Expression and functional significance of VE-cadherin in aggressive human melanoma cells: Role in vasculogenic mimicry

We recently have introduced the term vasculogenic mimicry to describe the unique ability of aggressive melanoma tumor cells to form tubular structures and patterned networks in three-dimensional culture, which “mimics” embryonic vasculogenic networks formed by differentiating endothelial cells. In the current study, we address the biological significance of several endothelial-associated molecules (revealed by microarray analysis) with respect to expression and function in highly aggressive and poorly aggressive human cutaneous melanoma cell lines (established from the same patient). In a comparative analysis, CD31 was not expressed by any of the melanoma cell lines, whereas TIE-1 (tyrosine kinase with Ig and epidermal growth factor homology domains-1) was strongly expressed in the highly aggressive tumor cells with a low level of expression in one of the poorly aggressive cell lines. Vascular endothelial (VE)-cadherin was exclusively expressed by highly aggressive melanoma cells and was undetectable in the poorly aggressive tumor cells, suggesting the possibility of a vasculogenic switch. Down-regulation of VE-cadherin expression in the aggressive melanoma cells abrogated their ability to form vasculogenic networks and directly tested the hypothesis that VE-cadherin is critical in melanoma vasculogenic mimicry. These results highlight the plasticity of aggressive melanoma cells and call into question their possible genetic reversion to an embryonic phenotype. This finding could pose a significant clinical challenge in targeting tumor cells that may masquerade as circulating endothelial cells or other embryonic-like stem cells.

Molecular cloning of a human melanoma-associated chondroitin sulfate proteoglycan.

A human melanoma-associated chondroitin sulfate proteoglycan (MCSP), recognized by mAb 9.2.27, plays a role in stabilizing cell-substratum interactions during early events of melanoma cell spreading on endothelial basement membranes. We report here the molecular cloning and nucleotide sequencing of cDNA encoding the entire core protein of human MCSP and provide its deduced amino acid sequence. This core protein contains an open reading frame of 2322 aa, encompassing a large extracellular domain, a hydrophobic transmembrane region, and a relatively short cytoplasmic tail. Northern blot analysis indicated that MCSP cDNA probes detect a single 8.0-kb RNA species expressed in human melanoma cell lines. In situ hybridization experiments with a segment of the MCSP coding sequence localized MCSP mRNA in biopsies prepared from melanoma skin metastases. Multiple human Northern blots with an MCSP-specific probe revealed a strong hybridization signal only with melanoma cells and not with other human cancer cells or a variety of human fetal and adult tissues. These data indicate that MCSP represents an integral membrane chondroitin sulfate proteoglycan expressed by human malignant melanoma cells. The availability of cDNAs encoding MCSP should facilitate studies designed to establish correlations between structure and function of this molecule and help to establish its role in the progression of human malignant melanoma.

The activation of human gene MAGE-1 in tumor cells is correlated with genome-wide demethylation.

Human gene MAGE-1 encodes tumor-specific antigens that are recognized on melanoma cells by autologous cytolytic T lymphocytes. This gene is expressed in a significant proportion of tumors of various histological types, but not in normal tissues except male germ-line cells. We reported previously that reporter genes driven by the MAGE-1 promoter are active not only in the tumor cell lines that express MAGE-1 but also in those that do not. This suggests that the critical factor causing the activation of MAGE-1 in certain tumors is not the presence of the appropriate transcription factors. The two major MAGE-1 promoter elements have an Ets binding site, which contains a CpG dinucleotide. We report here that these CpG are demethylated in the tumor cell lines that express MAGE-1, and are methylated in those that do not express the gene. Methylation of these CpG inhibits the binding of transcription factors, as seen by mobility shift assay. Treatment with the demethylating agent 5-aza-2'-deoxycytidine activated gene MAGE-1 not only in tumor cell lines but also in primary fibroblasts. Finally, the overall level of CpG methylation was evaluated in 20 different tumor cell lines. It was inversely correlated with the expression of MAGE-1. We conclude that the activation of MAGE-1 in cancer cells is due to the demethylation of the promoter. This appears to be a consequence of a genome-wide demethylation process that occurs in many cancers and is correlated with tumor progression.

Surface-epitope masking and expression cloning identifies the human prostate carcinoma tumor antigen gene PCTA-1 a member of the galectin gene family.

The selective production of monoclonal antibodies (mAbs) reacting with defined cell surface-expressed molecules is now readily accomplished with an immunological subtraction approach, surface-epitope masking (SEM). Using SEM, prostate carcinoma (Pro 1.5) mAbs have been developed that react with tumor-associated antigens expressed on human prostate cancer cell lines and patient-derived carcinomas. Screening a human LNCaP prostate cancer cDNA expression library with the Pro 1.5 mAb identifies a gene, prostate carcinoma tumor antigen-1 (PCTA-1). PCTA-1 encodes a secreted protein of approximately 35 kDa that shares approximately 40% sequence homology with the N-amino terminal region of members of the S-type galactose-binding lectin (galectin) gene family. Specific galectins are found on the surface of human and marine neoplastic cells and have been implicated in tumorigenesis and metastasis. Primer pairs within the 3' untranslated region of PCTA-1 and reverse transcription-PCR demonstrate selective expression of PCTA-1 by prostate carcinomas versus normal prostate and benign prostatic hypertrophy. These findings document the use of the SEM procedure for generating mAbs reacting with tumor-associated antigens expressed on human prostate cancers. The SEM-derived mAbs have been used for expression cloning the gene encoding this human tumor antigen. The approaches described in this paper, SEM combined with expression cloning, should prove of wide utility for developing immunological reagents specific for and identifying genes relevant to human cancer.

Loss of oncogenic ras expression does not correlate with loss of tumorigenicity in human cells.

ras oncogenes are mutated in at variety of human tumors, which suggests that they play an important role in human carcinogenesis. To determine whether continued oncogenic ras expression is necessary to maintain the malignant phenotype, we studied the human fibrosarcoma cell line, HT1080, which contains one mutated and one wild-type N-ras allele. We isolated a variant of this cell line that no longer contained the mutated copy of the N-ras gene. Loss of mutant N-ras resulted in cells that displayed a less transformed phenotype characterized by a flat morphology, decreased growth rate, organized actin stress fibers, and loss of anchorage-independent growth. The transformed phenotype was restored following reintroduction of mutant N-ras. Although loss of the oncogenic N-ras drastically affected in vitro growth parameters, the variant remained tumorigenic in nude mice indicating that mutated N-ras expression is not necessary for maintenance of the tumorigenic phenotype. We confirmed this latter observation in colon carcinoma cell lines that have lost activated K-ras expression via targeted knockout of the mutant K-ras gene.

Serological analysis of Melan-A(MART-1), a melanocyte-specific protein homogeneously expressed in human melanomas.

Recent progress in the structural identification of human melanoma antigens recognized by autologous cytotoxic T cells has led to the recognition of a new melanocyte differentiation antigen, Melan-A(MART-1). To determine the properties of the Melan-A gene product, Melan-A recombinant protein was produced in Escherichia coli and used to generate mouse monoclonal antibodies (mAbs). Two prototype mAbs, A103 and A355, were selected for detailed study. Immunoblotting results with A103 showed a 20-22-kDa doublet In Melan-A mRNA positive melanoma cell lines and no reactivity with Melan-A mRNA-negative cell lines. A355, in addition to the 20-22-kDa doublet, recognized several other protein species in Melan-A mRNA-positive cell lines. Immunocytochemical assays on cultured melanoma cells showed specific and uniform cytoplasmic staining in Melan-A mRNA-positive cell lines. Immunohistochemical analysis of normal human tissues with both mAbs showed staining of adult melanocytes and no reactivity with the other normal tissues tested. Analysis of 21 melanoma specimens showed homogenous staining of tumor cell cytoplasm in 16 of 17 Melan-A mRNA-positive cases and no reactivity with the three Melan-A mRNA-negative cases.

Cell cycle-dependent modulation of telomerase activity in tumor cells.

Telomerase is a ribonucleoprotein complex that is thought to add telomeric repeats onto the ends of chromosomes during the replicative phase of the cell cycle. We tested this hypothesis by arresting human tumor cell lines at different stages of the cell cycle. Induction of quiescence by serum deprivation did not affect telomerase activity. Cells arrested at the G1/S phase of the cell cycle showed similar levels of telomerase to asynchronous cultures; progression through the S phase was associated with increased telomerase activity. The highest level of telomerase activity was detected in S-phase cells. In contrast, cells arrested at G2/M phase of the cell cycle were almost devoid of telomerase activity. Diverse cell cycle blockers, including transforming growth factor beta1 and cytotoxic agents, also caused inhibition of telomerase activity. These results establish a direct link between telomerase activity and progression through the cell cycle.

Prostaglandin H synthase 2 is expressed abnormally in human colon cancer: evidence for a transcriptional effect.

Evidence from epidemiological studies, clinical trials, and animal experiments indicates that inhibitors of prostaglandin synthesis lower the risk of colon cancer. We tested the hypothesis that abnormal expression of prostaglandin H synthase 2 (PHS-2), which can be induced by oncogenes and tumor promoters, occurs during colon carcinogenesis by examining its level in colon tumors. Human colon cancers were found to have an increased expression of PHS-2 mRNA compared with normal colon specimens from the same patient (n = 5). In situ hybridization showed that the neoplastic colonocytes had increased expression of PHS-2 (n = 4). Additionally, five colon cancer cell lines were shown to express high levels of PHS-2 mRNA even in the absence of a known inducer of PHS-2. To study the basis for this increased gene expression, we transfected a colon cancer cell line, HCT-116, with a reporter gene containing 2.0 kb of the 5' regulatory sequence of the PHS-2 gene. Constitutive transcription of the reporter gene was observed, whereas normal control cell lines transcribed the reporter only in response to an exogenous agonist. We conclude that PHS-2 is transcribed abnormally in human colon cancers and that this may be one mechanism by which prostaglandins or related compounds that support carcinogenesis are generated.

Normal human serum contains a natural IgM antibody cytotoxic for human neuroblastoma cells.

Neuroblastoma (NB) is characterized by the second highest spontaneous regression of any human malignant disorder, a phenomenon that remains to be elucidated. In this study, a survey of 94 normal human adult sera revealed a considerable natural humoral cytotoxicity against human NB cell lines in approximately one-third of the tested sera of both genders. Specific cell killing by these sera was in the range of 40% to 95%. Serum cytotoxicity was dependent on an intact classical pathway of complement. By several lines of evidence, IgM antibodies were identified as the cytotoxic factor in the sera. Further analyses revealed that a 260-kDa protein was recognized by natural IgM of cytotoxic sera in Western blots of NB cell extracts. The antigen was expressed on the surface of seven human NB cell lines but not on human melanoma or other control tumor cell lines derived from kidney, pancreas, colon, bone, skeletal muscle, lymphatic system, and bone marrow. Furthermore, no reactivity was observed with normal human fibroblasts, melanocytes, and epidermal keratinocytes. The antigen was expressed in vivo as detected by immunohistochemistry in both the tumor of a NB patient and NB tumors established in nude rats from human NB cell lines. Most interestingly, the IgM anti-NB antibody was absent from the sera of 11 human NB patients with active disease. The anti-NB IgM also could not be detected in tumor tissue obtained from a NB patient. Collectively, our data suggest the existence of a natural humoral immunological tumor defense mechanism, which could account for the in vivo phenomenon of spontaneous NB tumor regression.

Stimulation of intrachromosomal homologous recombination in human cells by electroporation with site-specific endonucleases.

In somatic mammalian cells, homologous recombination is a rare event. To study the effects of chromosomal breaks on frequency of homologous recombination, site-specific endonucleases were introduced into human cells by electroporation. Cell lines with a partial duplication within the HPRT (hypoxanthine phosphoribosyltransferase) gene were created through gene targeting. Homologous intrachromosomal recombination between the repeated regions of the gene can reconstruct a functioning, wild-type gene. Treatment of these cells with the restriction endonuclease Xba I, which has a recognition site within the repeated region of HPRT homology, increased the frequency or homologous recombination bv more than 10-fold. Recombination frequency was similarly increased by treatment with the rare-cutting yeast endonuclease PI-Sce I when a cleavage site was placed within the repeated region of HPRT. In contrast, four restriction enzymes that cut at positions either outside of the repeated regions or between them produced no change in recombination frequency. The results suggest that homologous recombination between intrachromosomal repeats can be specifically initiated by a double-strand break occurring within regions of homology, consistent with the predictions of a model.

Amplification of AKT2 in human pancreatic cells and inhibition of AKT2 expression and tumorigenicity by antisense RNA.

We previously demonstrated that the putative oncogene AKT2 is amplified and overexpressed in some human ovarian carcinomas. We have now identified amplification of AKT2 in approximately 10% of pancreatic carcinomas (2 of 18 cell lines and 1 of 10 primary tumor specimens). The two cell lines with altered AKT2 (PANC1 and ASPC1) exhibited 30-fold and 50-fold amplification of AKT2, respectively, and highly elevated levels of AKT2 RNA and protein. PANC1 cells were transfected with antisense AKT2, and several clones were established after G418 selection. The expression of AKT2 protein in these clones was greatly decreased by the antisense RNA. Furthermore, tumorigenicity in nude mice was markedly reduced in PANC1 cells expressing antisense AKT2 RNA. To examine further whether overexpression of AKT2 plays a significant role in pancreatic tumorigenesis, PANC1 cells and ASPC1 cells, as well as pancreatic carcinoma cells that do not overexpress AKT2 (COLO 357), were transfected with antisense AKT2, and their growth and invasiveness were characterized by a rat tracheal xenotransplant assay. ASPC1 and PANC1 cells expressing antisense AKT2 RNA remained confined to the tracheal lumen, whereas the respective parental cells invaded the tracheal wall. In contrast, no difference was seen in the growth pattern between parental and antisense-treated COLO 357 cells. These data suggest that overexpression of AKT2 contributes to the malignant phenotype of a subset of human ductal pancreatic cancers.

Eradication of human hepatic and pulmonary melanoma metastases in SCID mice by antibody-interleukin 2 fusion proteins.

Antibody-cytokine fusion proteins combine the unique targeting ability of antibodies with the multifunctional activity of cytokines. Here, we demonstrate the therapeutic efficacy of such constructs for the treatment of hepatic and pulmonary metastases of different melanoma cell lines. Two antibody-interleukin 2 (IL-2) fusion proteins, ch225-IL2 and ch14.18-IL2, constructed by fusion of a synthetic sequence coding for human IL-2 to the carboxyl end of the Cgamma1 gene of the corresponding antibodies, were tested for their therapeutic efficacy against xenografted human melanoma in vivo. Tumor-specific fusion proteins completely inhibited the growth of hepatic and pulmonary metastases in C.B-17 scid/scid mice previously reconstituted with human lymphokine-activated killer cells, whereas treatment with combinations of the corresponding antibodies plus recombinant IL-2 only reduced the tumor load. Even when treatment with fusion proteins was delayed up to 8 days after inoculation of tumor cells, it still resulted in complete eradication of micrometastases that were established at that time point. Selection of tumor cell lines expressing or lacking the targeted antigen of the administered fusion protein proved the specificity of the observed antitumor effect. Biodistribution analysis demonstrated that the tumor-specific fusion protein accumulated not only in subcutaneous tumors but also in lungs and livers affected with micrometastases. Survival times of animals treated with the fusion protein were more than doubled as compared to those treated with the combination of the corresponding antibody plus IL-2. Our data demonstrate that an immunotherapeutic approach using cytokines targeted by antibodies to tumor sites has potent effects against disseminated human melanoma.

Mutagenesis by third-strand-directed psoralen adducts in repair-deficient human cells: high frequency and altered spectrum in a xeroderma pigmentosum variant.

Psoralen-conjugated triple-helix-forming oligonucleotides have been used to generate site-specific mutations within mammalian cells. To investigate factors influencing the efficiency of oligonucleotide-mediated gene targeting, the processing of third-strand-directed psoralen adducts was compared in normal and repair-deficient human cells. An unusually high mutation frequency and an altered mutation pattern were seen in xeroderma pigmentosum variant (XPV) cells compared with normal, xeroderma pigmentosum group A (XPA), and Fanconi anemia cells. In XPV, targeted mutations were produced in the supF reporter gene carried in a simian virus 40 vector at a frequency of 30%, 3-fold above that in normal or Fanconi anemia cells and 6-fold above that in XPA. The mutations generated by targeted psoralen crosslinks and monoadducts in the XPV cells formed a pattern distinct from that in the other three cell lines, with mutations occurring not just at the damaged site but also at adjacent base pairs. Hence, the XPV cells may have an abnormality in trans-lesion bypass synthesis during repair and/or replication, implicating a DNA polymerase or an accessory factor as a basis of the defect in XPV. These results may help to elucidate the repair deficiency in XPV, and they raise the possibility that genetic manipulation via triplex-targeted mutagenesis may be enhanced by modulation of the XPV-associated activity in normal cells.

In vitro synthesis of oncogenic human papillomaviruses requires episomal genomes for differentiation-dependent late expression.

Human papillomavirus (HPV) types 16, 18, 31, and 51 are the etiologic agents of many anogenital cancers including those of the cervix. These "high risk" HPVs specifically target genital squamous epithelia, and their lytic life cycle is closely linked to epithelial differentiation. We have developed a genetic assay for HPV functions during pathogenesis using recircularized cloned HPV 31 genomes that were transfected together with a drug resistance marker into monolayer cultures of normal human foreskin keratinocytes, the natural host cell. After drug selection, cell lines were isolated that stably maintained HPV 31 DNA as episomes and underwent terminal differentiation when grown in organotypic raft cultures. In differentiated rafts, the expression of late viral genes, amplification of viral DNA, and production of viral particles were detected in suprabasal cells. This demonstrated the ability to synthesize HPV 31 virions from transfected DNA templates and allowed an examination of HPV functions during the vegetative viral life cycle. We then used this system to investigate whether an episomal genome was required for the induction of late viral gene expression. When an HPV 31 genome (31E1*) containing a missense mutation in the E1 open reading frame was transfected into normal human keratinocytes, the mutant viral sequences were found to integrate into the host cell chromosomal DNA with both early and late regions intact. While high levels of early viral gene transcription were observed, no late gene expression was detected in rafts of cell lines containing the mutant viral genome despite evidence of terminal differentiation. Therefore, the induction of late viral gene expression required that the viral genomes be maintained as extrachromosomal elements, and terminal differentiation alone was not sufficient. These studies provide the basis for a detailed examination of HPV functions during viral pathogenesis.

Expression of the fructose transporter GLUT5 in human breast cancer.

The primary metabolic characteristic of malignant cells is an increased uptake of glucose and its anaerobic metabolism. We studied the expression and function of the glucose transporters in human breast cancer cell lines and analyzed their expression in normal and neoplastic primary human breast tissue. Hexose uptake assays and immunoblotting experiments revealed that the breast carcinoma cell lines MCF-7 and MDA-468 express the glucose transporters GLUT1 and GLUT2, isoforms expressed in both normal and neoplastic breast tissue. We also found that the breast cancer cell lines transport fructose and express the fructose transporter GLUT5. Immunolocalization studies revealed that GLUT5 is highly expressed in vivo in human breast cancer but is absent in normal human breast tissue. These findings indicate that human breast cancer cells have a specialized capacity to transport fructose, a metabolic substrate believed to be used by few human tissues. Identification of a high-affinity fructose transporter on human breast cancer cells opens opportunities to develop novel strategies for early diagnosis and treatment of breast cancer.