Identification of genetic fingerprint of type I interferon therapy in visceral metastases of melanoma | Scientific Reports

Scientific Reports volume 14, Article number: 26540 (2024) Cite this article Metrics details Malignant melanoma is a difficult-to-treat skin cancer with increasing incidence worldwide. Although type-I

Scientific Reports volume 14, Article number: 26540 (2024) Cite this article

Metrics details

Malignant melanoma is a difficult-to-treat skin cancer with increasing incidence worldwide. Although type-I interferon (IFN) is no longer part of guidelines, several melanoma patients are treated with type-I interferon (IFN) at some point of the disease, potentially affecting its genetic progression. We run genome-wide copy number variation (CNV) analysis on previously type-I IFN-treated (n = 17) and control (n = 11) visceral metastases of melanoma patients. Results were completed with data from the TCGA and MM500 databases. We identified metastasis- and brain metastasis-specific gene signatures mostly affected by CN gains. Some cases were genetically resistant to IFN showing characteristic gene alterations (e.g. ABCA4 or ZEB2 gain and alterations of DNA repair genes). Analysis of a previously identified type-I IFN resistance gene set indicates that only a proportion of these genes was exclusive for the IFN-treated metastases reflecting a possible selective genomic pressure of endogenous IFNs during progression. Our data suggest that previous type-I IFN treatment and/or endogenous IFN production by immune response affect genomic progression of melanoma which may have clinical relevance, potentially influence immune checkpoint regulation in the tumor microenvironment.

Interferons can be categorized into three groups based on their receptors. Type I interferons (α, β, ε, κ, and ω) are recognized by IFN-α/β receptors (IFNAR) including heterodimeric subunits, IFNAR1 and IFNAR2. Type II (γ) binds IFN-γ receptor 1 (IFNGR1), and type III (λ), which binds the heterodimeric receptor consisting of an IFN-λ receptor 1 and IL10 receptor β subunits1. In case of IFN-I binding IFNAR subunits dimerize resulting in the stimulation of STAT1/STAT2 heterodimers through the phosphorylation of several STAT activator kinases (e.g. JAK1, TYK2). STAT1/2 dimers then have the ability to form a complex with IRF9, which translocate to the cell’s nucleus to stimulate the expression of thousands of genes having interferon signaling response elements in their DNA sequence according to the Interferome database2. Type I interferons are also able to activate signaling by STAT3-6 and some other sovereign pathways, such as extracellular signal regulated kinase (ERK), p38 mitogen-activated protein kinase (MAPK), c-Jun N-terminal kinase (JNK) and mammalian target of rapamycin (mTOR)3. Both types of IFNs are involved in antitumoral immunity, but type-I IFN predominates in innate immunity4.

Melanoma is a chemo- and radiation therapy resistant tumor due to its large genetic burden often occurring in highly heterogenous tumor cell populations and to its resistance to programmed cell death5. On the other hand, melanoma is one of the most metastatic human malignancies. Cytokine therapy was long considered the only adjuvant therapeutic approach in cutaneous melanoma using type I IFN or combination of type I IFN and IL2 in case of high-risk cases6. In case of ulcerated melanomas, the efficacy of the IFN-therapy was 10–20%, in non-ulcerated tumors it was much lower and the average 5-year disease-free survival of IFN-treated melanoma patients was found to be ~ 6%, while the overall survival was 3%, suggesting that the overwhelming majority of skin melanoma can be considered as relatively IFN-resistant to this therapeutic protocol as well6. A previous study on large melanoma databases identified a type-I IFN stimulated gene expression signature (GES) which had strong prognostic power. Furthermore, it was also discovered that in low GES melanomas copy number loss of the type-I IFN gene cluster occurred in melanomas connecting type-I IFN production by melanoma cells to disease progression7.

Type I IFNs and IFN receptor gene deletion and downregulation is a frequently occurring phenomena in cancer. Majority of the studies investigating melanoma reported that CNAs of the IFN-γ pathway genes (IFNGR1, IRF-1, CXCL10, IFIT1, IFIT2 and IFIT3) correlate with shorter survival in patients with metastatic melanoma after immune checkpoint inhibitor (ICI) therapy8,9,10,11,12. Another study has been identified incomplete IFNγ pathway activity in melanoma cells through the analysis of the expression of IFNγ targets pre- and post-IFNγ stimulation13. However, the importance of IFNγ pathway is widely investigated in melanoma, there is not many publications on the type I IFN signaling in melanoma14,15. More recently we have identified a 79-gene type-I IFN resistance GES in preclinical human melanoma model which was validated in silico in TCGA. This analysis revealed a 5-gene set of the original GES which was associated with IFN treatment16. Furthermore, a 13-gene component of this resistance signature was also found to be associated with the efficacy of IFN-therapy16. Since we have established a frozen tissue metastatic melanoma cohort collected at autopsy which contains IFN treated cases as well, we raised the question if previous IFN treatment can affect the genetic progression of melanoma.

To explore the presence of a unique molecular signature linked to type-I interferon (IFN) treatment at the functional proteome level, we utilized proteomic data from a postmortem melanoma cohort that was part of the MM500 project17,18. Our analysis involved 67 hematogenous metastases, which were divided into two groups based on treatment status: metastases from patients who received previously type-I IFN (n = 28) and metastases from treatment-naive or patients receiving other therapies (n = 39). Through bioinformatic analysis of the differentially expressed proteome between the two groups, we identified a distinct signature associated with IFN treatment. Samples from IFN-treated patients exhibited significant downregulation of the transcription, splicing, and translation machineries, as well as a downregulation of SUMOylation. In contrast, immune system-related proteins, particularly those associated with the innate immunity, including neutrophil degranulation, were induced in IFN-treated tumors. Notably, IFN-treated samples showed an upregulation of phagosomal proteins, together with reactive oxygen species (ROS) and reactive nitrogen species (RNOS) production machineries (Fig. 1a). We confirmed our findings using a different bioinformatic approach at the global proteome level. The proteome dynamics between the two groups underwent 1D functional annotation enrichment analysis (see “Materials and methods”). Consistent with our previous observations, IFN treatment induced downregulation of proliferation-related pathways, including mitochondrial translation, which has recently been linked to high proliferation and disease progression19. This analysis also revealed the upregulation of immune system-related processes, as well as fatty acid- and lipid metabolisms including peroxisomal lipid metabolism in IFN-treated tumors (Fig. 1b). A feature selection analysis in a prospective study was also performed to identify proteins that most effectively distinguish between naive (N = 26) and interferon-treated local lymph node metastases (N = 7). A total of 65 proteins (23 up- and 42 downregulated) were significantly altered in treated samples, involved in cell adhesion and several pathways that influence tumor progression, such as Ras-MAPK, PI3K-Akt and Rap1 signaling (Fig. 1c, Table S1).

Functional molecular signature associated with interferon treatment. (a) Functional enrichment analysis of significantly dysregulated proteins between interferon treatment status groups. (b) 1D functional annotation enrichment analysis on the global proteome differences between both groups. (c) Feature selection analysis of prospectively collected naive (N = 26) and IFNα-treated (N = 7) lymph node melanoma metastases developed without therapeutic intervention or during or following IFN treatment.

Elucidating the molecular mechanisms underlying the tumor response to interferon treatment at the functional proteome level initiated a global genomic CNV analysis of type-I IFN treated metastatic samples of an independent cohort.

We collected a frozen sample cohort of 16 cases of metastatic melanoma derived from a previously established autopsy cohort (Table 1)20. Six cases were IFN therapy naive, while 10 patients were treated, beside other therapies, with IFNα2a. The frozen sample collection contained a good quality metastatic sample of lung-, liver- and brain metastases. The majority of cases were classical BRAFV600E/K mutants, there were 3 cases of NRAS mutants and only one IFN therapy naive case was triple negative (BRAF/NRAS/KIT wild type). Only one patient from the IFN-naive group and 2 ones from the IFN-treated group received BRAFi therapy. DTIC (dacarbazine) was administered in the IFN-treated group exclusively. It is of note, that this cohort derived from a pre-immunotherapy era, accordingly none of the patients received immune checkpoint inhibitors.

We applied seventeen IFN-treated and eleven IFN-naive metastatic samples to Affymetrix chips with array comparative genomic hybridization (aCGH) technique to determine the underlying copy number variation (CNV) landscape. Genome duplication was not observed in any of the included cases. We proceeded to further analysis only with those altered regions where difference of frequency was above the predetermined differential threshold (25%) with the statistical significance of p ≤ 0.05, and their alterations were consistent within the sample group analyzed. First, we checked whether type-I IFN genes were altered by CNV in melanoma metastases independently from the administered therapies. IFNA1/2/8/13/E genes were commonly altered by homozygous loss (hoCNL) in brain, liver and lung metastases, while other IFN type I genes were affected by hoCNL exclusively in lung metastases (Table S2A). Comparing the IFN-treated to naive group we did not find any differences in the loss of type-I IFN genes, indicating that this process is an early event during melanoma progression paralleled by the alterations of the nearby CDKN2A gene. Co-deletion of the IFN and CDKN2A genes was more frequent in the treated samples, but not exclusive.

Comparing the IFN-treated samples to naive ones, there were characteristic copy number gains (CNG) and copy number losses (CNL) on chromosomal arms 6p, 6q, 8p, 12p, 13q and 20p affecting 39 genes, while IFN-naive specific alterations were found in 288 genes localizing mainly on chromosomal arms 1p, 3p, 3q, 5p, 5q, 6p, 6q, 8p, 8q, 9p, 9q, 10p, 10q, 11q, 12p, 12q, 13q, 14q, 17p, 20p and 20q (Fig. 2a, Table S2B). In case of low CNGs none of the gene alterations were the result of chromosomal trisomy. By removing genes of regulatory RNAs and genes with uncertain function (LOC symbols) 22 and 204 in IFN positive and naive samples were further investigated, respectively. Next, we eliminated those CNV alterations which were inconsistent throughout the cohort (random changes) ended up in a 17-gene list of CNGs, only 5 of which (29.4%) were regulated by IFN type I (Table 2).

CNV landscape of interferon-treated melanoma metastases. (a) Frequency of copy number alteration types in IFN-naive and IFN-treated pooled metastases. Chromosomes and their regions can be found on the horizontal axis. (b) Proportion of genes altered by different copy number variation events in pooled melanoma metastases and distant visceral sites. IFN: interferon; LOH: loss-of-heterozygosity; CNG: copy number gain; CNL: copy number loss; CNV: copy number variation.

When we segmented our dataset according to the localization of visceral metastases and analyzed the CNV patterns using the beforementioned filtering criteria no IFN therapy specific alterations were observed in liver metastases (Table S2C). Regarding lung metastases we found CNGs affecting two genes (FMN2, GREM2) on 1q43 and another one at the 11q23.3 locus (DSCAML1) altered by CNL (Table 2, Table S2D). Out of them only GREM2 was proven to be IFN type I regulated. In case of IFN treated brain metastases, 37 genes were altered by CNG on 1q, 6p and 8p. In case of low CNGs none of the gene alterations were the result of chromosomal trisomy. After the exclusion of lincRNA, miRNA and LOC genes, the analysis was continued focusing on 27 CNG altered genes (Fig. 2b, Table S2E). Elimination of the inconsistent CNV changes (random effect of genetic instability) we ended up in a 24-gene set of brain-specific CNV alterations, characteristically CNGs (Table 2) where 10/24 (41.7%) genes were IRGs. Six out of them are regulated by IFN type I or both by type I and II.

In a previous study we identified DEGs of IFN-resistance which was predicted immune checkpoint inhibition efficacy in a large in-silico melanoma database16. The 79-gene set was tested on our IFN-treated metastasis cohort and found that 43 genes were altered in our distant metastasis cohort affected preferably by CNG (21/43, 72,1%) in the brain and lung, but not in liver metastases. Twenty-eight out of them (28/43, 65.1%) belong to the IRGs and 53,6% (15/28) of IRGs are IFN type I regulated (Table S3). Among those 43 genes only 15 was specific to IFN-treated samples (Table 3) while the waste majority of them (28/43) was also present to some extent in the non-treated samples as well (Table S3). In IFN-treated metastases 7/15 genes were above the prespecified 25% threshold where highly frequent (~ 50%) losses of BCORL1, HDAC8, DOCK11 were predominant beside the CPXM1, ZNF703 and FGF20 gains. It is also important to note, that the frequent IFN-resistance genes of the IFN-treated metastases barely IRGs unlike the rare events of this signature (Table 3). On the other hand, among common IFN-resistant genes TTTY2, PDE1C, NDRG1, AQP1, HSF1, DKK1, SDC2 and TPD52L1 were highly frequent (~ 50%) (Table S3). Nine of those common genes, BCORL1, AQP1, SDC2, DEK, SOX4, SLC17A3, DDX10, TNFSF10 and PTPB1 were associated with ICI therapy efficacy, but only BCORL1, DDX10 and PTBP1 were present exclusively in the IFN-treated cohort (Table 3)16. Furthermore, 5 out of 15 IFN-treatment-specific and 6 out of 28 common CNVs were rare events since they occurred only in a single case (Tables 3, S3).

Since we did not observe many genes with CN modifications in liver or lung metastases, we further investigated only brain metastases separately or all metastases from different visceral sites as a pooled sample population. We applied Kyoto Encyclopedia of Genes and Genomes (KEGG) and Reactome’s overrepresentation analysis on the beforementioned gene sets using hypergeometric distribution21. The generatedprobability score wascorrected by Benjamini–Hochberg method for false discovery rate. Regarding pooled metastases, there weren’t any significantly (FDR ≤ 0.05) enriched pathway, but interestingly most of them were associated with function of the endoplasmic reticulum, such as regulation of neuronal receptors and channels by MECP2 or unfolded protein response (Fig. S1A). In case of brain metastases, interesting pathways were affected by CNVs, such as defensins (CNG of DEFA4/6, DEFB1). L1CAM interactions, involved in cell adhesion, was also overrepresented by CNG events affecting CNTN2 and NFASC (Fig. S1B). In the KEGG Database, distinct signaling pathways important in melanoma progression such as Ras-MAPK-, PI3K-Akt or Rap1 were also affected (CNG ANGPT2, ELK4).

Copy number data from the TCGA Database was analyzed for interferon regulated genes (IRGs) in two different cohorts (Table S4A): (1) interferon (IFN; N = 18) vs. other, not immunotherapy (N = 51) treated metastases and (2) IFN-treated metastases with known therapy response (stable/progressive disease, N = 4 vs. complete/partial response, N = 4). Regarding genes exclusively altered in the different sample groups, in the first cohort 7.5 times more CN altered IRGs were detected in the IFN-treated samples, whereas in the second cohort 1,132 times more CN altered IRGs were detected in the none responder melanomas (Table S4B). These differences were highly significant at the level of 0.00001. Comparing to other treated samples, in case of IFN-treated ones a remarkably higher proportion of altered IRGs was regulated by IFN type I. However, in case of therapy response no such comparison was made, because only quite a few IRGs (N = 3, 0.04%) altered exclusively in the CR/PR group (Table S4B).

Comparing our strictly filtered data to the above gene lists and excluding the inconsistently altered genes, FKBP5, NRSN2, PRNP and SOX12 genes showed high-copy gain (average 4.3–4.4 copies) in the TCGA IFN-treated samples (Table S4C). In case of therapy response TCGA samples with SD/PD, except for DSCAML1, all the genes altered consistently in our original datasets (Table 2) were CN gained as well (average 4.0–5.5 copies), 26 (60.5%) of which were exclusive for this group. The rest of them were also present in the CR/PR group with a lower frequency (Table S4C).

By analyzing the distant metastasis- and brain metastasis-specific IFN-induced CNV patterns in the individual cases, we found cases where both types of genomic aberrations were present (IFN-sensitive/reactive cases 8, 12–14, 16), and there were cases where these signature alterations were completely absent (IFN-resistant cases 7, 9, 11), while in two cases these signature alterations were partially present in certain types of metastases (IFN-low responder cases 10, 15) (Tables 2, 4). The extent of genome affected by CNV changes was significantly smaller in IFN-resistant samples as compared to IFN-sensitive ones (9.3% ± 4.2% vs. 35.5% ± 3.9%, p = 0.007, Mann–Whitney U-test). Low-copy gain (avg. CN = 3) of ABCA4 (4/5, 80%) and ZEB2 (3/5, 60%) genes and heterozygous copy loss (avg. CN = 1) of ADAM5 (4/5, 80%) and ADAM3A (4/5, 80%) genes were exclusive for them. A higher number of LOHs at several chromosomes was also detected affecting 170 genes, and among them 75 (44.1%) were type-I IFN regulated (Table S5A). Based on Reactome pathway analysis, these type-I IFN regulated genes affected by LOH were mainly enriched in the interferon alpha/beta signaling pathway (p < 0.0001, FDR = 0.006) including OAS1/2/3 and PTPN11, and cell–cell communication (p < 0.001, FDR = 0.025) including ACTB, PIK3CB, CD2AP, PTPN11 and MAGI2 (Table S5B). However, the IFN-treated cases did not differ in respect of the CN loss of type-I IFN genes (Table S2A). CN alterations of a previously found “immunogenic mimicry” genes22,23 and DDR genes, predominantly the homologous directed repair (HDR; p = 0.004) and to a lesser extent mismatch (MMR; p = 0.034), base excision repair (BER; p = 0.028) and non-homologous end joining (NHEJ; p = 0.034) genes, were also significantly lower in IFN-resistant cases (Fig. S2, Table S5C).

The NCBI Gene Expression Omnibus (GEO) Database was screened for mRNA level data to evaluate the above IFN-related CNV patterns. Dataset selection, the inclusion and exclusion criteria and analysis workflow are detailed in the “Materials and methods”. Pre- (N = 227) and on-treatment (N = 87) primary melanoma samples were analyzed, as there is very little data available from metastases, with the vast majority of studies not indicating treatment and often the exact site of metastasis (local or distant organ). Higher expression of CNV pattern genes was significantly associated with shorter survival in anti-PD1 pretreated patients (HR = 1.69, p = 0.026; Figs. 3, S3A). Whereas an opposite, but unsignificant trend was observed for anti-CTLA4 treatment.

Association of the IFN signature mRNA expression with the overall survival (OS) in immune checkpoint inhibitor (ICI) pre-treatment primary melanoma samples (N = 227). Data was extracted from the GEO Database under the following GSE accession numbers: GSE91061, GSE115821, GSE2019, GSE165278 and GSE158403. Kaplan–Meier and Cox regression analyses were used, and p ≤ 0.05 considered statistically significant.

At the protein level, we have only limited data from local lymph node metastases developed without therapeutic intervention (N = 26) or during or after IFN (N = 7) or ICI (N = 5) treatments. When looking at the IFN signature genes separately, only downregulation of the BCORL1 gene was found to be significant in ICI-treated patients (Fig. S3B). However, the sample size did not allow a breakdown by individual therapies and no data is available from visceral metastases. In the above prospective IFN treatment study on localized lymph nodes, the significantly different genes were involved in pathways that show high overlap with the pathways affected by our IFN pattern (Table S1B).

Our global proteome analysis of an autopsy cohort of the MM500 project17,18 identified characteristic proteomic fingerprint of type-I IFN treated metastatic samples which included upregulation of those associated with innate immunity. It also revealed the upregulation of proteins involved in the ROS/rNOS production. During phagocytosis, NAPDPH oxidase (NOX) enzymes are activated, leading to the production of superoxide anions that are subsequently converted into other ROS, including hydrogen peroxide and hydroxyl radicals24. Although ROS generated within the phagosome play an important role in killing engulfed microorganisms and regulating intracellular signaling pathways, it may be partly responsible for the increased chromosomal instability associated with IFN-treatments. These observations paved the way for our genomic studies of an independent metastatic melanoma cohort. The global CNV analysis of visceral melanoma metastases treated with type-I IFN identified a 17-gene signature mostly affected by CNGs. It is not a surprise that 52.9% of these genes are IFN-regulated according to the Interferome v2.0 database2. These genes have divers functions, they are cell adhesion receptors (FLRT3 and SGCZ), signal transducers (FKBP5 and RAKGAPA2) and several of them are neural genes. The most interesting is that this gene set contains a transcription factor SOX12 and a DNA repair enzyme XRN2. It is important to note, that in this signature ARMC12 is an oncogene of neuroblastoma, while CSMD1 is a metastasis suppressor gene25,26.

It was interesting that a smaller proportion of type-I IFN treated melanoma cases (3/8) did not contain IFN-treatment related specific CNVs (type-I IFN-resistant cases). These cases were characterized by a lower level of chromosomal instability, less CNV involvement of DNA repair genes and the so-called immunogenic mimicry genes22 but a higher number of LOHs of several IFN-regulated genes. This group was characterized only by the gain of ABCA4, an ABC transporter gene previously not identified in melanoma, and ZEB2, of which increased expression in melanoma induces proliferation and growth at both primary and secondary sites27, along with the loss of ADAM3A/5, two pseudogenes of which CN loss is associated with high-risk and worse outcome in other tumors28.

Previously we identified an IFN-type I resistance GES in a preclinical model of human melanoma, validated in TCGA16. Our analysis of visceral metastases of melanoma patients indicated that 43 of those genes are affected by CNV, mostly by CNGs. However, only 15 genes were present exclusively in type-I IFN treated samples. The most commonly altered genes of this signature; BCORL1 is a transcriptional repressor which is involved in BRAF inhibitor sensitivity of melanoma29, HDAC8 is a histone modifier which is also involved in BRAF inhibitor resistance of melanoma30, CPXM1 is a carboxypeptidase, DOCK11 is a GEF, ZNF703 is an oncogene of the luminal B breast cancer31, DDX10 is a tumor suppressor in ovarian cancer and osteosarcoma and FGF20 is a cytokine. Among the “rare events” of this signature, PDE9A is a phosphodiesterase, SYDE1 is a Rho-GTPase activator, neither was reported to be involved in melanoma. MX1 is an IRG involved in IFN-resistance of melanoma32. ZNF444/EZH2 is a transcription factor involved in the metastatization of melanoma33. PTBP1 is a hnRNA-binding protein in mRNA processing reported to be involved in the tumorigenesis of melanoma34. PLA2G15 belongs to the phospholipase family, and is known to be involved in cancer progression including melanoma35. CALM3 is a Ca-signaling molecule which was accidentally found to be differentially expressed by melanoma36. Looking at the list of these genes specific to IFN-type I treated melanoma samples, none of them was previously connected to tumor immunology, rather to tumor biology in general. On the other hand, the presence of the large number of common “IFN-resistance” CNVs in the non-treated metastases may reflect the selective genomic pressure of the local, endogenous IFNs during progression of the disease.

Analysis of the individual metastasis types revealed other type-I IFN specific CNVs. Interestingly, such CNVs were completely missing in liver metastases, only 3 additional genes were affected in lung metastases but the brain metastases contained a 23-gene set. Several of those genes belong to the defensin family of antimicrobial peptides (DEFA and DEFB), this signature contains also an ETS type transcription factor EML4, a GTPase (RAB28), and an angiogenic cytokine ANGPT2, beside several neural genes involved in brain development. Furthermore, similar to the visceral signature, the brain signature also contains a DNA repair enzyme, NUCKS1, shown to be involved in melanoma metastasis37. It is tempting to speculate that type-I IFN treatment may specifically affect the progression of the disease to brain.

Based on the TCGA gene level copy number data, interferon-regulated genes have been enriched in IFN-treated distant metastases, but especially in non-responders having stable or progressive disease after IFN treatment. Comparison of data from our samples with the altered IRG gene set from interferon-treated TCGA SKCM samples, consisting mainly of distant dermal metastases, only 4 genes (FKBP5, NRSN2, PRNP and SOX12) were affected by high CNG in both sample sets. The immunophilin FKBP5 is capable of immunosuppression in melanoma38. And some studies provide data that both PRNP and SOX12 play a role in melanoma progression, including cell migration39,40. Meanwhile alterations of NRSN2 is previously not reported in melanoma. Furthermore, of the 43 genes that were altered in our visceral metastasis cohort, 26 were also exclusively characterized by copy number gain in the TCGA samples with SD/PD response.

All these data support the idea that type-I IFN treatment of melanoma patients significantly affect the genomic progression of the disease. In preclinical models, it was demonstrated that type-II IFN and CTLs can induce genomic instability in tumor cells by modulating the expression of DNA damage sensors (such as ATR) or repair enzymes (such as APOBEC) resulting in progressive chromosomal instability41. It is interesting that genomic progression of melanoma after type-I IFN treatment results in genomic gains in genes of various transcription factors (SOX4, SOX12, DEK), DNA repair enzymes (XRN2, NUCKS1, DCLRE1B) and of those involved in the innate immune escape (defensins, proteoglycans, CSMD1). Furthermore, recent studies revealed a “long-distance” modulation of type-II IFN on gene expression of tumor cells42,43. If locally produced type-II IFN has such an effect, one can hypothesize that therapeutically administered type-I IFN might also has such.

IFN-treated metastatic melanoma samples were also tested for a GES identified in type-I IFN resistant melanoma in a preclinical model16. We identified that the majority of this IFN-resistance signature genes have been affected by CNVs, mostly by gains. However, only a minority of this signature was specific to IFN-treated samples while the majority was common in both sample types. Since 19/30 of these genes are IRGs, we suspect that the common IFN-resistance signature gene alteration may be induced by endogenous IFN-exposure of the progressing tumor, while in case of the type-I IFN treated samples this signature was completed by another set of gene alterations. Our IFN-resistance GES identified in human melanoma was tested in a large ICI-treated melanoma database which revealed that 11 genes are predictive for treatment efficacy16. Nine out of those 11 genes (BCORL1, AQP1, SCD2, DEK, SOX4, SLC17A3, DDX10, TNFSF10 and PTBP1) are affected by CNG in our visceral metastases but only BCORL1, DDX10 and PTBP1 were unique in IFN-treated melanoma metastases. None of these genes were linked clinically to IFN-treatment or immunotherapy of melanoma. IFN-treated samples in which the IFN pattern was undetectable (“IFN-resistant”) were all characterized by the heterozygous loss of BCORL1, HDAC8 and DOCK11 genes. BCORL1, a subunit of the epigenetic modifier polycomb repressive complex, is frequently mutated mainly in human leukemias44. But its co-mutation with BRAF in melanoma contributes to inhibitor vemurafenib resistance29. HDAC8, which catalyzes the deacetylation of both histones and nonhistone proteins, regulates a wide variety of cellular processes45. HDAC8 inhibitor therapy appears to reduce resistance to PD1 and BRAFi treatments30. Although it is considered oncogenic, in our cohort its genetic loss was characteristic for “IFN-resistant” samples, suggesting a complex role in the pathogenesis of these melanoma samples. These alterations suggest a role for epigenetic landscape rearrangements in IFN therapeutic response and raises the importance of synthetic lethality. Retrospective analysis of treatment outcome of melanoma patients exposed or not to previous type-I IFN can answer the question whether such therapy may affect the efficacy of immunotherapy. Type-I IFN treatment of melanoma patients is not part of guidelines anymore. However, there is a quite large cohort of melanoma patients which was treated at some point of their disease; accordingly, the genomic status of their metastatic tumors can be a prognostic or predictive factor when the immunotherapeutic or any type of target therapies are considered. On the other hand, there are numerous clinical combination trials on immune checkpoint inhibitors and type-I IFN which are based on strong preclinical data that type-I IFN has immunogenic effects in melanoma cells46,47,48.

The used IFN-treated melanoma sample size is relatively low. The majority of studies on distant metastasis use a similar sample size to ours49,50,51,52. This is because visceral metastasis samples are fundamentally difficult to obtain. The main reasons are (1) usually no surgery, (2) late autopsy, (3) very distant from the primary in time, (4) very heterogeneous sites typical of melanoma, (5) no adequate information about the therapy or metastatic site, etc. Mainly distant skin, lymph node, or possibly brain metastases occur, in which latter case the melanoma disease is usually detected because of the metastasis.

The research is mainly based on autopsy samples. Sample quality was affected by the time between death and autopsy. Typically, the limit was 48 h for the organs examined. However, other death-related processes cannot be excluded. Accordingly, our results are primarily descriptive. Future functional validation in the form of either in vitro experimental or prospective studies is needed.

Additional limitation is that only a small number of patients with metastatic melanoma receive monotherapy, so the effect of other therapies cannot be completely ruled out. Our priority was to include patients who did not receive any other types of immunotherapies and received long-termed IFN treatment (more than 1 year), as the development of de novo genetic changes takes time. For other (co)therapies, samples started shortly before death or received for a relatively short period were considered.

Frozen autopsy melanoma specimens were collected from the 2nd Department of Pathology and the 1st Department of Pathology and Experimental Cancer Research at the Semmelweis University (Budapest, Hungary) within 48 h after death according to the relevant regulations. FFPE (formalin-fixed paraffin-archived) archived surgical samples were gathered from several institutions including the Semmelweis University (2nd Department of Pathology, 1st Department of Pathology and Experimental Cancer Research and Department of Dermatology, Venereology and Dermatooncology), the Szent György University Teaching Hospital, Department of Dermatology (Székesfehérvár, Hungary), the Nyírő Gyula and Honvéd Hospitals (Budapest, Hungary), the National Institute of Oncology (Budapest, Hungary) and the National Institute of Clinical Neurosciences.

This study was accepted by the Semmelweis University Regional and Institutional Committee of Science and Research Ethics [document no.: 191-4/2014], and was performed in accordance with the Declarations of Helsinki and all relevant regulations and guidelines. A written informed consent was obtained from the patients when it was relevant. All lesions were diagnosed, and the percentage of tumor cells was determined by using FFPE tissue sections stained with hematoxylin–eosin. Tumor to normal cell ratio was 76% (± 20) and 78.9% (± 18) of the investigated primary tumors (N = 10) and melanoma metastases (N = 28), respectively.

Genomic DNA was purified using QIAamp DNA FFPE Tissue and QIAamp DNA Mini Kits (BioMarker Kft., Gödöllő, Budapest, Hungary) by following of the provided protocol of the manufacturer. Quantity and quality of DNA were both checked by using NanoDrop ND-1000 UV–visible spectrophotometer (Thermo Fisher Scientific, Inc., Waltham MA, USA) and Qubit dsDNA HS Assay on a Qubit 1.0 fluorometer (Thermo Fisher Scientific, Inc., Waltham MA, USA). DNA samples with 260/280 ratio more than 1.8 were used for further analysis. Fragment analysis of the previously isolated DNA was carried out applying 1% agarose gel electrophoresis as well as BioAnalyzer 2100 using High Sensitivity DNA Kit (Agilent Technologies, Inc., Santa Clara CA, USA). Samples with sufficient quality were labeled and hybridized to CytoScan HD or OncoScan FFPE microarrays (Affymetrix Inc., Santa Clara, California, USA). Labeling, hybridization, and imaging setup were performed by UD GenoMed Medical Genomic Technologies Ltd. (Clinical Genomic Center, University of Debrecen, Debrecen, Hungary) according to the contemporary laboratory protocol using 600 ng of sample DNA.

Molecular inversion probes (MIP) based arrays were used to identify genome-wide CNV in DNA extracted from FFPE tumor samples and its matched distant metastasis pairs. Reference DNA was provided by the manufacturer and was used as controls in each array batch. Quality scores were calculated for each sample, and specimens with poor data quality were disapproved for further analysis. CNVs were then segmented using the SNP-FASST2 (Fast Adaptive States Segmentation Technique) algorithm in ChAS (Affymetrix, Inc., Santa Clara CA, USA). DNA samples with 120–200 bp fragment length were hybridized onto OncoScan FFPE arrays (N = 20), while less fragmented DNA samples, mainly from snap-forzen tissues were hybridized onto CytoScan HD arrays (N = 18). Visualization and statistical analysis to determine copy number alterations (CNAs) and LOH were performed using both Nexus Copy Number Discovery 10.0 (BioDiscovery, Inc., El Segundo CA, USA) and ChAS 2.1.0.16 (r6634) (Affymetrix, Inc., Santa Clara CA, USA) software working with the processed CHP files. We determined a significance threshold of 0.05, differential threshold of 25% and specified 1 kbp as the maximum spacing between adjoining probes. To eliminate noise derived from small CNAs, the minimum number of probes per segment was fixed at 25. To detect gains and losses, the following log2 ratio thresholds were set: ± 0.3 for gains and losses, 0.7 for high CN gains and − 1.0 for homozygous deletions. In case of allelic events, LOH calls smaller than 2000 bp and not overlapped by a minimum of 25 probes were banned for further analysis. All probes on chromosomes X and Y were excluded to avoid sex bias. Altered regions of 100% overlap with segmental duplications were also removed.

Gene level copy number data of skin cutaneous melanoma (SKCM) tissue samples (1) primary (N = 103) vs. distant organ metastases (N = 67), (2) filtered for metastatic samples with known therapy for the uploaded tissue data [IFN (N = 18) versus other (N = 51)] or (3) IFN treated metastases with known therapy response [stable/progressive disease (N = 4) vs. complete/partial response (N = 4)] were used for analysis (Table S4). Immunotherapy other than IFN was excluded from the analysis. Genes were excluded if: (1) not protein coded, (2) altered either direction (gain or loss) in less than 50% of samples of the given groups or (3) not interferon regulated.

For database establishment, we screened the NCBI Gene Expression Omnibus (GEO) repository for relevant datasets using the following keywords: “melanoma” AND “human [organism] AND (anti-PD-1 OR anti-PD-L1 OR anti-CTLA-4)” and “human [organism] AND (list of the available immunotherapy agents, including pembrolizumab, nivolumab, atezolizumab, durvalumab, avelumab, cemiplimab, ipilimumab, camrelizumab, cintilimab, tislelizumab, and toripalimab)”. We excluded datasets that lacked available gene expression or clinical data, as well as those containing single-cell RNA-sequencing, T or B cell receptor sequencing, non-mRNA-sequencing data, or studies involving cell lines, stem cells, or sorted peripheral blood mononuclear cells. Additionally, studies in mice, non-cancer studies, and GEO SuperSeries files were omitted. To further expand our dataset, we conducted a thorough literature review.

For data integration, we combined gene expression data from all eligible studies (GSE91061, GSE115821, GSE2019, GSE165278 and GSE158403) into a single dataset, performed quantile normalization, and scaled the values to 1000. Clinical data were annotated by saving the reported clinical and pathological data and survival times. Tumor specimens collected prior to therapy initiation were designated as “pre-treatment,” while those obtained during or after therapy were classified as “on-treatment.”

To investigate the association between gene expression and relapse-free survival in immunotherapy-treated melanoma patients, we conducted a Cox regression analysis. For this analysis, we utilized the gene set described above. In the survival analysis, rather than selecting a fixed cut-off point for gene expression, we systematically tested all possible values between the lowest and highest quartiles of expression levels for the combined signature. Given the extensive number of tests performed, we addressed the potential for false positives by applying a False Discovery Rate (FDR) correction for multiple hypothesis testing. To visually illustrate the impact of gene expression on survival outcomes, Kaplan–Meier survival curves were generated. These plots were based on the cut-off points determined through Cox regression, demonstrating the relationship between gene expression levels and patient survival.

The regulation of genes by any type of interferon (I, II or III) was checked using the Interferome Database v2.012 with p-value < 0.05 and fold-change ≥ 2.0 restricted to data from human experiments.

The experiments were conducted within the “Human Melanoma Proteome Atlas” project as described previously17,18. Briefly, fresh-frozen tumor samples and matched blood samples were collected from distinct University clinics in Sweden and Hungary following the same biobanking protocol. All studies were approved by the local ethical committees, including the Regional Ethical Committee at Lund University, Southern Sweden (DNR191/2007, 101/2013 (BioMEL biobank), 2015/266 and 2015/618), Semmelweis University, Hungary (191–4/2014), University of Szeged, Hungary (MEL-PROTEO-001). The study was carried out according to the Declarations of Helsinki and all relevant regulations and guidelines. All patients provided written informed consent. Histopathological evaluation was then performed on formalin-fixed paraffin-embedded (FFPE) specimens using hematoxylin–eosin staining. Mass spectrometry (LC–MS/MS) with two distinct setups was used to identify and quantify protein composition of samples using Proteome Discoverer software (Thermo Fisher Scientific, MA USA) against Uniprot Human dataset. Bioinformatic analysis was performed applying Perseus software using log2-normalized protein intensities and standardization by subtracting individual intensity values by the median in each sample. Pview software was used to identify and calculate the site‐specific acetylation occupancy, and motifeR was used both to align the phosphopeptide sequences with the background database, and enrich phosphorylation motifs and retrieve kinase‐substrate annotation. A feature selection analysis was further performed to identify proteins that most effectively distinguish between naive and interferon-treated lymph node metastatic samples developed without therapeutic intervention or during or following IFN treatment. The results of this analysis were displayed in a heatmap, highlighting both upregulated and downregulated proteins.

The data interpreted in this study have been deposited in NCBI’s Gene Expression Omnibus (GEO) Database and are accessible through GEO Series accession number GSE185165 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE185165).

Lee, A. J. & Ashkar, A. A. The dual nature of type I and type II interferons. Front. Immunol. 9, 2061. https://doi.org/10.3389/fimmu.2018.02061 (2018).

Article CAS PubMed PubMed Central Google Scholar

Rusinova, I. et al. Interferome v2.0: an updated database of annotated interferon-regulated genes. Nucleic Acids Res. 41, D1040-1046. https://doi.org/10.1093/nar/gks1215 (2013).

Article CAS PubMed Google Scholar

Lu, M. et al. Interferon-alpha targets JAK2V617F-positive hematopoietic progenitor cells and acts through the p38 MAPK pathway. Exp. Hematol. 38, 472–480. https://doi.org/10.1016/j.exphem.2010.03.005 (2010).

Article CAS PubMed PubMed Central Google Scholar

Zitvogel, L., Galluzzi, L., Kepp, O., Smyth, M. J. & Kroemer, G. Type I interferons in anticancer immunity. Nat. Rev. Immunol. 15, 405–414. https://doi.org/10.1038/nri3845 (2015).

Article CAS PubMed Google Scholar

Kang, K., Xie, F., Mao, J., Bai, Y. & Wang, X. Significance of tumor mutation burden in immune infiltration and prognosis in cutaneous melanoma. Front. Oncol. 10, 573141. https://doi.org/10.3389/fonc.2020.573141 (2020).

Article PubMed PubMed Central Google Scholar

Di Trolio, R., Simeone, E., Di Lorenzo, G., Buonerba, C. & Ascierto, P. A. The use of interferon in melanoma patients: a systematic review. Cytokine Growth Factor Rev. 26, 203–212. https://doi.org/10.1016/j.cytogfr.2014.11.008 (2015).

Article CAS PubMed Google Scholar

Rozeman, E. A. et al. Survival and biomarker analyses from the OpACIN-neo and OpACIN neoadjuvant immunotherapy trials in stage III melanoma. Nat. Med. 27, 256–263. https://doi.org/10.1038/s41591-020-01211-7 (2021).

Article CAS PubMed Google Scholar

Certa, U., Seiler, M., Padovan, E. & Spagnoli, G. C. High density oligonucleotide array analysis of interferon- alpha2a sensitivity and transcriptional response in melanoma cells. Br. J. Cancer. 85, 107–114. https://doi.org/10.1054/bjoc.2001.1865 (2001).

Article CAS PubMed PubMed Central Google Scholar

Krepler, C. et al. Pegylated and conventional interferon-alpha induce comparable transcriptional responses and inhibition of tumor growth in a human melanoma SCID mouse xenotransplantation model. J. Investig. Dermatol. 123, 664–669. https://doi.org/10.1111/j.0022-202X.2004.23433.x (2004).

Article CAS PubMed Google Scholar

Katlinskaya, Y. V. et al. Suppression of type I interferon signaling overcomes oncogene-induced senescence and mediates melanoma development and progression. Cell Rep. 15, 171–180. https://doi.org/10.1016/j.celrep.2016.03.006 (2016).

Article CAS PubMed PubMed Central Google Scholar

Gao, J. et al. Loss of IFN-gamma pathway genes in tumor cells as a mechanism of resistance to anti-CTLA-4 therapy. Cell. 167, 397-404.e399. https://doi.org/10.1016/j.cell.2016.08.069 (2016).

Article CAS PubMed PubMed Central Google Scholar

Rodriguez, T. et al. Distinct mechanisms of loss of IFN-gamma mediated HLA class I inducibility in two melanoma cell lines. BMC Cancer. 7, 34. https://doi.org/10.1186/1471-2407-7-34 (2007).

Article CAS PubMed PubMed Central Google Scholar

Alavi, S. et al. Interferon signaling is frequently downregulated in melanoma. Front Immunol. 9, 1414. https://doi.org/10.3389/fimmu.2018.01414 (2018).

Article CAS PubMed PubMed Central Google Scholar

Buzzai, A. C. et al. Diverse anti-tumor immune potential driven by individual IFNalpha subtypes. Front. Immunol. 11, 542. https://doi.org/10.3389/fimmu.2020.00542 (2020).

Article CAS PubMed PubMed Central Google Scholar

Heise, R. et al. Interferon alpha signalling and its relevance for the upregulatory effect of transporter proteins associated with antigen processing (TAP) in patients with malignant melanoma. PLoS ONE. 11, e0146325. https://doi.org/10.1371/journal.pone.0146325 (2016).

Article PubMed PubMed Central Google Scholar

Ladanyi, A. et al. Identification of a tumor cell associated type I IFN resistance gene expression signature of human melanoma, the components of which have a predictive potential for immunotherapy. Int. J. Mol. Sci. https://doi.org/10.3390/ijms23052704 (2022).

Article PubMed PubMed Central Google Scholar

Betancourt, L. H. et al. The human melanoma proteome atlas-defining the molecular pathology. Clin Transl. Med. 11, e473. https://doi.org/10.1002/ctm2.473 (2021).

Article CAS PubMed PubMed Central Google Scholar

Betancourt, L. H. et al. The human melanoma proteome atlas-complementing the melanoma transcriptome. Clin. Transl. Med. 11, e451. https://doi.org/10.1002/ctm2.451 (2021).

Article CAS PubMed PubMed Central Google Scholar

Gil, J. et al. An observational study on the molecular profiling of primary melanomas reveals a progression dependence on mitochondrial activation. Cancers. https://doi.org/10.3390/cancers13236066 (2021).

Article PubMed PubMed Central Google Scholar

Doma, V., Karpati, S., Raso, E., Barbai, T. & Timar, J. Dynamic and unpredictable changes in mutant allele fractions of BRAF and NRAS during visceral progression of cutaneous malignant melanoma. BMC Cancer. 19, 786. https://doi.org/10.1186/s12885-019-5990-9 (2019).

Article CAS PubMed PubMed Central Google Scholar

Milacic, M. et al. The Reactome Pathway Knowledgebase 2024. Nucleic Acids Res. 52, D672–D678. https://doi.org/10.1093/nar/gkad1025 (2024).

Article PubMed Google Scholar

Papp, O. et al. Organ specific copy number variations in visceral metastases of human melanoma. Cancers. https://doi.org/10.3390/cancers13235984 (2021).

Article PubMed PubMed Central Google Scholar

Timar, J., Honn, K. V., Hendrix, M. J. C., Marko-Varga, G. & Jalkanen, S. Newly identified form of phenotypic plasticity of cancer: immunogenic mimicry. Cancer Metastasis Rev. 42, 323–334. https://doi.org/10.1007/s10555-023-10087-1 (2023).

Article CAS PubMed PubMed Central Google Scholar

Marchi, L. F., Sesti-Costa, R., Ignacchiti, M. D., Chedraoui-Silva, S. & Mantovani, B. In vitro activation of mouse neutrophils by recombinant human interferon-gamma: increased phagocytosis and release of reactive oxygen species and pro-inflammatory cytokines. Int. Immunopharmacol. 18, 228–235. https://doi.org/10.1016/j.intimp.2013.12.010 (2014).

Article CAS PubMed Google Scholar

Ermis Akyuz, E. & Bell, S. M. The diverse role of CUB and Sushi Multiple Domains 1 (CSMD1) in human diseases. Genes. https://doi.org/10.3390/genes13122332 (2022).

Article PubMed PubMed Central Google Scholar

Li, D. et al. Armadillo repeat containing 12 promotes neuroblastoma progression through interaction with retinoblastoma binding protein 4. Nat. Commun. 9, 2829. https://doi.org/10.1038/s41467-018-05286-2 (2018).

Article ADS CAS PubMed PubMed Central Google Scholar

Vandamme, N. et al. The EMT transcription factor ZEB2 promotes proliferation of primary and metastatic melanoma while suppressing an invasive, mesenchymal-like phenotype. Cancer Res. 80, 2983–2995. https://doi.org/10.1158/0008-5472.CAN-19-2373 (2020).

Article CAS PubMed Google Scholar

Barrow, J. et al. Homozygous loss of ADAM3A revealed by genome-wide analysis of pediatric high-grade glioma and diffuse intrinsic pontine gliomas. Neuro Oncol. 13, 212–222. https://doi.org/10.1093/neuonc/noq158 (2011).

Article CAS PubMed Google Scholar

Mologni, L. et al. Concomitant BCORL1 and BRAF mutations in vemurafenib-resistant melanoma cells. Neoplasia. 20, 467–477. https://doi.org/10.1016/j.neo.2018.02.009 (2018).

Article CAS PubMed PubMed Central Google Scholar

Emmons, M. F. et al. HDAC8 regulates a stress response pathway in melanoma to mediate escape from BRAF inhibitor therapy. Cancer Res. 79, 2947–2961. https://doi.org/10.1158/0008-5472.CAN-19-0040 (2019).

Article CAS PubMed PubMed Central Google Scholar

Klaestad, E. et al. ZNF703 gene copy number and protein expression in breast cancer; associations with proliferation, prognosis and luminal subtypes. Breast Cancer Res. Treat. 186, 65–77. https://doi.org/10.1007/s10549-020-06035-0 (2021).

Article CAS PubMed Google Scholar

Lucarini, V. et al. Combining type I interferons and 5-Aza-2’-deoxycitidine to improve anti-tumor response against melanoma. J. Investig. Dermatol. 137, 159–169. https://doi.org/10.1016/j.jid.2016.08.024 (2017).

Article CAS PubMed Google Scholar

White, J. R. et al. AP-2alpha-mediated activation of E2F and EZH2 drives melanoma metastasis. Cancer Res. 81, 4455–4470. https://doi.org/10.1158/0008-5472.CAN-21-0772 (2021).

Article CAS PubMed PubMed Central Google Scholar

Zhang, S., Wan, H. & Zhang, X. LncRNA LHFPL3-AS1 contributes to tumorigenesis of melanoma stem cells via the miR-181a-5p/BCL2 pathway. Cell Death Dis. 11, 950. https://doi.org/10.1038/s41419-020-03141-1 (2020).

Article CAS PubMed PubMed Central Google Scholar

Tigyi, G. et al. Anti-cancer strategies targeting the autotaxin-lysophosphatidic acid receptor axis: is there a path forward?. Cancer Metastasis Rev. 40, 3–5. https://doi.org/10.1007/s10555-021-09955-5 (2021).

Article CAS PubMed PubMed Central Google Scholar

Huang, C., Sheng, Y., Jia, J. & Chen, L. Identification of melanoma biomarkers based on network modules by integrating the human signaling network with microarrays. J. Cancer Res. Ther. 10(Suppl), C114-124. https://doi.org/10.4103/0973-1482.145816 (2014).

Article CAS PubMed Google Scholar

Xu, X. et al. Gain of Chromosome 1q perturbs a competitive endogenous RNA network to promote melanoma metastasis. Cancer Res. 82, 3016–3031. https://doi.org/10.1158/0008-5472.CAN-22-0283 (2022).

Article CAS PubMed PubMed Central Google Scholar

Romano, S. et al. Immunomodulatory pathways regulate expression of a spliced FKBP51 isoform in lymphocytes of melanoma patients. Pigment Cell Melanoma Res. 28, 442–452. https://doi.org/10.1111/pcmr.12378 (2015).

Article CAS PubMed Google Scholar

Ke, J. et al. Melanoma migration is promoted by prion protein via Akt-hsp27 signaling axis. Biochem. Biophys. Res. Commun. 523, 375–381. https://doi.org/10.1016/j.bbrc.2019.12.042 (2020).

Article CAS PubMed Google Scholar

Xu, J. et al. LINC01063 functions as an oncogene in melanoma through regulation of miR-5194-mediated SOX12 expression. Melanoma Res. 32, 218–230. https://doi.org/10.1097/CMR.0000000000000803 (2022).

Article CAS PubMed Google Scholar

Takeda, K. et al. IFN-gamma is required for cytotoxic T cell-dependent cancer genome immunoediting. Nat. Commun. 8, 14607. https://doi.org/10.1038/ncomms14607 (2017).

Article ADS CAS PubMed PubMed Central Google Scholar

Hoekstra, M. E. et al. Long-distance modulation of bystander tumor cells by CD8(+) T cell-secreted IFNgamma. Nat. Cancer. 1, 291–301. https://doi.org/10.1038/s43018-020-0036-4 (2020).

Article CAS PubMed PubMed Central Google Scholar

Thibaut, R. et al. Bystander IFN-gamma activity promotes widespread and sustained cytokine signaling altering the tumor microenvironment. Nat. Cancer. 1, 302–314. https://doi.org/10.1038/s43018-020-0038-2 (2020).

Article CAS PubMed PubMed Central Google Scholar

Schaefer, E. J. et al. BCOR and BCORL1 mutations drive epigenetic reprogramming and oncogenic signaling by unlinking PRC1.1 from target genes. Blood Cancer Discov. 3, 116–135. https://doi.org/10.1158/2643-3230.BCD-21-0115 (2022).

Article PubMed PubMed Central Google Scholar

Meng, J. et al. Rb selectively inhibits innate IFN-beta production by enhancing deacetylation of IFN-beta promoter through HDAC1 and HDAC8. J. Autoimmun. 73, 42–53. https://doi.org/10.1016/j.jaut.2016.05.012 (2016).

Article CAS PubMed Google Scholar

Bald, T. et al. Immune cell-poor melanomas benefit from PD-1 blockade after targeted type I IFN activation. Cancer Discov. 4, 674–687. https://doi.org/10.1158/2159-8290.CD-13-0458 (2014).

Article CAS PubMed Google Scholar

Fragale, A. et al. Reversing vemurafenib-resistance in primary melanoma cells by combined romidepsin and type I IFN treatment through blocking of tumorigenic signals and induction of immunogenic effects. Int. J. Cancer. 153, 1080–1095. https://doi.org/10.1002/ijc.34602 (2023).

Article CAS PubMed Google Scholar

Razaghi, A., Durand-Dubief, M., Brusselaers, N. & Bjornstedt, M. Combining PD-1/PD-L1 blockade with type I interferon in cancer therapy. Front. Immunol. 14, 1249330. https://doi.org/10.3389/fimmu.2023.1249330 (2023).

Article CAS PubMed PubMed Central Google Scholar

Marzese, D. M., Huynh, J. L., Kawas, N. P. & Hoon, D. S. Multi-platform genome-wide analysis of melanoma progression to brain metastasis. Genom. Data. 2, 150–152. https://doi.org/10.1016/j.gdata.2014.06.007 (2014).

Article PubMed PubMed Central Google Scholar

Rabbie, R. et al. The mutational landscape of melanoma brain metastases presenting as the first visceral site of recurrence. Br. J. Cancer. 124, 156–160. https://doi.org/10.1038/s41416-020-01090-2 (2021).

Article CAS PubMed Google Scholar

Manca, A. et al. Mutational concordance between primary and metastatic melanoma: a next-generation sequencing approach. J. Transl. Med. 17, 289. https://doi.org/10.1186/s12967-019-2039-4 (2019).

Article CAS PubMed PubMed Central Google Scholar

Forschner, A. et al. Response and outcome of patients with melanoma skin metastases and immune checkpoint inhibition. Int. J. Cancer. https://doi.org/10.1002/ijc.35103 (2024).

Article PubMed Google Scholar

Download references

This study was supported by the Hungarian National Research, Development and Innovation Office (NKFIH), Hungary: K-135540 (JT), TKP21-EGA25 (JT) and N-OTKA 114460 (SK). This work was done under the auspices of a Memorandum of Understanding between the European Cancer Moonshot Center in Lund and the U.S. National Cancer Institute’s International Cancer Proteogenome Consortium (ICPC). ICPC encourages international cooperation among institutions and nations in proteogenomic cancer research in which proteogenomic datasets are made available to the public.

These authors contributed equally: Laura Vízkeleti and Orsolya Papp.

Department of Bioinformatics, Faculty of Medicine, Semmelweis University, 1094, Budapest, Hungary

Laura Vízkeleti, Szonja A. Kovács & Balázs Győrffy

Department of Pathology, Forensic and Insurance Medicine, Faculty of Medicine, Semmelweis University, Üllői Str. 93., 1091, Budapest, Hungary

Laura Vízkeleti, Orsolya Papp, Viktória Doma & József Tímár

Turbine Simulated Cell Technologies, Budapest, 1027, Hungary

Orsolya Papp

Department of Dermatology, Venerology and Dermato-Oncology, Faculty of Medicine, Semmelweis University, 1085, Budapest, Hungary

Viktória Doma & Sarolta Kárpáti

Clinical Protein Science & Imaging, Biomedical Centre, Department of Biomedical Engineering, Lund University, 223 63, Lund, Sweden

Jeovanis Gil & György Markó-Varga

Chemical Genomics Global Research Lab, Department of Biotechnology, College of Life Science and Biotechnology, Yonsei University, Seoul, 03722, South Korea

György Markó-Varga

1St Department of Surgery, Tokyo Medical University, Tokyo, 160-8582, Japan

György Markó-Varga

Doctoral School of Pathological Sciences, Semmelweis University, 1085, Budapest, Hungary

Szonja A. Kovács

National Laboratory for Drug Research and Development, 1117, Budapest, Hungary

Szonja A. Kovács

You can also search for this author in PubMed Google Scholar

The study was designed by and the concept was created by J.T. and L.V. Samples were collected by O.P., V.D. and L.V. The experimental work was carried out by O.P. and L.V. Data was analysed by L.V., O.P., G.M.V., J.G., S.A.K. and B.G. The manuscript was written and revised by L.V., O.P., J.T., G.M.V. and S.K. Figure 1 was prepared by G.M.V., Figs. 2, 3 were prepared by O.P. and Fig. 4 was prepared by B.G. All authors have read and agreed to the published version of the manuscript.

Correspondence to József Tímár.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Vízkeleti, L., Papp, O., Doma, V. et al. Identification of genetic fingerprint of type I interferon therapy in visceral metastases of melanoma. Sci Rep 14, 26540 (2024). https://doi.org/10.1038/s41598-024-77285-x

Download citation

Received: 17 April 2024

Accepted: 21 October 2024

Published: 03 November 2024

DOI: https://doi.org/10.1038/s41598-024-77285-x

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Engineering immunogens that select for specific mutations in HIV broadly neutralizing antibodies | Nature Communications

Atypical chemokine receptor 2 expression is directly regulated by hypoxia inducible factor-1 alpha in cancer cells under hypoxia | Scientific Reports

News

Identification of genetic fingerprint of type I interferon therapy in visceral metastases of melanoma | Scientific Reports