WHEN Robert Holt started tearing apart colorectal tumours two years ago, he was looking for evidence of microbes at work behind the scenes. He never expected his team would find the tissue crawling with a germ linked to gum disease.
Holt, a North Vancouver resident and a senior scientist at the B.C. Cancer Agency, is one of many researchers around the globe using new DNA sequencing techniques to expose the inner workings of carcinoma. His most recent discovery, however, has drawn international attention. It's not that the bacterium itself is remarkable, but that its appearance inside colorectal tumours raises the tantalizing possibility that one of the world's most feared cancers might someday be brought to its knees by a vaccine.
Anatomy of a killer
Although it's hard to rank one cancer against another, colon cancer is widely considered to be among the deadliest on the planet, killing more than 600,000 people worldwide annually, according to estimates from the World Health Organization.
Like other cancers, the disease arises when something goes wrong with the genes that keep cellular reproduction in check, causing a cell to replicate out of control and, in extreme cases, kill the host.
There are countless things, from cigarette smoke to sunlight to asbestos, that can trigger this process in the human body. In some cases, the culprit appears to be an infectious agent - a virus or a bacterium. Cervical cancer can be set off by the human papilloma virus, for instance, and stomach cancer by a bacterium called Heliobacter pylori.
So far, researchers have linked about 15 per cent of the world's cancer load to infectious agents, but Holt believes the real proportion may be much higher.
"Those are just the handful that are actually known already without looking terribly hard," he said. "For some time, we've been thinking that there are probably a lot more."
To the average non-expert, it might seem strange that we don't know. Surely, one would think, we could just look at various tumours under a microscope, see which viruses or bacteria show up in them most often, and then inject those microbes into rats or whatever to see if they're to blame. The reality is far more complex, however.
Historically, to find bacteria in a given cancer, scientists have had to take samples from a tumour and then attempt to coax whatever microbes might be there into growing in a Petri dish for identification. Unfortunately, only about one per cent of bacterial species cooperate with this process; the others just die or fail to reproduce, leaving researchers largely blind as to which germs might be present in a given piece of tissue.
Recently, however, that has changed.
A new weapon
In the last few years, a new and powerful tool has emerged from genomics, an area of study that examines a living thing's genetic information in its entirety. DNA, the astonishingly long molecule at the centre of our cells that provides instructions for the construction of the body, has been broadly understood for decades. Generally speaking, scientists have known how DNA is structured, how it functions and how it encodes information. They have been sequencing - spelling out - individual genes for many years. But sequencing a full strand of DNA with its tens of thousands of genes and billions of chemical base pairs (the letters, so to speak, in a genetic sentence) was until very recently out of reach, simply because of the scale of the task.
It was only in 2003 that a massive endeavour by two competing teams finally produced a complete reading of all the DNA in a single human cell. With technology being what it was at the time, the effort (by one of the teams, anyway) took 13 years and cost, by some estimates, upward of $3 billion.
Over the last eight years, however, technology has continued to advance rapidly. Now, sequencing has sped up so much that it has become possible for a small research team such as Holt's to read vast quantities of DNA in a short time and at minimal expense. His whole project lasted just 18 months and cost only $100,000.
Since every individual has a distinct set of genes, and since those genes vary even more markedly between species, Holt recognized that the new techniques could be used to take a kind of role call of all the living cells - including bacteria - that may be lurking in or on a tumour.
Exposing a cancer's secrets
Two year's ago, Holt's team went to the B.C. Cancer Agency's tumour tissue repository in Victoria and obtained samples drawn from 11 patients suffering from colorectal carcinoma. The deep-frozen slices came in pairs, with one piece taken from a patient's tumour and the second from the healthy tissue right next to it. The latter would serve as a control for the experiment.
The team mashed up each sample, processed it, extracted all the RNA (short molecules closely related to DNA which contain copies of a strand's active genes) and ran it through an instrument that makes use of the new sequencing technology to speed-read base pairs.
What came out, as one might expect, was the genetic signature of the patient's own cells as well as the signatures of the microbes that were living in and around them. The team then compared the extracted sequences with databases of bacterial and human genomes - another outgrowth of the new DNA-crunching technology - and eliminated all the human strands and all the microbial strands that appeared in both the healthy and cancerous tissue simultaneously (these were the innocent germs that happened to be in the area on both types of tissue, minding their own business). What they were looking for was a microbial signature that appeared in just the tumours.
They found it.
A surprise suspect
To the team's shock, a large majority of the cancerous samples were infested with a bacterium called Fusobacterium nucleatum, a well-known species linked to periodontitis (diseases around the teeth), appendicitis and a handful of other disorders, but which is not usually found in any significant numbers in the large intestine. What's more, the microbe was actually inside many of the cancer cells, rather than just on the surface, behaviour typical of ill-intentioned bugs.
Armed with this clue, Holt and his colleagues obtained samples from another 88 patients and put them through the same process, but this time looked for F. nucleatum specifically. Lo and behold, roughly two-thirds of those tumours were infested as well.
What's more, the discovery was strengthened by the simultaneous findings of another team in Boston that appeared to show the same thing. The second group, led by a researcher called Matthew Meyerson, had looked at colon tumours using a very similar method and had found the bacterium in similar abundance. Their efforts had been completely independent of Holt's, and their samples had been drawn from farther afield - from Asia, North America and Europe - lending credence to Holt's team's findings.
"What I find amazing is how similar our results were," said Holt. "It's a reciprocal validation. We were both pleased by that."
Both studies were published simultaneously in the journal Genome Research on Oct. 18. They were featured in the New York Times the same day.
The result doesn't mean, Holt is careful to point out, that the bacterium is necessarily the culprit; rather it just shows the bacterium tends to be associated with the diseased tissue. It could be that the species just happens to be well adapted to that type of cancerous cell, that it was nothing more than an innocent passerby that saw a nice, cozy already-cancerous place to set up shop and reproduce. But the finding is nonetheless important.
MAYBE, MAYBE, MAYBE
If it turns out down the road that there is no causal link between F. nucleatum and colorectal carcinoma, that the bacterium is not actually triggering the disease, the discovery could nonetheless save lives. Right now, colon cancer is detected by colonoscopy, a highly invasive procedure in which a long snaking device is inserted into the body cavity. Holt's finding hints at an alternative in which the cancer could be detected from a stool sample: if the patient's stool contains traces of the bacterium associated with gum disease, it would raise a red flag. The approach, if it worked, would be far less invasive and might encourage more people to get tested, said Holt. That could in turn lead to earlier detection and perhaps a higher survival rate.
"Nobody I know is enthusiastic about (colonoscopies)," said Holt. "If you can have some of other way of establishing the presence of a tumour, I think that would be welcomed."
The more enticing possibility, however, is that F. nucleatum is actually causing the cancer, perhaps via inflammation, because it would mean we might ultimately be able to immunize people against it. Much as health authorities have begun immunizing populations against HPV to stymie cervical cancer, so we might be able to protect people against colon cancer - at least reducing their odds of contracting it - by vaccinating them against F. nucleatum. If it worked, the technique could conceivably save hundreds of thousands of lives annually.
That, however, is a long way down the road, Holt noted. First, a causal link would have to be established. Then a vaccine would have to be developed, tested, and approved for use, a process that can take decades.
BACK TO THE LAB
As with so many other aspects of cancer research, proving a given microbe guilty of causing cancer is no easy task. Different teams would likely have to tackle the problem from different angles, said Holt, but one approach, put simply, would be to attempt to give a lab animal colorectal carcinoma by infecting it with F. nucleatum, and then - just to double-check - see if the bacterium could be isolated again from the resulting tumours.
"There is sort of a roadmap, but in reality a lot of those steps are really difficult," he said.
This would not be as easy as it sounds, he said. Bacteria are often highly specialized, and just because a given germ can grow in one species - humans, in this case - doesn't mean it can grow in another - mice, say. So a first step down this road would be to establish whether such a suitable unwitting host exists. There are no guarantees that it does.
It's not a task Holt plans to take on; rather, his next step will likely be to study F. nucleatum further, to take a census of the strains that are out there, and which, if any, are overrepresented in tumours. That's just one possibility, however. For now, he's weighing his options.
"There's not a lot of opportunity to go down blind alleys and recover from that," said Holt. "We're going to be cautious."
