BRN2 is a non-canonical melanoma tumor-suppressor

Hamm, Michael; Sohier, Pierre; Petit, Valérie; Raymond, Jérémy H.; Delmas, Véronique; Le Coz, Madeleine; Gesbert, Franck; Kenny, Colin; Aktary, Zackie; Pouteaux, Marie; Rambow, Florian; Sarasin, Alain; Charoenchon, Nisamanee; Bellacosa, Alfonso; Sanchez-del-Campo, Luis; Mosteo, Laura; Lauss, Martin; Meijer, Dies; Steingrimsson, Eirikur; Jönsson, Göran B.; Cornell, Robert A.; Davidson, Irwin; Goding, Colin R.; Larue, Lionel

doi:10.1038/s41467-021-23973-5

Download PDF

Article
Open access
Published: 17 June 2021

BRN2 is a non-canonical melanoma tumor-suppressor

Michael Hamm^1,2,3^na1,
Pierre Sohier ORCID: orcid.org/0000-0002-2341-6848^1,2,3^na1,
Valérie Petit^1,2,3^na1,
Jérémy H. Raymond^1,2,3,
Véronique Delmas ORCID: orcid.org/0000-0001-7368-3664^1,2,3,
Madeleine Le Coz^1,2,3,
Franck Gesbert^1,2,3,
Colin Kenny⁴,
Zackie Aktary^1,2,3,
Marie Pouteaux^1,2,3,
Florian Rambow^1,2,3,
Alain Sarasin⁵,
Nisamanee Charoenchon^1,2,3,6,
Alfonso Bellacosa⁷,
Luis Sanchez-del-Campo ORCID: orcid.org/0000-0002-9537-7761⁸,
Laura Mosteo⁸,
Martin Lauss⁹,
Dies Meijer ORCID: orcid.org/0000-0002-8461-6341¹⁰,
Eirikur Steingrimsson¹¹,
Göran B. Jönsson⁹,
Robert A. Cornell⁴,
Irwin Davidson ORCID: orcid.org/0000-0001-5533-1171^4,12,
Colin R. Goding ORCID: orcid.org/0000-0002-1614-3909⁸ &
…
Lionel Larue ORCID: orcid.org/0000-0002-2010-6107^1,2,3

Nature Communications volume 12, Article number: 3707 (2021) Cite this article

5160 Accesses
9 Citations
12 Altmetric
Metrics details

Subjects

Abstract

While the major drivers of melanoma initiation, including activation of NRAS/BRAF and loss of PTEN or CDKN2A, have been identified, the role of key transcription factors that impose altered transcriptional states in response to deregulated signaling is not well understood. The POU domain transcription factor BRN2 is a key regulator of melanoma invasion, yet its role in melanoma initiation remains unknown. Here, in a Braf^V600E Pten^F/+ context, we show that BRN2 haplo-insufficiency promotes melanoma initiation and metastasis. However, metastatic colonization is less efficient in the absence of Brn2. Mechanistically, BRN2 directly induces PTEN expression and in consequence represses PI3K signaling. Moreover, MITF, a BRN2 target, represses PTEN transcription. Collectively, our results suggest that on a PTEN heterozygous background somatic deletion of one BRN2 allele and temporal regulation of the other allele elicits melanoma initiation and progression.

Enhanced BRAF engagement by NRAS mutants capable of promoting melanoma initiation

Article Open access 07 June 2022

BRN2 expression increases anoikis resistance in melanoma

Article Open access 06 July 2020

CDK12 is hyperactivated and a synthetic-lethal target in BRAF-mutated melanoma

Article Open access 29 October 2022

Introduction

Cancer initiation is triggered by the activation of oncogenic signaling combined with senescence bypass. Yet while many typical oncogenes and tumor suppressors that affect cancer initiation have been identified, cancer initiation is likely to be modulated by additional genetic events. Understanding how non-classical driver mutations may impact cancer initiation is a key issue that has been relatively underexplored. Melanoma, a highly aggressive skin cancer, arises through the acquisition of well-defined genetic and epigenetic modifications in oncogenes and tumor suppressors and represents an excellent model system to address this key question.

As a highly genetically unstable cancer type, the initiation of melanoma requires the induction of melanocyte proliferation, which is mediated by several major founder mutations, the most common of which are BRAF^V600E and NRAS^Q61K/R^1,2. However, activation of BRAF or NRAS is insufficient to promote melanoma initiation without senescence bypass mediated by additional founder mutations or expression changes of several genes including p16^INK4A, CTNNB1, PTEN, or MDM4^3,4,5,6,7.

The transcription factor BRN2, also known as POU3F2 and N-OCT3, plays a critical role in neurogenesis and drives proliferation in a range of cancer types with neural or neuroendocrine origins, including glioblastoma, neuroblastoma, small cell lung cancer, and neuroendocrine prostate cancer^8,9,10. In the melanocyte lineage, BRN2 is not detected in melanoblasts in vivo but is heterogeneously expressed in naevi and melanoma^11,12,13,14. In vitro studies have shown that BRN2 expression is induced by a range of melanoma-associated signaling pathways including activation of the mitogen-activated protein kinase (MAPK) pathway downstream from BRAF, the PI3K pathway, the LEF-β-catenin axis, as well as FGF, TNF-α, EDN3, and SCF signaling^14,15,16,17. Consistent with BRN2 being expressed in a predominantly mutually exclusive pattern with the Microphthalmia-associated transcription factor (MITF)¹³ that plays a crucial role in melanoma proliferation¹⁸, BRN2 is repressed by MITF via miR-211¹⁹. However, the relationship between MITF and BRN2 is complex. For example, BRN2 was recently found to be regulated by E2F1, a cell cycle-regulated transcription factor that is also a target for MITF, and both BRN2 and MITF are regulated by PAX3 and WNT/β-catenin^{15,16,20,21,22,23,24}. Indeed, both BRN2 and MITF can regulate expression of AXL^21,25, with BRN2 repressing AXL expression, thus enabling some cells in human melanoma to adopt an AXL^High, MITF^Low and BRN2^Low state²⁵. The activation of BRN2 expression in a specific subset of melanoma cells in response to all three major signaling pathways (MAPK, PI3K/PTEN, and β-catenin) linked to melanoma initiation (early proliferation and bypass/escape senescence) and progression (including late proliferation and metastatic dissemmination) suggests that Brn2 is likely to have a critical role in disease progression²⁶. Most notably, BRN2 has been associated with MITF^Low senescent or slow-cycling cells¹¹, and identified as a key regulator of melanoma invasion and anoikis in vitro^13,27,28 and in in vivo xenograft experiments^20,29,30. Mechanistically, the ability of BRN2 to promote invasion has been linked to its ability to control expression of PDE5A-mediated cell contractility, phosphorylation of myosin light chain 2, repression of MITF and PAX3, and cooperation with bi-allelic loss of CDKN2A^13,20,27,30. However, despite abundant information linking BRN2 to melanoma proliferation and invasiveness in vitro and in xenograft experiments, the impact of BRN2 on melanoma initiation and progression in vivo has never been assessed.

In this work, we show that BRN2 acts as a tumor suppressor during melanoma initiation and progression in a BRAF-PTEN context since BRN2 and MITF regulate positively and negatively the transcription of PTEN, respectively.

Results

BRN2 loss or low expression correlates with reduced survival and worse prognosis

Although BRN2 has been implicated in melanoma invasiveness, and its expression is highly regulated, whether and how it may contribute to melanoma initiation or incidence is not understood. To evaluate the prevalence of BRN2 loss in human skin cutaneous melanoma (SKCM), we retrieved copy-number alteration (CNA) data for BRN2 in SKCM metastases (stage IV) from the Cancer Genome Atlas (TCGA, https://cancergenome.nih.gov/). The BRN2 locus showed mono-allelic loss in 53% and bi-allelic loss in 2.7% of all patient samples (n = 367, Fig. 1A). Only a minority (n = 29 of 367, corresponding to 7.9%) of SKCM samples showed a copy-number gain (n = 27)/amplification (n = 2) for BRN2. We did not further analyze these samples since the expression of BRN2 was slightly increased but not statistically different (p = 0.26, test of Kruskall-Wallis with a Dunn correction) between tumors that gained and/or amplified this locus compared to the normal situation. We screened a panel of human melanoma cell lines available in our laboratory (n = 23) for deletions that affect the BRN2 locus by comparative genomic hybridization. The BRN2 locus showed mono-allelic loss in 48% (11 out 23) of the human melanoma cell lines and no bi-allelic loss, comparable to the TCGA-data (Supplementary Fig. 1A, Supplementary Data 1). Notably, BRN2 mRNA levels were significantly lower in SKCM metastases with bi-allelic BRN2 loss (Supplementary Fig. 1B). The mono- and bi-allelic loss of BRN2 was frequently associated with a large segmental deletion of the long arm of chromosome 6 (Chr.6q) in SKCM metastases and in our cell line panel (Fig. 1B, Supplementary Fig. 1C, Supplementary Data 2). From the TCGA, patients carrying the monoallelic loss of BRN2 in metastases displayed a trend to have a shorter overall survival than those with diploid status (Fig. 1C). These results were validated using an independent cohort of 108 regional metastases previously described³¹ (Fig. 1D). Moreover, we evaluated the number of BRN2 alleles in nevi and melanoma that arose from these nevi using publically available data³². It appears that 28% (5 out of 18) or 22% (4 out of 18) of melanomas presented either a mono-allelic loss or a gain of BRN2 respectively compared to nevi (Supplementary Fig. 1D). The situation is clearly complex, but we may conclude that BRN2 mono-allelic loss can occur during the early steps of melanomagenesis.

**Fig. 1: One *BRN2* allele is frequently lost in human melanoma and reduced BRN2 mRNA level correlates with reduced overall survival.**

We next assessed the correlation between BRN2 mRNA levels and overall patient survival to evaluate the effect of BRN2 mono-allelic loss on melanoma progression. We established “BRN2-high” and “BRN2-low” patient groups based on RNA-seq data available from the TCGA (BRN2 subgroups defined as BRN2-low (≤1 transcript per million reads [TPM]) and BRN2 expressed/high [>1 TPM]). Patients in the “BRN2-low” group displayed significantly shorter overall survival than those of the “BRN2-high” group (Fig. 1E). Overall, the BRN2 locus, frequently associated with a large segmental deletion, is lost (mono- and bi-allelic) in ≈60% of human SKCM metastases and correlates with significantly reduced overall survival. Since CCNC, ROS1 and ARID1B loci are distal to BRN2 on chromosome 6, and are known to be involved in melanomagenesis, we evaluated the overall survival of these patients according to the presence and the level of mRNA expression of the corresponding genes. We observed no significant difference between the presence or absence of these three genes or their expression (Supplementary Fig. 1E–L). Finally, we compared the loss of 6q with the mono-allelic loss (MAL) of BRN2 (Supplementary Fig. 1M) and found that 6q loss is associated with a worse prognostic than BRN2 mono-allelic loss. This result indicates, as suspected, that other gene(s) located on 6q are of importance in melanomagenesis. Taken together, these data indicate that in human melanoma BRN2 loss/low expression is associated with an adverse outcome for the patient.

Co-occurrence of BRN2 loss with mono-allelic loss of PTEN

We next determined whether BRN2 loss co-occurs with melanoma driver mutations by examining the TCGA CNA-data set and human melanoma cell-line panel. There was no significant correlation between BRAF or NRAS mutation and BRN2 loss (mono- or bi-allelic), neither in human melanoma samples nor the human cell-line panel (Supplementary Fig. 2A,B). We then searched for co-occurring CNAs of other known melanoma-associated genes and found that mono-allelic loss of BRN2 co-occurred with mono-allelic loss of PTEN in approximately 40% of the human melanoma samples in TCGA and in the human cell-line panel (Supplementary Fig. 2C,D). We next evaluated the concomitant BRN2 locus alterations, BRAF/NRAS mutations, and CDKN2A/PTEN alterations and found the most frequent genetic constellation that co-occurs with BRN2 loss in melanoma to be BRAF^V600X mutation together with mono-allelic PTEN loss (Supplementary Fig. 2E). Finally, we compared the overall survival of human patients with a loss of one allele of PTEN who also had a loss of BRN2 (monoallelic loss = MAL) with a loss of one allele of PTEN and no loss of BRN2 (BRN2-normal). In this context, patients with loss of BRN2 showed significantly lower overall survival than BRN2-normal patients (Supplementary Fig. 1N).

In conclusion, in human melanoma, the loss of BRN2 is preferentially associated with BRAF mutation together with PTEN loss.

Loss of BRN2 drives melanomagenesis in vivo

These data suggesting that the loss of BRN2 might be of importance in melanoma prompted us to evaluate the potential causal role of BRN2 in melanomagenesis in vivo by examining whether heterozygous (het) or homozygous (hom) loss of Brn2 affects melanoma initiation and/or progression in a mouse model. Note however, that while genetic loss of BRN2 might be important, the complex regulation of BRN2 expression driven by melanoma-associated signaling pathways might also play a major role, especially given that melanoma cells within a single tumor can exhibit both high and very low BRN2 expression^13,20. We therefore developed an inducible genetically engineered mouse model system for generating Brn2-deficient melanoma driven by the most common alterations in human SKCM (Braf^V600E and Pten loss). Specifically, we used Tyr::Cre^ERt2/°-Lar; Braf^V600E/+ (called Braf from hereon) and Tyr::Cre^ERt2/°-Lar; Braf^V600E/+; Pten^F/+ (called Braf-Pten from hereon) mice carrying a tamoxifen-inducible Cre-recombinase under the control of the tyrosinase promoter^33,34,35. This model system allows melanocyte lineage-specific induction of a BRAF^V600E mutation and mono-allelic deletion of Pten for Braf-Pten mice. Cre-mediated defloxing leads to activation of the Braf^V600E oncogene, inducing nevus and spontaneous melanoma formation in Braf mice, reproducing many of the cardinal histological and molecular features of human melanoma³⁶. Bi-allelic and mono-allelic loss of PTEN reduces tumor latency in Braf^V600E- and NRAS^Q61K-driven mouse melanoma models^3,37.

Using these models, we studied the effect of Brn2 insufficiency (het and hom) on in vivo melanomagenesis by introducing the floxed Brn2 locus into the genome by appropriate crossings (Supplementary Fig. 3A)³⁸. Specifically, we generated the mouse lines Tyr::Cre^ERT2/°; Braf^V600E/+; Brn2^+/+ (Braf-Brn2-WT), Tyr::Cre^ERT2/°; Braf^V600E/+; Brn2^F/+ (Braf-Brn2-het), Tyr::Cre^ERT2/°; Braf^V600E/+; Brn2^F/F (Braf-Brn2-hom), Tyr::Cre^ERt2/°; Braf^V600E/+; Pten^F/+; Brn2^+/+ (Braf-Pten-Brn2-WT), Tyr::Cre^ERt2/°; Braf^V600E/+; Pten^F/+; Brn2^F/+ (Braf-Pten-Brn2-het) and Tyr::Cre^ERt2/°; Braf^V600E/+; Pten^F/+; Brn2^F/F (Braf-Pten-Brn2-hom). Cre-mediated defloxing of Braf, Pten, and Brn2 loci was induced by topical application of tamoxifen during the first 3 days after birth (Supplementary Fig. 3B). All mice were monitored for the appearance and growth rate of the first tumor, as well as for the number of tumors/mouse. Note that the ability to generate either homo- or heterozygous Brn2 KOs will mimic not only mono or biallelic loss in humans, but also reflect the variable levels of BRN2 observed within human tumors^13,20. In the absence of PTEN (Pten^F/F) on a Braf^V600E background, the appearance of the tumors was too rapid to observe any difference between Brn2 + /+, Brn2 F/ + or Brn2 F/F mice.

Braf-Brn2-WT/het/hom mice showed no differences in the appearance of the first tumor, number of tumors/mouse or the tumor growth rate from those of Braf-WT mice (Fig. 2, Supplementary Fig. 3C). However, compared to Braf-Pten-Brn2-WT, Braf-Pten-Brn2-het/hom mice significantly increased the number of tumors/mouse and the tumor growth rate, but not the timing of the appearance of the first tumor (Fig. 2, Supplementary Fig. 3D). We verified that Brn2, Braf, and Pten were correctly defloxed in the resulting melanomas (Supplementary Fig. 3E,F). In summary, our data show that Brn2 acts as a tumor suppressor in vivo. As it has been shown in humans and mice, induction of early proliferation induced by the presence of the BRAF^V600E mutation leads to senescence, but it can be bypassed when the level of Pten is reduced^3,37,39,40. It is important to note that in the absence of Pten in the physiological context, proliferation of melanoblasts and melanocytes is not induced as the reduction/lack of Pten promotes proliferation once the cells are transformed^3,41. On a Braf-Pten context the loss/reduction of Brn2 appears to induce melanoma initiation after promoting proliferation and bypass/escape of senescence, and then allows the tumor growth as was previously showed in vivo^3,37.

**Fig. 2: Brn2 loss potentiates melanomagenesis in Braf-Pten mice.**

Reduction of BRN2 levels increases proliferation in vivo and in vitro

The effect of Brn2 loss on tumor growth prompted us to investigate whether Brn2 loss increases intra-tumor proliferation. Staining sections for Ki-67, a marker of cycling cells, revealed that melanomas from Braf-Pten-Brn2-het/hom mice displayed a significantly higher number of Ki-67⁺ cells than Braf-Pten-Brn2-WT melanomas (Fig. 3A,B). To confirm this result, we injected Braf-Pten-Brn2 mice with bromodeoxyuridine (BrdU) two hours prior to euthanization, to determine whether melanoma cells were slow or fast-dividing. Braf-Pten-Brn2 melanomas had a significantly higher number of BrdU⁺ cells when Brn2 was heterozygous or homozygous (Fig. 3C,D). These results indicated that heterozygous loss of Pten combined with heterozygous/homozygous loss of Brn2 promotes melanoma proliferation in vivo.

**Fig. 3: BRN2-het/hom induces proliferation in vitro and in vivo.**

We next assessed whether Brn2 knockdown favors proliferation in vitro and whether this mechanism is conserved (i) between human and mouse and (ii) between transformed and non-transformed cells using Dauv-1 human melanoma cell line, and the Melan-a mouse melanocyte cell line. These cell lines express both Pten and Brn2 mRNA and protein (Supplementary Table 1). Dauv-1 cells carry a BRAF^V600E/+ mutation identical to that used in the mouse melanoma model system, and Melan-a cells are WT for Braf. siRNA-mediated knockdown of Brn2 significantly increased cell number 72 h after transfection of both cell lines (Fig. 3E–G). Brn2 knockdown, assessed by western blot, led to an increase in cyclin D1 protein levels in both cell lines, but did not alter cyclin D1 mRNA levels, suggesting a regulation of cyclin D1 at the protein level (Fig. 3F–H). Overall, the reduction of the Brn2 protein induces cell proliferation of the melanocytic lineage in vivo and in vitro.

Colony formation assays in vitro are classically used to show the importance of a gene in tumorigenesis, and indicate that a single cell may survive in vitro and proliferate to form a colony. Previous work has demonstrated that a reduction in Brn2 levels in melanoma cells has no effect on colony formation^42,43. However, to test this in our model, mouse melanoma cell lines were established and characterized from the different independent Braf-Pten-Brn2 C57BL/6 J mouse melanoma; the m50 and m6 cell lines were Brn2-WT, m59, and m36 were Brn2-het, and m82 and m8 were Brn2-hom (Supplementary Fig. 4A–F). The presence or absence of Brn2 did not decrease the ability of these melanoma cell lines to grow in syngeneic mice (Supplementary Table 2). In other words, it appears that the absence of Brn2 in these melanoma cells does not affect the implantation of the cells on the body wall, the proliferation after their transformation or the induction of angiogenesis in an immunocompetent environment. We evaluated the capacity of the m50 (Brn2 WT), m59 (het), and m82 (hom) cells to generate colonies and observed, in agreement with previous observations obtained using BRN2 depletion, that the three cell lines have similar abilities to form colonies (Supplementary Fig. 4G). Moreover, re-expression of Brn2 using lentivirus infection in two independent m8 and m82 Braf-Pten-Brn2-hom mouse melanoma cell lines, also indicated that BRN2 does not affect their capacity to form colonies under these conditions (Supplementary Fig. 4H, I). In conclusion, as shown in human cell lines, the number of colony forming units is independent of the presence or absence of Brn2 and consequently this assay is uninformative regarding the role of Brn2 in melanomagenesis.

We also evaluated the activity of BRAF^V600 (PLX4720), MEK (Binimetinib), and PI3K (LY294002) inhibitors on the capacities of m50, m59 and m82 mouse melanoma cell lines to form CFU and determined the IC50 of these drugs (Supplementary Fig. 4J-O). Consistent with observations suggesting that BRN2 may suppress cell death⁴³, Brn2-het/hom cells are more sensitive than WT cells to these three drugs. We also evaluated the cooperativity of PLX4720 and LY294002 using various concentrations of each drug, but we did not observe any cooperation/synergy of these two drugs.

Mono-allelic loss of Brn2 induces melanoma metastasis

We next evaluated the effects of Brn2 on metastasis formation in vivo. Since human SKCM is known to spread to proximal lymph nodes (LNs), we assessed the presence of pigmented cells in the inguinal LNs of tumor-bearing Braf-Pten-Brn2 mice. Specifically, we estimated the volume of the various metastases present in LNs and the number of pigmented areas after haematoxylin & eosin (HE) staining of LN sections (Fig. 4A–C). All Braf-Pten-Brn2 mice, irrespective of BRN2 status, showed the presence of pigmented cells in both inguinal LNs. However, the LN volume of Braf-Pten-Brn2-het mice was significantly higher than that of Braf-Pten-Brn2-WT and Braf-Pten-Brn2-hom (Fig. 4A, B). Similarly, Braf-Pten-Brn2-het LNs showed a higher number of pigmented areas per mm² than Braf-Pten-Brn2-WT/hom mice (Fig. 4C).

To verify that the pigmented cells in the lymph nodes did not arise from cells in which the Cre recombinase had not worked efficiently, we tested whether these pigmented cells were properly defloxed for Brn2. The targeted Brn2-flox allele, used in our mouse model, has an eGFP-cassette inserted downstream of the floxed Brn2 locus (Fig. 4D). Thus the production of eGFP occurs once the upstream Brn2 gene is defloxed. Consistent with correct defloxing of Brn2, pigmented areas of LNs expressed GFP in Braf-Pten-Brn2-het/hom mice, but not Braf-Pten-Brn2-WT mice (Fig. 4E). The pigmented cells present in Braf-Pten-Brn2-WT and Braf-Pten-Brn2-het LNs also expressed Sox10, a melanocytic marker (Fig. 4E), that co-localized with eGFP-expression in Braf-Pten-Brn2-het mice confirming the melanocytic origin of the pigmented cells observed.

To get a better understanding of melanomagenesis in the various Brn2-WT/het/hom situations, we performed a transcriptomic analysis of the various Brn2 tumors and cell lines (Supplementary Data 3 and 4). The ontology enrichment analysis indicated that the Braf-Pten-Brn2-het tumors were more inflamed than the Braf-Pten-Brn2-WT tumors with increased neutrophil-associated gene expression (Supplementary Fig. 5A). It also suggests that the extracellular-matrix was actively remodeled and that these Braf-Pten-Brn2-het tumors were more subject to angiogenesis (Supplementary Fig. 5A). Pathway enrichment analysis supported these results as inflammatory gene expression signatures are enriched from both the WikiPathways and KEGG databases (Supplementary Fig. 5B,C). Significantly, the PI3K/AKT pathway was enriched in the Braf-Pten-Brn2-het tumors compared to the Braf-Pten-Brn2-WT in both WikiPathways and KEGG databases (Supplementary Fig. 5B,C), suggesting that AKT could be more phosphorylated in the Braf-Pten-Brn2-het tumors. This assumption was verified by Western-blot analysis where phosphorylation of AKT on S473, a surrogate of the AKT activity, and the phosphorylation of S6 on S235/236 were significantly increased in Braf-Pten-Brn2-het/hom tumors compared to Braf-Pten-Brn2-WT tumors (Supplementary Fig. 5D).

Gene Set Enrichment Analysis (GSEA) using the melanoma invasive signature from Verfaillie⁴⁴ indicated that Braf-Pten-Brn2-het tumors and cell lines were more invasive than Braf-Pten-Brn2-hom tumors and cell lines (Fig. 4F). Moreover, from the GO, WikiPathways and KEGG 2019 analyses of the mouse melanoma tumors, the immune system (cytokine, neutrophil, macrophage) and angiogenesis are induced in Braf-Pten-Brn2-het tumors compared to Braf-Pten-Brn2-WT tumors (Supplementary Fig. 5A–C). This information suggests that Braf-Pten-Brn2-het cells have more potential to metastasize than Braf-Pten-Brn2-WT and Braf-Pten-Brn2-hom cells. In this respect, we tested the capacity of Braf-Pten-Brn2-hom (m82) mouse melanoma cell lines re-expressing or not Brn2 (m82 and m82 + Brn2) to invade matrigel in 3D. In the presence of ectopic Brn2, the m82 Brn2 KO melanoma cell lines have more ability to invade (Fig. 4G and Supplementary Fig. 4I). Since AXL, a receptor tyrosine kinase (RTK), is associated with melanoma metastasis⁴⁵, we also evaluated the level of Axl in mouse melanoma tumors and cells. Braf-Pten-Brn2-het cells produce slightly more Axl mRNA than Brn2-WT/hom (Fig. 4H). This upregulation does not affect genes that are in Cis of Axl suggesting a specific regulation of Axl in a Braf-Pten-Brn2-het context (Supplementary Fig. 5E–G). Moreover, when the levels of both BRN2 and MITF are reduced in human melanoma cell lines, the level of AXL mRNA is induced (Supplementary Fig. 6D,H,L,P,T). More precisely the level of AXL is slightly, but significantly, higher in Gerlach and DAUV-1 cell lines in which the level of AXL is already very high, and is higher in SK28 and 501Mel cell lines in which the level of AXL is much lower than in Gerlach and DAUV-1 cell lines.

In conclusion, compared with Braf-Pten-Brn2-WT mice, melanoma initiation is promoted in both Braf-Pten-Brn2-het/hom mice, but metastasis is promoted only in Braf-Pten-Brn2-het mice. These observations are consistent with the fact that on a Brn2-het/hom background melanoma initiation (proliferation and bypass/escape of senescence) is induced and melanoma invasion and/or survival is inhibited on a Brn2-hom background.

BRN2 binds to the PTEN promoter and BRN2 loss leads to the reduction of PTEN transcription

The PI3K-AKT pathway is induced in melanoma and its induction abrogates BRAF^V600E-induced senescence^46,47. The loss of Pten, a suppressor of the PI3K pathway, induces melanoma initiation and proliferation in vivo^3,37. Since our Braf-Pten mouse melanoma model retained one functional allele of Pten, we hypothesized that Brn2 loss would induce less expression from the WT Pten allele, leading to the increased PI3K-AKT signaling observed (Supplementary Fig. 5D) and consequent melanoma initiation and proliferation. We therefore evaluated the number of Pten positive cells in the various Braf-Pten tumors by immunohistochemistry and found that Braf-Pten-Brn2-het/hom tumors showed fewer Pten-pos cells than Braf-Pten-Brn2-WT tumors (Fig. 5A). This result was verified by western blotting of Braf-Pten tumor samples. The reduction of Brn2 correlates with the reduction of Pten in Braf-Pten-Brn2-het tumors compared to the Braf-Pten-Brn2-WT tumors (Fig. 5B). In accordance with the reduced protein levels, the mRNA levels of Brn2 and Pten were significantly lower in Braf-Pten-Brn2-het/hom tumors than in Braf-Pten-Brn2-WT tumors, suggesting regulation of Pten at the transcriptional level (Fig. 5C).

**Fig. 5: Brn2 binds to the Pten promoter and Brn2 loss leads to Pten transcription reduction.**

Next, we evaluated the mechanism of Pten repression mediated by reduced levels of Brn2 in the human Dauv-1, Gerlach, SK28, and 501Mel melanoma cell lines and in the non-transformed mouse Melan-a cell line. siRNA-mediated BRN2 knockdown led to significantly reduced PTEN protein and mRNA levels in cell lines having high BRN2 expression (Fig. 5D, E and Supplementary Fig. 6). In cells with lower level of BRN2 (501Mel), the reduction of PTEN was not observed.

We examined the PTEN promoter for BRN2 binding sites conserved between humans and mice to determine whether BRN2 acts directly on PTEN and detected two potential Brn2 binding sites (BS1 and BS2) at the positions—2049 and +761 (numbering relative to TSS). Chromatin-immunoprecipitation (ChIP) for BRN2, performed on Dauv-1 melanoma cell extracts, followed by qPCR, revealed quantitative BRN2 binding to both binding sites, comparable to BRN2 binding on the PAX3 promoter (Fig. 5F–H). IgG was used as a negative control.

To evaluate Pten promoter activity in response to BRN2, we cloned a 3.2-kb-human Pten promoter fragment containing the two BRN2 binding sites upstream the luciferase gene to generate the hsPten::Luciferase reporter construct (Fig. 5H). The analysis of human PTEN promoter activity showed an intrinsic activity of hsPten::Luciferase in human melanoma cells (Supplementary Fig. 7A, B). A consistent repression of PTEN transcription was observed when a smart-pool of siBRN2 was added. In contrast, PTEN transcription was activated following co-transfection of a BRN2 expression vector compared to empty vector in Dauv-1 and SK28 cell lines in luciferase assay (Fig. 5I–L). We performed similar experiments with the mouse PTEN promoter, which was cloned upstream of the luciferase construct (Supplementary Fig. 7C–F); Brn2 activated Pten transcription in three mouse melanoma cell lines (m82, m59, and m50) (Fig. 5M, N). In conclusion, BRN2 directly induces PTEN transcription.

MITF binds the PTEN locus and represses PTEN transcription

The MITF gene encodes a key transcription factor that plays a major role in melanocyte and melanoma biology¹⁸. Several studies have reported that MITF transcription is directly repressed by BRN2 whereas a reduction in BRN2 levels leads to increased MITF levels^13,27,48, while another study showed that BRN2 induces MITF⁴⁹. Likely both results are correct due to the versatile role of BRN2 as a transcription regulator whose activity may be dependent on context including genetic status and/or the environment^21,26,27.

Knowing the importance of MITF in the melanocyte lineage, we analyzed the consequences of modulation of MITF expression on PTEN. Using siRNA-mediated MITF knockdown in three human melanoma cell lines expressing high levels of MITF (501mel, SK28, HBL) led to a significant increase of PTEN protein and mRNA levels (Fig. 6A, B). In cells expressing lower levels of MITF, such as Gerlach and Dauv-1, we did not observe an increase of PTEN (Supplementary Fig. 6C, G). Next, to determine if MITF directly regulates PTEN expression through proximal enhancers we performed Cleavage Under Targets and Release Using Nuclease (CUT&RUN) with antibodies to MITF, as previously described,⁵⁰ and to H3K27Ac, revealing active chromatin, and against H3K4me3, revealing active and poised promoters, in SK28 cell lines that were wild type for MITF (MITF-WT) or had loss of function mutations in all alleles of it (ΔMITF) (Fig. 6C)⁵⁰. An MITF peak is present within the gene body of PTEN (intron 4), and a second one is 114 kb downstream of the PTEN transcriptional start site (TSS) (i.e., +114 kb). Both MITF peaks were centered on an M-box motif (5ʹ-TCATGTG-3ʹ). At both intron-4 and +114 kb MITF peaks, the H3K27Ac signal was unchanged in ΔMITF mutant cell lines. However, six distal enhancers (painted light blue) exhibited a twofold greater H3K27ac signal in MITF mutant cell lines compared to wild-type cell lines (+140 kb, +210 kb, +230 kb, +253 kb, +301 kb and +317 kb), suggesting increased transcription of PTEN in the former. We confirmed the +140 kb MITF peak after performing a quantitative ChIP experiments on 501Mel cells expressing MITF-HA (Fig. 6D, E), using Tyrosinase (TYR) and PRM1 as positive and negative controls respectively. Overall, reduction of MITF results in an increase in PTEN mRNA expression. As such, it is plausible that Brn2 induces Pten transcription through two different, but concurrent mechanisms: (i) directly through BRN2 binding to the PTEN promoter to induce its transcription and (ii) indirectly via BRN2 modulating (repressing/inducing depending of the situation) MITF expression, with MITF binding to the 3ʹ end of PTEN to inhibit its transcription. In the absence of BRN2, these mechanisms are disrupted and PTEN transcription is downregulated in our conditions. In conclusion, MITF represses directly PTEN transcription.

**Fig. 6: MITF binds downstream of *PTEN* gene, MITF loss induces its transcription, and enhancers flanking *PTEN* are activated in MITF-depleted cells.**

Overall, our results are consistent with a model in which reduction of BRN2 reduces PTEN transcription in vitro and in vivo, thus ramping up PI3K signaling and inducing both the initiation of melanoma and the formation of metastases.

Discussion

A well-established principle of cancer biology is that tumors are initiated by a combination of oncogene activation together with loss of tumor suppressor expression or activity. In melanoma, key oncogenic drivers, such as BRAF and NRAS, have been well defined. Loss of P16 or PTEN tumor suppressor activity is required to bypass oncogene-induced senescence and permit melanoma initiation. However, while inactivation of tumor suppressors by mutation has been extensively studied, it is less well understood how their activity may be modulated by changes in their mRNA expression mediated by key melanoma-associated transcription factors. At this point, it is important to note that factors involved in melanoma initiation will not automatically reflect/translate into overall survival. Here, we identify BRN2, a key transcription factor lying downstream of three melanoma-associated signaling pathways (WNT/β-catenin, MAPK, and PI3K), as a tumor suppressor that functions to regulate PTEN expression. Thus, monoallelic loss of Brn2 promotes melanoma initiation in a Braf^V600E/Pten^F/+ background where mono-allelic loss of Pten sensitizes cells to loss of Brn2.

Previous work has primarily linked BRN2 to melanoma migration and invasion in vitro and in xenograft experiments^13,20,42,51, but its role during melanoma initiation and proliferation in vivo and in normal melanocytes had not been determined. We report that, consistent with BRN2 playing a key role as a tumor suppressor in melanomagenesis, its locus is frequently lost in human skin cutaneous melanoma (SKCM) metastases, independently of their NRAS or BRAF status, and that BRN2 status contributes to overall patient survival. Significantly, the overall survival of patients with a mono-allelic loss of PTEN is higher when the BRN2 locus is intact.

Although these observations are consistent with BRN2 affecting human melanoma initiation and progression, loss of the BRN2 locus is frequently associated with large segmental deletions that affect the long arm of chromosome 6 (6q) as it was observed in 14 out of 53, and 9 out of 32 out of primary melanoma and in 17 out of 21 melanoma cell lines^52,53,54, and confirmed in this study on a total of 205 out of 338 melanomas, and 11 out of 25 melanoma cell lines. It has already been shown that the loss of 6q was associated with a worse prognosis⁵². According to our observations, the loss of 6q is more detrimental for the overall survival than the focal loss of the BRN2 locus (Fig. 1 and Supplementary Fig. 1). Moreover, in vitro studies have shown that several genes are linked to melanomagenesis in the co-deleted region, including ARID1B, CCNC, and ROS1, although none have yet been shown to be functionally important in melanoma in vivo^{48,55,56,57,58,59,60,61}. Similarly, we analysed the SKCM TCGA study to evaluate the overall survival comparing the mono allelic loss and the diploid state for CCNC, ROS1 and ARID1B (Fig. S1). There is no statistical significance for CCNC (p = 0.086), ROS1 (p = 0.27), or ARID1B (p = 0.66).

Thus, it might be argued that any of the genes located in this frequently deleted region may be acting to modulate melanoma initiation or progression. However, our functional mouse molecular genetics models show conclusively that the heterozygous or homozygous loss of Brn2 promotes melanoma initiation and initial growth. Similarly, our transcriptomic analysis showed that the mRNA levels of the seven genes in the region co-deleted in humans (Arid1b, Mchr2, Ccnc, Cdk19, Dll1, Ros1, and Crybg1/Aim1) were not affected in the mouse tumors (Braf-Pten-Brn2-WT, Braf-Pten-Brn2-het and Braf-Pten-Brn2-hom). Collectively, these results strongly suggest that the reduction/loss of Brn2 levels is a critical event that cooperates with heterozygous Pten to promote the initiation and growth of melanoma, independently of the co-deleted genes in melanoma. Finally, independent initiation events are promoted when the level of Brn2 is lower than normal -heterozygous or homozygous- since the number of independent melanoma is higher in Braf-Pten-Brn2-het/hom mice than in Braf-Pten-Brn2-WT mice (Fig. 2B). Our observations are therefore consistent with BRN2 acting as a tumor suppressor in melanoma, and are in full agreement with the predominantly mutually exclusive pattern of BRN2 and Ki-67 in situ staining of invasive melanoma²⁰.

The presence of Braf^V600E promotes proliferation prior to inducing senescence and the loss of Pten results in senescence bypass^3,34,37. As such, we believe that the increased proliferation and tumor-initiation frequency observed in our Braf^V600E/Pten^F/+ model arising as a consequence of the reduction of Brn2, is likely to occur as a consequence of the ability of Brn2 to activate Pten expression and suppress PI3K signaling either directly or potentially indirectly via Mitf repression during early melanomagenesis. In other words, inactivation of Brn2 or a reduction in its expression would lead to low expression of the remaining Pten allele and as a consequence increase the probability of senescence bypass. This explanation fits with the fact that in the absence of Pten on a Braf^V600E background, the appearance of the tumors is too rapid to observe any difference between Brn2 + /+, Brn2 F/ + or Brn2 F/F. According to our model the transcription regulation by Brn2 and Mitf would not affect the level of Pten since it is already lost. Moreover, in a context in which Pten is diploid (WT) and Brn2 is reduced (het or hom), the downregulation of Pten would be limited and not sufficient to act efficiently as a tumor-suppressor.

Although it might be argued that the increase in visible tumor number in a Braf-Pten-Brn2-het and -hom might be a consequence of the increased proliferation caused by reduction/loss of Brn2, as evidenced by the increased proportion of Ki-67 positive cells within tumors, we feel this is unlikely. For welfare reasons the Braf-Pten-Brn2-WT mice were euthanized with tumors (total volume 2 cm³) after 4 weeks with about 8 tumors per mouse; by contrast the Braf-Pten-Brn2-het/hom mice were euthanized after 1.3 weeks with about 17 tumors per mouse of a similar size to the WT. Since tumor size is similar in the WT and Brn2 mutants, this indicates that the total number of cells in the WT and mutant tumors is similar and have undergone a similar number of cell divisions (though this occurred in a shorter time in the mutant). Since the WT and mutants have undergone a similar number of cell divisions, if proliferation were responsible for any increase in number of visible tumors then at 4 weeks the numbers of tumors in the WT should be the same as in the mutant at 1.3 weeks, whereas in fact, the tumor numbers in the WT were around 50% of those in the Brn2 mutants. Although we do not want to rule out a contribution of proliferation, the more likely explanation is that the reduction of BRN2 promotes the bypass of senescence (by reducing the level of Pten) or/and promotes survival of proliferating melanoma cells at the early stages of initiation.

Our in vivo data therefore reveal that melanocyte-specific Brn2 reduction in Braf-Pten mice promotes the initiation and progression of melanoma. Melanoma initiation is promoted after proliferation is induced through various proteins including Mitf and senescence bypassed in this case through reduction of the level of Pten. Melanoma progression is induced by promoting invasion when the level of Brn2 is intermediate after inducing Axl and modulating the immune system. In the future, we will have to evaluate the kinase activity of Axl in this context, and the consequences of its inhibition in cellulo and in vivo to understand the Mitf/Brn2/Axl ménage à trois. Altogether, we establish Brn2 as a central tumor suppressor, acting at different steps of melanomagenesis, and complement numerous other studies showing the effect of Brn2 on invasion^{13,14,15,20,29,51}.

Brn2 heterozygous mice were more prone to form LN metastases efficiently than mice that were Brn2 WT when Pten was already heterozygous. The efficiency of the metastasis process depends on the status of Brn2 as the number and size of each micro-metastasis was greater in Brn2 heterozygous than Brn2 wild-type melanoma. This is important as mono-allelic loss of BRN2 in human, corresponding to BRN2 heterozygous melanoma, occurs in 53% of human melanoma. In Brn2 wild-type melanoma, the change of the level/activity of Brn2 is possible but the amount of protein found in Brn2 wild-type melanoma absorbs better transient Brn2 depletion than in Brn2 heterozygous melanoma.

Some residual Brn2 activity might be required for efficient melanoma progression. On one hand, Braf-Pten-Brn2-het and Braf-Pten-Brn2-hom melanomas proliferate faster, and on the other hand, according to our results (Fig. 4F,G) Braf-Pten-Brn2-hom melanomas have a reduced ability to invade compared to Braf-Pten-Brn2-het melanomas. Moreover, the lack of BRN2 reduces migration and increases the rate of apoptosis and/or anoikis^20,28,43. In a more speculative way, one may think that optimal melanoma progression may be associated with a series of proliferative and invasive phases. In the absence of BRN2, cells are “fixed” in one stage and cannot switch from “proliferative” to “invasive” states. These cells are mainly proliferative and poorly invasive as we observed in Braf-Pten-Brn2-hom situation. In the presence of Brn2 as WT [two alleles] or heterozygous [one allele], the corresponding mRNA and protein concentrations can be positively and negatively modulated by external/internal factors. The modulation of the level of Brn2 is more sensitive on a heterozygous background than on a WT background.

Altogether, one may understand that melanoma grows faster and forms melanoma when Brn2 expression is not too high (proliferation handicap) but at the same time not too low (migration/invasion/survival handicap). Indeed, on the one hand, Brn2-het and Brn2-hom melanomas proliferate faster, and on the other hand, Brn2-hom melanomas are handicapped to invade (Fig. 4F,G), to migrate²⁰ and to die by anoikis²⁸.

The formation of metastases is a multistep process in which cells proliferating in the primary tumor, and surviving the metastatic process, must undergo a switch to an invasive phenotype prior to a switch to a proliferative phenotype on site. Since the switch from proliferation to invasion, and back, has been associated with the activity of MITF, it is possible that for efficient metastatic colonization cells must be able to modulate MITF expression via BRN2. In this respect, it is especially important to note that BRN2 has been identified as a key regulator of MITF^13,49. Consistent with the observation that MITF and BRN2 are frequently observed in mutually exclusive populations in melanoma, and that BRN2 may act in vivo as an MITF repressor¹³, we have observed that in a non-tumoral context, the specific knock-out of Brn2 in vivo in melanocytes increases the level of Mitf (publication in preparation). Importantly, during progression of mouse Braf-Pten melanoma Mitf levels are modulated. During the initial phase of growth, melanoma cells are pigmented, indicative of Mitf activity, but later in situ they lose pigmentation and ability to produce Mitf⁶². Although the reduction of Brn2 in Braf-Pten-Brn2 primary melanoma is not sufficient to re-induce Mitf mRNA and pigmentation in all cells of these primary melanomas, the Braf-Pten melanoma cells that formed LN metastases were pigmented and re-expressed Sox10, a key transcription activator of Mitf. Thus, during progression of Braf-Pten melanomas Mitf is produced during initial growth and subsequently repressed during the second phase, before being re-expressed in LN metastases. It seems therefore likely that one role of the residual BRN2 in the heterozygotes may be to facilitate the modulation of MITF expression during metastatic spread.

It has been shown that BRAF and PI3K induce BRN2^15,16. In consequence, it was expected that the level of BRN2 would decrease in the presence of such inhibitors. In addition, in the presence or absence of BRN2, melanoma cells are either more resistant or more sensitive to BRAF inhibitors, respectively^28,43. This sensitivity would be associated with the function of BRN2 in DNA repair⁴³.

Although here we have focused on the role of BRN2 in melanoma, BRN2 is also expressed in a number of other cancer types including small cell lung cancer, neuroblastoma, glioblastoma and neuroendocrine prostate cancer^8,9,10. While BRN2-mediated regulation of MITF is not likely to be important for non-melanoma cancers that do not express MITF, the ability of BRN2 to modulate PTEN expression uncovered here may play an equally important role in promoting the initiation and progression of these cancer types. In this respect the inducible knockout mice described here may represent an important tool to examine the role of BRN2 in non-melanoma cancers.

In conclusion, our results identify Brn2 as a key tumor suppressor through its ability to modulate Pten expression that, given the high prevalence of monoallelic mutations, is likely to play a key role in initiation of human melanoma and likely other BRN2-expressing cancer types. Since BRN2 expression is activated by PI3K signaling via PAX3¹⁶, its ability to suppress PI3K signaling by increasing PTEN expression may also provide cells with a negative feedback loop to control the PI3K pathway. Moreover activation of BRN2 by MAPK signaling downstream BRAF¹⁵ as well as WNT/β-catenin signaling¹⁴, may also permit coordination between these pathways and the PTEN/PI3K axis. Finally, given the importance of BRN2 in melanomagenesis identified here as well as its frequent heterozygosity, it may be important to further explore whether tumors with low BRN2 expression may be more susceptible to PI3K pathway inhibition, as it was shown for MAPK inhibitors. In this respect, this mouse model of BRN2-deficient melanoma could be useful for the pre-clinical testing of inhibitors for clinical development especially since it has been shown that BRN2 is involved in DNA repair⁴³.

Methods

TCGA data mining

All TCGA data sets for somatic mutations, copy number alterations (CNAs), RNA levels, and clinical data for skin cutaneous melanoma and other cancers were retrieved from http://www.cbioportal.org on August, 2019. CNAs were calculated using the GISTIC algorithm. Samples with GISTIC copy-number values of “−1” were considered to have mono-allelic loss and those with GISTIC copy number values of “−2” to have bi-allelic loss. GISTIC copy number values > ”+1” were considered as amplification. The TCGA datasets used were the: CNA-data set (n = 367), Seq-Data set (n = 473), Clark level data set (n = 461), and Breslow index data set (n = 316). mRNA levels were calculated from RNA sequencing read counts using RNA-Seq V2 RSEM and normalized to transcripts per million reads (TPM).

Copy number analysis

Copy number data for the regional metastatic cohort used in Fig. 1D was obtained from a previous study⁶³. Briefly, DNA sequencing data of 1,500 cancer genes were used to generate copy number data. Copy number log ratios of sequenced exons were generated from bam files of tumor–normal pairs using CONTRA 2.03⁶⁴, with default parameters. Exons with insufficient coverage in the normal sample were removed. Copy number data were segmented using GLAD⁶⁵. We used cut-off of −0.3 to determine mono-allelic loss of BRN2.

Copy number data with matched nevus-melanoma pairs used in Supplementary Fig. 1D was obtained from a previous study³². Log2 copy ratio values were processed using the PureCN R package⁶⁶ to estimate tumor purity and copy number.

Level of expression of BRN2 mRNA in melanoma patients

BRN2 mRNA levels were obtained from RNA sequencing data analyzed using the RNA Seq V2 RSEM pipeline as transcripts per million reads (TPM).

Mouse models

Mice were bred and maintained in the specific pathogen-free mouse colony of the Institut Curie, in accordance with the institute’s regulations and French and European Union laws. Mice were bred and maintained in the specific pathogen-free mouse colony of the Institut Curie, in accordance with the institute’s regulations and French and European Union laws. The transgenic Tyr::Cre^ERT2 (031281—B6N.Cg-Tg(Tyr-cre/ERT2)1Lru/J), Braf^V600E/+, Pten (006440 - B6.129S4-Ptentm1Hwu/J) and Brn2 mice have been described and characterized elsewhere^{33,34,35,38,67}. All mouse lines were backcrossed onto a C57BL/6 background for more than ten generations. All desired combinations of genotypes were obtained through crosses. Mice were born with the expected ratio of Mendelian inheritance and no changes in gender ratios were observed. Experimental mice were of both genders and no apparent phenotypic differences between genders were observed. No statistical methods were used to predetermine sample size. The sample size was sufficient to measure the effect size for all experiments presented in this study. The experiments were not randomized, and the investigators were not blinded to allocation during the experiments and outcome assessment. Mice were housed in a certified animal facility with a 12-hour light/dark cycle in a temperature-controlled room (22 ± 1 °C) with free access to water and food.

Growth of the mouse melanoma cell lines in syngeneic mice

C57BL/6 mice were purchased from Charles River Laboratories. Twenty-four C57BL/6 mice were injected with Brn2 + /+ [m6 and m50], Brn2 F/ + [m36 and m59], and Brn2 F/F [m8 and m82] mouse melanoma cell lines. The cells were resuspended in PBS and 10⁵ cells (100 μL) subcutaneously implanted into the flanks of seven-week-old C57BL/6 mice using a 27-gauge needle. Presence of tumors was detected from day 15 to 50, it was independent of the genotypes.

Ethical rules

Animal care, use, and experimental procedures were conducted in accordance with recommendations of the European Community (86/609/EEC) and Union (2010/63/UE) and the French National Committee (87/848). Animal care and use were approved by the ethics committee of the Curie Institute in compliance with the institutional guidelines. Experimental procedures were specifically approved by the ethics committee of the Institut Curie CEEA-IC #118 (CEEA-IC 2016-001) in compliance with the international guidelines.

Mouse genotyping

Mouse biopsies were digested overnight at 55°C using 200 ng proteinase K (Roche, #11 243 233 001) in 500 µL lysis buffer containing 16 mM [NH4]2 SO4, 67 mM Tris-HCl [pH 8.8 at 25 °C], 0.01% [v/v] Tween-20, in deionized H₂O. Proteinase K was inactivated for 20 min at 95 °C. Primers and PCR conditions are described (Supplementary Tables 3 and 4). PCR products were separated by agarose (Invitrogen, #15510027) gel electrophoresis. Genotyping were performed accordingly (Supplementary Tables 3 and 4).

In vivo gene activation/deletion and melanoma monitoring

Newborn mice were treated dorsally with 20 µL/day/mouse tamoxifen (Sigma, T5648, working concentration 20 μg/mL in DMSO) for the first three consecutive days after birth. Non-tamoxifen-induced mice of the same genotype were used as controls. After the application of tamoxifen, the mice were evaluated for the appearance of tumors and their progression once per week or more frequently if required. Developing skin excrescences > 3 mm diameter were considered to be melanomas, and validated after growth. Mice were euthanized and autopsied four weeks after tumor appearance or once the tumors reached 2 cm³. Melanoma-specific survival curves were estimated from the day of euthanasia. Mouse melanomas were excised, rinsed in cold phosphate-buffered saline (PBS, Euromedex, ET330-A) and divided into four parts, two snap-frozen in liquid nitrogen for subsequent transcriptomic and western blot analysis and two fixed in 4% paraformaldehyde (PFA, Euromedex, 15714-S) and embedded in paraffin or OCT (VWR, #00411243) for histological analysis and immunostaining. Inguinal lymph nodes were fixed in 4% PFA and embedded in paraffin or OCT for histological analysis and immunostaining.

Detection of defloxing from mouse melanoma tissue

DNA extraction from paraffin-embedded melanoma sections (10 μm) was performed using the QIAamp DNA FFPE Tissue Kit (Qiagen, #56404), according to the manufacturer’s instructions. A PCR using DNA extracted from each tumor was performed to verify the mouse genotype and proper defloxing of the modified genes (see Key Resource Table and Supplementary Tables 3 and 4 for PCR conditions and primers sequences).

Immunohistochemistry of mouse melanoma and inguinal lymph nodes

Paraffin-embedded mouse melanomas were sectioned into 7-μm-thick transverse sections and stained with hematoxylin/eosin (H&E), as previously described⁶⁸. For immunostaining, sections were deparaffinized, rinsed in Tris-buffered saline (TBS; 20 mM Tris (Sigma: T-1503) pH7.6), 150 mM NaCl (OSI, A4321152), containing 0.1% [v/v] Tween-20 (VWR 8221840500) (TBST), depigmented with H₂O₂ (Sigma, H1009) for 15 min, boiled for 20 min in 10 mM sodium citrate (VWR, 1120051000), and blocked with TBST containing 3% bovine serum albumin (BSA, Sigma A9418). Sections were incubated overnight at 4 °C in TBST containing 3% BSA with antibodies against Ki-67 (Nova-Costra, NCL-Ki67p), BrdU (BD Biosciences, #555627), or PTEN (Cell signaling, #9559). The sections were then incubated in secondary biotinylated anti-rabbit or anti-mouse antibodies for 2 h at room temperature (RT). AEC (Sigma-Aldrich, A6926) was used to reveal bound antibody according to the manufacturer’s instructions. All sections were counterstained with hematoxylin. Images were captured using a ZEISS Axio Imager 2 with Axiocam 506 color cameras. Image analysis was performed using ZEISS ZEN, Adobe Photoshop, and ImageJ software. Quantifications of Ki-67 and BrdU stainings were determined as a percentage. The percentage of Ki-67⁺ and BrdU⁺ cells from three fields (1000–2000 cells/field) from three independent tumors per genotype was determined and normalized.

Immunofluorescence of mouse inguinal lymph nodes

OCT-embedded lymph nodes were sectioned into 7-μm-thick transverse sections and rinsed in TBST, depigmented with H₂O₂ (Sigma, H1009) for 15 min, boiled for 20 min in 10 mM sodium citrate (VWR, 1120051000) and blocked with TBST containing 3% BSA (Sigma A9418). Sections were incubated overnight at 4 °C in TBST containing 3% BSA with antibodies against SOX10 (Abcam, ab155279). Sections were then incubated in secondary donkey anti-mouse 647 (Abcam, ab150107) and goat anti-rabbit IgG (H&L)-Affinity Pure, DyLight® 650 Conjugate (GtxRb-003-E650NHSX) antibodies for 2 h at RT. All sections were counterstained with DAPI (0.5 μg/mL in ethanol, Sigma, D9542). Images were captured using a ZEISS Axio Imager 2 with Axiocam 506 color cameras. Image analysis was performed using ZEISS ZEN, Adobe Photoshop, and ImageJ software.

Protein extraction from tumors

All steps were performed at 4 °C. Tissues were transferred to a tube containing 2.8-mm stainless steel beads and 1 mL RIPA buffer supplemented with sodium orthovanadate (Sigma, S6508), Complete protease inhibitor (Roche, #11873580001) and Phostop (Roche, #04906837001) were added. Tissues were homogenized three times for 3 min using the BeadBug™ homogenizer at 4 °C, followed by a brief centrifugation at 17,949 g for 1 min at 4 °C. Supernatants were transferred and centrifuged for 20 min at 15,294 × g. Supernatants were collected and incubated with 200 µL previously PBS-washed G-Sepharose beads (GE Healthcare, #17-0618-01) for 2 h. Samples were centrifuged at 15,294 × g for 5 min and quantified using the Bradford assay.

Microarray analysis

Only one tumor per mouse was considered and corresponding to the biggest one. Tumors had the same size since we harvested tumors for transcriptomic analyses when they reached a size of 1cm³. RNA from mouse melanomas (5 Braf-Pten-Brn2-WT, 10 Braf-Pten-Brn2-het, and 10 Braf-Pten-Brn2-hom) and mouse melanoma cell lines (4 Braf-Pten-Brn2-WT, 2 Braf-Pten-Brn2-het, and 3 Braf-Pten-Brn2-hom) established from the mouse melanoma tumors was extracted using the miRNeasy Kit (Qiagen, #217004). RNA Integrity was assessed using an Agilent BioAnalyser 2100 (Agilent Technologies), only RNA with a RIN > 7 were kept for the analysis. RNA concentrations were measured using a NanoDrop (NanoDrop Technologies). Complementary RNA (cRNA) was synthesized using the GeneChip 3′IVT Plus reagent Kit (Thermofisher, #902415), according to the manufacturer’s protocol. In brief, total RNA was first reverse transcribed using a T7-Oligo(dT) promoter primer for first-strand cDNA synthesis. After RNase H treatment and second-strand cDNA synthesis, the double-stranded cDNA was purified and served as template for subsequent in vitro transcription (IVT). The IVT reaction was carried out in the presence of T7 RNA polymerase and a biotinylated nucleotide analog/ribonucleotide mix for cRNA amplification and biotin labeling. The biotinylated cRNA targets were then cleaned up, fragmented, and 11 μg cRNA hybridized to mouse MOE430 gene expression Affymetrix microarrays (Affymetrix, #900443). After washing and staining, using the Affymetrix fluidics station 450 (Affymetrix, # 00-0079), the probe arrays were scanned using an Affymetrix GeneChip Scanner 3000 (Affymetrix, # 00-0210).

The microarray data were normalized using the RMA (Robust Multichip Average) function of the edgeR package⁶⁹. For genes targeted by multiple probesets, only the most variant probeset was kept. Differentially expressed gene analysis was performed with R, for the cell lines and the tumors using the limma package⁷⁰ available from Bioconductor (http://www.bioconductor.org). An enriched gene-ontology (Gene-Ontology Biological-Process, 2018) and pathway (WikiPathways human, 2019 & KEGG human, 2019) analysis was performed on the 296 genes found overexpressed in the Braf-Pten-Brn2-het compared to the Braf-Pten-Brn2-WT using Enrichr⁷¹. Gene-set enrichment analysis (GSEA) was performed using the “Verfaillie” signature⁴⁴. These signatures originally containing too much genes (>1,000) were reduced to fit to the GSEA algorithm by selecting the most differentially expressed genes based on their fold change, a threshold of 4 standard deviation was selected. For the tumors analysis the GSEA was ran with one thousand permutations gene set based. For the analysis of the cell lines, the run pre-ranked function of the GSEA software was used using one thousand permutations gene set based. The genes were ranked according to their statistic t coming from the differential analysis results. The enrichment score (ES) reflects the degree to which a given gene set is represented in a ranked list of genes. Calculation of the ES is based on walking down a ranked list of genes and adjusting a running-sum statistic based on the presence of absence of a gene in the gene set. The magnitude of the increment represents the correlation of the gene with the phenotype. P values were estimated by gene-based permutation. GSEA normalizes the enrichment score for each gene set to account for the variation in set sizes, yielding a normalized enrichment score (NES). Only gene set with a NES absolute > 1.7 and a FDR < 0.01 were considered.

Cell lines

Melan-a and C57BL/6 9 v cells were grown in Ham’s F12 media (GIBCO 11765054) supplemented with 10% fetal calf serum (FCS, GIBCO, 10270106), 1% Penicillin-Streptomycin (GIBCO, 15140), and 200 nM TPA (Sigma, P 8139)⁴. 501mel, 501mel-MITF-HA, HBL, SK-Mel-28, SK-Mel-28 Δex6, and Dauv-1 cells were grown in RPMI 1640 media (GIBCO, 11875101) supplemented with 10% FCS (GIBCO, 10270106) and 1% Penicillin-Streptomycin (GIBCO, 15140)^50,72,73,74. SK28Δex6 melanoma cells lacking Mitf was previously produced⁵⁰. Cells are routinely tested for the absence of mycoplasmas using MycoAlert (Lonza). Mouse melanoma cell lines were established as previously described⁷⁵. The genetic status and the level of expression of key genes of these cell lines is given in Supplementary Table 1.

Genomic DNA extraction from cell culture

Genomic DNA was extracted from melanoma cell lines using the AllPrep DNAMini Kit (Qiagen, #80204) or QIAamp Kit (QIAGEN) according to manufacturer’s instructions. The DNA region for BRAF and NRAS was amplified by PCR and submitted for sequencing (see for primers and conditions in the Key Resource Table and Supplementary Tables 3 and 4).

siRNA-mediated knock-down

siRNA targeting human BRN2 and MITF was purchased from Dharmacon as a SMART-pool mix of four sequences. siRNA targeting PTEN was purchased from Santa Cruz Biotech. Si Scramble (siSCR), with no known human or mouse targets, was purchased from Eurofins Genomics (Supplementary Table 5). All sequences and product references are listed in Supplementary Table 6. Briefly, cells were transfected with Lipofectamin2000 with 200 pmol siRNA or siSCR and assayed for mRNA expression or protein content 48 or 72 h post-transfection.

Cell counting

Phase-contrast pictures of cells were taken using a Zeiss Axiovert 135 microscope with an Axiocam MRC camera. Cells were counted using a LUNA automated cell counter (L10001) and cell-counting slides (L12001).

Clonogenic assays

For clonogenic assays, six-well tissue culture plates were seeded with 500 cells in complete medium. The medium was changed 24 h after cell seeding and replaced with complete medium containing the indicated concentrations of Binimetinib (Selleck), MK-2206 2HCl (Selleck), LY294002 (Calbiochem) or PLX4720 (Axon Medchem). After 9 days of incubation, colonies were fixed with 4% PFA (paraformaldehyde), stained with Crystal violet in 10% ethanol, and counted on images. IC50 were determined for each pharmacological agent and for each cell line using the number of colonies in mock treated condition as the top response. Cell lines have been treated with Binimetinib from 10⁻³ to 100 μM, with PLX4720 from 10⁻³ to 100 μM, with LY294002 from 5.10⁻³ to 50 μM and with MK-2206 from 10⁻³ to 10 μM. We used the resulted sigmoidal curves to calculate the IC50 with Graphpad Prism. Experiments were performed at least in triplicate.

Western blotting and detection

Whole-cell lysate was prepared from human melanoma cell lines using RIPA buffer supplemented with sodium orthovanadate (Sigma S 6508), complete inhibitor (Roche, #11873580001), and Phostop (Roche, #04906837001). SDS-PAGE was carried out on homemade 10% polyacrylamide protein gels. Following the transfer of the proteins, the nitrocellulose membranes were blocked in TBST with 5% non-fat dry milk for 1.5 h at RT and then probed with antibodies against BRN2 (Cell signaling, #12137), MITF (Abcam, ab12039), β-actin (Sigma, A5441), cyclin D1 (Cell signaling, #2926), PTEN (Cell signaling, #9559), phospho-S6 (Cell signaling, #4857), S6 (Cell signaling, #2317), phospho-AKT (Cell signaling, #3787), AKT (Cell signaling, #4685), or vinculin (Sigma, V9131). Primary antibodies were applied in TBST/5% non-fat dry milk overnight at 4 °C and visualized using secondary antibodies (HRP-conjugated goat anti-rabbit IgG, Jackson, 111-035-003 and HRP-conjugated goat anti-mouse IgG, Jackson, 115-035-003) in TBST/5% non-fat dry milk for 1 h at RT. Blots were incubated in ECL (Pierce, #34075) and revealed in the dark using ECL hyperfilm (GE Healthcare, RPN3103K). All primary antibodies were used at a dilution of 1/1,000, except β-actin and vinculin (1/5,000). All secondary antibodies were used at a dilution of 1/20,000. Pageruler (ThermoFisher, #26616) was used as the molecular marker. Quantification of the western blots was performed using ImageJ software. Quantification of the western blots was performed using ImageJ software. See Supplementary table 7.

Chromatin immunoprecipitation

ChIP experiments were performed as previously described⁴, and see Supplementary Table 6. ChIP assays of BRN2 binding to the PTEN promoter. ChIP assays were performed using an antibody against BRN2 and analyzed after 30-cycle PCR in exponentially growing phase of Dauv-1 melanoma cells. PAX3 promoter (prom.) and Brn2 coding sequences (CDS) were used as positive and negative controls, respectively. Input represents approximately 0.4% of the input used for the ChIP assay. H3 (histone H3) and IgG (Immunoglobulin G) were used as positive and negative controls for each region of interest, respectively. The oligonucleotides, their position on the genome, and the sizes of the amplified fragments are given in Supplementary Tables 3 and 4.

RNA extraction and (ChIP) RT-qPCR

Tissues were crushed with a mortar and pestle, and stainless steel beads. Qiazol was used to homogenize the samples prior extracting RNA using the miRNeasy Kit. Purified RNA was reversed transcribed using M-MLV Reverse Transcriptase. Real-time quantitative PCR (qPCR) was performed using iTaq™ Universal SYBR Green Supermix. Each sample was run in technical triplicates and the quantified RNA normalized against TBP (human) or Gapdh (mouse) as housekeeping transcripts (Supplementary Tables 3 and 4).

CUT&RUN

Anti-MITF, anti-H3K27Ac and anti-H3K4Me3 Cleavage Under Targets and Release Using Nuclease (CUT&RUN) sequencing was performed in SK28 melanoma cell lines that are MITF-WT or null (ΔMITF) as described⁷⁶, with minor modifications. Cells (~75–80% confluent) were harvested by cell scraping (Corning), centrifuged at 600 × g (Eppendorf, centrifuge 5424) and washed in calcium-free wash-buffer (20 mM HEPES, pH7.5, 150 mM NaCl, 0.5 mM spermidine and protease inhibitor cocktail, cOmplete Mini, EDTA-free Roche). Pre-activated Concanavalin A-coated magnetic beads (Bangs Laboratories Inc) were added to 100uL cell suspensions (2 × 10⁵ cells) and incubated at 4 ^oC for 15 min. Antibody dilution buffer (wash-buffer with 2 mM EDTA and 0.03% digitonin) containing anti-MITF (Sigma, HPA003259, 1:100), anti-H3K27Ac (Millipore, 07-360, 1:100), anti-H3K4Me3 (Millipore, 05-745 R, 1:100) and Rabbit IgG (Millipore, 12-370, 1:100) was added and cells were incubated at 4 ^oC overnight. The next day, cells were placed on a magnetic rack and washed twice in dig-wash buffer (wash buffer containing 0.025% digitonin). pAG-MNase at a concentration of 500 μg/ mL was added and cells were incubated at 4^oC for 30 mins (pAG-MNase was purified in Dr. Robert Cornell’s research group at the University of Iowa). The pAG-MNase reactions were quenched with 2X Stop buffer (340 mM NaCl, 867 20 mM EDTA, 4 mM EGTA, 0.05% Digitonin, 100 μg/ mL RNAse A and 50 μg/ mL Glycogen). Released DNA fragments were Phosphatase K (1 μL/mL, Thermo Fisher Scientific) treated at 50^oC for 1 hr and purified by phenol/chloroform-extracted and ethanol-precipitated. Fragment sizes were analyzed using a 2100 Bioanalyzer (Agilent). All CUT&RUN experiments were performed in duplicate.

Library preparation and data analysis

CUT&RUN-seq libraries were prepared using the KAPA Hyper Prep Kit (Roche). Quality control post-library amplification was conducted using the 2100 Bioanalyzer (Agilent). Fragment analysis and fragments sizes were compared pre- and post-library amplification to insure correct size selection. Libraries were pooled to equimolar concentrations and sequenced with paired-end 100 bp reads on an Illumina HiSeq X platform. Paired-end FastQ files were processed through FastQC (Babraham Bioinformatics) for quality control. Reads were trimmed using Trim Galore Version 0.6.3 (Developed by Felix Krueger at the Babraham Institute), Bowtie2 version 2.1.0⁷⁷ was used to map the reads against the hg19 genome assembly and MACS2 Version 2.1.1.20160309.6 was used to call peaks. The mapping parameters and peak calling analysis was performed as previously described^76,78 using IgG samples as background control. The Deeptools function “BAMcoverage”⁷⁹ was used to generate normalized (–RPKM) BigWigs files for visualization on Integrative Genomics Viewer (IGV)⁸⁰.

Luciferase assays

Human (−2,375 to +840) and murine (−2136 to +936) Pten-promoter fragments were cloned upstream 56 bp of CMV promoter conferring a very weak basal activity upstream luciferase (miniCMV::Luc, VectorBuilder) to generate hsPTEN::Luc (#1282) and mmPten::Luc (#1283) reporter vectors. Cells at 70% confluence were transiently transfected in twelve-well plates, using 2μL of Lipofectamin2000 (Invitrogen), 500 ng of total plasmid DNA 200 ng Pten::Luc reporter plasmids (#1282 or #1283) or miniCMV::Luc reporter (#1281), 200 ng of the expression vectors CMV::EGFP-Brn2 (#896) or CMV::EGFP (#1042) (equimolar) as a control²⁷; 100 ng of the HSV-TK::renilla luciferase construct (#894) in Opti-MEM medium (Gibco). Luciferase activity and renilla luciferase activity were determined 48 h after transfection. Luciferase activity was normalized against renilla luciferase activity.

Software

GraphPad PRISM, R version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria), Adobe Illustrator, Adobe Photoshop, and Microsoft Power Point software were used to analyze data and generate all graphs and figures.

Quantification and statistical analysis

Cell culture-based experiments were performed in at least biological triplicate and validated three times as technical triplicates. P values for the comparison of two groups were calculated using the unpaired Student t-test or Mann-Whitney test. P values for the comparison of multiple groups were calculated using the analysis of variance (ANOVA) and Fisher’s least significant difference tests. P-values for categorical data were calculated using the Chi-square test. P-values for the comparison of Kaplan–Meier curves were calculated using the log-rank (Mantel–Cox) or Gehan-Breslow-Wilcoxon test giving weight to the early events. P values were reported as computed by Prism 6.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

Microarray gene expression data that support the findings of this study have been deposited in Gene Expression Omnibus (GEO) with the accession codes GSE126524 and GSE163085. The TCGA Skin Cutaneous Melanoma data referenced during the study are available in a public repository from the National Cancer Institute (NCI) Genomic Data Commons (GDC) website (https://portal.gdc.cancer.gov). Cut&Run assay data that support the findings of this study have been deposited in Gene Expression Omnibus (GEO) with the accession codes GSE153020. The accession number for the sequencing data reported in this paper is dbGaP: phs001550.v1.p1. All the other data supporting the findings of this study are available within the article and its supplementary information/data files and from the corresponding author upon reasonable request. A reporting summary for this article is available as a Supplementary Information file. The data that support the findings of this study are available from the corresponding author upon reasonable request. All details concerning antibodies, chemicals, critical commercial assays, cell lines, model organisms, oligonucleotides, and software and algorithms can be found in Supplementary Table 7a–g. Source data are provided with this paper.

References

Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417, 949–954 (2002).
Article ADS CAS PubMed Google Scholar
Eskandarpour, M. et al. Frequency of UV-inducible NRAS mutations in melanomas of patients with germline CDKN2A mutations. J. Natl. Cancer Inst. 95, 790–798 (2003).
Article CAS PubMed Google Scholar
Conde-Perez, A. et al. A caveolin-dependent and PI3K/AKT-independent role of PTEN in beta-catenin transcriptional activity. Nat. Commun. 6, 8093 (2015).
Article ADS CAS PubMed Google Scholar
Delmas, V. et al. Beta-catenin induces immortalization of melanocytes by suppressing p16INK4a expression and cooperates with N-Ras in melanoma development. Genes Dev. 21, 2923–2935 (2007).
Article CAS PubMed PubMed Central Google Scholar
Bennett, D. C. Genetics of melanoma progression: the rise and fall of cell senescence. Pigment Cell Melanoma Res. 29, 122–140 (2016).
Article CAS PubMed Google Scholar
Gray-Schopfer, V. C. et al. Cellular senescence in naevi and immortalisation in melanoma: a role for p16? Br. J. Cancer 95, 496–505 (2006).
Article CAS PubMed PubMed Central Google Scholar
Gembarska, A. et al. MDM4 is a key therapeutic target in cutaneous melanoma. Nat. Med. 18, 1239–1247 (2012).
Article CAS PubMed Google Scholar
Schreiber, E. et al. Astrocytes and glioblastoma cells express novel octamer-DNA binding proteins distinct from the ubiquitous Oct-1 and B cell type Oct-2 proteins. Nucleic Acids Res. 18, 5495–5503 (1990).
Article CAS PubMed PubMed Central Google Scholar
Bishop, J. L. et al. The master neural transcription factor BRN2 is an androgen receptor-suppressed driver of neuroendocrine differentiation in prostate cancer. Cancer Disco. 7, 54–71 (2017).
Article CAS Google Scholar
Ishii, J. et al. POU domain transcription factor BRN2 is crucial for expression of ASCL1, ND1 and neuroendocrine marker molecules and cell growth in small cell lung cancer. Pathol. Int. 63, 158–168 (2013).
Article CAS PubMed Google Scholar
Chitsazan, A. et al. Unexpected high levels of BRN2/POU3F2 expression in human dermal melanocytic nevi. J. Invest. Dermatol 140, 1299–1302 e4 (2020).
Article CAS PubMed Google Scholar
Colombo, S., Champeval, D., Rambow, F. & Larue, L. Transcriptomic analysis of mouse embryonic skin cells reveals previously unreported genes expressed in melanoblasts. J. Investig. Dermatol. 132, 170–178 (2012).
Article CAS PubMed Google Scholar
Goodall, J. et al. Brn-2 represses microphthalmia-associated transcription factor expression and marks a distinct subpopulation of microphthalmia-associated transcription factor-negative melanoma cells. Cancer Res. 68, 7788–7794 (2008).
Article CAS PubMed Google Scholar
Goodall, J. et al. Brn-2 expression controls melanoma proliferation and is directly regulated by beta-catenin. Mol. Cell Biol. 24, 2915–2922 (2004).
Article CAS PubMed PubMed Central Google Scholar
Goodall, J. et al. The Brn-2 transcription factor links activated BRAF to melanoma proliferation. Mol. Cell Biol. 24, 2923–2931 (2004).
Article CAS PubMed PubMed Central Google Scholar
Bonvin, E., Falletta, P., Shaw, H., Delmas, V. & Goding, C. R. A phosphatidylinositol 3-kinase-Pax3 axis regulates Brn-2 expression in melanoma. Mol. Cell Biol. 32, 4674–4683 (2012).
Article CAS PubMed PubMed Central Google Scholar
Cook, A. L., Smith, A. G., Smit, D. J., Leonard, J. H. & Sturm, R. A. Co-expression of SOX9 and SOX10 during melanocytic differentiation in vitro. Exp. Cell Res. 308, 222–235 (2005).
Article CAS PubMed Google Scholar
Goding, C. R. & Arnheiter, H. MITF-the first 25 years. Genes Dev. 33, 983–1007 (2019).
Article CAS PubMed PubMed Central Google Scholar
Boyle, G. M. et al. Melanoma cell invasiveness is regulated by miR-211 suppression of the BRN2 transcription factor. Pigment Cell Melanoma Res. 24, 525–537 (2011).
Article CAS PubMed Google Scholar
Zeng, H. et al. Bi-allelic Loss of CDKN2A initiates melanoma invasion via BRN2 activation. Cancer Cell 34, 56–68 e9 (2018).
Article CAS PubMed PubMed Central Google Scholar
Smith, M. P. et al. A PAX3/BRN2 rheostat controls the dynamics of BRAF mediated MITF regulation in MITF(high) /AXL(low) melanoma. Pigment Cell Melanoma Res. 32, 280–291 (2019).
Article CAS PubMed Google Scholar
Potterf, S. B., Furumura, M., Dunn, K. J., Arnheiter, H. & Pavan, W. J. Transcription factor hierarchy in Waardenburg syndrome: regulation of MITF expression by SOX10 and PAX3. Hum. Genet. 107, 1–6 (2000).
Article CAS PubMed Google Scholar
Bondurand, N. et al. Interaction among SOX10, PAX3 and MITF, three genes altered in Waardenburg syndrome. Hum. Mol. Genet. 9, 1907–1917 (2000).
Article CAS PubMed Google Scholar
Vivas-Garcia, Y. et al. Lineage-restricted regulation of SCD and fatty acid saturation by MITF controls melanoma phenotypic plasticity. Mol. Cell 77, 120–139 (2020).
Article CAS PubMed Google Scholar
Simmons, J. L., Neuendorf, H. M. & Boyle, G. M. BRN2 and MITF together impact AXL expression in melanoma. Exp. Dermatol. https://doi.org/10.1111/exd.14225 (2020).
Article PubMed PubMed Central Google Scholar
Fane, M. E., Chhabra, Y., Smith, A. G. & Sturm, R. A. BRN2, a POUerful driver of melanoma phenotype switching and metastasis. Pigment Cell Melanoma Res. 32, 9–24 (2019).
Article CAS PubMed Google Scholar
Berlin, I. et al. Phosphorylation of BRN2 modulates its interaction with the Pax3 promoter to control melanocyte migration and proliferation. Mol. Cell. Biol. 32, 1237–1247 (2012).
Article CAS PubMed PubMed Central Google Scholar
Pierce, C. J. et al. BRN2 expression increases anoikis resistance in melanoma. Oncogenesis 9, 64 (2020).
Article CAS PubMed PubMed Central Google Scholar
Pinner, S. et al. Intravital imaging reveals transient changes in pigment production and Brn2 expression during metastatic melanoma dissemination. Cancer Res. 69, 7969–7977 (2009).
Article CAS PubMed PubMed Central Google Scholar
Arozarena, I. et al. In melanoma, beta-catenin is a suppressor of invasion. Oncogene 30, 4531–4543 (2011).
Article CAS PubMed PubMed Central Google Scholar
Cirenajwis, H. et al. NF1-mutated melanoma tumors harbor distinct clinical and biological characteristics. Mol. Oncol. 11, 438–451 (2017).
Article CAS PubMed PubMed Central Google Scholar
Shain, A. H. et al. Genomic and transcriptomic analysis reveals incremental disruption of key signaling pathways during melanoma evolution. Cancer Cell 34, 45–55 (2018).
Article CAS PubMed PubMed Central Google Scholar
Yajima, I. et al. Spatiotemporal gene control by the Cre-ERT2 system in melanocytes. Genesis 44, 34–43 (2006).
Article CAS PubMed Google Scholar
Dhomen, N. et al. Oncogenic Braf induces melanocyte senescence and melanoma in mice. Cancer Cell 15, 294–303 (2009).
Article CAS PubMed Google Scholar
Lesche, R. et al. Cre/loxP-mediated inactivation of the murine Pten tumor suppressor gene. Genesis 32, 148–149 (2002).
Article CAS PubMed Google Scholar
Dhomen, N. et al. Inducible expression of (V600E) Braf using tyrosinase-driven Cre recombinase results in embryonic lethality. Pigment Cell Melanoma Res. 23, 112–120 (2010).
Article CAS PubMed Google Scholar
Dankort, D. et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat. Genet. 41, 544–552 (2009).
Article CAS PubMed PubMed Central Google Scholar
Jaegle, M. et al. The POU proteins Brn-2 and Oct-6 share important functions in Schwann cell development. Genes Dev. 17, 1380–1391 (2003).
Article CAS PubMed PubMed Central Google Scholar
Bennett, D. C. Human melanocyte senescence and melanoma susceptibility genes. Oncogene 22, 3063–3069 (2003).
Article CAS PubMed Google Scholar
Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005).
Article ADS CAS PubMed Google Scholar
Puig, I. et al. Deletion of Pten in the mouse enteric nervous system induces ganglioneuromatosis and mimics intestinal pseudoobstruction. J. Clin. Invest. 119, 3586–3596 (2009).
Article CAS PubMed PubMed Central Google Scholar
Simmons, J. L., Pierce, C. J., Al-Ejeh, F. & Boyle, G. M. MITF and BRN2 contribute to metastatic growth after dissemination of melanoma. Sci. Rep. 7, 10909 (2017).
Article ADS PubMed PubMed Central CAS Google Scholar
Herbert, K. et al. BRN2 suppresses apoptosis, reprograms DNA damage repair, and is associated with a high somatic mutation burden in melanoma. Genes Dev. 33, 310–332 (2019).
Article CAS PubMed PubMed Central Google Scholar
Verfaillie, A. et al. Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state. Nat. Commun. 6, 6683 (2015).
Article ADS CAS PubMed Google Scholar
Muller, J. et al. Low MITF/AXL ratio predicts early resistance to multiple targeted drugs in melanoma. Nat. Commun. 5, 5712 (2014).
Article ADS PubMed CAS Google Scholar
Davies, M. A. et al. Integrated molecular and clinical analysis of AKT activation in metastatic melanoma. Clin. Cancer Res. 15, 7538–7546 (2009).
Article CAS PubMed PubMed Central Google Scholar
Vredeveld, L. C. et al. Abrogation of BRAFV600E-induced senescence by PI3K pathway activation contributes to melanomagenesis. Genes Dev. 26, 1055–1069 (2012).
Article CAS PubMed PubMed Central Google Scholar
Thurber, A. E. et al. Inverse expression states of the BRN2 and MITF transcription factors in melanoma spheres and tumour xenografts regulate the NOTCH pathway. Oncogene 30, 3036–3048 (2011).
Article CAS PubMed PubMed Central Google Scholar
Wellbrock, C. et al. Oncogenic BRAF regulates melanoma proliferation through the lineage specific factor MITF. PLoS ONE 3, e2734 (2008).
Article ADS PubMed PubMed Central CAS Google Scholar
Dilshat, R. et al. MITF reprograms the extracellular matrix and focal adhesion in melanoma. Elife 10, e63093 (2021).
Article PubMed PubMed Central Google Scholar
Thomson, J. A. et al. The Brn-2 gene regulates the melanocytic phenotype and tumorigenic potential of human melanoma cells. Oncogene 11, 691–700 (1995).
CAS PubMed Google Scholar
Healy, E. et al. Prognostic significance of allelic losses in primary melanoma. Oncogene 16, 2213–2218 (1998).
Article CAS PubMed Google Scholar
Bastian, B. C., LeBoit, P. E., Hamm, H., Brocker, E. B. & Pinkel, D. Chromosomal gains and losses in primary cutaneous melanomas detected by comparative genomic hybridization. Cancer Res. 58, 2170–2175 (1998).
CAS PubMed Google Scholar
Guan, X. Y. et al. Detection of chromosome 6 abnormalities in melanoma cell lines by chromosome arm painting probes. Cancer Genet. Cytogenet. 107, 89–92 (1998).
Article CAS PubMed Google Scholar
Lee, J. J. et al. Targeted next-generation sequencing reveals high frequency of mutations in epigenetic regulators across treatment-naive patient melanomas. Clin. Epigenetics 7, 59 (2015).
Article PubMed PubMed Central CAS Google Scholar
Shain, A. H. et al. Exome sequencing of desmoplastic melanoma identifies recurrent NFKBIE promoter mutations and diverse activating mutations in the MAPK pathway. Nat. Genet. 47, 1194–1199 (2015).
Article CAS PubMed PubMed Central Google Scholar
Saito, Y. et al. Endogenous melanin-concentrating hormone receptor SLC-1 in human melanoma SK-MEL-37 cells. Biochem. Biophys. Res. Commun. 289, 44–50 (2001).
Article ADS CAS PubMed Google Scholar
Cifola, I. et al. Comprehensive genomic characterization of cutaneous malignant melanoma cell lines derived from metastatic lesions by whole-exome sequencing and SNP array profiling. PLoS One 8, e63597 (2013).
Article ADS PubMed PubMed Central Google Scholar
Zhang, J. P. et al. Notch ligand Delta-like 1 promotes the metastasis of melanoma by enhancing tumor adhesion. Braz. J. Med Biol. Res. 47, 299–306 (2014).
Article CAS PubMed PubMed Central Google Scholar
Wiesner, T. et al. Kinase fusions are frequent in Spitz tumours and spitzoid melanomas. Nat. Commun. 5, 3116 (2014).
Article ADS PubMed CAS Google Scholar
Ray, M. E., Su, Y. A., Meltzer, P. S. & Trent, J. M. Isolation and characterization of genes associated with chromosome-6 mediated tumor suppression in human malignant melanoma. Oncogene 12, 2527–2533 (1996).
CAS PubMed Google Scholar
Laurette, P. et al. Transcription factor MITF and remodeller BRG1 define chromatin organisation at regulatory elements in melanoma cells. Elife 4 (2015).
Cirenajwis, H. et al. Molecular stratification of metastatic melanoma using gene expression profiling: Prediction of survival outcome and benefit from molecular targeted therapy. Oncotarget 6, 12297–12309 (2015).
Article PubMed PubMed Central Google Scholar
Li, J. et al. CONTRA: copy number analysis for targeted resequencing. Bioinformatics 28, 1307–1313 (2012).
Article PubMed PubMed Central CAS Google Scholar
Hupe, P., Stransky, N., Thiery, J. P., Radvanyi, F. & Barillot, E. Analysis of array CGH data: from signal ratio to gain and loss of DNA regions. Bioinformatics 20, 3413–3422 (2004).
Article CAS PubMed Google Scholar
Riester, M. et al. PureCN: copy number calling and SNVclassification using targeted short readsequencing. Source Code Biol. Med. 11, 13 (2016).
Article PubMed PubMed Central Google Scholar
Aktary, Z., Corvelo, A., Estrin, C. & Larue, L. Sequencing two Tyr::CreER(T2) transgenic mouse lines. Pigment Cell Melanoma Res. 33, 426–434 (2020).
Article CAS PubMed Google Scholar
Gallagher, S. J. et al. General strategy to analyse melanoma in mice. Pigment Cell Melanoma Res. 24, 987–988 (2011).
Article PubMed Google Scholar
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Article CAS PubMed Google Scholar
Ritchie, M. E. et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).
Article PubMed PubMed Central CAS Google Scholar
Chen, E. Y. et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinforma. 14, 128 (2013).
Article Google Scholar
Strub, T. et al. Essential role of microphthalmia transcription factor for DNA replication, mitosis and genomic stability in melanoma. Oncogene 30, 2319–2332 (2011).
Article CAS PubMed Google Scholar
Ghanem, G. E., Comunale, G., Libert, A., Vercammen-Grandjean, A. & Lejeune, F. J. Evidence for alpha-melanocyte-stimulating hormone (alpha-MSH) receptors on human malignant melanoma cells. Int. J. Cancer 41, 248–255 (1988).
Article CAS PubMed Google Scholar
Rambow, F. et al. New functional signatures for understanding melanoma biology from tumor cell lineage-specific analysis. Cell Rep. 13, 840–853 (2015).
Article CAS PubMed PubMed Central Google Scholar
Petit, V. et al. C57BL/6 congenic mouse NRAS(Q61K) melanoma cell lines are highly sensitive to the combination of Mek and Akt inhibitors in vitro and in vivo. Pigment Cell Melanoma Res. 32, 829–841 (2019).
Article CAS PubMed Google Scholar
Skene, P. J. & Henikoff, S. An efficient targeted nuclease strategy for high-resolution mapping of DNA binding sites. Elife 6, e21856 (2017).
Article PubMed PubMed Central Google Scholar
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Article CAS PubMed PubMed Central Google Scholar
Meers, M. P., Tenenbaum, D. & Henikoff, S. Peak calling by sparse enrichment analysis for CUT&RUN chromatin profiling. Epigenetics Chromatin 12, 42 (2019).
Article PubMed PubMed Central CAS Google Scholar
Ramírez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 44, W160–W165 (2016).
Article PubMed PubMed Central CAS Google Scholar
Robinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We are grateful to Dorothy C Bennett, Ghanem Ghanem, Florence Faure, and Meenhard Herlyn for providing cell lines. Hong Wu for providing Pten flox mice. We thank the Institut Curie staff responsible for the animal colony (especially P. Dubreuil), and the histology (S. Leboucher), FACS (C. Lasgi), and PICT-IBiSA imaging (C. Lovo) facilities. This work was supported by the Ligue nationale contre le cancer, INCa, ITMO Cancer, Fondation ARC (PGA), and is under the program «Investissements d’Avenir» launched by the French Government and implemented by ANR Labex CelTisPhyBio (ANR-11-LBX-0038 and ANR-10-IDEX-0001-02 PSL). M.H. had a fellowship from PSL and FRM, P.S. had a fellowship from INSERM, and M.L.C. had a fellowship from FRM. C.R.G. was supported by the Ludwig Institute for Cancer Research. RC was supported in part by a grant from the National Institutes of Health (AR062547, RAC) and a postdoctoral fellowship from the American Association for Anatomy (CK). E.S. was supported from the Research Fund of Iceland (184861-053). L.L. and E.S. are supported by a Jules Verne grant.

Author information

These authors contributed equally: M Hamm, P Sohier, V Petit.

Authors and Affiliations

Institut Curie, Université PSL, CNRS UMR3347, Inserm U1021, Normal and Pathological Development of Melanocytes, Orsay, France
Michael Hamm, Pierre Sohier, Valérie Petit, Jérémy H. Raymond, Véronique Delmas, Madeleine Le Coz, Franck Gesbert, Zackie Aktary, Marie Pouteaux, Florian Rambow, Nisamanee Charoenchon & Lionel Larue
Université Paris-Saclay, CNRS UMR3347, Inserm U1021, Signalisation radiobiologie et cancer, Orsay, France
Michael Hamm, Pierre Sohier, Valérie Petit, Jérémy H. Raymond, Véronique Delmas, Madeleine Le Coz, Franck Gesbert, Zackie Aktary, Marie Pouteaux, Florian Rambow, Nisamanee Charoenchon & Lionel Larue
Equipes Labellisées Ligue Contre le Cancer, Paris, France
Michael Hamm, Pierre Sohier, Valérie Petit, Jérémy H. Raymond, Véronique Delmas, Madeleine Le Coz, Franck Gesbert, Zackie Aktary, Marie Pouteaux, Florian Rambow, Nisamanee Charoenchon & Lionel Larue
Department of Anatomy and Cell biology, Carver College of Medicine, University of Iowa, Iowa City, IA, USA
Colin Kenny, Robert A. Cornell & Irwin Davidson
Laboratory of Genetic Instability and Oncogenesis, UMR8200 CNRS, Gustave Roussy, Université Paris-Sud, Villejuif, France
Alain Sarasin
Department of Pathobiology, Faculty of Science, Mahidol University, Bangkok, Thailand
Nisamanee Charoenchon
Cancer Epigenetics Program, Fox Chase Cancer Center, Philadelphia, PA, USA
Alfonso Bellacosa
Ludwig Institute for Cancer Research, Nuffield Department of Clinical Medicine, University of Oxford, Headington, Oxford, UK
Luis Sanchez-del-Campo, Laura Mosteo & Colin R. Goding
Department of Oncology, Clinical Sciences Lund, Lund University and Skåne University Hospital, Lund, Sweden
Martin Lauss & Göran B. Jönsson
Centre of Neuroregeneration, University of Edinburgh, Edinburgh, UK
Dies Meijer
Department of Biochemistry and Molecular Biology, and Department of Anatomy, BioMedical Center, Faculty of Medicine, University of Iceland, Reykjavik, Iceland
Eirikur Steingrimsson
Department of Functional Genomics and Cancer, Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/UNISTRA, 1 Rue Laurent Fries, 67404, Illkirch, Cedex, France
Irwin Davidson

Authors

Michael Hamm
View author publications
You can also search for this author in PubMed Google Scholar
Pierre Sohier
View author publications
You can also search for this author in PubMed Google Scholar
Valérie Petit
View author publications
You can also search for this author in PubMed Google Scholar
Jérémy H. Raymond
View author publications
You can also search for this author in PubMed Google Scholar
Véronique Delmas
View author publications
You can also search for this author in PubMed Google Scholar
Madeleine Le Coz
View author publications
You can also search for this author in PubMed Google Scholar
Franck Gesbert
View author publications
You can also search for this author in PubMed Google Scholar
Colin Kenny
View author publications
You can also search for this author in PubMed Google Scholar
Zackie Aktary
View author publications
You can also search for this author in PubMed Google Scholar
Marie Pouteaux
View author publications
You can also search for this author in PubMed Google Scholar
Florian Rambow
View author publications
You can also search for this author in PubMed Google Scholar
Alain Sarasin
View author publications
You can also search for this author in PubMed Google Scholar
Nisamanee Charoenchon
View author publications
You can also search for this author in PubMed Google Scholar
Alfonso Bellacosa
View author publications
You can also search for this author in PubMed Google Scholar
Luis Sanchez-del-Campo
View author publications
You can also search for this author in PubMed Google Scholar
Laura Mosteo
View author publications
You can also search for this author in PubMed Google Scholar
Martin Lauss
View author publications
You can also search for this author in PubMed Google Scholar
Dies Meijer
View author publications
You can also search for this author in PubMed Google Scholar
Eirikur Steingrimsson
View author publications
You can also search for this author in PubMed Google Scholar
Göran B. Jönsson
View author publications
You can also search for this author in PubMed Google Scholar
Robert A. Cornell
View author publications
You can also search for this author in PubMed Google Scholar
Irwin Davidson
View author publications
You can also search for this author in PubMed Google Scholar
Colin R. Goding
View author publications
You can also search for this author in PubMed Google Scholar
Lionel Larue
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

M.H., P.S., V.P., J.R., V.D., M.L.C., C.K., M.P., F.R., N.C., A.S., L.S.C., L.M., M.L., and G.B.J. conducted the experiments M.H., P.S., R.A.C, C.R.G. and L.L. designed the experiments M.H., P.S., V.P., F.G., V.D., Z.A., A.S., A.B., I.D., C.R.G., and L.L. wrote and reviewed the manuscript C.R.G., and L.L. secured the funding D.M. and E.S. provided reagents I.D., A.S., C.R.G. provided expertise and feedback

Corresponding authors

Correspondence to Colin R. Goding or Lionel Larue.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks Rugang Zhang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

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

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, 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 license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Hamm, M., Sohier, P., Petit, V. et al. BRN2 is a non-canonical melanoma tumor-suppressor. Nat Commun 12, 3707 (2021). https://doi.org/10.1038/s41467-021-23973-5

Download citation

Received: 24 April 2019
Accepted: 27 May 2021
Published: 17 June 2021
DOI: https://doi.org/10.1038/s41467-021-23973-5

This article is cited by

In silico development and experimental validation of a novel 7-gene signature based on PI3K pathway-related genes in bladder cancer
- Linhui Wang
- Yutao Wang
- Jianbin Bi
Functional & Integrative Genomics (2022)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.