The TNF-a Gene Modifies Stroke Risk in Children with Sickle Cell Anemia: Results from the STOP Trial.

Background: The etiology of stroke in sickle cell anemia (SCA) is complex and multiple genetic factors outside of the sickle gene undoubtedly contribute to stroke risk in SCA. Simultaneous analyses of multiple genetic variants may therefore be more informative than evaluation of individual candidate genes. Moreover, stroke in SCA is phenotypically heterogeneous, varying from large cortical vessel (LV) distribution infarcts to milder lacunar infarcts due to small vessel disease, suggesting distinct pathogenetic mechanisms for stroke in SCA.

We previously reported risk associations with particular TNFA, IL4R and ADRB2 polymorphisms in children enrolled in the CSSCD.

Methods: To confirm our previous findings from the CSSCD and identify other potential “susceptibility” loci that may influence the development of LV stroke, we extended our investigations to SCA children who participated in the Stroke Prevention Trial in Sickle Cell Anemia (STOP). Children with MRA evidence of obstruction or stenosis of one or more large vessels (LV disease) were compared with untransfused children with a normal MRA. Ninety-seven patients were included in this study. Of the 119 patients originally enrolled in STOP, 77 (30 MRA+, 47 MRA−) were eligible for this study and were analyzed together with a local sample of 20 MRA(+)patients. Genotyping for 110 polymorphisms among 65 candidate cardiovascular and proinflammatory genes was performed using multiplex linear array methods. Of the 110 variant sites examined, 95 were sufficiently polymorphic to permit statistical testing, using 2 x 2 exact tests.

Results: When comparing MRA(+) patients with MRA(−) patients, 9 of the polymorphic loci revealed nominally significant differences (P<0.10): TNF(−308), IL4(−589), ADRB2 16, ICAM1 56, APOC3(−482), LPL 93, AGT 235, and MMP3(−1171). For TNF G(−308)A, the frequency of the G allele was 80%, consistent with allele frequency data in African-Americans. In the current study, there was a relative excess of children with the GG genotype in the MRA(+) group compared to the MRA(−) group (OR= 2.48; p = 0.037). These results, showing the TNF(−308) locus to be predictive of LV vasculopathy, are virtually identical to our previous findings in the CSSCD population and confirm that the TNF gene plays a significant role in stroke risk. The TNF gene lies within the HLA region, making this finding particularly intriguing in light of our previous data indicating HLA associations with stroke risk. Detailed studies of this region are now necessary to determine if it is indeed the TNF locus or other genes in linkage disequilibrium that are responsible for this association.