key: cord-0007829-phw8nzkw
authors: Honey, Neville K.; Shows, Thomas B.
title: The tumor phenotype and the human gene map()
date: 2003-01-17
journal: Cancer Genet Cytogenet
DOI: 10.1016/0165-4608(83)90058-4
sha: 5dc4c6029cfbee4f39da4ff0b89c3ed1d263b445
doc_id: 7829
cord_uid: phw8nzkw

The tumor phenotype is associated with the rearrangement of genetic information and the altered expression of many gene products. In this review, genes associated with the tumor phenotype have been arranged on the human gene map and indicate the extent to which the tumor phenotype involves the human genome. Nonrandom chromosomal aberrations that are frequently observed in tumors are presented. Altered metabolic demands of the tumor cell are reflected in altered gene expressions of a wide range of enzymes and other proteins, and these changed enzyme patterns are described. The study of oncogenes increasingly suggests that they may be significant in certain cancers, and the assignment of these genes has been tabulated. The biochemical and metabolic changes observed in tumors are complex; studying the patterns and interactions of these changes will aid our genetic understanding of the origins and development of tumors.

Tumors are associated with a large number of chromosomal and biochemical changes, including altered patterns of enzyme activities and isozyme expression, cell surface and secreted proteins, and various abnormalities in chromosome structure and number. These changes reflect the altered expression of many genes; however, it is not known which, if any, trigger tumor development or are a consequence of tumorigenesis.

Individual tumors are likely caused by a single or small number of genetic changes, while the initiation of tumorigenesis is followed by numerous biochemical changes that probably interact in complex ways to produce the final tumor phenotype. These biochemical changes reflect altered gene expressions, and it is important to determine which genes have altered expressions, how they are altered, and to look for patterns of altered expression in order to understand the ~arious interactions between genes and gene products that result in the tumor phenotype.

Some tumors are associated with specific chromosomal rearrangements, and these are included in this review. The chromosomal locations of those genes with altered expressions of gene products in tumors are also described. The study of oncogenes may provide new insights into the genetic control of tumors, and those oncogenes that have been mapped to specific chromosomes are also reported. With this information, a map of human genes and chromosomal regions involved in tu-From the Department of Human Genetics, Roswell Park Memorial Institute, New York State Department of Health, Buffalo, NY. morigenesis has been generated. The study of biochemical changes in tumors, of chromosomal regions of particular interest in tumor genetics, and of the involvement of oncogenes in the disease will aid further investigation and understanding of tumorigenesis.

Tumor cells commonly express chromosomal changes either with the loss or gain of whole chromosomes or with their rearrangement by translocations, deletions, etc. Although some tumors are associated with nonrandom chromosomal changes, other changes appear to be nonspecific, and most tumors have a range of different chromosomal defects associated with them. Indeed, tumor cells grown in culture often accumulate multiple chromosomal changes. However, some cancers have specific associations with certain mutations or chromosomal changes ( Table 1 ) [1] [2] [3] . For a number of those chromosomal abnormalities that have been localized to specific chromosome bands, certain genes have also been mapped to the same regions [4] . Analysis of these genes may therefore be valuable in characterizing the specific chromosomal breakpoints, and may perhaps aid the study of the various tumors.

The Philadelphia (Ph 1) chromosome in chronic myeloid leukemia is the result of a t(9;22)(q34;q11) translocation; although in a small percentage of cases, the Ph 1 chromosome is associated with unusual or complex translocations. The gene for aconitase-2 (ACO2) is on chromosome #22 (q11--~q13), as are the proto-oncogene c-sis and the genes for the immunoglobulin h-chain, constant and variable regions (IGLC, IGLV) [4] . Some genes have been mapped to the q34 band of chromosome #9, namely, the ABO blood group (ABO), adenylate kinase-1 (AK1), and the nailpatella syndrome (NPS1).

Burkitt's lymphoma is primarily associated with a t(8;14)(q24;q32) translocation. However, other translocations that have been reported in a minority of cases are t(2;8)(p12;q24) and t(8;22)(q24;q11), suggesting that a breakpoint in 8q24 is crucial for the disease [1] [2] [3] . Two human proto-oncogenes that have been mapped to chro- Table 1 Cancers associated with specific chromosomal abnormalities [1] [2] [3] [4] aGene markers: The genes potentially involved in tumorigenesis together with those localized to the chromosomal regions involved in the translocations or deletions are listed here. MPI, mannose phosphate isomerase; B2M, [32-microglobulin; ACO2, aconitase (mitochondrial); ABO, ABO blood group; NPS1, nailpatella syndrome type 1; ACY1, aminoacylase-1; FN, FN*, FNS, genes in the fibronectin system. The designations of the other gene markers are described in Table 12 . For additional genes mapped to these chromosomes, see refs. 80 and 93. mosome #8 are c-mos (8q22) and c-myc (8q24) [5, 6, 87] . A gene in the fibronectin system (FNS) has also been mapped to chromosome #8. The translocations with chromosome #8 are not random, but involve chromosome #14, and to a lesser extent, chromosomes #2 and #22. It is of interest to note that an Epstein-Barr virus nuclear antigen (EBNA) has been mapped to chromosome #14 [4] , with Epstein-Barr virus being associated with Burkitt's lymphoma in Africans [7] . Also, the genes for the immunoglobulin heavy chains (IGH) are to be found on chromosome #14; and for at least one Burkitt's lymphoma cell line, the breakpoint on chromosome #14 occurs in the IGH region [8] . The genes for the immunoglobulin kappa chains (IGK) are on chromosome #2 (with IGKV on 2p12-*cen), while the immunoglobulin lambda chains (IGLC and IGLV) are on chromosome #22 [4] . Another human proto-oncogene, c-sis, is also found on chromosome #22 [9] . Further study should determine what significance these loci might have in Burkitt's lymphoma.

Acute promyelocytic leukemia (APL) is associated with translocation t(15;17)(q26;q21) [2] [3] [4] . Mannosephosphate isomerase (MPI), pyruvate kinase (M2) (PKM2), and ~2-microglobulin (~2m) are all on the region 15q22--~qter, whereas thymidine kinase-1 (TK1), adenovirus-12 chromosome modification site 17 (A12M4), and galactokinase (GALK) are all on chromosome #17q21-~q22.

The Wilms' tumor--aniridia, genitourinary abormalities, and mental retardation-disorder (WAGR) is associated in some cases with a constitutional (i.e., inborn) deletion in 11p13 [4] . The gene for catalase (CAT) is also found on the 11p13 band, with the enzyme being deficient in some cases of WAGR [10] . The use of catalase may help in the study of this disorder. Insulin (INS) maps to the same region (11p14-~p15) [11] , and another gene in the fibronectin (FN) system is also on chromosome #11 [86] .

There is a specific association between retinoblastoma (RB1) and a constitutional deletion in the 13q14 band [12] . The gene for esterase-D (ESD) has also been mapped to 13q14, with the enzyme being deficient in RB1 patients [13] . The analysis of esterase-D should therefore be of value in the study and diagnosis of retinoblastoma.

Whang-Peng et al. [14] report that, in human small cell lung cancer (SLC), there is a specific deletion in the p14--~p23 region of chromosome #3. Marker genes also found in this region are aminoacylase-1 (ACY1, 3p21) and ~-galactosidase (GLB1, 3p21--*q21). ACY1 is deleted in SCLC and could therefore be used as a marker for the tumor.

Wake et al. [15] report a specific translocation, t(6;14)(q21;q24), in ovarian cancer. Superoxide dismutase-2 (SOD-2) is also found on 6q21, while other genes mapped to 14q24 have been described above.

A family with inherited renal cell carcinoma has been reported that exhibit a t(3;8)(p21;q24) translocation [16] . The gene markers associated with the 3p21 and 8q24 regions have already been discussed above. However, other cases of familial or nonfamilial renal cell carcinoma studied do not have a t(3;8) translocation.

For a number of tumors there are nonrandom associations with particular chromosomal abnormalities (Table 2) . Alterations in chromosome #1 are associated with a large number of different cancers (Table 2 ). In a study of 218 cancers with breakpoints in chromosome #1 [19] , 50% of cases had breakpoints in the 1p21--~q21 region, 15% in the 1p22--~p32 region, 14% in the 1q31---~q32 region, and 8% in the lp36 band. Genes mapped to these regions include adenovirus-12 modification site 1A (A12M2) and enolase-1 (ENO1, both lp36), uridine monophosphate kinase (UMPK, lp32), phosphoglucomutase-1 (PGM1, lp22.2), phosphogluconate dehydrogenase (PGD, 1pter--~p34), DNA satellite 3 (DIZ1, lq12), adenovirus-12 modification site 1B (A12M3, lq21), UDP glucose pyrophosphorylase-1 (UGP1, 1q21---~q22), c~-amylase (AMY1, AMY2, lp21), and guanylate kinase-1 (GUK1, QGene markers mapped to each of these chromosomes have been tabulated [80, 93] .

1q32--~q42) [4, 80] . Use of these markers should aid the further study of the significances of these chromosome regions.

About 10% of cases with acute myeloid leukemia (AML) have a t(8;21) (q22;q22) translocation [1, 4] . The proto-oncogene c-mos is on 8q22, while superoxide dismutase (SOD1) has been localized to region 21q22. The various subtypes of AML reflect different stages of differentiation of granulocytic cells, and they are associated with different chromosomal abnormalities [e.g., the myeloblast with t(8;21) and the promyelocyte with t(15;17)]. Thus, different chromosomes, and by implication different genes, are significant in tumorigenesis at different stages of granulocyte differentiation.

In about 70% of the meningioma cases studied, a deletion of chromosome #22 is observed [20] . Genes that have been mapped to chromosome #22 are described above. There is also a nonrandom association between chromosome lp-deletions and neuroblastoma [18] .

Characteristic changes in various enzymes are observed in tumors as compared to the corresponding normal tissues [for reviews, see references [21] [22] [23] [24] . These changes occur in both specific enzyme activities and in isozyme patterns. The metabolic changes are associated with altered growth patterns of tumors, specifically with the change of a differentiated slow-growing tissue to a dedifferentiated fast-growing tumor. The chromosomal locations of the genes of only a few of the affected enzymes are known, and these are scattered widely over the whole genome, as shown in Table 3 . The genes with expressions affected in tumors that are not presently mapped (Tables 4 and 5 ) are also expected to be distributed widely over the genome.

Many of the metabolic changes observed in tumors more closely resemble the corresponding fetal tissue than the original adult tissue (Tables 3-5) [31] . Knox [43] studied 82 enzymes present in normal rat liver and found that 24 enzymes that had low activities in fast-growing hepatomas also had low activities in fetal liver. Fortytwo of the 50 enzymes with high activities in hepatomas also had high activities in fetal liver. Tumor Phenotype 295 OTHER A good model system for the study of metabolic patterns in tumors is provided by a series of rat Morris hepatomas that range from slow-growing well differentiated tumors with more nearly normal adult isozyme patterns to fast-growing poorly differentiated tumors with fetal-like isozyme patterns [44] . Much of the work on enzyme activities and isozyme patterns has been done on rat and human hepatomas, which exhibit very similar changes in enzyme patterns. The study of hepatomas is valuable because the liver expresses many differentiated enzyme functions that are often changed in tumors. The enzyme changes observed for hepatomas mostly seem to be generalized metabolic changes, reflecting the process of transformation and increased growth demands on the tumor ceils. Hepatomas and other tumors express different isozyme patterns for some enzymes, but these mostly reflect the different fetal and adult isozyme patterns of the respective tissues. The generalized response of metabolic enzymes to tumorigenesis is of interest, especially considering that the initial defects of the various tumors most likely occur in different loci scattered throughout the genome.

The metabolic changes in tumors occur in the key rate-limiting enzymes of different metabolic pathways, such that there is increased growth and decreased catabolism [29, 30] . The isozyme patterns of these enzymes tend to change from the differentiated forms to the dedifferentiated fetal forms. The key enzymes are affected by isozyme and activity changes, such that there is increased growth, e.g., increased glycolysis (with increased hexokinase, phosphofructokinase, pyruvate kinase activities} and decreased gluconeogenesis (with decreased glucose-6-phosphatase, fructose diphosphatase, phosphoenolpyruvate carboxykinase, and pyruvate carboxylase activities). There is also increased protein synthesis and decreased catabolism, increased nucleic acid synthesis and decreased degradation, and so on. In general, the metabolic pathways involved in synthesis and growth increase over the catabolic degradative pathways as tumor growth rates increase. Some of these known changes are listed in Tables 3-5. As more of these and other genes altered in tumors are mapped, our understanding of the patterns and interactions of genetic change in tumorigenesis will increase. It is readily apparent, however, that there is no obvious pattern to gene location in the genome. Enzymes from the same metabolic pathways are found on different chromosomes, as are enzymes with similar functions. The level of control of isozyme changes in tumors is not at the chromosomal level, but is apparently at the individual gene level.

Other gene products are also implicated in the tumor phenotype, either in altered gene expression or in the involvement of the gene product with some aspect of the tumor cell. Many of these genes have been assigned to specific chromosomes.

The abnormal production of polypeptide hormones is associated with certain tumors, with the pattern of hormone synthesis apparently reflecting the embryologic background of the tumor in question [45] . Increased hormone synthesis may be a secondary effect of tumorigenesis rather than a primary cause. Thus, tumors whose origins derive from the neural crest tend to secrete hormones such as insulin, calcitonin, proopiocortin, ACTH, vasopressin, glucagnn, serotonin, catecholamines, and others. Tumors with embryologic origins in the endoderm or mesoderm, however, mostly secrete hormones such as parathyroid hormone, the gonadotropins, chorionic somatomammotropin, prolactin, growth hormone, etc. The chromosomal locations of the hormone genes that have been mapped are listed in Table 6 . There is no apparent correlation between the two groups of hormones and the chromosomal locations of their genes.

Many cell surface changes occur during malignant transformation, generally resulting in the disorganization of the ordered networks of cell surface proteins. These changes are important, as they mediate the interaction of the tumor cell with its environment. Much of the work on cell surface changes has been performed on virally transformed cells cultured in vitro, although the data have general relevance for other tumor cells [47] . The amounts of many membrane-associated proteins are altered following transformation. Fibronectin (chromosomes #2, #8, and #11) [4, 48, 86] , collagen (chromosomes #7 and #17) [92] , and membrane-associated actin and myosin are all decreased in many transformed cell lines (Table 7) . Certain tumor-specific transplantation antigens (TSTAs) appear following viral transformation. Examples of TSTAs include SV40, feline leukemia and feline sarcoma virus, and Moloney leukemia virus. There are increases in nutrient transport molecules, including the transport of glucose, phosphate, and certain amino acids, cAMP is often decreased in transformed cells. In some cases, this is related to a decreased activity of the plasma membrane adenylate cyclase, and in others to a defect in the regulation of cAMP phosphodiesterase [47] . Hormone receptors may be decreased following transformation, including receptors for epidermal growth factor (chromosome #7), a melanocyte-stimulating activity, transferrin (chromosome #3), catecholamines, and prostaglandin ( Table 7 ). The glycosylation of many glycoproteins are also altered, which may be related to an increased sialylation of the oligosaccharide moieties [49] . Secreted proteins may also be altered following transformation. These changes may include increases in activity of plasminogen activator (chromosome #5) and other proteases and of glycosidases. Plasminogen activators may be responsible for tumor invasiveness and are found in the early embryo, in addition to transformed cells. It is evident that a large number of changes in cell surface proteins may occur following tumorigenesis, and the genes for some of these have now been mapped (Table 7) . These proteins are important for cell adhesion and cell-cell recognition, for the uptake of nutrients, and for growth control. Alterations in their expression may therefore be instrumental in the changed ability of the malignant cell to grow and to migrate in the tissue environment. However, the question remains as to whether they are merely a generalized response to the increased metabolic and growth requirements of a tumor.

Certain other proteins with a role in the protection of the body against disease include interferon and the immunoglobulin and HLA molecules [4, 50] . These proteins may also have some association with protection against cancer; the chromosomal assignments of these genes are listed in Table 8 .

Some proteins, usually detected antigenically, are commonly found in the fetus and in tumors, but not in adult tissues [51] . These proteins are referred to as carcinoembryonic proteins. This reflects the general phenomenon of dedifferentiation of a slow-growing normal cell to a fast-growing tumor cell with many aspects of fetal [51] . However, none of the genes of these proteins have yet been mapped, although ~-fetoprotein (AFP) has been suggested to be on chromosome #4 by homology with the mouse genome [4] .

When hybrids are formed between tumor cells and normal cells of the same or different species, the transformed phenotype is generally suppressed [52] [53] [54] . Suggestions that tumorigenicity is not suppressed in hybrids may reflect rapid chromosomal loss of the normal parental cell in those hybrids. Suppression of tumorigenicity indicates that the cancer defect is recessive in these cases and that it can be corrected by the normal genome regardless of species. When suppressor chromosomes are lost in hybrids, the transformed phenotype may reappear. In principle, the chromosomal site of a defect causing a specific tumor could thus be mapped by following chromosomal loss, and some have been characterized in this way (Table 9 ). However, the rapid segregation of chromosomes in interspecific hybrids complicates this strategy [54] . Also, there does not always seem to be an absolute correlation between a single specific chromosome and tu- morigenicity; instead, combinations of several chromosomes seem to be involved in tumorigenicity.

Stanbridge et al. [55] generated HeLa x human fibroblast hybrids and found reexpression of tumorigenicity associated with the loss of chromosomes #11 and #14, presumably from the fibroblast parent. The parental origins of the lost chromosomes could not be determined. In human x tumorigenic Chinese hamster hybrids, Klinger and Shows [52] found that human chromosome #2 was lost in all clones that acquired the tumorigenic phenotype. However, chromosome #2 was also preferentially lost in nontumorigenic and low tumorigenicity hybrids, and its possible role in tumorigenicity suppression could not be determined. The loss of chromosomes #9, #10, #11, or #17 were also implicated in the reappearance of tumorigenicity, as were certain combinations of pairs of chromosomes (involving chromosomes #4, #7, #8, #9, #10, #11, #13, and #17). The reappearance of tumorigenicity was enhanced by increased numbers of chromosomes #6, or #6 and #12.

When human tumor cell lines of different origins were fused together, tumorigenicity was sometimes suppressed [54] . Carcinoma × carcinoma and carcinoma × lymphoblastoid hybrids were fully tumorigenic. Carcinoma × sarcoma and carcinoma × melanoma hybrids were not tumorigenic. Thus, at least some tumors have distinct genetic defects that are complementary. HeLa × human fibroblast hybrids were also generated, and a number of phenotypic characteristics typical of tumors could be dissociated from tumorigenicity [54] , that is, these characteristics are under separate genetic control and are only secondarily related to the expression of the tumor.

Somatic cell genetics is a valuable tool in the dissection of the chromosomal involvement in the expression of particular tumors. This technique may be used to generate a strategy to dissect the different aspects of tumorigenicity and their interaction. The primary causes of tumorigenesis and some of the secondary effects may thus be distinguished. In suppressed hybrids, the loss of chromosomes involved with both the original defect or with some of the early subsequent events in tumor growth may result in the reappearance of tumorigenicity. This may result in the somewhat complex patterns of chromosomal loss reported for tumor × nontumor hybrids.

The significance of viral infection in the etiology of human cancer is generally not known, although an association has been demonstrated for Burldtt's lymphoma and Epstein-Barr virus in Africans [7] , for adult T-cell leukemia and ATLV [56, 57, 89] , and has been implicated in some other cases [58] . However, many animal viruses have oncogenic potential in their natural hosts [59] , and a number of sites on the human genome are known to be related to viral function or chromosomal integration. Their possible primary or secondary involvement in tumorigenesis is not known and is most unlikely in some cases, but their chromosome locations have been listed in Table 10 .

In a rapidly increasing number of cases, single genes have been associated with the ability to transform cell cultures and to induce tumors in experimental animals [60] . These genes have been termed oncogenes and have been characterized in viruses and in tumors of viral and nonviral origin. The DNA tumor viruses (such as SV40, EBV, polyoma) transform cells by means of viral gene products that are essential to the viral replication cycle [61] . However, most of the work on viral oncogenes has been done on retroviruses, and at least 15 retrovirus oncogenes are now known [62] . A number of the human cellular homologs of these retroviral onco- genes have now been mapped [ Table 11 ). These oncogenes all seem to have been derived originally from the normal genome of the viral hosts, and for each oncogene studied, homologous cellular genes have been detected in various tissues in widely different vertebrate species [66] . Some, but not all, of the cellular homologs (cellular proto-oncogenes) may have a low degree of transcriptional activity in normal cells and seem to play a normal role within the cellular metabolism [60] . Proto-oncogenes from normal human tissues will not transform NIH 3T3 cells, and there is evidence that homologous proto-oncogenes and v-onc's have different DNA sequences [60] . Proto-oncogenes may have to undergo a recombination or mutationlike event before they can transform cells [69, 90, 91] . During transformation by viral oncogenes, the concentration of oncogene product in the cell may increase 100-1000-fold [70] . This overloading of the cell with an excess of an essentially normal cellular protein may be a factor in triggering transformation.

There is some indication that the total number of proto-oncogenes is limited-duplicate and closely related v-onc's have already been identified among the limited number of independently isolated retroviruses analyzed [60] . The proto-oncogenes may comprise a family of genes with related functions, perhaps involved in the regulation of growth and development (since v-onc's disrupt patterns of cellular differentiation during transformation). It may be significant that at least some, if not most, of the v-onc's characterized are tyrosine-specific protein kinases. [60] .

Recent recombinant DNA techniques are amenable to the chromosomal mapping of oncogenes (Table 11) , and further work in this field is currently in progress. DNA 20 67 from animal and human tumors may be used to transform NIH 3T3 fibroblasts that are consequently tumorigenic in nude mice [71] [72] [73] [74] [75] . The presence of oncogenes are thus indicated in tumors independent of the involvement of viruses. Only a minority of human tumors apparently contain oncogenes that are detectable by the NIH 3T3 assay [71, 73] . Oncogenes from different tumors may be the same or highly related, as characterized by restriction endonuclease patterns, as in the case of two colon and two lung carcinomas [73] and for seven human and mouse mammary carcinomas [76] . On the other hand, unique oncogenes were detected in a colon and bladder carcinoma and a promyelocytic leukemia cell line [72] and in at least two bladder carcinomas [73] . In general, however, dissimilar tumor cell lines have unique oncogenes, whereas similar tumors tend to have very similar oncogenes. Lane et al. [77] examined the oncogenes from 20 B and T lymphocytic neoplasms of human and mouse origin. They found that tumors of the same lineage at the same stage of cellular differentiation had very similar oncogenes, whereas tumors at different stages of normal cell differentiation had different oncogenes. Thus, sp.ecific transforming genes are activated in tumors at discrete stages of differentiation within these cell lineages, suggesting that these oncogenes may have normal roles in cellular differentiation. The oncogenes characterized in tumors are most likely derived from normal cells, where they had normal metabolic functions, but were induced, by an unknown series of events, to trigger tumorigenesis. Chang et al. [70] demonstrated that a normal human gone homologous to the viral oncogene Harvey v-ras can induce transformation if made to produce high levels of the gene product. The oncogenes apparently represent a relatively limited class of genes, with similar tumors being caused by the same or very similar oncogenes. However, we may only be observing the limited class of oncogenes that may be detected by the NIH 3T3 assay. At least two of these oncogenes isolated from tumors seem to be analogous to the retrovirus oncogenes [78, 79, 90, 91] , and it is of interest to speculate whether tumor and retroviral oncogenes may comprise the same or overlapping classes of genes. A colon/lung oncogene is homologous to the Kirsten sarcoma virus v-ras gone [78] , while Parada et al. [79] demonstrated that a human bladder carcinoma oncogene is homologous to the Harvey v-ras.

A large number of gene products are altered in tumor cells, involving many aspects of cellular metabolism and growth. Many genes with altered expression in tumors have been mapped and are summarized in Tables 3 and 12 . Chromosomal changes are frequently observed in tumors. Certain tumors have very specific chromosome alterations associated with them {such as Burkitt's lymphoma), and these indicate that the tumors are associated with defects in specific gene loci. In other tumors, however, there is no absolute correlation between the tumor type and particular chromosomal abnormalities; instead, there is only a tendency for certain abnormalities to occur relatively frequently. The accumulation of chromosome abnormalities may be a secondary response to loss of growth controls in a cell during tumorigenesis, or perhaps the appearance of chromosome abnormalities disrupts the genetic regulation of a cell, resulting in tumorigenesis. At present, we cannot distinguish between these possibilities.

Many metabolic enzymes have altered expression in tumors {Tables 3-5}. These changes mostly seem to reflect altered growth patterns in a tumor cell. Thus, enzyme patterns often change from that of the slowly or nongrowing differentiated adult cell to that of the rapidly growing undifferentiated fetal cell. Tumors from different tissue types often display different enzyme patterns, but these differences often reflect the isozyme patterns of the normal tissues and of their respective fetal tissues. However, among the multiple enzyme changes observed, it is possible that the altered synthesis of one or a few specific enzymes may in some way affect the metabolic balance of the cell and thus induce tumorigenesis. Oncogenes represent a class of genes on several chromosomes (Tables 11 and 12 ) that may, under suitable circumstances, disrupt the metabolism of a cell to induce tumorigenesis. The role of oncogenes in the growth of tumors is not known, but increasing evidence suggests that they may be significant in a number of cases. The study of oncogenes provides an illustration of the potential value of recombinant DNA technology in future investigations of tumors at the molecular level.

All chromosomes are involved in different aspects of tumorigenesis, with either altered gene expression or chromosome abnormalities associated with various tumors (Table 12) . However, certain specific chromosome regions seem to be associated with a number of different tumors and with a number of different tumor phenotypes, including chromosomes #3, #8, #14, and #22. For example, small cell lung cancer is associated with a deletion in the 3p14-~p23 bands. A translocation involving 3p21 is associated with the familial renal cell carcinoma studied. Chro- mosome #3 is also associated with various other tumors, such as malignant lymphomas, carcinomas, melanomas (Table 2 ). In chromosome #8, the q24 band, the location of the c-myc proto-oncogene, is associated with both Burkitt's lymphomas and familial renal cell carcinoma. About 10% of AML cases have a translocation involving 8q22, and many other tumors are also associated with chromosome #8 abnormalities. The c-mos proto-oncogene has also been localized to the 8q22 band. Chromosome #14 is associated with Burkitt's lymphoma (q32) and various other tumors. Epstein-Barr virus nuclear antigen (EBNA) and the immunoglobulin heavy chains are also on chromosome #14. Chromosome 22q is associated with the Philadelphia chromosome of CML, with about 70% of meningioma patients, in a few cases of Burkitt's lymphoma, and with a number of other tumors. The immunoglobulin lambda light chain genes are found on chromosome #22, as is the human proto-oncogene c-sis. It is not known whether the association of different tumors to specific chromosome regions occurs only by chance or whether they reflect certain gene loci, defects in which have profound effects resulting in different tumors. It is also not known whether oncogenes are associated with these specific loci.

Each tumor is probably initiated by defects at a single or a limited number of sites. Defects in these sites have wide-ranging pleiotropic effects involving many genes scattered throughout the genome, resulting in a cascade of primary, second-ary, and tertiary effects. The expression of a large number of genes controlling the general metabolism and various other cellular processes are thus changed, with transformation and the growth of the tumor consequences of all these changes. Many of the changes in gene expression observed in tumors probably reflect generalized metabolic changes caused by the increased growth of the cell. Although tumors display extensive changes in gene expression, there are no obvious patterns at the chromosome level--the general regulation of tumor growth occurs at a finer level. However, as more genes altered in cancer are mapped and their genetic regulation characterized, their interactions to produce the tumor phenotype will be better understood.

Supported in part by NIH Grants GM 20454, HD 05196, and ACS Grant CD-62.

Clustering of aberrations to specific chromosomes in human neoplasms IV. A survey of 1,871 cases

The Chromosomes in Human Cancer and Leukemia

Chromosome abnormalities in malignant hematologic diseases

Molecular cloning and chromosomal mapping of a human locus related to the transforming gene of Moloney murine sarcoma virus

Genetic organization of human proto-oncogenes

The relationship of the virus to Burkitt's lymphma

Translocation of immunoglobulin VH genes in Burkitt lymphoma

Chromosomal mapping of the simian sarcoma virus onc gene analogue in human cells

Catalase levels in patients with aniridia and/or Wilms' tumor: utility and limitations

High resolution in situ hybridization: Localization of a DNA restriction polymorphism, the human proopiomelanocortin gene, and the human insulin gene

Retinoblastoma and its association with a deletion in chromosome #13: A survey using high-resolution chromosome techniques

Regional assignment of genes for human esterase D and retinoblastoma to chromosome band 13q14

A nonrandom chromosomal abnormality, del 3p(14 -23), in human small cell lung cancer (SCLC)

Specific chromosome change in ovarian cancer

Hereditary renal-cell carcinoma associated with a chromosomal translocation

Anomalies chromosomiques numeriquest et structurales dans 36 lignees de melanoma malin humain

Cytogenetic features of human neuroblastomas and cell lines

Break points in chromosome #1 abnormalities of 218 human neoplasms

Correlations between clinical and cytogenetical data in 180 human meningiomas

A review of isozymes in cancer

Developmental gene expression in cancer

Isozymes and cancer

Alteration of isozymes in tumors

Isozymes of carbohydrate enzymes

Enzymology of pyrimidine and carbohydrate metabolism in human colon carcinomas

Copper-and zinc-containing superoxide dismutase and manganese-containing superoxide dismutase in human tissues and human malignant tumors

Role of superoxide dismutase in cancer: A review

Enzymology of cancer cells (first of two parts)

Enzymology of cancer cells (second of two parts)

Enzymes and isoenzymes in cancer

Oxidoreductase activities in normal rat liver, tumor-bearing rat liver, and hepatoma HC-252

Purine pathway enzyme abnormalities in acute lymphoblastic leukemia

Aryl hydrocarbon hydrolase activity in mouse, rat, and human mammary tumors

Elevated activities and properties of arylsulfatases A and B and B-variant in human lung tumors

P~urification~ properties, and immunotitration of hepatoma glutamine phosphoribosyl-pyrophosphate amidotransferase

Elevation of lysosomal enzymes in primary Lewis lung tumor correlated with the initiation of metastasis

Alteration of hexosaminidase isozymes in human renal carcinoma

Biochemical activities of lysosomal acid hydrolases in leukemic cells

Molecular species of plasminogen activators secreted by normal and neoplastic human cells

A re'evaluation of studies on the distribution of glyoxylases in animal and tumor tissues

Carbamoyl phosphate synthetase (glutamine-hydrolyzing): Increased activity in cancer cells

Enzyme Patterns in Fetal, Adult and Neoplastic Rat Tissues

Isozyme composition, gene regulation and metabolism of experimental hepatomas

A classification of ectopic hormone-producing tumors

Human genomic organization of glycoprotein and polypeptide hormone genes

Cell surface glycoproteins and malignant transformation

Assignment of fibronectin gene to human chromosome 2 using monoclonal antibodies

Metastatic potential is positively correlated with cell surface sialylation of cultured murine tumor cell lines

Human immune interferon {IFN-~/) gene is located on chromosome 12

Carcinoembryonic proteins

Suppression of tumorigenicity in somatic cell hybrids II. Human chromosomes implicated as suppressors of tumorigenicity in hybrids with Chinese hamster

Suppression of malignancy in human cancer cells: Issues and challenges

Human cell hybrids: Analysis of transformation and tumorigenicity

Specific chromosome loss associated with the expression of tumorigenicity in human cell hybrids

Adult T-cell leukemia: Antigen in an ATL cell line and detection of antibodies to the antigen in human sara

Isolation and characterization of retrovirus from cell line of human adult T-cell leukemia and its implication in the disease

zur Hausen H (eds) {1980): Viruses in Naturally Occurring Cancer

Viruses and leukemia

Retroviruses and cancer genes

The role of gene dosage and genetic transpositions in carcinogenesis

Proposal for naming host cell-derived inserts in retrovirus genomes

Chromosomal assignment of the human homologues of feline sarcoma virus and avian myeloblastosis virus onc genes

Bodmer WF {1982): Chromosomal localization of human cellular homolognes of two viral oncogenes

Oncogene from human EJ Bladder carcinoma is located on the short arm of chromosome 11

Weinberg RA {in press): Human c-Ki-ras2 proto-oncogene on chromosome 12

Shows TB {in press): A human gene homologous to RSV v-src is on chromosome 20

Transcripts from the cellular homologs of retroviral oncogenes: Distribution among chicken tissues

At least two regions of the viral genome determine the oncoganic potential of avian leukosis viruses

Tumorigenic transformation of mammalian cells induced by a normal human gene homologous to the oncogene of Harvey murine sarcoma virus

Transforming activity of human tumor DNAs

Three different human tumor cell lines contain different oncogenes

Human-tumorderived cell lines contain common and different transforming genes

Passage of phenotypes of chemically transformed cells via transfection of DNA and chromatin

Transforming genes of carcinomas and neuroblastomas introduced into mouse fibroblasts

Activation of related transforming genes in mouse and human mammary carcinomas

Stage-specific transforming genes of human and mouse B-and T-lymphocyte neoplasms

Transforming genes of human bladder and lung carcinoma cell lines are homologous to the ras genes of Harvey and Kirsten sarcoma viruses

Human EJ bladder carcinoma oncogene is homologue of Harvey sarcoma virus ms gene

The 1981 catalogue of assigned human genetic markers and report of the nomenclature committee

Leukocyte and fibroblast interferon genes ~e located on human chromosome 9

The prolactin gene is located on chromosome 6 in humans

The insulin gene is located on the short arm of chromosome 11 in humans

Genes for growth hormone, chorionic somatomammotropin, and growth hormone-like gene on chromosome 17 in humans

Chromosomal localization of the human placental lactogen-growth hormone gene cluster to 17q22-~24

Human chromosome 11 is syntenic with human specific fibronectin production in human × mouse cell hybrids

Two human c-onc genes are located on the long arm of chromosome 8

Chromosomal localization of the human homolog (c-sis} of the simian sarcoma virus onc gene

The virus of Japanese adult Tcell leukaemia is a member of the human T-cell leukaemia virus group

Mechanism of activation of a human oncogene

A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene

Human type 1 procollagen genes are located on different chromosomes

Mapping the human genome, cloned genes, DNA polymorphisms, and inherited disease