Chromosomes Blown to Smithereens

Where do cancers and cancer relapses come from?

DNA is a treasure trove that keeps on giving. The human genome sequence was a milestone that may not have been self-interpreting, but has provided grist for leaps of technical advancement and knowledge. Ancestry studies are one example, but disease studies are of more immediate interest. Cancer is now understood to be a molecular disease where the DNA suffers mutations that release various brakes on cell proliferation. One of the most influential types of mutations are gene fusions, where one gene that has roles in proliferation is broken from its normal regulatory controls, either within its coding sequence (such as a repressing protein domain) or its upstream expression controls, and hooked up with some other gene that drives its expression in new places and high levels. A recent paper studied several cancers in detail, sequencing samples from various time points and locations, coming up with very interesting findings about the origins of these mutations and the nature of metastasis.

One example of a genome blow-up, called "chromoplexy". A few regions of the genome got caught in some kind of spindle, and came out with several breaks which then were repaired to form re-joined fusions. In this diagram (right) of one resulting fusion, of genes BCLAF1 and GRM1, the chromosome 6 parts on the outside have rejoined, while the broken parts between the rejoined ends have fused to each other and then to chromosome 16, with one small bit unassigned and perhaps ending up somewhere else. The diagram seems to indicate that GRM1 ends up upstream of BCLAF1, (these are divergently transcribed in the native chromosome), which I think is an error.

Chromoplexy is one form of a genome blowup, one that is restricted in scope (at least compared with the even more destructive chromothripsis). The best theory about its origin posits that the affected portion of the genome (typically an early-replicating and transcriptionally active region) gets caught outside the normal nucleus, forming a temporary mini-nucleus which is cut off from normal controls, causing the trapped DNA to break up. The cell has strong controls against free DNA ends, and uses end-joining DNA repair to patch things up, pasting ends together essentially at random. This is obviously quite dangerous, and leads to unexpected gene fusions, of which hundreds of different examples are now known that drive various cancers. One such fusion, diagrammed above, is between genes BCLAF1 (upstream) and GRM1 (downstream).  GRM1 is a receptor for glutamate, the most prevalent excitatory neurotransmitter. While most highly expressed in the brain, glutamate receptors act throughout the body, and malfunctions are connected with a variety of diseases. Increased expression and activation can drive cell proliferation. The other fusion partner, BCLAF1, is a promoter of cell suicide, or apoptosis. That function will be lost in the fusion, which might have some importance to the disease (though a second copy presumably remains intact elsewhere). The important part is that it is very widely expressed, especially in bone marrow. An earlier paper describing this fusion states:

"The GRM1 coding region remains intact, and 18 of 20 CMFs (90%) showed a more than 100-fold and up to 1,400-fold increase in GRM1 expression levels compared to control tissues. Our findings unequivocally demonstrate that direct targeting of GRM1 is a necessary and highly specific driver event for CMF [bone tumor chondromyxoid fibroma] development."

This pattern of mutation, and the specific fusions that resulted, became apparent due to the deep sequencing the researchers did, taking samples from the patient's tumors and from normal tissues. An important concept here is of mutational signatures. Each mechanism of mutation has its characteristic pattern of mutations left in the genome. Exposure to UV light, which causes C->T mutations, will leave a much different pattern in the genome than the localized chromoplexy blowup mentioned above. So a forensic analysis of the patient's DNA can tell what happened, in some mechanistic detail. For example, the various fusions seen in these samples were not part of extensive copy number variations- reduplications that are common in cancerous cells, which indicated that this blowup took place once as a discrete event, not repeatedly or slowly over a long period of time.

It can also tell when it happened, and here we get to a particularly interesting message from this paper. When they sequenced primary and relapsed tumors, (with comparisons to normal tissue), such tumors shared some key mutations, those which drove the overall cancer. But they failed to share many others. Indeed, the metastatic tumors carried none of several mutations that were uniformly present in the primary tumor. This says that metastases or relapse cancers, (this part of the study was specific only to Ewing's sarcoma, a bone cancer typically arising around ages 1-20), typically do not develop from the primary tumor, but from cells that carry the same driver mutation, but diverged before primary tumor formation. They are independent events, and metastatic prognosis has little to do with the fate of the primary tumor.

The author's proposed time course of Ewing's sarcoma evolution, placing the origin of metastatic and relapsing tumors well before and outside of the primary tumor at the time of diagnosis.

Whether this observation about metastisis applies to other tumors is naturally important to follow up. It would alter significantly how we deal with primary tumors, and informs the kind of conservative treatments (lump-ectomies, for instance) that are becoming more common. As sequencing becomes cheaper and more common for all kinds of tumors, the particular drivers, from whatever mutational source, can be identified and used to direct specific, (buzzword: "precision") treatments. GRM1 can be targeted by direct or indirect means. But if one has Ewing's sarcoma, typically associated with a fusion of EWSR1-FLI1, where FLI1 is a transcription factor that drives growth factor production and hence cell proliferation, a different set of therapies would be indicated.

• The Green dream is on.
• Our cold war with China.
• AIPAC is the NRA of foreign policy.
• Bill Mitchell on economics education, especially macro.
• Election crisis.. no problem.
• Give to the rich, take from the poor.
• Notes on experts.