Cancer: One word, many diseases

Out in the real world it’s common to hear people talk about cancer as one disease.  But in the scientific community, cancer is not one disease; it is many.  Each cancer type, each patient is unique.  It’s part of the inherent complexity of cancer.  And so cancer scientists spend their time characterizing different tumor types to determine their intrinsic similarities and differences.  These similarities and differences can – hopefully – be exploited to tailor treatments for each specific disease.    A recent publication in Nature, as part of a larger initiative to understand the cancer genome, focuses on these similarities and differences by outlining the mutational landscape across 12 major cancer types (Find article here) (1). 

The methods are relatively straightforward.  The researchers took 3281 tumors representing 12 different cancer types including breast, two different types of lung cancer, colon, kidney, bladder, ovarian cancers, as well as acute myelogenous leukemia (AML) among others, extracted and purified the DNA, and then subsequently sequenced each tumor DNA sample.  Since we already know the sequence of the human genome, they compared these tumor samples to matched normal tissue to identify regions, or specific genes, where alterations such as mutations or small insertions or deletions, occurred.   Using powerful software and statistical methods, they investigated the mutational frequency for each cancer as well as identified common and/or unique genes that were mutated.

As expected, these 12 major cancer types differ genetically.  Importantly, they also share some similarities. 

Similarities:

The principal result generated from this study was the identification of 127 significantly mutated genes involved in a broad array of cellular functions including some of the obvious functions such as DNA repair, cell cycle control, and regulation of known key signaling transduction pathways including the MAPK pathway and TGFb pathways as well as less understood roles such as metabolism, histone modification and proteolysis.  Two things become apparent from this list of genes and their functions: 1) The list is small.  Of the ~25000 genes in the human genome, only 127 genes make the list of significantly mutated genes in different cancer types.  What makes these genes primary targets for mutagenesis?  I’m not sure we know the answer to that question. 2) Although the list of mutated genes is small, the functions they perform are broad and cover most of the processes necessary for maintaining homeostasis, highlighting the importance of maintaining cellular homeostasis for disease avoidance.  Dysregulation of these cellular functions leads to tumor progression no matter the cancer type.

Of these 127 SMGs, several of them can be found across multiple cancers. The most commonly mutated gene, not surprisingly, is TP53, occurring in 42% of samples.  We’ve known about TP53 for many years now and the vital role it plays as a tumor suppressor by functioning as a transcription factor to control cell proliferation and apoptosis.   Next to TP53 is PIK3CA, which is mutated in greater than 10% of samples.  This gene codes for a subunit of the P13K protein which is essential for maintaining cell survival through induction of cell growth and inhibition of cell death.  The fact that these genes, among other, become mutated in many cancers suggests that these proteins are vital for maintaining cellular homeostasis.  Can they serve as therapeutic targets?

           

Differences

Although some overarching themes emerge that characterize cancer as a whole, the specific details defining each cancer are unique.  For example, the mutation frequency varies between tumor types, with AML having the fewest mutations and lung squamous cell carcinoma having the highest.  Not only does the mutation frequency differ, but the types of mutations are also different between cancer types.  Together, this suggests that factors other than age could contribute to the development of certain cancers.  For example, a certain type of mutation called a C>A transversion is most common in lung cancer, and not surprisingly, this type of mutation can be caused by cigarette smoke. 

           

More specifically, different tumors have different mutation signatures.  In other words, each cancer has different genes and their encoded proteins altered.  This is not a new concept, but this study takes this idea a step further by identifying driver mutations in different tumor types.  Driver mutations are identified by their early appearance in the progression of cancer and in their ability to provide the cancer cell with a selective advantage for continued growth.  By looking at the variant allele fraction (or the fraction of alleles that carry a mutation, VAF), the researchers assumed that genes with high VAF in specific cancers could be considered driver mutations.  The tumor suppressor TP53 has a high VAF across tumor types suggesting mutations in this gene are universal for driving cancer progression.  However, other genes differ in different tumor types.  In breast cancer, AKT1, CBFB, MAP2K4, ARID1A, FOXA1, and PIKCA carry mutations, whereas PIK3CA, PIK3R1, PTEN, FOXA2, ARIK1A are mutated in endometrial cancer. 

Furthermore, even within each tumor type, specific clusters can be separated out based on genetic signatures.  Therefore, not only do cancer differ, but variability also exists within that cancer type.  Thus, even breast cancer is more than one disease!

Where do we go from here?

I would not characterize this publication as one that identified novel concepts in cancer research.  We already knew that cancer is complex and differs by type.  However, by systematically analyzing the mutational landscape across 12 different cancer types, this study identified key genes and pathways involved in tumor progression.

At a scientific level, it is clear where this study should go.  We need to look at the list of 127 significantly mutated genes and assess their ability to act as therapeutic targets for cancer treatments.  Some of these can already be effectively targeted in cancer treatments including EGFR (Cetuximab, Lapatinab, for example) (2).  Moreover, the identification of rational drug combinations, targeting two of the key pathways identified in this study, may help to avoid drug resistance and disease recurrence.

At a more personal level, you could consider the implications of knowing your gene signature.  If you knew you carried a “driver mutation”, what would you do?  With companies, particularly in the United States, (23andMe) (3) offering to sequence your genome, the ability to do this is a realistic possibility albeit at a high financial cost.  In certain instances, for example in hereditary breast and ovarian cancer, knowing that you carry a BRCA1/2 mutation could cause you to undergo a radical mastectomy or hysterectomy.  Carrying a BRCA1 mutation prompted Angelina Jolie to undergo such a procedure.  What would you do? 

   

Uncovered cancer morse for today: cancer is complex: it is not merely one disease but a group of related yet distinct diseases.

References:

1. Kandoth C, McLellan MD, Vandin F, et al. Mutational landscape and significance across 12 major cancer types. Nature. 2013;502(7471):333-339. doi: 10.1038/nature12634; 10.1038/nature12634.

2. Camp ER, Summy J, Bauer TW, Liu W, Gallick GE, Ellis LM. Molecular mechanisms of resistance to therapies targeting the epidermal growth factor receptor. Clin Cancer Res. 2005;11(1):397-405.

3. https://www.23andme.com/. Updated 2013.