Background
Mucosal melanoma (MM) is a highly aggressive and rare form of melanoma arising from melanocytes at any mucosal surface [
1]. Due to their hidden sites of origin and the lack of early and specific symptoms, these tumors are often late diagnosed [
2,
3]. The majority of the patients undergo surgical resection as a first-line treatment, followed by adjuvant radiotherapy to reduce loco-regional recurrence; however, this approach results in a negligible clinical benefit [
3‐
5]. Local recurrence and distant metastasis are common and associated with a very poor outcome in terms of overall survival rate [
2,
6], and the mechanisms leading to this aggressive clinical behavior are only partially understood.
Compared to cutaneous melanoma (CM), MM represent a different entity also in terms of biological features. A diverse cell of origin (mucosal vs cutaneous melanocytes) and a lack of UV-exposure are associated with a distinct molecular profile. Moreover, limited data are available on the existence of MM precursor lesions as well as on their molecular hubs sustaining the progression trajectories [
7]. MM rarely show
BRAF V600 mutations [
8,
9], thus ruling out treatment with BRAF inhibitors (BRAFi) for the majority of MM patients. Among MM drivers, somatic mutations of
NRAS and the
TERT promoter are reported, as well as mutually exclusive loss of function mutation of the tumor suppressors
NF1 and
SPRED1, the latter co-occurring with activating
KIT mutations [
10‐
12]. Recent studies indicate that MM have significant genomic instability and display structural variants including amplifications (e.g.,
CDK4,
MDM2,
TERT, and
cyclin D1), deletions (e.g.,
CDKN2A/B,
PTEN and
TP53) and fusion genes [
10,
13,
14] displaying a site-specific variation [
13,
15]. As we reported on a retrospective cohort, many MM cases show enhanced expression of pAkt and pErk, suggesting the activation of both PI3K/Akt/mTOR and RAS/MAPK pathways [
16]; but the molecular basis of this activation remains unknown.
A fraction of MM cases might obtain clinical benefits from immunotherapy [
17]. In recent years, the role of immune checkpoint inhibitors (ICI) has been evaluated in several monocentric and multicentric retrospective studies [
18‐
20] and clinical trials [
21,
22]. To date, ICIs represent the front-line therapy for patients with unresectable advanced or metastatic MM; however, outcomes to this approach remain poor compared to CM [
23‐
30]. The tumor mutational burden (TMB) is substantially low in MM [
14,
31] and is more comparable to poorly immunogenic cancers not associated with exposure to mutagens [
31]. These genomic and immunogenic differences might explain the lower response rate of MM to ICI blockade [
32].
This set of data, although of potential relevance, offers a still limited support to decision making for the clinical management of these patients. Limited understanding of this rare form of cancer is due to the lack of pre-clinical models for genetic and functional studies. Dogs are frequently affected by aggressive spontaneous mucosal melanomas of the oral cavity, providing a unique but still invalidated working model for their human counterparts [
33‐
35].
A better understanding of the immunobiology of MM may benefit from novel cellular models of human origin. Starting from patient-derived fresh biopsies of MM arising in the sinonasal tract, we could generate and successfully propagate five novel cell lines (SN-MM1 to 5). Sinonasal tract MM are among the most frequent MM and on endoscopy they present as expansile or polypoid lesions with different degrees of pigmentation. Patients suffer from nasal obstruction and discharge, epistaxis, facial pain, olfaction disturbance [
1]. Additional signs and symptoms (exophthalmos, visual disturbance, headache), suggesting intraorbital and intracranial extension, are in keeping with the heterogeneity of growth patterns (i.e., superficial spreading over the mucosal and submucosal lining of the sinonasal tract
versus deep infiltration) and multifocality [
1,
36]. SN-MM1 to 5 display a melanocytic identity and tumorigenic potential. Moreover, their proteomic analysis revealed activation of cancer-associated pathways involved in cell transformation and cancer progression, including PI3K/Akt/mTOR pathway; as a proof of concept, we could also demonstrate a role for this pathway in MM cell fitness. Finally, microscopic analysis of derived mouse xenograft identified a subpopulation of MM cells coherent with melanoma-initiating cells.
Methods
Human tumor biopsies
Human tissues were obtained from surgical specimens from patients undergoing surgery with radical or palliative intent. All samples included in this study were obtained under the approval of the institutional Ethics Board of ASST Spedali Civili of Brescia (IRB code: NP 2066/2015) and ASST Sette Laghi of Varese (IRB code: NP 33025/2015), Italy. Informed consent was obtained from all patients for the manipulation of human tissue and cell culture. Diagnosis of sinonasal mucosal melanoma was performed according to the American Joint Committee on Cancer/Union for International Cancer control staging system for mucosal melanoma of the upper aerodigestive tract [
37]. Demographics (gender, age) and selected oncologic informations (site of origin and recurrence, TNM classification, treatment strategy and outcome) [
2,
36,
38] are summarized in Additional file
6: Table S1. Briefly, 80% of patients were female and older than 60 years. All patients were in advanced stage disease (T3-T4) and underwent endoscopic endonasal resection followed by radiation therapy (4/5). All patients were died of disease.
Generation of primary SN-MM cell lines
SN-MM cell lines were generated from fresh surgically collected tumor samples. Surgical specimens were placed in RPMI 1640 medium (cat.n. F1215, Biochrom, Berlin, Germany) supplemented with 20% Fetal Bovine Serum (FBS) (cat.n. S0115, Biochrom), 1% Penicillin/Streptomycin (cat.n. 15070-063, Thermo Fisher Scientific, Waltham, MA, USA) on ice. Samples were finely minced with a scalpel in a 100 mm cell culture dish in sterile conditions. After mechanical dissociation, samples were transferred in 15 mL sterile tube with RPMI 1640 medium supplemented with 10% FBS, 2 mM Glutammine (cat.n. 25030149, Thermo Fisher Scientific), 0.5% Penicillin/Streptomycin, and 200 U/mL Collagenase Type II (cat.n. LS004174, Worthington, Biochemical Corporation, USA). Samples were incubated at 37 ℃ in a swinging water bath for 3 h and vortexed every 30 min. The enzymatic digestion was stopped by adding DPBS (cat.n. 14190144, Thermo Fisher Scientific) and cell suspension was filtered through 40 µm cell strainer and centrifuged at 10 min at 300 g. Red blood cells lysis was achieved with 1X RBC Lysis Buffer (cat.n. 420301, Biolegend, San Diego, CA, USA) following manufacturer’s instructions. Resulting cells were resuspended in RPMI 1640 medium supplemented with 10% FBS, 2 mM Glutammine, 0.5% Penicillin/Streptomycin and 100 µg/mL Primocin (InvivoGen, San Diego, CA, USA), seeded in T25 culture flask at density of 5 × 105 cells/mL and maintained at 37 ℃ in a 5% CO2 humidified incubator. The following day, cell culture medium was replaced with fresh medium. When adherent cells reached 80% of confluence, cells were detached using the Tryple Express Enzyme (cat.n. 12604013, Thermo Fisher Scientific) and sub-cultured at 1:2 or 1:3 ratio. Cell lines were named SN-MM1-5 and sub-cultured for at least 5–10 times before further analysis.
SN-MM F1 cell lines were generated from fresh cell-derived xenografts (CDX) samples harvested from mice, as described above.
Cell culture
The SN-MM cell lines were culture with complete RPMI medium consisting of RPMI 1640 supplemented with 10% FBS, 2 mM Glutammine, 0.5% Penicillin/Streptomycin. Normal human epidermal melanocytes (NHEM) M2 (cat.n. C-12400, PromoCell, Heidelberg, Germany) and M3 (cat.n. C-12413, PromoCell) were cultured in Melanocytes Growth Medium M2 (cat.n. C-24300, PromoCell) and M3 (cat.n. C-24310B, PromoCell), respectively, following manufacturer’s instructions. Mycoplasma contamination were excluded by routinely testing with Universal Mycoplasma detection kit (cat. N. 30-1012K ATCC, Manassas, VA) according to manufacturer’s suggestions.
Cell block preparation
For cell-block preparation, at least 1 × 106 cells were harvested and processed as follows. Cell suspensions of SN-MM cell lines were centrifuged for 10 min at 3,000 rpm. A solution of plasma (100 mL, kindly provided by Centro Trasfusionale, ASST Spedali Civili) and HemosIL RecombiPlasTin 2G (200 mL, Instrumentation Laboratory; cat. no. 0020003050; 1:2) were added to cell pellets, mixed until the formation of a clot, then placed into a labeled cassette (Bio-Optica; cat. no. 07-7350). The samples were fixed in 10% formalin (Bio-Optica; cat. no. 05-K01004) for 1 h followed by paraffin inclusion.
Histology and Immunohistochemistry
Formalin fixed and paraffin embedded (FFPE) tumor biopsies, xenografted tumors and cell-block sections were used for histology and immunohistochemistry (IHC).
Haematoxylin–eosin (H&E) staining was performed using standard protocols on 4 µm sections. Histologic evaluation of H&E-stained sections of parental tumor samples, cell-blocks and xenografts was performed by two expert pathologists. Four-micron thick tissue sections were used for immunohistochemical staining, heat mediated antigen retrieval was performed in microwave oven and endogenous peroxidase activity was quenched using 0.3% hydrogen peroxide diluted with methanol during rehydration. After washing with Tris-Buffered Saline (TBS) solution, slides were incubated with primary antibody for 1 h at room temperature and revealed by incubation with horseradish-peroxidase polymer Novolink Polymer Detection System (cat.n. RE7159, Leica Biosystems, Wetzlar, Germany), Envision + System-HRP Labelled Polymer Anti-mouse (cat.n. K4001, Dako) or Anti-Rabbit (cat.n. K4003, Dako), followed by 3,3ʹ-diaminobenzidine as chromogen. Sections were counterstained with Mayer’s haematoxylin (Bioptica, Milano, Italy). Primary antibody used are listed in Additional file
6: Table S2.
IHC evaluation of melanocytic markers was performed using a four-tiered scoring system (score 0 = 0% of positive cells; score 1 = 1–10% of positive cells; score 2 = 10–50% of positive cells; score 3 > 50% positive cells).
For double sequential immunostains, the first reaction is deleted after first chromogen de-stain and stripping. Anti-CDH1 was used for the first immune reaction, revealed using Novolink Polymer (Leica) and developed in 3-amino-9-ethylcarbazole chromogen (AEC), counterstained with hematoxylin and cover-slipped using gelatin. Subsequently, the slides were digitally scanned, using Aperio Scanscope CS (Leica Microsystems). After cover slip removal, AEC was washed out and the slides were eluted using a 2-Mercaptoethanol/SDS solution (20 mL 10% w/v SDS with 12.5 mL 0.5 M Tris–HCl, pH6.8, 67.5 mL distilled water and 0.8 mL 2-ME). Slides were subsequently incubated in this solution in a water-bath pre-heated at 56 ℃ for 30 min. Sections were washed for 1 h in distilled water. After unmasking in microwave, anti-ZEB1, was revealed using Novolink Polymer (Leica) and AEC CD271 was revealed using Mach 4 MR-AP and Ferangi blue as chromogen and slides were counterstained with hematoxylin, cover-slipped and digitally scanned. The subsequent immunostains for anti-MITF1 was developed analogously. The digital slides were processed using ImageScope. Slides were synchronized and corresponding tissue regions were analysed.
Immunofluorescence
SN-MM cells (3 × 10
5 cells) were seeded in Nunc
™ Lab-Tek
™ II Chamber Slide
™ System (cat.n. 154453 Thermo Fisher Scientific). Cells were fixed in ethanol 95% and were used for immunofluorescence (IF) staining. Heat mediated antigen retrieval was performed in microwave oven. After washing with TBS solution, slides were blocked with TRIS plus 5% BSA solution for 30 min. Then, slides were incubated with primary antibody for 1 h at room temperature and revealed by incubation with horseradish-peroxidase polymer Novolink Polymer Detection System (cat.n. RE7159, Leica Biosystems, Wetzlar, Germany), followed by Alexa Fluor TM 488 tyramide reagent (cat. n. B40953, Invitrogen, Waltham, MA, USA) and cyanine TM 555 tyramide reagent (cat. n. 96020, Biotium, Fremont, CA, USA). Fluorescence Mounting Medium (cat. n. S3023, Dako) was used. Primary antibody used are listed in Additional file
6: Table S2.
The slides were digitalized using Axioscan7 (Zeiss) and processed using Zen Blue software (Zeiss).
Transmission electron microscopy
2 × 106 cells were collected and washed once with DPBS. Cell pellets were fixed with 2.5% glutaraldehyde (Sigma-Aldrich, Cat. N. G5882) in Sorensen’s Phosphate Buffer pH 7.2 and kept at 4 ℃ for at least 2 h. After fixation, pellets were washed three times with Sorensen’s phosphate buffer for and post-fixed for at least 2 h with 1% Osmium Tetroxide (Electron Microscopy Sciences, Cat.#19134). Samples were washed in distilled water and dehydrated in progressive ethanol concentration (50%, 70%, 90% and 100%) for 15 min at each step. Cell pellets were soaked in acetone for 15 min and then transferred in a 1:1 mixture of acetone and Epoxidic resin for at least 3 h. Embedding was accomplished in Epoxidic resin. Semithin (1 µm-thick) sections cut with glass sharp blade using an ultramicrotome (Ultracut E- Leica, Microsystems S.r.l., Milan, Italy) were stained with toluidine blue and examined by light microscopy. For ultrastructural analysis, 80 nm-thick sections, cut with diamond glass were obtained from selected areas and were collected on 200 meshes formvar coated copper grids, double stained with uranyl acetate (Electron Microscopy Sciences Cat. N. 22409) and lead citrate (Electron Microscopy Sciences Cat. N. 22410) and examined using FEI Tecnai G2 Spirit or CM12 TEM transmission electron microscopes (FEI Instrumentation Company, Hillsboro, Oregon, USA) operating at 85kV. Photographs were taken using a Veleta integrated Digital Camera (Olympus Soft Imaging Solutions, Munster, Germany).
DNA isolation
Genomic DNA was extracted from SN-MM cell lines using QIAamp DNA Blood Mini Kit (cat. n. 51104, Qiagen, Redwood City, CA, USA) according to manufacturer's protocol. Automated DNA Extraction was performed on the FFPE tissues using Maxwell® RSC DNA FFPE Kit (cat. n. AS1450, Promega, Madison, WI, USA) according to manufacturer’s protocol.
Array comparative genomic hybridisation (aCGH)
aCGH experiments were performed on gDNA from SN-MM cell lines. To avoid false negative results due to the sensitivity of this assay, only samples displaying a tumor cells content greater than 50% were analyzed. The genomic DNA was prepared and purified to a standard where the A260/A280 ratio exceeded 1.8 and A260/A230 ratio exceeded 1.5. The amount of DNA requested to perform the test was 1.0 µg. Samples and references were labelled using the CytoSure Genomic DNA Labelling Kit (Oxford Gene Technology). The CytoSure Array 8 × 60 k CGH were scanned by Agilent Surescan C scanner with 2 μm resolution; features were extracted with Feature Extraction software and log2 ratio data were imported and analyzed by Cytosure Interpret Software 4.11.36 (Oxford Gene Technology) for the identification of copy number variation (CNV). Copy number changes below 5Mb was not reported unless affecting a gene/region relevant to cancer disease.
Fluorescence in situ hybridization (FISH)
Interphasic FISH Test IGH-MYC-CEP8 and FISH Test CCND1/CEP11 were performed on 4 μm FFPE (Formalin Fixed Paraffin Embedded), probes IGH-MYC-CEP8 were provided by Abbott Molecular (Des Plaines, USA), while CCND1/CEP11 by Zytovision GmbH (Bremerhaven, Germany) and were used following the manufacturer’s suggested protocol. For each case, a minimum of 50 nuclei were observed using the Leica DM6000B System (Leica Microsystems, Buccinasco, MI, Italy). A total of 100 cells were evaluated for signal pattern for all probes. Nuclei with at least 15% with rearranged patterns were considered positive.
Sanger Sequencing
PCR was performed to amplify NRAS exon 2 using the specific primers (NRAS_Forward: 5ʹCAACAGGTTCTTGCTGGTGT3ʹ; NRAS_Reverse: 5ʹCCTCACCTCTATGGTGGGAT3ʹ). The PCR products were purified using the Amicon Ultra 0.5 mL Centrifugal Filters 30 K (cat. n. UFC5030BK, Millipore, Burlington, MA, USA), according to the manufacturer’s instructions, and sequenced in both directions using the BigDye™ Terminator v3.1 Cycle Sequencing Kit (cat. n. 4337455; Applied Biosystems). The excess of fluorescent terminator dye was removed using Performa® DTR Gel Filtration Cartridges (cat. n. 4408228, Edge Bio Systems, San Jose, CA, USA). The samples were resuspended in formamide, denatured at 95 ℃ for 3 min and processed through SeqStudio™ Genetic Analyzer (ThermoFisher Scientific).
Flow cytometry
SN-MM cells were detached using the Tryple Express Enzyme (cat.n. 12604013, Thermo Fisher Scientific), washed in 1 mL cold DPBS w/o proteins and stained using Live/Dead Fixable Red Dead Cell Stain Kit (cat. n. L23102, Life Technologies) and anti-CD271 (clone REA844) PE-conjugated antibody (cat.n. 130-112-601, Miltenyi Biotec, Bergisch Gladbach, Germany) following manufacturer’s instructions. Unstained cells were used as negative controls. Results were reported as the percentage of positive cells after gating on single live cells. Samples were acquired with MACS Quant® Analyzer 16 (Miltenyi Biotec) and results were analyzed by FlowJo X software v10.8 (Tree Star Inc, Wilmington, NC, USA).
MTS assay
15 × 103 cells were plated in flat bottom 96-well plates in 125 µL complete RPMI medium. The LY294002 (1–100 µM; cat.n. s1105, Selleckchem, Houston, TX, USA) was used and cells were cultured for 24 h, 48 h and 72 h. The DMSO was used as vehicle control at the final concentration of 0.2%. The cisplatin (0.625–20 µM; cat.n. s1166, Selleckchem) and temozolomide (3.125–200 µM; cat.n. s1237, Selleckchem) were used and cells were cultured for 72 h or 120 h. The 0.1% DMF and 0.4% DMSO were used as vehicle controls, respectively. Cell proliferation was evaluated using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (cat.n. G3580, Promega), according to the manufacturer's instructions. Absorbance was determined at 490 nm with EnSight™ multimode plate reader (PerkinElmer, Waltham, MA, USA). The sigmoidal concentration–response function was applied to calculate the IC50 values.
Western blotting
1–1.5 × 10
5 cells were seeded in 12-well plates in 1 mL complete RPMI medium. After eight hours from seeding, cells were starved in RPMI medium with 1% FBS overnight (o/n), then treated with 50 µM LY294002 for 4 h and 24 h. Cells were lysed in RIPA lysis buffer (cat.n. 89900, Thermo Fisher Scientific) supplemented with Protease Inhibitor Cocktail (cat.n. 78440, Sigma-Aldrich) and incubated on ice for 20 min. Protein concentration was determined by Bradford assay and 20 µg of total proteins were loaded on 4–12% NuPAGE
® Bis–Tris Mini Gels (cat.n. NP0335, Invitrogen) under reducing condition and transferred onto a PVDF membrane (cat.n. LC2007, Invitrogen). Membranes were blocked with 5% milk (cat.n. 22012, Biotium, Fremont, CA) in TBS-T (TBS with 0.05% Tween 20; cat.n. 28360, Invitrogen) for 1 h at room temperature. Primary antibodies were incubated o/n at 4 ℃ in TBS-T with 5% BSA (cat.n. A3059, Sigma-Aldrich). Primary antibodies are listed in Additional file
6: Table S2. The anti-Rabbit (cat.n. 31460, Thermo Fisher Scientific) or the anti-mouse (cat.n. 7076, CST) secondary antibodies conjugated with horseradish peroxidase were incubated for 1 h at room temperature. Detection was performed using the SuperSignal
™ West Pico Chemiluminescent Substrate (cat.n. 34577, Thermo Fisher Scientific) and visualized by autoradiography.
Tumorigenic assay
6 weeks old NOD/SCID mice (Envigo, Udine, Italy), were subcutaneously injected in the dorsolateral right flank with 8 × 106 cells in 1:1 200 µL mixture of PBS and Matrigel (Cultrex BME, R&D Systems). Mice were euthanized when tumors 300 mm3 volume and tumors were harvested and paraffin included for histological examination. Animal experiments were approved by the local animal ethics committee and were performed in accordance with national guidelines and regulations. Procedures involving animals and their care are conformed with institutional guidelines that comply with national and international laws and policies (EEC Council Directive 86/609, OJ L 358, 12 December 1987) and with “ARRIVE” guidelines (Animals in Research Reporting In Vivo Experiments).
Proteomic analysis
4 × 10
4 M3-NHEM and 1–1.5 × 10
5 SN-MM cells were seeded in 12-well plates in 1 mL complete medium. Cells were collected, washed, and lysed, as described in the
Western blotting Materials and Methods Section. After cell lysis, the proteins were digested with trypsin. 40 µg of protein were reduced in 25 µL of 100 mM NH4HCO3 with 2.5 μL of 200 mM DTT (Sigma) at 60 ℃ for 45 min and next alkylated with 10 μL 200 mM iodoacetamide (Sigma) for 1 h at RT in dark conditions. Iodoacetamide excess was removed by the addition of 200 mM DTT. The digests were dried by Speed Vacuum and then desalted [
40,
41].
Trypsin-digested sample proteins were analyzed with a micro-LC (Eksigent Technologies, Dublin, CA, USA) system coupled with a 5600 + TripleTOF system (Sciex, Concord, ON, Canada) equipped with DuoSpray Ion Source. Stationary phase was a Halo C18 column (0.5 × 100 mm, 2.7 µm; Eksigent Technologies, Dublin, CA, USA). Mobile phase was a mixture of 0.1% (v/v) formic acid in water (A) and 0.1% (v/v) formic acid in acetonitrile (B), eluting at a flowrate of 15.0 µL min − 1 at an increasing concentration of solvent B from 2 to 40% in 30 min. For identification purposes the samples were subjected to a data dependent acquisition (DDA) while the quantification was performed through a data independent analysis (DIA) approach. The DDA files were searched using Protein Pilot software v4.2 (Sciex, Concord, ON, Canada) and Mascot v2.4 (Matrix Science Inc., Boston, MA, USA) using trypsin as enzyme, with 2 missed cleavages, a search tolerance of 50 ppm for the peptide mass tolerance, and 0.1 Da for the MS/MS tolerance. The UniProt Swiss-Prot reviewed database containing human proteins (version 01/02/2018, containing 42271 sequence entries), with a false discovery rate fixed at 1%.
Label-free quantification was carried out with PeakView 2.0 and MarkerView 1.2 (Sciex, Concord, ON, Canada). Six peptides per protein and six transitions per peptide were extracted from the DIA files. Shared peptides were excluded as well as peptides with modifications. Peptides with FDR lower than 1.0% were exported in MarkerView for the t-test (p-value < 0.05 and fold change > 1.3). Multivariate statistical analysis was performed through MetaboAnalyst 5.0 (
www.metaboanalyst.org). Cluster analysis was done by k-means after assessing the best number of clusters maximizing the average silhouette width. Bioinformatic analysis was carried out using Ingenuity Pathways Analysis (IPA) software (Qiagen), Gene Set Enrichment Analysis v7.4 (GSEA), DAVID tools (
https://david.ncifcrf.gov), STRING software (
https://string-db.org) [
39] and R version 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria) [
40] was used for statistical analysis including the use of the following packages: (“ComplexHeatmap” [
41], “ggplot2” [
42], “ggpubr” [
43], “PCAtools” [
44], “factoextra” [
45]. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD037551.
Discussion
Compared to the cutaneous counterpart, MM has been poorly investigated due to its rarity and lack of pre-clinical models. Generation of MM cell lines from the primary mucosal site is challenging with very little published. Moreover, none of the reported cell lines originated from MM of the sinonasal cavity [
75‐
77]. The first well-validated MM cell line, generated from an oral cavity MM, has been very recently proposed [
78]. This MM line shows low tumor-mutational burden, a lack of known oncogenic drivers, and is chemo-resistant to the vast majority of compounds currently approved for the melanoma treatment (BRAFi/MEKi, C-KIT inhibitors, anti-angiogenic kinase inhibitors, DNA-alkylating agents) [
78]. Very recently, a set of organoids have been established from human oral MM suggesting NGFR as relevant biomarker for resistance to anti–PD-1 therapy [
79]. In the present study, we could generate five novel human SN-MM cell lines (success rate over 10%) and confirm their identity through cell morphology, ultrastructural features, cell phenotype, and proteomic profile. SN-MM cell lines were tumorigenic and contain clusters of poorly differentiated MICs, resembling parental SN-MM tumors. Of clinical relevance, they resulted sensitive to cisplatin, but not to temozolomide, at variance with the proposed treatment protocol [
67,
80]. The introduction of immunotherapies has led to a shift in the systemic treatment of MM, in the adjuvant setting and for advanced disease. Their proteomic profile is consistent with transformed melanocytes showing a heterogeneous degree of melanocytic differentiation along with activation of cancer-related genes and pathways. By chemical inhibition, we could demonstrate a functional role for PI3K/Akt/mTOR to sustain tumor cell viability. These cellular systems might represent unique preclinical tools for a better understanding of the molecular landscape and the immunobiology of these neoplasms (manuscript in preparation) and, as an extension, to MM from other sites. These studies, will likely results in innovative treatment options and corresponding tissue biomarkers. Although with different levels of differentiation [
81], SN-MM are
bona fide melanocytic, as demonstrated by their expression of conventional melanocytic markers, largely comparable to the parental tumor. SOX10 and PRAME resulted positive in all tumor cells, whereas the other melanocytic markers showed extreme heterogeneity. It should be noted that phenotypic plasticity and undifferentiation (i.e. loss of diagnostic immunomarkers) within primary and metastatic malignant melanomas is an uncommon but well-documented phenomenon [
82,
83]. TEM analysis unveiled that SN-MM cells are endowed with melanosomes at different stages of development. Specifically, SN-MM5 is enriched in mature and immature melanosomes, whereas the melanosomes are rare or mainly immature in the remaining SN-MM cell lines. Accordingly, only SN-MM5 is melanotic in vitro and maintains strong and diffuse expression of HMB-45 and tyrosinase [
50]. In line with these data, one-third of head&neck-MM appear as amelanotic [
49]. It should be noted that TEM landmark studies confirmed various degrees of differentiation also in primary CM [
84].
How this level of de-differentiation account for the biological features of human MM is unknown. The malignant transformation of skin melanocytes occurs through a well-established sequential accumulation of genetic and molecular alterations involving precursors [
7,
85]. On the contrary, limited data are available on the cellular and molecular features sustaining mucosal melanocyte transformation and MM progression [
86].
As revealed by aCGH, structural rearrangements on SN-MM cell lines targeted a wide number of chromosome regions, including 1q, 5p, 6p, 6q, 8q and 9p, with a common consequence being gain of oncogenes, such as
TERT,
MYC,
KRAS, or loss of tumor suppressor genes, such as
ARID1B and
CDKN2A. The loss of p16
INK4a protein expression further support these data, suggesting that MM might benefit from therapeutic strategies targeting
CDKN2A loss, including the pharmacological inhibition of CDK4/6 (i.e. palbociclib and ribociclib) that are targets of p16
INK4a [
87]. Copy number gain of
BRAF and copy number deletion of
NF1 were also observed in most SN-MM cell lines, suggesting potential drivers for these melanocytic tumors. Moreover, copy number alterations involving components of the PI3K/Akt/mTOR pathway were found in SN-MM cell lines, including recurrent gains of
AKT,
RICTOR,
RPTOR,
NDRG1,
RPS6KB1,
RHEB, and loss of
SGK1, supporting a relevant role for PI3K/Akt/mTOR pathway in controlling cell growth in these tumors. Based on the proteomic analysis performed in this study, several pathways are involved in the SN-MM cells maintenance, including those controlling cell cycle and proliferation, senescence, cell motility, and metastatic spreading, the epigenetic and transcriptional regulation of gene expression, cancer cell metabolism and cellular response to oxidative stress as well as immune response [
10,
88].
SN-MM cell lines demonstrated different tumorigenic properties in vivo, that could be partially explained by their intrinsic fitness and differentiation trajectories, including a stem-like quiescent state, likely regulated by the surrounding microenvironment and marked by MITF expression [
89]. Invasive MITF
− cells require a longer period to initiate in vivo growing tumor by comparison with MITF
+ proliferative cells [
89]. Of note, based on the microscopic analysis of mouse skin xenografts of SN-MM cells, we could identify a proliferation of MITF
low/CDH1
−/CDH2
+/ZEB1
+/CD271
+ melanoma cell clusters losing most melanocytic markers, thus suggesting a MIC identity, as proposed in CM [
64,
90‐
92]. Although this outcome requires a set of additional experiments to be completely understood, this is in keeping with the existence of a dedifferentiated MM component endowed with an increased aggressiveness [
93,
94] and tumor-initiating cell features [
34,
95]. The malignant transformation of melanocytes involves the EMT that convey invasiveness and tumor-initiating potential. The “cadherin-switch” is critical for malignant melanocytes to escape from the primary tumor mass [
96]; cell–cell and cell–extracellular matrix adherence junctions are remodeled leading to cancer cell dissemination, and a new transcriptional program is activated. Due to epigenetic plasticity, the EMT is a reversible process and can give rise to cancer stem cells during the metastatic dissemination.
SN-MM lines display a partial melanocytic differentiation also loosing MITF; moreover, clusters of MITF
low/CDH1
−/CDH2
+/ZEB1
+/CD271
+ cells have been identified in mouse xenografts and parental biopsies of SN-MM. This population is likely driven by their MITF
low phenotype in their MIC functions and might undergo transition to differentiated melanoma cells. The upstream regulator analysis revealed inhibition of the MiT/TFE family of transcription factors, including MITF, TFEB, and TFE3, representing critical hubs in melanocytes differentiation, melanosome biogenesis, melanin biosynthesis, and melanosome transport [
97]. MITF coordinates many biological properties of melanocytes and might be used as a key molecular marker to distinguish between differentiated, proliferative (MITF
+/high), or invasive (MITF
−/low) melanoma phenotypic states [
98]. Specifically, MITF
high promotes melanocyte differentiation, proliferation, and survival, whereas MITF
low leads to invasion and senescence [
95,
97]. Moreover, MITF can regulate DNA damage repair, telomere maintenance [
99], and cell metabolism as mitochondrial biogenesis and oxidative phosphorylation [
100], activates SIRT1 expression, and together with TFEB controls lysosome biogenesis and autophagy [
95,
97,
101]; therefore, the MITF
low phenotype might justify some of the enriched pathways of proteomic analysis.
Previous data from our group [
16] and results from this study suggest a relevant role for the activation of the PI3K-Akt-mTOR pathway in MM. Activation might rely on PTEN loss in at least one-third of MM [
16]. Accordingly, co-existing PTEN loss and constitutive Akt activation were observed in two out of five cell lines. Novel treatment options have been recently proposed for PTEN-deficient cancer [
102]. PTEN loss sustain stem cell self-renewal and has been associated with cadherin switch [
103] during CM progression. The cadherin switch is regulated via PI3K/Akt by transcriptional up-regulation of Twist and Snail transcription factors [
103]. The observed outcome of PI3K-Akt-mTOR inhibition on SN-MM cells viability might thus rely on a direct effect also on the MIC reservoir. By flow cytometry analysis, we found that the NGFR/CD271
+ cells correspond to a significant fraction of the SN-MM population; however, this population contain also cells lacking a MIC phenotype. The MIC are the major contributors to tumor dissemination and persistence and tightly linked to drug resistance. Therefore, effort to identify and characterize MIC within the CD271
+ population are mandatory to identify appropriate therapeutic targets for this neoplasm [
104]. Furthermore, it should be noted that CD271/NGFR has been emerged as a relevant factor in the resistance to anti–PD-1 therapy in oral MMs, paving the way for new combination therapies with anti-PD-1 [
79]. The combination of PI3Ki with selective blocking of mTORC2 activity, by knocking-down RICTOR [
105], could avoid feedback loops and overcome chemo-resistance mechanism.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.