Pharma Strategy Blog

This was a stunner via a tweet from the Roswell Park Cancer Institute yesterday.  Of course, I clicked on the link because my suspicion was that it wouldn't be the most common O blood that is linked to pancreatic cancer, and no, I don't have blood group O.  A little voyeuristic, perhaps, but sometimes curiousity literally kills the cat.

What was interesting about the link was that it came from the NCI, a well respected institution in the US, which was reporting on a study that has just been published in Nature Genetics confirming epidemiology studies on stomach and other GI cancers going back to the 1950's.  It's not new?  How come we haven't heard much about this fascinating story then?  That's the trouble with science, a lot of great stuff gets lost in the river of noise in daily life and finding the good snippets can be like searching for needles in a haystack, to mix metaphors.

Previously, we have heard that a high fat diet, alcohol and heavy smoking have all been linked with the disease, but that isn't always the case as Prof Randy Pausch and these incredible patient stories have shown.

The NIH summarised the new study succinctly:

"The researchers discovered that genetic variation in a region of chromosome 9 that contains the gene for ABO blood type was associated with pancreatic cancer risk. Individuals with the variant that results in blood types A, B, or AB were at an increased risk of pancreatic cancer, compared to those with the variant for blood type O. This finding is consistent with previous research, some of it dating back to the 1950s and 1960s, that had shown increased risks of gastric and pancreatic cancer among individuals of the A and B blood groups (i.e., blood types A, B, and AB). The latest results provide a genetic basis for those earlier observations."

So what does that mean from a scientific basis?  The NIH continued:

"A person's blood type depends on which form or forms of the ABO gene they inherit from their parents. The protein produced by the ABO gene determines the type of carbohydrates (complex sugars) that are present on the surface of red blood cells and other cells, including cells of the pancreas. The proteins encoded by the A and B forms of the gene transfer different carbohydrates onto the cell surfaces to make A and B blood types. The O form encodes a protein that is unable to transfer carbohydrates. Studies by other researchers have shown that ABO protein encoding in pancreatic tumor cells is different than in normal pancreatic cells."

Most pancreatic cancers are diagnosed late, ie in stage IV, meaning that the prognosis and long term survival is poor but these new findings may help enable earlier detection of the disease.  The researchers performed arrived at their conclusions by conducting a genome-wide association study (GWAS). In a GWAS, researchers analyze common variants, called single-nucleotide polymorphisms (SNPs), in the genomes of people with disease and in a control group (people without the disease).

The team genotyped 558,542 SNPs in 1,896 individuals with pancreatic cancer and 1,939 controls drawn from 12 prospective cohorts plus one hospital-based case-control study.  They also conducted a combined analysis of these groups plus an additional 2,457 affected individuals and 2,654 controls from eight case-control studies, adjusting for study, sex, ancestry and five principal components.

The end result?

They were able to identify several SNPs on the long arm of chromosome
9 that were associated with pancreatic cancer risk and mapped to the
ABO gene.

Still, the big questions in my mind are:

a) Why do people with blood group O appear to have a lower risk? What protective effect is at play?

b) How can we use the knowledge of higher risk factors in blood groups A, B and AB to screen and diagnose pancreatic patients earlier?  Would the spittoon type tests offered by commercial genetic testing companies such as 23andme help with this at all?

It's all very well science finding new relationships and genetic associations with disease, but it would be very sad indeed if we went another 50 years with no improvement in early detection and mortality associated with the fatal disease.

Amundadottir, L., Kraft, P., Stolzenberg-Solomon, R., Fuchs, C., Petersen, G., Arslan, A., Bueno-de-Mesquita, H., Gross, M., Helzlsouer, K., Jacobs, E., LaCroix, A., Zheng, W., Albanes, D., Bamlet, W., Berg, C., Berrino, F., Bingham, S., Buring, J., Bracci, P., Canzian, F., Clavel-Chapelon, F., Clipp, S., Cotterchio, M., de Andrade, M., Duell, E., Fox Jr, J., Gallinger, S., Gaziano, J., Giovannucci, E., Goggins, M., González, C., Hallmans, G., Hankinson, S., Hassan, M., Holly, E., Hunter, D., Hutchinson, A., Jackson, R., Jacobs, K., Jenab, M., Kaaks, R., Klein, A., Kooperberg, C., Kurtz, R., Li, D., Lynch, S., Mandelson, M., McWilliams, R., Mendelsohn, J., Michaud, D., Olson, S., Overvad, K., Patel, A., Peeters, P., Rajkovic, A., Riboli, E., Risch, H., Shu, X., Thomas, G., Tobias, G., Trichopoulos, D., Van Den Eeden, S., Virtamo, J., Wactawski-Wende, J., Wolpin, B., Yu, H., Yu, K., Zeleniuch-Jacquotte, A., Chanock, S., Hartge, P., & Hoover, R. (2009). Genome-wide association study identifies variants in the ABO locus associated with susceptibility to pancreatic cancer Nature Genetics DOI: 10.1038/ng.429