Studies of both chromosomes and nucleic acids support Boveri's 1914 hypothesis that cancer is caused by a change in genetic material at the cellular level. Three classes of genes can be involved in neoplastic transformation.
Oncogenes arise from preexisting normal genes (protooncogenes) that have been altered by both viral and nonviral factors. As a result, the cells synthesize either normal proteins in inappropriate amounts or proteins that are aberrant in structure and function. Many of these proteins are cellular growth factors or receptors for growth factors. The net result of oncogene activation is unregulated cell division. Mutations that activate oncogenes almost always arise in somatic cells and are not usually inherited. Although some oncogenes are more likely to be activated in certain tumors, in general the same mutations may be found in neoplasia arising in different cells and tissues.
Tumor suppressor genes can be viewed as the antithesis of oncogenes. Their normal function is to suppress transformation; mutation in both alleles is necessary to obliterate this important function. The first mutant allele at any tumor suppressor gene might arise spontaneously or might be inherited; mutation in the other allele (the "second hit") almost always arises spontaneously but by any of a number of molecular mechanisms. These genes show considerably more tumor specificity than do oncogenes; however, although some specific mutations are necessary for certain tumors to arise, no loss of single tumor suppressor function is sufficient. Clearly, a person who inherits one copy of a mutant tumor suppressor gene is at increased risk that in some susceptible cell, at some time during life, the function of that gene will be lost. This susceptibility is inherited as an autosomal dominant trait. For example, mutation in one allele of the p53 locus results in the Li-Fraumeni syndrome, in which susceptibility before age 45 years to sarcomas and other tumors occurs in males and females in successive generations. Inherited mutations in this locus also increase the risk that a second tumor will develop following radiation or chemotherapy for the first tumor, suggesting that the initial treatment may induce a "second hit" in a p53 locus in another tissue. However, inheriting a p53 mutation is not a guarantee that cancer will develop at an early age. Much more needs to be learned about the pathogenesis of neoplasia before the genetic counseling of families with a molecular predisposition to cancer is clarified. BRCA1, a gene that predisposes women to breast and ovarian cancer, is another example of a tumor suppressor gene. Women who inherit one mutant allele of BRCA1 have, on average, a 60–80% lifetime risk of developing breast cancer, and the average age of tumor detection is in the fifth decade. Their risk of developing ovarian cancer is 34–45%. For both females and males with certain mutations in BRCA1, the risks of colon and pancreatic cancer are increased several-fold over that of the general population.
In selected cases, a patient's DNA can be analyzed for the presence of a mutated gene and thereby that individual's risk for developing a tumor can be assessed. Examples are retinoblastoma, certain forms of Wilms tumor, breast cancer, and familial colon cancer. To illustrate how noninvasive and sensitive the methodology has become, it is possible to analyze stool for the presence of mutations in tumor suppressor genes that might indicate the presence of a clinically undetected adenocarcinoma of the colon. The analysis—not yet in general use—depends on the ability of the PCR to amplify minute quantities of the mutant DNA present in epithelial cells shed from the tumor.
A third class of genes that predispose to malignancy is the so-called mutator of DNA repair genes. Mutator genes normally function to repair damage to DNA that occurs from environmental insults such as exposure to carcinogens and ultraviolet irradiation. When a mutator gene is itself mutated, DNA damage accumulates and eventually affects oncogenes and tumor suppressor genes, thereby making cancer more likely. Hereditary nonpolyposis colon cancer (HNPCC) is one familial syndrome due to mutations in one or another of the five mutator genes identified thus far (MSH2 and MLH1 being the most commonly responsible for HNPCC).
This exciting work on the molecular nature of oncogenesis was preceded by years of study of the cytogenetics of tumors. The first important discovery was the 'Philadelphia chromosome,' an apparently shortened G-group chromosome in many lymphocytes of patients with chronic myelogenous leukemia. Eventually, this was shown to be the result of a translocation between the long arm of chromosome 9 and the long arm of chromosome 22. This led to the recognition that the ableson oncogene (chromosome 9) is highly expressed because of the juxtaposition of the BCR region of chromosome 22. Additionally, the retinoblastoma tumor suppressor gene was ultimately isolated because a small number of patients with this tumor have a constitutive deletion of chromosome 13 where this gene maps. Other chromosomal aberrations have been found to be highly characteristic of—or even specific for—certain tumors (eTable 40–7). Detection of one of these cytogenetic aberrations can thus aid in diagnosis.
eTable 40–7.Chromosome aberrations associated with representative solid tumors. ||Download (.pdf) eTable 40–7. Chromosome aberrations associated with representative solid tumors.
|Tumor ||Chromosome Aberration |
|Meningioma ||del(22)(q11)1 |
|Neuroblastoma ||del(1)(p36), del(11)(q23) |
|Renal cell carcinoma ||del(3)(p14.2–p25) or translocation of this region |
|Retinoblastoma, osteosarcoma ||del(13)(q14.1) or translocation of this region |
|Small-cell lung carcinoma ||del(3)(p14–p23) |
|Wilms tumor ||del(11)(p15) |
Hematologic malignancies are especially amenable to study because of the relative ease of accessing tumor cells for cytogenomic analysis. Such malignancies are associated with over 100 specific genomic rearrangements, primarily translocations. Most of these rearrangements are restricted to a specific type of cancer (eTable 40–8), and the remainder occur with many cancers.
eTable 40–8.Chromosomal aberrations associated with representative hematologic malignancies. ||Download (.pdf) eTable 40–8. Chromosomal aberrations associated with representative hematologic malignancies.
|Tumor ||Chromosomal Aberration |
|Acute myeloblastic ||t(8;21)(q22;q11)1 |
|Acute promyelocytic ||t(15;17)(q22;q11–q12) |
|Acute monocytic ||t(10;11)(p15–p11;q23) |
|Chronic myelogenous ||t(9;22)(q34;q11) |
|Burkitt ||t(8;14)(q24.1;q32.3) |
|B cell ||t(1;14)(q42;q43) |
|T cell ||inv, del, and t of 1p13–p12 |
|Polycythemia vera ||Del(20)(q11) |
In the leukemias, the chromosomal aberration is the basis of one of the subclassifications of the disease. When cytogenomic information is combined with the histologic classification, it is possible to define subsets of patients in whom response to therapy, clinical course, and prognosis are predictable. If at the time of diagnosis there are no chromosomal changes in the bone marrow cells, the survival time is longer than if any or all of the bone marrow cells have abnormal cytogenomic characteristics. As secondary chromosomal changes occur, the leukemia becomes more aggressive, often associated with drug resistance and a reduced chance for complete or prolonged remission. The least ominous chromosomal change is numerical alteration without morphologic abnormality.
Less cytogenomic information is available for lymphomas and premalignant hematologic disorders than for leukemia. In Hodgkin disease, studies have been limited by the low yield of dividing cells and the low number of clear-cut aneuploid clones, so that complete cytogenomic analyses are available for far fewer patients with Hodgkin disease than for any other type of lymphoma. In Hodgkin disease, the modal chromosomal number tends to be triploid or tetraploid. About one-third of the samples have a 14q+ chromosome. In non-Hodgkin lymphomas, cytogenomic abnormalities occur in 95% of cases. Cytogenomic findings are now being correlated with the immunologic and histologic features and with prognosis.
In Burkitt lymphoma, a solid tumor of B cell origin, 90% of patients have a translocation between the long arm of chromosome 8 and the long arm of chromosome 14, with chromosomal breakage sites being at or near immunoglobulin and oncogene loci.
Instability of chromosomes also predisposes to the development of some malignancies. In certain autosomal recessive diseases such as ataxia-telangiectasia, Bloom syndrome, and Fanconi anemia, the cells have a tendency for genetic instability, ie, to chromosomal breakage and rearrangement in vitro. These diseases are associated with a high incidence of neoplasia, particularly leukemia and lymphoma.
Some chromosomal aberrations, better known for their effect on phenotype, also predispose to tumors. For example, patients with Down syndrome (trisomy 21) have a 20-fold increase in the risk of childhood leukemia, 47,XXY males (Klinefelter syndrome) have a 30-fold increase in the risk of breast cancer, and XY phenotypic females have a heightened risk of developing ovarian cancer, primarily gonadoblastoma.
The indications for cytogenomic analysis of neoplasia continue to evolve. Not all tumors require study. However, in cases of tumors of unclear type (especially leukemias and lymphomas), with a strong family history of early neoplasia, or for certain tumors associated with potential generalized chromosomal defects (present in nonneoplastic cells), cytogenomic analysis should be strongly considered.
Increasingly, the specific genetic mutations that are present in a person's tumor tissue are being used, not just for prognosis, but to guide therapy. Several examples among many include mutations in BRAF in melanoma, expression of HER-2/neu in breast cancer, and mutation of EGFR in non–small cell lung cancer.
et al. Twenty-first century precision medicine in oncology: genomic profiling in patients with cancer. Mayo Clin Proc. 2017 Oct;92(10):1583–91.
et al. Clinical actionability of multigene panel testing for hereditary breast and ovarian cancer risk assessment. JAMA Oncol. 2015 Oct;1(7):943–51.
et al. Circulating tumor DNA as biomarkers for cancer detection. Genomics Proteomics Bioinformatics. 2017 Apr;15(2):59–72.
et al. The NCI Genomic Data Commons as an engine for precision medicine. Blood. 2017 Jul 27;130(4):453–9.
et al. Factors affecting the adoption and use of gene expression profiling by oncologists in clinical practice. Per Med. 2015;12(1):333–42.
et al. Long and short noncoding RNAs in lung cancer precision medicine: opportunities and challenges. Tumour Biol. 2017 Apr;39(4):1010428317697578.
et al. Multigene testing of moderate-risk genes: be mindful of the missense. J Med Genet. 2016 Jun:53(6):366–76.