Genome-Wide Methylation Profiles of Primary and Matched Distant Metastasis: Insights from the Dutch Early-Stage Melanoma (D-Esmel) Study

Genome-Wide Methylation Profiles of Primary and Matched Distant Metastasis: Insights from the Dutch Early-Stage Melanoma (D-Esmel) Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Genome-Wide Methylation Profiles of Primary and Matched Distant Metastasis: Insights from the Dutch Early-Stage Melanoma (D-Esmel) Study Jasper Ouwerkerk, Thamila Kerkour, Antien Mooyaart, Ruben Boers, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7642695/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Dec, 2025 Read the published version in Human Genomics → Version 1 posted 12 You are reading this latest preprint version Abstract Background Early-stage (stage I-II) cutaneous melanoma accounts for the majority of melanoma diagnoses, but more than 40% of patients who die due to melanoma were initially diagnosed with an early-stage melanoma. Methods The aim of this study was to identify prognostic genome-wide methylation markers of metastasized primary early-stage melanomas and retrieving biological insights from its matched distant metastasis. We selected samples from the Dutch Early-Stage Melanoma (D-ESMEL) study, representing case-control sets where the primary melanoma of each metastatic case is matched to a primary melanoma of a control based on known clinical risk factors. Matched distant metastasis were also retrieved. Laser capture microdissection was performed to isolate the tumor tissue, where after a genome-wide methylated DNA sequencing (MeD-seq) was conducted. Differentially methylated regions (DMR) between primary tumors of the cases-control sets and the tumor of the primary case and its metastasis were tested using Chi-squared test with a genome-wide sliding window analysis, as well as a paired t-test in predefined promotor, gene body, and CpG-island regions. Results MeD-seq analyses did not reveal prognostic methylation markers in primary melanomas, which have additional prognostic value on top of known clinical risk factors We identified eight protein coding genes with the largest methylation difference between primary melanomas of patients with and without metastasis and between primary melanomas and matched distant metastasis: CYP2E1 , PTPRN2 , CHCHD2, NDRG2 , EDN2 , GC , USP17L1 , and SERPINB8. Conclusion This study found 8 genes that have been implicated in primary tumors or metastasis of other cancers which require further investigation into their involvement of metastasis in melanoma. Melanoma Metastasis Methylation Genome-wide D-ESMEL Figures Figure 1 Figure 2 Figure 3 Introduction Early-stage (stage I-II) cutaneous melanoma accounts for the majority of melanoma diagnoses and is generally associated with a favorable outcome. However, more than 40% of patients who die due to melanoma were initially diagnosed with an early-stage melanoma [ 1 – 3 ]. Therefore, current Tumor Node Metastasis (TNM) staging, which is based on Breslow thickness and ulceration, is not sufficient to identify these high-risk patients. With the availability of adjuvant therapy, there is a high need for a precise biomarker which can identify patients at risk for metastasis who may benefit from these therapies [ 4 , 5 ]. A potential biomarker for a prognostic model could be DNA methylation, which is well known to affect gene expression and pathways shown to be related to response to treatment [ 6 ] and metastasis [ 7 ]. Previous studies found DNA methylation markers in melanoma predictive for survival independent of clinical histopathological factors using Illumina methylation arrays [ 8 – 10 ]. However, array-based analyses are limited in the number of methylation sites. For example the Illumina methylation arrays, 450K and 850K, can only detect less than 2% and 4% of CpG sites respectively. The genome-wide methylated DNA sequencing (MeD-seq) is an advanced technique compared to traditional array-based methylation analysis, enabling comprehensive profiling of over 50% of all CpG sites across the genome [ 11 ]. This essay enables genome-wide profiling of methylated CpG regions by selectively sequencing methylated DNA using the DNA-methylation dependent LpnPI restriction enzyme and identifying differentially methylated regions (DMRs). This approach could potentially uncover novel methylation sites related to metastasis in melanoma, as evidenced by its successful application in studying prognostic signatures in non-cirrhotic hepatocellular carcinoma [ 12 ], survival in renal cell carcinoma [ 13 ], and metastasis in uveal melanoma [ 14 ]. Furthermore, the previous studies mainly compared unmatched primary and metastatic early-stage melanomas. Therefore, we used the Dutch Early-Stage Melanoma (D-ESMEL) study [ 15 ], which is designed to identify new prognostic factors in addition to known prognostic factors (such as Breslow thickness and ulceration) and includes the largest collection worldwide of matched primary stage I/II melanoma, which lead to distant metastasis. The discovery set of the D-ESMEL study also includes matched distant metastasis from primary melanomas [ 15 ]. In this study, we aimed to investigate if the DNA methylation patterns differ between matched early-stage primary melanomas that metastasize and those that do not, and whether these patterns are retained in matched distant metastases. To this end, we applied the MeD-seq technique of primary melanomas and matched distant metastasis of the D-ESMEL study. Materials and Methods D-ESMEL Study The D-ESMEL study was designed to identify new prognostic markers in early-stage melanoma [15]. Early-stage melanoma patients who developed a distant metastasis (cases) were matched to those who did not (controls) based on age, sex, Breslow thickness, ulceration and follow-up time (i.e. the control was required to have at least the same amount of follow-up time as the time between the primary melanoma and the metastasis of the case). In the D-ESMEL study, formalin-fixed, paraffin-embedded (FFPE) specimens of the primary melanoma of the controls and the primary melanoma and metastasis of the case were collected. A detailed description of the D-ESMEL study design has been published elsewhere [15]. Sample size Previous studies using the MeD-seq technology showed that a group size of at least 7 was sufficient to detect reliable DMRs [16]. This is based on a database of thousands of general and cancer-specific DMRs from gynecological cancers and healthy tissues analyzed with MeD-seq using a genome-wide sliding window with Bonferroni-corrected Chi-squared testing, requiring DMR consistency in > 70% of samples per group [16]. The D-ESMEL study was sufficiently large to start with at least 15 samples per group, including primary melanomas that did not metastasize, primary melanomas that did, and the corresponding metastatic tissue. Including at least 7 case-control sets should be sufficient to reduce DMRs that are detected due to technical variation and remove the false positives. As a start, we selected 15 sets, in order to take sample failures into account. The D-ESMEL study was sufficiently large to make a careful selection in order to include multiple primary tumor stages and multiple distant metastatic sites. The selection of 15 sets included 45 samples (15 primary melanomas of cases, 15 primary melanomas of controls and 15 matched metastasis of cases) of the discovery set of the D-ESMEL study. Laser capture microdissection for target cell enrichment FFPE blocks of the 15 sets were sectioned according to the sandwich cutting procedure: a 4 µm section for diagnosis (H&E before); 16 µm sections for LCM-MeD-seq analysis; a set of 3–6 × 8 µm (24 µm) sections for DNA quality determination, and finally, a 4 µm section for pathological confirmation (H&E after). The tumor tissue was annotated based on the H&E by a dermatopathologist (AM) to determine which samples were of sufficient quality for LCM, containing at least 5 mm 2 tumor tissue without necrosis or blood and 30% tumor cellularity. This resulted in 33 samples eligible for Laser capture microdissection (LCM), see Fig. 1. For high-quality genome-wide DNA methylation experiments, it is crucial to have the purest possible population of target cells. LCM is an accurate method for isolating these specific target cells out of complex heterogeneous tissue [17]. To improve tissue adhesion to the membrane, PALM membrane slides 1.0 PEN were prior to use UV treated (254 nm) and coated with poly-L-lysine (0.1%) according to manufacturer’s protocol. 16 µm sections of FFPE tissues were cut, mounted on the UV and poly-L-lysine pre-treated membrane slides and incubated at 60°C for 2 hours. Next, the sections were dewaxed, dehydrated and stained with hematoxylin. After digital imaging of the stained slides, regions of interest were annotated by a pathologist. The annotated regions must contain the highest possible percentage of desired target cells (preferably > 70%) surrounded by only normal tissue cells (and preferably no other pathological structures). Using LCM, the percentage of desired cells in the region of interest (cancer lesion) can be increased and regions with unwanted cell types/cell structures can be removed. The annotated regions were cut by a laser beam using the Leica Laser Microdissection system and collected in tubes. For the MeD-seq analysis we need at least 10 (preferably 50) ng DNA per 8.66 µl of a sample. For this we dissected one or more regions that together were 5–20 mm2. DNA extraction FFPE tissue DNA was extracted using our Proteinase K protocol. To each LCM sample 50 µl of 1 mg/ml Proteinase K and 200 µl of mineral oil were added. Thereafter, all sample types were incubated 1–2 hours at 70˚C. The samples were then gently mixed and incubated overnight for 16 hours at 70˚C. If the tissue was not completely dissolved, 5 µl 10 mg/ml Proteinase K was added and the samples were incubated for 6 hours at 70˚C. Afterwards, Proteinase K was inactivated by 10 min at 95˚C and the mineral oil was removed. Samples were stored at -20˚C. After the extraction of genomic, the concentrations and quality of the extracted DNA were determined using the DNA QC qPCR. MeD-seq sample preparation MeD-seq assays were performed as previously described [11]. In short, genomic DNA and plasma-derived cfDNA were digested with the methylation-dependent restriction enzyme LpnPI (New England Biolabs, Ipswich, MA) generating 32 bp DNA fragments containing the methylated CpG in the middle. All samples were sequenced on the Illumina NextSeq2000 platform. MeD-seq was performed on the 33 LCM tumor tissues of which 3 samples were excluded due to low yield of methylated reads (< 20% of total reads). The remaining samples consisted of 11 primary melanomas of cases, 10 matched metastatic tissues, and 9 primary melanomas of controls. Resulting in 8 complete case-control sets (i.e. 1 primary melanoma of the case and 1 primary melanoma of the control) and 6 complete case-metastasis sets (i.e. 1 primary melanoma of the case + 1 matched metastasis), see Fig. 1. MeD-seq data analysis Custom python scripts were used to process acquired DNA methylation profiles using the following python version and packages: python v3.8.5, numpy v1.19.1, scikit-learn v0.23.2, scipy v1.5.2, matplotlib v3.3.1, seaborn v0.11.0, umap-learn v0.5.1. Raw fastq files were subjected to Illumina adaptor trimming and reads were filtered based on LpnPI restriction site occurrence between 13–17 bp from either the 5′ or 3′ end of the read. Reads that passed the filter were mapped to hg38 using bowtie2 version 2.3.3 and visualized in IGV. Genome-wide individual LpnPI site scores were used to generate read count scores for the following annotated regions: promoter (1 kb before and 1 kb after the transcription start sites (TSS)), CpG-islands and gene bodies (1 kb after TSS till transcription end site). Gene and CpG-island annotations were downloaded from ENSEMBL (Homo_sapiens_hg38.GRCh38.79.gtf, www.ensembl.org). MeD-seq analysis: Genome-wide sliding window To detect DMRs, a genome-wide sliding window was used to detect sample-wise and group-wise sequentially differentially methylated LpnPI sites. Genome-wide read counts were normalized, reads per million, for coverage and compared using the Chi-Squared test, with significance set at p < 0.05 and a Bonferroni correction for multiple testing. Neighboring significantly called LpnPI sites were binned and reported. Overlap of genome-wide detected DMRs was reported for promoter, CpG-islands or gene body regions using the annotations of the UCSC database (Hg38) [18]. DMR thresholds were based on LpnPI site count, DMR sizes (in bp) and mean log10 FCs of read counts between group a and b (\(\:FC=log10\left(\frac{\mu\:a}{\mu\:b}\right)\)). All comparisons are made between primary case and control (case-control) sets and primary case and matched metastatic material (case-metastasis) sets. MeD-seq analysis: Genome-wide promoter, gene body, and CpG-island regions Alternatively, group-wise DMRs were detected in the known promoter, CpG-island, and gene body regions of the Hg38 genome using a paired t-test with significance set at p < 0.05 and a FDR-BH correction for multiple testing. All comparisons are made between primary case and control (case-control) sets and primary case and matched metastatic material (case-metastasis) sets. Unmatched analyses Next to the matched sample analysis, DMRs between the unmatched samples were also analyzed in order to assess the impact of two matched control samples which developed a metastases after their matched cases. In this sensitivity analysis these primary controls are changed to primary cases. In this analysis we assess DMRs between primary cases (n = 11) and controls (n = 9), primary case tumors (n = 11). Additionally, DMRs primary tumors (n = 20) and metastatic tumors (n = 11) to gain biological insight and DMRs within the metastatic samples were analyzed to detect DMRs between metastatic lung (n = 3) or GI (n = 3) and all other metastatic sites (n = 7). All DMRs were detected with an unpaired t-test in the known promoter, gene body, and CpG-island regions and corrected for multiple testing with FDR-BH and significance set at p < 0.05. Results Patients characteristics To identify prognostic differentially methylated regions (DMRs) in early-stage melanoma, we performed the MeD-seq technique on early-stage melanomas from the D-ESMEL study. In this study early-stage melanoma patients who developed a distant metastasis (case) were matched to those who did not (controls) based on age, sex, Breslow thickness, and ulceration. We aimed to include 10 complete sets, as this should be sufficient to reduce the biological variation and identify robust prognostic DMRs [16]. Each complete set includes 3 samples: a primary melanoma from the case, the matched distant metastasis, and a primary melanoma from the control. Initially 15 candidate sets, i.e. 45 samples, were selected from the discovery set of the D-ESMEL study for MeD-seq analysis, prioritizing primary tumors >1 mm thick to ensure sufficient material for laser capture microdissection (LCM). Following the LCM quality control, a total of 12 samples were excluded due to insufficient tumor area (<5mm 2 tumor area or <30% tumor cellularity) or poor DNA quality. After applying MeD-seq on the remaining 33 samples, three samples were excluded because of low quality, since only 20% of the total reads were methylated. This resulted in 30 total unmatched samples of which 20 primary tumor tissues and 10 metastatic tissues, including 11 primary melanomas from cases, 10 matched metastasis, and 9 primary melanomas of controls ( Additional file 1 ). These samples are matched on age, sex, ulceration, and Breslow thickness resulting in 8 complete matched case-control sets (primary melanomas of the case and the control) and 6 matched case-metastasis sets (primary melanoma of the case and the matched distant metastasis) ( Table 1 ). A complete overview of the inclusion criteria is shown in Figure 1 . The primary analyses were conditional analyses of matched sets. As a sensitivity analyses, unmatched samples were analyzed as well to take all available samples into account. Table 1. Clinical-histopathological characteristics of the matched MeD-seq samples. Characteristic Case-control sets (primary melanomas) Control (n=8) Case (n=8) Total (n=8) of which with matched metastatic material (n=6) Age, median (IQR) 54 (50-60) 51 (45-62) 51 (46-59) Sex Female 2 2 1 Male 6 6 5 Breslow ≤1.0 mm 1 0 0 >1.0-2.0 mm 1 2 2 >2.0-4.0 mm 2 2 2 >4.0 mm 4 4 2 Ulceration Absent 5 5 4 Present 3 3 2 Stage 1A 1 1 1 1B 1 1 1 2A 1 1 1 2B 4 4 3 2C 1 1 0 Distant metastatic site collected Lung n/a 1 1 Gastro-intestinal n/a 2 2 Adrenal gland n/a 1 1 Lymph node n/a 1 1 Muscle n/a 1 1 Skin n/a 2 0 Time until distant metastasis < 2 years n/a 6 4 2-5 years n/a 2 2 Location Primary Melanoma Scalp 1 1 1 Trunk 3 3 2 Lower extremities 2 1 1 Upper extremities 2 3 2 WHO subtype Superficial Spreading 6 8 6 Nodular 2 0 0 Differentially methylated regions between matched case-controls and case-metastasis sets Genome-wide sliding window analysis reveals a DMR in PTPRN2 DMRs were identified between case-control and case-metastasis sets by comparing the normalized methylated read counts at every LpnP1-site for each matched sample and per group using a genome-wide sliding window and Chi-squared test. After a Bonferroni correction no DMRs were found between both the control-case and case-metastasis groups. By comparing the sets sample-wise (e.g. the primary melanoma of the case of set A vs the primary melanoma of the control of set A) we find 20,752 DMRs between the case-control sets and 7,759 DMRs between the case-metastasis sets. In the sample-wise comparisons a large number of DMRs are identified due to 1:1 comparisons giving an overestimation of the true DMRs because of biological variation, e.g. due to methylated repetitive regions. We found one DMR which was significant within all samples for both the case-control and case-metastasis sets (fold change (FC) > 2; p <0.05) in the Protein Tyrosine Phosphatase Receptor Type N2 (PTPRN2) gene. We also found a DMR in a CpG-island (CpG20827; chr17:81341464-81342175) which was significant in all case-control sets ( Additional file 2 ). Genome-wide promoter, gene body, and CpG-island regions matched analysis reveals methylated protein coding genes as potential prognostic markers Alternatively, to the genome-wide sliding window analysis we investigated the known promoter, CpG-island, and gene body methylation regions and tested for differences using the paired t-test followed by a false discovery rate Benjamini-Hochberg (FDR-BH) correction. In this analysis PTPRN2 was not detected as a DMR. We did find 4660 DMRs between case-control sets and 8192 DMRs between case-metastasis sets. However, after FDR-BH correction no significant DMRs remained in both analyses. The most significant site between the case-control sets, before correction, is in the promoter region of the Coiled-Coil-Helix-Coiled-Coil-Helix Domain Containing 2 ( CHCHD2 ) pseudogene 8 ( CHCHD2P8 ), see Table 2 . Methylation of CHCHD2P8 decreased in all controls (FC=-0.122; p =1.71e-05) when comparing it to the matched cases ( Figure 2a ). In the top 10 most significant sites ( Additional file 3 ) two DMRs in the protein coding genes N-Myc downstream-regulated gene family member 2 ( NDRG2 ) and Endothelin 2 ( EDN2 ) were found. NDRG2 showed an increase (FC=0.365; p =3.17e-05) and EDN2 a decrease (FC=-0.223; p =4.62e-05) in methylation in all metastatic sites compared to the primary site, see Figure 2b and c respectively. The remaining top 10 DMRs mostly included long non-coding RNAs (lncRNA), pseudogenes, unknown, or uncharacterized proteins. Interestingly, no single DMR was identified in both the case-control and case-metastasis sets in the top 10 most significant DMRs ( Additional file 3 ). Table 2. Most significant relevant sites across matched and unmatched primary controls, cases, and metastatic tumors. Analysis Comparison Site Region p-value † Fold change ‡ Function Matched Case vs Control CHCHD2P8 Promoter 1.71-05 -0.122 Pseudogene NDRG2 Promoter 3.17e-05 0.365 Protein coding Case vs Metastasis EDN2 Promoter 4.62e-05 -0.223 Protein coding Unmatched Case vs Control CYP2E1 Promoter 2.50e-05 0.857 Protein coding Case vs Metastasis SERPINB8 Promoter 6.25e-05 1.078 Protein coding Metastatic lung vs other TAS2R50 Promoter 4.87e-03 Present in Lung Protein coding Metastatic GI vs other SPAG11B Promoter 3.81e-02 Absent in GI Protein coding † Uncorrected p-value ‡ Case is taken as reference When investigating loss of methylation we found no methylation site that was uniquely present in either the case, control, or metastasis samples. However, we did find four sites in which methylation was completely lost in the metastasis compared to the primary melanoma of the same case. These are found in the promoter region in the GC vitamin D binding protein ( GC ) gene ( p =4.05e-04) ( Figure 2d ), and the gene body of ubiquitin-specific peptidase 17-like family member 1 ( USP17L1 ) ( p =1.72e-03) ( Figure 2e ). Matched analysis identifies more DMRs in protein-coding genes than unmatched analysis To assess the impact of our matching strategy, we conducted a sensitivity analysis using the unmatched case-control and case-metastatic samples ( Additional file 1 ). This is of particular importance as, in the matched analysis, two control subjects developed metastases after the matched case. DMRs were detected using an unpaired t-test on the promoter, gene body, and CpG-island regions between the unmatched case control primary tumors. This resulted in no DMRs after FDR-BH correction. However, we report 5561 DMRs before correction. This included DMRs in the promoter of the protein coding genes and ligand of arginyl-tRNA--protein transferase 1 ( LIAT1 ) (FC=-0.299; p =1.70e-04) and cytochrome P450 family 2 subfamily E member 1 ( CYP2E1 ) (FC=0.857; p =2.50e-05) ( Table 2 ). CYP2E1 showed a general reduction in methylation in the primary cases ( Figure 3a ). The top 10 ( Additional file 3 ) also included one long intergenic non-protein coding RNA (lincRNA) which is an RNA type known to play a role in cancer development. Interestingly, none of the top 10 DMRs overlapped with DMRs found in the eight matched case-control sets. An unpaired t-test on the promoter, gene body, and CpG-island regions between the tumor of primary cases and metastatic tumors resulted in no DMRs after FDR-BH correction. However, we report 5980 DMRs before correction. This included one DMR in a protein coding gene, namely in the promoter region of serpin family B member 8 ( SERPINB8 ) (FC=1.078; p =6.25e-05) ( Table 2 ), in which a reduction of methylation was detected in all the primary cases ( Figure 3b ). Again, the top 10 list of DMRs do not overlap with DMRs found in the 6 matched case-metastasis sets ( Additional file 3 ). DMRs in lincRNAs are detected which were also found between the unmatched case-controls. The DMR with the lowest p -value is located in the promoter region of the nuclear pore associated protein 1 ( NPAP1 ) pseudogene 6 ( NPAP1P6 ) (FC=-0.19; p =3.97e-07). Aside from this pseudogene, three other pseudogenes were identified, including glycine-N-acyltransferase like 1 ( GLYATL1 ) pseudogene 2 ( GLYATL1P2 ) (Absent in metastasis; p =2.74e-05), cutaneous T cell lymphoma-associated antigen family, member 5 ( CTAGE5 ) pseudogene (FC=-0.838; p =3.65e-05), and RP4-640E24.1 (FC=-1.01; p =4.46e-05). Unmatched comparison of primary and metastatic tumors fails to detect meaningful DMRs To gain biological insight between the tumor of the primary case and its metastasis we performed an unpaired t-test on the promoter, gene body, and CpG-island regions. This resulted in no DMRs after FDR-BH correction. However, we report 12,472 DMRs before correction. No DMRs in protein coding genes were detected in the ten most significant DMRs ( Additional file 3 ). The top 10 did include DMRs in the promoter region of four pseudogenes, namely RP11-72M10.8 (FC=0.402; p =5.42e-06), family with sequence similarity 27-like (FAM27L) pseudogene (FC=0.155; p =8.95e-06), family with sequence similarity 27 (FAM27) pseudogene (FC=0.147; p =1.64e-05), and RNA, U6 small nuclear 162, pseudogene (RNU6-162P) (FC=1.17; p =1.39e-05). TAS2R50 and SPAG11B are differentially methylated in metastatic lung and GI An unpaired t-test between the methylated regions of lung metastasis (n=3), gastro-intestinal (GI) metastasis (n=3), and all other metastatic samples (n=7) in the promoter, gene body, and CpG-island region resulted in four DMRs in lung metastasis and three DMRs in GI metastasis after FDR-BH correction. In total, we report 2911 DMRs in lung metastasis and 10,805 DMRs in GI metastasis. In the lung metastasis one DMR was found to be present in a protein coding gene in the promoter region of the taste 2 receptor member 50 ( TAS2R50 ) gene (p =4.87e-03) . In the GI metastasis one DMR was found to be absent in a protein coding gene in the promoter region of sperm associated antigen 11B ( SPAG11B ) gene (p =3.81e-02) ( Table 2 ) . Most other DMRs are located in CpG-islands and one DMR was found in the promoter region of a lincRNA ( Additional file 3 ). Discussion In this study, we performed genome-wide DNA methylation analyses to uncover potential DNA methylation patterns that are associated with the development of distant metastasis in early-stage cutaneous melanoma. We applied MeD-seq essay which allows higher coverage (50%) and resolution compared to traditional array-based techniques (2–4%). Primary melanomas and metastatic samples from the D-ESMEL study were used to identify DMRs that are related with melanoma progression and not associated with known prognostic features (Breslow thickness, ulceration, age, sex). Our analysis revealed several differently methylation regions in promoters, coding regions of protein coding genes, in addition to non-coding RNAs that warrant further investigation for their role in melanoma progression. When comparing metastasized primary tumors (cases) to their matched primary melanoma of controls we found eight candidate DMRs, located in the promotor regions of the genes CYP2E1 , PTPRN2 , CHCHD2, NDRG2 , EDN2 , GC , USP17L1 , and SERPINB8 , that may reflect the metastasis transition from the primary to distant organs. Firstly, we found a general decrease in DNA methylation in the promotor of the CYP2E1 gene in unmatched primary cases compared to controls. One study found that forced expression of CYP2E1 inhibited the growth of malignant melanoma cells in vivo using bone marrow-derived mesenchymal stem cells to deliver CYP2E1 transduced with a pAd5-CMV- CYP2E1 recombinant adenovirus [ 19 ]. These findings may contradict our observed decreased methylation in the CYP2E1 promotor region, which may lead to increased expression of CYP2E1, and has been reported to increase expression of CYP2E1 [ 20 ] in other diseases. However, it is unlikely that growth of melanoma cells is inhibited in primary melanomas that metastasized, as larger tumor size is associated with poorer outcomes [ 21 ]. A decrease in methylation of the CYP2E1 promotor region has also been reported to increase expression of CYP2E1 [ 20 ] in other diseases. Secondly, for PTPRN2 , we found methylation differences in the promoter in each case compared to its matched control. PTPRN2 has been reported to be overexpressed in tumors such as breast, pancreatic, and colon cancer. In addition, it is reported to be prognostic for overall survival in pancreatic and colon cancer [ 22 – 24 ] and might also be overexpressed and prognostic in melanoma. Finally, the functional gene CHCHD2 of the pseudogene CHCHD2P8 has been discussed as potential prognostic factor for cancer and target for cancer therapy, as it triggers oxidative phosphorylation causing increased proliferation and metastasis [ 25 ]. However, no functionality of CHCHD2P8 itself has been reported. Additionally, we identified six DMRs in potentially relevant genes which might help to better understand the development of metastasis in early-stage melanoma. Four of these genes ( PTPRN2 , NDRG2 , EDN2 , and GC ) have been reported to be related to cancer in general but not specifically for melanoma. It was found that over-expression of NDRG2 could inhibit tumor growth and invasion [ 26 , 27 ]. Additionally, methylation of NDRG2 was found to be the main cause for its down regulation in gastric cancer [ 28 ]. EDN2 is a protein coding gene and part of the endothelin protein family. Endothelins and their receptors have been associated with melanoma progression through the alteration of tumor-host interactions [ 29 ]. A decrease in methylation in the promoter of EDN2 was reported in the metastatic site compared to the primary site. How the methylation of EDN2 affects its expression in melanoma remains uncertain. The GC protein binds to circulating vitamin D and activates macrophages [ 30 ] and an abundance of the GC protein decreases tumor spreading [ 31 ]. The exact relevance and effect of methylation of GC in melanoma remains uncertain since no methylation was reported in the promoter of any of the metastatic samples. Lastly, we reported DMRs in USP17L1 and SERPINB8 which have been known to be involved in melanomas. In USP17L1, a loss of methylation was detected at the metastatic site compared to the primary in the gene body region. Methylation of the gene body could have different effects on USP17L1 such as an increase of expression, impact on the splicing of the gene, and possibly cause a loss of regulation of the Ras processing pathway [ 32 ]. In SERPINB8 , a DMR was detected in the promoter region. In vitro experiments of BRAF V600E mutant melanoma found that SERPINB8 regulates the expression of integrin alpha x ( ITGAX ) in melanoma cells, a gene which significantly promotes the invasion and proliferation of melanoma cells [ 33 ]. However, methylation of SERPINB8 and its effect on its expression has not been studied. No relevant DMRs were found when comparing primary and metastatic tumors as we report mostly DMRs in pseudogenes with unknown functions. When comparing methylation patterns between metastatic sites (lung, GI, and others) two significant DMRs were reported TAR2R50 in lung metastasis and SPAG11B in GI metastasis. A paralog of SPAG11B , SPAG11A , was found to be differentially expressed in gastric carcinoma [ 34 ] and TAR2R50 encodes a bitter taste receptors part of the taste family 2 receptors (T2R) [ 35 ], which were associated, together with family 1 receptors (T1R), to survival differences in 12 solid tumor subtypes, including one in lung adenocarcinoma and melanoma. The strength of our study is that we analyze the differences between matched (metastatic) melanomas to uncover truly relevant DMRs in melanoma covering a wide-range of CpG sites. Matching case and controls on known prognostic factors allows a more robust identification of new prognostic factors, rather than identifying methylation patterns related to known prognostic factors. Moreover, matching primary tumors and their corresponding metastases is particularly robust, as both tissues originate from the same patient, minimizing variability due to genetic background differences. In addition, we used MeD-seq allowing us to uncover CpG sites genome-wide instead of a limited number of CpG sites using array-based methods. To our standing, this is the first study analyzing matched primary case-control and matched primary case and metastasis sets across such a wide number of methylated regions. While the matched design of the D-ESMEL study is a strength for detecting new prognostic factors, the limitation of this design is that methylation patterns related to important biological processes could not be detected. Small differences in methylation patterns were observed between the matched primary tumors of cases and controls, which are likely due to the matching on known prognostic factors. As a result, we did not observe DMRs that were prognostic over and above Breslow thickness, ulceration, age and sex. While this is important information for clinical practice, the consequence of this study design, is that any methylation patterns that are related to either Breslow thickness or ulceration cannot be detected in our matched sets. Although our sample size was sufficient to avoid detecting false positive DMRs, possibly additional samples may help with detecting very small differences in DMRs, independent of known prognostic factors. However, the clinical relevance of those small differences is unknown. We also observed small differences between primary tumors and its matched metastasis, which could be attributed to intratumoral heterogeneity introduced during the metastatic process. Another aspect of the study design that is both a strength and a limitation, is the matching of the controls on follow-up time. The control needed to have at least the same amount of follow-up as the case. While this design assures that risk estimates are comparable to a cohort study (as controls are sampled from the risk set of each case in a cohort study), a limitation of this design is that controls can develop metastasis later during follow-up. During this study, we updated the follow-up of all patients and we observed, that two control samples developed a metastases after their matched cases. These controls may have been at lower risk compared to its matched case as the controls developed a metastasis later than its matched case. Therefore, these are still relevant samples to investigate to further understand the development of metastasis. On the other hand, the inclusion of these controls may have further reduced the small differences between cases and controls. Therefore, we also performed a sensitivity analysis where the matching was not included and in which those 2 controls were considered cases. These analysis also did not reveal prognostic DMRs. Conclusions To our knowledge, this is the first study to investigate the methylation of matched primary and metastatic early-stage melanomas using the MeD-seq essay. We did not identify DMRs in early-stage primary melanomas that are prognostic for developing distant metastasis independent from known prognostic factors. We identified 8 DMRs that may play a biological role in the development of metastasis and which can be further investigated as possible epigenetic drivers of melanoma progression. Declarations Acknowledgments We would like to thank IKNL for providing the clinical data and Catherine Zhou for collecting and curating the clinical data and FFPE blocks. Author Contributions Conceptualization: LH, RB; Data Curation: AM, TK; Formal analysis: JO, RB; Investigation: TK, JO, RB; Methodology: JO, RB; Supervision: LH, YL, RB; Writing – Original Draft: JO, TK, RB; Writing – Reviewing & Editing: LH, YL, MW, AM, JG, JB, RB. Data Availability Statement The clinical data from these patients was provided to the authors by the Netherland Comprehensive Cancer Organisation (IKNL). These data are not publicly available and restrictions apply to the availability of the data used for the current study. However, the MeD-seq data and clinical data are available upon reasonable request to the authors and with permission of the Netherlands Comprehensive Cancer Organisation. Funding No funding was provided for this work. Clinical Trial Number Clinical trial number: not applicable. Competing interest The authors declare no conflict of interest or financial interests except for RB, JB and JG, who report being shareholder in Methylomics B.V., a commercial company that applies MeD-seq to develop methylation markers for cancer staging. Consent for publication Not applicable. Ethics approval statement and consent to participate The use of leftover diagnostic tissue samples for scientific research is based on the ‘no objection’ principle, as outlined in the Code of Conduct for Health Research by the Committee on Regulation of Health Research. References Zhou C, Louwman M, Wakkee M, van der Veldt A, Grünhagen D, Verhoef C, et al. Primary Melanoma Characteristics of Metastatic Disease: A Nationwide Cancer Registry Study. Cancers (Basel). 2021;13:4431. https://doi.org/10.3390/cancers13174431. Enninga EAL, Moser JC, Weaver AL, Markovic SN, Brewer JD, Leontovich AA, et al. Survival of cutaneous melanoma based on sex, age, and stage in the United States, 1992–2011. Cancer Med. 2017;6:2203–12. https://doi.org/10.1002/cam4.1152. Isaksson K, Mikiver R, Eriksson H, Lapins J, Nielsen K, Ingvar C, et al. Survival in 31 670 patients with thin melanomas: a Swedish population‐based study*. British Journal of Dermatology. 2021;184:60–7. https://doi.org/10.1111/bjd.19015. Luke JJ, Rutkowski P, Queirolo P, Del Vecchio M, Mackiewicz J, Chiarion-Sileni V, et al. Pembrolizumab versus placebo as adjuvant therapy in completely resected stage IIB or IIC melanoma (KEYNOTE-716): a randomised, double-blind, phase 3 trial. The Lancet. 2022;399:1718–29. https://doi.org/10.1016/S0140-6736(22)00562-1. Kirkwood JM, Del Vecchio M, Weber J, Hoeller C, Grob J-J, Mohr P, et al. Adjuvant nivolumab in resected stage IIB/C melanoma: primary results from the randomized, phase 3 CheckMate 76K trial. Nat Med. 2023;29:2835–43. https://doi.org/10.1038/s41591-023-02583-2. Fietz S, Zarbl R, Niebel D, Posch C, Brossart P, Gielen GH, et al. CTLA4 promoter methylation predicts response and progression-free survival in stage IV melanoma treated with anti-CTLA-4 immunotherapy (ipilimumab). Cancer Immunology, Immunotherapy. 2021;70:1781–8. https://doi.org/10.1007/s00262-020-02777-4. Marzese DM, Scolyer RA, Huynh JL, Huang SK, Hirose H, Chong KK, et al. Epigenome-wide DNA methylation landscape of melanoma progression to brain metastasis reveals aberrations on homeobox D cluster associated with prognosis. Hum Mol Genet. 2014;23:226–38. https://doi.org/10.1093/hmg/ddt420. Akbani R, Akdemir KC, Aksoy BA, Albert M, Ally A, Amin SB, et al. Genomic Classification of Cutaneous Melanoma. Cell. 2015;161:1681–96. https://doi.org/10.1016/j.cell.2015.05.044. Wouters J, Vizoso M, Martinez-Cardus A, Carmona FJ, Govaere O, Laguna T, et al. Comprehensive DNA methylation study identifies novel progression-related and prognostic markers for cutaneous melanoma. BMC Med. 2017;15:101. https://doi.org/10.1186/s12916-017-0851-3. Conway K, Edmiston SN, Vondras A, Reiner A, Corcoran DL, Shen R, et al. DNA Methylation Classes of Stage II and III Primary Melanomas and Their Clinical and Prognostic Significance. JCO Precis Oncol. 2024. https://doi.org/10.1200/PO-24-00375. Boers R, Boers J, De Hoon B, Kockx C, Ozgur Z, Molijn A, et al. Genome-wide DNA methylation profiling using the methylation-dependent restriction enzyme LpnPI. Genome Res. 2018;28:88–99. https://doi.org/10.1101/gr.222885.117. Fu S, Deger T, Boers RG, Boers JB, Doukas M, Gribnau J, et al. Hypermethylation of DNA Methylation Markers in Non-Cirrhotic Hepatocellular Carcinoma. Cancers (Basel). 2023;15:4784. https://doi.org/10.3390/cancers15194784. Bos MK, Verhoeff SR, Oosting SF, Menke-van der Houven van Oordt WC, Boers RG, Boers JB, et al. Methylated Cell-Free DNA Sequencing (MeD-seq) of LpnPI Digested Fragments to Identify Early Progression in Metastatic Renal Cell Carcinoma Patients on Watchful Waiting. Cancers (Basel). 2023;15:1374. https://doi.org/10.3390/cancers15051374. Smit KN, Boers R, Vaarwater J, Boers J, Brands T, Mensink H, et al. Genome-wide aberrant methylation in primary metastatic UM and their matched metastases. Sci Rep. 2022;12:42. https://doi.org/10.1038/s41598-021-03964-8. Zhou C, Mooyaart AL, Kerkour T, Louwman MWJ, Wakkee M, Li Y, et al. The Dutch Early-Stage Melanoma (D-ESMEL) study: a discovery set and validation cohort to predict the absolute risk of distant metastases in stage I/II cutaneous melanoma. Eur J Epidemiol. 2025;40:27–42. https://doi.org/10.1007/s10654-024-01188-4. Boers J, Boers R, Sakoltchik J, Dasgupta S, Martens L, Tadema KCD, et al. DNA methylation database for gynecological cancer detection, classification and assay development. 2024. https://doi.org/10.1101/2024.07.01.601485. Quint W, Jenkins D, Molijn A, Struijk L, van de Sandt M, Doorbar J, et al. One virus, one lesion—individual components of CIN lesions contain a specific HPV type. J Pathol. 2012;227:62–71. https://doi.org/10.1002/path.3970. Perez G, Barber GP, Benet-Pages A, Casper J, Clawson H, Diekhans M, et al. The UCSC Genome Browser database: 2025 update. Nucleic Acids Res. 2025;53:D1243–9. https://doi.org/10.1093/nar/gkae974. Wang J, Ma D, Li Y, Yang Y, Hu X, Zhang W, et al. Targeted delivery of CYP2E1 recombinant adenovirus to malignant melanoma by bone marrow-derived mesenchymal stem cells as vehicles. Anticancer Drugs. 2014;25:303–14. https://doi.org/10.1097/CAD.0000000000000046. Kaut O, Schmitt I, Wüllner U. Genome-scale methylation analysis of Parkinson’s disease patients’ brains reveals DNA hypomethylation and increased mRNA expression of cytochrome P450 2E1. Neurogenetics. 2012;13:87–91. https://doi.org/10.1007/s10048-011-0308-3. Gershenwald JE, Scolyer RA, Hess KR, Sondak VK, Long G V., Ross MI, et al. Melanoma staging: Evidence‐based changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA Cancer J Clin. 2017;67:472–92. https://doi.org/10.3322/caac.21409. Sengelaub CA, Navrazhina K, Ross JB, Halberg N, Tavazoie SF. PTPRN2 and PLCβ1 promote metastatic breast cancer cell migration through PI(4,5)P2‐dependent actin remodeling. EMBO J. 2016;35:62–76. https://doi.org/10.15252/embj.201591973. Feigin ME, Garvin T, Bailey P, Waddell N, Chang DK, Kelley DR, et al. Recurrent noncoding regulatory mutations in pancreatic ductal adenocarcinoma. Nat Genet. 2017;49:825–33. https://doi.org/10.1038/ng.3861. Yin J, Guo Y. HOXD13 promotes the malignant progression of colon cancer by upregulating PTPRN2. Cancer Med. 2021;10:5524–33. https://doi.org/10.1002/cam4.4078. Gundamaraju R, Lu W, Manikam R. CHCHD2: The Power House’s Potential Prognostic Factor for Cancer? Front Cell Dev Biol. 2021;8. https://doi.org/10.3389/fcell.2020.620816. Liu N, Wang L, Li X, Yang Q, Liu X, Zhang J, et al. N-Myc downstream-regulated gene 2 is involved in p53-mediated apoptosis. Nucleic Acids Res. 2008;36:5335–49. https://doi.org/10.1093/nar/gkn504. Choi S-C, Yoon SR, Park YP, Song EY, Kim JW, Kim WH, et al. Expression of NDRG2 is related to tumor progression and survival of gastric cancer patients through Fas-mediated cell death. Exp Mol Med. 2007;39:705–14. https://doi.org/10.1038/emm.2007.77. Chang X, Li Z, Ma J, Deng P, Zhang S, Zhi Y, et al. DNA Methylation of NDRG2 in Gastric Cancer and Its Clinical Significance. Dig Dis Sci. 2013;58:715–23. https://doi.org/10.1007/s10620-012-2393-z. Saldana-Caboverde A, Kos L. Roles of endothelin signaling in melanocyte development and melanoma. Pigment Cell Melanoma Res. 2010;23:160–70. https://doi.org/10.1111/j.1755-148X.2010.00678.x. Saburi E, Saburi A, Ghanei M. Promising role for Gc-MAF in cancer immunotherapy: From bench to bedside. Caspian Journal of Internal Medicine. 2017;8:228–38. https://doi.org/10.22088/cjim.8.4.228. Danahy L, Long C, Hofmann TJ, Tara Z, Mark J, Roizen JD. Dietary vitamin D is a novel modulator of tumor engraftment through regulation of GC protein abundance. 2024. https://doi.org/10.21203/rs.3.rs-3911213/v1. Wang Q, Xiong F, Wu G, Liu W, Chen J, Wang B, et al. Gene body methylation in cancer: molecular mechanisms and clinical applications. Clin Epigenetics. 2022;14:154. https://doi.org/10.1186/s13148-022-01382-9. Ni L, Li P, Li M, Huang S, Dang N. SERPINB8 and furin regulate ITGAX expression and affect the proliferation and invasion of melanoma cells. Exp Dermatol. 2023;32:24–9. https://doi.org/10.1111/exd.14677. Li DF, Wang NN, Chang X, Wang SL, Wang LS, Yao J, et al. Bioinformatics analysis suggests that COL4A1 may play an important role in gastric carcinoma recurrence. J Dig Dis. 2019;20:391–400. https://doi.org/10.1111/1751-2980.12758. Wei K, Hill BL, Thompson JC, Miller ZA, Mueller A, Lee RJ, et al. Bitter Taste Receptor Agonists Induce Apoptosis in Papillary Thyroid Cancer. Head Neck. 2025. https://doi.org/10.1002/hed.28120. Additional Declarations Competing interest reported. The authors declare no conflict of interest or financial interests except for RB, JB and JG, who report being shareholder in Methylomics B.V., a commercial company that applies MeD-seq to develop methylation markers for cancer staging. Supplementary Files Additionalfile1.xlsx Supplementary Information Additional file 1.xlsx, Unmatched patient characteristics, table containing patient characteristics of all samples. Additionalfile2.ods Additional file 2.ods, Sliding window DMRs, tables containing DMRs identified using a sliding window. Additionalfile3.xlsx Additional file 3.xlsx, DMRs in predefined CpG sites, tables containing DMRs identified in predefined CpG sites. Cite Share Download PDF Status: Published Journal Publication published 06 Dec, 2025 Read the published version in Human Genomics → Version 1 posted Editorial decision: Revision requested 15 Oct, 2025 Reviews received at journal 10 Oct, 2025 Reviews received at journal 10 Oct, 2025 Reviewers agreed at journal 29 Sep, 2025 Reviewers agreed at journal 28 Sep, 2025 Reviewers agreed at journal 26 Sep, 2025 Reviewers agreed at journal 26 Sep, 2025 Reviewers agreed at journal 25 Sep, 2025 Reviewers invited by journal 24 Sep, 2025 Editor assigned by journal 22 Sep, 2025 Submission checks completed at journal 22 Sep, 2025 First submitted to journal 17 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7642695","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":522594187,"identity":"e024b925-b1e5-46b2-b42b-0f773cc5b3ba","order_by":0,"name":"Jasper Ouwerkerk","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Jasper","middleName":"","lastName":"Ouwerkerk","suffix":""},{"id":522594189,"identity":"a24b7bc9-da88-43e4-b2bc-76342f2e8cf2","order_by":1,"name":"Thamila Kerkour","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Thamila","middleName":"","lastName":"Kerkour","suffix":""},{"id":522594190,"identity":"6d230804-4294-4c3e-8426-175c58c682a8","order_by":2,"name":"Antien Mooyaart","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Antien","middleName":"","lastName":"Mooyaart","suffix":""},{"id":522594192,"identity":"fc154371-43b3-494f-9e9f-d6cb5a58a4dc","order_by":3,"name":"Ruben Boers","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Ruben","middleName":"","lastName":"Boers","suffix":""},{"id":522594193,"identity":"70d44a90-01f5-4fbe-a18d-b8d74e27c62b","order_by":4,"name":"Joachim Boers","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Joachim","middleName":"","lastName":"Boers","suffix":""},{"id":522594194,"identity":"3433b578-e056-46fa-a360-a027c084b41d","order_by":5,"name":"Joost Gribnau","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Joost","middleName":"","lastName":"Gribnau","suffix":""},{"id":522594195,"identity":"32b1de82-a84b-4d6c-a283-4a9ea1b6c779","order_by":6,"name":"Marlies Wakkee","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Marlies","middleName":"","lastName":"Wakkee","suffix":""},{"id":522594196,"identity":"49e67599-4b77-4bb1-a0b1-51d9f88ec82b","order_by":7,"name":"Yunlei Li","email":"","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Yunlei","middleName":"","lastName":"Li","suffix":""},{"id":522594197,"identity":"71c7e63f-5ebc-4e94-ad04-50c418b7f32d","order_by":8,"name":"Loes Hollestein","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+0lEQVRIiWNgGAWjYBACPgYeBoYHBSAmYwPDB1Qp7IANpCXBgAFIMTYwzkCVIqiFgYGZhzgtvAcfJBjY2NuzN7c9tqm4Jy8/7Yzxpxs1DHm4tfAlGyQYpCX28BxsN845U2y44XaOmXTOMYZiPA4zk0gwOJzAI5HYJp3blsC4QTrHjDm3gSGxDb+W//Y88g/bpC3/JdjPn51j/JkILQcYeyQY26QZGxISG27nGEjj1cLMYwz0S3Jiz5nENsmeYwnJG26nlQH9IoFTCz97j+GDDxV29uztx59J/KhJsJ0/O3nz55wam8R+HFoYmHGIS+DSMApGwSgYBaOACAAAv4VNBoaf3/wAAAAASUVORK5CYII=","orcid":"","institution":"Erasmus Medical Center","correspondingAuthor":true,"prefix":"","firstName":"Loes","middleName":"","lastName":"Hollestein","suffix":""}],"badges":[],"createdAt":"2025-09-17 17:38:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7642695/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7642695/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40246-025-00871-1","type":"published","date":"2025-12-06T15:57:02+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":93012587,"identity":"38660470-eb2d-446b-935f-85ec32d1934c","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"jpg","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":308378,"visible":true,"origin":"","legend":"","description":"","filename":"figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/e2e21ce90ea9df461944ff1f.jpg"},{"id":93010740,"identity":"4d5b3ef2-a035-4190-bcf4-1d18d7ae2b89","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":557180,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/ba214b13d0a94601ea81b1b7.docx"},{"id":93012583,"identity":"f93bbee8-cca6-4fa3-aba5-ccb178a09a35","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":101363,"visible":true,"origin":"","legend":"","description":"","filename":"figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/550991ee5f7dd265c7c53e10.jpg"},{"id":93010734,"identity":"8e83e1ae-f2c0-4326-b058-1442373b62b9","added_by":"auto","created_at":"2025-10-08 07:15:16","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43012,"visible":true,"origin":"","legend":"","description":"","filename":"figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/c53fe1cdfed74e572e5b36cc.jpg"},{"id":93010747,"identity":"261f7f47-c77c-4b50-b192-af1755590a06","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18757,"visible":true,"origin":"","legend":"","description":"","filename":"table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/12f36f25b7f82f040f88affc.xlsx"},{"id":93012598,"identity":"5d3e86cb-7583-4cea-bba2-7b11fee8d4a3","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18915,"visible":true,"origin":"","legend":"","description":"","filename":"table2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/0e615e008d6c1ef4e3ba1c1e.xlsx"},{"id":93012595,"identity":"fd701994-4ad0-4b7c-a020-33b6137bad55","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"json","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10614,"visible":true,"origin":"","legend":"","description":"","filename":"553c433cc502415a88ce9aa9e52efac2.json","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/35169deaa5eb361f28e89ffe.json"},{"id":93010743,"identity":"8fb4e052-9e6e-4cf5-9af6-8b5c917a3cb1","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19008,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/db6f8eb7e7bbc64f85d8bd60.xlsx"},{"id":93010751,"identity":"2eed7a65-0e0f-4dfd-a427-448282bbd71c","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"ods","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":621511,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile2.ods","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/77648b372214ae431f6e7ab2.ods"},{"id":93010749,"identity":"28b0600d-8132-4720-820c-aaffd6c1e7a3","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":25915,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/d4d431c2f6747f632441a75d.xlsx"},{"id":93010745,"identity":"ab84b04c-423c-4f7a-a489-a4c89a301885","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":129389,"visible":true,"origin":"","legend":"","description":"","filename":"553c433cc502415a88ce9aa9e52efac21enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/4674a9abc3b1521e5f88bf7d.xml"},{"id":93014673,"identity":"f15506f7-7086-4316-a5fa-2fb6277b61c7","added_by":"auto","created_at":"2025-10-08 07:39:17","extension":"jpg","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":308378,"visible":true,"origin":"","legend":"","description":"","filename":"figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/748a498c97f244ab99030331.jpg"},{"id":93010760,"identity":"dbffc2b7-4b47-4497-b1cd-7b502cc86787","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"jpg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":101363,"visible":true,"origin":"","legend":"","description":"","filename":"figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/3095e672edaae6986bd98b57.jpg"},{"id":93010756,"identity":"13201bef-007c-4d43-9fdb-91a2526abfb2","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"jpg","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43012,"visible":true,"origin":"","legend":"","description":"","filename":"figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/c67c2d9123f9ffcca9e94933.jpg"},{"id":93012596,"identity":"940bc41c-64aa-407f-9318-be5f318b3b29","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"jpeg","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":114562,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/c0fed653524077368f984e87.jpeg"},{"id":93010757,"identity":"90928d52-7505-45bb-96bc-341fbe8ef168","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"jpeg","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82702,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/4236dfd8ded6fbc49ce02082.jpeg"},{"id":93010755,"identity":"f0d3e21a-c8f1-4105-bc23-6301267132ef","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"jpeg","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37812,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/420b7b305d5af7cb13626744.jpeg"},{"id":93010762,"identity":"2484513a-72a8-48f0-a5a1-a17871e0b140","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":86344,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/eff15a4c8d18a76cdeceb298.png"},{"id":93012594,"identity":"78a2862d-51a8-438d-bb1c-36a7770579e7","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44653,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/9f2b93d31db4e12e17565741.png"},{"id":93010742,"identity":"99fea22a-d7a7-4ab6-a1bc-4cc4be510f4e","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16550,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/650c229c571a07f66445de94.png"},{"id":93010764,"identity":"9da9c7f5-07d8-4b3b-a972-f581b538c45c","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78095,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/05c563fda33af5cc889689b8.png"},{"id":93013873,"identity":"0010269c-0e45-40a9-9f12-7f9a22029c01","added_by":"auto","created_at":"2025-10-08 07:31:17","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30877,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/31f42e26e7ee9ef181c155e1.png"},{"id":93010761,"identity":"4783a959-95c5-4560-8405-1e960df5dc63","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9856,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/8531c91f317b1cb90f55ddf8.png"},{"id":93012590,"identity":"7719bc62-e88e-490d-9c52-87d93c217938","added_by":"auto","created_at":"2025-10-08 07:23:17","extension":"xml","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":129368,"visible":true,"origin":"","legend":"","description":"","filename":"553c433cc502415a88ce9aa9e52efac21structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/c9bbf49ae3835a0a8cb7e735.xml"},{"id":93010758,"identity":"fff2aca3-f8be-4a19-b8d8-f603220dceaa","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"html","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140617,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/d647124a7873ddfa31ed53ae.html"},{"id":93010733,"identity":"38c24829-146c-4ff2-961c-2b7b7cbdc93d","added_by":"auto","created_at":"2025-10-08 07:15:16","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":308378,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlowchart of sample selection and inclusion criteria at the different steps of sample preparation and analysis. \u003c/strong\u003e15 case-control sets were selected from the discovery set of the D-ESMEL study based on available matched metastatic material, while prioritizing primary tumors with sufficient material for LCM. After LCM 12 tissues were excluded due to low tumor material or tumor cellularity. On the remaining samples the MeD-seq technique was applied resulting in 30 high quality sequenced samples of which 8 matched case-controls and 6 case-metastatic sets.\u003c/p\u003e","description":"","filename":"figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/44c145bdac186fc6c7f24a68.jpg"},{"id":93010739,"identity":"f506e88b-e012-46a2-bd7b-19719e095fba","added_by":"auto","created_at":"2025-10-08 07:15:17","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":101363,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBoxplots of normalized methylated read counts in the most relevant DMRs between matched samples.\u003c/strong\u003e The lines between the boxplots indicate the difference between the matched sets. (\u003cstrong\u003ea\u003c/strong\u003e) Shows the most significant DMR between the case-control sets in the \u003cem\u003eCHCHD2P8\u003c/em\u003e gene. (\u003cstrong\u003eb\u0026amp;c\u003c/strong\u003e) Shows the \u003cem\u003eNDRG2\u003c/em\u003e and \u003cem\u003eEDN2\u003c/em\u003e genes which are one of the most significant DMRs between the case-metastasis sets in protein coding genes. (\u003cstrong\u003ed\u003c/strong\u003e) Shows the \u003cem\u003eGC\u003c/em\u003e gene in which a loss of methylation was detected between all the case-metastasis sets. (\u003cstrong\u003ee\u003c/strong\u003e) Shows the \u003cem\u003eUSP17L1\u003c/em\u003e gene in which a loss of methylation was detected between all the case-metastasis sets.\u003c/p\u003e","description":"","filename":"figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/8c4757b4400b7e86a6d4ba6a.jpg"},{"id":93010737,"identity":"3373163f-59f0-4c33-8654-b107c5987ecc","added_by":"auto","created_at":"2025-10-08 07:15:16","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":43012,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBoxplots of normalized methylated read counts in the most relevant DMRs between unmatched samples.\u003c/strong\u003e (\u003cstrong\u003ea\u003c/strong\u003e) Shows the \u003cem\u003eCYP2E1\u003c/em\u003e gene in which a general reduction of methylation was found between the unmatched cases and controls. (\u003cstrong\u003eb\u003c/strong\u003e) Shows the \u003cem\u003eSERPINB8\u003c/em\u003e gene in which a reduction of methylation was detected between the tumor of the primary case and its metastasis.\u003c/p\u003e","description":"","filename":"figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/452a7493c64318040be97822.jpg"},{"id":97723818,"identity":"a83030b9-5465-49d9-8878-0679a89d17b9","added_by":"auto","created_at":"2025-12-08 16:08:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1732372,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/b9820742-0067-49a9-8241-f2e392a9fb59.pdf"},{"id":93012582,"identity":"7b793500-1e34-4b47-afaf-3c8932354787","added_by":"auto","created_at":"2025-10-08 07:23:16","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":19008,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary Information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAdditional file 1.xlsx, Unmatched patient characteristics, table containing patient characteristics of all samples.\u003c/p\u003e","description":"","filename":"Additionalfile1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/6ad1ba96e5e689edeec7a06a.xlsx"},{"id":93013872,"identity":"e6a69bdf-646f-4275-af20-4e4dd0f90a18","added_by":"auto","created_at":"2025-10-08 07:31:17","extension":"ods","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":621511,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 2.ods, Sliding window DMRs, tables containing DMRs identified using a sliding window.\u003c/p\u003e","description":"","filename":"Additionalfile2.ods","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/1d15e7796d471f03ef43c58b.ods"},{"id":93010735,"identity":"d463b647-7d8a-48bd-859b-1289b81194d9","added_by":"auto","created_at":"2025-10-08 07:15:16","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":25915,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 3.xlsx, DMRs in predefined CpG sites, tables containing DMRs identified in predefined CpG sites.\u003c/p\u003e","description":"","filename":"Additionalfile3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7642695/v1/e25ba16304e97fee249bbe95.xlsx"}],"financialInterests":"Competing interest reported. The authors declare no conflict of interest or financial interests except for RB, JB and JG, who report being shareholder in Methylomics B.V., a commercial company that applies MeD-seq to develop methylation markers for cancer staging.","formattedTitle":"\u003cp\u003eGenome-Wide Methylation Profiles of Primary and Matched Distant Metastasis: Insights from the Dutch Early-Stage Melanoma (D-Esmel) Study\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEarly-stage (stage I-II) cutaneous melanoma accounts for the majority of melanoma diagnoses and is generally associated with a favorable outcome. However, more than 40% of patients who die due to melanoma were initially diagnosed with an early-stage melanoma [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Therefore, current Tumor Node Metastasis (TNM) staging, which is based on Breslow thickness and ulceration, is not sufficient to identify these high-risk patients. With the availability of adjuvant therapy, there is a high need for a precise biomarker which can identify patients at risk for metastasis who may benefit from these therapies [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eA potential biomarker for a prognostic model could be DNA methylation, which is well known to affect gene expression and pathways shown to be related to response to treatment [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e] and metastasis [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Previous studies found DNA methylation markers in melanoma predictive for survival independent of clinical histopathological factors using Illumina methylation arrays [\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. However, array-based analyses are limited in the number of methylation sites. For example the Illumina methylation arrays, 450K and 850K, can only detect less than 2% and 4% of CpG sites respectively.\u003c/p\u003e\u003cp\u003eThe genome-wide methylated DNA sequencing (MeD-seq) is an advanced technique compared to traditional array-based methylation analysis, enabling comprehensive profiling of over 50% of all CpG sites across the genome [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. This essay enables genome-wide profiling of methylated CpG regions by selectively sequencing methylated DNA using the DNA-methylation dependent LpnPI restriction enzyme and identifying differentially methylated regions (DMRs). This approach could potentially uncover novel methylation sites related to metastasis in melanoma, as evidenced by its successful application in studying prognostic signatures in non-cirrhotic hepatocellular carcinoma [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e], survival in renal cell carcinoma [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], and metastasis in uveal melanoma [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eFurthermore, the previous studies mainly compared unmatched primary and metastatic early-stage melanomas. Therefore, we used the Dutch Early-Stage Melanoma (D-ESMEL) study [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], which is designed to identify new prognostic factors in addition to known prognostic factors (such as Breslow thickness and ulceration) and includes the largest collection worldwide of matched primary stage I/II melanoma, which lead to distant metastasis. The discovery set of the D-ESMEL study also includes matched distant metastasis from primary melanomas [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn this study, we aimed to investigate if the DNA methylation patterns differ between matched early-stage primary melanomas that metastasize and those that do not, and whether these patterns are retained in matched distant metastases. To this end, we applied the MeD-seq technique of primary melanomas and matched distant metastasis of the D-ESMEL study.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec3\"\u003e\n \u003ch2\u003eD-ESMEL Study\u003c/h2\u003e\n \u003cp\u003eThe D-ESMEL study was designed to identify new prognostic markers in early-stage melanoma [15]. Early-stage melanoma patients who developed a distant metastasis (cases) were matched to those who did not (controls) based on age, sex, Breslow thickness, ulceration and follow-up time (i.e. the control was required to have at least the same amount of follow-up time as the time between the primary melanoma and the metastasis of the case). In the D-ESMEL study, formalin-fixed, paraffin-embedded (FFPE) specimens of the primary melanoma of the controls and the primary melanoma and metastasis of the case were collected. A detailed description of the D-ESMEL study design has been published elsewhere [15].\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eSample size\u003c/h3\u003e\n\u003cp\u003ePrevious studies using the MeD-seq technology showed that a group size of at least 7 was sufficient to detect reliable DMRs [16]. This is based on a database of thousands of general and cancer-specific DMRs from gynecological cancers and healthy tissues analyzed with MeD-seq using a genome-wide sliding window with Bonferroni-corrected Chi-squared testing, requiring DMR consistency in \u0026gt; 70% of samples per group [16]. The D-ESMEL study was sufficiently large to start with at least 15 samples per group, including primary melanomas that did not metastasize, primary melanomas that did, and the corresponding metastatic tissue. Including at least 7 case-control sets should be sufficient to reduce DMRs that are detected due to technical variation and remove the false positives. As a start, we selected 15 sets, in order to take sample failures into account. The D-ESMEL study was sufficiently large to make a careful selection in order to include multiple primary tumor stages and multiple distant metastatic sites. The selection of 15 sets included 45 samples (15 primary melanomas of cases, 15 primary melanomas of controls and 15 matched metastasis of cases) of the discovery set of the D-ESMEL study.\u003c/p\u003e\n\u003ch3\u003eLaser capture microdissection for target cell enrichment\u003c/h3\u003e\n\u003cp\u003eFFPE blocks of the 15 sets were sectioned according to the sandwich cutting procedure: a 4 µm section for diagnosis (H\u0026amp;E before); 16 µm sections for LCM-MeD-seq analysis; a set of 3–6 × 8 µm (24 µm) sections for DNA quality determination, and finally, a 4 µm section for pathological confirmation (H\u0026amp;E after). The tumor tissue was annotated based on the H\u0026amp;E by a dermatopathologist (AM) to determine which samples were of sufficient quality for LCM, containing at least 5 mm\u003csup\u003e2\u003c/sup\u003e tumor tissue without necrosis or blood and 30% tumor cellularity. This resulted in 33 samples eligible for Laser capture microdissection (LCM), see Fig.\u0026nbsp;1. For high-quality genome-wide DNA methylation experiments, it is crucial to have the purest possible population of target cells. LCM is an accurate method for isolating these specific target cells out of complex heterogeneous tissue [17]. To improve tissue adhesion to the membrane, PALM membrane slides 1.0 PEN were prior to use UV treated (254 nm) and coated with poly-L-lysine (0.1%) according to manufacturer’s protocol. 16 µm sections of FFPE tissues were cut, mounted on the UV and poly-L-lysine pre-treated membrane slides and incubated at 60°C for 2 hours. Next, the sections were dewaxed, dehydrated and stained with hematoxylin. After digital imaging of the stained slides, regions of interest were annotated by a pathologist. The annotated regions must contain the highest possible percentage of desired target cells (preferably \u0026gt; 70%) surrounded by only normal tissue cells (and preferably no other pathological structures). Using LCM, the percentage of desired cells in the region of interest (cancer lesion) can be increased and regions with unwanted cell types/cell structures can be removed. The annotated regions were cut by a laser beam using the Leica Laser Microdissection system and collected in tubes. For the MeD-seq analysis we need at least 10 (preferably 50) ng DNA per 8.66 µl of a sample. For this we dissected one or more regions that together were 5–20 mm2.\u003c/p\u003e\n\u003ch3\u003eDNA extraction FFPE tissue\u003c/h3\u003e\n\u003cp\u003eDNA was extracted using our Proteinase K protocol. To each LCM sample 50 µl of 1 mg/ml Proteinase K and 200 µl of mineral oil were added. Thereafter, all sample types were incubated 1–2 hours at 70˚C. The samples were then gently mixed and incubated overnight for 16 hours at 70˚C. If the tissue was not completely dissolved, 5 µl 10 mg/ml Proteinase K was added and the samples were incubated for 6 hours at 70˚C. Afterwards, Proteinase K was inactivated by 10 min at 95˚C and the mineral oil was removed. Samples were stored at -20˚C. After the extraction of genomic, the concentrations and quality of the extracted DNA were determined using the DNA QC qPCR.\u003c/p\u003e\n\u003ch3\u003eMeD-seq sample preparation\u003c/h3\u003e\n\u003cp\u003eMeD-seq assays were performed as previously described [11]. In short, genomic DNA and plasma-derived cfDNA were digested with the methylation-dependent restriction enzyme LpnPI (New England Biolabs, Ipswich, MA) generating 32 bp DNA fragments containing the methylated CpG in the middle. All samples were sequenced on the Illumina NextSeq2000 platform. MeD-seq was performed on the 33 LCM tumor tissues of which 3 samples were excluded due to low yield of methylated reads (\u0026lt; 20% of total reads). The remaining samples consisted of 11 primary melanomas of cases, 10 matched metastatic tissues, and 9 primary melanomas of controls. Resulting in 8 complete case-control sets (i.e. 1 primary melanoma of the case and 1 primary melanoma of the control) and 6 complete case-metastasis sets (i.e. 1 primary melanoma of the case + 1 matched metastasis), see Fig.\u0026nbsp;1.\u003c/p\u003e\n\u003cdiv id=\"Sec8\"\u003e\n \u003ch2\u003eMeD-seq data analysis\u003c/h2\u003e\n \u003cp\u003eCustom python scripts were used to process acquired DNA methylation profiles using the following python version and packages: python v3.8.5, numpy v1.19.1, scikit-learn v0.23.2, scipy v1.5.2, matplotlib v3.3.1, seaborn v0.11.0, umap-learn v0.5.1. Raw fastq files were subjected to Illumina adaptor trimming and reads were filtered based on LpnPI restriction site occurrence between 13–17 bp from either the 5′ or 3′ end of the read. Reads that passed the filter were mapped to hg38 using bowtie2 version 2.3.3 and visualized in IGV. Genome-wide individual LpnPI site scores were used to generate read count scores for the following annotated regions: promoter (1 kb before and 1 kb after the transcription start sites (TSS)), CpG-islands and gene bodies (1 kb after TSS till transcription end site). Gene and CpG-island annotations were downloaded from ENSEMBL (Homo_sapiens_hg38.GRCh38.79.gtf, www.ensembl.org).\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eMeD-seq analysis: Genome-wide sliding window\u003c/h3\u003e\n\u003cp\u003eTo detect DMRs, a genome-wide sliding window was used to detect sample-wise and group-wise sequentially differentially methylated LpnPI sites. Genome-wide read counts were normalized, reads per million, for coverage and compared using the Chi-Squared test, with significance set at \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05 and a Bonferroni correction for multiple testing. Neighboring significantly called LpnPI sites were binned and reported. Overlap of genome-wide detected DMRs was reported for promoter, CpG-islands or gene body regions using the annotations of the UCSC database (Hg38) [18]. DMR thresholds were based on LpnPI site count, DMR sizes (in bp) and mean log10 FCs of read counts between group \u003cem\u003ea\u003c/em\u003e and \u003cem\u003eb\u003c/em\u003e (\\(\\:FC=log10\\left(\\frac{\\mu\\:a}{\\mu\\:b}\\right)\\)). All comparisons are made between primary case and control (case-control) sets and primary case and matched metastatic material (case-metastasis) sets.\u003c/p\u003e\n\u003ch3\u003eMeD-seq analysis: Genome-wide promoter, gene body, and CpG-island regions\u003c/h3\u003e\n\u003cp\u003eAlternatively, group-wise DMRs were detected in the known promoter, CpG-island, and gene body regions of the Hg38 genome using a paired t-test with significance set at \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05 and a FDR-BH correction for multiple testing. All comparisons are made between primary case and control (case-control) sets and primary case and matched metastatic material (case-metastasis) sets.\u003c/p\u003e\n\u003cdiv id=\"Sec11\"\u003e\n \u003ch2\u003eUnmatched analyses\u003c/h2\u003e\n \u003cp\u003eNext to the matched sample analysis, DMRs between the unmatched samples were also analyzed in order to assess the impact of two matched control samples which developed a metastases after their matched cases. In this sensitivity analysis these primary controls are changed to primary cases. In this analysis we assess DMRs between primary cases (n = 11) and controls (n = 9), primary case tumors (n = 11). Additionally, DMRs primary tumors (n = 20) and metastatic tumors (n = 11) to gain biological insight and DMRs within the metastatic samples were analyzed to detect DMRs between metastatic lung (n = 3) or GI (n = 3) and all other metastatic sites (n = 7). All DMRs were detected with an unpaired t-test in the known promoter, gene body, and CpG-island regions and corrected for multiple testing with FDR-BH and significance set at \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Results","content":"\u003ch2\u003ePatients characteristics\u003c/h2\u003e\n\u003cp\u003eTo identify prognostic differentially methylated regions (DMRs) in early-stage melanoma, we performed the MeD-seq technique on early-stage melanomas from the D-ESMEL study. In this study early-stage melanoma patients who developed a distant metastasis (case) were matched to those who did not (controls) based on age, sex, Breslow thickness, and ulceration. We aimed to include 10 complete sets, as this should be sufficient to reduce the biological variation and identify robust prognostic DMRs [16]. Each complete set includes 3 samples: a primary melanoma from the case, the matched distant metastasis, and a primary melanoma from the control. \u0026nbsp;Initially 15 candidate sets, i.e. 45 samples, were selected from the discovery set of the D-ESMEL study for MeD-seq analysis, prioritizing primary tumors \u0026gt;1 mm thick to ensure sufficient material for laser capture microdissection (LCM). Following the LCM quality control, a total of 12 samples were excluded due to insufficient tumor area (\u0026lt;5mm\u003csup\u003e2\u003c/sup\u003e tumor area or \u0026lt;30% tumor cellularity) or poor DNA quality. After applying MeD-seq on the remaining 33 samples, three samples were excluded because of low quality, since only 20% of the total reads were methylated. This resulted in 30 total unmatched samples of which 20 primary tumor tissues and 10 metastatic tissues, including 11 primary melanomas from cases, 10 matched metastasis, and 9 primary melanomas of controls (\u003cstrong\u003eAdditional file 1\u003c/strong\u003e). \u0026nbsp;These samples are matched on age, sex, ulceration, and Breslow thickness resulting in 8 complete matched case-control sets (primary melanomas of the case and the control) and 6 matched case-metastasis sets (primary melanoma of the case and the matched distant metastasis) (\u003cstrong\u003eTable 1\u003c/strong\u003e). A complete overview of the inclusion criteria is shown in \u003cstrong\u003eFigure 1\u003c/strong\u003e. The primary analyses were conditional analyses of matched sets. As a sensitivity analyses, unmatched samples were analyzed as well to take all available samples into account.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1.\u0026nbsp;\u003c/strong\u003eClinical-histopathological characteristics of the matched MeD-seq samples.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"635\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" style=\"width: 280px;\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" rowspan=\"2\" style=\"width: 356px;\"\u003e\n \u003cp\u003eCase-control sets\u003cbr\u003e\u0026nbsp;(primary melanomas)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"55\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd height=\"48\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"bottom\" style=\"width: 90px;\"\u003e\n \u003cp\u003eControl\u003cbr\u003e\u0026nbsp;(n=8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 265px;\"\u003e\n \u003cp\u003eCase (n=8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"5\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003cp\u003e(n=8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003eof which with matched metastatic material (n=6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eAge, median (IQR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e54 (50-60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e51 (45-62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e51 (46-59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eSex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eBreslow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e\u0026le;1.0 mm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e\u0026gt;1.0-2.0 mm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e\u0026gt;2.0-4.0 mm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e\u0026gt;4.0 mm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eUlceration\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eAbsent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003ePresent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eStage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e1A\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e1B\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e2A\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e2B\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e2C\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eDistant metastatic site collected\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eLung\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eGastro-intestinal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eAdrenal gland\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eLymph node\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eMuscle\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eSkin\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eTime until distant metastasis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e\u0026lt; 2 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003e2-5 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003en/a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eLocation Primary Melanoma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eScalp\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eTrunk\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eLower extremities\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eUpper extremities\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eWHO subtype\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eSuperficial Spreading\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 280px;\"\u003e\n \u003cp\u003eNodular\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 90px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 85px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 1px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"2\" style=\"width: 0px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eDifferentially methylated regions between matched case-controls and case-metastasis sets\u003c/h2\u003e\n\u003ch3\u003eGenome-wide sliding window analysis reveals a DMR in PTPRN2\u003c/h3\u003e\n\u003cp\u003eDMRs were identified between case-control and case-metastasis sets by comparing the normalized methylated read counts at every LpnP1-site for each matched sample and per group using a genome-wide sliding window and Chi-squared test. After a Bonferroni correction no DMRs were found between both the control-case and case-metastasis groups. By comparing the sets sample-wise (e.g. the primary melanoma of the case of set A vs the primary melanoma of the control of set A) we find 20,752 DMRs between the case-control sets and 7,759 DMRs between the case-metastasis sets. In the sample-wise comparisons a large number of DMRs are identified due to 1:1 comparisons giving an overestimation of the true DMRs because of biological variation, e.g. due to methylated repetitive regions. We found one DMR which was significant within all samples for both the case-control and case-metastasis sets (fold change (FC) \u0026gt; 2; \u003cem\u003ep\u003c/em\u003e\u0026lt;0.05) in the Protein Tyrosine Phosphatase Receptor Type N2 (PTPRN2) gene. We also found a DMR in a CpG-island (CpG20827; chr17:81341464-81342175) which was significant in all case-control sets (\u003cstrong\u003eAdditional file 2\u003c/strong\u003e).\u003c/p\u003e\n\u003ch3\u003eGenome-wide promoter, gene body, and CpG-island regions matched analysis reveals methylated protein coding genes as potential prognostic markers\u003c/h3\u003e\n\u003cp\u003eAlternatively, to the genome-wide sliding window analysis we investigated the known promoter, CpG-island, and gene body methylation regions and tested for differences using the paired t-test followed by a false discovery rate Benjamini-Hochberg (FDR-BH) correction. In this analysis PTPRN2 was not detected as a DMR. We did find 4660 DMRs between case-control sets and 8192 DMRs between case-metastasis sets. However, after\u0026nbsp;FDR-BH correction no significant DMRs remained in both analyses.\u0026nbsp;The most significant site between the case-control sets, before correction, is in the promoter region of the Coiled-Coil-Helix-Coiled-Coil-Helix Domain Containing 2 (\u003cem\u003eCHCHD2\u003c/em\u003e) pseudogene 8 (\u003cem\u003eCHCHD2P8\u003c/em\u003e), see \u003cstrong\u003eTable 2\u003c/strong\u003e. Methylation of CHCHD2P8 decreased in all controls (FC=-0.122; \u003cem\u003ep\u003c/em\u003e=1.71e-05) when comparing it to the matched cases (\u003cstrong\u003eFigure 2a\u003c/strong\u003e). In the top 10 most significant sites (\u003cstrong\u003eAdditional file 3\u003c/strong\u003e) two DMRs in the protein coding genes N-Myc downstream-regulated gene family member 2 (\u003cem\u003eNDRG2\u003c/em\u003e) and Endothelin 2 (\u003cem\u003eEDN2\u003c/em\u003e) were found. \u003cem\u003eNDRG2\u003c/em\u003e showed an increase (FC=0.365; \u003cem\u003ep\u003c/em\u003e=3.17e-05) and EDN2 a decrease (FC=-0.223; \u003cem\u003ep\u003c/em\u003e=4.62e-05) in methylation in all metastatic sites compared to the primary site, see \u003cstrong\u003eFigure 2b and c\u0026nbsp;\u003c/strong\u003erespectively. The remaining top 10 DMRs mostly included long non-coding RNAs (lncRNA), pseudogenes, unknown, or uncharacterized proteins. Interestingly, no single DMR was identified in both the case-control and case-metastasis sets in the top 10 most significant DMRs (\u003cstrong\u003eAdditional file 3\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u0026nbsp;\u003c/strong\u003eMost significant relevant sites across matched and unmatched primary controls, cases, and metastatic tumors.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"642\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003eAnalysis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 93px;\"\u003e\n \u003cp\u003eComparison\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eSite\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eRegion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003ep-value\u003csup\u003e\u0026dagger;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003eFold change\u003csup\u003e\u0026Dagger;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eFunction\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 86px;\"\u003e\n \u003cp\u003eMatched\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 93px;\"\u003e\n \u003cp\u003eCase vs\u003cbr\u003e\u0026nbsp;Control\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eCHCHD2P8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e1.71-05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e-0.122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003ePseudogene\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eNDRG2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e3.17e-05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e0.365\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eProtein coding\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 93px;\"\u003e\n \u003cp\u003eCase vs Metastasis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eEDN2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e4.62e-05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e-0.223\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eProtein coding\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" style=\"width: 86px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eUnmatched\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 93px;\"\u003e\n \u003cp\u003eCase vs\u003cbr\u003e\u0026nbsp;Control\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eCYP2E1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e2.50e-05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e0.857\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eProtein coding\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 93px;\"\u003e\n \u003cp\u003eCase vs Metastasis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eSERPINB8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e6.25e-05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e1.078\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eProtein coding\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 93px;\"\u003e\n \u003cp\u003eMetastatic\u003cbr\u003e\u0026nbsp;lung vs other\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eTAS2R50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e4.87e-03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003ePresent in Lung\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eProtein coding\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 93px;\"\u003e\n \u003cp\u003eMetastatic\u0026nbsp;\u003cbr\u003e\u0026nbsp;GI vs other\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 96px;\"\u003e\n \u003cp\u003eSPAG11B\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003ePromoter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e3.81e-02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003eAbsent in GI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 108px;\"\u003e\n \u003cp\u003eProtein coding\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026dagger; Uncorrected p-value\u003c/p\u003e\n\u003cp\u003e\u0026Dagger; Case is taken as reference\u003c/p\u003e\n\u003cp\u003eWhen investigating loss of methylation we found no methylation site that was uniquely present in either the case, control, or metastasis samples. However, we did find four sites in which methylation was completely lost in the metastasis compared to the primary melanoma of the same case. These are found in the promoter region in the GC vitamin D binding protein (\u003cem\u003eGC\u003c/em\u003e) gene (\u003cem\u003ep\u003c/em\u003e=4.05e-04) (\u003cstrong\u003eFigure 2d\u003c/strong\u003e), and the gene body of ubiquitin-specific peptidase 17-like family member 1 (\u003cem\u003eUSP17L1\u003c/em\u003e) (\u003cem\u003ep\u003c/em\u003e=1.72e-03) (\u003cstrong\u003eFigure 2e\u003c/strong\u003e).\u003c/p\u003e\n\u003ch2\u003eMatched analysis identifies more DMRs in protein-coding genes than unmatched analysis\u003c/h2\u003e\n\u003cp\u003eTo assess the impact of our matching strategy, we conducted a sensitivity analysis using the unmatched case-control and case-metastatic samples (\u003cstrong\u003eAdditional file 1\u003c/strong\u003e). This is of particular importance as, in the matched analysis, two control subjects developed metastases after the matched case. DMRs were detected using an unpaired t-test on the\u0026nbsp;promoter, gene body, and CpG-island regions\u0026nbsp;between the unmatched case control primary tumors. This resulted in no DMRs after FDR-BH correction. However, we report 5561 DMRs before correction. This included DMRs in the promoter of the protein coding genes and ligand of arginyl-tRNA--protein transferase 1 (\u003cem\u003eLIAT1\u003c/em\u003e) (FC=-0.299; \u003cem\u003ep\u003c/em\u003e=1.70e-04) and cytochrome P450 family 2 subfamily E member 1 (\u003cem\u003eCYP2E1\u003c/em\u003e) (FC=0.857; \u003cem\u003ep\u003c/em\u003e=2.50e-05)\u0026nbsp;(\u003cstrong\u003eTable 2\u003c/strong\u003e). \u003cem\u003eCYP2E1\u003c/em\u003e showed a general reduction in methylation in the primary cases (\u003cstrong\u003eFigure 3a\u003c/strong\u003e). \u0026nbsp;The top 10 (\u003cstrong\u003eAdditional file 3\u003c/strong\u003e) also included one long intergenic non-protein coding RNA (lincRNA) which is an RNA type known to play a role in cancer development. Interestingly, none of the top 10 DMRs overlapped with DMRs found in the eight matched case-control sets.\u003c/p\u003e\n\u003cp\u003eAn unpaired t-test on the promoter, gene body, and CpG-island regions between the tumor of primary cases and metastatic tumors resulted in no DMRs after FDR-BH correction. However, we report 5980 DMRs before correction. This included one DMR in a protein coding gene, namely in the promoter region of serpin family B member 8 (\u003cem\u003eSERPINB8\u003c/em\u003e) (FC=1.078; \u003cem\u003ep\u003c/em\u003e=6.25e-05) (\u003cstrong\u003eTable 2\u003c/strong\u003e), in which a reduction of methylation was detected in all the primary cases (\u003cstrong\u003eFigure 3b\u003c/strong\u003e). Again, the top 10 list of DMRs do not overlap with DMRs found in the 6 matched case-metastasis sets (\u003cstrong\u003eAdditional file 3\u003c/strong\u003e). DMRs in lincRNAs are detected which were also found between the unmatched case-controls. The DMR with the lowest \u003cem\u003ep\u003c/em\u003e-value is located in the promoter region of the nuclear pore associated protein 1 (\u003cem\u003eNPAP1\u003c/em\u003e) pseudogene 6 (\u003cem\u003eNPAP1P6\u003c/em\u003e) (FC=-0.19; \u003cem\u003ep\u003c/em\u003e=3.97e-07). Aside from this pseudogene, three other pseudogenes were identified, including glycine-N-acyltransferase like 1 (\u003cem\u003eGLYATL1\u003c/em\u003e) pseudogene 2 (\u003cem\u003eGLYATL1P2\u003c/em\u003e) (Absent in metastasis; \u003cem\u003ep\u003c/em\u003e=2.74e-05), cutaneous T cell lymphoma-associated antigen family, member 5 (\u003cem\u003eCTAGE5\u003c/em\u003e) pseudogene (FC=-0.838; \u003cem\u003ep\u003c/em\u003e=3.65e-05), and RP4-640E24.1 (FC=-1.01; \u003cem\u003ep\u003c/em\u003e=4.46e-05).\u003c/p\u003e\n\u003cp\u003eUnmatched comparison of primary and metastatic tumors fails to detect meaningful DMRs\u003c/p\u003e\n\u003cp\u003eTo gain biological insight between the tumor of the primary case and its metastasis we performed an unpaired t-test on the promoter, gene body, and CpG-island regions. This resulted in no DMRs after FDR-BH correction. However, we report 12,472 DMRs before correction. No DMRs in protein coding genes were detected in the ten most significant DMRs (\u003cstrong\u003eAdditional file 3\u003c/strong\u003e). The top 10 did include DMRs in the promoter region of four pseudogenes, namely RP11-72M10.8 \u0026nbsp;(FC=0.402; \u003cem\u003ep\u003c/em\u003e=5.42e-06), family with sequence similarity 27-like (FAM27L) pseudogene (FC=0.155; \u003cem\u003ep\u003c/em\u003e=8.95e-06), family with sequence similarity 27 (FAM27) pseudogene (FC=0.147; \u003cem\u003ep\u003c/em\u003e=1.64e-05), and RNA, U6 small nuclear 162, pseudogene (RNU6-162P) (FC=1.17; \u003cem\u003ep\u003c/em\u003e=1.39e-05).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;TAS2R50 and SPAG11B are differentially methylated in metastatic lung and GI\u003c/p\u003e\n\u003cp\u003eAn unpaired t-test between the methylated regions of lung metastasis (n=3), gastro-intestinal (GI) metastasis (n=3), and all other metastatic samples (n=7) in the promoter, gene body, and CpG-island region resulted in four DMRs in lung metastasis and three DMRs in GI metastasis after FDR-BH correction. In total, we report 2911 DMRs in lung metastasis and 10,805 DMRs in GI metastasis. In the \u0026nbsp;lung metastasis one DMR was found to be present in a protein coding gene in the promoter region of the taste 2 receptor member 50 (\u003cem\u003eTAS2R50\u003c/em\u003e) gene \u003cem\u003e(p\u003c/em\u003e=4.87e-03)\u003cem\u003e.\u003c/em\u003e In the GI metastasis one DMR was found to be absent in a protein coding gene in the promoter region of sperm associated antigen 11B (\u003cem\u003eSPAG11B\u003c/em\u003e) gene \u003cem\u003e(p\u003c/em\u003e=3.81e-02) (\u003cstrong\u003eTable 2\u003c/strong\u003e)\u003cem\u003e.\u0026nbsp;\u003c/em\u003eMost other DMRs are located in CpG-islands and one DMR was found in the promoter region of a lincRNA (\u003cstrong\u003eAdditional file 3\u003c/strong\u003e).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we performed genome-wide DNA methylation analyses to uncover potential DNA methylation patterns that are associated with the development of distant metastasis in early-stage cutaneous melanoma. We applied MeD-seq essay which allows higher coverage (50%) and resolution compared to traditional array-based techniques (2\u0026ndash;4%). Primary melanomas and metastatic samples from the D-ESMEL study were used to identify DMRs that are related with melanoma progression and not associated with known prognostic features (Breslow thickness, ulceration, age, sex). Our analysis revealed several differently methylation regions in promoters, coding regions of protein coding genes, in addition to non-coding RNAs that warrant further investigation for their role in melanoma progression.\u003c/p\u003e\u003cp\u003eWhen comparing metastasized primary tumors (cases) to their matched primary melanoma of controls we found eight candidate DMRs, located in the promotor regions of the genes \u003cem\u003eCYP2E1\u003c/em\u003e, \u003cem\u003ePTPRN2\u003c/em\u003e, \u003cem\u003eCHCHD2, NDRG2\u003c/em\u003e, \u003cem\u003eEDN2\u003c/em\u003e, \u003cem\u003eGC\u003c/em\u003e, \u003cem\u003eUSP17L1\u003c/em\u003e, and \u003cem\u003eSERPINB8\u003c/em\u003e, that may reflect the metastasis transition from the primary to distant organs. Firstly, we found a general decrease in DNA methylation in the promotor of the \u003cem\u003eCYP2E1\u003c/em\u003e gene in unmatched primary cases compared to controls. One study found that forced expression of \u003cem\u003eCYP2E1\u003c/em\u003e inhibited the growth of malignant melanoma cells in vivo using bone marrow-derived mesenchymal stem cells to deliver \u003cem\u003eCYP2E1\u003c/em\u003e transduced with a pAd5-CMV-\u003cem\u003eCYP2E1\u003c/em\u003e recombinant adenovirus [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. These findings may contradict our observed decreased methylation in the \u003cem\u003eCYP2E1\u003c/em\u003e promotor region, which may lead to increased expression of CYP2E1, and has been reported to increase expression of CYP2E1 [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] in other diseases. However, it is unlikely that growth of melanoma cells is inhibited in primary melanomas that metastasized, as larger tumor size is associated with poorer outcomes [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. A decrease in methylation of the \u003cem\u003eCYP2E1\u003c/em\u003e promotor region has also been reported to increase expression of CYP2E1 [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] in other diseases. Secondly, for \u003cem\u003ePTPRN2\u003c/em\u003e, we found methylation differences in the promoter in each case compared to its matched control. \u003cem\u003ePTPRN2\u003c/em\u003e has been reported to be overexpressed in tumors such as breast, pancreatic, and colon cancer. In addition, it is reported to be prognostic for overall survival in pancreatic and colon cancer [\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] and might also be overexpressed and prognostic in melanoma. Finally, the functional gene \u003cem\u003eCHCHD2\u003c/em\u003e of the pseudogene \u003cem\u003eCHCHD2P8\u003c/em\u003e has been discussed as potential prognostic factor for cancer and target for cancer therapy, as it triggers oxidative phosphorylation causing increased proliferation and metastasis [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. However, no functionality of \u003cem\u003eCHCHD2P8\u003c/em\u003e itself has been reported.\u003c/p\u003e\u003cp\u003eAdditionally, we identified six DMRs in potentially relevant genes which might help to better understand the development of metastasis in early-stage melanoma. Four of these genes (\u003cem\u003ePTPRN2\u003c/em\u003e, \u003cem\u003eNDRG2\u003c/em\u003e, \u003cem\u003eEDN2\u003c/em\u003e, \u003cem\u003eand GC\u003c/em\u003e) have been reported to be related to cancer in general but not specifically for melanoma. It was found that over-expression of \u003cem\u003eNDRG2\u003c/em\u003e could inhibit tumor growth and invasion [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Additionally, methylation of \u003cem\u003eNDRG2\u003c/em\u003e was found to be the main cause for its down regulation in gastric cancer [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. \u003cem\u003eEDN2\u003c/em\u003e is a protein coding gene and part of the endothelin protein family. Endothelins and their receptors have been associated with melanoma progression through the alteration of tumor-host interactions [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. A decrease in methylation in the promoter of EDN2 was reported in the metastatic site compared to the primary site. How the methylation of EDN2 affects its expression in melanoma remains uncertain. The GC protein binds to circulating vitamin D and activates macrophages [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e] and an abundance of the GC protein decreases tumor spreading [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. The exact relevance and effect of methylation of GC in melanoma remains uncertain since no methylation was reported in the promoter of any of the metastatic samples. Lastly, we reported DMRs in \u003cem\u003eUSP17L1\u003c/em\u003e and \u003cem\u003eSERPINB8\u003c/em\u003e which have been known to be involved in melanomas. In USP17L1, a loss of methylation was detected at the metastatic site compared to the primary in the gene body region. Methylation of the gene body could have different effects on \u003cem\u003eUSP17L1\u003c/em\u003e such as an increase of expression, impact on the splicing of the gene, and possibly cause a loss of regulation of the Ras processing pathway [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. In \u003cem\u003eSERPINB8\u003c/em\u003e, a DMR was detected in the promoter region. \u003cem\u003eIn vitro\u003c/em\u003e experiments of BRAF V600E mutant melanoma found that SERPINB8 regulates the expression of integrin alpha x (\u003cem\u003eITGAX\u003c/em\u003e) in melanoma cells, a gene which significantly promotes the invasion and proliferation of melanoma cells [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. However, methylation of SERPINB8 and its effect on its expression has not been studied.\u003c/p\u003e\u003cp\u003eNo relevant DMRs were found when comparing primary and metastatic tumors as we report mostly DMRs in pseudogenes with unknown functions. When comparing methylation patterns between metastatic sites (lung, GI, and others) two significant DMRs were reported \u003cem\u003eTAR2R50\u003c/em\u003e in lung metastasis and \u003cem\u003eSPAG11B\u003c/em\u003e in GI metastasis. A paralog of \u003cem\u003eSPAG11B\u003c/em\u003e, \u003cem\u003eSPAG11A\u003c/em\u003e, was found to be differentially expressed in gastric carcinoma [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] and \u003cem\u003eTAR2R50\u003c/em\u003e encodes a bitter taste receptors part of the taste family 2 receptors (T2R) [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e], which were associated, together with family 1 receptors (T1R), to survival differences in 12 solid tumor subtypes, including one in lung adenocarcinoma and melanoma.\u003c/p\u003e\u003cp\u003eThe strength of our study is that we analyze the differences between matched (metastatic) melanomas to uncover truly relevant DMRs in melanoma covering a wide-range of CpG sites. Matching case and controls on known prognostic factors allows a more robust identification of new prognostic factors, rather than identifying methylation patterns related to known prognostic factors. Moreover, matching primary tumors and their corresponding metastases is particularly robust, as both tissues originate from the same patient, minimizing variability due to genetic background differences. In addition, we used MeD-seq allowing us to uncover CpG sites genome-wide instead of a limited number of CpG sites using array-based methods. To our standing, this is the first study analyzing matched primary case-control and matched primary case and metastasis sets across such a wide number of methylated regions.\u003c/p\u003e\u003cp\u003eWhile the matched design of the D-ESMEL study is a strength for detecting new prognostic factors, the limitation of this design is that methylation patterns related to important biological processes could not be detected. Small differences in methylation patterns were observed between the matched primary tumors of cases and controls, which are likely due to the matching on known prognostic factors. As a result, we did not observe DMRs that were prognostic over and above Breslow thickness, ulceration, age and sex. While this is important information for clinical practice, the consequence of this study design, is that any methylation patterns that are related to either Breslow thickness or ulceration cannot be detected in our matched sets. Although our sample size was sufficient to avoid detecting false positive DMRs, possibly additional samples may help with detecting very small differences in DMRs, independent of known prognostic factors. However, the clinical relevance of those small differences is unknown. We also observed small differences between primary tumors and its matched metastasis, which could be attributed to intratumoral heterogeneity introduced during the metastatic process.\u003c/p\u003e\u003cp\u003eAnother aspect of the study design that is both a strength and a limitation, is the matching of the controls on follow-up time. The control needed to have at least the same amount of follow-up as the case. While this design assures that risk estimates are comparable to a cohort study (as controls are sampled from the risk set of each case in a cohort study), a limitation of this design is that controls can develop metastasis later during follow-up. During this study, we updated the follow-up of all patients and we observed, that two control samples developed a metastases after their matched cases. These controls may have been at lower risk compared to its matched case as the controls developed a metastasis later than its matched case. Therefore, these are still relevant samples to investigate to further understand the development of metastasis. On the other hand, the inclusion of these controls may have further reduced the small differences between cases and controls. Therefore, we also performed a sensitivity analysis where the matching was not included and in which those 2 controls were considered cases. These analysis also did not reveal prognostic DMRs.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eTo our knowledge, this is the first study to investigate the methylation of matched primary and metastatic early-stage melanomas using the MeD-seq essay. We did not identify DMRs in early-stage primary melanomas that are prognostic for developing distant metastasis independent from known prognostic factors. We identified 8 DMRs that may play a biological role in the development of metastasis and which can be further investigated as possible epigenetic drivers of melanoma progression.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgments\u003c/h2\u003e\n\u003cp\u003eWe would like to thank IKNL for providing the clinical data and Catherine Zhou for collecting and curating the clinical data and FFPE blocks.\u003c/p\u003e\n\u003ch2\u003eAuthor Contributions\u003c/h2\u003e\n\u003cp\u003eConceptualization: LH, RB; Data Curation: AM, TK; Formal analysis: JO, RB; Investigation: TK, JO, RB; Methodology: JO, RB; Supervision: LH, YL, RB; Writing \u0026ndash; Original Draft: JO, TK, RB; Writing \u0026ndash; Reviewing \u0026amp; Editing: LH, YL, MW, AM, JG, JB, RB.\u003c/p\u003e\n\u003ch2\u003eData Availability Statement\u003c/h2\u003e\n\u003cp\u003eThe clinical data from these patients was provided to the authors by the Netherland Comprehensive Cancer Organisation (IKNL). These data are not publicly available and restrictions apply to the availability of the data used for the current study. However, the MeD-seq data and clinical data are available upon reasonable request to the authors and with permission of the Netherlands Comprehensive Cancer Organisation.\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eNo funding was provided for this work.\u003c/p\u003e\n\u003ch2\u003eClinical Trial Number\u003c/h2\u003e\n\u003cp\u003eClinical trial number: not applicable.\u003c/p\u003e\n\u003ch2\u003eCompeting interest\u003c/h2\u003e\n\u003cp\u003eThe authors declare no conflict of interest or financial interests except for RB, JB and JG, who report being shareholder in Methylomics B.V., a commercial company that applies MeD-seq to develop methylation markers for cancer staging.\u003c/p\u003e\n\u003ch2\u003eConsent for publication\u003c/h2\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch2\u003eEthics approval statement and consent to participate\u003c/h2\u003e\n\u003cp\u003eThe use of leftover diagnostic tissue samples for scientific research is based on the \u0026lsquo;no objection\u0026rsquo; principle, as outlined in the Code of Conduct for Health Research by the Committee on Regulation of Health Research.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eZhou C, Louwman M, Wakkee M, van der Veldt A, Gr\u0026uuml;nhagen D, Verhoef C, et al. Primary Melanoma Characteristics of Metastatic Disease: A Nationwide Cancer Registry Study. Cancers (Basel). 2021;13:4431. https://doi.org/10.3390/cancers13174431.\u003c/li\u003e\n\u003cli\u003eEnninga EAL, Moser JC, Weaver AL, Markovic SN, Brewer JD, Leontovich AA, et al. Survival of cutaneous melanoma based on sex, age, and stage in the United States, 1992\u0026ndash;2011. Cancer Med. 2017;6:2203\u0026ndash;12. https://doi.org/10.1002/cam4.1152.\u003c/li\u003e\n\u003cli\u003eIsaksson K, Mikiver R, Eriksson H, Lapins J, Nielsen K, Ingvar C, et al. Survival in 31 670 patients with thin melanomas: a Swedish population‐based study*. British Journal of Dermatology. 2021;184:60\u0026ndash;7. https://doi.org/10.1111/bjd.19015.\u003c/li\u003e\n\u003cli\u003eLuke JJ, Rutkowski P, Queirolo P, Del Vecchio M, Mackiewicz J, Chiarion-Sileni V, et al. Pembrolizumab versus placebo as adjuvant therapy in completely resected stage IIB or IIC melanoma (KEYNOTE-716): a randomised, double-blind, phase 3 trial. The Lancet. 2022;399:1718\u0026ndash;29. https://doi.org/10.1016/S0140-6736(22)00562-1.\u003c/li\u003e\n\u003cli\u003eKirkwood JM, Del Vecchio M, Weber J, Hoeller C, Grob J-J, Mohr P, et al. Adjuvant nivolumab in resected stage IIB/C melanoma: primary results from the randomized, phase 3 CheckMate 76K trial. Nat Med. 2023;29:2835\u0026ndash;43. https://doi.org/10.1038/s41591-023-02583-2.\u003c/li\u003e\n\u003cli\u003eFietz S, Zarbl R, Niebel D, Posch C, Brossart P, Gielen GH, et al. CTLA4 promoter methylation predicts response and progression-free survival in stage IV melanoma treated with anti-CTLA-4 immunotherapy (ipilimumab). Cancer Immunology, Immunotherapy. 2021;70:1781\u0026ndash;8. https://doi.org/10.1007/s00262-020-02777-4.\u003c/li\u003e\n\u003cli\u003eMarzese DM, Scolyer RA, Huynh JL, Huang SK, Hirose H, Chong KK, et al. Epigenome-wide DNA methylation landscape of melanoma progression to brain metastasis reveals aberrations on homeobox D cluster associated with prognosis. Hum Mol Genet. 2014;23:226\u0026ndash;38. https://doi.org/10.1093/hmg/ddt420.\u003c/li\u003e\n\u003cli\u003eAkbani R, Akdemir KC, Aksoy BA, Albert M, Ally A, Amin SB, et al. Genomic Classification of Cutaneous Melanoma. Cell. 2015;161:1681\u0026ndash;96. https://doi.org/10.1016/j.cell.2015.05.044.\u003c/li\u003e\n\u003cli\u003eWouters J, Vizoso M, Martinez-Cardus A, Carmona FJ, Govaere O, Laguna T, et al. Comprehensive DNA methylation study identifies novel progression-related and prognostic markers for cutaneous melanoma. BMC Med. 2017;15:101. https://doi.org/10.1186/s12916-017-0851-3.\u003c/li\u003e\n\u003cli\u003eConway K, Edmiston SN, Vondras A, Reiner A, Corcoran DL, Shen R, et al. DNA Methylation Classes of Stage II and III Primary Melanomas and Their Clinical and Prognostic Significance. JCO Precis Oncol. 2024. https://doi.org/10.1200/PO-24-00375.\u003c/li\u003e\n\u003cli\u003eBoers R, Boers J, De Hoon B, Kockx C, Ozgur Z, Molijn A, et al. Genome-wide DNA methylation profiling using the methylation-dependent restriction enzyme LpnPI. Genome Res. 2018;28:88\u0026ndash;99. https://doi.org/10.1101/gr.222885.117.\u003c/li\u003e\n\u003cli\u003eFu S, Deger T, Boers RG, Boers JB, Doukas M, Gribnau J, et al. Hypermethylation of DNA Methylation Markers in Non-Cirrhotic Hepatocellular Carcinoma. Cancers (Basel). 2023;15:4784. https://doi.org/10.3390/cancers15194784.\u003c/li\u003e\n\u003cli\u003eBos MK, Verhoeff SR, Oosting SF, Menke-van der Houven van Oordt WC, Boers RG, Boers JB, et al. Methylated Cell-Free DNA Sequencing (MeD-seq) of LpnPI Digested Fragments to Identify Early Progression in Metastatic Renal Cell Carcinoma Patients on Watchful Waiting. Cancers (Basel). 2023;15:1374. https://doi.org/10.3390/cancers15051374.\u003c/li\u003e\n\u003cli\u003eSmit KN, Boers R, Vaarwater J, Boers J, Brands T, Mensink H, et al. Genome-wide aberrant methylation in primary metastatic UM and their matched metastases. Sci Rep. 2022;12:42. https://doi.org/10.1038/s41598-021-03964-8.\u003c/li\u003e\n\u003cli\u003eZhou C, Mooyaart AL, Kerkour T, Louwman MWJ, Wakkee M, Li Y, et al. The Dutch Early-Stage Melanoma (D-ESMEL) study: a discovery set and validation cohort to predict the absolute risk of distant metastases in stage I/II cutaneous melanoma. Eur J Epidemiol. 2025;40:27\u0026ndash;42. https://doi.org/10.1007/s10654-024-01188-4.\u003c/li\u003e\n\u003cli\u003eBoers J, Boers R, Sakoltchik J, Dasgupta S, Martens L, Tadema KCD, et al. DNA methylation database for gynecological cancer detection, classification and assay development. 2024. https://doi.org/10.1101/2024.07.01.601485.\u003c/li\u003e\n\u003cli\u003eQuint W, Jenkins D, Molijn A, Struijk L, van de Sandt M, Doorbar J, et al. One virus, one lesion\u0026mdash;individual components of CIN lesions contain a specific HPV type. J Pathol. 2012;227:62\u0026ndash;71. https://doi.org/10.1002/path.3970.\u003c/li\u003e\n\u003cli\u003ePerez G, Barber GP, Benet-Pages A, Casper J, Clawson H, Diekhans M, et al. The UCSC Genome Browser database: 2025 update. Nucleic Acids Res. 2025;53:D1243\u0026ndash;9. https://doi.org/10.1093/nar/gkae974.\u003c/li\u003e\n\u003cli\u003eWang J, Ma D, Li Y, Yang Y, Hu X, Zhang W, et al. Targeted delivery of CYP2E1 recombinant adenovirus to malignant melanoma by bone marrow-derived mesenchymal stem cells as vehicles. Anticancer Drugs. 2014;25:303\u0026ndash;14. https://doi.org/10.1097/CAD.0000000000000046.\u003c/li\u003e\n\u003cli\u003eKaut O, Schmitt I, W\u0026uuml;llner U. Genome-scale methylation analysis of Parkinson\u0026rsquo;s disease patients\u0026rsquo; brains reveals DNA hypomethylation and increased mRNA expression of cytochrome P450 2E1. Neurogenetics. 2012;13:87\u0026ndash;91. https://doi.org/10.1007/s10048-011-0308-3.\u003c/li\u003e\n\u003cli\u003eGershenwald JE, Scolyer RA, Hess KR, Sondak VK, Long G V., Ross MI, et al. Melanoma staging: Evidence‐based changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA Cancer J Clin. 2017;67:472\u0026ndash;92. https://doi.org/10.3322/caac.21409.\u003c/li\u003e\n\u003cli\u003eSengelaub CA, Navrazhina K, Ross JB, Halberg N, Tavazoie SF. PTPRN2 and PLC\u0026beta;1 promote metastatic breast cancer cell migration through PI(4,5)P2‐dependent actin remodeling. EMBO J. 2016;35:62\u0026ndash;76. https://doi.org/10.15252/embj.201591973.\u003c/li\u003e\n\u003cli\u003eFeigin ME, Garvin T, Bailey P, Waddell N, Chang DK, Kelley DR, et al. Recurrent noncoding regulatory mutations in pancreatic ductal adenocarcinoma. Nat Genet. 2017;49:825\u0026ndash;33. https://doi.org/10.1038/ng.3861.\u003c/li\u003e\n\u003cli\u003eYin J, Guo Y. HOXD13 promotes the malignant progression of colon cancer by upregulating PTPRN2. Cancer Med. 2021;10:5524\u0026ndash;33. https://doi.org/10.1002/cam4.4078.\u003c/li\u003e\n\u003cli\u003eGundamaraju R, Lu W, Manikam R. CHCHD2: The Power House\u0026rsquo;s Potential Prognostic Factor for Cancer? Front Cell Dev Biol. 2021;8. https://doi.org/10.3389/fcell.2020.620816.\u003c/li\u003e\n\u003cli\u003eLiu N, Wang L, Li X, Yang Q, Liu X, Zhang J, et al. N-Myc downstream-regulated gene 2 is involved in p53-mediated apoptosis. Nucleic Acids Res. 2008;36:5335\u0026ndash;49. https://doi.org/10.1093/nar/gkn504.\u003c/li\u003e\n\u003cli\u003eChoi S-C, Yoon SR, Park YP, Song EY, Kim JW, Kim WH, et al. Expression of NDRG2 is related to tumor progression and survival of gastric cancer patients through Fas-mediated cell death. Exp Mol Med. 2007;39:705\u0026ndash;14. https://doi.org/10.1038/emm.2007.77.\u003c/li\u003e\n\u003cli\u003eChang X, Li Z, Ma J, Deng P, Zhang S, Zhi Y, et al. DNA Methylation of NDRG2 in Gastric Cancer and Its Clinical Significance. Dig Dis Sci. 2013;58:715\u0026ndash;23. https://doi.org/10.1007/s10620-012-2393-z.\u003c/li\u003e\n\u003cli\u003eSaldana-Caboverde A, Kos L. Roles of endothelin signaling in melanocyte development and melanoma. Pigment Cell Melanoma Res. 2010;23:160\u0026ndash;70. https://doi.org/10.1111/j.1755-148X.2010.00678.x.\u003c/li\u003e\n\u003cli\u003eSaburi E, Saburi A, Ghanei M. Promising role for Gc-MAF in cancer immunotherapy: From bench to bedside. Caspian Journal of Internal Medicine. 2017;8:228\u0026ndash;38. https://doi.org/10.22088/cjim.8.4.228.\u003c/li\u003e\n\u003cli\u003eDanahy L, Long C, Hofmann TJ, Tara Z, Mark J, Roizen JD. Dietary vitamin D is a novel modulator of tumor engraftment through regulation of GC protein abundance. 2024. https://doi.org/10.21203/rs.3.rs-3911213/v1.\u003c/li\u003e\n\u003cli\u003eWang Q, Xiong F, Wu G, Liu W, Chen J, Wang B, et al. Gene body methylation in cancer: molecular mechanisms and clinical applications. Clin Epigenetics. 2022;14:154. https://doi.org/10.1186/s13148-022-01382-9.\u003c/li\u003e\n\u003cli\u003eNi L, Li P, Li M, Huang S, Dang N. SERPINB8 and furin regulate ITGAX expression and affect the proliferation and invasion of melanoma cells. Exp Dermatol. 2023;32:24\u0026ndash;9. https://doi.org/10.1111/exd.14677.\u003c/li\u003e\n\u003cli\u003eLi DF, Wang NN, Chang X, Wang SL, Wang LS, Yao J, et al. Bioinformatics analysis suggests that COL4A1 may play an important role in gastric carcinoma recurrence. J Dig Dis. 2019;20:391\u0026ndash;400. https://doi.org/10.1111/1751-2980.12758.\u003c/li\u003e\n\u003cli\u003eWei K, Hill BL, Thompson JC, Miller ZA, Mueller A, Lee RJ, et al. Bitter Taste Receptor Agonists Induce Apoptosis in Papillary Thyroid Cancer. Head Neck. 2025. https://doi.org/10.1002/hed.28120.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"human-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"hugm","sideBox":"Learn more about [Human Genomics](http://humgenomics.biomedcentral.com/)","snPcode":"40246","submissionUrl":"https://submission.nature.com/new-submission/40246/3","title":"Human Genomics","twitterHandle":"@OAgenetics","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Melanoma, Metastasis, Methylation, Genome-wide, D-ESMEL","lastPublishedDoi":"10.21203/rs.3.rs-7642695/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7642695/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eEarly-stage (stage I-II) cutaneous melanoma accounts for the majority of melanoma diagnoses, but more than 40% of patients who die due to melanoma were initially diagnosed with an early-stage melanoma.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eThe aim of this study was to identify prognostic genome-wide methylation markers of metastasized primary early-stage melanomas and retrieving biological insights from its matched distant metastasis. We selected samples from the Dutch Early-Stage Melanoma (D-ESMEL) study, representing case-control sets where the primary melanoma of each metastatic case is matched to a primary melanoma of a control based on known clinical risk factors. Matched distant metastasis were also retrieved. Laser capture microdissection was performed to isolate the tumor tissue, where after a genome-wide methylated DNA sequencing (MeD-seq) was conducted. Differentially methylated regions (DMR) between primary tumors of the cases-control sets and the tumor of the primary case and its metastasis were tested using Chi-squared test with a genome-wide sliding window analysis, as well as a paired t-test in predefined promotor, gene body, and CpG-island regions.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eMeD-seq analyses did not reveal prognostic methylation markers in primary melanomas, which have additional prognostic value on top of known clinical risk factors We identified eight protein coding genes with the largest methylation difference between primary melanomas of patients with and without metastasis and between primary melanomas and matched distant metastasis: \u003cem\u003eCYP2E1\u003c/em\u003e, \u003cem\u003ePTPRN2\u003c/em\u003e, \u003cem\u003eCHCHD2, NDRG2\u003c/em\u003e, \u003cem\u003eEDN2\u003c/em\u003e, \u003cem\u003eGC\u003c/em\u003e, \u003cem\u003eUSP17L1\u003c/em\u003e, and \u003cem\u003eSERPINB8.\u003c/em\u003e\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003eThis study found 8 genes that have been implicated in primary tumors or metastasis of other cancers which require further investigation into their involvement of metastasis in melanoma.\u003c/p\u003e","manuscriptTitle":"Genome-Wide Methylation Profiles of Primary and Matched Distant Metastasis: Insights from the Dutch Early-Stage Melanoma (D-Esmel) Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-08 07:15:12","doi":"10.21203/rs.3.rs-7642695/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-15T07:06:13+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-10T19:35:43+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-10T07:25:02+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"61353603070526745934424236463744998736","date":"2025-09-29T08:43:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"187117209277819466435275403530755341084","date":"2025-09-28T16:50:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"31876390881340925020916473652433840596","date":"2025-09-26T15:16:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"67352357666051121321379513473506206025","date":"2025-09-26T10:59:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"135642147511193507443444705020933037310","date":"2025-09-25T09:32:07+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-24T11:04:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-22T12:18:40+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-22T12:18:12+00:00","index":"","fulltext":""},{"type":"submitted","content":"Human Genomics","date":"2025-09-17T17:36:16+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"human-genomics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"hugm","sideBox":"Learn more about [Human Genomics](http://humgenomics.biomedcentral.com/)","snPcode":"40246","submissionUrl":"https://submission.nature.com/new-submission/40246/3","title":"Human Genomics","twitterHandle":"@OAgenetics","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"be733f25-261b-469c-ae1e-45c95e483dcd","owner":[],"postedDate":"October 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-08T16:00:46+00:00","versionOfRecord":{"articleIdentity":"rs-7642695","link":"https://doi.org/10.1186/s40246-025-00871-1","journal":{"identity":"human-genomics","isVorOnly":false,"title":"Human Genomics"},"publishedOn":"2025-12-06 15:57:02","publishedOnDateReadable":"December 6th, 2025"},"versionCreatedAt":"2025-10-08 07:15:12","video":"","vorDoi":"10.1186/s40246-025-00871-1","vorDoiUrl":"https://doi.org/10.1186/s40246-025-00871-1","workflowStages":[]},"version":"v1","identity":"rs-7642695","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7642695","identity":"rs-7642695","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}