Deregulated expression of the transcription factor PAX3 was observed previously in several tumors like rhabdomyosarcoma and Ewing’s sarcoma. Because PAX3 expression is also found in pluripotent neural crest cells, we investigated whether melanomas, tumors derived mostly from cutaneous intraepidermal melanocytes, might show deregulated PAX3 expression. Using a specific and sensitive reverse transcription-PCR, we detected PAX3 mRNA in 77% (27 of 35) of primary cultured melanomas. These results could be confirmed by direct in situ hybridization on the corresponding tissue sections where PAX3 expression was unambiguously confined to tumor cells and not detected in surrounding normal tissue, normal skin sections, or sections of benign lesions. Furthermore, down-regulation of PAX3 expression achieved through a specific antisense oligonucleotide-based treatment resulted in >70% of dead cells specifically in PAX3-positive melanomas. Annexin V staining confirmed that primary melanoma cells underwent apoptosis after treatment. These experiments suggest that in situ hybridization of PAX3 on paraffin-embedded tissue may represent a novel means to identify melanoma cell lesions, which appear to become dependent on expression of this deregulated transcription factor.

Melanoma is a tumor with increasing incidence whereby advanced disease is associated with a poor prognosis and responds very poorly to treatment modalities including chemotherapy and immunotherapy(1). Despite the availability of a fair number of marker genes/proteins (e.g., melanA/MART-1, MAGE-3, and tyrosinase), there is a need to identify proteins involved in the conversion from benign to malignant disease and in tumor cell survival. These factors should have diagnostic or prognostic implications and might also be functionally important for tumor development.

PAX proteins are developmentally expressed transcription factors that play a fundamental role in the establishment of cell lineages. Their importance has been underscored by several loss-of-function mutations that usually lead to a lack of the specific structures or organs where a PAX protein is normally expressed (2). For example, PAX3 is expressed during normal development in specific areas of the neural tube, sensory organs, and the dermomyotome. A naturally occurring PAX3 loss-of-function mutation in mice (Splotch Sp) exhibits severe pigmentation defects and fails to establish hypaxial skeletal muscle cells (3). Homozygous affected mice die shortly after birth. Mutations in the human PAX3 gene cause Waardenburg syndrome, which is characterized by pigmentation abnormalities and hearing impairment attributable to the absence of melanocytes (4).

Apart from the normal physiological role during development, several PAX genes are reexpressed in malignant neoplasms(5), e.g.,PAX3 in rhabdomyosarcoma and in Ewing’s sarcoma (6), PAX5 in medulloblastoma and lymphomas, and PAX2 and PAX8in renal cell carcinoma and Wilms’ tumor. Because PAX3 is expressed during development in proliferating melanocyte precursor cells, it might be reexpressed in melanomas, analogous to the situation in rhabdomyosarcoma. In this report, we demonstrate that PAX3 is indeed deregulated in the majority of melanomas analyzed by RT-PCR3as well as in situ hybridization on paraffin-embedded tissue sections and contributes to tumor cell survival.

Cell Culture.

A total of 35 primary cultures established from melanoma lesions of patients in different stages and five melanoma cell lines were cultured as monolayers at 37°C and 5% CO2 in RPMI 1640(Life Technologies, Inc.) supplemented with 10% fetal bovine serum and 100 units/ml penicillin/100 μg/ml streptomycin as described previously (7).

Analysis of RNA Transcripts.

Total RNA extraction and RT-PCR reactions were performed as described previously (8). RNA integrity was checked by amplification of the ubiquitous FKHR cDNA. RT-PCR was also performed to monitor the presence of the melanoma-associated antigens MAGE-3, melanA, and tyrosinase mRNA, as described previously (7, 9).

In Situ Hybridization.

The phPAX3-sp6 plasmid containing 152 bp of human PAX3(10) was linearized with EcoRI to synthesize biotin-14-CTP-labeled antisense RNA (MAXIscript; Ambion, Inc.). In situ hybridization of paraffin sections was performed as detailed in the instruction manuals of the mRNAlocator-Hyb and mRNAlocator-biotin kit (Ambion, Inc.). Briefly, sections were hybridized overnight at 55°C, and the detection reaction was performed for 6 h at 37°C.

ODN Incubation.

ODN treatment was performed as reported earlier (11). The ODNs were synthesized as phosphothioate molecules by Microsynth(Balgach, Switzerland), purified over a reversed phase column and resuspended in TE (10 μm Tris-HCl, 1 mm EDTA,pH 7.5).

Apoptosis Assay.

An Annexin V-FITC kit (BenderMedSystems, Vienna, Austria) was used to assess apoptotic cells. Sixty-five h after ODN treatment, cells were washed with DMEM and stained for 10 min at room temperature with Annexin V-FITC according to the manufacturer’s instructions.

PAX3 Expression in Melanoma Cell Lines.

To test initially whether PAX3 mRNA is expressed in melanomas,several cell lines were screened by RT-PCR. Indeed, three of the five investigated cell lines were found to be positive for PAX3 expression, i.e., MEL15, G-361, and UKRV-Mel2 (Fig. 1 A). No PAX3 expression was detected in the other two melanoma-derived cell lines (A365 and Küng-A375). Hence, 60% of the investigated melanoma cell lines expressed PAX3 mRNA.

PAX3 Expression in Cultured Melanomas.

Because expression of PAX3 was found only in some of the melanoma cell lines analyzed, we next asked whether PAX3 expression might depend on a given tumor stage or is a property of established cell lines only. To this end, cultured primary melanomas established from different stages and locations from a cohort of patients were analyzed (7). From 35 melanoma cultures analyzed totally, 27 were found to express PAX3 representing 77% of all investigated tumors (Table 1). The highest prevalence of PAX3 expression was found in stage III tumors, where 11 of 12 (92%) of all cultures analyzed were PAX3 positive. In stage I, 9 of 11 (82%), and in stage II, 7 of 12 (58%),melanoma cultures expressed detectable levels of PAX3.

To investigate whether PAX3 expression changes with time of tumor progression and to test whether a correlation exists with other melanoma markers, cultured melanomas established from the same patients at different disease stages were analyzed (Table 2). In four of eight patients analyzed, an increase in PAX3 expression was found in more advanced tumors (patients II, III, V, and VII), whereas three patients had no alterations in the levels of PAX3 expression (patients IV, VI, and VIII). Finally, one patient lost PAX3 expression in a more advanced stage (patient I). However, there was no correlation either with expression of the melanoma-specific antigens melanA, MAGE-3, nor with tyrosinase. Hence,PAX3 expression identifies a distinct subset of melanomas.

PAX3 Expression in Tumor Sections.

To ensure that expression of PAX3 in short-time melanoma cultures does accurately reflect expression in the tumor tissue, in situ hybridization of paraffin tissue sections with a PAX3-specific antisense probe was performed. Sections from the original tumors of the same biopsy that were used to establish melanoma cultures from a total of 10 different lesions were hybridized. Six of these were found to be positive for PAX3 (950728, 961127, 960306, 960819, 961209,and 980513), whereas four tumors showed no PAX3 expression (950504,961205, 970917, and 980229). Hence, the in situhybridization data correlate perfectly with the RT-PCR results of the cultured melanomas (Table 1). As an example, a PAX3-positive (980513)tumor is shown hybridized with antisense or sense probe (Fig. 2,A and B, respectively). A PAX3-negative (970917)metastasis is shown in Fig. 2,C. Intriguingly, PAX3 expression was confined to malignant metastases and not observed in the surrounding normal tissue (Fig. 2,E). Also, in PAX3-negative tumors, only melanin deposits can be seen, and tumor cells are clearly negative (Fig. 2,F). When tumor and normal skin are present in the same sections and tumor cells are clearly positive (Fig. 2,D), epidermis including melanocytes is negative (Fig. 2,G). To further confirm that PAX3 expression is confined to malignant cells, several benign nevi have been analyzed including dysplastic, compound, dermal, and junctional nevi (Fig. 2, H and I). No PAX3 expression was detected in these lesions. Hence, the expression of PAX3 is restricted to malignant melanomas already at early stage I as well as their distant metastases and not detected in normal skin melanocytes or benign lesions.

Survival of Melanoma Cells Depends on PAX3 Expression.

We described previously that PAX3 expression in rhabdomyosarcoma, a childhood tumor associated with myogenic precursor cells (10, 12), provides a critical survival function (11). Therefore, we hypothesized that expression of PAX3 in melanomas might also affect cell survival. To test this, different melanoma cell lines were treated with oligonucleotides designed to down-regulate PAX3 expression, and cell survival was monitored by counting after trypan blue exclusion. To determine liposome and ODN toxicity, cells were incubated with lipofectin alone and with MS-ODN (Fig. 1,B). Incubating PAX3-positive cells with specific AS-ODNs resulted in dramatically reduced cell viability (down to 28–32%; Fig. 1,B). In contrast, AS-ODN treatment of PAX3-negative cells had no effect on cell viability compared with lipofectin-treated cells. The cellular morphology observed during ODN treatment of the PAX3-positive melanoma culture 980928 and the PAX3-negative cell line Küng-A375 is illustrated in Fig. 1,C. AS-ODN-treated 980928 cells show nuclear condensation, rounding up, and finally detachment from the culture dish, indicating an apoptotic process. To assess this observation on the molecular level, the same cells were stained for Annexin V after incubation with either MS-ODN or AS-ODN(Fig. 1,D). Sixty-five h after ODN incubation, the few 980928 cells still remaining after AS-ODN treatment stained to 30% (29 of 96)positive for Annexin V, whereas incubation with MS-ODN did not yield in any Annexin V-positive cells. In contrast, in the PAX3-negative cell line A365, neither incubation with AS-ODN nor MS-ODN resulted in any Annexin V-stained cells (Fig. 1 D). These results demonstrate that treatment of PAX3-expressing melanomas with AS-ODN targeting PAX3 results in substantial reduction of viability via an apoptotic pathway.

PAX3 appears to play an essential role in the establishment of melanogenic and myogenic cell lineages, as suggested by the severe phenotypes in mice and in humans with reduced functional PAX3 protein. Because PAX3 is implicated in the pediatric tumor rhabdomyosarcoma,both because of the occurrence of a specific translocation involving PAX as well as its overexpression (5), we speculated that reexpression of PAX3 might also occur in melanoma. Indeed, in this study, we demonstrate by sensitive and specific RT-PCR analysis in cultured melanomas, as well as by in situ hybridization on tissue sections, that PAX3 is expressed in a substantial fraction of melanomas. Importantly, PAX3-positive and -negative cultured melanomas could be unambiguously distinguished by RT-PCR and confirmed by in situ hybridization on the corresponding tissue sections. These cultured primary melanomas were additionally analyzed for known melanoma markers (Melan A, MAGE-3, and tyrosinase) to ensure the presence of melanoma cells. Nevertheless, any contaminating fibroblasts in these cultures would not have been detected in our RT-PCR analysis,because fibroblasts are negative for PAX3 expression (data not shown). However, whereas PAX3 expression was confined to malignant melanomas and not detected in benign lesions or normal skin by in situhybridization, sensitive RT-PCR methods were also able to generate a signal in foreskin (data not shown). Hence, nonquantitative RT-PCR should be used cautiously when directly analyzing biopsy specimens. For paraffin-embedded tissue sections, in situ hybridization will therefore be the method of choice for the detection of tumor-specific PAX3 expression, because available antibodies against PAX3 are not sufficiently sensitive to detect the endogenous protein.

One evident interpretation of these results might be that PAX3 expression simply indicates progressive dedifferentiation of melanocytes. However, some interesting parallels exist between myogenesis and melanogenesis. PAX3 seems to be necessary for the development of both lineages and in each case activates tissue-specific transcription factors of the basic helix-loop-helix class, which are important in cell determination and differentiation (myoD family members in muscle and MITF in melanoblasts; Refs. 13 and14). How the activity of PAX3 is regulated toward either one or the other lineage is unknown at present.

In tumor cells, expression of developmentally regulated genes occurs fairly often. However, it is unclear if the reexpressed genes play any role in the development or maintenance of the tumorigenic phenotype. Therefore, an antisense strategy to investigate the functional role of PAX3 expression in melanomas was applied. Indeed, PAX3-expressing melanoma cell lines and cultures were susceptible to AS-ODN treatment and showed reduced viability compared with control cells. Because dying cells could readily be stained with Annexin V and show clear nuclear condensation, we conclude that these cells die by apoptosis. The effect is specific because PAX3-negative cells did not respond to AS-ODN treatment, nor did cells treated with control MS-ODN. These experiments suggest that PAX3 has a conserved antiapoptotic function in both melanoma and rhabdomyosarcoma.

The PAX3-dependent survival pathway is not yet fully understood. Because PAX3 is a transcription factor, either activation or repression of target genes might be involved. Recently, several direct targets for PAX3 have been identified. The best-studied PAX3 target gene is the proto-oncogene c-met, which encodes the tyrosine kinase receptor for HGF/SF (15). The importance of the c-met/HGF signaling pathway for tumor development and cell survival is well documented, and indeed strong c-met expression is found in both melanomas (16) and rhabdomyosarcomas (17). Of specific interest might also be that transgenic mice ectopically expressing HGF/SF develop malignant melanomas, which were shown to overexpress both HGF/SF and c-met (18). Hence, stimulation of c-met expression by PAX3 might be an important pathway in melanoma development. Additional PAX3 target genes that might play a role in cell survival include another member of the tyrosine kinase receptor family, the IGF-1R,4as well as the antiapoptotic survival guide gene bcl-XL(19). The IGF-1R has been implicated to play a crucial role in melanoma development,because growth of melanomas in xenotransplanted nude mice could be inhibited by blocking the IGF-1R signalling pathway with neutralizing antibodies (20). Less is known about the role of the antiapoptotic protein bcl-XL in melanoma development. Future work will have to address the relationship between PAX3 and these putative target genes and its potential significance for melanoma progression and diagnosis.

Expression of PAX3 in the majority of melanomas might participate in their development and/or maintenance because cell survival appears to be dependent on PAX3 expression. PAX3 might therefore represent a possible novel target for therapeutic molecular interventions in melanomas.

Fig. 1.

A, expression of PAX3 in human melanoma cell lines. Total RNA (100 ng) for each RT-PCR reaction was used,whereas FKHR was amplified as an internal control for RNA quality. RT-PCR products were loaded on 2% agarose gels and stained with ethidium bromide. As a marker, a 100-bp ladder was used. B, viability of melanoma cell cultures treated with lipofectin (▪), MS-ODN (), and AS-ODN (□). Cells were treated with 1 μm ODN and 10 μg of lipofectin or with lipofectin alone and counted after 72 h. At least three independent experiments were carried out for each treatment, and each experiment was performed in triplicate; bars, SD. C, morphology of melanoma cells 72 h after start of the ODN treatment. Küng A375 and 980928 melanoma cells were treated with either lipofectin alone or in combination with 1μ m MS-ODNs or AS-ODNs. D, detection of apoptosis after AS-ODN treatment. Annexin V-FITC labeling of 980928(A–D) and A365 (E–H) cells 65 h after incubation with 1 μm of either AS-ODN(C, D, G, and H) or MS-ODN (A, B, E, and F). A, C, E, and G are phase contrast pictures.

Fig. 1.

A, expression of PAX3 in human melanoma cell lines. Total RNA (100 ng) for each RT-PCR reaction was used,whereas FKHR was amplified as an internal control for RNA quality. RT-PCR products were loaded on 2% agarose gels and stained with ethidium bromide. As a marker, a 100-bp ladder was used. B, viability of melanoma cell cultures treated with lipofectin (▪), MS-ODN (), and AS-ODN (□). Cells were treated with 1 μm ODN and 10 μg of lipofectin or with lipofectin alone and counted after 72 h. At least three independent experiments were carried out for each treatment, and each experiment was performed in triplicate; bars, SD. C, morphology of melanoma cells 72 h after start of the ODN treatment. Küng A375 and 980928 melanoma cells were treated with either lipofectin alone or in combination with 1μ m MS-ODNs or AS-ODNs. D, detection of apoptosis after AS-ODN treatment. Annexin V-FITC labeling of 980928(A–D) and A365 (E–H) cells 65 h after incubation with 1 μm of either AS-ODN(C, D, G, and H) or MS-ODN (A, B, E, and F). A, C, E, and G are phase contrast pictures.

Close modal
Fig. 2.

Detection of PAX3 expression in melanomas by in situ hybridization. Tissue sections from malignant metastasis(A, positive sample 980513; B, control staining with sense probe of the same tumor; C, negative sample 970917; D, positive sample 960819; and E, positive sample 950728) and primary tumor(F, negative sample 961205) were hybridized with a biotinylated PAX3-specific antisense riboprobe. G, a different portion from the same section as shown in D. H and I, benign lesions, dysplastic, and compound nevi, respectively. Positive reaction is indicated by the blue stain; the brown stain is caused by melanin deposits. AC and E, ×200; D, F, and GI, ×100.

Fig. 2.

Detection of PAX3 expression in melanomas by in situ hybridization. Tissue sections from malignant metastasis(A, positive sample 980513; B, control staining with sense probe of the same tumor; C, negative sample 970917; D, positive sample 960819; and E, positive sample 950728) and primary tumor(F, negative sample 961205) were hybridized with a biotinylated PAX3-specific antisense riboprobe. G, a different portion from the same section as shown in D. H and I, benign lesions, dysplastic, and compound nevi, respectively. Positive reaction is indicated by the blue stain; the brown stain is caused by melanin deposits. AC and E, ×200; D, F, and GI, ×100.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by Grants 31-46886.96 and 31-56869.99 from the Swiss National Science Foundation, the Krebsliga of the Kanton Zurich, and the Hartmann Muller Foundation.

3

The abbreviations used are: RT-PCR, reverse transcription-PCR; ODN, oligonucleotide; MS-ODN, missense ODN;AS-ODN, antisense ODN; HGF/SF, hepatocyte growth factor/scatter factor;IGF-1R, insulin-like growth factor receptor 1.

4

O. V. Murmann, C. M. Masque, D. Le Roith, F. G. Barr, F. Niggli, B. W. Schäfer. Paired-box transcription factors PAX3 and PAX3/FKHR modulate expression of the type 1 insulin-like growth factor receptor, submitted for publication.

Table 1

Expression of PAX3 in melanomas

PatientStage, TNMaPAX3
RT-PCRbIn situ              c
950504 I, T3N0M0 − − 
950803 I, T2N0M0 ++  
951010 I, T3N0M0 ++  
951019 I, T3N0M0 ++  
960514 I, T4N0M0 ++  
960704 I, T1N0M0 ++  
960724 I, T1N0M0 ++  
961008 I, T1N0M0 ±  
961121 I, T4N0M0 ++  
961122 I, T1N0M0 ++  
961205 I, T3N0M0 − − 
941209 II, TxN2aM0 −  
950322 II, TxN2M0 ++  
950526 II, TxN2aM0 −  
950728 II, TxN2aM0 ++ ++ 
950822 II, TxN2aM0 ++  
960104 II, TxN2bM0 ++  
960312 II, TxN1M0 −  
961127 II, TxN2bM0 ++ ++ 
970325 II, TxN1M0 ++  
970917 II, T4N2aM0 − − 
980229 II, TxN2aM0 − − 
980325 II, TxN2aM0 ±  
960306 III, TxNxM1a ++ ++ 
960618 III, TxNxM1a ++  
960819 III, TxNxM1a ++ ++ 
960924 III, TxNxM1a ++  
961209 III, TxNxM1b ++ ++ 
970604 III, TxNxM1a ++  
970703 III, TxNxM1b ±  
970723 III, TxNxM1b ++  
980409 III, TxNxM1b ++  
980513 III, TxNxM1b ++ ++ 
980924 III, TxNxM1b −  
980928 III, TxNxM1b ++  
PatientStage, TNMaPAX3
RT-PCRbIn situ              c
950504 I, T3N0M0 − − 
950803 I, T2N0M0 ++  
951010 I, T3N0M0 ++  
951019 I, T3N0M0 ++  
960514 I, T4N0M0 ++  
960704 I, T1N0M0 ++  
960724 I, T1N0M0 ++  
961008 I, T1N0M0 ±  
961121 I, T4N0M0 ++  
961122 I, T1N0M0 ++  
961205 I, T3N0M0 − − 
941209 II, TxN2aM0 −  
950322 II, TxN2M0 ++  
950526 II, TxN2aM0 −  
950728 II, TxN2aM0 ++ ++ 
950822 II, TxN2aM0 ++  
960104 II, TxN2bM0 ++  
960312 II, TxN1M0 −  
961127 II, TxN2bM0 ++ ++ 
970325 II, TxN1M0 ++  
970917 II, T4N2aM0 − − 
980229 II, TxN2aM0 − − 
980325 II, TxN2aM0 ±  
960306 III, TxNxM1a ++ ++ 
960618 III, TxNxM1a ++  
960819 III, TxNxM1a ++ ++ 
960924 III, TxNxM1a ++  
961209 III, TxNxM1b ++ ++ 
970604 III, TxNxM1a ++  
970703 III, TxNxM1b ±  
970723 III, TxNxM1b ++  
980409 III, TxNxM1b ++  
980513 III, TxNxM1b ++ ++ 
980924 III, TxNxM1b −  
980928 III, TxNxM1b ++  
a

TNM, Tumor-Node-Metastasis; I, patients with primary melanoma: II, patients with locoregional lymph node metastases: III,patients with remote metastases. TNM classification according to UICC.−, no PAX3 expression: ±, just detectable PAX3 signal; ++, PAX3 expression.

b

Cultured melanomas.

c

tissue sections.

Table 2

PAX3 expression in melanomas of the same patient

PatientMelanoma cultureStage, TNMaPAX3bMAGE-3bcMelanAbcTyrosinasebc
961121 I, T4N0M0 ++ 
 970917 II, T4N2aM0 − ND ND ND 
II 941209 II, TxN2aM0 − − − d 
 950526 II, TxN2aM0 − − 
 950822 II, TxN2aM0 ++ 
III 960312 II, TxN1M0 − − − − 
 970325 II, TxN1M0 ++ − − − 
IV 980229 II, TxN2aM0 − d d +e 
 980924 III, TxNxM1b − − e 
980325 II, TxN2aM0 ± d +e 
 980928 III, TxNxM1b ++ +d +d d 
VI 950322 II, TxN2M0 ++ − − 
 961209 III, TxNxM1b ++ d d d 
VII 970703 III, TxNxM1b ± d d +d 
 970723 III, TxNxM1b ++ d +d d 
VIII 980409 III, TxNxM1b ++ +d d +d 
 980513 III, TxNxM1b ++ +e +e +e 
PatientMelanoma cultureStage, TNMaPAX3bMAGE-3bcMelanAbcTyrosinasebc
961121 I, T4N0M0 ++ 
 970917 II, T4N2aM0 − ND ND ND 
II 941209 II, TxN2aM0 − − − d 
 950526 II, TxN2aM0 − − 
 950822 II, TxN2aM0 ++ 
III 960312 II, TxN1M0 − − − − 
 970325 II, TxN1M0 ++ − − − 
IV 980229 II, TxN2aM0 − d d +e 
 980924 III, TxNxM1b − − e 
980325 II, TxN2aM0 ± d +e 
 980928 III, TxNxM1b ++ +d +d d 
VI 950322 II, TxN2M0 ++ − − 
 961209 III, TxNxM1b ++ d d d 
VII 970703 III, TxNxM1b ± d d +d 
 970723 III, TxNxM1b ++ d +d d 
VIII 980409 III, TxNxM1b ++ +d d +d 
 980513 III, TxNxM1b ++ +e +e +e 
a

TNM, Tumor-Node-Metastasis; I, patients with primary melanoma; II, patients with locoregional lymph node metastasis; III,patients with remote metastasis. TNM classification according to UICC.

b

ND, not done; −, no PAX3 expression; ±, just detectable PAX3 signal; ++, PAX3 expression.

c

\

N

e

Expression was analyzed by RT-PCR or

d

FACS;or

e

immunohistochemistry.

We are thankful for the continuous support of C. W. Heizmann.

1
Nestle F. O., Burg G., Dummer R. New perspectives on immunobiology and immunotherapy of melanoma..
Immunol. Today
,
20
:
5
-7,  
1999
.
2
Dahl E., Koseki H., Balling R. Pax genes and organogenesis..
BioEssays
,
19
:
755
-765,  
1997
.
3
Bober E., Franz T., Arnold H. H., Gruss P., Tremblay P. PAX3 is required for the development of limb muscles: a possible role for the migration of dermomyotomal muscle progenitor cells..
Development (Camb.)
,
120
:
603
-612,  
1994
.
4
Tassabehji M., Newton V. E., Leverton K., Turnbull K., Seemanova E., Kunze J., Sperling K., Strachan T., Read A. P. PAX3 gene structure and mutations: close analogies between Waardenburg syndrome and the Splotch mouse..
Hum. Mol. Genet.
,
3
:
1069
-1074,  
1994
.
5
Schäfer B. W. Emerging roles for PAX transcription factors in cancer biology..
Gen. Physiol. Biophys.
,
17
:
211
-224,  
1998
.
6
Schulte T. W., Toretsky J. A., Ress E., Helman L., Neckers L. M. Expression of PAX3 in Ewing’s sarcoma family of tumors..
Biochem. Mol. Med.
,
60
:
121
-126,  
1997
.
7
Geertsen R. C., Hofbauer G. F., Yue F. Y., Manolio S., Burg G., Dummer R. Higher frequency of selective losses of HLA-A and -B allospecificities in metastasis than in primary melanoma lesions..
J. Investig. Dermatol.
,
111
:
497
-502,  
1998
.
8
Scholl F. A., Betts D. R., Niggli F. K., Schäfer B. W. Molecular features of a human rhabdomyosarcoma cell line with spontaneous metastatic progression..
Br. J. Cancer
,
82
:
1239
-1245,  
2000
.
9
Geertsen R., Zenklusen R., Kamarashev J., Burg G., Dummer R. Inverse regulation of neuronal cellular adhesion molecule (NCAM) by IFN-γ in melanoma cell cultures established from CNS lesions..
Int. J. Cancer
,
83
:
135
-140,  
1999
.
10
Schäfer B. W., Czerny T., Bernasconi M., Genini M., Busslinger M. Molecular cloning and characterization of a human PAX-7 cDNA expressed in normal and neoplastic myocytes..
Nucleic Acids Res.
,
22
:
4574
-4582,  
1994
.
11
Bernasconi M., Remppis A., Fredericks W. J., Rauscher F. J. , III, and Schäfer, B..
W. Induction of apoptosis in rhabdomyosarcoma cells through down-regulation of PAX proteins. Proc. Natl. Acad. Sci. USA
,
93
:
13164
-13169,  
1996
.
12
Frascella E., Toffolatti L., Rosolen A. Normal and rearranged PAX3 expression in human rhabdomyosarcoma..
Cancer Genet. Cytogenet.
,
102
:
104
-109,  
1998
.
13
Tajbakhsh S., Rocancourt D., Cossu G., Buckingham M. Redefining the genetic hierarchies controlling skeletal myogenesis: Pax-3 and Myf-5 act upstream of MyoD..
Cell
,
89
:
127
-138,  
1997
.
14
Watanabe A., Takeda K., Ploplis B., Tachibana M. Epistatic relationship between Waardenburg syndrome genes MITF and PAX3..
Nat. Genet.
,
18
:
283
-286,  
1998
.
15
Epstein J. A., Shapiro D. N., Cheng J., Lam P. Y. P., Maas R. L. PAX3 modulates expression of the c-Met receptor during limb muscle development..
Proc. Natl. Acad. Sci. USA
,
93
:
4213
-4218,  
1996
.
16
Saitoh K., Takahashi H., Sawada N., Parsons P. G. Detection of the c-met proto-oncogene product in normal skin and tumors of melanocytic origin..
J. Pathol.
,
174
:
191
-199,  
1994
.
17
Ginsberg J. P., Davis R. J., Bennicelli J. L., Nauta L. E., Barr F. G. Up-regulation of MET but not neural cell adhesion molecule expression by the PAX3-FKHR fusion protein in alveolar rhabdomyosarcoma..
Cancer Res.
,
58
:
3542
-3546,  
1998
.
18
Otsuka T., Takayama H., Sharp R., Celli G., LaRochelle W. J., Bottaro D. P., Ellmore N., Vieira W., Owens J. W., Anver M., Merlino G. c-Met autocrine activation induces development of malignant melanoma and acquisition of the metastatic phenotype..
Cancer Res.
,
58
:
5157
-5167,  
1998
.
19
Margue C. M., Bernasconi M., Barr F. G., Schäfer B. W. Transcriptional modulation of the anti-apoptotic protein BCL-XL by the paired box transcription factors PAX3 and PAX3/FKHR..
Oncogene
,
19
:
2919
-2929,  
2000
.
20
Furlanetto R. W., Harwell S. E., Baggs R. B. Effects of insulin-like growth factor receptor inhibition on human melanomas in culture and in athymic mice..
Cancer Res.
,
53
:
2522
-2526,  
1993
.