Claudine Geffrotin

Antifibrotic action of Cu/Zn SOD is mediated by TGF-β1 repression and phenotypic reversion of myofibroblasts

Skin fibrosis is characterized by the proliferation and accumulation of activated fibroblasts called myofibroblasts. They exhibit specific cytoskeletal differentiation, overexpress the fibrogenic cytokine TGF-beta1, synthesize excess extracellular matrix compounds and exhibit a depleted antioxidant metabolism. Recently, SOD was successfully used as an antifibrotic agent in vivo, thus challenging the postulate of established fibrosis irreversibility. We postulated that myofibroblasts could be a direct target for this therapeutic effect. To test this hypothesis, we used three-dimensional co-culture models of skin, in which specific phenotypes of normal fibroblasts versus myofibroblasts are retained. These 3-D models were treated with liposomal and carrier-free Cu/Zn SOD, and examined for their effects on cell number, cell death, and phenotypic differentiation. The results show that SOD did not induce myofibroblast cell death, whereas it significantly reduced TGF-beta1 expression, thus demonstrating that SOD might be proposed as a potent antagonist of this major fibrogenic growth factor. We also found that SOD significantly lowered the levels of the myofibroblast marker alpha-sm actin, of beta-actin, and of the extracellular matrix components alpha1(I) collagen and tenascin-C. In conclusion, our results suggest that SOD antifibrotic action occurred in vitro through the reversion of myofibroblasts into normal fibroblasts.

Type I collagen expression contributes to angiogenesis and the development of deeply invasive cutaneous melanoma

Tumors are complex tissues composed of neoplastic cells, soluble and insoluble matrix components and stromal cells. Here we report that in melanoma, turn‐over of type I collagen (Col(I)), the predominant matrix protein in dermal stroma affects melanoma progression. Fibroblasts juxtaposed to melanoma cell nests within the papillary dermis display high levels of Col(I) mRNA expression. These nests are enveloped by collagen fibers. In contrast, melanoma‐associated fibroblasts within the reticular dermis express Col(I) mRNA at a level that is comparable to its expression in uninvolved dermis and reduced amount of collagen protein can be observed. To determine the significance of Col(I) expression in melanoma, we pharmacologically inhibited its transcription in a porcine cutaneous melanoma model by oral administration of halofuginone. When administered before melanoma development, it reduced melanoma incidence and diminished the transition from microinvasive toward deeply invasive growth by limiting the development of a tumor vasculature. Whereas invasive melanoma growth has been correlated with increased blood vessel density previously, our data for the first time demonstrate that the proangiogenic effect of Col(I) expression by fibroblasts and vascular cells precedes the development of invasive melanomas in a de novo tumor model.

Transcription analysis in the MeLiM swine model identifies RACK1 as a potential marker of malignancy for human melanocytic proliferation

BackgroundMetastatic melanoma is a severe disease. Few experimental animal models of metastatic melanoma exist. MeLiM minipigs exhibit spontaneous melanoma. Cutaneous and metastatic lesions are histologically similar to humans. However, most of them eventually spontaneously regress. Our purpose was to investigate whether the MeLiM model could reveal markers of malignancy in human melanocytic proliferations.ResultsWe compared the serial analysis of gene expression (SAGE) between normal pig skin melanocytes and melanoma cells from an early pulmonary metastasis of MeLiM minipigs. Tag identification revealed 55 regulated genes, including GNB2L1 which was found upregulated in the melanoma library. In situ hybridisation confirmed GNB2L1 overexpression in MeLiM melanocytic lesions. GNB2L1 encodes the adaptor protein RACK1, recently shown to influence melanoma cell lines tumorigenicity. We studied the expression of RACK1 by immunofluorescence and confocal microscopy in tissues specimens of normal skin, in cutaneous and metastatic melanoma developped in MeLiM minipigs and in human patients. In pig and human samples, the results were similar. RACK1 protein was not detected in normal epidermal melanocytes. By contrast, RACK1 signal was highly increased in the cytoplasm of all melanocytic cells of superficial spreading melanoma, recurrent dermal lesions and metastatic melanoma. RACK1 partially colocalised with activated PKCαβ. In pig metastases, additional nuclear RACK1 did not associate to BDNF expression. In human nevi, the RACK1 signal was low.ConclusionRACK1 overexpression detected in situ in human melanoma specimens characterized cutaneous and metastatic melanoma raising the possibility that RACK1 can be a potential marker of malignancy in human melanoma. The MeLiM strain provides a relevant model for exploring mechanisms of melanocytic malignant transformation in humans. This study may contribute to a better understanding of melanoma pathophysiology and to progress in diagnosis.

Opposite regulation of tenascin-C and tenascin-X in MeLiM swine heritable cutaneous malignant melanoma.

Interactions between tumour cells and surrounding extracellular matrix (ECM) influence the growth of tumour cells and their ability to metastasise. It is thus interesting to compare ECM composition in tumours and healthy tissues. Using the recently described MeLiM miniature pig model of heritable cutaneous malignant melanoma, we studied the expression of two ECM glycoproteins, the tenascin-C (TN-C) and tenascin-X (TN-X), in normal skin and melanoma. Using semiquantitative RT-PCR, we observed a 3.6-fold mean increase of TN-C RNAs in melanoma compared to normal skin. Both stromal and tumour cells synthesise TN-C. On the contrary, TN-X RNAs decreased 30-fold on average in melanoma. This opposite regulation of TN-C and TN-X RNAs was confirmed at the protein level by indirect immunofluorescence. Whereas pig normal skin displayed a discrete TN-C signal at the dermo-epidermal junction, around blood vessels and hair bulbs, the swine tumour showed enhanced expression of TN-C in these areas and around stromal and tumour cells. In contrast, normal skin showed a strong TN-X staining at the dermo-epidermal junction and in the dermis, whereas this signal almost completely disappeared in the tumour. The results presented here describe a dramatic alteration of the ECM composition in swine malignant melanoma which might have a large influence on tumourigenesis or invasion and metastasis of melanoma cells.

Identification of five chromosomal regions involved in predisposition to melanoma by genome-wide scan in the MeLiM swine model.

In human familial melanoma, 3 risk susceptibility genes are already known, CDKN2A, CDK4 and MC1R. However, various observations suggest that other melanoma susceptibility genes have not yet been identified. To search for new susceptibility loci, we used the MeLiM swine as an animal model of hereditary melanoma to perform a genome scan for linkage to melanoma. Founders of the affected MeLiM stock were crossed with each other and with healthy Duroc pigs, generating MeLiM, F1 and backcross families. As we had previously excluded the MeLiM CDKN2A gene, we paid special attention to CDK4 and MC1R, as well as to other candidates such as BRAF and the SLA complex, mapping them on the swine radiation hybrid map and/or isolating close microsatellite markers to introduce them into the genome scan. The results revealed, first, that swine melanoma was inherited as an autosomal dominant trait with incomplete penetrance, preferably in black animals. Second, 4 chromosomal regions potentially involved in melanoma susceptibility were identified on Sus Scrofa chromosomes (SSC) 1, 2, 7 and 8, respectively, in intervals 44–103, 1.9–18, 59–73 and 47–62 cM. A fifth region close to MC1R was revealed on SSC 6 by analyzing an individual marker located at position 7.5 cM. Lastly, CDK4 and BRAF were unlikely to be melanoma susceptibility genes in the MeLiM swine model. The 3 regions on SSC 1, 6 and 7, respectively, have counterparts on human chromosomes (HSA) 9p, 16q and 6p, harboring melanoma candidate loci. The 2 others, on SSC 2 and 8, have counterparts on HSA 11 and 4, which might therefore be of interest for human studies.

Identification of differentially expressed genes in spontaneously regressing melanoma using the MeLiM swine model.

Partial and some few cases of complete spontaneous regression have been observed in cutaneous melanoma patients but little is known about the molecular mechanisms involved. The Melanoblastoma‐bearing Libechov Minipig (MeLiM) is a suitable animal model to study the phenomenon of spontaneous regression because MeLiM pigs exhibit naturally occurring melanomas which regress completely 6 months after birth. In this study, we used suppression subtractive hybridization (SSH) to identify molecular determinants of melanoma regression within swine melanoma tissues and melanoma cell cultures. Several markers involved in cell‐adhesion, ‐communication, ‐motility, signal transduction, negative regulation of cell proliferation, transport and immune response were identified that correlated with melanoma regression whereas the main genes involved in melanin synthesis showed a strong downregulation. For the most differentially expressed genes, we validated the results obtained by SSH with qRT‐PCR and with immunohistochemistry for some of them (CD9, MITF, RARRES1). Most notable, for the first time in melanoma, we identified the retinoic acid responder 1 gene (RARRES1) as a main actor of the regression process in melanoma. This first gene expression study in swine melanoma regression, may contribute to the finding of new therapeutic targets for human melanoma treatment.

A new animal model for the imaging of melanoma: correlation of FDG PET with clinical outcome, macroscopic aspect and histological classification in Melanoblastoma-bearing Libechov Minipigs

The aim of this study was to evaluate the Melanoblastoma-bearing Libechov Minipigs (MeLiM) as an animal model of melanoma for in vivo imaging. Serial whole-body 2-deoxy-2-[18F]fluoro-d-glucose positron emission tomography (FDG PET) scans were conducted on five MeLiM. In order to explore different clinical stages of the tumoural lesions, each animal was scanned two to four times, at intervals of 30–155 days. PET images were analysed by a semiquantitative method based on the tumour to muscle metabolic ratio. Histology was performed on biopsies taken between or after the scans and the histological grading of the tumours was compared with the FDG uptake. The overall sensitivity of FDG PET for the detection of cutaneous melanoma was 75%; 62.5% of involved lymph nodes were positive. Sensitivity was better for tumours with vertical growth than for flat lesions. FDG PET did not detect tumours with epidermal involvement only, nor did it detect small metastatic foci. The metabolic ratio was correlated with the evolution of the melanoma. FDG PET is effective in the staging of cutaneous melanoma and the follow-up of tumoural extension and regression in Melanoblastoma-bearing Libechov Minipigs. The results obtained in this animal model correlate well with those described in human melanoma. Accordingly, this model may be useful in testing new tracers specific for melanoma and in helping to detect molecules expressed early during tumoural regression.

Swine cytosolic malic enzyme: cDNA cloning, sequencing, and localization

A highly significant genetic association has been found between some alleles of the swine Major Histocompatibility Complex SLA (Swine Leukocyte Antigen genetic complex) and the cytosolic malic enzymatic activity level in muscles. The aim of this study was to find out whether this genetic association was due to a close linkage of the SLA region and the gene coding for the enzyme. Since no swine cytosolic malic enzyme sequence (ME1) was available, we isolated several overlapping fragments that spanned the almost entire malic enzyme transcript both by screening of a swine cDNA library and by RT-PCR. The results indicated the existence of two transcripts of 2.0 and 3.1 kb, which probably correspond to two alternative forms of one gene. The sequence of the transcript was highly similar to the other published mammalian cytosolic NADP+-dependent malic enzyme cDNA, especially within the four functional domains. Two major bands at 3.7 and 2.4 kb were detected on Northern blots containing the RNA from 25 tissues from fetuses and adult pigs. A high expression level was found in the adrenal gland, muscle, liver, and peripheral nerves. The analysis of malic enzyme RFLPs in five SLA informative families revealed an independent segregation of the ME1 gene from the SLA region. In situ hybridization results localized the cytosolic malic enzyme on the swine Chromosome (Chr) lpl.2, except that the association between SLA and the malic enzyme activity level was due to a physical genetic linkage. Thus, the mechanisms underlying this association remain to be elucidated.

Comparative genomic hybridization analysis of hereditary swine cutaneous melanoma revealed loss of the swine 13q36-49 chromosomal region in the nodular melanoma subtype.

Genetic alterations implicated in malignant melanoma are still poorly understood. Malignant melanomas present highly variable histologic and cytologic patterns. The aim of the present study is to define genomic imbalances associated with the development of 2 histologic types of swine hereditary cutaneous melanoma. We have investigated 11 swine tumors by comparative genomic hybridization (CGH), 4 superficial spreading melanomas (SSMs) and 7 nodular melanomas (NMs). Following laser capture microdissection and degenerate oligonucleotide primed‐polymerase chain reaction, we were able to isolate and then amplify DNA from the 2 histologic subtypes. Consensus regions of chromosome gains were identified on both histologic subtypes, on swine chromosomes 3p13‐p17 (75% of the SSMs and 71% of the NMs), 12q (100% of the SSMs and 57% of the NMs) and 14q11‐q21 (75% of the SSMs and 42% of NMs). Chromosomal loss was restricted to NM lesions and the swine 13q36‐49 region was lost in 100% of the NMs. Interphase fluorescence in situ hybridization with a probe mapping to the 13q41‐q42 region indicates loss of the corresponding region on NM lesions. Taking into account this CGH analysis and the comparative genomic data between swine and human genomes, we suggest that a role for the human chromosomes 3p11‐qter and chromosome 21 losses should be investigated in human nodular melanoma progression.

Identification and mapping of swine cyclin-dependent kinase inhibitor CDKN2A and CDKN2B exon 2 sequences

We have isolated the swine homologs of human CDKN2A and CDKN2B exon 2 sequences. As in the human and mouse genomes, the exon 2 sequences of these two genes present a high level of sequence homology and are tightly linked. Using fluorescence in situ hybridization, we have mapped swine CDKN2A and CDKN2B to chromosome 1q25. This confirms the comparative mapping data among man, mouse, and swine, showing a conserved synteny among chromosome segments 9p21, 4C3-C6, and 1q25, respectively.