Responsiveness of stem-like human glioma cells to all-trans retinoic acid and requirement of retinoic acid receptor isotypes α, ß and γ.

Retinoic acid (RA) is required for development and homeostasis of the normal mammalian brain and may play a role in the initiation and progression of malignant brain tumors, such as the glioblastoma multiforme (GBM) and the gliosarcoma (Gsarc). The subpopulation of stem-like glioma cells (SLGCs) was shown to resist standard glioma radio-/chemotherapy and to propagate tumor regrowth. We used phenotypically distinct, self-renewing SLGC lines from six human GBMs, two Gsarcs, and two subcloned SLGC derivatives in order to investigate their responsiveness to all-trans retinoic acid (atRA) and to identify the RA-receptor (RAR) isotypes involved. In general, atRA exerted a pro-proliferative and pro-survival effect on SLGCs, though the efficacy was distinct. By means of RAR isotype-selective retinoids we disclosed that these effects were mediated by RARα and RARγ, except for one SLGC line, in which the pro-proliferative signal was induced by the RARß-selective retinoid. Only one GBM-derived cell line (T1338) and a subpopulation of another (T1389) displayed neural differentiation in response to atRA. Differentiation of T1338 was induced by RARα and RARγ isotype-selective retinoids, associated with down-regulation of Sox2, and the failure to induce orthotopic tumors in the brains of SCID mice. The differential responsiveness of the SLGC lines appeared unrelated to the expression of RARß, as (i) atRA augmented RAR isotype mRNA expression and particularly rarß mRNA in all SLGC lines, (ii) rarß promoter hypomethylation in the SLGC lines was not related to differentiation and (iii) the induction of T1338 differentiation was by RARα- and RARγ-selective ligands.

The biologic lower detection limit of six ultrasensitive PSA assays.

BACKGROUND: The benefit of monitoring patients with prostate cancer (PCA) by ultrasensitive measurement of prostate specific antigen (PSA) is frequently discussed. Usually, the analytic lower detection limit of an ultrasensitive assay is determined by the manufacturer. As the analytic lower detection limit does not take into account interfering factors of human serum, the biologic lower detection limit, which is defined as PSA concentration detected in PSA-free human serum, plus 3 standard deviations, is of greater interest. MATERIALS AND METHODS: We investigated the biologic lower detection limit of six ultrasensitive PSA assays. Sera from 15 men with bladder cancer after radical cystoprostatectomy and from 30 healthy women were applied. Hence, we expected no PSA of prostatic origin. RESULTS: The biologic lower detection limit obtained using these sera was up to 30 fold higher (men, 0.29-0.63 ng/ml; women, 0.03-0.69 ng/ml) than the analytic lower detection limit (0.01-0.09 ng/ml). CONCLUSIONS: PSA measurement in sera obtained from men without prostate and women results in PSA values above the ultrasensitive range. Therefore, advantages provided by ultrasensitive PSA measurement in monitoring PCA patients after radical prostatectomy are limited.

Different classes of coactivators recognize distinct but overlapping binding sites on the estrogen receptor ligand binding domain.

We have analyzed interaction of coactivators with the wild-type estrogen receptor alpha (ER), HEG0, and a mutant, L536P-HEG0, which is constitutively active in several transiently transfected cells and a HeLa line that stably propagates an estrogen-sensitive reporter gene. Different classes of coactivators do not recognize the ER ligand binding domain (LBD) in the same manner. Steroid receptor coactivator-1 (SRC-1), amplified in breast cancer-1 (AIB-1), transcriptional intermediary factor-1 (TIF-1), transcriptional intermediary factor-2 (TIF-2), and receptor interacting protein 140 (RIP140) interacted with HEG0 and L536P-HEG0 in the presence of estradiol, but generally not in the presence of anti-estrogens. However, ICI164,384 stimulated some interaction of RIP140 with LBDs. SRC-1, AIB-1, and RIP140 interacted constitutively with the L536P ER, whereas TIF-1 and TIF-2 interacted only weakly in the absence of hormone. Reciprocal competition for binding to the ER LBD was observed between different classes of coactivators. Moreover, coexpression of RIP140 blocked enhanced transactivation by HEG0 observed in the presence of TIF-2, suggesting that RIP140 may play a negative role in ER signaling. We conclude that constitutive activity of L536P-HEG0 is manifested to similar degrees in different cell types and likely arises from constitutive coactivator binding; different classes of coactivators recognize distinct but overlapping binding sites on the ER LBD. Finally, the observation that L536P-HEG0 interacted constitutively with AIB-1, a coactivator that has been implicated in ER signaling in breast and ovarian cancer, suggests that similar mutations in the ER may contribute to hormone-independent proliferation of breast and ovarian cells.

Comparison of prostate-specific antigen (PSA) measured by four combinations of free PSA and total PSA assays.

We compared prostate-specific antigen (PSA) assay systems [i.e., free PSA (f-PSA) and the corresponding total PSA (t-PSA) assay] from four different manufacturers as well as the f-PSA/t-PSA ratios with regard to their ability to discriminate between benign prostate hyperplasia (BPH) and prostate cancer (PCA). ROC analysis showed similar areas under the curves (AUCs) with different assay systems. For the entire patient population the AUCs of the f-PSA/t-PSA ratio were not or slightly increased compared with the sole measurement of t-PSA (t-PSA, 0.792-0.820; f-PSA/t-PSA ratio, 0.685-0.859). In contrast, for only those patients who showed t-PSA concentrations within the diagnostic gray area of 4-25 micrograms/L t-PSA, the AUCs were greater for the f-PSA/ t-PSA ratio than for measurement of t-PSA alone (t-PSA, 0.608-0.647; f-PSA/t-PSA ratio, 0.690-0.806). These results were confirmed by the predictive values of the negative results (NPVs) of the t-PSA assays and the f-PSA/t-PSA ratios (assay thresholds corresponding to a 95% detection limit). Compared with the sole t-PSA measurement there was no mentionable increase in the NPVs due to the f-PSA/t-PSA ratio for the entire patient population, but an increase up to 49% when limited to t-PSA concentrations within 4-25 micrograms/L. We therefore conclude that the f-PSA/t-PSA ratio may be helpful for differential diagnosis of BPH and PCA within the diagnostic gray area of 4-25 micrograms/L t-PSA.

Impact of free prostate-specific antigen on discordant measurement results of assays for total prostate-specific antigen.

OBJECTIVES: To determine why various assays for total PSA (t-PSA) produce discordant results in identical serum samples. METHODS: A total of 84 sera from 40 patients with histologically confirmed benign prostatic hyperplasia and from 44 patients with untreated prostate cancer were analyzed with seven assays for t-PSA and the Hybritech research assay for free prostate-specific antigen (f-PSA). Comparison between assays was performed by linear regression of the t-PSA concentrations as well as between the t-PSA concentrations and the f/t-PSA ratios. RESULTS: The coefficients of correlation for the investigated assays versus Hybritech Tandem-E range from 0.96 to 0.99. Nevertheless average PSA concentrations differed significantly from the Tandem-E assay in all assays. Despite a good correlation, some assays showed a regression line with a slope notably different from 1. In these assays, elevated concentrations were observed in sera with a high proportion of f-PSA. CONCLUSIONS: The study illustrates a significant and clinically relevant discordance between reported t-PSA concentrations for identical samples, depending on the assay used and on the contents of f-PSA in the sample. The interpretation of t-PSA concentrations requires awareness of the applied assay as well as the establishment of an assay-specific reference range in order to avoid inappropriate clinical consequences, such as unnecessary biopsies. Respective details must be contained in the laboratory reports. A change of assays without specifically reassessing previously valid reference ranges or the uncritical use of a customarily applied limit of < 4 ng/mL will otherwise be harmful to the patient.

TIF2, a 160 kDa transcriptional mediator for the ligand-dependent activation function AF-2 of nuclear receptors.

Nuclear receptors (NRs) act as ligand-inducible transcription factors which regulate the expression of target genes upon binding to cognate response elements. The ligand-dependent activity of the NR activation function AF-2 is believed to be mediated to the transcription machinery through transcriptional mediators/intermediary factors (TIFs). We report here the cloning of the 160 kDa human nuclear protein TIF2, which exhibits all properties expected for a mediator of AF-2: (i) it interacts in vivo with NRs in an agonist-dependent manner; (ii) it binds directly to the ligand-binding domains (LBDs) of NRs in an agonist- and AF-2-integrity-dependent manner in vitro; (iii) it harbours an autonomous transcriptional activation function; (iv) it relieves nuclear receptor autosquelching; and (v) it enhances the activity of some nuclear receptor AF-2s when overexpressed in mammalian cells. TIF2 exhibits partial sequence homology with the recently isolated steroid receptor coactivator SRC-1, indicating the existence of a novel gene family of nuclear receptor transcriptional mediators.

Ligand-dependent interaction of nuclear receptors with potential transcriptional intermediary factors (mediators).

The activity of the ligand-inducible activation function 2 (AF-2) contained in the ligand binding domain (LBD) of nuclear receptors (NRs) is thought to be mediated by transcriptional intermediary factors (TIFs). We have recently reported the isolation and characterization of two novel mouse proteins, designated TIF1 and mSUG1, that interact in a ligand-dependent fashion with the LBD (region E) of several NRs in vivo as well as in vitro. Remarkably, these interactions require the conserved core motif of the AF-2 activating domain (AF-2 AD) and can be blocked by AF-2 antagonists. TIF1 and mSUG1 might therefore represent TIFs/mediators for the ligand-dependent AF-2 of NRs. By comparing the interaction properties of these two putative TIFs with different NRs including the oestrogen (ER), thyroid hormone (TR), vitamin D3 (VDR), retinoic acid (RAR alpha) and retinoid X (RXR) receptors, we demonstrate that: (i) RXR alpha efficiently interacts with TIF1, but not with mSUG1, whereas TR alpha interacts much more efficiently with mSUG1 than with TIF1, and RAR alpha, VDR and ER efficiently interact with both TIF1 and mSUG1; (ii) the amphipathic alpha helix core of AF-2 AD is differentially involved in the interactions of RAR alpha with TIF1 and mSUG1; and (iii) the AF-2 AD cores of RAR alpha and ER are similarly involved in their interaction with TIF1, but not with mSUG1. Thus the interaction interfaces between the various NRs and either TIF1 or mSUG1 may vary depending on the nature of both the receptor and the putative mediator of its AF-2 function. We discuss the possible roles of TIF1 and mSUG1 as mediators of the transcriptional activity of the AF-2 of NRs.

Differential ligand-dependent interactions between the AF-2 activating domain of nuclear receptors and the putative transcriptional intermediary factors mSUG1 and TIF1.

Using a yeast two-hybrid system we report the isolation of a novel mouse protein, mSUG1, that interacts with retinoic acid receptor alpha (RAR alpha) both in yeast cells and in vitro in a ligand- and AF-2 activating domain (AF-2 AD)-dependent manner and show that it is a structural and functional homologue of the essential yeast protein SUG1. mSUG1 also efficiently interacts with other nuclear receptors, including oestrogen (ER), thyroid hormone (TR), Vitamin D3 (VDR) and retinoid X (RXR) receptors. By comparing the interaction properties of these receptors with mSUG1 and TIF1, we demonstrate that: (i) RXR alpha efficiently interacts with TIF1, but not with mSUG1, whereas TR alpha interacts much more efficiently with mSUG1 than with TIF1, and RAR alpha, VDR and ER efficiently interact with mSUG1 and TIF1; (ii) the amphipathic alpha-helix core of the AF-2 AD is differentially involved in interactions of RAR alpha with mSUG1 and TIF1; (iii) the AF-2 AD cores of RAR alpha and ER are similarly involved in their interaction with TIF1, but not with mSUG1. Thus, the interaction interfaces between the different receptors and either mSUG1 or TIF1 may vary depending on the nature of the receptor and the putative mediator of its AF-2 function. We discuss the possibility that mSUG1 and TIF1 may mediate the transcriptional activity of the AF-2 of nuclear receptors through different mechanisms.

The N-terminal part of TIF1, a putative mediator of the ligand-dependent activation function (AF-2) of nuclear receptors, is fused to B-raf in the oncogenic protein T18.

Nuclear receptors (NRs) bound to response elements mediate the effects of cognate ligands on gene expression. Their ligand-dependent activation function, AF-2, presumably acts on the basal transcription machinery through intermediary proteins/mediators. We have isolated a mouse nuclear protein, TIF1, which enhances RXR and RAR AF-2 in yeast and interacts in a ligand-dependent manner with several NRs in yeast and mammalian cells, as well as in vitro. Remarkably, these interactions require the amino acids constituting the AF-2 activating domain conserved in all active NRs. Moreover, the oestrogen receptor (ER) AF-2 antagonist hydroxytamoxifen cannot promote ER-TIF1 interaction. We propose that TIF1, which contains several conserved domains found in transcriptional regulatory proteins, is a mediator of ligand-dependent AF-2. Interestingly, the TIF1 N-terminal moiety is fused to B-raf in the mouse oncoprotein T18.

Expression of genes related to the human erbB, erbA, pdgf and pdgf-r in tumors of different etiology in Xiphophorus.

The melanoma determining Tu locus of the teleost Xiphophorus contains an accessory gene, x-erbB*a, which is closely related to the EGF receptor gene family, and is probably oncogenic. x-erbB*a exists in allelic forms that are specific for distinct Tu-loci, and shows high homology to a non-allelic non-oncogenic counterpart x-erbB*i which is transcribed into mRNA of 4.6 kb in non-tumorous and tumorous tissues of fish harboring and lacking Tu. Expression of a 4.0-kb mRNA in tumors (melanoma and fibrosarcoma) of different etiology is strictly correlated with the inheritance of X. maculatus x-erbB*a alleles; transcripts of 8.0 kb were detected in melanoma and carcinoma of fish harboring a certain x-erbB*a of X. variatus. The expression of the putative x-erbB*a transcripts parallels the stage of malignancy of the tumor. The expression of the xiphophorine EGF receptor gene (x-erbB) was detected in almost all tumors, is strongly enhanced in carcinoma, and is positively correlated with the degree of malignancy of melanoma and fibrosarcoma. Some tumors show expression of erbA-related genes. The PDGF receptor mRNA is expressed in all tumors analyzed and shows enhanced expression in malignant tumors of neurogenic, epithelial and mesenchymal origin. Expression of x-pdgf was observed in several cases of melanoma, but more frequently in carcinoma and fibrosarcoma. We conclude that x-erbB*a might be involved in initiation of tumors of different cellular origin and etiology in fish harboring Tu, as well as in the determination of the malignancy of the tumor. Furthermore, we assume that x-erbB*i, x-erbB, x-pdgf and x-pdgf-r play a role in secondary events in tumorigenesis by, e.g., conferring a selective growth advantage to the tumor cells.

Discrimination of initiating and promoting carcinogens in fish.

Xiphophorine fish from wild populations are insusceptible of developing neoplasia. In contrast, certain backcrosses of Xiphophorus maculatus (platyfish) with Xiphophorus helleri (swordtail) as the recurrent parent produce offspring that develop neoplasia in a Mendelian fashion. We concentrated our research on melanoma. To construct a fish strain which is highly susceptible to mutagenic carcinogens, a particular regulatory gene, ie an oncosuppressor gene (Bs), was introduced into the fish developing the Mendelian inherited melanoma by introgression. Bs prevents the progeny from developing melanoma. However, Bs can be impaired by carcinogen-induced somatic mutation which gives rise to the development of clonal melanoma. Activity of the oncogene x-src (measured on pp60x-src kinase activity) and inositol lipid turnover is elevated in the tumor but, in contrast to the animals bearing the inherited melanoma, not in the brain. Tumor promoters do not induce melanoma in this strain. Similarly, in order to breed a fish strain which is highly susceptible to tumor promoters we introduced a regulatory gene, for instance an oncostatic gene (g) coding for a pretransformational arrest of pigment cell differentiation in the stem cell stage of the fish that develop the Mendelian inherited melanoma. The new strain is incapable of developing melanoma. Its x-src kinase activity and inositol lipid turnover is elevated in the brain, indicating that the biochemical processes which were found to be correlated with the hereditary melanoma formation, operate without the occurrence of melanoma. Following treatment of these animals with tumor promoters, melanoma develops within a very short latent period. Our tester strain can discriminate between tumor-initiating and tumor-promoting activities of agents of unknown carcinogenic potential.

Genetic and molecular approach for breeding and use of laboratory fish for the detection of agents with carcinogenic and/or promoting activity.

Xiphophorine fish from wild populations are insusceptible to develop neoplasia. In contrast, certain backcrosses of Xiphophorus maculatus (platyfish) with Xiphophorus helleri (swordtail) as the recurrent parent, produce offspring that develop neoplasia in a Mendelian fashion. We concentrated our research on melanoma. The starting signal for the development of melanoma comes from an accessory v-erbB-related oncogene, x-erbB*, which is highly homologous to the human EGF receptor gene, and is part of a platyfish-specific tumor gene-complex designated as Tu. Normally, the platyfish is protected from its own Tu by Tu-specific regulatory gene systems. The swordtail has neither evolved the Tu-complex nor the regulatory gene systems. Therefore, the backcross procedure dismantles the regulatory gene systems thus permitting Tu-directed melanoma formation. x-erbB* derived from the platyfish, together with the swordtail-derived oncogenes src, sis, pdgf-r, ras, myc, erbA are expressed or overexpressed in the melanoma, and inositol lipid turnover is considerably elevated. x-src and inositol lipid turnover have also been found elevated in the healthy tissues (eg brain) of the tumourous fish. To construct a fish strain which is highly susceptible to mutagenic carcinogens, we introduced a particular regulatory gene, ie an oncosuppressor gene (Bs), into the genome of the animals developing Mendelian inherited melanoma, by introgressive breeding. Bs prevents the strain from germ line-inherited melanoma but, following carcinogen-induced impairment in a somatic cell gives rise to the Tu-directed development of clonal melanoma in a particular fish, x-src activity and inositol lipid turnover are elevated in the tumor but, in contrast to the animals bearing the inherited melanoma, are not elevated in the brain. Promoting carcinogens (tumor promoters) do not induce melanoma in this strain. Similarly, in order to breed a fish strain susceptible to tumor promoters, we introduced a regulatory gene (the oncostatic gene g, golden) coding for a pre-transformational arrest of pigment cell differentiation in the stem cell stage of the fish that develop the Mendelian inherited melanoma. The new strain is incapable of developing melanoma. Its x-src activity and inositol lipid turnover is elevated in the brain indicating that the biochemical processes which were found to be correlated with the hereditary melanoma formation, are running in the new strain without the occurrence of melanoma. Following treatment with tumor promoters that overcome the arrest of pigment cell differentiation, melanoma develops within a very short latent period.

[Attempts at analyzing the initiation of the initial processes of carcinogenesis based on the Xiphophorus melanoma model]. / Ansätze zur Analyse der Initiation von Initialprozessen der Krebsbildung am Melanom-Modell von Xiphophorus.

Certain backcross hybrids (BC8-22) of a spotted X. maculatus (platyfish) and a non-spotted X. helleri (swordtail; recurrent parent) are highly sensitive to mutagenic carcinogens and, after a latent period of 8 to 12 months, develop melanoma of unicellular origin that is genealogically related to the spots of the platyfish. Sensitivity to the carcinogen or susceptibility to melanoma, respectively, are inherited in a Mendelian fashion and can be assigned to a "tumor gene-complex" (Tu-complex) consisting probably of almost 20 genes. The Tu-complex is located at the end of an autosome or sex chromosome, and is largely deregulated by crossing conditioned replacement of platyfish chromosome carrying regulatory genes (tumor suppressor genes, oncostatic genes, antioncogenes) for the Tu-complex by swordtail chromosomes lacking them. The melanoma-free condition of these BC-hybrids depends upon the skin-specific regulatory gene Bs (body side) that requires impairment in a pigment cell precursor for the outgrowth of melanoma. Structural mutations involving different breakpoints indicate that the signal for melanoma formation comes from a particular region of the Tu-complex where an accessory v-erb B related oncogene (x-erb B*a; 85% homology to the human EGF receptor gene) is located. Northern blot analyses of the melanoma cell line showed an about 20-fold overexpression of x-erbB*a. Both the inositol lipid turnover [(3H)inositol incorporated into phosphoinositides], and the xiphophorine pp60x-src kinase activity that are assumed to be causally involved in tumor formation showed a remarkable elevation in the melanoma as compared to the normal tissue (brain) of the tumorous and non-tumourous (with or without the Tu-complex) segregants. Other BC hybrids carrying the Tu-complex but lacking the linked regulatory gene develop melanoma "spontaneously". This kind of melanoma occurs early in the course of life, is of multicellular origin, and is inherited as a Mendelian character. In contrast to the BC hybrids requiring somatic mutation for melanoma formation, both inositol, lipid turnover and x-src activity are remarkable enhanced in both melanoma and normal tissues. A mutant of the laller BC hybrids carrying in addition of the Tu-complex the homozygous oncostatic gene g (g/g, "golden") that arrests pigment cell differentiation in the stem cell stage is incapable to develop melanoma spontaneously. Nevertheless it shows the elevation of inositol lipid turnover and x-src activity in its always healthy tissues. Following treatment with tumor promoters such as TPA and steroid hormones pigment cell differentiation recovers and melanoma of multicellular origin develops within 4 to 8 weeks.(ABSTRACT TRUNCATED AT 400 WORDS)

Search for genes critical for the early and/or late events in carcinogenesis: studies in Xiphophorus (Pisces, Teleostei).

Southern blot analyses of the xiphophorine genome with probes specific for 15 viral and cellular oncogenes revealed that only three v-erbB related EcoRI fragments comprising 4.9 kb of a certain X, 11.5 kb of another X, and 6.7 kb of both a Y and a Z chromosome are inherited in parallel with the Tu complex and melanoma formation. They are accessory in the genome, and are highly homologous with each other and with an ubiquitous autosomal 7.5-kb fragment. The latter one is probably linked to the indispensable Tu complex that is postulated to be present in all individuals of Xiphophorus irrespective of whether they possess or lack the capacity to form melanoma in interspecific hybrids. Three restriction fragments, the X-chromosomal 4.9-kb, the Y-chromosomal 6.7-kb and the ubiquitous Tu-nonlinked 5.5-kb EcoRI fragments were cloned and sequenced. The X- and the Y-chromosomal fragments show perfect identity in the regions of the putative exons C and D of the EGF receptor gene and minor but significant differences to the putative exon C (exon D not identified) of the Tu-nonlinked fragment of 5.5 kb, indicating that at least two different types of x-erb B genes coding for slightly different EGF-receptors exist in the fish. Northern blot analyses revealed expression of the Tu-linked x-erbB genes (x-gfrB genes) in both transformed and nontransformed tissue, suggesting their essential role in regulation of normal cell proliferation and in carcinogenesis. We conclude that the indispensable x-egfrB genes remain unchanged and strictly regulated, while the sex chromosomal accessory x-egfrB genes possibly undergo dramatic changes in structure and/or function (e.g., unscheduled expression, ectopic expression, point mutations, truncation) leading to activation of the oncogenic potential of these genes, which in turn could induce several cellular events involved in the switch from the normal to the transformed state of the cell. In contrast, none of the x-erbA restriction fragments could be assigned to the Tu-complex or to any regulatory gene (R or S). These results, however, do not exclude the existence of a structural and/or functional relation between x-erbA genes and R and S genes. We therefore analyzed x-erbA genes by cloning, sequencing, and expression studies.(ABSTRACT TRUNCATED AT 400 WORDS)

v-erbB related sequences in Xiphophorus that map to melanoma determining Mendelian loci and overexpress in a melanoma cell line.

Certain genotypes of Xiphophorus harbour an accessory Mendelian gene which following loss, impairment, or insufficiency of its regulatory genes in the germ line, mediates the hereditary capacity to develop neoplasia spontaneously or following induction with initiating and promoting carcinogens. Together with its linked regulatory genes it forms a 'tumor gene-complex' (Tu-complex). We concentrated on accessory sex chromosomal Tu-complexes that are responsible for sex chromosome-linked melanoma formation. Southern analyses of the xiphophorine genome with 15 authentic oncogene probes revealed so far that only three v-erbB related EcoRI fragments comprising 4.9 kb of a certain X-, 11.5 kb of another X-, and 6.7 kb of both a Y- and Z-chromosome are inherited in parallel with the Tu-complex and melanoma formation. They are accessory in the genome, and are highly homologous with each other. The sequence of the X-chromosomal 4.9 kb fragment shows minor but significant differences to that of the invariably present autosomal xiphophorine erbB (x-erbB) fragment of 5.5 kb, indicating that at least two different x-erbB genes coding for different EGF receptors can exist in the fish. Northern analyses showed expression of both genes in a fibroblast cell line, and overexpression of the sex chromosomal x-erbB in a melanoma cell line. The co-segregation of the hereditary trait of melanoma with the sex chromosomal x-erbB fragments, suggests that the accessory x-erbB gene may be responsible for the switch from the normal to the neoplastic state of the pigment cells.