Andre Goy, MD
Editor-in-Chief of Oncology & Biotech News
Chairman and Director Chief of Lymphoma Director, Clinical and Translational Cancer Research
John Theurer Cancer Center at Hackensack University Medical Center
While it is recognized that cancer is a collection of complex and diverse pathological and biological entities, at the genetic level, cancer arises from a clone that has accumulated enough requisite somaticallyacquired modified genes that force normal cells to grow abnormally. A mutation is a single base change in the DNA, like a “tax” that must be paid every time a cell replicates. While such spontaneous mutations normally arise stochastically at a very low frequency, they become fairly common over time in an adult body. For the most part, however, they have no clear effect on the affected cell behavior due to either redundancy in the cell and/or the silent nature of the mutation (ie, bystander or passenger mutations).
In contrast, oncogenesis results form the cumulative effect of driver mutations, which appear to occur at a surprisingly small number in any cancer cell subtype—perhaps three to six in most cases. The development of the original clone and its progression to disseminated disease has previously been thought to be a linear process driven by serially acquired new mutations that continue to give growth advantage to the cell. Such changes—which include base pair substitutions, small insertions or deletions, chromosomal rearrangements, and gains and losses in gene copy number—occur either spontaneously or due to selection pressure (eg, acquired chemoresistance).
While traditional gene-focused characterization studies using PCR-amplification and conventional sequencing (Sangerbased capillary sequencing methodology) have shed light on cancer cell biology, identifying frequently mutated genes across different cancer subtypes, they were limited by cost and throughput. The advances in technological tools for massively parallel, high-throughput sequencing of DNA has enabled the comprehensive characterization of somatic mutations in large number of tumor samples, illustrating an emerging shift toward a potential new classification of cancers, as I have previously discussed in this column (Hoadley KA, et al. Cell
. 2014;158:929-944). In particular, whole genome sequencing and whole exome sequencing have revealed the extraordinary and unanticipated diversity of cancer, not only among patients, but also within a patient through impressive clonal heterogeneity (not just a linear evolution of the original clone).
Although a mutational landscape review for each tumor subtype is well beyond the scope of this column, the ability to identify which mutations are likely to be “drivers” in cancer pathogenesis and to elucidate how mutated genes affect the biology of a given tumor are fundamental current challenges that might provide hope in cancer genomics in the future.
First, the number of mutations varies enormously among cancer subtypes: for example, the mutation rates for pediatric and hematological cancers show the lowest mutation rates (about 1 mutation/Mb for chronic lymphocytic leukemia), while much higher (10-15 times) mutation rates are seen in melanoma and lung cancer, likely reflecting exposure to environmental mutagens. Not only the mutation rate but also the spectrums of mutations vary among cell types based on exposure to carcinogens and particularly, mutations in mismatch repair genes. These factors likely explain in part why hematological cancers, particularly in pediatric patients, are more sensitive to chemotherapy compared with other tumor types, as well as hematologic cancers in elderly patients. On the other hand, in the era of checkpoint inhibitors, such as those used in melanoma, patients with more genetically complex tumors seem to be more immunogenic, and hence respond better to immunotherapy.
Second, what is the role of chance versus environmental/ behavioral factors versus genetic predisposition in the development of cancer? One of the most unexplained issues in oncology is why some tissue types give rise to most human cancers (as reflected in breast, colon, lung, and prostate cancers), as opposed to some very rare tumors, such as sarcomas. A report earlier this year suggested that the lifetime risk of cancers among different tissue types is strongly correlated with the total number of divisions of the normal self-renewing cells maintaining that specific tissue’s homeostasis (Tomasetti C, Vogelstein B. Science. 2015;347:78-81). Such mutations could be explained in large part by how many stem-cell divisions—each of which associated with the risk of random mutations—have accumulated by a certain age or over a lifetime of stem-cell divisions. This report also suggested that only one-third of the variation in cancer risk among tissues is attributable to environmental factors or inherited predispositions.
Not surprisingly, this assertion generated a huge reaction online and across the medical community with the perception that the majority of cancer-associated mutations are due to “bad luck,” through random mutations arising during DNA replication in normal, noncancerous stem cells. Numerous authors argued about those conclusions, particularly regarding the difficulty to assess the actual number of stem-cell divisions for each organ or tissue and the risk to discourage prevention measures against cancer. Tomasetti and Vogelstein replied with an interesting analogy using a car journey: the biggest factor being “the longer the trip (ie, the longest renewal of stem cells), the higher the risk of accidents,” with bad driving (environmental/behavioral factors) or poor maintenance of the car (inherited genetic factors) both increasing the risk of problems during that journey.
Though knowing the mutational landscape appears essential to better understanding and stratification of patients and a required effort toward precision medicine, these mutations are often too many/too complex to stop or control. Understanding the hierarchy among such mutations seems a prerequisite to using them as real target. A recent report suggests that interfering with a single gene can turn colorectal cancer cells into normal cells in experimental models (Dow LE, et al. Cell. 2015;161:1539-1552). The adenomatous polyposis coli (APC) tumor suppressor is mutated in the vast majority (up to 90%) of human colorectal cancer cells. A mouse model where APC could be conditionally suppressed showed regression of established tumors through widespread tumor-cell differentiation. Though this research is only preclinical, it is very encouraging and could apply to other gastrointestinal tumors.
Though some argue that high throughput approach to cancer has not delivered, I maintain that we are at a new dawn and should embrace customized Cura Personalis!