Q: What are the similarities and differences between “driver” mutations and “passenger” mutations and in what common malignancies is that distinction most important?
A: The commonly accepted definition of a driver mutation is a mutation within a gene that confers a selective growth advantage (thus promoting cancer development), while passenger mutations are those that do not provide a growth advantage. Independent of context, the type of mutation observed is not a factor in differentiating between a driver versus a passenger mutation. However depending upon whether the driver gene is classified as an oncogene or tumor suppressor, the type of mutation observed can play a role in determining whether it is a driver or passenger. For instance, driver mutations in oncogenes tend to be missense mutations at specific codons or focal amplifications, while nonsense or frameshift mutations or focal deletions are often the hallmark driver mutation type in tumor suppressors. Driver mutations have a tendency to occur in protein-coding regions of genes and within important functional domains of the protein, although it’s increasingly being recognized that non-coding mutations, like splice-site or promoter mutations, can also be driver mutations. Thus, using mutation location as a discriminatory factor may be becoming a less reliable indicator of whether a mutation is a driver or passenger. Additionally, driver mutations are often somatic in origin, with germline mutations often fast-tracked to the passenger bucket; however, a cautionary note should be inserted here as there are very clear examples of where germline mutations are driver mutations (e.g. BRCA1/2 in familial breast and ovarian cancer or TP53 mutations in Li-Fraumeni syndrome).
In terms of the ‘how’, there are generally two methods or approaches to classifying a mutation as a driver or passenger: 1) by frequency (driver mutations should be mutated in a greater proportion of cancer samples than would be expected from the background mutation rate) and/or 2) by prediction of functional impact (either via in-silico algorithms or cell/model-based assays). Each method is fraught with caveats and disadvantages or challenges, however the gold standard of evidence that a mutation is a driver is experimental evidence demonstrating that the mutation produces a cellular phenotype that provides a selective growth advantage to the cell. Thus, importantly, bioinformatic methods cannot provide definitive classification of mutations as either driver or passenger but can be a means by which to prioritize mutations for functional testing.
Because driver mutations are by definition those resulting in cancer initiation and/or progression, they are seen as the ‘achilles’ heel’ of tumors, sought after as targets for drugs, and used in making therapeutic decisions. Thus, being able to make distinctions about whether a mutation is driver or passenger is important for any malignancy. However, being able to make this distinction is harder in some cancer types than others. For example, lung cancer has a much higher mutational burden than acute myeloid leukemia (AML), which makes the identification of driver mutations in lung cancer more difficult than in AML. Passenger mutations perhaps shouldn’t be dismissed entirely, as emerging data and theories suggest that passenger mutations can transform into driver mutations (so-called “latent drivers” or “mini-drivers”, amongst other proposed terms), especially within the context of resistant and/or recurrent disease. However given the high number of passenger mutations usually present in tumors, it will be hard to discriminate between those passengers likely to ‘stay put’ and those with hidden driver potential, requiring more investigative studies.