Mou S, Soetikno R, Shimoda T, Rouse R, Kaltenbach T. Pathologic predictive factors for lymph node metastasis in submucosal invasive (T1) colorectal cancer: a systematic review and meta-analysis. Surg Endosc (2013) 27(8):2692–703. doi: 10.1007/s00464-013-2835-5

PubMed Abstract | CrossRef Full Text | Google Scholar

6. Luo H, Shen K, Li B, Li R, Wang Z, Xie Z. Clinical significance and diagnostic value of serum NSE, CEA, CA19-9, CA125 and CA242 levels in colorectal cancer. Oncol Lett (2020) 20(1):742–50. doi: 10.3892/ol.2020.11633

PubMed Abstract | CrossRef Full Text | Google Scholar

7. Esteller M. Epigenetics in cancer. N Engl J Med (2008) 358(11):1148–59. doi: 10.1056/NEJMra072067

PubMed Abstract | CrossRef Full Text | Google Scholar

8. Wang HL, Zhang Y, Liu P, Zhou PY. Aberrant promoter methylation of RASSF1A gene may be correlated with colorectal carcinogenesis: A meta-analysis. Mol Biol Rep (2014) 41(6):3991–9. doi: 10.1007/s11033-014-3267-6

PubMed Abstract | CrossRef Full Text | Google Scholar

9. Lin SY, Yeh KT, Chen WT, Chen HC, Chen ST, Chang JG, et al. Promoter CpG methylation of caveolin-1 in sporadic colorectal cancer. Anticancer Res (2004) 24(3a):1645–50.

PubMed Abstract | Google Scholar

10. Wu Y, Gong L, Xu J, Mou Y, Xu X, Qian Z. The clinicopathological significance of HES1 promoter hypomethylation in patients with colorectal cancer. Onco Targets Ther (2017) 10:5827–34. doi: 10.2147/OTT.S151857

PubMed Abstract | CrossRef Full Text | Google Scholar

11. Li B, Pan R, Zhou C, Dai J, Mao Y, Chen M, et al. SMYD3 promoter hypomethylation is associated with the risk of colorectal cancer. Future Oncol (2018) 14(18):1825–34. doi: 10.2217/fon-2017-0682

PubMed Abstract | CrossRef Full Text | Google Scholar

12. Graule J, Uth K, Fischer E, Centeno I, Galván JA, Eichmann M, et al. CDX2 in colorectal cancer is an independent prognostic factor and regulated by promoter methylation and histone deacetylation in tumors of the serrated pathway. Clin Epigenet (2018) 10(1):120. doi: 10.1186/s13148-018-0548-2

CrossRef Full Text | Google Scholar

13. Yan W, Wu K, Herman JG, Xu X, Yang Y, Dai G, et al. Retinoic acid-induced 2 (RAI2) is a novel tumor suppressor, and promoter region methylation of RAI2 is a poor prognostic marker in colorectal cancer. Clin Epigenet (2018) 10:69. doi: 10.1186/s13148-018-0501-4

CrossRef Full Text | Google Scholar

14. Zhang X, Wan S, Yu Y, Ruan W, Wang H, Xu L, et al. Identifying potential DNA methylation markers in early-stage colorectal cancer. Genomics (2020) 112(5):3365–73. doi: 10.1016/j.ygeno.2020.06.007

PubMed Abstract | CrossRef Full Text | Google Scholar

15. Bilimoria KY, Bentrem DJ, Stewart AK, Talamonti MS, Winchester DP, Russell TR, et al. Lymph node evaluation as a colon cancer quality measure: a national hospital report card. J Natl Cancer Inst (2008) 100(18):1310–7. doi: 10.1093/jnci/djn293

PubMed Abstract | CrossRef Full Text | Google Scholar

16. Li L, Ye Z, Yang S, Yang H, Jin J, Zhu Y, et al. Diagnosis of pulmonary nodules by DNA methylation analysis in bronchoalveolar lavage fluids. Clin Epigenet (2021) 13(1):185. doi: 10.1186/s13148-021-01163-w

CrossRef Full Text | Google Scholar

17. Berdasco M, Esteller M. Clinical epigenetics: seizing opportunities for translation. Nat Rev Genet (2019) 20(2):109–27. doi: 10.1038/s41576-018-0074-2

PubMed Abstract | CrossRef Full Text | Google Scholar

18. Ozawa T, Kandimalla R, Gao F, Nozawa H, Hata K, Nagata H, et al. A MicroRNA signature associated with metastasis of T1 colorectal cancers to lymph nodes. Gastroenterology (2018) 154(4):844–48.e7. doi: 10.1053/j.gastro.2017.11.275

PubMed Abstract | CrossRef Full Text | Google Scholar

19. Qu A, Yang Y, Zhang X, Wang W, Liu Y, Zheng G, et al. Development of a preoperative prediction nomogram for lymph node metastasis in colorectal cancer based on a novel serum miRNA signature and CT scans. EBioMedicine (2018) 37:125–33. doi: 10.1016/j.ebiom.2018.09.052

PubMed Abstract | CrossRef Full Text | Google Scholar

20. Peng L, Chen Z, Wang G, Tian S, Kong S, Xu T, et al. Long noncoding RNA LBX2-AS1-modulated miR-4766-5p regulates gastric cancer development through targeting CXCL5. Cancer Cell Int (2020) 20:497. doi: 10.1186/s12935-020-01579-z

PubMed Abstract | CrossRef Full Text | Google Scholar

21. Wang Y, Zhao Y, Zhang X, Zhang A, Ma J. Long noncoding RNA LBX2-AS1 drives the progression of hepatocellular carcinoma by sponging microRNA-384 and thereby positively regulating IRS1 expression. Pathol Res Pract (2020) 216(4):152903. doi: 10.1016/j.prp.2020.152903

PubMed Abstract | CrossRef Full Text | Google Scholar

22. Zhang Y, Chen W, Pan T, Wang H, Zhang Y, Li C. LBX2-AS1 is activated by ZEB1 and promotes the development of esophageal squamous cell carcinoma by interacting with HNRNPC to enhance the stability of ZEB1 and ZEB2 mRNAs. Biochem Biophys Res Commun (2019) 511(3):566–72. doi: 10.1016/j.bbrc.2019.02.079

PubMed Abstract | CrossRef Full Text | Google Scholar

23. Lie DC, Colamarino SA, Song HJ, Désiré L, Mira H, Consiglio A, et al. Wnt signalling regulates adult hippocampal neurogenesis. Nature (2005) 437(7063):1370–5. doi: 10.1038/nature04108

PubMed Abstract | CrossRef Full Text | Google Scholar

24. Taciak B, Pruszynska I, Kiraga L, Bialasek M, Krol M. Wnt signaling pathway in development and cancer. J Physiol Pharmacol (2018) 69(2). doi: 10.26402/jpp.2018.2.07

PubMed Abstract | CrossRef Full Text | Google Scholar