Lorne 2009 profile: Kenneth Kinzler

The names Kinzler, K.W. and Vogelstein, B. – often joined by Velculescu, V.E. – will be familiar to any cancer geneticist working in the field for the last 25 years.

Drs Kenneth Kinzler and Bert Vogelstein have been enormously prolific authors of outstanding scientific papers for much of that time, beginning with their work on APC mutations in colon cancer, the development of the SAGE methodology (led by Victor Velculescu) in the mid 90s, and their recent work on integrated genetic analyses of colon, breast, pancreatic and brain cancers.

Between 1995 and 2005, Kinzler was the fifth most cited author in clinical medicine in the world. Vogelstein was the fourth.

Kinzler, who spoke at the Lorne Cancer conference last week, has spent his entire career at Johns Hopkins University in Baltimore. “They refused to give me a degree so I’m staying,” he jokes.

He joined as a graduate student in 1983 and there he met Vogelstein, who seemed to share a complementary set of interests and skills with Kinzler that has made them a team of such renown.

Kinzler and Vogelstein are best known for their work in colon cancer, but only entered the field due to some timely funding from a philanthropist named Benjamin Baker Jr, who set up a fund for colon cancer research at Johns Hopkins that enabled Vogelstein to start his own lab.

“Money made Bert glance in that direction, but then we saw all of the advantages of colon cancer as a system to study,” Kinzler says. “It’s frequent – the second leading cause of cancer deaths in the US – it was recognised even then that there were inherited predispositions to it, and you can obtain samples at various stages very easily.”

In the late 80s and early 90s, Kinzler and Vogelstein published extensively on mutations of the adenomatous polyposis coli (APC) gene, which are strongly associated with both inherited and sporadic cases of colon cancer and which they and others first identified.

The team developed methodology for diagnostic testing for the mutation in both familial adenomatous polyposis as well as hereditary nonpolyposis colorectal cancer.

The team expanded their work on colon cancer, investigating not only APC but the tumour suppressor gene p53, and also in other gene families which could be mutated to the same effect, including how amplification of Mdm2 inhibits p53 and how a mutation in beta-catenin mimics the effect of mutated APC.

In 1998, they found a connection between APC and the oncogene c-MYC, which they believe initiated most colon cancers. Even then, Kinzler and co understood that while individual genes were important, so were the many pathways involved.

“The idea that it wasn’t a specific gene that was mutated but that pathways were important was coming about in the late 90s, well before we started sequencing whole families of genes or whole genomes,” he says.

“However, we did think there would be more genes mutated more frequently than we are seeing now. We are surprised that when you look at colon cancer, essentially every gene that is mutated, we have previously found more than a third of them. We didn’t expect that our early studies were that efficient.”

---PB--- Breast and colon cancer maps

In 2006, Kinzler, Vogelstein and Velculescu led a project to map the genetic code of breast and colon cancers. They found about 100 genes in each tumour type that were significantly altered, and suspected that about 20 of these were likely to be crucial in each.

In October 2008, they updated this study with an integrated analysis looking at copy number changes in breast and colon cancers, and found that about 17 genes on average were altered by major copy number changes in each type of tumour.

A month earlier, they published integrated genomic analyses of pancreatic cancer and of glioblastoma multiforme, where they found on average 12 core signalling pathways for each tumour type.

The studies used microarrays to identify copy number variations and sequencing technologies to evaluate gene expression. These data were then integrated using newly developed statistical algorithms, and the ‘driver’ mutations separated from the ‘passengers’.

This is the sort of work that Kinzler says he and Vogelstein have been trying to achieve for their whole careers: to define what goes wrong genetically in a tumour.

“We hoped that we would have found more of what we call mountains – genes that are frequently mutated in cancers – than we did,” he says. “But we really just wanted to understand what the genetic changes are so we could stop thinking about looking for genetic changes and start thinking about how to apply them.

“When we did the colon cancer study, the most frequently mutated genes were APC, p53, K-ras and PIK3CA. In the end we sequenced nearly 20,000 genes, or well over 15,000, and they remained the same genes that we could have found 10 years earlier.

“And we did breast because we thought that wasn’t as well characterised, but we still didn’t find genes that were more frequently mutated than what had been previously described. So that was a little bit surprising.”

While the most frequently mutated genes will continue to be studied in depth, what Kinzler has learned is that it will help understanding if research looks at the effect of the mutations and the pathways involved, rather than the effect of that particular gene. The pathways may be better targets than the individual gene, he says, as the end result is still the same.

“With passenger mutations, which are random changes that are thought not to contribute in a significant way to the phenotype of the tumour – they are going to be different in different patients and different individuals – but what was surprising was, before we did these studies we could point to many genes that were mutated, but what we were surprised to find is that there weren’t more of the drivers being shared.

“We see differences in the drivers between tumours, both the minor drivers and the major drivers varied between tumour types, but a lot more mutations varied even within a tumour type than we thought. There’s more than one way to skin a cat.”

---PB--- Pancreas and brain

Following the breast and colon cancer studies, Kinzler and his colleagues decided to take a look at pancreas and glioblastoma because they tend to be two of the most lethal cancers.

“There’s not much you can do about those for most patients who are identified with these diseases,” he says. “We felt that any progress that we could make in terms of understanding them, any insights into developing new therapies or any insights into early detection methods would be a major improvement on what we have now.”

In the pancreas study, the team found 12 core signalling pathways that were implicated to a greater or lesser extend. This is a lot fewer than the 70 or so genetic alterations that occur in each cancer, he says.

“And there is some overlap between each pathway. But again, a lot of little hills may look to be disparate but if you can group them by function then you can start to see patterns.

“We’re somewhat limited in our understanding of how the over 20,000 genes in our cells interact, but that’s something that we are continuing to expand at an amazing rate. How these pathways interact and what these core pathways will be, will continue to evolve as we learn more and more about these cancers.

“We will also learn more about the normal function of these pathways.”

Kinzler likes to use the analogy of cancer as a continent or a land mass, dotted with hills and mountains. If you survey that continent, you will see the biggest mountains first, and that’s what the team saw in their studies. But the closer you get, the more you realise that while the mountains are easy to find, most of the continent is flat or populated with hills.

Does this make cancer as a landscape all too complex?

“I don’t think so as we can always go back to the pathways,” he says. “Although there are lots of genes that are being mutated, they are producing the same effects. We just have to figure out how they all fit together. Can we distinguish the ones that we think work together from the ones that really work together to produce the neoplastic effect?

“Some people have asked, does this information make it harder to treat cancer? It doesn’t make it harder – it’s just that we now know what we face. The mountains are still there, but knowing what’s going on with the other genes is going to make things a lot easier in the long run.”