Supplementary Table 2 from Metastasizing Melanoma Formation Caused by Expression of Activated N-Ras^Q61K on an INK4a-Deficient Background

Supplementary Methods from Pten Deficiency in Melanocytes Results in Resistance to Hair Graying and Susceptibility to Carcinogen-Induced Melanomagenesis

Supplementary Table 2 from Metastasizing Melanoma Formation Caused by Expression of Activated N-RasQ61K on an INK4a-Deficient Background

Metastatic melanoma is hallmarked by its ability of phenotype switching to more slowly proliferating, but highly invasive cells. Here, we tested the impact of signal transducer and activator of transcription 3 (STAT3) on melanoma progression in association with melanocyte inducing transcription factor (MITF) expression levels. We established a mouse melanoma model for deleting Stat3 in melanocytes with specific expression of human hyperactive NRASQ61K in an Ink4a-deficient background, two frequent driver mutations in human melanoma. Mice devoid of Stat3 showed early disease onset with higher proliferation in primary tumors, but displayed significantly diminished lung, brain, and liver metastases. Whole-genome expression profiling of tumor-derived cells also showed a reduced invasion phenotype, which was further corroborated by 3D melanoma model analysis. Notably, loss or knockdown of STAT3 in mouse or human cells resulted in the upregulation of MITF and induction of cell proliferation. Mechanistically we show that STAT3-induced CAAT Box Enhancer Binding Protein (CEBP) expression was sufficient to suppress MITF transcription. Epigenetic analysis by ATAC-seq confirmed that CEBPa/b binding to the MITF enhancer region silenced the MITF locus. Finally, by classification of patient-derived melanoma samples, we show that STAT3 and MITF act antagonistically and hence contribute differentially to melanoma progression. We conclude that STAT3 is a driver of the metastatic process in melanoma and able to antagonize MITF via direct induction of CEBP family member transcription.

Although c-Myc is essential for melanocyte development, its role in cutaneous melanoma, the most aggressive skin cancer, is only partly understood. Here we used the NrasQ61KINK4a-/- mouse melanoma model to show that c-Myc is essential for tumor initiation, maintenance, and metastasis. c-Myc-expressing melanoma cells were preferentially found at metastatic sites, correlated with increased tumor aggressiveness and high tumor initiation potential. Abrogation of c-Myc caused apoptosis in primary murine and human melanoma cells. Mechanistically, c-Myc-positive melanoma cells activated and became dependent on the metabolic energy sensor AMP-activated protein kinase (AMPK), a metabolic checkpoint kinase that plays an important role in energy and redox homeostasis under stress conditions. AMPK pathway inhibition caused apoptosis of c-Myc-expressing melanoma cells, while AMPK activation protected against cell death of c-Myc-depleted melanoma cells through suppression of oxidative stress. Furthermore, TCGA database analysis of early-stage human melanoma samples revealed an inverse correlation between C-MYC and patient survival, suggesting that C-MYC expression levels could serve as a prognostic marker for early-stage disease.

Cutaneous metastatic melanoma remains one of the deadliest forms of cancer despite recent progress with treatments. New treatment options include (i) immune checkpoint inhibitors (anti-PD1, anti-PDL1 and anti-CTLA4) which are effective for long term but for a low percentage of patients, and (ii) targeted therapy against known tumour molecular abnormalities, such as kinase inhibitors. In cutaneous melanoma, four main classes of driver mutations have been identified after genetic analysis: "BRAF" (mutated in 50% of patients), "NRAS" (mutated in 25% of patients), "NF1" (mutated in 10% of patients), and "wild-type". The most common BRAF mutation is located on V600, where kinase inhibitors, like Vemurafenib® are the most active. Unfortunately, the clinical effects of BRAFV600E kinase inhibitors cannot persist long enough owing to the resistance of metastatic melanoma. To date, there is no efficient targeted therapy for "NRAS", "NF1" and "other" melanomas. The aim of our work was to establish in vivo and in vitro models, which are relevant for cutaneous melanomas mutated for NRAS, to test and discover novel combined therapies for these patients. For this purpose, we established NRAS melanocyte cell lines from NRAS (Q61K) mice and NRAS melanoma cell lines from NRAS (Q61K) murine tumours, on a pure C57BL/6 background. These melanoma cell lines can be re-implanted in syngeneic C57BL/6 mice and give rise to new tumour in an immuno-competent environment. We will present results showing in vitro and in vivo cellular and molecular characterisation of these cell lines and their reactions to different therapies. These cellular tools will be useful to study melanoma biology, and to evaluate new therapeutic approaches.

NRAS and its effector BRAF are frequently mutated in melanoma. Paradoxically, CRAF but not BRAF was shown to be critical for various RAS-driven cancers, raising the question of the role of RAF proteins in NRAS-induced melanoma. Here, using conditional ablation of Raf genes in NRAS-induced mouse melanoma models, we investigate their contribution in tumour progression, from the onset of benign tumours to malignant tumour maintenance. We show that BRAF expression is required for ERK activation and nevi development, demonstrating a critical role in the early stages of NRAS-driven melanoma. After melanoma formation, single Braf or Craf ablation is not sufficient to block tumour growth, showing redundant functions for RAF kinases. Finally, proliferation of resistant cells emerging in the absence of BRAF and CRAF remains dependent on ARAF-mediated ERK activation. These results reveal specific and compensatory functions for BRAF and CRAF and highlight an addiction to RAF signalling in NRAS-driven melanoma.

The role of the Pax3 gene in embryonic development of pigment cells is well characterized. By contrast, the function of Pax3 in melanoma development is controversial. Indeed, data obtained from cultured cells suggest that PAX3 may contribute to melanomagenesis. PAX3 is found to be overexpressed in melanomas and also in nevi compared with normal skin samples. Pax3 homozygous loss of function is embryonic lethal. To assess the role of Pax3 in melanoma development in vivo, we analyzed Pax3 haploinsufficiency in a mouse model of melanoma predisposition. The Pax3(GFP/+) knock-in reporter system was combined with the Tyr::NRAS(Q61K); Cdkn2a(-/-) mouse melanoma model. Melanoma development was followed over 18 months. Histopathological, immunohistochemical, and molecular analyses of lesions at different stages of melanoma progression were carried out. Fluorescence-activated cell sorting on GFP of cells from primary or metastatic melanoma was followed by ex-vivo transformation tests and in-vivo passaging. We report here that Tyr::NRAS(Q61K); Cdkn2a(-/-); Pax3(GFP/+) mice developed metastasizing melanoma as their Tyr::NRAS(Q61K); Cdkn2a(-/-) littermates. Histopathology showed no differences between the two genotypes, although Pax3 mRNA and PAX3 protein levels in Pax3(GFP/+) lesions were reduced by half. The Pax3(GFP) allele proved to be a convenient marker to identify and directly sort heterogeneous populations of melanoma cells within the tumor bulk at each stage of melanoma progression. This new mouse model represents an accurate and reproducible means for identifying melanoma cells in vivo to study the mechanisms of melanoma development.

Figure S1. Normal skin of TyrNRas/Ing1−/− mice, and melanoma incidence and percentage of lung or lymph node metastasis in TyrNRas/Ink4a/Arf−/− mice. Figure S2. Macroscopic appearance of melanocytic lesions in Ing1−/− mice and TRP2 immunohistochemistry of locally invasive TyrNRas/Ing1−/− lesions. TRP2 and Ki-67 staining of representative TyrNRas/wt and TyrNRas/Ing1+/− melanomas. Figure S3. Specificity of p19Arf IHC in melanin-rich melanomas. Figure S4. Western blot of p19Arf and p53 in melanomas of different Ing1 genotypes. Figure S5. Representative Western blot and quantification of p33ING1 protein in human melanoma cell lines. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

The ciliary body and iris are pigmented epithelial structures in the anterior eye segment that function to maintain correct intra-ocular pressure and regulate exposure of the internal eye structures to light, respectively. The cellular and molecular factors that mediate the development of the ciliary body and iris from the ocular pigmented epithelium remain to be fully elucidated. Here, we have investigated the role of Notch signaling during the development of the anterior pigmented epithelium by using genetic loss- and gain-of-function approaches. Loss of canonical Notch signaling results in normal iris development but absence of the ciliary body. This causes progressive hypotony and over time leads to phthisis bulbi, a condition characterized by shrinkage of the eye and loss of structure/function. Conversely, Notch gain-of-function results in aniridia and profound ciliary body hyperplasia, which causes ocular hypertension and glaucoma-like disease. Collectively, these data indicate that Notch signaling promotes ciliary body development at the expense of iris formation and reveals novel animal models of human ocular pathologies.

Background Patent ductus arteriosus is a life-threatening condition frequent in premature newborns but also present in some term infants. Current mouse models of this malformation generally lead to perinatal death, not reproducing the full phenotypic spectrum in humans, in whom genetic inheritance appears complex. The ductus arteriosus (DA), a temporary fetal vessel that bypasses the lungs by shunting the aortic arch to the pulmonary artery, is constituted by smooth muscle cells of distinct origins (SMC1 and SMC2) and many fewer melanocytes. To understand novel mechanisms preventing DA closure at birth, we evaluated the importance of cell fate specification in SMC that form the DA during embryonic development. Upon specific Tyr::Cre-driven activation of Wnt/β-catenin signaling at the time of cell fate specification, melanocytes replaced the SMC2 population of the DA, suggesting that SMC2 and melanocytes have a common precursor. The number of SMC1 in the DA remained similar to that in controls, but insufficient to allow full DA closure at birth. Thus, there was no cellular compensation by SMC1 for the loss of SMC2. Mice in which only melanocytes were genetically ablated after specification from their potential common precursor with SMC2, demonstrated that differentiated melanocytes themselves do not affect DA closure. Loss of the SMC2 population, independent of the presence of melanocytes, is therefore a cause of patent ductus arteriosus and premature death in the first months of life. Our results indicate that patent ductus arteriosus can result from the insufficient differentiation, proliferation, or contractility of a specific smooth muscle subpopulation that shares a common neural crest precursor with cardiovascular melanocytes.

In humans, cutaneous melanoma (CM) is the deadliest cutaneous cancer. In recent years, the identification of recurrent mutations in CM allowed partial understanding of the molecular pathogenesis of CM. Mutations in the proto-oncogene NRAS occur in 18% to 20% of CM and are particularly frequent in the nodular subtype (Lee et al., 2011). NRAS mutations mostly occur in codon 61 of exon 2 where a lysine replaces a glutamine (Q61K), leading to a constitutively activated NRAS protein that promotes both proliferation and survival of melanoma cells (Ellerhorst et al., 2011; Lee et al., 2011). Interestingly, NRAS mutations are associated with distinct clinical and pathological features such as worse prognosis and shorter melanoma-specific survival (Devitt et al., 2011). Inactivation of the CDKN2A locus, which encodes the two tumor suppressor proteins p16INK4A and p14ARF, is also frequently encountered in CM (Serrano et al., 1996) and most often occurs through deletion (Funk et al., 1998). Both mono- and bi-allelic deletions are found and are associated with shorter median survival (Grafstrom et al., 2005). The need for a reliable model of human CM led to the generation of transgenic mice combining both dominant-active NRAS targeted to the melanocyte lineage (Tyr::NRasQ61K) and partial or total deletion of the Cdkn2a locus. These Tyr::NRasQ61K; Cdkn2a−/+ and Tyr::NRasQ61K; Cdkn2a−/− mice spontaneously develop metastasizing CM (Ackermann et al., 2005). Over the years, this model gained wide popularity in preclinical studies investigating the links between causative mutations, tumor progression and response to new therapeutics. However, to date, histopathological description of this model remains scarce and the terminology quite inconsistent among authors. Accurate pathological description and classification of melanocytic lesions (ML) is essential to evaluate animal models of melanoma (Sellers and Ward, 2012; Walker et al., 2011). Here, we propose a detailed histopathological atlas and classification of melanocytic lesions encountered in Tyr::NRasQ61K; Cdkn2a−/− or Cdkn2a−/+ transgenic mice along with their main microscopic features, with a particular emphasis on cutaneous ML. Our data highlight the histopathological specificity of melanoma in comparison to the other ML encountered in this model. Comparison to human ML is briefly discussed. Melanocytic lesions in Tyr::NRasQ61K; Cdkn2a−/− or Cdkn2a−/+ mice were commonly observed in the skin, lymph nodes, brain, eyes, and lungs. Infrequent sites included liver, spleen, heart, harderian gland and epididymis. The most common lesions and their frequency in the 35 necropsied mice are summarized in Table S1. Types of lesions and incidence did not vary between sexes. Gross features are summarized in Figure S1. Primary cutaneous ML were invariably present in the skin and were observed on gross examination as early as 2 months of age (Figures 1 and 2; Table 1). At that time, cutaneous thickness progressively increased and tousled, disarranged hairs arose giving a dull appearance to the mouse fur (Figure S1a-c). The distribution of ML ranged from patchy, with well-individualized lesions, to continuous when ML were numerous and coalesced. The type, number, size, distribution and density of ML could vary among different samples from the same animal. Overall, microscopic small cutaneous ML occurred all over the body, while no gross lesion was found on ventral skin. Usually < 200 μm (not grossly visible) May be composed of only few melanocytes Usually > 200 μm (usually grossly visible) May coalesce horizontally and form plaques Usually upper and mid dermis Large lesions may occupy entire dermis Ill-defined contours that blend in the dermis Stellate shape Usually well-defined contours Roughly nodular or piriform shape Usually ill-defined contours Rarely, cells invade the epidermis Forms nodule or plaques Usually well-defined contours Nodular shape Spindle High NCR Spindle Moderate NCR Spindle to epithelioid Moderate to low NCR Spindle to epithelioid Moderate to low NCR Small with fine/dusty chromatin Small/absent nucleolus Round with fine chromatin Small nucleolus Round with clear chromatin Round and large nucleolus Round with clear chromatin Round and large nucleolus Ulceration is rare Necrosis is absent Microscopically, ML encompassed a morphological spectrum in which we identified four main types of ML. In such a spectrum, a type was defined only when significantly distinctive morphological features were observed and could be relevant for the model. Occasionally, some lesions could not be assigned to a single type due to intermediate features. In terms of clinical behavior, these four types of ML also formed a spectrum ranging from benign lesions (nevus type A and B) to malignant lesions (melanoma) with an intermediate category of melanocytic lesions of unknown malignant potential (atypical nevus). Melanin-containing macrophages, called melanophages afterwards, were present in the dermis and the hypodermis and were associated with all types of ML. They were distinguished from melanocytes by their larger size and round to polygonal shape. When melanophagic infiltration was severe, ML were difficult to individualize. The intensity of melanophagic infiltration could vary among different samples from the same animal. ML-A (classical blue nevus) (Figure 1A-F) were the predominant type of ML. Because of their small size (usually less than 200 μm) they were not grossly discernible. They were observed in mice from 21 days of age (Figure 1A-C). The smallest ML-A were composed of only few melanocytes. ML-A occurred in the dermis, usually in the vicinity of hair follicles. They were frequently stellate shaped due to ill-defined contours that blended in the surrounding dermis. Their cellularity was moderate as melanocytes were separated by collagen fibers. Melanocytes were small, spindle-shaped, heavily pigmented and had little cytoplasm (high nucleo-cytoplasmic ratio (NCR)). Their nucleus was small and poorly discernible due to abundant pigmentation. When visible, it harbored a fine and dusty chromatin and usually no nucleolus. Mitoses were absent. These morphological features supported a benign behavior (nevi). ML-B (cellular blue nevus) (Figure 1G-I) were usually larger than ML-A but remained grossly indistinguishable. They were detected from 3 months onwards. They were well-demarcated, nodular to piriform and highly cellular lesions with densely packed melanocytes. Melanocytes were spindle-shaped and well pigmented but had more cytoplasm (moderate NCR) than in ML-A. Some epithelioid cells might be present. The nucleus, usually discernible, was round with a fine chromatin and a small nucleolus. Small capillaries could be present. These morphological features supported a benign behavior (nevi). ML-C (atypical blue nevus) (Figure 2A-C) were usually grossly discernible and greater than 200 μm but very small ML-C also occurred. Grossly, they appeared as small black nodules or plaques. They mainly occurred in the upper dermis but large lesions extended in the deeper dermis. They were apparent in 5.5-month-old mice. They usually had ill-defined contours and frequently coalesced horizontally to form large plaques that lifted the epidermis. The overlying epidermis could be hyperplastic and form deep rete ridge. Epidermal invasion was rare. ML-C had high cellularity with densely packed melanocytes. Melanocytes were plump, spindle-shaped to epithelioid and had an abundant acidophilic cytoplasm (moderate to low NCR) that tended to be poorly pigmented in upper-dermis cells. The nucleus was round, large and frequently had a large acidophilic nucleolus. Atypias were moderate, and mitoses were rarely present (usually less than 1 mitosis per 10 high power fields (HPF) or per lesion). Small vessels were usually well discernible. Ulceration was rare. The malignant potential of ML-C was equivocal and difficult to determine based solely on morphological features. ML-D (melanoma) (Figure 2D-I) were the less frequent but also the largest lesions. They first appeared in 7.5-month-old mice. They appeared as black, alopecic and frequently ulcerated nodules. They mostly arose from furry part of the skin and rarely on genitalia. They usually formed well-demarcated hypodermal nodules that occasionally encroached on the overlying dermis. These lesions were highly cellular with densely packed melanocytes that arranged in bundles or nests and frequently formed perivascular pseudorosettes. Artifactual clefts were frequently observed. Melanocytes were spindle to epithelioid with a moderate to abundant acidophilic cytoplasm (low to moderate NCR). Pigmentation was variable with the majority of lesions being poorly pigmented. Immunofluorescence for Melan-A and PEP-1 differentiation antigens performed on non-pigmented lesions or areas showed positive labeling confirming their melanocytic nature (Figure S2). The nucleus was round with a clear chromatin and an acidophilic nucleolus. Atypias were moderate to high and mitoses were commonly observed. Although their frequency was variable between lesions, ML-D usually displayed more than 5 mitoses per 10 HPF. More specifically, the presence of more than 1 mitosis in a HPF was virtually restricted to ML-D (Figure 2I). Necrosis and ulceration were frequent. Vascular invasion or emboli were almost never observed. These morphologic features were typical of a malignant melanocytic tumor (melanoma). Table 1 summarizes histological features to be used for the diagnosis of cutaneous ML in Tyr::NRasQ61K; Cdkn2a−/−. Because lymph nodes are the first metastatic site for cutaneous melanoma, they were systematically sampled (Figure 3 and Figure S1d-f). In young animals, some pigmented cells were observed and interpreted as interdigitating dendritic cells (Figure 3A-C). With age, lymph nodes progressively enlarged and became infiltrated by melanophages, resembling dermatopathic lymphadenopathy in humans (Figure 3D-F) (Ioachim, 2009). In mice diagnosed with cutaneous melanoma (ML-D), metastases could be observed in satellite lymph nodes (Figure 3G-L). Melanoma cells could be found in the lymph node subcapsular sinus, particularly in early metastases (Figure 3G, I). ML resembling cutaneous ML, ranging from ML-A to ML-D, were observed in lymph node (Figure 3E,F). Surprisingly, nearly 35% of these nodal ML occurred in the absence of concurrent cutaneous melanoma (ML-D). In these cases, melanocytic cells were hardly detected in the subcapsular sinus. In young animals, melanocytes were observed in the leptomeninges, particularly in the rostral region of the brain, a well-known site for meningeal melanosis in mice (Figures S1 g-i and S3). From 7.5 months onwards, some animals developed cerebral ML. Grossly, they appeared as a unique ill-demarcated pigmented lesion in the rostral region that eventually distorted the brain due to a mass effect. On histopathological examination, these lesions were diagnosed as melanoma and closely resembled cutaneous melanoma (ML-D). Some pigmented areas and nodules were occasionally observed on the corresponding region on the calvarium and revealed to be extensions from cerebral melanoma. Animals with cerebral melanoma did not systematically have concurrent cutaneous melanoma. The lesions were considered to be primary cerebral melanoma. In young animals, melanocytes were observed in the choroid layers of the eye where they formed a thin and regular layer (Figure S4). From 3 months onwards, some animals developed ocular ML resembling cutaneous ML with lesions ranging from ML-A to ML-D. ML-D were diagnosed as ocular melanomas and occurred from 7.5 months onwards. Animals with ocular melanoma did not systematically have concurrent cutaneous melanoma. The lesions were considered to be primary ocular melanocytic lesions. After lymph nodes, lungs are considered to be the most common metastatic site for melanoma (Figures S1j-l and S5). On gross examination, melanoma metastases typically presented as multiple black spots (Figure S1j-l). On histopathological examination, metastases were frequently observed in the lungs and were typically associated with vessels. In one case, despite the absence of cutaneous gross features suggestive of malignancy, lung metastases were detected and histopathological analysis of the skin confirmed the presence of cutaneous ML-C/ML-D lesions. The liver and the heart were infrequent sites for metastases with only 3 and 1 cases respectively. Melanocytic lesions were observed in the spleen (2 cases), the harderian gland (1 case), and the epididymis (1 case) (Figure S6). Splenic and harderian lesions exhibited malignant features favoring a diagnosis of melanoma whereas the epididymal lesion had a benign aspect and resembled cutaneous ML-B. Although a metastatic nature could not been ruled out as concurrent cutaneous melanomas were observed in these mice, these lesions were considered to be primary rather than metastatic. In Tyr::NRasQ61K; Cdkn2a−/− mice, a large spectrum of ML ranging from benign to malignant was observed in the skin. Four main types of lesions were identified and named as ML-A, ML-B, ML-C and ML-D, in a putative increasing order of malignancy. Indeed, ML-A and ML-B had a benign appearance and were considered to be the equivalent of the so-called melanocytic hyperplasia and nevi described previously (Ackermann et al., 2005). On the other hand, ML-D had typical malignant features and were considered to be equivalent to melanoma. ML-C appeared to be morphologically intermediate between nevi and melanoma, and were therefore difficult to categorize as benign or malignant. It is crucial to precisely identify this ML-C to correctly interpret changes in modified Tyr::NRasQ61K; Cdkn2a−/− mice. From ML-A to ML-D, there was a morphological progression characterized by both increase in size of lesions and cells, change in shape of cells from spindle to plump and epithelioid, increased atypias, mitotic index and vascularization, and a trend towards a loss of pigmentation. Furthermore, a chronological progression was also observed with ML-A being the first lesion to appear in young animals and ML-D being the ultimate lesion to develop in adults. We also noticed that the frequency of lesions decreased from ML-A to ML-D. All in all, these data strongly suggest that the sequence from ML-A to ML-D not only represents a morphological and chronological spectrum but also the multistep sequence of progression of ML in the Tyr::NRasQ61K mice. This is further supported by the presence of lesions with intermediate features that could not be assigned to a definitive type (transition forms). Comparison with human pathology revealed that ML in this model had similar features to the blue nevus family although exact equivalence, in terms of pathogenesis, morphology and clinical behavior may not be complete. In this analogy, ML-A, ML-B, ML-C and ML-D would correspond to classical blue nevus, epithelioid blue nevus, atypical blue nevus and malignant blue nevus respectively. One of the most common locations of these tumors in humans is the scalp (Zembowicz and Phadke, 2011), reminiscent of fur in animals. Interestingly, atypical blue nevus is a particular entity for which biological behavior is difficult to predict and is part of lesions often referred to as melanocytic tumors of uncertain malignant potential (MELTUMP) (Barnhill and Gupta, 2009; Murali et al., 2009; Phadke and Zembowicz, 2011; Zembowicz and Phadke, 2011). In Tyr::CreERT2; Braf +/V600E and Tyr::CreERT2; KrasG12V mice models, blue nevi were also described (Dhomen et al., 2009; Milagre et al., 2010). As Tyr::NRasQ61K mice, these mice present skin hyperpigmentation in a C57BL/6 background. The tendency of these mice to develop blue nevi type lesions could be due to the predominance of dermal melanocytes. Here, we extend the comparison and provide evidence to link the lesions of Tyr::NRasQ61K; Cdkn2a−/− mice to the blue nevi family of lesions in humans. Because Tyr::NRasQ61K mice harbor numerous nevi with NRas mutations that can progress to melanoma, they were recently reported to be a model for giant congenital nevi (CGN) (Shakhova et al., 2012). Although this comparison is interesting because patients with CGN can develop neuromelanosis (Alikhan et al., 2012) which is reminiscent of the primary cerebral ML observed in these mice, our histological analysis showed that nevi in CGN are morphologically distinct from the ML observed in Tyr::NRasQ61K mice. Mice lacking Cdkn2a are reported to develop sarcomas and lymphomas with a rate of 70%, which may cause diagnostic difficulties (Serrano et al., 1996). In our colony however, only 4 out of 26 (15%) Tyr::NRasQ61K; Cdkn2a−/− mice were diagnosed with non-melanocytic tumors (Data S1) suggesting that additional genes may cooperate with Cdkn2a for tumor suppression (Takeuchi et al., 2010). Non-melanocytic tumors can be differentiated from ML by their morphological and immunohistochemical characteristics. The distinction is particularly important with ML-D that tend to be poorly pigmented. Unexpected results came from the histopathological examination of lymph nodes. Besides typical melanoma metastasis, lesions resembling cutaneous ML and ranging from ML-A-like to ML-D-like were observed in some lymph nodes without associated cutaneous melanoma. We propose different hypotheses to explain these observations: (i) primary nodal ML could develop in the Tyr::NRasQ61K model as for other extra cutaneous site (see below); (ii) other ML than ML-D, especially ML-C, could have malignant potential and produce metastasis; (iii) a combination of these hypotheses. Primary cerebral melanomas were diagnosed. A metastatic nature is considered unlikely because (i) lesions invariably developed in the rostral brain, a typical site for meningeal melanosis in C57BL/6 mice; this specific distribution was previously reported (Lindsay et al., 2011); (ii) unlike metastases, lesions were large and unique rather than small and multifocal; (iii) unlike metastases, lesions appeared to originate from the leptomeninges rather than from vessels at the interface between white and grey matters (Brat et al., 1999; Das et al., 2010; Kusters-Vandevelde et al., 2010); 4) lesions were observed in animals without cutaneous melanoma. Similarly, primary ocular ML, including nevi and melanomas were frequently diagnosed and originated from the choroid where melanocytes normally exist. Infrequently, ML were diagnosed in the spleen, the harderian gland and the epididymis in animals with concurrent cutaneous melanomas (ML-D). Interestingly, a potential primary nature of the lesions could not be ruled out since ectopic melanocytes are present in C57BL/6 mice in the spleen, the heart or the harderian gland (Percy and Barthold, 2007; Yajima and Larue, 2008) as in humans (Plonka et al., 2009). The results suggest that, because extra-cutaneous melanocytes are common in C57BL/6 mice and because all melanocytes harbor the NRasQ61K activating mutation in the Tyr::NRasQ61K mice, primary ML including melanomas can develop in non-cutaneous sites. Although it is unclear whether ectopic melanocytes are present or not in lymph nodes, we believe that the existence of primary nodal ML should be considered in the Tyr::NRasQ61K model. Interestingly, intracapsular primary nodal melanocytic nevi have been described in humans, though they were not reported to progress to melanoma as we observed in the Tyr::NRasQ61K model (Dohse and Ferringer, 2010; Holt et al., 2004; Misago et al., 2008). Accurate histopathological phenotyping of mice models is a challenge (Sellers and Ward, 2012). Accordingly, the Tyr::NRasQ61K; Cdkn2a−/− model appeared to be much more complex than expected. With at least four types of cutaneous ML identified, this model offers opportunity to study the steps of melanomagenesis. The presence of extra-cutaneous melanocytes harbouring the same mutations as cutaneous melanocytes is responsible for the development of primary extra-cutaneous ML that co-exist with primary cutaneous ML and metastases. This feature should not be underestimated and should prompt to practice thorough necropsies. As a typical site for melanophagic infiltration and metastases, and a putative site for primary ML, lymph nodes appear to be the most challenging non-cutaneous site to evaluate. Histopathological evaluation of lymph node should be systematic and a metastatic status should not be assigned based solely on gross features. We hope that the proposed classification and atlas will be the basis for a common language between pathologists and researchers of the melanoma community working with Tyr::NRASQ61K; Cdkn2a−/− transgenic mice. The present classification is intended to evolve with the improvement of the comprehension of the molecular mechanisms of melanomagenesis. We thank C. Koënen for mice care, as well as A. Champeix, R. Nkosi and P. Wattier for histological technical assistance, M. Serrano for providing Cdkn2a−/+ mice, L. Larue for the backcrosses in C57BL/6J background and V. Hearing for providing the Tyrp1 (PEP-1) antibody. This work was supported by Institut National de la Recherche Agronomique, Agence Nationale de la Recherche Emergence Bio and Association pour la Recherche contre le Cancer grants. CC was granted by Allocation de Recherche MENRT from the French Ministry of Research (2009–2012). Table S1. Frequency of the most common ML encountered in Tyr::NRasQ61K; Cdkn2a−/− transgenic mice. Data S1. Materials and methods, supplementary data and figure legends. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Autophagy is linked to cell death, yet the associated mechanisms are largely undercharacterized. We discovered that melanoma, which is generally resistant to drug-induced apoptosis, can undergo autophagic cell death with the participation of orphan nuclear receptor TR3. A sequence of molecular events leading to cellular demise is launched by a specific chemical compound, 1-(3,4,5-trihydroxyphenyl)nonan-1-one, newly acquired from screening a library of TR3-targeting compounds. The autophagic cascade comprises TR3 translocation to mitochondria through interaction with the mitochondrial outer membrane protein Nix, crossing into the mitochondrial inner membrane through Tom40 and Tom70 channel proteins, dissipation of mitochondrial membrane potential by the permeability transition pore complex ANT1-VDAC1 and induction of autophagy. This process leads to excessive mitochondria clearance and irreversible cell death. It implicates a new approach to melanoma therapy through activation of a mitochondrial signaling pathway that integrates a nuclear receptor with autophagy for cell death.

Reference EPFL-CONF-189492View record in Web of Science Record created on 2013-10-01, modified on 2016-08-09

The serine protease CAP1/Prss8 is crucial for skin barrier function, lung alveolar fluid clearance and has been unveiled as diagnostic marker for specific cancer types. Here, we show that a constitutive knockout of CAP1/Prss8 leads to embryonic lethality. These embryos presented no specific defects, but it is during this period, and in particular at E13.5, that wildtype placentas show an increased expression of CAP1/Prss8, thus suggesting a placental defect in the knockout situation. The placentas of knockout embryos exhibited significantly reduced vascular development and incomplete cellular maturation. In contrary, epiblast-specific deletion of CAP1/Prss8 allowed development until birth. These CAP1/Prss8-deficient newborns presented abnormal epidermis, and died soon after birth due to impaired skin function. We thus conclude that a late placental insufficiency might be the primary cause of embryonic lethality in CAP1/Prss8 knockouts. This study highlights a novel and crucial role for CAP1/Prss8 in placental development and function.

Melanoma antigen recognized by T cells 1 (MART-1) is a melanoma-specific antigen, which has been thoroughly studied in the context of immunotherapy against malignant melanoma and which is found only in the pigment cell lineage. However, its exact function and involvement in pigmentation is not clearly understood. Melanoma antigen recognized by T cells 1 has been shown to interact with the melanosomal proteins Pmel17 and OA1. To understand the function of MART-1 in pigmentation, we developed a new knockout mouse model. Mice deficient in MART-1 are viable, but loss of MART-1 leads to a coat color phenotype, with a reduction in total melanin content of the skin and hair. Lack of MART-1 did not affect localization of melanocyte-specific proteins nor maturation of Pmel17. Melanosomes of hair follicle melanocytes in MART-1 knockout mice displayed morphological abnormalities, which were exclusive to stage III and IV melanosomes. In conclusion, our results suggest that MART-1 is a pigmentation gene that is required for melanosome biogenesis and/or maintenance.