Tyrosine levels regulate the melanogenic response to alpha-melanocyte-stimulating hormone in human melanocytes: implications for pigmentation and proliferation.

Melanocyte-stimulating hormone (alpha-MSH) increases cytosolic levels of cAMP as well as tyrosinase activity in murine melanocytes. These activities depend upon the presence of melanin precursors and may differ in human melanocytes. In this study, we demonstrate that high levels of tyrosine (3.7 mM), the chief melanin precursor, reduced the proliferative effect of alpha-MSH and altered human melanocyte morphology as compared to treatment with low (25-30 microM, half-physiological) levels of tyrosine. The anti-proliferative effect of high levels of tyrosine was not restricted to alpha-MSH; tyrosine also reduced proliferation induced by forskolin, a direct activator of the cAMP pathway. Exposure to low tyrosine levels and alpha-MSH induced a dendritic morphology; in the presence of high tyrosine and alpha-MSH, melanocytes displayed large, pigmented cell bodies and less dendricity. Exposure to alpha-MSH in the presence of low tyrosine for up to 9 days did not appreciably increase melanin levels, but culturing the human melanocytes in high levels of tyrosine with alpha-MSH increased melanin levels 10-50-fold, depending on the pigmentation background of the donor. A greater induction of melanin accumulation was observed in melanocytes derived from light-skinned donors than was observed in cells obtained from dark-skinned donors. The poor ability of alpha-MSH to stimulate melanin synthesis was not caused by a lack of induction of melanogenic proteins, as alpha-MSH increased the expression of microphthalmia (MITF), tyrosinase, dopachrome tautomerase (DCT), and Pmel-17, compared to untreated cells or cells stimulated by phorbol ester alone, regardless of tyrosine levels. DCT levels were greatly induced by low tyrosine with alpha-MSH, but were dramatically decreased by high tyrosine with alpha-MSH. Interestingly, in this same medium (high tyrosine), MITF levels also decreased after 2 weeks and were barely detectable by the third week. Despite the absence of MITF at 3 weeks of treatment in high tyrosine medium, tyrosinase levels remained high, thereby suggesting that additional factors must be responsible for tyrosinase transcription in human melanocytes. Our results indicate that tyrosine levels can regulate the proliferative activity induced by alpha-MSH, as well as the extent of melanogenesis in normal human melanocytes. The significance of this work is that tyrosine levels may be part of the mechanism that switches melanocytes out of a proliferative status and into a melanin-synthesizing, terminally differentiated phenotype.

High levels of expression of p27KIP1 and cyclin E in invasive primary malignant melanomas.

Cancer cells have abnormal cell cycle regulation which favors accelerated proliferation, chromosomal instability, and resistance to the senescence response. Although the p16INK4a locus is the most prominent susceptibility locus for familial melanomas, the low frequency of p16 mutations in sporadic melanomas suggests additional alterations in other cell cycle regulatory genes. Here we used primary melanoma tumors to reveal early cell cycle alterations that could be masked in advanced metastatic lesions due to their inherently high genetic instability. Unexpectedly, the cyclin-dependent kinase inhibitors p27KIP1 and/or p21Waf-1/SDI-1 were found to be expressed in 13 of 18 (72%) of the primary melanomas with a Breslow thickness greater than 0.076 mm. In general, p27 and/or p21 staining in the primary tumors correlated with low Ki-67 index. Importantly, most of the p21- and p27-positive tumors expressed high levels of cyclin D1 and cyclin E. In proliferating cells p27 is predominantly associated with cyclin D-CDK4 complexes, but does not inhibit the kinase activity, whereas in quiescent cells p27 is found associated with inactive CDK2 complexes. p27 was also expressed at high levels in proliferating primary melanomas in culture, and found to be associated with active cyclin E-CDK2 complexes containing high levels of cyclin E. It is thus likely that accumulation of cyclin E overcomes the potent inhibitory activity of p27 and p21 in CDK2 complexes. Of the primary melanomas with no indication of invasiveness, only three of 15 (20%) were positive for p27 and/or p21. We propose that high levels of p27 and p21 may confer upon melanoma tumors their characteristic resistance to conventional therapies. In turn, high levels of cyclins E and D1 may contribute to unlimited proliferation in primary melanomas that express the tumor suppressor p16INK4. J Invest Dermatol 113:1039-1046 1999

Antisense oligonucleotide blockade of connexin expression during embryonic bone formation: evidence of functional compensation within a multigene family.

Prior studies in our laboratory demonstrated the presence of gap junction proteins (connexins) throughout intramembranous bone formation [Minkoff et al. (1994) Anat Embryol 190:231-241]. In addition, two members of the connexin family of gap junction proteins, connexin 43 (Cx43; Gj alpha 1) and connexin 45 (Cx45; Gj alpha 6), were found by Civitelli et al. [1993; J Clin Invest 91:1888-1896] to be associated, specifically, with osteogenesis. Recently, however, a null mutation in the gene encoding Gj alpha 1 in mice has been produced by Reaume et al. [1995; Science 267:1831-1834]. Gj alpha 1 null homozygotes survived to term but died at birth of heart abnormalities. Examination of the null homozygous embryos, surprisingly, did not reveal overt histological or anatomical abnormalities in any organ system other than the heart. In view of this, the present investigation was initiated in order to evaluate bone formation under conditions in which the expression of Gj alpha 1 and Gj alpha 6, the connexins specifically associated with osteogenesis, had been perturbed, individually as well as in combination. An in vitro system employing organ cultures of dissociated embryonic chick mandibular mesenchyme was employed. Mesenchyme was cultured in the presence and absence of sense and antisense oligodeoxynucleotides (ODN), ranging in length from 15 to 24 mer and containing sequences that included the initiation codon of Gj alpha 1 and of Gj alpha 6. In cultures of mesenchyme, grown for 6 to 13 days in the presence of the combined antisense ODNs to Gj alpha 1 and Gj alpha 6, bone formation was markedly reduced or absent. By contrast, in cultures grown in medium containing the combination of corresponding sense ODNs to both Gj alpha 1 and Gj alpha 6, bone formation was evident. In addition, when cultures were grown in the presence of antisense or sense ODNs to either Gj alpha 1 or Gj alpha 6, individually, bone formation was seen. Immunohistochemical analysis of connexin expression revealed intense immunoreactive signal to Gj alpha 1 and Gj alpha 6 in bone of the control explants, in which no ODNs were present; in those cultures in which either Gj alpha 1 and Gj alpha 6 antisense ODNs were present, however, the expression of the respective connexin protein was either significantly reduced or absent. Further, in those explants in which Gj alpha 1 expression was blocked, immunoreactive signal to Gj alpha 6 appeared to have been amplified in regions of developing bone. These results suggest that, in avian osteogenic tissue, when Gj alpha 1 protein expression has been impeded another related connexin protein (Gj alpha 6) may subserve the functions of the missing connexin. The findings of this study, therefore, support the hypothesis that, within the connexin gene family, functional compensation can occur.

Association of cell and substrate adhesion molecules with connexin43 during intramembranous bone formation.

Prior studies in our laboratory have demonstrated an association of specific gap junction proteins with intramembranous bone formation in the avian mandible. The purpose of the present study was to extend these observations by determining if there was a relationship between the expression of one of the gap junction proteins examined previously (connexin43) and the expression of specific cell adhesion (CAM) and/or substrate adhesion (SAM) molecules [i.e. NCAM, A-CAM (N-cadherin) and tenascin (tenascin-C)] that have previously been shown to be associated with bone formation. Immunohistochemical localization of connexin43, tenascin, NCAM and N-cadherin was performed on serial sections of mandibles of chick embryos from 6 to 12 days of incubation. Analysis of adjacent serial sections revealed that the NCAM and tenascin immunostaining that appeared initially on the lateral aspect of Meckel's cartilage preceded the overt expression of trabecular bone. At subsequent stages, NCAM and tenascin staining gradually overlapped the region of connexin43 expression. In contrast, the expression of N-cadherin was found to colocalize with that of connexin43 from the first appearance of connexin43 expression. Most significantly, although the domains of NCAM and tenascin expression were initially separate from that of connexin43, bone formation originated only in the region where these domains intersected. These findings suggest that, of the CAMs and SAMs examined, N-cadherin appears to be associated with the establishment of cell contacts responsible for the presence and/or maintenance of connexin43-mediated gap junctional communication, while tenascin and NCAM appear to be associated, in a more specific manner, with processes that accompany the overt expression of the osteogenic phenotype.

Specific association between the human DNA repair proteins XPA and ERCC1.

Processing of DNA damage by the nucleotide-excision repair pathway in eukaryotic cells is most likely accomplished by multiprotein complexes. However, the nature of these complexes and the details of the molecular interactions between DNA repair factors are for the most part unknown. Here, we demonstrate both in vivo, using the two-hybrid system, and in vitro, using recombinant proteins, that the human repair factors XPA and ERCC1 specifically interact. In addition, we report an initial determination of the domains in ERCC1 and XPA that mediate this interaction. These results suggest that XPA may play a role in the localization or loading of an incision complex, composed of ERCC1 and possibly other repair factors, onto a damaged site.

Characterization of molecular defects in xeroderma pigmentosum group C.

Xeroderma pigmentosum (XP) is a rare autosomal recessive disease of humans characterized by an accelerated chronic degeneration of sun-exposed areas of the body, including an elevated risk of developing cancers of the skin. We recently reported the isolation of a gene XPCC that complements the repair deficiency of cultured XP-C cells. Here we report the results of a characterization of XPCC at the nucleotide level in five XP-C cell lines. Each cell line exhibited a unique mutation that correlated well with the cellular DNA repair deficiency and the clinical severity of the disease. These results extend our previous observations and indicate that defects in XPCC cause Xeroderma pigmentosum group C.

Correlation between Ha-ras gene amplification and spontaneous metastasis in NIH 3T3 cells transfected with genomic DNA from human skin cancers.

Our previous studies have shown that DNA from some human skin cancers contained activated Ha-ras oncogenes capable of inducing tumorigenic transformation when introduced into NIH 3T3 cells by DNA-mediated gene transfer. In addition, we found that NIH 3T3 cells transfected with DNA from one of the human skin cancers not only induced s.c. tumors at the site of injection but also metastasized spontaneously to the lungs in 100 per cent of nude mice injected. In this present study we examined the relationship between Ha-ras oncogene amplification and metastatic potential in tumors induced by various human skin cancer DNA-transfectants. Total cellular RNA was extracted from nude mouse tumor cell lines and analyzed by northern blot hybridization to a 32P-labeled, nick-translated Ha-ras probe. The metastatic potential of nude mouse tumor cell lines was assessed by their ability to form lung colonies after i.v. or s.c. injection. It was found that only the tumors expressing high levels of Ha-ras gene transcripts induced spontaneous metastasis after s.c. injection. There appeared to be little correlation between the level of Ha-ras oncogene amplification and experimental metastasis. These results suggest that amplification and overexpression of Ha-ras oncogene may play a role in the escape of cells from the primary tumor rather than in the ability of cells to survive in the circulatory system and colonize secondary sites.

Deletion of the c-Ha-ras-1 allele in human skin cancers.

Previous studies have shown that c-Ha-ras-1 and other genes located on the short arm of chromosome 11 are frequently lost in a number of human tumors. We investigated whether similar losses of the c-Ha-ras-1 allele occurred in human squamous cell carcinomas (SCCs) and basal cell carcinomas (BCCs). DNAs were isolated from 35 pairs of skin tumors (25 BCCs and 10 SCCs) and matching normal skin from the same patients and analyzed for c-Ha-ras gene polymorphism by Southern blot hybridization. Sixteen BCC patients and 1 SCC patient were constitutionally heterozygous for the c-Ha-ras gene in their normal skin DNA. Of these 17 patients, five patients (four with BCC and one with SCC) (29%) showed loss of one of the c-Ha-ras alleles in their tumor DNA. One of the constitutionally heterozygous BCC patients exhibited deletion of the 6.6-kb c-Ha-ras allele and an extra copy of the 7.8-kb allele. In summary, loss of heterozygosity at the c-Ha-ras locus occurred frequently (29%) in the 17 human skin cancers studied. However, our finding that 90% of the patients with SCC, as opposed to 36% of the patients with BCC, had only one of the c-Ha-ras alleles in their normal skin tissue requires further study. Whether c-Ha-ras homozygosity has any bearing on genetic susceptibility to SCC remains to be established.

Detection and identification of activated oncogenes in human skin cancers occurring on sun-exposed body sites.

High-molecular-weight DNA isolated from eight fresh human skin cancers occurring on sun-exposed body sites were assayed for their ability to transform NIH 3T3 cells. A cotransfection protocol using pSV2-neo DNA, which confers resistance to the antibiotic G418, was used to select cells that had taken up the transfected DNA. About 2 weeks after transfection, G418-resistant colonies were pooled and injected s.c. into athymic nude mice. The NIH 3T3 cells transfected with DNA from six of the human skin cancers induced tumors in nude mice. DNAs from all six tumor cell lines contained human alu sequences. Southern blot hybridization with ras-specific probes revealed that DNAs from the four alu-rich tumors contained the human Ha-ras oncogene, in addition to that of the NIH 3T3 controls. In contrast, DNAs from the other two tumors did not contain any of the known oncogenes tested, except those endogenous to NIH 3T3 cells. DNAs from three of four first cycle tumorigenic transformants gave rise to morphologically transformed foci when assayed in a second cycle of transfection. DNAs from all three secondary transformants contained discrete human alu sequences, and in addition, contained Ha-ras sequences similar to those present in their respective primary transformants. Interestingly, DNA from both primary and secondary transformants of one particular human squamous cell carcinoma contained highly amplified copies of the Ha-ras oncogene. These results suggest that activation of the Ha-ras oncogene may be common in human skin cancers originating on sun-exposed body sites. Further characterization of the Ha-ras oncogenes present in these human skin cancers may provide information on the molecular mechanisms by which UV radiation of the sun induces human neoplasms on exposed body sites.

Simultaneous transfer of tumorigenic and metastatic phenotypes by transfection with genomic DNA from a human cutaneous squamous cell carcinoma.

High-molecular-weight genomic DNA isolated from a human cutaneous squamous cell carcinoma (AS) was assayed for its ability to induce tumorigenic transformation of NIH 3T3 cells. Subcutaneous injection of NIH 3T3 cells cotransfected with DNAs from AS tumor and pSV2-neo plasmid not only induced tumors at the site of injection, but also metastasized spontaneously to the lungs in 100% of nude mice injected. DNA isolated from a representative primary tumor and a metastasis was again used in a second round of transfection. Injection of secondary transfectants into nude mice again resulted in induction of both subcutaneous tumors and spontaneous long metastases. Southern blot hybridization with ras-specific probes revealed that DNA from both primary tumors and metastases induced by AS tumor DNA contained highly amplified Ha-ras oncogene. Furthermore, DNAs from secondary tumors and metastases induced by DNA from a primary tumor and a metastasis also contained similar highly amplified Ha-ras oncogene. These results suggest that the amplified Ha-ras oncogene may be responsible for induction of both tumorigenic and metastatic phenotypes in NIH 3T3 cells transfected with DNA from AS tumor.