Response: associations between activating KIR genes and childhood leukemia

We are surprised to know that Babor et al did not detect any significant differences in the frequencies of activating KIR genes between acute lymphoblastic leukemia (ALL) patients and healthy controls.¹  Obviously, these results contradict those reported recently by us.²  Interestingly, the frequencies of these genes in controls are more or less comparable in both studies. Babor and colleagues mentioned technical challenges associated with PCR-based KIR genotyping due to strong similarities between KIR genes and their many alleles. They further note that because of the archived nature of patient samples, the quality of the DNA may be compromised and hence it may be difficult to amplify KIR genes. Therefore, the authors imply that this could be one reason for decreased KIR gene frequencies in our patient samples. We have experienced such difficulties in amplifying KIR genes in both patient and control samples with longer amplifications. Such difficulties were alleviated when smaller segments of the genes were amplified. All of our PCR reactions for the KIR genes amplified ≤ 300 bp bands. Furthermore, we have used the same genotyping methods to determine gene frequencies of activating KIR genes in other diseases (eg, Crohn disease) and have found increased frequencies of several of the activating KIR genes in the patients compared with the controls (data not shown). We receive samples from both patients and controls for these studies from DNA banks. Furthermore, all of the patient and control DNA samples yielded bands in positive control reactions. Therefore, we do not believe that the decreased frequencies of activating KIR genes in leukemia patients are due to inherent unsuitability of our DNA samples for PCR-based amplifications. After receiving an invitation to respond to the letter by Babor et al, we regenotyped a subset of our patients and controls. Again, we found significantly decreased frequencies of all activating KIR genes in our cohorts. It is noteworthy that decreased frequencies of some activating KIR genes in ALL patients have been described.^3,4 

The reasons for the discordant results in the 2 studies are not immediately apparent. However, we speculate that these differences may result from the use of different PCR primer sets. We would encourage Babor and colleagues to verify their results with our primer pairs. We can provide them or send them their exact sequences. We would also be willing to genotype our DNA samples (a subset or all) in our laboratory. Furthermore, it could also be possible that these associations may not be present in the German population. Literature is replete with examples where genetic associations with a disease in one population were not replicated in another.

Concerning inhibitory KIR genes in our ALL patients, we are in the process of genotyping them. It is premature to draw definitive conclusions. However, we can say with certainty that they are showing a trend opposite to that of activating genes. This trend also verifies the suitability of our DNA samples for PCR-based genotyping. Concerning KIR haplotypes, our results do suggest that B haplotypes reduce risk for ALL. However, we have not directly determined them in our patient and control samples.