Two Independent Effects of Immunoglobulin-Like Receptor (KIR) Allele Matching between Siblings: Inhibitory KIR (iKIR) Mismatches Are Associated with Graft Versus Host Disease (GVHD) While Activatory KIR Matches (aKIR) and cGVHD Are Associated with Graft Versus Leukemia (GVL).

Background: Activatory KIR receptors are observed less frequently than inhibitory KIRs (iKIR). Although the role of donor iKIRs and recipient ligands have been analyzed extensively, the effect of mismatches(mm) at aKIR or iKIR between donors(D) and recipients (R) have been addressed in only two studies (Gagne, Hum Immunol.2002 and Verheyden, Leukemia 2005 ): an aGVHD inducing effect of aKIR mm between unrelated donors and a protective effect of aKIRs: 2DS1 and 2DS2 against relapse were reported. The evaluation of other factors ie GVHD on GVL (related transplants) or GVL effects(MUD study) were lacking in these studies.

Aim: In this prospective study we aimed to analyze the role of both D and R iKIR, aKIR and KIR-ligand match/mismatches on OS and DFS and made a multivariate comparison of all factors effecting outcome.

Methods: A total of 79 patients with a median age of 34 (M/F: 42/37, AML/CML: 37/33, PBSCT/BMT: 59/20, sex mm: 49 %, ablative/nonablative conditioning: 63/16, BuCy: 72 %) transplanted from their HLA matched siblings. All D and R were typed for KIR genes (2DL1, 2DL2, 2DL3, 2DL4, 2DL5a, 2DL5b, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DL1, 3DL2, 3DL2, 3DS1) using the KIR Genotyping SSP Kit (Pel-Freeze, Dynal Biotech, USA). The frequency of GVHD was acute:44/76, chronic: 54/74. Statistical analysis were done using the SPSS 13.0 for Windows. The frequency of relapse in relation to m/mm at iKIR or aKIR alleles are summarized in table 1.

Results: Overall KIR mm was observed in 75% of R-D pairs. 37 pairs had mm at the six iKIR loci (33% 2DL5a), 57 pairs had mm at the seven aKIR loci. The analysis on D iKIR/aKIR and the relevant R-ligand m/mm didnot reveal any effect on the frequency of GVHD or DFS. However when we compared the R and D KIR genotypes we were able to show a correlation between cGVHD and m vs mm at iKIR (62.5% vs 85%, p=0.041) and aGVHD ( 50% vs 67%) but not aKIR. The effect of aKIR mm was not influenced by stem cell source or diagnosis. Among the aKIR only 2DS5 and 3DS1, alone (20/30, 17/25 ) or together(16/21), resulted with more frequent aGVHD than pairs with other aKIRs (23/45). GVHD was associated with a decrease (aGVHD: p=0.032) or increase (cGVHD: p=0.076) in survival. cGVHD resulted with a decrease in relapse rate(−): 11/20 vs (+): 9/54: p=0.001). aKIRm was associated with an increase on DFS (p:0,031). Factors, other than KIR, known to influence outcome ie stem cell source, sex mismatch, conditioning regimen, disease type were analyzed in the multiple logistic regression and did not reveal any significant results.

Conclusion: This study material enabled us to minimize the effect of HLA but PBSCT being the major source of stem cells potentiated the role of alloreactive T cells and cGVHD. Although only two of the donor aKIRs, 3DS1 and/or 2DS5 were associated with aGVHD, D-R match between all aKIR and cGVHD exerted a protective effect against relapse(0/15). Mismatching for iKIR was also associated with GVHD but GvL wasnot independent of GVHD. Thus, we may conclude that D-R aKIR, iKIR genotyping may help to predict GvL in related transplants.