BRN2 is a non-canonical melanoma tumor-suppressor

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.

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

A Bar graph showing the status of the BRN2 locus in human skin cutaneous melanoma (SKCM) metastases (stage IV). Copy-number alterations (CNAs) were estimated using the GISTIC algorithm. Two alleles (2a in black), one allele (1a in red), no allele (0a in orange), and gain and/or amplification (G/Aa in blue) of the BRN2 locus are given. B Pictogram showing the extent of segmental deletions (red or orange vertical lines) that affect the BRN2 locus on Chr.6q16 (dashed blue horizontal line) in SKCM metastases. C Kaplan–Meier curves comparing 10-year overall survival of SKCM patients diploid for BRN2 (black line, n = 106) or those with mono-allelic (red n = 156). The TCGA CNA-data set was analyzed (n = 309). Diploid vs. Mono-allelic loss: log-rank (Mantel-cox) test (p = 0.067). Data were retrieved from TCGA on August 8, 2019. D Kaplan–Meier curves comparing melanoma patients with diploid status or mono-allelic loss of BRN2 in 108 regional metastatic melanoma patients (p = 0.001, log-rank test) and unpublished data. E Kaplan–Meier curves comparing 30-year overall survival of SKCM patients to BRN2 mRNA levels. Log-rank (Mantel-Cox) test (p = 0.03). Data were retrieved from TCGA on August 8, 2019. Significance was defined as * (p < 0.05) and *** (p ≤ 0.001).

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

A Macroscopic pictures of the dorsal view of mice with cutaneous melanomas carrying mutations in the melanocyte lineage for Braf, Pten, and Brn2 after tamoxifen induction at birth (p1, p2, and p3 – see Supplementary Fig. 3). Tyr::CreER^T2/°; Braf^V600E/+; Pten^+/+ (= Braf), Pten^F/+ (= Pten), Brn2^+/+ (=Brn2-WT), Brn2^F/+ (= Brn2-het), and Brn2^F/F ( = Brn2-hom). Tumors are highlighted with arrows and the sizes of the first growing tumors to appear are proportional to the diameters of the circles. F means floxed allele. B All Braf (n = 9), Braf-Brn2-het (n = 8), and Braf-Brn2-hom (n = 4) mice produced cutaneous melanomas and their number was similar (1 to 2 tumors/mouse). All Braf-Pten-Brn2-WT (n = 7), Braf-Pten-Brn2-het (n = 21), and Braf-Pten-Brn2-hom (n = 11) mice produced cutaneous melanomas. Note that in the absence of Pten (Pten^F/F), the appearance of the melanoma was too rapid to observe any difference between Brn2-WT, Brn2-het, and Brn2-hom mice. Each dot corresponds to an individual mouse. As control, mice of different genetic backgrounds were produced and not induced with tamoxifen Braf [n = 12], Braf-Brn2-het [n = 25], Braf-Brn2-hom [n = 11], Braf-Pten [n = 7], Braf-Pten-Brn2-het [n = 13], and Braf-Pten-Brn2-hom [n = 6]; none of them developed melanoma after 18 months, except one Braf-Pten-Brn2-het mouse that developed one melanoma after 12 months. None of the mice that were wild-type for Braf displayed any obvious phenotype, irrespective of the status of Pten or Brn2, including melanomagenesis and hyperpigmentation. C Growth rates of the first tumor appearing in each mouse for Braf-Brn2-WT, Braf-Brn2-het, Braf-Brn2-hom, Braf-Pten-Brn2-WT, Braf-Pten-Brn2-het, and Braf-Pten-het-Brn2-hom mice. The number of tumors is determined all along the life of the mouse by checking the mice a minimum of twice a week. Statistical analysis was performed using the two-tailed unpaired t test. ns = non-significant, *p < 0.05, **p < 0.01, and ***p < 0.001. Data are presented as mean values ± SEM. Braf-Pten-Brn2-het mice were euthanized in average 1.3 weeks after appearance of the first tumors with an average of 16 tumors/mouse. Similar results were obtained with Braf-Pten-Brn2-hom mice. Braf-Pten-Brn2-WT mice were euthanized at 4 weeks with an average of 8 tumors even though they did not reach a total volume of 2 cm³ except for one mouse that was euthanized earlier (3 weeks).

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

A–D Representative photomicrographs of Ki-67 (a) and BrdU (c) stainings of Braf-Pten-Brn2-WT/het/hom tumors. Ki-67⁺ cells are stained in red. Nuclei are stained in blue. Scale bar = 40 μm. Quantification of (n = 3) Ki-67+ (B) and BrdU (D) stainings of (A) and (C), respectively. Scale bar = 40 µm. Each dot represents the result for one tumor. (e) Growth rate is induced in Dauv-1 and Melan-a cell lines after reduction of Brn2 using siBRN2 and siScr as control (Scr = scramble). Three independent biological and technical experiments were performed for each cell line and for each condition. F–H Brn2 knock down induces Cyclin D1 protein but not its mRNA in melanocytes. F Western blot analysis for Brn2, Cyclin D1, and actin after reduction of Brn2 in Dauv-1 and Melan-a cells. Experiments were performed independently three times. One representative western blot is shown (raw data are presented in Supplementary Fig. 8). Quantification of protein (G) and mRNA (H) levels for Dauv-1 cells after siRNA-mediated knockdown (n = 3, independent experiments). For the proteins, all values were normalized against the background and corresponding actin loading control for each sample. Quantification was performed using Image-J software. For mRNA, all values were normalized against those of TBP. au = arbitrary units. Statistical analysis was performed using the two-tailed unpaired (B, D, E, H) and paired (G) t tests. ns = non-significant, *p < 0.05, **p < 0.01, and ***p < 0.001. Data are presented as mean values ± SEM.

Fig. 4. Mono-allelic loss of Brn2 induces melanoma metastasis.

A Upper panel: Representative photomicrographs of in situ inguinal lymph nodes (LN) of Braf-Pten-Brn2-WT/het/hom mice. Scale bar = 1 mm. The pigmented volume (mm³) was estimated for each LN. Lower panel: Representative photomicrographs of haematoxylin & eosin (H&E) staining of LNs containing pigmented cells. Scale bar = 20 µm. B Quantification of the pigmented volume of inguinal LNs in the upper panel of Figure (A). n = 3, 5, and 5 for WT, het, and hom. (c) Quantification of the pigmented areas per mm² of inguinal LNs in the lower panel of Figure (A). Pigmented areas > 50 µm² were considered. n = 3, 5, and 3 for WT, het, and hom. D Scheme showing the defloxing strategy of Brn2 in melanocytes of the primary tumor that releases eGFP expression upon the defloxing of Brn2. E Representative photomicrographs of serial LN sections of Braf-Pten-Brn2-WT and Braf-Pten-Brn2-het mice stained with H&E and the melanocyte marker Sox10. H&E staining was evaluated for one section and GFP (green channel) and Sox10 staining (red channel) evaluated for an adjacent section. Scale bar = 20 µm. F A melanoma invasive signature was significantly enriched in the Braf-Pten-Brn2-het tumors (left) and in the Braf-Pten-Brn2-het melanoma cell lines (right) compared to their Braf-Pten-Brn2-hom counterparts. G Left: photomicrographs of m82 and m82-BRN2 mouse melanoma cells embedded as spheroids in 600 μg/mL matrigel at t0 (H0) and 18 hrs after (H18). Right: Boxes and plots represent the area of invasion (red lines on photomicrographs) quantified with ImageJ (n = 54 for m82 cells and n = 56 for m82-BRN2 cells). P value < 0.0001. au = arbitrary unit. h Axl mRNA is significantly overexpressed in Braf-Pten-Brn2-het melanoma and melanoma cell lines (n = 10 and 2, respectively) compared to the Braf-Pten-Brn2-WT (n = 5 and 4, respectively) and Braf-Pten-Brn2-hom (n = 10 and 3, respectively) tumors. Statistical analysis was performed using the two-tailed unpaired t test for (B), (C), (G), and (H) (tumors) and an Anova test for H (cell lines). Data are presented as mean values ± SEM for (B) and (C), sd for (H), and Box and whiskers (median, min to max) for (G). ns = non-significant, *p < 0.05, ***p < 0.001, ****p < 0.0001.

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

A Representative photomicrographs of immunohistochemistry staining of Pten (red) in Braf-Pten-Brn2-WT and Braf-Pten-Brn2 mouse melanomas are shown. Scale bar = 40 µm. Three independent tumors for each genotype were used for these experiments and three independent sections were used for each tumor. A 2-way ANOVA with Dunnett’s multiple comparisons tests were performed. The percentage of Pten⁺ cells in WT and mutant tumors is shown. B Western blot analysis of Brn2, Pten and actin for Braf-Pten-WT and Braf-Pten-Brn2 from at least three tumors of each genotype. One representative example is presented, raw data are presented in Supplementary Fig. 8. The relative intensities of the band were estimated with ImageJ. C RT-qPCR of Brn2 and Pten from Braf-Pten-WT and Braf-Pten-Brn2 melanomas. Three independent mouse melanomas per genotype were analyzed. Data were normalized against the values of Gapdh. au = arbitrary unit. D Western blot analysis of BRN2, PTEN and ACTIN from Dauv-1 human melanoma cells and Melan-a mouse melanocytes after siRNA mediated knockdown. A representative western blot is shown, raw data are presented in Supplementary Fig. 8. Scr = Scramble. E RT-qPCR of BRN2 and PTEN from human melanoma cells (Dauv-1) and mouse melanocytes (Melan-a) after siRNA-mediated knockdown. Specific primers were used for human and mouse samples. Dauv-1 (n = 6), Melan-a (n = 4), independent experiments. Data were normalized against the values for TBP (Dauv-1) or Gapdh (Melan-a). F ChIP assays of BRN2 binding to the PTEN promoter in Dauv-1 melanoma cells. All data shown are representative of at least three independent assays. G Quantification of the ChIP-qPCR, plotted and normalized against IgG as the reference. au = arbitrary unit. n = 6, 3, 4, and 3 for BRN2 CDS, PAX3 prom, BS1 and BS2, respectively. H Scheme of the human PTEN promoter containing two BRN2 binding sites (BS) represented as colored circles. Note that BS are conserved between humans and mice. TSS = transcription start site. Exons (X) 1 and 2 are shown as horizontal rectangles. The translation start site (ATG) and the end of exon 1 are indicated. All numbering is relative to the TSS (+1). Representation of the reporter luciferase (luc) construct with or without PTEN promoter. I–N Human and mouse PTEN promoter activities were evaluated in human Dauv-1 (I, K), SK28 (J, L), and in mouse m82 [Brn2-hom], m59 [Brn2-het], and m50 [Brn2-WT] (M, N) melanoma cell lines either in the presence of siScr (scramble) or siBrn2 (smart-pool) (I, J, M) or in the presence of expression vector of BRN2 (CMV::BRN2) (L, N). The experiments were independently performed four (I, J, K) and three (L) times. They were performed three independent times for m82 and m50, seven for m59 (M), and seven times for m82, six for m59 and four for m50 (N). Statistical analysis was performed using the two-tailed unpaired t test for (C), (E), (G) and paired t test for (I–N). Data are presented as mean values ± SEM for (C), (E), (G), I–M Box and whiskers (median, min to max) for N. ns = non-significant, *p < 0.05, ***p < 0.001, and ****p < 0.0001.

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

A Western blot analysis of MITF and PTEN from human melanoma cells (501mel, SK28, and HBL) after siRNA-mediated knockdown. Actin was used as a loading control. A representative western blot is shown, raw data are presented in Supplementary Fig. 8. The molecular weight is indicated in kDa. Scr = Scrambled. B RT-qPCR of MITF and PTEN from human melanoma cells (501mel and SK28) after siRNA-MITF and Scr knockdown. All values were normalized against TBP. The analysis was performed on three independent experiments with technical triplicates, au = arbitrary unit. C Screenshot of IGV genome browser (GRCH37/hg19) visualization of MITF, H3K27Ac and H3K4me3 binding to the PTEN locus in SkMel28 cell lines that are MITF-WT or mutant (i.e., MITF-ΔX6 = ΔMITF). Blue boxes below MITF and H3K27Ac tracks: signal above IgG background (i.e., peaks) called by MACS2. PTEN and downstream regions are shown, blue arrows indicate strand orientation and horizontal rectangles the exons. Y-axes are scaled per antibody sample. Anti-MITF CUT&RUN peaks present in WT cells that harbor an M-Box binding motif are painted gray. Six distal enhancers (painted light blue) exhibited a twofold greater H3K27ac signal in MITF mutant cell lines compared to wild-type cell lines. At least two CUT&RUN biological replicates were performed for MITF, H3K27ac, and H3K4me3. D ChIP assays of MITF binding downstream of PTEN in 501mel human melanoma cells stably expressing HA-Tagged MITF (location + 140 kb). ChIP assays are performed using an antibody against HA and analyzed after a 30-cycle PCR (exponential phase). The tyrosinase promoter (TYR) and PRM1 were used as positive and negative controls, respectively. Input represents approximately 3% of the input used for the ChIP assay. H3 (histone H3) and IgG (Immunoglobulin G) were used as positive and negative technical controls for each region of interest, respectively. The oligonucleotides, their positions on the genome, and sizes of the amplified fragments are shown in Supplementary Tables 3 and 4. All data shown are representative of three independent assays. * corresponds to the remaining oligonucleotides. E Quantification of the ChIP–qPCR presented in (D) is plotted and normalized against IgG as a reference. au = arbitrary unit. PRM1 (−): PRM1 (negative control), TYR (+): tyrosinase promoter (positive control). Statistical analysis was performed using the two-tailed unpaired t test. Data are presented as mean values ± SD. ns = non-significant, *p < 0.05, **p < 0.01, and ***p < 0.001.