Our study revealed that shorter HO-1 promoter genotype has the protective effect on ischemic stroke, especially in the patients with lower HDL-C levels. The effects and interactions were demonstrated in univariate and multivariate models for all subjects, and the protective gene effect appeared in the stratified high risk group, not in the low risk group. The significant increases of genotype L than S indicated the protective genetic effects of shorter HO-1 gene repeats seem work up well in high risk group like low HDL-C level, and the risk was decreased to the level near the low risk group like appropriate HDL-C level. All of patients in our study had no prior CAD or stroke, so our observation was not confounded by selection bias, such as survivor for severe disease. It is less considered in previous studies.
Interestingly, we also described more risk effects of genotype L on stroke in those with normal TC or LDL-C levels than those of genotype S, although the effect was no longer present in multivariate analysis. Before our study, one study reported shorter repeats in HO-1 gene exerted a protective effect on the development of ischemic cerebrovascular events definitely in patients without hypercholesterolemia . The finding was similar with that in our study, but they did not include HDL-C measure. In summary of our findings, the protective effect of HO-1 genotype on ischemic stroke depended on the presence of lipid conditions, that is, the levels of HDL-C, it may explain the controversial findings in the literatures.
Similar as the previous CAD studies, we did not find the significant difference of averaged (GT)n repeats in HO-1 gene promoter between IS and NS patients. Instead, the lengths of (GT)n repeat seem associated with the ischemic stroke status only in those individuals with lowered HDL-C level from our observation and some previous studies. Chen and his coauthors have shown a protection of HO-1 genotype in diabetic CAD patients with at least 1 coronary narrowing (>75%) and in restenosis patients after coronary stenting [24, 35]. Another study has shown that the longer (GT)n repeat in HO-1 gene promoter was related to CAD risk in diabetic and hypercholesterolemic patients . Interestingly, in the other study, HO-1 gene exerted a protective effect on ischemic cerebrovascular events in patients with normal cholesterol level . No effect of HO-1 genotype could be observed in total population, they only have been observed in high risk group or low risk group. It implies that long (GT)n repeats alone may not be sufficient to cause diseases. But it may contribute to the development of the disease when certain conditions of enhanced oxidative stress coexist, including the abnormal composition of cholesterol levels in patients with normal total cholesterol levels.
Many studies reported that HO-1 gene involved in the mechanism against the development of atherosclerosis. Animal studies reported that products of HO pathway such as bilirubin act as a significant protective factor for atherosclerosis. The modulation of HO-1 gene expression in LDL-C receptor-deficient mice influence the progression of atherosclerosis , and mice treated with the HO-1 inducer exhibited reduced atherosclerotic lesion formation. Vascular proliferation was inhibited by transferring HO-1 gene. These observations support that HO-1 functions as an intrinsic protective factor against atherosclerotic lesion formation and may be an anti-atherogenic role in vascular wall .
Chen et al conducted a transient transfection experiment in rat aortic smooth muscle cells to show that longer (GT)n repeats in HO-1 promoter decreased luciferase promoter activity, indicating decrease in gene transcription in vascular cells. They also found that genotype L/L carriers are associated with higher extent of sever lipid peroxidation, supporting the genetic influence of HO-1 on oxidative stress. Lipid abnormalities like hypercholesterolemia correlate with enhanced oxidative stress. HDL-C acts as an anti-oxidant through its capability of inhibiting LDL-C oxidation, preventing the formation of lipid hydroperoxides [36, 37]. Excessive oxidative stress was considered as a potential cause of the vascular disease and other complications in hyperglycemic patients [24, 38]. S genotype in the HO-1 gene promoter may increase the induction of HO-1 by reactive oxygen species in patients with low HDL-C concentrations. The insufficient effect of anti-oxidative stress due to lower HDL-C levels may be reversed by S genotype in HO-1 gene promoter, thereby reducing the risk of cerebral ischemia.
In this study, the allelic frequency distribution of the lengths of (GT)n repeats in the HO-1 promoter in recruited subjects (range from 15 to 39) was similar with that in the previous reports[23–25, 34, 35]. The previous studies defined L and S alleles first and constructed genotypes SS, SL, and LL to examine the disease risk. We demonstrated the results assuming co-dominant (additive) model and using averaged length of two alleles to define genotype L and S, since the latter is more powerful and fits well the characteristics of complex model. We also obtained consistent results using the traditional (former) classification method: the age, sex & plaque score-adjusted OR of 2.26 (p = 0.0263) and multivariate OR of 1.82 (p = 0.0691) in those with lower HDL-C level while those carriers with homozygous S allele (≦26 GT repeats) as genotype SS compared with otherwise. The borderline significant findings on HO-1 genotype and cerebral ischemia were found in those with low HDL-C level or in those with high TC/HDL-C ratio. Therefore, the role of HDL-C associated with developing ischemic stroke exists identically but underestimated. Such finding focused the additive effect implied the equal importance in double helix, then decreasing number of GT repeats act additively with the increasing protective effects, as some previous reports described[39, 40].
The limitations of this study should be mentioned. Our study is a case control study. The controls were recruited from outpatients of the same hospital who seemed to have higher levels of several CVD risk factors than general population. On the other hand, although we only included the first case without the history of CAD or stroke to reduce the selection bias, however, we still not avoid the loss in the stroke patients died before admission. Fortunately, the number is few. Taken together, the effect of HO-1 genotype on ischemic stroke may have been underestimated.