Loss of LYN is frequent in targeted-therapy resistant melanoma cells and favors metastatic properties.
2024; Elsevier BV; Linguagem: Inglês
10.1016/j.jid.2024.06.1286
ISSN1523-1747
AutoresCéline Gaudel, Céline Pisibon, Frédéric Soysouvanh, Serena Giuliano, Alexandra Picard‐Gauci, J. C. Leclerc, Paul Hofman, Philippe Bahadoran, Corine Bertolotto, Robert Ballotti,
Tópico(s)Cutaneous Melanoma Detection and Management
ResumoDespite the undeniable progress made with new therapies, notably targeted therapies (TT), and immunotherapies (Luke et al., 2017Luke J.J. Flaherty K.T. Ribas A. Long G.V. Targeted agents and immunotherapies: optimizing outcomes in melanoma.Nat Rev Clin Oncol. 2017; 14: 463-482Crossref PubMed Scopus (898) Google Scholar), prognosis of metastatic cutaneous melanoma remains poor. In patients with BRAFV600 mutation, TT using BRAF inhibitors, now in combination with MEK inhibitors show dramatic results in terms of response rates (70% of objective response), but frequent acquired secondary resistances lead to recurrence, decreasing the OS at 4 years around 20%. The mechanisms of resistance to TT have been extensively studied. In addition to new mutations, resistance to TT frequently involves a de-differentiation process (Falletta et al., 2017Falletta P. Sanchez-del-Campo L. Chauhan J. Effern M. Kenyon A. Kershaw C.J. et al.Translation reprogramming is an evolutionarily conserved driver of phenotypic plasticity and therapeutic resistance in melanoma.Genes Dev. 2017; 31: 18-33Crossref PubMed Scopus (80) Google Scholar), associated with increased mesenchymal and neural crest-like phenotypes. In this work we focused on SRC family kinases that comprises 8 members and have long been known to be oncogenes (Brickell, 1992Brickell P.M. The p60c-src family of protein-tyrosine kinases: structure, regulation, and function.Crit Rev Oncog. 1992; 3: 401-446PubMed Google Scholar). Several reports showed that resistance to BRAF inhibitors can be bypassed by a combination of BRAF and SRC inhibitors (Paris et al., 2022Paris A. Tardif N. Baietti F.M. Berra C. Leclair H.M. Leucci E. et al.The AhR-SRC axis as a therapeutic vulnerability in BRAFi-resistant melanoma.EMBO Mol Med. 2022; 14e15677Crossref PubMed Scopus (7) Google Scholar). However, the role of each member of the SRC family has not been studied so far. First, we isolated melanoma cell cultures from human melanoma tumors. Patients gave a written informed consent for participating to this research project that was approved by the Nice University Hospital Committee for Protection of Persons. Transcriptomic analysis from 6 melanoma cell cultures isolated in our laboratory showed high expression level of FYN, LYN, SRC and YES1 (sup Fig1.a). Focusing on these 4 kinases, we first analyzed 4 pairs of sensitive (S) and BRAF inhibitor (BRAFi) resistant (R) melanoma cells. As expected, all resistant cells showed a drastic decrease in the effect of BRAFi on cell viability compared to the sensitive counterpart (sup Fig 1.b). Western blot analysis of these cells showed a decrease in LYN expression in 3 out of 4 resistant cells (P#1R, M229R, WM9R). The other SRC kinases (YES1, FYN, SRC) did not show consistent changes between sensitive and resistant cells (Fig.1.a). Quantification confirmed the significant inhibition of LYN expression in resistant cells (sup Fig 1.c). Further, immunofluorescence studies in patient cells (P#1) and M229 cell line confirmed the loss of LYN in resistant cells, that have also lost SOX10 (sup Fig.1.c, d). Then, immunofluorescence on frozen sections of 2 melanoma tumors from patients P#3 and P#8, showed LYN expression in melanoma cells identified by SOX10 (Fig.1.b), confirming that LYN is expressed in melanoma cells in situ. These observations prompted us to evaluate the role of LYN in resistance of melanoma cells to TT. However, using cells from patient 1 (P#1) we showed that neither siRNA induced loss of LYN in P#1S (sup.Fig.2.b) nor adenovirus forced LYN expression in P#1R (sup.Fig.2.c) altered the sensitivity of melanoma cells to BRAF inhibitor (sup.Fig.2.a). These data indicate that the level of LYN expression alone is not sufficient to influence the sensitivity of melanoma cells to BRAFi. Then, we evaluated if the loss of LYN was involved in the aggressiveness and metastatic propensity of TT resistant melanoma cells. Analysis of skin cutaneous melanoma TCGA cohort, showed that low levels of LYN were associated with decreased survival (Fig.1.c) and GSEA comparative analysis of the CCLE top 20 melanoma cell lines expressing high or low LYN level, revealed that low LYN melanoma cells are enriched in invasive and dedifferentiated signature described by Tsoi et al. (Fig.1.d) (Tsoi et al., 2018Tsoi J. Robert L. Paraiso K. Galvan C. Sheu K.M. Lay J. et al.Multi-stage Differentiation Defines Melanoma Subtypes with Differential Vulnerability to Drug-Induced Iron-Dependent Oxidative Stress.Cancer Cell. 2018; 33: 890-904.e5Abstract Full Text Full Text PDF PubMed Scopus (521) Google Scholar). These observations suggest a role of LYN invasive properties of melanoma cells. Using Boyden chamber approach, we observed that inhibition of LYN expression by specific siRNA, increased migration of melanoma cells from patients, P#2 (Fig.1.e). Quantification of the effect of LYN silencing performed on 3 additional melanoma cells, P#4, P#5 and SKMel5 (Fig.1. f) confirmed that LYN inhibition increased migration of melanoma cells. The effect of siLYN on LYN expression was verified by western blot (Fig.1.g). Next, we demonstrated that MLR1023, a LYN kinase activator, induced phosphorylation of Tyr396 in the activation loop of LYN that generates a highly active LYN (Fig.2.a) (Gringeri et al., 2010Gringeri E. Carraro A. Tibaldi E. D'Amico F.E. Mancon M. Toninello A. et al.Lyn-mediated mitochondrial tyrosine phosphorylation is required to preserve mitochondrial integrity in early liver regeneration.Biochemical Journal. 2010; 425: 401-412Crossref Scopus (21) Google Scholar). MLR1023 treatment inhibited migration of melanoma cells from P#2 (Fig.2.b) and P#4 (sup.Fig.3.a, b). Quantification of MLR1023 effect in 5 melanoma cells cultures demonstrated a statistically significant effect of LYN activation on 3 cell cultures (Fig.2.c). The two cell cultures that do not respond to MLR1023 do not express LYN, strengthening the specificity of the observed effect (sup.Fig.3.c). Additionally, adenovirus forced expression of LYN, in melanoma cells (P#1R) induced a dose dependent inhibition of migration as shown by images of the Boyden chambers (Fig.2.d). Quantification (Fig.2.e) clearly showed up to 70% inhibition evoked by LYN forced expression. The expression of LYN, was shown in Fig.2.f. Forced LYN expression also increased p-LYN, but decreased p-FAK, in agreement with an inhibition of migration. This observation was confirmed in 3 additional melanoma cell cultures (sup.Fig.4.a) and correlated with the level of LYN expression (sup.Fig.4.b). Then using in vivo lung extravasation model, we observed that melanoma cells (P#1R) overexpressing LYN (Green) do not remain in lungs 24 hours after tail vein injection (Fig.2.g, h), while cells transduced by control adenovirus (red) were retained in lungs. This suggests that LYN-expressing cells are unable to cross the blood vessel wall to nest and grow in the lung parenchyma, confirming that forced expression and activation of LYN impairs the metastatic properties of melanoma cells, in vivo. SRC kinases, mainly SRC itself, have also been implicated in melanoma development. However, clinical trials using Dasatinib, the first SRC inhibitor, or saracatinib have not yielded significant results. (Gangadhar et al., 2013Gangadhar T.C. Clark J.I. Karrison T. Gajewski T.F. Phase II study of the Src kinase inhibitor saracatinib (AZD0530) in metastatic melanoma.Invest New Drugs. 2013; 31: 769-773Crossref PubMed Scopus (40) Google Scholar). LYN plays a key role in myeloid and B cell functions and triggers either pro-inflammatory or inhibitory signalling pathways. LYN was also described to play a protumoral role in some solid cancer such as breast and prostate cancer (Brian and Freedman, 2021Brian B.F. Freedman T.S. The Src-family Kinase Lyn in Immunoreceptor Signaling.Endocrinology. 2021; 162bqab152Crossref PubMed Scopus (30) Google Scholar). In this work, LYN silencing with specific siRNA increased melanoma cell migration, while LYN activation, or forced expression inhibited migration. LYN forced expression also inhibited in vivo lung colonisation and induces focal adhesion kinase (FAK) dephosphorylation that is associated with a decrease in cell motility. Therefore, counterintuitively to the pro-oncogenic role ascribed to LYN and SRC kinases in general, we demonstrated that LYN acted negatively on metastatic properties of melanoma cells. Loss of LYN in targeted therapy resistant melanoma cells is consistent with increased invasive properties associated with resistance that has been largely documented (Falletta et al., 2017Falletta P. Sanchez-del-Campo L. Chauhan J. Effern M. Kenyon A. Kershaw C.J. et al.Translation reprogramming is an evolutionarily conserved driver of phenotypic plasticity and therapeutic resistance in melanoma.Genes Dev. 2017; 31: 18-33Crossref PubMed Scopus (80) Google Scholar). One report, in glioblastoma also showed an anti-tumorigenic function of LYN, but the molecular mechanisms behind these specific functions remain unclear (Lewis-Tuffin et al., 2015Lewis-Tuffin L.J. Feathers R. Hari P. Durand N. Li Z. Rodriguez F.J. et al.Src family kinases differentially influence glioma growth and motility.Mol Oncol. 2015; 9: 1783-1798Crossref PubMed Google Scholar). As SRC and FYN are more tumorigenic than LYN in prostate cells (Cai et al., 2011Cai H. Smith D.A. Memarzadeh S. Lowell C.A. Cooper J.A. Witte O.N. Differential transformation capacity of Src family kinases during the initiation of prostate cancer.Proc Natl Acad Sci U S A. 2011; 108: 6579-6584Crossref PubMed Scopus (79) Google Scholar), we can argue that LYN competes with the other SRC kinases for the interaction with pro-metastatic partners and therefore dampen the pro-metastatic functions of the other SRC kinases. Increased LYN expression or LYN activation by the allosteric activator MLR1023, currently in clinical trial for type 2 diabetes, decreases metastatic properties of melanoma cells. It is tempting to propose a repositioning of the MLR1023 in the treatment of metastatic melanomas that express LYN to block further metastatic spreading or as adjuvant therapy to prevent metastasis development after resection of primary or lymph node metastasis melanoma. Further studies are required to evaluate the pharmacokinetic properties of MLR1023, and the blood concentration required for an efficient inhibition of melanoma cell metastasis. All original data generated in this study are available on request. Arrays original data can be accessed at https://www.ncbi.nlm.nih.gov/gds, with the accession number GSE270698. The authors state no conflict of interest. Investigation, Data Curation, Visualization and Writing: CG and CP equally contributed. Investigation, Data Curation: FS, SG, JL and APG. Resources: PB, PH. Conceptualization, Funding Acquisition, Supervision, Writing : CB and RB equally contributed. Melanoma cell lines, A375, SKmel5, WM9, M229 (Nazarian et al., 2010Nazarian R. Shi H. Wang Q. Kong X. Koya R.C. Lee H. et al.Melanomas acquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS upregulation.Nature. 2010; 468: 973-977Crossref PubMed Scopus (1841) Google Scholar) were cultured in RPMI medium (Invitrogen) supplemented with 7% FBS (Thermo Scientific) and Penicillin/Streptomycin (Invitrogen). Cells were maintained in 75 cm2 flasks in a humidified cell culture incubator with 5% CO2 at 37°C. Cells were authenticated before freezing and used before passage 20. The absence of mycoplasma was checked every other week. Cells were isolated from melanoma metastasis biopsies by digestion with an enzymatic cocktail of Collagenase A (Roche Diagnostics), Dispase II (Roche Diagnostics) and DNase I (Sigma) dissolved in serum-free RPMI medium supplemented with Penicillin/Streptomycin. After two hours, cell suspension was filtered through a 70μm filter (BD Falcon) and centrifuged for five minutes at 2000rpm. Cell pellet was resuspended in RPMI medium supplemented with 7% FBS and Penicillin/Streptomycin and cultured as described above. Cell cultures used in this work were used between passage 10 and 15. These cells were referred to as" short term culture" or "patient cells" (P# followed by an arbitrary patient number) and are listed in Supplementary Table 1. P#1S and P#1R cells were isolated respectively from patient#1, before treatment and in progression during treatment with targeted therapy. All patients gave a signed informed consent for participating to the research project that was approved by the Nice University Hospital Committee for Protection of Persons. The samples are collected by the laboratory of experimental and clinical pathology (accredited ISO 15189 norm) and the biobank BB-0033-00025 (certified S96-900 norm). Cell viability was determined by a colorimetric method using XTT, a tetrazolium salt cleaved by mitochondrial dehydrogenases in viable cells. Cells were seeded in 96-well plates at 0.2x105 cells per well in complete RPMI medium. 15h later, medium was supplemented with the indicated concentration of Dabrafenib or PLX4032. After 72 hours, XTT reagent (Roche Diagnostics) was added to the cells and optical density was measured at 450nm to measure relative cell viability. As mentioned above, P#1S and P#1R cells were isolated respectively from a patient before treatment and in progression during treatment with targeted therapy, respectively. All other cell lines M229, A375 and WM9 were rendered resistant by culturing them in increasing concentrations of PLX4032, until they proliferate in 5μM PLX4032. Cells were seeded in 60mm culture dishes in complete RPMI medium and subjected to the indicated treatments. Then, medium was removed, and cells were washed with PBS before adding lysis buffer (Tris 50mM pH 7.4, NaCl 150mM, Triton 1%) supplemented with Complete protease inhibitor cocktail (Roche Diagnostics) and PhosSTOP phosphatase inhibitor cocktail (Roche Diagnostics). Protein concentration in cell lysates was determined using Micro BCA Protein Assay (Pierce). 25μg of proteins were analyzed by SDS-PAGE and transferred on PolyVinyliDene Fluoride membranes (Immobilon) using the TransBlot Turbo transfer system (Biorad). The antibodies used were as follows: Src family antibody sampler kit (#9320), P-Src family (Tyr416) (#6943), P-FAK (#3283) all from Cell Signaling Technology, FAK (05-537) from Millipore, P-Lyn (Tyr396) (AJ1451b) from AbGent. The ubiquitously expressed Heat Shock Protein 90 (Hsp90) or beta-actin were detected using antibodies from Santa Cruz (sc-13119) and AbCam (ab8226) respectively and used as loading control. Signals were detected with horseradish peroxidase conjugated mouse or rabbit polyclonal antibodies (Cell Signaling Technologies) using ECL detection kit (ThermoFisher) and visualized by digital imaging (Fuji LAS4000) or autoradiography. Cells were seeded in 6-well plates at 4x105 cells per well in RPMI medium and then transfected with Lyn siRNA or non-targeted siRNA (Flexitube, Qiagen) at 25nM final using HiPerfect transfection reagent (Qiagen). After 48 hours, medium was removed, and cells were washed with PBS before being used for cell viability, Western Blot, and migration analyses. Cells were seeded in 6-well plates at 4x105 cells per well in RPMI medium and transduced with LYN or GFP adenovirus (Vector Biolabs) at 10 or 30 MOI. After 48 hours, medium was removed, and cells were washed with PBS before being used for cell viability, Western Blot, and migration analyses. To analyze migration, cells were seeded on 8μm cell culture inserts (Corning) at 300 000 cells in serum-free RPMI medium placed on top of RPMI medium containing 7% FBS in 24-well plates. After 24 hours, medium was removed, and cells were fixed with 4% PFA. The non-migrating cells were removed from the upper surface of the cell culture insert membrane by scrubbing. Afterwards cells were stained with a 0.2% Crystal Violet solution and washed with PBS. For short-term lung colonization assays, P#1R melanoma cells were transduced with LYN-GFP or control (Ctl) adenovirus. The cells were detached using HyQtase and LYN adenovirus transduced cells were labelled with 10 μM Cell Tracker red CMRA for 30 min, and an equal number of LYN or Ctl adenovirus transduced cells was injected into the tail veins of nude mice. The mice were euthanized after 30 min or 24 h, and lungs were examined by immunofluorescence microscopy. Animal experiments were performed in accordance with French law and approved by CIEPAL-AZUR and French ministry of research (APAFIS #35436-202203240948453 v1). Melanoma cells were seeded on coverslips at a confluence of 12 500 cells/cm2 in RPMI medium for 24 hours. Coverslips were rinsed with PBS, fixed with 4% PFA for 15 minutes, permeabilized with 0.1% Triton X-100 in PBS for 10 minutes and then blocked with blocking solution (0.1% Triton X-100 in PBS, 0.5% BSA, 1% goat serum (S26, Sigma)) for one hour. After successive washes in PBS, coverslips were incubated with blocking solution and primary antibodies LYN (1:200, 18135-1 ProteinTech) and SOX10 (1:200, MAB2864 R&Dsystems) overnight at 4°C in humid atmosphere. Coverslips were rinsed and then incubated with blocking solution and respective secondary antibodies (1:400, AlexaFluor 488 goat anti-rabbit IgG highly cross-adsorbed (Invitrogen) and AlexaFluor 594 goat anti-mouse IgG highly cross-adsorbed (Invitrogen)) for 90 minutes. After successive washes in PBS, coverslips were incubated with Hoechst solution in PBS (1:1000, H3570 Life Technologies) for nuclear staining for 5 minutes. Lastly, coverslips were rinsed with PBS and mounted in ProLong Antifade Mounting Reagent (P36934 Invitrogen). Representative fluorescent images were acquired using Nikon Confocal A1R with a 60x objective. Total RNA from 6 different melanoma cell cultures (P#1S, P#1R, P#4, P#6, P#7, P#8) was extracted using the RNeasy kit (Qiagen, Hilden, Germany). Integrity of RNA was assessed using an Agilent BioAnalyzer 2100 (Agilent Technologies) (RIN greater than 9). RNA samples were then labelled with Cy3 dye using the Low RNA Input QuickAmp kit (Agilent) as recommended by the supplier. Labelled cRNA probe was hybridized onto 8 × 60 K high-density SurePrint G3 gene expression human Agilent microarrays. Normalization and statistical analysis of microarray data were performed using the Limma package available from Bioconductor (http://www.bioconductor.org). Inter-slide normalization was performed using quantile methods. Expression data of the kinase of the SRC family were extracted and represented as a histogram. The 61 melanoma cell lines of the CCLE, https://portals.broadinstitute.org/ccle) data base were ranked by LYN expression, then GSEA (http://www.broad.mit.edu/gsea/) analyses were performed on the comparison of the 20 highest versus lowest LYN expressing cell lines, using the signatures described by Tsoi et al (Tsoi et al., 2018Tsoi J. Robert L. Paraiso K. Galvan C. Sheu K.M. Lay J. et al.Multi-stage Differentiation Defines Melanoma Subtypes with Differential Vulnerability to Drug-Induced Iron-Dependent Oxidative Stress.Cancer Cell. 2018; 33: 890-904.e5Abstract Full Text Full Text PDF PubMed Scopus (521) Google Scholar). The Kaplan Meier analysis of TCGA skin cutaneous melanoma cohort according to LYN expression level was performed on Xena (https://xenabrowser.net). Statistical analyses were performed with PRISM®, using the methods indicated in the figure legends.Tabled 1Supplemental table 1. List and characteristic of melanoma tumors used to isolate cells used in this work.REFPrimarySampleMutationP#1RC-11.33BackCutaneous MetBraf V600EP#1SC-10.12BackLymph NodeBraf V600EP#2SC-09.10LegCutaneous MetBraf V600EP#3C-13.10?Cutaneous MetBraf V600EP#4C-10.21BreastLymph NodeBRAF, NRas WTP#5C-13.08VertexLymph NodeBRAF, NRas WTP#6C-12.38?Cutaneous MetNRas Q61RP#7C-13.12AnkleLymph NodeBraf V600EP#8C-09.02Scalp (spitzoide)Cutaneous MetBRAF, NRas WT Open table in a new tab We thank Marylin Allegra and Grégory Salez for isolation and culturing of patients' melanoma cells. We are indebted to Tanesha Naiken for her help in animal experiments. We thank the C3M animal, image, and genomic facilities. This work received financial support from the French Institute of Cancer (InCA, PAIR Mélanome 2013), the foundation ARC (Labellisation 2022 to RB, PhD fellowship to CP), the French league against Cancer (Labellisation 2020 to CB) and the French Society of Dermatology (Fond de dotation 2020 to CB). Supplemental Figure 2: LYN levels do not affect BRAFi sensitivity of melanoma cellsShow full captiona) High-LYN expressing P#1S cells were transfected with control (siCtl) or LYN siRNA (siLYN). Low-LYN expressing P#1R cells were transduced with control (adCtl) or LYN (adLYN) adenovirus. Cell viability was evaluated by MTT after 72h exposure of 10nM of dabrafenib (n=4). Statistical analysis was performed by a two-tailed Mann-Whitney test. b) siLYN efficiency was analysed by western blot of P#1S cells after transfection with siCtl or siLYN. c) adLYN efficiency was analysed by western blot of P#1R cells after transduction with adCtl or ad LYN. Molecular weights in kDa are shown on the right.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplemental Figure 3. LYN activation by MLR1023 inhibits melanoma cell migration.Show full captiona) Cells isolated from human melanoma samples (P#4) were exposed to increasing doses of a LYN activator, MLR1023. LYN activation was monitored by western blot using anti-phospho-LYN (Tyr386) and total anti-LYN antibodies. HSP90 was used as loading control. Molecular weights in kDa are shown on the right. b) Images of Boyden chambers experiment with P#4 cells exposed or not to MLR1023 (10μM). Bar=150μM. c) Western blot analysis of LYN expression in the different cells using in experiments of figure 2. a, b, c. and supplemental figure 3. HSP90 was used as loading control. Molecular weights in kDa are shown on the right.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Supplemental Figure 4: LYN forced expression inhibits melanoma cells migration in vitro. a) P#3, P#6 and P#7 cells were transduced with GFP adenovirus (ad-GFP, 30 MOI) or LYN adenovirus (adLYN, 10 or 30 MOI) and subjected to migration in Boyden chambers. Quantification of the effect of ad LYN on migration (n=3) is shown. Statistical analysis was performed by a 2-way Anova and Tukey's multiple comparison test. b) Cells were lysed, and protein were analysed by western blot with antibodies to LYN and HSP90 was used as loading control. Molecular weights in kDa are shown on the right.View Large Image Figure ViewerDownload Hi-res image Download (PPT) a) High-LYN expressing P#1S cells were transfected with control (siCtl) or LYN siRNA (siLYN). Low-LYN expressing P#1R cells were transduced with control (adCtl) or LYN (adLYN) adenovirus. Cell viability was evaluated by MTT after 72h exposure of 10nM of dabrafenib (n=4). Statistical analysis was performed by a two-tailed Mann-Whitney test. b) siLYN efficiency was analysed by western blot of P#1S cells after transfection with siCtl or siLYN. c) adLYN efficiency was analysed by western blot of P#1R cells after transduction with adCtl or ad LYN. Molecular weights in kDa are shown on the right. a) Cells isolated from human melanoma samples (P#4) were exposed to increasing doses of a LYN activator, MLR1023. LYN activation was monitored by western blot using anti-phospho-LYN (Tyr386) and total anti-LYN antibodies. HSP90 was used as loading control. Molecular weights in kDa are shown on the right. b) Images of Boyden chambers experiment with P#4 cells exposed or not to MLR1023 (10μM). Bar=150μM. c) Western blot analysis of LYN expression in the different cells using in experiments of figure 2. a, b, c. and supplemental figure 3. HSP90 was used as loading control. Molecular weights in kDa are shown on the right.
Referência(s)