This is the fourth time I have written about single nucleotide polymorphisms (SNPs) in relation to CLL and I do so in response to a paper from Richard Houlston and Daniel Catovsky about to be published in Nature Genetics. (Di Barnardo et al, Nature Genetics published on-line 31st August 2008; doi:10.1038/ng.219). SNPs are an extremely powerful tool to investigate genetic tendencies for diseases and that is what has been done here.
A SNP is where a single nucleotide (ie G, A, C or T) in the genetic code has been swapped from the usual pattern. The best known variation is in sickle cell anemia, where one change causes a major disease, but most changes have no apparent effect. However, many of these so-called silent changes can make a particular protein operate in a very slightly different way, and several such polymorphisms acting in a synergistic way might make a patient prone to a certain type of response. For this to become leukemia, might require the right sort of trigger, or even more than one trigger happening at the same time.
So far we really don’t have much evidence as to what causes CLL, even though we know it does occur in families. Last year workers at Ohio State did discover a polymorphism that occurred in one family which made the DAP kinase enzyme operate in an adverse way, but it clearly wasn’t the cause of other cases of familial CLL and only one sporadic case was found where the same polymorphism occurred.
This is a major study involving hundreds of patients with CLL. The large numbers make it possible to find SNPs that make a small difference. They found 7 different SNPs that did so.
The SNP that was most significantly different from the controls had an odds ratio of 1.59. Although this was very significant because of the very large number of samples, this is not a very large extra risk. People who have a particular one letter switch are one and a half times as likely to have CLL than the general population – the risk changes from 4 per 100,000 to 6 per 100,000. Unraveling this SNP leads us to MUM1, a gene that has shown up before in the related condition, myeloma, and is now known to be the same as IRF4 a gene necessary for the transformation of memory B cells to plasma cells. The presumption is that this SNP doesn't produce a nonsense protein but likely one that affects the smooth running of this process when stressed. So this abnormality is very believable as being involved in the genesis of CLL.
The second most significant difference leads us to the area near GRAMD1B which seems to have an unknown function, so that's not very helpful. The odds ratio for this is 1.45.
The third most significant leads to an area with no known genes. The fourth leads to SP140 which codes for a nuclear body protein present in lymphocytes. We know that interferon increases the number of nuclear bodies and that somehow they are involved in the response to viral infection. Since there is accumulating evidence that infections might be involved in triggering CLL and the CLL may result from an error in coping with infection, a link to SP140 is not out of the question.
The fifth strongest association leads very indirectly through linkage disequilibrium to BIM, a pro-apoptotic protein. Apoptosis is impaired in CLL cells, so, again, a link may well be there. The sixth strongest leads to PRKD2 which has been also linked to CLL.
The picture is emerging of a number of different polymorphisms that could disturb the way that B cells react to stimuli. I am surprised that there are not more.
I am afraid that we are no nearer to understanding why some people are more likely to get CLL than others, but it is clear that it a reaction between heredity and environment.