Search

Most large bowel cancers are moderately to well-differentiated adenocarcinomas comprised chiefly or entirely of glands lined by tall columnar cells. We have identified a subset of poorly differentiated colon carcinomas with a distinctive histopathological appearance that we term large cell minimally differentiated carcinomas (LCMDCs). These tumors likely include a group of poorly differentiated carcinomas previously described by others as medullary adenocarcinomas. To better understand the pathogenesis of these uncommon neoplasms, we compared molecular features of 15 LCMDCs to those present in 25 differentiated adenocarcinomas (DACs) of the colon. Tumors were examined for alterations commonly seen in typical colorectal carcinomas, including increased p53 and beta-catenin immunoreactivity, K-ras gene mutations, microsatellite instability, and loss of heterozygosity of markers on chromosomes 5q, 17p, and 18q. In addition, tumors were evaluated by immunohistochemistry for CDX2, a homeobox protein whose expression in normal adult tissues is restricted to intestinal and colonic epithelium. Markedly reduced or absent CDX2 expression was noted in 13 of 15 (87%) LCMDCs, whereas only 1 of the 25 (4%) DACs showed reduced CDX2 expression (P < 0.001). Nine of 15 (60%) LCMDCs had the high-frequency microsatellite instability phenotype, but only 2 of 25 (8%) DACs had the high-frequency microsatellite instability phenotype (P = 0.002). Our findings provide support for the hypothesis that the molecular pathogenesis of LCMDCs is distinct from that of most DACs. CDX2 alterations and DNA mismatch repair defects have particularly prominent roles in the development of LCMDCs.

Clinical and molecular findings suggest that the four major histological subtypes of ovarian carcinoma (serous, clear cell, mucinous, and endometrioid) likely represent distinct disease entities. Prior studies have shown that ovarian endometrioid adenocarcinomas (OEAs) often carry mutations in the CTNNB1 gene, which encodes beta-catenin, a critical component of the Wnt signaling pathway. However, the nature of other defects in the Wnt signaling pathway in ovarian carcinomas remains largely unknown. Thus, in 45 primary OEAs and two OEA-derived cell lines, we sought to comprehensively address the prevalence of and mechanisms underlying beta-catenin and Wnt pathway deregulation. CTNNB1 missense mutations were detected in 14 primary tumors. All mutations affected the NH(2)-terminal regulatory domain of beta-catenin, presumably rendering the mutant proteins resistant to degradation. Immunohistochemical studies revealed nuclear accumulation of beta-catenin in all but two tumors with CTNNB1 mutations. Two primary tumors lacking CTNNBI mutations showed strong nuclear immunoreactivity for beta-catenin. In one of the two tumors, biallelic inactivation of the APC gene was found. In the remaining 29 primary OEAs, unequivocal nuclear beta-catenin immunoreactivity was not observed, though a nonsense mutation in AXIN1 was observed in one tumor and a truncating frameshift mutation in AXIN2 was seen in another case. Both OEA-derived cell lines studied (TOV-112D and MDAH-2774) had elevated constitutive T-cell factor/lymphoid enhancer factor transcriptional activity. TOV-112D cells were shown to harbor mutant beta-catenin, whereas a missense AXIN1 sequence alteration was identified in MDAH-2774 cells. Collectively, our findings demonstrate frequent defects of the Wnt signaling pathway in a particular subtype of ovarian carcinomas, i.e., OEAs. Although mutations in the CTNNB1 gene are the most common mechanism of beta-catenin deregulation in OEAs, beta-catenin deregulation may also result from mutations in the APC, AXIN1, and AXIN2 genes.

Molecular classification of tumors based on their gene expression profiles promises to significantly refine diagnosis and management of cancer patients. The establishment of organ-specific gene expression patterns represents a crucial first step in the clinical application of the molecular approach. Here, we report on the gene expression profiles of 154 primary adenocarcinomas of the lung, colon, and ovary. Using high-density oligonucleotide arrays with 7129 gene probe sets, comprehensive gene expression profiles of 57 lung, 51 colon, and 46 ovary adenocarcinomas were generated and subjected to principle component analysis and to a cross-validated prediction analysis using nearest neighbor classification. These statistical analyses resulted in the classification of 152 of 154 of the adenocarcinomas in an organ-specific manner and identified genes expressed in a putative tissue-specific manner for each tumor type. Furthermore, two tumors were identified, one in the colon group and another in the ovarian group, that did not conform to their respective organ-specific cohorts. Investigation of these outlier tumors by immunohistochemical profiling revealed the ovarian tumor was consistent with a metastatic adenocarcinoma of colonic origin and the colonic tumor was a pleomorphic mesenchymal tumor, probably a leiomyosarcoma, rather than an epithelial tumor. Our results demonstrate the ability of gene expression profiles to classify tumors and suggest that determination of organ-specific gene expression profiles will play a significant role in a wide variety of clinical settings, including molecular diagnosis and classification.

The beta-catenin protein plays a critical role in embryonic development and mature tissue homeostasis through its effects on E-cadherin-mediated cell adhesion and Wnt-dependent signal transduction. In colon and other cancers, mutations of beta-catenin or the adenomatous polyposis coli (APC) tumor suppressor appear to stabilize beta-catenin and enhance its interaction with T cell factor (TCF) or lymphoid enhancer factor (Lef) transcription factors. At present, a complete picture of the means by which beta-catenin's interactions with TCF/Lef proteins contribute to neoplastic transformation is lacking. We report that the transcriptional coactivator p300 interacts with beta-catenin in vitro and in vivo and is critical for beta-catenin-mediated neoplastic transformation. p300 synergistically activates beta-catenin/TCF transcription, and their biochemical association requires the CH1 domain of p300 and a region of beta-catenin that includes its NH(2)-terminal transactivation domain and the first two armadillo repeats. Lowering of cellular p300 levels by using a ribozyme directed against p300 reduced TCF transcriptional activity and inhibited the neoplastic growth properties of a beta-catenin-transformed rat epithelial cell line and a human colon carcinoma line with a beta-catenin mutation. These findings demonstrate a critical role for p300 in beta-catenin/TCF transcription and in cancers arising from defects in beta-catenin regulation.

Inactivation of the E-cadherin cell adhesion molecule is believed critical in the development and behavior of many epithelial cancers, though mutations in the E-cadherin gene account for inactivation in only a fraction of cases. In many breast cancer lines, E-cadherin transcription is extinguished, but the role and significance of alterations in trans-acting transcription factors, promoter hypermethylation, and chromatin changes remain unresolved. To gain further insights into mechanisms underlying E-cadherin inactivation in breast cancer, we analysed somatic cell hybrids resulting from pairwise fusions between breast cancer lines with intact E-cadherin transcription (E-cad+) and lines lacking E-cadherin transcription (E-cad-). All hybrid lines failed to express E-cadherin transcripts and protein, despite the fact that E-cadherin alleles from E-cad+ lines were present in the hybrids. Elements in the proximal 108 bp of the E-cadherin promoter, when present in reporter gene constructs, were sufficient to direct strong transcription in E-cad+ breast lines, but displayed weak activity in E-cad- parental lines and hybrids. E-cadherin expression could not be restored in E-cad- lines or hybrids by treatment with a DNA demethylating agent and/or a histone deacetylase inhibitor. Our findings suggest loss of E-cadherin expression in some breast cancers may be due to dominant repression of the trans-acting pathways that regulate E-cadherin transcription.

Previous studies have demonstrated frequent allelic losses of distal chromosome 11q in ovarian carcinomas. The tumor suppressor gene(s) presumably targeted by these losses have not yet been identified. PPP2R1B is a candidate tumor suppressor gene at 11q23 that has recently been shown to be mutated in a subset of colorectal and lung cancers. We evaluated 5 ovarian carcinoma cell lines and 27 primary ovarian carcinomas for allelic losses of 11q23 and for mutations in the open reading frame of PPP2R1B. We also evaluated the primary tumors for allelic losses at 17p13, another chromosomal region frequently affected by losses of heterozygosity (LOH) in ovarian cancers. 11q23 and 17p13 allelic losses were identified in 25% and 74% of the carcinomas, respectively. No mutations within PPP2R1B coding sequences were found. These findings indicate that mutations of the PPP2R1B gene are infrequent in ovarian cancer and that deletions affecting the distal portion of chromosome 11q in ovarian cancer likely target inactivation of other genes.

Current models predict that beta-catenin (beta-cat) functions in Wnt signaling via activation of Tcf/Lef target genes and that its abundance is regulated by the adenomatous polyposis coli (APC) and glycogen synthase kinase 3beta (GSK3beta) proteins. In colon and other cancers, mutations in APC or presumptive GSK3beta phosphorylation sites of beta-cat are associated with constitutive activation of Tcf/Lef transcription. In spite of assumptions about its oncogenic potential, prior efforts to demonstrate that mutated beta-cat will induce neoplastic transformation have yielded equivocal results. We report here that mutated, but not wild-type, beta-cat proteins induced neoplastic transformation of RK3E, an adenovirus E1A-immortalized epithelial cell line. Analysis of the properties of mutant beta-cat proteins and studies with a dominant negative Tcf-4 mutant indicated that the ability of beta-cat to bind and activate Tcf/Lef factors is crucial for transformation. c-myc has recently been implicated as a critical Tcf-regulated target gene. However, c-myc was not consistently activated in beta-cat-transformed RK3E cells, and a dominant negative c-Myc mutant protein failed to inhibit beta-cat transformation. Our findings underscore the role of beta-cat mutations and Tcf/Lef activation in cancer and illustrate a useful system for defining critical factors in beta-cat transformation.

Beta-Catenin has a critical role in E-cadherin-mediated cell-cell adhesion, and it also functions as a downstream signaling molecule in the wnt pathway. Mutations in the putative glycogen synthase kinase 3beta phosphorylation sites near the beta-catenin amino terminus have been found in some cancers and cancer cell lines. The mutations render beta-catenin resistant to regulation by a complex containing the glycogen synthase kinase 3beta, adenomatous polyposis coli, and axin proteins. As a result, beta-catenin accumulates in the cytosol and nucleus and activates T-cell factor/ lymphoid enhancing factor transcription factors. Previously, 6 of 27 melanoma cell lines were found to have beta-catenin exon 3 mutations affecting the N-terminal phosphorylation sites (Rubinfeld B, Robbins P, Elgamil M, Albert I, Porfiri E, Polakis P: Stabilization of beta-catenin by genetic defects in melanoma cell lines. Science 1997, 275:1790-1792). To assess the role of beta-catenin defects in primary melanomas, we undertook immunohistochemical and DNA sequencing studies in 65 melanoma specimens. Nuclear and/or cytoplasmic localization of beta-catenin, a potential indicator of wnt pathway activation, was seen focally within roughly one third of the tumors, though a clonal somatic mutation in beta-catenin was found in only one case (codon 45 Ser-->Pro). Our findings demonstrate that beta-catenin mutations are rare in primary melanoma, in contrast to the situation in melanoma cell lines. Nonetheless, activation of beta-catenin, as indicated by its nuclear and/or cytoplasmic localization, appears to be frequent in melanoma, and in some cases, it may reflect focal and transient activation of the wnt pathway within the tumor.

The adenomatous polyposis coli (APC) tumor suppressor protein is inactivated by mutations in the majority of colorectal cancers. A recent study has revealed that alterations in the APC signaling pathway can result in the transcriptional activation of the c-MYC gene.

The Drosophila seven in absentia (sina) gene was initially discovered because its inactivation leads to R7 photoreceptor defects. Recent data indicate that Sina binds to the Sevenless pathway protein Phyllopod, and together they mediate degradation of Tramtrack, a transcriptional repressor of R7 cell fate. Independent studies have shown that Sina and its highly related mammalian homologues Siah-1 and Siah-2 bind to the DCC (deleted in colorectal cancer) protein and promote its proteolysis via the ubiquitin-proteasome pathway. To determine the roles of mammalian Siahs in proteolysis and their interactions with target proteins, we sought to define Siah-1 domains critical for regulation of DCC. Mutant Siah-1 proteins, harboring missense mutations in the carboxy (C)-terminal domain analogous to those present in Drosophila sina loss-of-function alleles, failed to promote DCC degradation. Point mutations and deletion of the amino (N)-terminal RING finger domain of Siah-1 abrogated its ability to promote DCC proteolysis. In the course of defining Siah-1 sequences required for DCC degradation, we found that Siah-1 is itself rapidly degraded via the proteasome pathway, and RING domain mutations stabilized the Siah-1 protein. Siah-1 was found to oligomerize with itself and other Sina and Siah proteins via C-terminal sequences. Finally, evidence that endogenous Siah-1 regulates DCC proteolysis in cells was obtained through studies of an apparent dominant negative mutant of Siah-1, as well as via an antisense approach. The data indicate that the Siah-1 N-terminal RING domain is required for its proteolysis function, while the C-terminal sequences regulate oligomerization and binding to target proteins, such as DCC.

Several studies have reported loss or alteration of expression of E-cadherin in breast cancer and more recently changes in levels of expression of the catenins. We used immunofluorescence to examine E-cadherin, alpha-catenin, beta-catenin, and p120ctn (formerly p120CAS) expression in 91 cases of invasive ductal carcinoma. As expected, all four proteins co-localize to the junctional regions of the cells. Although nuclear localization has been described for beta-catenin in colonic polyps, no examples were found in these breast cancer cases. We found that, although alteration is common in the catenins and E-cadherin, complete loss, as exemplified by E-cadherin in lobular carcinoma (where E-cadherin is frequently mutated), is rarely seen. In contrast, the catenin-related protein p120ctn shows an expression pattern that is significantly unrelated to the other catenins (or E-cadherin), including complete loss of expression in approximately 10% of the cases. No statistically significant correlations with traditional prognostic indicators were observed with any of these proteins. We conclude 1) that expression of E-cadherin and alpha- and beta-catenin are generally retained at the membrane although frequently reduced or altered, 2) that complete loss of p120ctn expression is seen in approximately 10% of the cases, and 3) that there is a significant correlation in the expression of E-cadherin and the catenins but no correlation between these molecules and p120ctn, suggesting an absence of coordinate regulation.

Studies of Drosophila photoreceptor development have illustrated the means by which signal transduction events regulate cell fate decisions in a multicellular organization. Development of the R7 photoreceptor is best understood, and its formation is dependent on the seven in absentia (sina) gene. We have characterized two highly conserved human homologs of sina, termed SIAH1 and SIAH2. SIAH1 maps to chromosome 16q12 and encodes a 282-amino-acid protein with 76% amino acid identity to the Drosophila SINA protein. SIAH2 maps to chromosome 3q25 and encodes a 324-amino-acid protein that shares 68% identity with Drosophila SINA and 77% identity with human SIAH1. SIAH1 and SIAH2 were expressed in many normal and neoplastic tissues, and only subtle differences in their expression were noted. However, one of three murine homologs, Siah1B, was strongly induced in fibroblasts undergoing apoptotic cell death. While a previous study suggested that SINA was a nuclear protein, epitope-tagged SINA and SIAH1 proteins were found in the cytoplasm of Drosophila and mammalian cells. Their substantial evolutionary conservation, role in specifying cell fate, and activation in apoptotic cells suggest the SIAH proteins have important roles in vertebrate development. Furthermore, given the role of sina in Drosophila photoreceptor development, SIAH2 is a candidate for the Usher syndrome type 3 gene at chromosome 3q21-q25.

More than 20 different hereditary cancer syndromes have now been defined and attributed to specific germline mutations in various inherited cancer genes. Collectively, the syndromes affect about 1 percent of cancer patients. An individual who carries a mutant allele of an inherited cancer gene has a variable risk of cancer that is influenced by the particular mutation, other cellular genes, and dietary, lifestyle, and environmental factors. Though hereditary cancer syndromes are rare, their study has provided powerful insights into more common forms of cancer. Somatic mutations in sporadic cancers frequently alter the inherited cancer genes, and the functions of cell signaling pathways have been illuminated by study of the affected genes. Further investigation of inherited mutations that affect susceptibility to cancer will aid efforts to effectively prevent, detect, and treat the disease.

NIH 3T3 fibroblasts stably transformed with a constitutively active isoform of p21(Ras), H-RasV12 (v-H-Ras or EJ-Ras), produced large amounts of the reactive oxygen species superoxide (.O2-). .O2- production was suppressed by the expression of dominant negative isoforms of Ras or Rac1, as well as by treatment with a farnesyltransferase inhibitor or with diphenylene iodonium, a flavoprotein inhibitor. The mitogenic activity of cells expressing H-RasV12 was inhibited by treatment with the chemical antioxidant N-acetyl-L-cysteine. Mitogen-activated protein kinase (MAPK) activity was decreased and c-Jun N-terminal kinase (JNK) was not activated in H-RasV12-transformed cells. Thus, H-RasV12-induced transformation can lead to the production of .O2- through one or more pathways involving a flavoprotein and Rac1. The implication of a reactive oxygen species, probably .O2-, as a mediator of Ras-induced cell cycle progression independent of MAPK and JNK suggests a possible mechanism for the effects of antioxidants against Ras-induced cellular transformation.

DCC (deleted in colorectal cancer), a candidate tumor suppressor gene located in chromosome band 18q21.2, encodes a transmembrane protein of 1447 amino acids. Neogenin, a protein with nearly 50% amino acid identity to DCC, was recently identified because of its dynamic expression in the developing nervous system and gastrointestinal tract of the chicken. To explore a role for the human neogenin (NGN) gene in cancer, we have isolated cDNAs for two alternatively spliced forms of NGN, encoding proteins of 1461 and 1408 amino acids. Fluorescence in situ hybridization studies (FISH) localized NGN in chromosome band 15q22, a region infrequently affected by alterations in cancer. NGN transcripts of about 7.5 and 5.5 kb were detected in all adult tissues studied. In contrast to the frequent loss of DCC expression, no alterations in NGN expression were observed in more than 50 cancers studied, including glioblastoma, medulloblastoma, neuroblastoma, colorectal, breast, cervical and pancreatic cancer cell lines and xenografts. Based on their sequence conservation and similar expression during development, DCC and NGN may have related functions. However, the chromosomal location and ubiquitous expression of NGN in various human tumors suggest it is infrequently altered in cancer.

The deleted in colorectal cancer (DCC) gene, a candidate tumor suppressor gene on chromosome 18q21, encodes a neural cell adhesion molecule family protein that is most highly expressed in the nervous system. To address the hypothesis that DCC may play a role in glioma development and/or progression, we examined DCC expression by immunohistochemistry in 57 resected human astrocytic tumors. Overall, low-grade astrocytomas were predominantly DCC positive (15 of 16, or 94%), whereas high-grade tumors significantly less often expressed the DCC protein (27 of 41, or 66%; P = 0.03). We were able to directly assess the relationship between DCC expression and tumor progression in 15 patients who initially presented with a low-grade astrocytoma and subsequently recurred with a glioblastoma. Within this panel of paired lesions from the same patient, 14 of 15 (93%) low-grade tumors expressed the DCC protein, whereas only 7 of 15 (47%) corresponding glioblastomas were DCC positive. We also observed that secondary glioblastomas resulting from malignant progression of low-grade astrocytomas were more often DCC negative (8 of 15, or 53%) compared with primary or de novo glioblastomas (6 of 26, or 23%; P = 0.05). These findings implicate DCC inactivation in glioma progression and also demonstrate that DCC expression is preferentially, but not exclusively, lost in the genetic pathway to secondary glioblastoma multiforme.

DCC, a candidate tumor suppressor gene from chromosome 18q21, is most highly expressed in the developing nervous system. In vitro studies suggest a role for DCC in neuronal differentiation, and 18q allelic loss occurs in a subset of neuroblastomas. To address the hypothesis that loss of DCC function may contribute to tumorigenesis in cells of neural origin, we utilized a combination of RNase protection, immunoblotting, and immunohistochemical approaches to characterize DCC expression in 62 primary neuroblastomas and 16 neuroblastoma cell lines. The DCC protein was undetectable in 38% of the primary tumors and 56% of the cell lines. Of note, primary tumors lacking DCC expression were more likely to have been obtained from patients with disseminated or stage D disease (P = 0.01). In addition, loss of DCC expression was observed in three of six primary tumors from stage DS patients. No consistent relationship between the loss of DCC expression and N-myc amplification was observed in our studies. Our findings suggest that loss of DCC expression may contribute to the dissemination of neuroblastoma cells, perhaps through alterations in growth and differentiation pathways distinct from those regulated by N-myc.

Transformation of fibroblast-like cells (NIH 3T3) by a constitutively activated GTP-bound isoform of p21ras (EJ-Ras) produces morphogenic changes characterized by decreased attachment to the substratum, with retraction and rounding of the cell body. Transformed fibroblasts lose their "stressed" conformation and adopt a "relaxed" morphology. The specific molecular mechanisms responsible for these changes remain uncharacterized. We found that EJ-Ras transformation of NIH 3T3 cells decreased the cellular content of polymerized actin, particularly at the expense of actin stress fibers, but induced the accumulation of actin filaments in peripheral ruffling membranes. Polymerization of actin could be induced in EJ-Ras-transformed cells by exposure to platelet-derived growth factor (PDGF)-BB to an extent similar to that observed in wild-type NIH 3T3 cells. In EJ-Ras cells, actin polymerization was independent of phospholipase C gamma 1 (PLC gamma 1) activity, because inositol tris-phosphate (IP3) production observed in control NIH 3T3 cells in response to PDGF-BB was absent. Although PDGF-BB did stimulate tyrosine phosphorylation of PLC gamma 1, the phospholipase was strongly inhibited by an inhibitory factor present in the cytoplasm of EJ-Ras-transformed cells. In addition, cytoplasmic extracts of EJ-Ras, but not of control cells, inhibited phosphatidylinositol 4,5-diphosphate (PIP2) hydrolysis catalyzed by a recombinant PLC gamma 1 in vitro. Although PIP2 hydrolysis could not contribute to the reorganization of the actin cytoskeleton induced by PDGF-BB in EJ-Ras-transformed cells, phosphatidylinositol 3-kinase (PI3-K) was necessary for actin polymerization. Wortmannin, a specific PI3-K inhibitor, not only blocked actin polymerization in both control and EJ-Ras-transformed cells but actually led to rapid actin depolymerization when these cells were exposed to PDGF-BB. Thus, in EJ-Ras-transformed cells, cell morphogenic changes in response to PDGF-BB rely importantly on PI3-K and can occur in the complete absence of IP3 production despite tyrosine phosphorylation of PLC gamma 1.

The gene deleted in colorectal cancer (DCC) is a candidate tumor suppressor gene encoding a neural cell adhesion molecule like transmembrane protein. Over the past year, data supporting DCC inactivation in multiple tumor types have continued to accumulate. Functional studies suggest that DCC may participate in signaling pathways that regulate cell proliferation and/or differentiation, two cellular processes that often go awry during tumorigenesis.

The DCC (deleted in colorectal cancer) gene was identified because it is affected by somatic mutations in colorectal tumors, including allelic losses in greater than 70% of cancers and localized mutations in a subset of cases. The DCC gene also may be inactivated in other tumor types, including cancers of the pancreas, stomach, breast, prostate, and brain, as well as some leukemias. We have characterized DCC complementary DNAs obtained from human fetal brain tissues and IMR32 human neuroblastoma cells. Based on the fetal brain complementary DNA sequence, the predicted transmembrane DCC protein product has 1447 amino acids. The extracellular domain of about 1100 amino acids has four immunoglobulin-like domains and six fibronectin type III-like domains; the 325-amino acid cytoplasmic domain does not show similarity to previously characterized proteins. Comparison of DCC complementary DNAs from IMR32 cells to those from fetal brain identified two potential alternative splice sites. Studies of adult mouse tissues revealed that DCC transcripts were present at very low levels in all tissues studied, and alternative splicing of DCC transcripts was seen in some tissues. Immunoblotting and immunoprecipitation studies with DCC-specific antisera identified protein species with molecular weights of approximately 175,000-190,000 in some rodent tissues and human tumor cell lines. DCC protein expression was highest in brain tissues and neural crest-derived cell lines and markedly reduced or absent in the majority of cancer cell lines studied. Treatment of DCC-expressing cells with tunicamycin decreased the apparent molecular weight of the immunoreactive proteins, establishing that DCC is a glycoprotein. The studies presented here demonstrate that the DCC gene encodes several related glycoprotein species that are likely to be expressed at very low levels in many normal adult tissues. Furthermore, the absence of DCC expression in some of the cancer cell lines studied may result from genetic inactivation of DCC.

The Deleted in Colorectal Cancer (DCC) gene is a candidate tumor suppressor gene that is predicted to encode a transmembrane polypeptide with strong similarity to the neural cell adhesion molecule (N-CAM) family. Previous studies have suggested that several different N-CAMs, when expressed in non-neuronal cell types can stimulate neurite outgrowth from PC12 rat pheochromocytoma cells. Based on the predicted structural similarity of DCC to N-CAMs, we sought to determine whether NIH3T3 cells expressing DCC could stimulate neurite outgrowth in PC12 cells. We found that NIH3T3 cell lines expressing DCC could stimulate PC12 cells to extend neurites. Supernatants from DCC-transfected NIH3T3 cells did not induce neurite outgrowth above background levels, suggesting that cell-cell interaction was required. NIH3T3 cells expressing a truncated form of DCC, lacking the majority of the cytoplasmic domain sequences, also failed to induce neurite outgrowth above the levels seen with control NIH3T3 cells, suggesting that the cytoplasmic domain of DCC was necessary for its neurite-promoting function. In contrast to NGF-mediated neurite outgrowth, the DCC-mediated response was inhibited by treatment with pertussis toxin or the combination of N- and L-type calcium channel blockers, and was unaffected by the transcriptional inhibitor cordycepin. The data suggest that the DCC protein can function in a fashion analogous to other N-CAMs to alter PC12 cell phenotype through intracellular pathways distinct from those involved in NGF signaling.

DCC is a candidate tumor-suppressor gene encoding a protein with sequence similarity to cell adhesion molecules such as N-CAM. A set of overlapping YAC clones that contains the entire DCC coding region was isolated. Studies of this YAC contig showed that the DCC gene spans approximately 1.4 Mb. For elucidation of exon-intron structure, lambda phage clones containing all known coding sequences were isolated from a genomic library. These clones were used to demonstrate the existence of 29 DCC exons, and the sequences of the exon-intron boundaries were determined for each. Twenty-three polymorphic markers from chromosome 18 were then studied in a panel of primary colorectal tumors that had lost some, but not all, of chromosome 18. In most of these tumors, the region that was lost included DCC. Finally, Southern blot and PCR-based approaches were used to search for subtle mutations in several DCC exons. One tumor that had a point mutation in exon 28 was found, resulting in a proline to histidine substitution. A second tumor with a point mutation in intron 13 was also found. The regional map and genomic structure of DCC should provide the means to more extensively study DCC gene alterations and protein function in normal and neoplastic cells.

Studies are described that allow for the in vivo detection of helix-loop-helix (HLH) protein-protein interaction. The assay used requires HLH protein-protein interaction to reconstitute a functional GAL4 transcriptional activator, which in turn activates a reporter gene placed downstream of GAL4 DNA binding sequences. Using this assay, we are able to detect intracellular heterodimerization but not homodimerization of the MyoD, E12, and Id gene products. In addition, using this system we are unable to detect stable heterodimerization between MyoD and c-Jun. We also show that expression of activated rasH gene product does not inhibit and may stabilize HLH protein-protein interaction. This system may be of general utility in studying the modulation of transcription factor interactions.

Mutations of the p53 gene occur commonly in colorectal carcinomas and the wild-type p53 allele is often concomitantly deleted. These findings suggest that the wild-type gene may act as a suppressor of colorectal carcinoma cell growth. To test this hypothesis, wild-type or mutant human p53 genes were transfected into human colorectal carcinoma cell lines. Cells transfected with the wild-type gene formed colonies five- to tenfold less efficiently than those transfected with a mutant p53 gene. In those colonies that did form after wild-type gene transfection, the p53 sequences were found to be deleted or rearranged, or both, and no exogenous p53 messenger RNA expression was observed. In contrast, transfection with the wild-type gene had no apparent effect on the growth of epithelial cells derived from a benign colorectal tumor that had only wild-type p53 alleles. Immunocytochemical techniques demonstrated that carcinoma cells expressing the wild-type gene did not progress through the cell cycle, as evidenced by their failure to incorporate thymidine into DNA. These studies show that the wild-type gene can specifically suppress the growth of human colorectal carcinoma cells in vitro and that an in vivo-derived mutation resulting in a single conservative amino acid substitution in the p53 gene product abrogates this suppressive ability.

Allelic deletions involving chromosome 18q occur in more than 70 percent of colorectal cancers. Such deletions are thought to signal the existence of a tumor suppressor gene in the affected region, but until now a candidate suppressor gene on this chromosomal arm had not been identified. A contiguous stretch of DNA comprising 370 kilobase pairs (kb) has now been cloned from a region of chromosome 18q suspected to reside near this gene. Potential exons in the 370-kb region were defined by human-rodent sequence identities, and the expression of potential exons was assessed by an "exon-connection" strategy based on the polymerase chain reaction. Expressed exons were used as probes for cDNA screening to obtain clones that encoded a portion of a gene termed DCC; this cDNA was encoded by at least eight exons within the 370-kb genomic region. The predicted amino acid sequence of the cDNA specified a protein with sequence similarity to neural cell adhesion molecules and other related cell surface glycoproteins. While the DCC gene was expressed in most normal tissues, including colonic mucosa, its expression was greatly reduced or absent in most colorectal carcinomas tested. Somatic mutations within the DCC gene observed in colorectal cancers included a homozygous deletion of the 5' end of the gene, a point mutation within one of the introns, and ten examples of DNA insertions within a 0.17-kb fragment immediately downstream of one of the exons. The DCC gene may play a role in the pathogenesis of human colorectal neoplasia, perhaps through alteration of the normal cell-cell interactions controlling growth.

Carbamyl phosphate synthetase I deficiency (CPSD) is an autosomal recessive disorder of ureagenesis characterized by hyperammonemic coma in the neonatal period. To study the genetic basis of CPSD we have performed a molecular analysis of the CPS I genes in CPSD patients from six unrelated families. Using a cDNA probe for the human CPS I gene and restriction endonuclease mapping techniques, we observed no abnormality in the number of size of the hybridizing DNA fragments from the seven affected individuals examined. These findings suggest that no gross alteration affected the CPS I genes. We did detect a frequent restriction fragment length polymorphism (RFLP) at the CPS I locus which we employed as a linkage marker. Our results suggest the polymorphic CPS I restriction fragments cosegregate with the CPSD phenotype, and that linkage disequilibrium exists between the CPSI RFLPs studied and the affected alleles. The RFLPs described may enable prenatal detection of CPSD in families where the coupling phases between CPSD alleles and RFLPs can be determined.

Recent studies have shown that normal cellular sequences on chromosome 13 are lost during the development of retinoblastomas and that sequences on chromosome 11 are similarly lost during the development of Wilms' kidney tumours and embryonal tumours. Cells from these tumors have been found to contain either the paternal or maternal copies of loci on the affected chromosome, but not both. Thus, the somatic loss of heterozygosity for sequences on chromosome 13 or 11 is hypothesized to result in homozygosity for a recessive mutant allele on these chromosomes, and in this way the chromosomal loss may contribute to the development of these tumours. We sought to investigate whether similar losses of heterozygosity for chromosome 11 sequences occurred in a common adult tumour. We chose to analyse bladder cancers, since such cancers are common in the adult population and are derived from urogenital tissue, as are Wilms' tumours. We examined constitutional and tumour genotypes at loci on the short arm of chromosome 11 (11p) in 12 patients with transitional cell carcinomas. In five tumours, we observed the somatic loss of genes on 11p resulting in homozygosity or hemizygosity of the non-deleted alleles in the tumour cells. Our results show that the frequency of loss of 11p sequences in bladder cancer approaches that seen in Wilms' tumour (42% compared with 55%), and suggest that recessive genetic changes involving sequences on 11p may contribute to the development of bladder neoplasms.

One of the most provocative findings in tumour biology is the relationship between chromosomal changes and embryonal cancers in children. For example, children with the rare paediatric syndrome AGR triad (aniridia, genito-urinary abnormalities and mental retardation) often develop Wilms' tumours at a very early age and carry a germ-line deletion on the short arm of chromosome 11 (11p13). It has been suggested that the germ-line deletion 11p is the first of two or more steps to cancer in AGR children. If this were true, one might expect a similar deletion to arise somatically in the far more common isolated Wilms' tumours of children without AGR, as suggested by Knudson from epidemiological data. However, a chromosomal deletion on 11p was observed in only two of five such cases, while it was absent or seen inconsistently in others. We have now used a molecular genetic approach to determine whether Wilms' tumour cells possess somatic alterations at 11p loci. We have found somatic deletions of specific genes in four of six Wilms' tumours. Surprisingly, in all four cases, the deletions were associated with duplications leading to homozygosity of the non-deleted alleles in the tumour cells. As analogous observations were recently reported in retinoblastoma, the genetic events reported here may underlie the development of many such embryonal tumours in children.

DNA sequence polymorphisms have been used to determine the linear order and recombinational distances separating the Harvey ras 1 oncogene (c-Ha-ras-1), beta-globin, insulin, and parathyroid hormone genes on the short arm of human chromosome 11. Our results indicate that c-Ha-ras-1 is closely linked to both the beta-globin locus (theta = .08 [8 centimorgans], lod score = 5.11) and the insulin locus (theta = .04 [4 centimorgans], lod score = 3.31). Furthermore, the probable order of these loci on chromosome 11p is centromere-parathyroid hormone-beta globin-c-Ha-ras-1-insulin.

We have used restriction endonuclease mapping to study a deletion involving the beta-globin gene cluster in a Mexican-American family with gamma delta beta-thalassemia. Analysis of DNA polymorphisms demonstrated deletion of the beta-globin gene from the affected chromosome. Using a DNA fragment that maps greater than 40 kilobases (kb) 5' to the epsilon-gene as a probe, reduced amounts of normal fragments were found in the DNA of affected family members. Similar analysis using radiolabeled DNA fragments located 3' to the beta-globin cluster has shown that the deletion extends more than 17 kb 3' to the beta-gene, but terminates before the 3' endpoint of the Ghanian HPFH deletion. Hence, this gamma delta beta-thalassemia deletion eliminates over 105 kb of DNA and is the first report of a deletion of the entire beta-globin gene cluster.

A combination of DNA hybridization analyses and tissue sectioning techniques demonstrate that ras gene mutations occur in over a third of human colorectal cancers, that most of the mutations are at codon 12 of the c-Ki-ras gene and that the mutations usually precede the development of malignancy.

An abnormal increase in numbers of CCGG sites methylated in the 5' region of the human calcitonin (CT) gene occurred in tumor cell DNA samples from 90% (17 of 19) of patients with non-Hodgkin's T and B cell lymphoid neoplasms and in 95% (21 of 22) of tumor cell DNA samples from patients with acute nonlymphocytic leukemia (ANLL). The changes were not seen in patients with chronic myelogenous leukemia (0 of 9). The abnormal methylation patterns appear to be a property only of transformed or malignant cells since they were not found in DNA from nonneoplastic adult tissues including sperm, early myeloid progenitor cells, benign lymphoid hyperplasia, peripheral lymphocytes stimulated to divide, or early myeloid progenitor cells (obtained by immunoaffinity using anti-My-10 antibody), but they did appear after Epstein-Barr virus transformation of lymphocytes. Moreover, during the course of therapy in patients with ANLL, the hypermethylation pattern reflects the presence of the leukemic clone even in normal-appearing granulocytes derived from this clone. The increased methylation of the CT gene may then provide an important molecular marker for biologic events in human cell transformation or tumor progression and may prove clinically useful in monitoring patients with lymphoid and acute myelogenous neoplasms.

The Wiskott-Aldrich syndrome (WAS) is an X-linked recessive disease characterized by immunodeficiency and severe thrombocytopenia in affected males, but no demonstrable clinical abnormalities in carrier females. Through analysis of the methylation patterns of X-linked genes that display restriction fragment length polymorphisms (RFLPs), we studied the pattern of X-chromosome inactivation in various cell populations from female relatives of patients with WAS. The peripheral blood T cells, granulocytes, and B cells of eight obligate WAS carriers were found to display specific patterns of X-chromosome inactivation clearly different from these of normal controls. Thus, carriers of WAS could be accurately identified using this analysis.