Development of a sense-antisense and dual-minor groove binder probe PCR assay in plasma for early detection of colorectal cancer

To the Editor: Since the introduction of Cologuard,[1] which targets hemoglobin, KRAS mutations, and the methylation of two genes in stool DNA, methods of colorectal cancer (CRC) screening based on stool DNA methylation have received increasing attention. In comparison to stool-based assays, tests based on blood cell-free DNA (cfDNA) methylation may be more readily accepted and better suited for clinical scenarios, although they are currently less widely available. Septin9 represents a promising initial approach; however, due to the limitations of insufficient early sensitivity, it is now not recommended by the American College of Gastroenterology (ACG) guidelines.[2] The development of blood methylation tests is hindered by several significant challenges, including the low concentration of cfDNA, issues with tissue traceability, and leukocyte interference. One potential solution to these issues is the identification of more promising markers, which are likely to be entirely new targets, distinct from stool markers due to the different mechanisms by which cancer signals are present in these sample types. Furthermore, the advancement of sensitive detection techniques capable of identifying small amounts of cfDNA is crucial for the success of these tests. In this study, we identified early CRC methylation markers in three steps via 17 public 450K/850K microarray datasets [Supplementary Table 1, https://links.lww.com/CM9/C390]. The 17 datasets comprised samples of tissue, healthy whole blood cells (WBCs), and plasma cfDNA. Step 1 involved the identification of differentially methylated CpG (DMC) loci from 14 Gene Expression Omnibus (GEO) datasets comprising CRC, adenoma, and normal adjacent tissue (NAT) samples. Step 2 involved the further filtering of DMCs from Step 1 through the application of cancer and NAT samples from 31 distinct cancer types sourced from The Cancer Genome Atlas (TCGA) database. The three datasets utilized in Step 3 were employed for the validation of the identified candidate markers [Figure 1A]. In Step 1, a total of 2238 DMCs were identified as being shared between the cancer vs. normal and the adenoma vs. normal comparisons [Figure 1A and Supplementary Figure 1A, https://links.lww.com/CM9/C390]. Among these genes, 75 were found to exhibit hypomethylated in 30 non-CRC cancers from TCGA [Figure 1A and Supplementary Figure 1B, https://links.lww.com/CM9/C390]. Subsequently, the number was reduced to eight through the application of LASSO regression [Figure 1A and Supplementary Figure 1C, https://links.lww.com/CM9/C390]. Two of these, cg14015706 and cg08247376, located in the first exon of NTMT1 and 200 base pairs upstream of the MAP3K14-AS1 transcription start site, respectively, were identified as suitable for primer/probe design. Additionally, these two probes exhibited hypermethylation in all three public validation dataset [Supplementary Figure 1D–F, https://links.lww.com/CM9/C390], thus qualifying them as potential candidate markers.Figure 1: The workflow and findings of this study. (A) The process of identifying potential DMCs from 17 public databases by applying the specified filtering criteria. The left column depicts the datasets utilized in each filtering step, the middle column illustrates the filtering conditions employed in each filtering step, and the right column depicts the number of candidate DMCs filtered. (B) Schematic representation of the SADMP technique. In the case of the NTMT1 gene, two pairs of primers and their corresponding MGB probes were designed in accordance with the sense and antisense BS DNA sequences and the CBS DNA. In the case of the MAP3K14-AS1 gene, one pair of primers and two MGB probes were designed in accordance with the antisense BS DNA and the CBS DNA. (C) The ROC curves of NTMT1 and MAP3K14-AS1 are illustrated using CRC + AA vs. control groups in the training set of plasma. (D) Sensitivity and specificity of NTMT1 and MAP3K14-AS1 when detected individually and jointly in the training set, validation set, and total dataset, and 95% confidence intervals for sensitivity and specificity. (E) The sensitivity of NTMT1 and MAP3K14-AS1 methylation in non-AA and AA when detected individually and in combination is shown. (F) The sensitivity of NTMT1 and MAP3K14-AS1 methylation in different stages of CRC when detected individually and in combination. (G) The specificity of NTMT1 and MAP3K14-AS1 methylation in healthy, NDD, ID, and polyp samples when tested alone and in combination. Error bars in (E–G) represent 95% confidence interval. AA: Advanced adenoma; BS: Bisulfite-converted; CBS: Complementary sequence of sense/antisense BS; CRC: Colorectal cancer; DMC: Differentially methylated CpG; FDR: False discovery rate; ID: Intestinal disease; LASSO: Least absolute shrinkage and selection operator; MGB: Minor groove binder; NATs: Normal adjacent tissues; NDD: Non-digestive disease; non-AA: Non-advanced adenoma; ROC: Receiver operating characteristic; SADMP: Sense-antisense and dual-MGB probe.To ascertain the clinical relevance of these two candidate markers, their methylation status was analyzed in 60 tissue and 709 plasma samples obtained from the First Affiliated Hospital of Zhengzhou University. This study was conducted in accordance with the ethical standards set forth by the institutional review board (IRB) and has been approved by the IRB of the First Affiliated Hospital of Zhengzhou University (No. 2022-KY-0631-002). Each participant signed an informed consent form. This study is a sub-project of the Clinical Study of Pan-cancer DNA Methylation Test in plasma (Clinical Trials, ID: NCT05685524). Sanger sequencing was conducted on the target regions of NTMT1 and MAP3K14-AS1 following bisulfite treatment to confirm their methylation status in 30 CRC tissue samples and 30 NATs. The sequences of the primers utilized are presented in Supplementary Table 2, https://links.lww.com/CM9/C390. The findings demonstrated that the CpG sites within the amplicons exhibited a higher frequency of methylation in CRC samples compared to NATs [Supplementary Figure 2, https://links.lww.com/CM9/C390]. To enhance the capacity of candidate markers to discern low-abundance cfDNA methylation signals in plasma, we endeavored to implement dual-strand and dual-minor groove binder (MGB) probe polymerase chain reaction (PCR) methodologies, which we designate the sense-antisense and dual-MGB probe (SADMP) technique. For the NTMT1 gene, two sets of primers and probes were employed to amplify the sense and antisense strands of the amplicon region (after bisulfite transformation) simultaneously. For MAP3K14-AS1, a pair of primers and two probes were used to amplify the antisense strand of the amplicon region [Figure 1B]. The standard curves demonstrated that the amplification efficiencies of these two genes were 98% and 110%, respectively [Supplementary Figure 3A, D, https://links.lww.com/CM9/C390], which was deemed satisfactory (within the range of 95–110%). In accordance with the theoretical expectations, the SADMP technique permitted the cycle threshold (Ct) values of both genes to be advanced by one cycle in comparison to the amplification of a single strand or the use of a single probe [Supplementary Figure 3B, E, https://links.lww.com/CM9/C390]. Furthermore, the SADMP tests for NTMT1 and MAP3K14-AS1 are highly sensitive. No amplification curves were observed for either target when no methylated DNA was present and a high proportion of unmethylated DNA (107 copies/μL) was present. Additionally, they can stably detect methylated DNA at a concentration of 10 copies/μL and 5 copies/μL, respectively [Supplementary Figure 3C, F, https://links.lww.com/CM9/C390]. Subsequently, the two biomarkers were validated in 2 mL of plasma from 709 participants via the SADMP method. Two milliliters of plasma-derived cfDNA underwent bisulfite transformation and were subsequently utilized for methylation-specific PCR (MSP). The locations and sequences of the MSP primers and probes are provided in Supplementary Table 3, https://links.lww.com/CM9/C390 for reference. The composition of the PCR solution is presented in Supplementary Table 4, https://links.lww.com/CM9/C390. The plasma samples were derived from individuals with CRC, advanced adenoma (AA), non-advanced adenoma (non-AA), polyp, intestinal disease (ID), non-digestive disease (NDD), non-CRC cancer, and healthy subjects. The demographic and clinicopathologic data for the 709 subjects are presented in Supplementary Table 5, https://links.lww.com/CM9/C390. The entire dataset was randomly divided into a training set and a validation set at a ratio of 2: 1. Initially, the training set was employed to ascertain the cutoff values of the Ct values for the two genes through receiver operating characteristic (ROC) curve analysis, with CRC and AA designated as positive samples and the remaining samples as negative samples [Figure 1C]. The cutoff values for the NTMT1 and MAP3K14-AS1 genes were determined to be 48.2 and 48.4, respectively. When the two genes are tested in combination, a sample is considered to exhibit methylation if either gene is positive or if both genes are positive. In accordance with the aforementioned interpretation criteria, the sensitivities of the dual-target methylation assay in the training set, validation set, and total dataset were 86.4%, 81.8%, and 84.8%, respectively, which were significantly greater than those of either single-target assay [Figure 1D]. The specificities were 91.0% (NTMT1), 92.5% (MAP3K14-AS1), and 91.5% (dual-target), respectively. The overall sensitivity did not differ significantly among CRCs of different ages, sexes, locations, and pathological types [Supplementary Table 6, https://links.lww.com/CM9/C390]. The dual-target assay demonstrated efficacy in the detection of early neoplasms, with sensitivities of 32.0%, 75.0%, and 81.2% for AA, stage I CRC, and stage II CRC, respectively [Figure 1E, F]. A head-to-head comparative study of NTMT1/MAP3K14-AS1 and Septin9 methylation assays was conducted on 120 CRC samples. The results demonstrated that the sensitivity of NTMT1/MAP3K14-AS1 was significantly greater than that of Septin9 (85.8% vs. 61.7%) [Supplementary Table 6, https://links.lww.com/CM9/C390]. The dual-target test exhibited the highest specificity (95.9%) in healthy samples, demonstrating consistent performance across different subgroups, including healthy individuals, NDD, ID, and polyp samples [Figure 1G]. This suggests that the test is resistant to interference from various benign conditions. Previous authors have reported that C9orf50[3] and MAP3K14-AS1[4] alone in CRC diagnosis and screening. In this study, we found that the sensitivity significantly increased when they were tested in combination, exceeding that of traditional septin9 detection. The plasma NTMT1/MAP3K14-AS1 dual-targeted methylation assay is expected to become a new tool for non-invasive screening of CRC. Funding This study was supported by Science and Technology Project of Henan Province (No. 232102310028). Conflicts of interest None.