Alteration of homeobox gene expression by N-ras transformation of PA-1 human teratocarcinoma cells.

We used a series of cell clones from a human teratocarcinoma cell line, PA-1, to study the effect of transformation by an activated N-ras oncogene on the expression of genes involved in retinoic acid (RA)-induced differentiation and growth regulation. Recently, it has been shown that expression of human HOX 2 genes is sequentially activated by RA beginning from Hox 2.9 at the 3' end of the HOX 2 cluster (A. Simeone, D. Acampora, L. Arcioni, P. W. Andrews, E. Boncinelli, and F. Mavilio, Nature [London] 346:763-766, 1990). We now report that six different genes of the cluster HOX 1 are sequentially induced by RA in a similar temporal pattern, beginning with genes at the 3' end of the cluster. However, in N-ras-transformed cell clones, RA-induced expression of these homeobox genes is delayed. Hox 1.4 and Hox 1.3, genes abundantly induced in nontransformed clones after 3 days of RA treatment, are expressed in N-ras-transformed cells only after 10 days of RA treatment. At this time, the cells' growth is arrested at very high density, and no differentiated morphologic characteristics are observed. Constitutive expression of a transfected Hox 1.4 gene under the control of a simian virus 40 promotor leads to differentiated cell morphology similar to that of the RA-induced phenotype and restores the growth-inhibitory effects of RA in N-ras-transformed cells. These observations provide evidence that enhanced proliferation in N-ras-transformed cells compromises teratocarcinoma cell differentiation by a mechanism that transiently suppresses homeobox gene induction and implies a central role for homeobox genes in RA-induced cell differentiation. We conclude that stimulation of a putative growth factor signal pathway, associated with ras-induced proliferation, transiently suppresses the induction of transcription factors functionally involved in cell growth and differentiation.

An alternatively spliced mRNA from the AP-2 gene encodes a negative regulator of transcriptional activation by AP-2.

AP-2 is a retinoic acid-inducible and developmentally regulated activator of transcription. We have cloned an alternative AP-2 transcript (AP-2B) from the human teratocarcinoma cell line PA-1, which encodes a protein differing in the C terminus from the previously isolated AP-2 protein (AP-2A). This protein contains the activation domain of AP-2 and part of the DNA binding domain but lacks the dimerization domain which is necessary for DNA binding. Analysis of overlapping genomic clones spanning the entire AP-2 gene proves that AP-2A and AP-2B transcripts are alternatively spliced from the same gene. Both transient and stable transfection experiments show that AP-2B inhibits AP-2 transactivator function, as measured by an AP-2-responsive chloramphenicol acetyltransferase reporter plasmid. Furthermore, constitutive AP-2B expression in PA-1 cells causes a retinoic acid-resistant phenotype, anchorage-independent growth in soft agar, and tumorigenicity in nude mice, in a fashion similar to transformation of these cells by oncogenes. To determine the mechanism by which AP-2B exerts its inhibitory function, we purified bacterially expressed AP-2A and AP-2B proteins. While bacterial AP-2B does not bind an AP-2 consensus site, it strongly inhibits binding of the endogenous AP-2 present in PA-1 cell nuclear extracts. However, DNA sequence-specific binding of bacterially expressed AP-2A cannot be inhibited by bacterially expressed AP-2B. Therefore, inhibition of AP-2 activity by the protein AP-2B may require an additional factor or modification supplied by nuclear extracts.

Human neoplastic and normal cells in tissue culture. I. Cell lines derived from malignant melanomas and normal melanocytes.

Forty-six cell lines derived from 31 human melanomas obtained from 28 patients were cultured. Fourteen of 16 lines have produced malignant tumors when injected into nude (thymus-deficient) mice. Tumors in 5 of the nude mice metastasized to distant lymph nodes and/or to the lungs of the mouse host. Extreme variability from line to line was observed for doubling time (34 to 106 hr), plating efficiency (0-86%), and melanin production. All tested lines had type B glucose-6-phosphate dehydrogenase, thereby excluding HeLa cell contamination. HeLa cells have been grown for some time in our laboratory. Our results clearly demonstrated that HeLa cell contamination does not occur invariably in heteroploid lines growing in a laboratory simultaneously with Hela cells, provided that proper care is taken to avoid such occurrence. Multiple cell lines derived from the same tumor had identical phosphoglucomutase enzyme phenotype, which suggested a lack of significant cross-contamination between the lines. Four long-term cultures of normal human uveal embryo melanocytes have also been established and characterized. Although all produced melanin after reaching saturation density, they differed from the melanoma cells morphologically; they were flat, not refringent, and lacked piling up and plating ability. When melanoma cells were exposed to bromodeoxyuridine (BUDR) for long periods, a phenotypic change toward non-neoplastic characteristics was observed. Cells became flat and not refringent and, when injected into nude mice, tumors appeared after a long latent period. These changes were completely reversible in vitro and in vivo. The BUDR-treated cultures were undistinguishable from the untreated mother cultures after 2 to 3 passages. Lines derived from tumors in nude mice (obtained by injection of BUDR-treated cells) were again indistinguishable from the untreated mother line. Normal melanocytes were mostly euploid; all the melanoma cells were aneuploid. All 29 cell lines derived from 14 patients had an average chromosome number higher than 46. Detailed group-by-group chromosome analysis always showed an excess of C chromosomes, which suggested that hyperreduplication of one or more C chromosomes is a specific characteristic of human melanomas.

Exceptional lethality for nude mice of cells derived from a primary human melanoma.

BRO human melanoma cells, obtained from a biopsy of a highly aggressive and malignant primary tumor, were grown as xenografts in nude mice and in cell culture. These cells were exceptionally tumorigenic and malignant for nude mice. NIH-II nude mice survived 11.0 +/- 0.4 (S.E.) and 14.1 +/- 0.4 days after i.p. inoculation of 10(7) or 10(6) BRO cells, respectively, and lethal tumors developed in all mice inoculated i.p. with only 10(3) cells. The doubling time (2.3 days) of the volume of tumors formed after s.c. inoculation was comparable to the doubling time of these cells in culture. After i.p. or s.c. inoculation, BRO cells metastasized to the diaphragm and lungs, causing respiratory failure in most of the host mice. The original tumor and the cell line derived from it had undifferentiated structures with prominent nuclei and very large nucleoli. Karyotype abnormalities included a gigantic A group chromosome, a large D group chromosome, and an unusual double centromere chromosome not found typically in human melanoma cells. Due to the short and reproducible survival times of nude mice inoculated i.p. with BRO cells, this model system may be useful for rapidly determining the effects of experimental treatment on the survival of hosts bearing human tumor cells.

Spontaneous abnormalities in normal fibroblasts from patients with Li-Fraumeni cancer syndrome: aneuploidy and immortalization.

Families of patients with the Li-Fraumeni cancer syndrome have an inherited pattern of sarcomas and various other types of cancers that follow a dominant mode of transmission, an early age of onset, and exhibit multiple primary tumors. As soft tissue sarcomas (including fibrosarcomas) are frequently observed with this syndrome, the in vitro growth characteristics of fibroblasts derived from skin biopsies of Li-Fraumeni syndrome patients were studied. Control fibroblasts maintained a normal morphology and eventually senesced in culture. Fibroblasts from seven of eight affected individuals developed changes in morphology, anchorage-independent growth, and chromosomal abnormalities. In a fashion similar to that of fibroblasts from normal donors they underwent a growth crisis during which their growth was slow, but they continued to grow past the point at which control samples had stopped dividing (35 population doublings). Fibroblasts from Li-Fraumeni cancer patients escape senescence, growing well beyond 35 population doublings with growth rates similar to early-passage cells. Patient fibroblasts maintain the morphology of a transformed cell but remain nontumorigenic in nude mice. These observations of the behavior of fibroblasts from patients with the Li-Fraumeni syndrome may have predictive value for the determination of gene carriers within these families who are at high risk of cancer.

Tumorigenic transformation of spontaneously immortalized fibroblasts from patients with a familial cancer syndrome.

Immortal cell lines arose spontaneously during in vitro culture of initially normal fibroblasts, MDAH041 and MDAH087, from patients with Li-Fraumeni familial cancer syndrome. Fibroblasts from a control donor, MDAH170, maintained a normal morphology and senesced at 31 population doublings. The immortal fibroblasts have several properties of transformed cells. In addition to having acquired an altered morphology and chromosomal anomalies, MDAH041 and MDAH087 have escaped from senescence, growing beyond 300 and 100 population doublings (pd), respectively. As early as 50 pd, these cells can be transformed by an activated H-ras oncogene to form tumors in nude mice. However, MDAH041 immortal cells were resistant to tumorigenic transformation by transfection with the v-abl oncogene.

Susceptibility to ras oncogene transformation is coregulated with signal transduction through growth factor receptors.

The human teratocarcinoma cell line PA-1 was derived from culturing ascites fluid cells from a patient with an ovarian germ line tumor. We previously described a non-neoplastic variant cloned from the PA-1 human teratocarcinoma cell line, clone 6, which at passage 40 was resistant to transformation by activated ras oncogenes. However, these cells could be transformed by a plasmid containing both myc and ras. Another PA-1 cell variant, clone 1, isolated at passage 63 and used 50 passages later becomes tumorigenic in nude mice after transfection with an activated ras oncogene (Tainsky et al., Anticancer Res., 8, 899-914, 1988). We report here that the progression from ras resistance to ras susceptibility occurs in both clone 1 and clone 6 cells during 25 passages in culture. In the presence of epidermal growth factor, transforming growth factor-alpha, and basic fibroblast growth factor, the ras-transformable cells exhibit anchorage independent growth, whereas the ras-resistant cells can not be growth stimulated by these growth factors. Similarly, ornithine decarboxylase (ODC) activity was inducible in ras susceptible and ras transformed cells by these growth factors, but not in the ras resistant cells. These differences are not due to the level and activity of epidermal growth factor receptor or to the level of expression of 25 proto-oncogenes.

Modulation of differentiation in PA-1 human teratocarcinoma cells after N-ras oncogene-induced tumorigenicity.

We have been studying the effect of oncogenes on differentiation using the human ovarian teratoma-derived cell line PA-1. From this study we have characterized variants representing four stages relevant to multistage carcinogenesis, two non-tumorigenic and two tumorigenic. The two non-tumorigenic cell variants differ in that one is resistant to transformation by ras oncogenes whereas the other can be transformed to tumorigenicity. When these non-tumorigenic PA-1 variants are treated with retinoic acid (RA), a morphogen, they stop dividing, begin to express homeobox genes, and change in morphology. Transfection of an activated N-ras oncogene into ras-resistant non-tumorigenic PA-1 cells does not alter the RA responsiveness of the cells, indicating that expression of the activated oncogene is not sufficient for blocking RA-induced differentiation. Spontaneous activation of an N-ras oncogene leading to tumorigenic transformants and gene transfer-induced N-ras transformants are resistant to these effects of RA. However, another spontaneous transformant of PA-1 cells that does not contain an activated N-ras is responsive to RA. We prepared somatic cell hybrids of the RA-non-responsive, N-ras-transformed and tumorigenic PA-1 cell and the RA-responsive, ras-resistant non-tumorigenic PA-1 cell; the hybrid cell lines continue to express the oncogene but are non-tumorigenic. These non-tumorigenic hybrids are responsive to RA with regard to morphological changes, growth arrest and induction of homeobox gene expression. Tumorigenic revertants of these hybrids arise as a result of the loss of some chromosomes; these hybrid cells express the oncogene but have lost RA responsiveness. These results indicate that tumorigenic transformation in general is not sufficient to induce RA resistance, and resistance to differentiation may be oncogene-specific. In addition, the expression of an activated N-ras oncogene alone is insufficient to induce resistance to RA and ras-induced tumorigenicity is necessary. Therefore, some feature of cellular metabolism that is altered by and discordantly segregates with tumorigenic transformation controls responsiveness to RA. This controlling element is presumably a tumor suppressor.

Telomerase activity during spontaneous immortalization of Li-Fraumeni syndrome skin fibroblasts.

Li-Fraumeni Syndrome (LFS) is characterized by heterozygous germline mutations in the p53 gene. Accompanied by genomic instability and loss or mutation of the remaining wild type p53 allele, a low frequency of spontaneous immortalization in LFS fibroblasts occurs. It is believed that the loss of p53 wild type function contributes to immortalization of these LFS fibroblasts, but it is not clear if this is sufficient. Because stabilization of telomere length is also thought to be a necessary step in immortalization, telomerase activity, expression of the telomerase RNA component (hTR) and telomere length were anlaysed at various passages during the spontaneous immortalization of LFS skin fibroblasts. One LFS strain which immortalized, MDAH087 (087), had no detectable telomerase activity whereas another LFS strain which immortalized, MDAH041 (041), had detectable telomerase activity. In preimmortal cells from both strains, hTR was not detected by in situ hybridization. Immortal 087 cells remained negative for hTR, while immortal 041 cells demonstrated strong hTR in situ hybridization signals. 087 cells had long and heterogenous telomeres whereas telomeres of 041 cells had short, stable telomere lengths. Tumorigenicity studies in nude mice with ras-transformed 087 and 041 cells resulted in both cell lines giving rise to tumors and retaining telomerase status. Overall these results suggest that strain specificity may be important in telomerase re-activation and that both abrogation of p53 function and a mechanism to maintain telomeres are necessary for immortalization.

Multiple karyotypic abnormalities, including structural rearrangements of 11p, in cell lines from malignant melanomas.

Cell lines were obtained from three malignant melanoma patients by culturing cell suspensions from tumor biopsies. A total of six lines (I to VI) were established. One line each was established from the first two cases. Lines III and IV were established from two different methyl cellulose colonies derived from the primary tumor of case 3; line III was from a non-pigmented and line IV from a pigmented colony. Cloning of line IV resulted in two highly malignant (IV Cl 1 and IV Cl 3) and one less malignant (IV Cl 2) clone. Clone IV Cl 1 was inoculated intracardially in nude mice and gave rise to adrenal and brain metastases. Lines V and VI were derived from such metastases. Multiple structural and/or numerical chromosome abnormalities were detected in all lines. Line I had 57-61 chromosomes, with structural changes affecting 1p, 2p, 3q, 7p, 7q, 11p, 14q, 17q, and 22q, as well as one unidentified marker. Line II had 40-48 chromosomes, with structural changes of 1p, 1q, 4q, 5p, 6p, 8p, 11p, 11q, 14p, 20p, and two unidentified markers. Line III had 45 chromosomes, 6q+, del(11p), and a centric fusion between chromosomes 14 and 15. Line IV had 45-46 chromosomes. The clonal changes included rearrangements of 1p, 9p, 11p, and the centric fusion of chromosomes 14 and 15. Line V was pseudodiploid and contained aberrations of 1p, 9p, 11p, 14q, 20q, an isochromosome for 21q, and an unidentified marker. Finally, the pseudodiploid line VI had changes of 9p, 11p, centric fusion of chromosomes 14 and 15, and an unidentified marker. Although no single identical aberration was shared by all six lines, structural abnormalities of 11p were invariably present and, hence, might constitute a common cytogenetic feature in melanoma development. The most consistent difference between the amelanotic and melanotic lines derived from case 3 was the presence of a 6q+ marker in the former and a 9p+ marker in the latter.

PA-1, a human cell model for multistage carcinogenesis: oncogenes and other factors.

We have developed a cell system which utilizes the human teratocarcinoma cell line PA-1, from which we have characterized four stages of tumor progression. Soon after establishment in culture PA-1 cells revert and are no longer tumorigenic in athymic nude mice. Later, PA-1 cells as they are passaged in culture, become tumorigenic at passage 100. The transition from nontumorigenic to tumorigenic is the result of the biological effects of an activated N-ras oncogene and can be reproduced by transfection of the cloned oncogene into preneoplastic PA-1 cells. Certain preneoplastic cells (prior to passage 100) in this series are susceptible to transformation by single oncogenes while others are not. In studying the basis of this susceptibility to single oncogene induced transformation we have found that somatic cell hybrids between preneoplastic cells which can suppress ras-induced transformation and ras-transformed cells are non-tumorigenic. Therefore, we believe that the progression from ras suppressing to ras susceptibility may be due to the inactivation of a trans-dominant suppressor gene. Our system has identified at least three steps which lead to tumorigenicity; establishment of growth past senesence, activation of a ras oncogene, and inactivation of an oncogene suppressor function. Further genetic alterations are necessary for tumor dissemination and metastasis.

Coordinate control of growth and cytokeratin 13 expression by retinoic acid.

Retinoic acid (RA) modulates the growth and differentiation of various normal and malignant cells. These effects are most likely mediated by changes in gene expression. Genes whose expression is modulated by RA may be useful as markers of growth responsiveness to retinoids. Using differential cDNA cloning we identified 10 genes regulated by RA in the head and neck squamous cell carcinoma cell line MDA886Ln. Keratin (K) 13 gene expression was the gene expression most related to the degree of sensitivity of growth to RA, as K13 was not expressed in a series of RA-resistant cell lines. Our data suggest that low K13 expression may be mechanistically related to resistance to RA-induced growth inhibition.

N-ras oncogene causes AP-2 transcriptional self-interference, which leads to transformation.

Genetic alterations in elements of normal signal transduction mechanisms are known to be oncogenic events often resulting in aberrant activation of programs of gene transcription. We have investigated the effect of N-ras oncogene-induced tumorigenic transformation on the transcription factor AP-2. N-ras oncogene-induced transformation of human teratocarcinoma cells PA-1 results in sixfold elevated AP-2 mRNA levels. However, the level of AP-2-mediated trans-activation is dramatically inhibited in these cells. We show here that the high-level expression of AP-2 ultimately results in transcriptional "self-interference". The activation domain of AP-2, when fused to the DNA-binding domain of GAL4, is sufficient for self-interference. Non-N-ras PA-1 cells constitutively expressing AP-2 or GAL4-AP-2 fusion protein from an SV40 promoter exhibit reduced AP-2-mediated transcriptional activation, inhibition of differentiation, and promotion of anchorage-independent growth, properties that are similar to N-ras-transformed PA-1 cells. Thus, AP-2 is placed in the N-ras signal transduction pathway, and many of the biological effects of N-ras can be accomplished by overexpression of AP-2. This is the first evidence that inhibition of the activity of a transcription factor by self-interference contributes to a physiological process.