KMT2C promoter methylation in plasma‐circulating tumor DNA is a prognostic biomarker in non‐small cell lung cancer

Liquid biopsies are emerging noninvasive diagnostic tools in cancer. MLL3 histone methyltransferase is a transcriptional regulator encoded by the tumor suppressor KMT2C gene. We evaluated the prognostic significance of KMT2C epigenetic changes in plasma of NSCLC patients. The detection of KMT2C promoter methylation in cfDNA predicts poor outcomes in NSCLC and merits further evaluation as a circulating epigenetic biomarker.


Introduction
Lung cancer remains one of the most common cancers and lethal diseases in the world [1]. Non-small cell lung carcinoma (NSCLC), the major subtype of lung cancer, can be subdivided into three subtypes: lung adenocarcinoma (ADC), squamous cell carcinoma (SCC), and large-cell carcinoma [2]. The detection of NSCLC at an early stage can favor prognosis; however, there is still a lack of good prognostic biomarkers that could predict clinical outcome. A significant number of actionable genetic driver mutations and translocations have been revealed for targeted therapy for NSCLC patients [3][4][5].
Nowadays, noninvasive or minimally invasive diagnostic tools based on liquid biopsy have significantly contributed to the successful management of lung cancer patients [6]. Liquid biopsy provides a noninvasive way for real-time monitoring of tumor evolution, and therapeutic efficacy, as well as an important insight into tumor heterogeneity [7]. In addition, liquid biopsy has recently revealed important biological mechanisms driving metastatic initiator cells in lung cancer [8][9][10]. Cell-free DNA (cfDNA) analysis is already being applied in routine clinical practice for the detection of EGFR mutations in plasma samples of NSCLC patients to stratify those patients who may be benefited from targeted therapies [11,12].
Beyond gene mutations, recent advances in lung cancer epigenetic research effectively provide a promising step toward the discovery of novel epigenetic biomarkers in NSCLC, since DNA methylation alterations are frequently observed and have diverse implications in carcinogenesis, diagnosis, and prediction [13][14][15]. DNA methylation is an ideal source of candidate biomarkers due to its stability, its virtually universal presence, and amenability to measurement [13,16]. Methylation of cancer-related genes in lung cancer patients has already been detected in various biological samples, including bronchoscopic washings/ brushings, sputum samples, plasma, and serum [13]. Analysis of DNA methylation markers in biological fluids of lung cancer patients could lead to the discovery of putative epigenetic biomarkers for diagnosis, prognosis, risk assessment, and disease monitoring [17,18]. We were the first to report epigenetic changes in circulating tumor cells (CTCs) [17,19,20] and to demonstrate a close correlation between DNA methylation in CTCs and paired plasma cfDNA [17,21]. We have also shown the prognostic value of BRMS1 and SOX17 in plasma cfDNA from NSCLC patients [22,23].
The KMT2 family [also known as mixed-lineage leukemia (MLL)] has important functions in transcriptional regulation; KMT2 enzymes bind to several promoters, enhancers, and other regulatory regions [24].
Specifically, the functionally redundant H3K4me1/2 methyltransferases MLL3 (KMT2C) and MLL4 (KMT2D) are enriched on both silent and active enhancers [25]. A recent study indicated that both these enzymes facilitate enhancer RNA synthesis and transcription from promoters independently of H3K4 monomethylation [26]. In normal cells, both MLL3 and MLL4 are essential for enhancer activation and cell-type-specific gene expression during cell differentiation [25,27]. MLL3/MLL4 is contributed to the formation of super-enhancers (SEs) by binding CBP/ p300 and that enhancer priming sequentially shapes the dynamic enhancer landscapes during adipogenesis [28].
A past study from Hu et al. in human HCT116 cells showed that 70% of putative enhancers present significant loss in H3K4me1 levels in MLL3 Dset /4 Dset cells. Gene ontology (GO) analysis of genes proximal to the MLL3/4-specific enhancers revealed that many of these genes are implicated in intracellular signaling, whereas MLL3/4-independent enhancers revealed include genes that are implicated in intracellular signaling [29]. In a more recent study from Rampias et al., KMT2C was found to bind to enhancers proximally located to 2808 genes, many of which were involved in cell adherence to the epithelial basement membrane (ITGB1, ITGB6, RHOB, MMP7); extracellular matrix organization (LOXL2, LOXL4, TIMP4); and epithelial development and differentiation (SMAD6, SOX2, EREG, WNT11, BMP2) [30]. KMT2C silencing led to significant H3K27ac loss in these enhancers and increased binding by the transcription factors JUNB, TEAD, RUNX1, and MAFA [30].
High KMT2C mutational burden in several epithelial cancers implies a role as a tumor suppressor. Exome sequencing of human tumors has revealed an increased number of disease-specific mutations in the KMT2/MLL gene family that places them among the most frequently mutated genes in human cancer [24]. Up to 60% of known KMT2C mutations have been specifically identified in carcinomas of the lungs, large intestine, breast [31], uterus, and urinary tract [32]. Recently, it has been shown that KMT2C genetic polymorphisms were also associated with larynx cancer in a Chinese population [33]. Many studies have also suggested that KMT2C expression could be used as a prognostic biomarker in gastric [34], pancreatic [35], and ER + breast cancer [36]. Finally, a role of KMT2C as a 7q haplo-insufficient tumor suppressor gene has been proposed in acute myeloid leukemia in mice, by impairing the differentiation of HSPC and cooperating with other established AML lesions [37].
A meta-analysis of RNA-seq data from The Cancer Genome Atlas (TCGA) Consortium has shown that KMT2C expression is downregulated in NSCLC in comparison with normal tissue, and reduced KMT2C expression in patients with NSCLC was correlated with DNA hypermethylation within a CpG island (chr7:152435133-152437025, assembly GRCh38/hg38, ENCODE) of the KMT2C proximal promoter [30].
In this study, we evaluated for the first time the prognostic significance of KMT2C promoter methylation in NSCLC. We analyzed KMT2C promoter methylation in DNA samples from fresh-frozen NSCLC primary tissues, matched adjacent noncancerous tissues and plasma cfDNA from operable NSCLC patients, plasma cfDNA from metastatic NSCLC patients, and healthy individuals. Our data strongly suggest that the detection of KMT2C promoter methylation in plasma cfDNA predicts poor prognosis in patients with both operable and metastatic NSCLCs.

Materials and methods
The workflow of the study is shown in Fig. 1.

Patients
Blood samples were collected from two different cohorts of patients with histologically confirmed NSCLC: (a) a group consisting of 48 patients with operable (early stage: IA-IIIA) and (b) a group consisting of 91 patients with metastatic NSCLC (stage IV). Moreover, a cohort of 60 age-matched male and female healthy individuals was used as control group. In this group, 48 fresh-frozen tissues, corresponding adjacent non-neoplastic tissues, and matched plasma samples (collected before surgery) were analyzed. In the metastatic NSCLC group, plasma was collected before the initiation of first-line chemotherapy. All tissue specimens and plasma samples were collected randomly and not on purpose to balance the smoking status. The study was conducted in accordance with the 1964 Declaration of Helsinki and was approved by the ethics and scientific committee of 'Sotiria' General Hospital for Chest Diseases, Athens. All participating patients signed an informed consent form to participate in the study. The collection of samples was performed between 2004 and 2006. In the operable NSCLC group, samples were collected from 37 men and 11 women (median age: 61 years), all diagnosed with operable (stages I-III) NSCLC; 28 patients were diagnosed with ADC, 19 with SCC, and one with undifferentiated NSCLC; in this group, the majority of patients (60.4%) were smokers and suffered from mild-to-moderate chronic obstructive pulmonary disease according to pulmonary function tests that were included as part of the standard preoperative evaluation of the patients. Moreover, all patients received adjuvant platinum-based chemotherapy according to the national guidelines. The main patients' characteristics are summarized in Table 1.

Sample preparation
To avoid PCR contamination, different rooms, dedicated labware, and dedicated areas were used for all procedures. All DNA preparation and handling steps took place in specific laminar-flow hoods under DNase-free conditions.

Analytical specificity
To evaluate the analytical specificity of the KMT2C real-time MSP, we initially tested the primers in silico and then in PCR, using gDNA (unconverted DNA) and SB-modified human placental gDNA samples that were not methylated; no amplification of the KMT2C was observed. In contrast, amplification was observed only when SB-treated DNA from the 100% methylated standard was used. The developed assay is highly specific since it can detect only SB-treated KMT2Cmethylated sequences [17,19,[21][22][23].

Analytical sensitivity
The analytical sensitivity of the developed real-time MSP was evaluated by using synthetic mixtures based on serial dilutions of SB-converted DNA control samples (0% and 100% methylated) at various percentages of methylation (0.1%, 1%, 10%, 30%, and 50%

Diagnostic specificity
The diagnostic specificity of the developed assay was evaluated by using sixty plasma samples from healthy donors (HDs). cfDNA was isolated from all these plasma samples, and the KMT2C methylation status was assessed by the developed specific and sensitive real-time MSP assay; none of the HD plasma samples were methylated at the KMT2C promoter (0/60, 0%).

Quality control
We have recently evaluated the effect of pre-analytical conditions on DNA methylation analyses in liquid biopsies, such as the time interval to plasma isolation, SB-converted DNA storage conditions, and the validity of whole-genome amplification for SB-converted DNA [16]. Quality control checks were performed in all steps prior to sample analysis. In each step of the analytical procedure, we included appropriate positive and negative controls in order to ensure the quality and reproducibility of results. Before proceeding to the SB treatment and real-time MSP, we assessed the integrity of gDNA in all samples by amplifying exon 20 of PIK3CA [38]. Only samples that were positive for PIK3CA exon 20 amplification were further processed to SB treatment.

Operable NSCLC
The methylation status of KMT2C promoter was first evaluated in 48 NSCLC fresh-frozen primary tumors and their adjacent non-neoplastic tissues. KMT2C was found methylated in 3/48 (6.3%) fresh-frozen tissues and in 3/48 (6.3%) of the corresponding adjacent noncancerous tissues. In 2/3 cases, KMT2C was methylated in both primary tumor and corresponding adjacent noncancerous tissue, while one sample was positive only in the primary tumor but not in the corresponding adjacent tissue, while another sample was positive for KMT2C methylation only in the adjacent tissue but not in the paired primary tumor. We further evaluated KMT2C methylation status in matched plasma cfDNA samples; KMT2C methylation was observed in 7/48 (14.6%) plasma samples. Fisher's exact test revealed a statistically significant correlation between KMT2C promoter methylation and NSCLC stage as well as the smoking status of patients (P < 0.001 and P = 0.034, respectively; Table 1). After To rule out the effect of advanced stage in predicting worse outcomes independently of KMT2C promoter methylation, we separated the unmethylated operable NSCLC group into early (N = 28, stages I-II) and advanced (N = 6, stages III-IV) patients and performed additional survival analysis. Although the end numbers of patients are low, our analysis revealed that stage on itself is insufficient to predict survival in the operable NSCLC cohort for both OS and DFS (P = 0.560 and P = 0.519, respectively; Table 1, Fig. S1). Lastly, multivariate Cox regression analysis including KMT2C methylation, LN status, smoking status, age, stage, tumor size, and gender in the operable NSCLC group showed that stage is not associated with a higher risk of progression or death, while KMT2C promoter methylation can potentially predict higher risk of death (results not shown). These findings further support our hypothesis that KMT2C promoter methylation in plasma can be used as a circulating prognostic biomarker, predictive of worse outcomes in NSCLC.

Metastatic NSCLC
Τhe prognostic significance of KMT2C promoter methylation in plasma cfDNA was subsequently evaluated in an independent cohort of 91 patients with metastatic NSCLC after a long-term follow-up. In this group, we found that KMT2C promoter was methylated in plasma-derived cfDNA in 18/91(19.8%) cases. More specifically, after a median follow-up of 52 months (range: 1-112 months), 87/91 (95.6%) patients relapsed and 83/91 (91.2%) patients died due to disease progression. All plasma samples that were found positive for KMT2C promoter methylation belonged to the group of patients who relapsed and died. Kaplan-Meier estimates of the cumulative PFS for patients with methylated and nonmethylated KMT2C promoter demonstrated a statistical significance (P = 0.004, log-rank test; Fig. 4A). In parallel, the Kaplan-Meier estimates of the cumulative OS were also significantly different between the two groups (P < 0.001, log-rank test; Fig. 4B). In addition, the univariate Cox regression analysis revealed that both PFS and OS were significantly different in patients with KMT2C promoter methylation (HR: 0.431, 95% CI: 0.239-0.779, P = 0.005, and HR: 0.306, 95% CI: 0.173-0.541, P < 0.001, respectively).

Discussion
Circulating tumor DNA provides a real-time monitoring of tumor progression and is a unique and minimally invasive way to monitor response of solid tumors to anticancer therapies [6]. Plasma cfDNA analysis is appealing due to the ease with which plasma can be collected and analyzed without prior need of enrichment or isolation of a rare cell population [7]. Beyond mutation analysis, DNA methylation prognostic biomarkers still need to be assessed in liquid biopsy analyses [39]. DNA methylation is considered as an early event in carcinogenesis; thus the detection of tumor-specific epigenetic alterations in plasma can be used as a screening and/or diagnostic tool in a noninvasive and cost-effective way. However, large-scale validation of DNA methylation biomarkers is necessary to boost their implementation into clinical practice [40]. A recent successful bench-to-bedside application of DNA methylation analysis is the first blood test based on SEPT9 promoter methylation that is an FDA-approved assay for screening and diagnosis of colorectal cancer [41].
It has been recently shown that KMT2C is downregulated in NSCLC compared with normal tissues (TCGA data) and that this transcriptional downregulation may be due to its promoter methylation [30]. Based on these very recent findings, we evaluated the clinical relevance of KMT2C promoter methylation, as a circulating epigenetic biomarker in plasma cfDNA from NSCLC patients.
Recent data have shown that KMT2C is highly mutated on macrodissected lung carcinoids [42]. It has also been recently reported that KMT2C is frequently mutated in certain populations with diffuse-type gastric adenocarcinoma and that KMT2C loss promotes epithelial-to-mesenchymal transition (EMT) and is associated with worse OS [43]. According to this study, KMT2C knockdown promoted EMT and was accompanied by increased expression of N-cadherin and Slug. This increased migration and invasion in in vitro by 47-to 88-fold [43]. These results strongly support our findings, since KMT2C loss could result not only through mutations, but DNA methylation as well. Our results, showing a prognostic significance of KMT2C promoter methylation in NSCLC, could be possibly explained by the promotion of EMT in these cases.
Epigenetic modifiers are frequently inactivated due to loss-of-function mutations and are promising drug targets for lung cancer treatment. In mouse models, lungspecific loss of Kmt2d promotes lung tumorigenesis and upregulates pathways such as glycolysis [44]. To this end, subsequent pharmacological inhibition of glycolysis suppressed tumor growth in vitro and in vivo [44]. The detection of low KMT2C activity, due to promoter methylation, could specifically have important therapeutic implications. Epigenetic state is altered in bladder cancer cells with low KMT2C activity, followed by the decreased expression of DNA damage/repair genes [30]. Therefore, cancer cells with low KMT2C activity may be effectively treated with PARP inhibitors that have synthetic lethal effects [45]. Moreover, MLL3 encoded by KMT2C enhances the transcription of PD-L1 and regulates antitumor immunity in mouse xenografts. According to this study, MLL3 interacts with an important immune checkpoint and gives prominence to a hidden therapeutic target [46].
Our findings clearly indicate that the detection of aberrant methylation of the KMT2C promoter in plasma-derived cfDNA is related to reduced DFS and OS in operable NSCLC and reduced PFS and OS in advanced NSCLC. We report a lack of concordance of KMT2C promoter methylation in paired primary tumors, their adjacent noncancerous tissues, and corresponding plasma samples at an early stage of NSCLC. In the same group of patients, we observed a higher frequency for KMT2C methylation in plasma-derived cfDNA compared with the corresponding tumor tissues. This could possibly be explained in the context of tumor tissue heterogeneity that is a limitation factor for a representative genomic analysis in primary tissues; tumor cells within the same tumor show distinct phenotypes attributed to genetic and epigenetic alterations. On the other hand, analysis in plasma can reflect tumor heterogeneity, since DNA from different sources is shed in plasma, thus recapitulating tissue's molecular characteristics. In our advanced NSCLC cohort, the Kaplan-Meier analysis has shown a strong correlation between KMT2C promoter methylation in plasma cfDNA and poor PFS and OS, which was also verified by the univariate Cox regression analysis.

Conclusions
To our knowledge, this is the first time that KMT2C methylation status is evaluated in NSCLC. Based on the aforementioned, the detection of KMT2C promoter methylation in plasma cfDNA provides important prognostic information for NSCLC patients at both early and advanced stages, whereas it is not detected in plasma of healthy individuals. We strongly believe that KMT2C promoter methylation in plasma cfDNA merits to be further evaluated and validated as a noninvasive circulating tumor biomarker in a large and well-defined patient cohort.