A 50-year-old male patient was diagnosed with stage IB cutaneous melanoma (according to the 8th Edition of the American Joint Committee on Cancer (15)) in 2009 after surgical resection of a 2 mm thick lesion on the right thigh. Sentinel node biopsy showed no metastasis, and the patient remained disease-free for 4 years. In 2013, a satellite cutaneous metastasis was excised. In 2014, a nodule on the right inner thigh led to a wide excision, revealing intranodal metastases in 5 lymph nodes. A subsequent right inguinal lymphadenectomy showed no further metastasis in 34 lymph nodes.

In 2015, a cutaneous nodule was excised from the thoracic wall, confirmed to be a melanoma metastasis harboring the BRAF V600E mutation. Additionally, right iliac lymph node metastases were excised, and the patient began treatment with Vemurafenib plus Cobimetinib in 2016. A 2017 PET scan showed right external iliac adenopathy, and the patient remained under surveillance, as he had already undergone surgery in that region.

In 2021, PET scan revealed right adrenal gland metastasis, prompting an adrenalectomy. Later that year, a bleeding cutaneous metastasis was excised from the abdominal wall, but a subsequent PET scan showed disease progression with lymph node, subcutaneous, intramuscular, and perirenal involvement. The patient discontinued targeted therapy and started anti-PD-1 with pembrolizumab.

By April 2022, further metastases in the left thorax and hypochondrium prompted a segmental resection of the colon. As the disease progressed, the patient underwent surgical resection of multiple metastatic foci in the intra-abdominal, retroperitoneal, cervical, gluteal and thoracic regions in early 2023. However, with continued disease progression, pembrolizumab treatment was discontinued in August, and the patient began a new regimen with ipilimumab plus nivolumab (induction treatment), followed by maintenance of nivolumab monotherapy to date.

The patient’s history is summarized in Figure 1 . This study was approved by the Instituto Português de Oncologia de Lisboa Francisco Gentil (IPOLFG) Ethics Board Committee (UIC/1310) and written informed consent was obtained from the patient.

During metastases excision in 2023, a surgical fragment from a retroperitoneal metastasis was taken for cell culture. The tumor fragment was mechanically digested under sterile conditions with a scalpel (~1 mm diameter pieces). Subsequently, fragments were seeded in 1 well of a 24-well tissue culture plate in 500 μL McCoy’s 5A Medium (American Type Culture Collection, ATCC, 30-2007) supplemented with 10% Fetal Bovine Serum (FBS) (Corning 35-079-CV) and 1% Pen Strep (P/S) (Gibco 15140-122, Thermo Fisher Scientific). The petri dish where the digestion took place was also washed with 1 mL McCoy’s medium and placed in 2 additional wells (500 μL per well). Media was replenished every other day until wells reached confluency, one week later. Cells were detached with TrypLE™ Express Enzyme (Gibco 12605-028, Thermo Fisher Scientific) for approximately 5 min and subsequently sub-cultured. This primary cell culture was designated MelT79.

Commercially available human malignant CM cell lines A375 (ATCC CRL-1619, BRAF V600E, non-pigmented) and WM115 (Rockland WM115-01-0001, BRAF V600D, non-pigmented) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) high glucose (Gibco 41965-039, Thermo Fisher Scientific), supplemented with 10% FBS and 1% P/S. G-361 cells (ATCC CRL-1424, BRAF V600E, lightly pigmented) were cultured in McCoy’s 5A Medium supplemented with 10% FBS and 1% P/S. MNT-1 cells (ATCC CRL-3450, BRAF V600E, highly pigmented) were cultured in DMEM high glucose with pyruvate (Gibco 21969-035, Thermo Fisher Scientific) supplemented with 10% FBS, 1% P/S, 1% L-glutamine (Gibco 25030-081, Thermo Fisher Scientific) and 1X MEM non-essential amino acids (Gibco 11140-050, Thermo Fisher Scientific). Cells were detached with 0.05% Trypsin 0.53 mM EDTA (Corning 25-052-CI) for approximately 5 min and split to new culture vessels according to the experimental standard procedures.

All cell cultures were maintained in a humidified incubator at 37°C, 5% CO₂, and were regularly checked for mycoplasma contamination using Universal Mycoplasma Detection Kit (ATCC 30-1012K).

Metaphases of the MelT79 cell line were obtained at passage 15 and multicolor fluorescence in situ hybridization (M-FISH) analysis was performed with 24XCyte Human Multicolor FISH Probe (MetaSystems) according to the manufacturer’s protocol. Fluorochromes were sequentially captured in a Zeiss Imager Z1 microscope linked to the M-FISH CytoVision software (version 7.4, Leica Biosystems). Karyotypes were described according to the International System for Human Cytogenomic Nomenclature (ISCN) 2020 (16).

High resolution-comparative genomic hybridization (HR-CGH) analysis of MelT79 tumor cells was performed as previously described (17).

DNA was extracted from MelT79 cells using the NZY Tissue gDNA Isolation Kit (NZYTech MB13502) according to the manufacturer’s instructions. DNA was also isolated from a 2015 formalin-fixed paraffin-embedded (FFPE) metastasis ( Figure 1 ), using the Maxwell^® RSC FFPE DNA kit (Promega) in the Maxwell^® RSC platform (Promega) according to the manufacturer’s instructions. After quantification using the Qubit^® 4 fluorimeter (Life Technologies), 50 ng of DNA were used to prepare the next-generation sequencing (NGS) library on the Magnis NGS Prep System (Agilent Technologies), following the manufacturer’s protocol. A custom multigene panel using oligonucleotide probes (SureSelect XT HS2 DNA Target Enrichment, Agilent Technologies), routinely used in our laboratory, was employed to target cancer-related genes, including those frequently altered in melanoma, such as BRAF, NRAS, CDKN2A, KIT, NF1, PTEN, and TERT (complete list of genes on request). NGS was performed on a MiSeq platform (Illumina) with 75bp paired-end reads with average target coverage of 100x for the FFPE sample and 560x for the MelT79 cell line. Alignment and annotation were performed on the SeqOne platform using the SomaVar v2.4 workset (SeqOne Genomics) to the human reference genome GRCh37. The criteria for selecting potentially relevant gene variants included: base coverage > 20x, variant allele frequency (VAF) > 5% and clinical evidence of actionable, prognostic and diagnostic variants, provided by the Cancer Knowledge Base (CKD) database.

BRAF mutational status was validated by Sanger Sequencing, as previously described (18), using forward primer 5’ AAA CTC TTC ATA ATG CTT GCT CTG 3’ and reverse primer 5’ GGC CAA AAA TTT AAT CAG TGG A 3’.

Total RNA was extracted and purified from all cell lines using NZY Total RNA Isolation kit (NZYTech MB13402), according to the manufacturer’s instructions, and quantified using Nanodrop 2000 (Thermo Fisher Scientific). Then, gene expression was assessed by reverse transcription quantitative PCR (RT-qPCR), as previously described (19). Primers used are presented in Supplementary Table 1 , and TBP and HPRT1 were selected as housekeeping genes.

Cells were seeded in duplicate in 24-well tissue culture plates containing glass coverslips pre-coated with 0.2% gelatin (Sigma-Aldrich G-1890), at a density of 1 × 10⁵ cells/well, and cultured for 48 h. At the endpoint, cells were washed three times with phosphate buffered saline 1x (PBS) and fixed with 2% paraformaldehyde for 20 mins at room temperature (RT). Permeabilization was performed with 0.1% Triton X-100 for 5 mins at RT, followed by three additional PBS 1x washes. Cells were then incubated for 10 mins with 50 mM NH₄Cl at 4°C to quench autofluorescence. Blocking and further permeabilization were performed for 40 mins at RT using PermBlock solution (PBS 0.5% bovine serum albumin [BSA] 0.1% saponin). Cells were incubated with primary antibodies rabbit anti-MITF (1:200, overnight, at 4°C; Assay Biotech B0512) and mouse anti-TRP1 (1:200, 2h, at RT; Abcam ab3312), followed by three 5 mins washes to remove unbound antibodies. Secondary antibody incubation was performed using Alexa Fluor™ 488 goat anti-rabbit (1:1000, 2h, at RT; Invitrogen A11008) and Alexa Fluor™ 568 goat anti-mouse (1:500, 2h, at RT; Invitrogen A11004), followed by three additional 5 mins washes. Cells were then rinsed once with PBS 1x for 5 mins before coverslips were mounted onto glass slides with Vectashield^® Antifade Mounting Medium with DAPI (Vector Laboratories H-1200-10). All antibody dilutions and washes were performed using PermBlock solution. Images were acquired in a Zeiss Imager Z1 microscope linked to the CytoVision software (version 7.1, Leica Biosystems), and corrected total cell fluorescence (CTCF) quantification was performed using ImageJ software (version 1.53e). Intra-cell line heterogeneity was assessed by calculating, for each imaged field, the ratio of the CTCF range (maximum minus minimum value within that image) divided by the mean CTCF. This calculation was only applied to cell lines with detectable protein expression.

Cells were seeded in duplicate in 6-well tissue culture plates at a density of 2 × 10⁵ cells/well. At each timepoint (24, 48 h), cells were detached as previously described in section 2.2. to create a cell suspension, which was mixed with Trypan blue solution 0.4% (Canvax Biotech CC007) in a 1:1 ratio. Viable cells were counted with a Neubauer improved cell counting chamber under the microscope. The cell doubling time was calculated using the following formula:

Cell proliferation was evaluated using a colorimetric BrdU ELISA kit (Roche 11647229001), following the manufacturer’s instructions. Briefly, cells were seeded in triplicate in 96-well tissue culture plates at a density of 1 × 10⁴ cells/well, and cultured for 24 h. Two hours before the endpoint, cells were labeled with BrdU. At the endpoint, cells were fixed and incubated with a peroxidase-conjugated anti-BrdU antibody for 90 mins. After washing, cells were incubated with substrate solution for 10 mins, and the reaction was subsequently stopped with 1M H₂SO₄. Absorbance was measured at 450 nm using the iMark™ Microplate Absorbance Reader (Bio-Rad). For each each cell line, background absorbance values (from wells incubated with anti-BrdU only) were subtracted from the corresponding test wells.

Cells were seeded in triplicates in 24-well tissue culture plates at a density of 5 × 10⁴ cells/well and allowed to adhere for 24 h. Then, cells were exposed to vehicle (dimethyl sulfoxide, DMSO, 1%), or different concentrations (0.1, 1, 10, 50, 100 nM, 1 μM) of trametinib (Selleckchem GSK1120212) or dabrafenib (Selleckchem GSK2118436) for 72 h. At the endpoint, Cell Counting Kit-8 reagent (Dojindo CK04-20) was added to each well in a 1:10 dilution, followed by a 30-minute incubation period at 37°C with 5% CO₂, in the dark. Absorbance values were measured at 450 nm in a 96-well plate, using the iMark™ Microplate Absorbance Reader (BioRad).

Assays were performed with at least three biological replicates. All statistical analyses and graphical representations were performed using GraphPad Prism version 8.4.3.

Z-score values were calculated for each cell line regarding the expression of each gene, considering the mean gene expression value of all cell lines and the corresponding standard deviation (SD).

Data for relative gene expression, CTCF fold change, CTCF range/mean, cell doubling time, and cell proliferation are presented as mean with SD. Differences between MelT79 and the other cell lines were determined by one-way analysis of variance (ANOVA), followed by Dunnett’s multiple comparisons test to compare individual means.

IC₅₀ values were calculated using a non-linear regression model, based on a dose-response slope, and are presented as best-fit IC₅₀ values with lower and upper 95% confidence limits. Viability percentages are presented as mean with SD. Differences in IC₅₀ values between MelT79 and commercially available cell lines were determined by extra sum-of-squares F test, with multiple comparison correction.

Our group has made several attempts to establish primary cell cultures from CM surgical fragments, which have been faced with 3 main challenges, often leading to cell culture loss: (i) low melanoma cell viability, with no cell adhesion; (ii) cell senescence; (iii) fibroblast overgrowth. Out of 34 attempts and optimizations, this was the first successful primary cell line established from a metastatic CM surgical fragment in our group ( Figure 2A ).

Regarding the first challenge, studies have shown that enzymatic digestion processes can significantly impact cell viability (20, 21). Therefore, similarly to Ścieżyńska et al. (22), we processed tissue fragments through mechanical digestion only, to ensure maximum CM cell viability.

Najem et al. (23) have described that primary cell cultures growing in media with low tyrosine levels could be propagated through multiple passages, while maintaining a clinically relevant and highly differentiated melanoma state. The authors noted that some primary cultures grown in high-tyrosine media became rapidly senescent, leading to culture loss. Therefore, to prevent cell culture loss, we used McCoy’s 5A medium, which has a lower tyrosine concentration (26.1 mg/L) compared to DMEM (104 mg/L), the media we grow most CM cell lines in. We also hypothesized that this approach would help the culture retain a phenotype closer to the original tumor.

With this specific tumor fragment, we encountered no issue with fibroblast growth. Raaijmakers et al. (14) have already demonstrated that with tissues with a low fibroblast content, a simple protocol can be used to obtain primary cultures that adequately mirror the original tumor. Guo & Jahoda (24) demonstrated that fibroblasts begin migrating from tissue fragments 7 days after culture initiation. We removed most tissue fragments from our culture 2 days after seeding, during media change, retaining only the melanoma cells that had already adhered to the culture vessel surface.

Overall, there is a lack of studies that empirically demonstrate the optimal method to establish primary cell cultures from CM patients, with methodologies used varying widely (14, 22, 23, 25–28). Most studies involving primary CM cell cultures do not discuss the methodology in detail, nor do they compare failed attempts with the successful ones to clearly define key steps in the protocol. However, such studies are extremely challenging due to the high heterogeneity in tissue type, fragment size, tumor purity, and patient clinicopathological characteristics. As such, a trial-and-error approach is often employed for each case.

M-FISH analysis revealed that the MelT79 cell line presents a complex karyotype with metaphase chromosomal number varying between near-diploid (n = 8) ( Figure 2B , Supplementary Table 2 ) to near-pentaploid (n = 6) ( Figure 2C , Supplementary Table 2 ). MelT79 cells are characterized by a high number of non-clustered structural chromosomal abnormalities such as deletions, inversions, and inter-chromosomal translocations. We also observed a high number of inter-chromatid chromosomal aberrations, such as the one depicted in Figure 2C . We have described this aberration affecting chromosome 4 as a der (4) t(X; 4;10;14) (?;?;?;?). These findings are in accordance with the observations made by Liu et al. (29) in therapy-resistant melanomas, who evidenced that these non-clustered structural aberrations were due to genetic defects in homologous recombination DNA repair and non-homologous end-joining (NHEJ) or alternative NHEJ recombination mechanisms.

Additionally, evaluation of MelT79 cells by HR-CGH revealed a highly unbalanced genome with multiple gains and losses ( Figure 2D , Supplementary Table 3 ). Interestingly, and as reported by Liu et al. (29), we found that the three most significantly deleted chromosomal regions in their tumors – 9p21 (51%), 15q14 (45%) and 15q21.1 (45%) – are also deleted in MelT79 cells.

To further characterize the alterations present in MelT79 cells, we have screened for gene mutations by NGS, using a multigene panel. Two gene mutations – BRAF V600E and RET S649L ( Table 1 ) – have been detected, each with a VAF of 100%, confirming the absence of fibroblast contamination in the culture. BRAF V600E had already been previously identified in a 2015 cutaneous metastasis, which led to the implementation of targeted therapy with Vemurafenib and Cobimetinib. This mutation is the most common and clinically significant targetable mutation in CM (1).

To determine whether RET S649L emerged as a mechanism of targeted or immunotherapy resistance, or as an early event in disease progression, we performed NGS on an FFPE sample of the 2015 treatment-naïve metastasis, as no primary tumor sample was available. Given the degraded DNA quality in older FFPE samples, initial analysis identified an extensive list of 279 variants, most of which were likely false positives caused by DNA fragmentation. Therefore, we focused our analysis on the variants with VAF > 10% ( Table 1 ), as previously discussed for low-quality FFPE samples (30). We confirmed the presence of both BRAF V600E and RET S649L in this earlier metastasis, indicating that RET S649L was an early event in tumor progression. Its persistence across different therapeutic approaches suggests that it may confer a selective advantage, making it a potential therapeutic target in this case, with a high likelihood of this mutation being present in most metastatic lesions of this patient.

Additional variants were identified in the 2015 FFPE metastatic sample, in genes SDHA, SDHC, PIK3CA, PTEN, ERBB2, POLE, and RET, but were absent in MelT79, which may reflect clonal evolution, highlighting CM’s dynamic genetic landscape. These genes are related with the PI3K/Akt and MAPK pathways (PIK3CA, PTEN, ERBB2) (1, 31), succinase dehydrogenase subunits (SDHA, SDHC) (32), and the DNA polymerase (POLE) (33). However, these variants are present at significantly lower VAF compared to RET S649L and BRAF V600E, which persisted throughout disease progression. This suggests that these variants are likely passenger mutations, artifacts, or subclonal events that were ultimately lost during progression or culture establishment. In contrast, the persistence of RET S649L and BRAF V600E underscores their potential roles as key drivers in this patient’s tumor evolution.

CM’s heterogeneity and plasticity enable tumor cells to adapt to microenvironmental changes and trigger metastasis and therapy resistance (34). Unlike other cancers, CM cells do not undergo epithelial-mesenchymal transition, since melanocytes are not epithelial cells, but instead use phenotype switching to transition between different states – from hyper-differentiated and slow-proliferating, to rapidly proliferating, or even slow-proliferating but highly invasive states (34–37).

Given the extensive and complex clinical history of the patient from whom MelT79 cells were derived, our goal was to better understand the phenotypic characteristics that might contribute to the persistence of these cells despite different therapeutic approaches. Thus, we analyzed the expression of several genes involved in CM differentiation and invasion processes and compared it to commercially available CM cell lines with different pigmentation levels and growth patterns ( Figure 3A ).

MelT79 cells share similarities with G-361 and MNT-1 cell lines, upregulating melanin production genes linked to a differentiated phenotype (MLANA, SOX10, PAX3, TYR) (38, 39), while downregulating genes associated with invasion (AXL, EGFR, CDH2, FGFR1, MMP2, WNT5A, ZEB1) (23, 36). In contrast, A375 and WM115 cells exhibit a more invasive gene expression profile. However, despite the similar profiles, MelT79 differs from G-361 and MNT-1 in some respects: it expresses intermediate MITF levels, significantly lower than MNT-1 but higher than A375 and WM115 ( Figure 3B ); it displays elevated expression of the invasion-associated gene MMP3; and reduced expression of differentiation markers RAB27A and ZEB2 (40–42). Despite these differences, TYRP1, a direct transcriptional target of MITF, closely linked to pigmentation (39), remains strongly expressed in MelT79 cells, consistent with the more differentiated profile shared with MNT-1 and G-361, and contrasting with A375 and WM115, which do not express it ( Figure 3C ).

To further characterize the differentiation status of MelT79 cells, we assessed MITF and TYRP1 protein levels by immunofluorescence. MITF protein expression in MelT79 is comparable to that in MNT-1 and G-361 cells, and markedly higher than in A375 and WM115, despite intermediate mRNA expression levels ( Figures 3D, E ). Notably, MITF expression varied widely from cell to cell within the same captured field, with some cells exhibiting high MITF levels, while others barely expressed it ( Figure 3D ). This heterogeneity was significantly greater in MelT79 when compared to the other cell lines ( Figure 3G ). TYRP1 protein levels were also elevated in MelT79, similar to MNT-1 and higher than G-361, consistent with mRNA trends ( Figure 3F ). However, TYRP1 expression showed no significant heterogeneity among MelT79 cells, indicating a more uniform protein expression ( Figure 3H ).

These expression patterns are closely tied to pigmentation status (39, 43) – pigmented cell lines G-361 and MNT-1 upregulate differentiation genes, while non-pigmented A375 and WM115 cell lines downregulate them. Despite the upregulation of most differentiation genes analyzed, with confirmed higher MITF and TYRP1 protein expression, MelT79 cells appear surprisingly non-pigmented ( Supplementary Figure 1 ).

Regarding doubling time, MelT79 exhibits significantly slower growth (49 h) compared to A375, WM115, and G-361. While their doubling time is closer to that of MNT-1 cells, there is still a trend indicating that MelT79 proliferates at a slower pace than MNT-1 cells ( Figure 4A ). Doubling times from patient-derived CM cell lines can vary widely (reports range from 25 to 104 h), depending on tissue origin, molecular alterations, and cell culture conditions (22, 28). These findings are supported by BrdU incorporation assay, which confirms that MelT79 cells proliferate significantly more slowly than G-361, WM115, and A375, and at a rate comparable to MNT-1 ( Figure 4B ).

Overall, the fact that MelT79 exhibits a differentiated gene expression profile, variable MITF expression when compared to other cell lines, lack of pigmentation, and reduced proliferation, further highlights this cell line’s heterogeneity.

In advanced CM, targeted therapy is a key therapeutic approach for patients with tumors harboring BRAF mutations. Currently, this involves combination of BRAFi/MEKi, namely dabrafenib/trametinib (7). To assess MelT79 sensitivity to these inhibitors, we calculated their IC₅₀ values and compared them to various commercially available CM cell lines ( Figure 5 , Supplementary Figure 2 ).

Our results show that IC₅₀ values for both dabrafenib and trametinib are significantly higher in MelT79 cells compared to MNT-1 cells, indicating greater resistance. However, we observed a trend for both dabrafenib and trametinib’s IC₅₀ to be lower in MelT79 cells compared to WM115. Since the transition from a differentiated to an invasive state is associated with increased resistance to targeted therapies (44, 45), it is expected that MNT-1 cells are the most sensitive, while WM115 cells seem to be the most resistant.

Interestingly, although MelT79 cells primarily exhibit a differentiated gene expression profile, our previous results show they also possess some dedifferentiated traits. This aligns with their intermediate sensitivity to BRAFi and MEKi when compared to the other cell lines – more resistant than highly differentiated MNT-1, but less resistant than the highly invasive WM115.