Reply to thread

[h=3]A Sardonic Sense of Humor[/b]Nobody saw this coming.  Between the spring of 2009 (when Watson penned his op-ed), and now, mountains of sequencing data has come in, reveling the mutational profile of many cancers, including ovarian, pancreatic, lung, melanoma, brain, breast, and several forms of leukemia, and the data is anything other than what was expected.  Rather, The Cancer Genome Atlas Project had revealed something completely unexpected.  The mutations that were always thought to sequentially sabotage critical cellular machinery – marching a cell, step by step, toward a chaotic, aggressive, uncontrolled, and invasive killer – simply made no sense.

Researchers believed the sequencing data would reveal a nice and orderly sequence of maybe 3 to 8 oncogenes that when mutated, manifested in a specific type of cancer – an identifying signature like a fingerprint – and they would work off this mutational signature with cures to follow, as Watson suggested.  But what they found instead was an almost random collection of mutations – not a single one, or any combination for that matter, being absolutely responsible for initiating the disease.  In 1976, after an arduous six decade long search for oncogenes –commenting on the vicissitudes, complexities, and surprises of cancer biology, renowned scientist Peyton Rous said, “Nature has a sardonic sense of humor”.  He had no idea how hauntingly prophetic that statement was to be.

[h=2]IF cancer researchers ever HAD a collective “STAND-with-your-MOUTH-WIDE-OPEN in SHOCK” moment, it is RIGHT NOW.[/b]You won't read about this in the newspapers yet, about the confusing data coming out of the TCGA.  Mostly because the data is still being collected, and right now the entire field of cancer biology is collectively in the middle of a hasty and massive reorganization.  If cancer researchers have ever had a collective “stand-with-your-mouth-wide-open in shock” moment, it is right now.  Some researchers are transfixed, sort of staring at each other in disbelief – looking to each other for clues as to what to do next.  Others, clinging onto a lifelong investment in the somatic mutation theory of cancer, are desperately trying to make the data work – modifying the current theory to account for the seemingly random data.  And yet others have moved on – embracing different theories to explain the obtuse sequence data.  But to be sure, if cancer biology was marching in a straight orderly line a short time ago, it has now converged into a cloud of chaos, with everyone running in different directions, picking new teams.  Not yet mainstream, this phase of the battle is still being fought in scientific journals – It's true you can't read a review of the data from the TCGA without encountering the words, “sobering”, “incredibly complex”, or “Immense therapeutic implications.”  Most mainstream publications seem to avoid the topic altogether.

Including Siddhartha Mukherjee's 2010 book on cancer, “The Emperor of all Maladies.”  Time magazine called it one of the 100 most influential books of the last 100 years.  It is a wonderfully written, rich historical journey of cancer, from its distant past all the way to the present.  The book portrays the excruciatingly difficult journey scientists have encountered in their efforts to understand and combat this disease, and it culminates with the Cancer Genome Atlas Project – a subject Mukherjee gives remarkably little attention.  In the previous chapters of Mukherjee's book he brings the characters to light in delightfully colorful detail and vivid texture, weaving an inspired and imaginative narrative. The progression of scientific discovery, in Mukherjee's book, all leads to the TCGA as the one tool needed to coalesce the randomly scattered pixels of data into a complete image of understanding.  Generations of effort culminating in one final unveiling, lifetimes of struggle, disappointment, and frustration would not have been in vain because the TCGA would completely reveal our foe.  It is strange because Mukherjee seems to just sort of gloss over it – only introducing a single scientist, Bert Vogelstein, to walk the readers through the data.  The one project destined to finally make a cure realistic seemed like it deserved so much more in such a comprehensive work on the story of cancer – but there is a good reason Mukherjee gives little attention to the TCGA – the data is virtually incomprehensible.

In fact, the grim details of the data from the TCGA are absent from many narratives – it's as if the cancer community is desperately waiting with their breath held for the data to make sense.  The most striking feature of the journal articles and reviews is what is not there, as if omitting something will just make it go away.

To truly appreciate the situation cancer biology is in, you have to take a brief walk through the data from the TCGA– it seems remarkably few have, and even fewer fully realize the implications and consequences of the data.

The unsettling data came in slowly at first. Between 2002 and 2003 the first large-scale efforts to systematically screen individual tumors from colon cancer samples for somatic mutations contained the first surprise for cancer researchers — remarkably few previously unknown oncogenes were identified.  It was sort of assumed that new key oncogenes would be identified – genes that would be implicated as causative when mutated.  But that was not to be the case – maybe the decades of work teasing out oncogenes had been more thorough than researchers realized.

The initial studies were also relatively limited – limited in the fact they did not sequence the entire 20,000 genes contained within the human genome.   The more comprehensive studies to come would surely reveal more.  The next cancers in line to be sequenced were breast and colon.  These studies would delve further into the genomes of these cancers than the previous work, hopefully culling out the handful of genes that cause these two types of common cancers.  But like before, when the results were published between 2006 and 2007 – that was not the case.

Again no new oncogenes were found, but far more unsettling than that, was the beginning realization that none of the mutations found were conclusively determined to be responsible for the origin of the disease.  In order for the somatic mutation theory to work, mutational patterns must be found that explain the origin of a given type of cancer – cause must precede, and explain effect.  Critically, the mutations determined to start and drive the disease were different from person to person – vastly different.  No single mutation could be identified that was required for the disease to start, no combination of mutations, for that matter, could be found that initiated the disease.  Other than a few commonly mutated oncogenes, the mutational pattern appeared to be largely random.  These studies sequenced the tumors from 11 different individuals with breast cancer, and 11 different individuals with colon cancer.  Over 18,000 genes were sequenced, almost 40 times the amount in the initial studies – the most exhaustive sequencing to date

In the meantime the technology continued to improve.  Sequencing technology became faster, more accurate, and cheaper.  Armed, reinvigorated, and determined, pancreatic cancer was on deck.  This time, in 2008, teams of researchers would again sequence over 20,000 genes, nearly all of the predicted protein coding genes in the human genome from the tumors of 24 individuals suffering from pancreatic cancer.  But it was more of the same.  Again no new mutations of any significance were found, and again the mutations they did find were unable to be assigned as definitely causative.  The somatic mutation theory was in trouble – a modification was needed to make the theory continue to work.

 

[h=3]The Search for Dark Matter[/b]This is where Bert Vogelstein, the scientist introduced to us in Mukherjee's book, the one chosen to walk the readers through the results of the Cancer Genome Atlas Project, returns to the story.  Vogelstein knew the Somatic Mutation Theory was in trouble and needed a modification.  Enough data was compiled to conclusively determine that the idea of a nice and tidy series of sequential mutations as the cause of cancer could be scrapped, an idea Vogelstein had championed for decades.  In its place, Vogelstein slightly tweaked the original theory, proclaiming that rather than a defined set of specific mutations being the cause of a given cancer, cancer is caused by mutations that render certain biological systems dysfunctional – systems involved in the qualitative aspects of cancer, like uncontrolled proliferation, inhibition of programmed cell death, and tissue invasion.  In other words; cancer was a cellular systems disease.  A given system might have say 20 or so constituent genes required for it to operate – so the theory goes – if any single one of the constituent genes was rendered dysfunction by a mutation, then the whole system was made non-operational, marching the cell one step closer to malignancy.

Some criticisms by other cancer biologists claim this was simply and ad hoc modification necessary to make a failed theory continue to fit the data.  But I don't think so.  The modification to a systems disease seemed reasonable to me.  For sure it is a broadening, or a dilution of definition, for sure it would make the data easier to fit.  That is not a reason alone to discard the new modified theory however.   But the data would have to validate it.  Time and more sequencing data would tell.  The authors of the pancreatic cancer study said this about the somatic mutation theory's new paradigm shift, “From an intellectual viewpoint, the pathway perspective helps bring order and rudimentary understanding to a very complex disease.”

Applying the new modified theory to the pancreatic cancer study determined that pancreatic cancer was caused by the dysfunction of 12 different biological systems.  Now a critical eye must be cast on just how diluted this new, modified theory had become.  In this case it seems it was pretty watered down.  It turns out, the authors had to use some imagination in order to assign some of the mutations to one of the 12 systems implicated in the pathogenesis of pancreatic cancer.  It appeared that some of the mutated genes were friends, of a friend, of a friend, that was definitely part of the implicated system.  By the authors own omission, “Although we cannot be certain that every identified mutation plays a functional role in the pathway or process in which it is implicated.”  Rather than bringing order and rudimentary understanding to a very complex disease – it seemed like the authors were manufacturing order and understanding to a very complex disease.

Despite the confusion, the TCGA soldiered on.  Glioblastom Multiforme was next – brain cancer.  Glioblastom is a mean aggressive-cancer; most will succumb to it within a year even with treatment.  Again, teams of researchers sequenced over 20,000 genes from 22 tumor samples.  This time a novel gene was found to be mutated in 12% of the samples – a big accomplishment.  Its discovery was cited as a validation of the utility of genome-wide genetic analysis of tumors. The authors concluded that GBM was caused by mutations that rendered 3 important biological processes dysfunctional.  However, as with pancreatic cancer, a close look at the data reveled something else.  The disturbing trend continued – none of these studies were able to validate the somatic mutation theory of cancer, not even the new modified version.  None of these studies were able to conclude that mutations were even the cause of the disease at all.  Of the 22 samples only 4 had mutations involving all 3 systems implicated as necessary for GBM to occur.  Nine samples had mutations in 2 of the 3 systems, 5 had mutations in 1 of the 3, and most significant, one sample (sample labeled Br20P) had no mutations in any of the 3 systems yet was a living, growing, aggressive case of GBM.  The profound silence with regard to these inconsistencies in the new and modified somatic theory of cancer speaks volumes.  For the theory to work, the original theory, or the new modified theory, samples like Br20P simply cannot exist.

[h=2]For the GENETIC theory of cancer to work, the original theory, or the new modified theory, samples like Br20P simply CANNOT exist.[/b]A little over a year ago, the sequence data was released on over 21,000 genes from 100 breast cancer samples, the most comprehensive to date, and for the somatic mutation theory of cancer; the most damning to date.  Like the other studies, the theory itself is not questioned.  Just silence.  The authors do again pay homage to the complexity of the sequence data, declaring, “The panorama of mutated cancer genes and mutational processes in breast cancer is becoming clearer, and a sobering perspective on the complexity and diversity of the disease is emerging.  Driver mutations are operative in many cancer genes.  A few are commonly mutated, but many infrequently mutated genes collectively make a substantial contribution in myriad different combinations.”

That statement does not even approach a realistic description of the complexity found in the mutation-profile of breast cancer, or most types of cancer.  From the 100 samples sequenced, 44 genes were implicated as being involved in the tumorigenesis of breast cancer.  The maximum number of mutated cancer genes in an individual breast cancer was 6, but 28 cases showed only a single driver mutation.  If you were to ask 100 oncologists, or cancer research scientists 10 years ago if breast cancer could be caused by a plethora of different single mutations, all 100 probably would have laughed at you.

Much worse, in yet another glaring omission, the authors failed to even make mention of five samples that had no mutations at all – no driver mutations found, yet these were living, breathing, aggressive killer cancer cells.  Again, for the somatic mutation theory of cancer to work – samples like these can't exist.

When I asked Dr. Larry Loeb of the University of Washington, one of the key players in the CGAP, to summarize in a few sentences what has been learned so far from the sequence data –he spoke slowly and deliberately, “There are enormous numbers of mutations present in each tumor – and it is very, very difficult to determine which ones are causative.  We do not have an adequate armament of effective drugs to target the spectrum of mutant genes within individual tumors.  The mutational complexity found in cancer is truly daunting.”  In many ways the somatic mutation theory of cancer seems like a grand-scale example of groupthink.  There is no-way mutations can be completely responsible for the origin of cancer – yet so few seem willing to say it – maybe it's because the discovery of DNA, and its central role as the dictator of life's processes was such a profound intellectual achievement, that nobody is willing to question its primacy in the etiology of cancer.  Maybe this is just how slow entrenched, dogmatic belief-systems are to change course.  Whatever the case may be, billions are still being spent chasing down and cataloguing the mutations thought to cause cancer.  And billions again are spent developing one failed drug after the next that target these mutations.  Drugs that typically cost up to 100,000 per treatment giving patients maybe a few months at best – many offering no increase in survival time at all.  We must all remember this is no intellectual exercise, this is not theoretical physics or astronomy where one theory slowly discards another, after taking careful consideration, and there is no need for a sense of great urgency.  But in the case of cancer research, there needs to be a sense of great urgency – this is war.  People are still dying.  Time is not a luxury many have.

I emailed Dr. Vogelstein, asking him about the inconsistencies of the data.  Specifically I asked him how he explained samples like Br20P, the brain cancer sample with no mutations in any of the 3 broad systems determined by Vogelstein to be required for the formation of cancer.  He politely referred me to his latest review in the highly esteemed journal Science.

In his review, Vogelstein does attempt to address the problems with the data from the CGAP.  First off, he explains that genomic wide sequencing technology is still far from perfect and has been shown to have a false-negative error rate of up to 37%.  However, even if one takes into account the potential error rate of sequencing the data still doesn't work – another explanation is needed.  And Vogelstein offers one up in a section titled “dark matter.”

In the 1930's it was noticed that the orbital velocities of galaxies, including our own Milky Way, didn't make sense.  Galaxies were rotating much faster than predicted by classic Newtonian mechanics – something else was at work here, something that could not be seen.  The explanation came in the postulated existence of an invisible material termed “dark matter”, an ephemeral, undetected-material that was physically influencing the world around us, and physicists are still hunting for this material today.  In fact, 40 miles from my home is the latest incarnation of this 80 year search for dark matter.  In a now abandoned goldmine in the Black Hills of South Dakota a colossal effort is underway to build the infrastructure necessary to capture just one of these elusive particles, furthering humanities understanding of the universe we live in.

Vogelstein borrowed the term dark matter from astrophysics and applied it to the gaping hole in understanding revealed by the Cancer Genome Atlas Project.  Vogelstein is well aware that some nebulous, presumptive-process is preventing the complete picture of cancer from being realized.  He just has to find the dark matter – it's just that he might be looking in the wrong place.


Back
Top