Malignant hypertension and the angiotensinogen gene

Malignant hypertension is one of the most severe consequences of high blood pressure. The disease is rare in comparison to the prevalence of high blood pressure but continues to challenge clinicians, and to puzzle researchers. In this issue of the Journal of Hypertension, van den Born and coworkers [1] present results of an association of malignant hypertension with the M235T polymorphism of the angiotensinogen gene (AGT). Among 38 white patients with malignant hypertension identified between 1995 and 2005 in an academic referral centre in Amsterdam, approximately two-thirds were homozygous carriers of the T allele. This astonishing frequency of TT homozygotes contrasted with the findings in two control populations of normotensive and hypertensive subjects: in both populations, between 10 and 20% of white participants were homozygous for the T allele, a frequency that is in agreement with numerous studies in similar populations. At first glance, the report by van den Born et al.[1] may appear as yet another underpowered association study presenting a potentially spurious finding [2,3]. While such criticisms cannot be dismissed easily, we believe that the results nonetheless deserve a closer look. The number of cases is small for a genetic association study but there is a large difference in the frequency of TT homozygotes between cases and controls. Assembling a larger cohort of patients with malignant hypertension from a single center will not be easy, and many published cohorts are of similar size [4]. Others have studied somewhat larger cohorts of patients with ‘hypertensive crisis’ [5], a diagnosis that is more frequently made but is not characterized by the rapid development of target organ damage seen in malignant hypertension. Lip and coworkers [6] retrospectively identified 315 patients with malignant hypertension at a single referral center over a 30-year period (including many years when far fewer treatment options used to be available for patients with hypertension). The problem of sample size is further aggravated by the need for a separate analysis of participants with African versus non-African descent: the frequency of the T allele is far higher in Africans than in non-Africans [7] which precludes an analysis of the pooled sample. Cohorts of patients with malignant hypertension, however, usually comprise a substantial proportion of patients with African descent [6], and the study by van den Born et al. is no exception: almost half the patients with malignant hypertension were of West African descent. As expected, the frequency of the T allele was more than 90% in African subjects, regardless of whether hypertension was present or not. Consequently, a meaningful analysis of the M235T polymorphism was possible only in subjects of non-African descent, further limiting the size of the cohort. Based on this observation, one can understand that the authors are tempted to speculate that ‘the frequency of the 235T allele observed in populations of West African descent may help explain the higher frequency of malignant hypertension in this group’ [1]. It is important, however, to realize that this is really nothing but speculation: the results cannot be interpreted to support this notion. The study design does not permit to discern the role of the M235T polymorphism relative to other genetic or environmental factors. On the other hand, the multi-ethnic cohort is an important limitation of the study. It is not clear how precisely the descent of the patients was evaluated. Even minor uncertainties bear the potential to affect the results, given the large differences in the frequency of the T-allele. The effect of the AGT genotype was large in the study by van den Born et al.[1], and there is a plausible link between the M235T polymorphism and malignant hypertension. Many experimental studies have shown that the renin–angiotensin system plays an essential role in animal models of malignant hypertension [8,9]. Since the report by Jeunemaitre et al.[10], numerous studies reported an association of the 235T allele with elevated plasma angiotensinogen levels and hypertension. Reviewing these studies is beyond the scope of this comment but it should be noted that negative findings have been published as well [11]. Whether these controversial reports indicate a different role of the polymorphism in different settings (including genetic ‘background’ and environmental factors), or a lack of confirmation of spurious findings, remains open to debate. Nevertheless, much effort has been devoted to the investigation of the potential molecular mechanism of the association of the M235T allele with plasma angiotensinogen levels. The amino acid exchange (threonine instead of methionine at position 235 of angiotensinogen) caused by the M235T polymorphism appears to be of little consequence but the polymorphism is in linkage disequilibrium with other single nucleotide polymorphisms of the AGT promoter, which can affect the transcription rate [12,13]. The first report by Inoue et al.[13] focused on a G/A polymorphism at position -6 whereas more recent studies emphasized the role of single nucleotide polymorphisms at position -217 and -20 of the AGT promoter [12]. Unfortunately, van den Born et al.[1] did not attempt to determine these polymorphisms, AGT promoter haplotypes, or angiotensinogen plasma levels. By means of gene duplication by homologous recombination, Kim, Smithies and coworkers created a mouse model of Agt gene titration [14,15]. Depending on the number of Agt copies, these mice exhibit a similar degree of elevation of plasma angiotensinogen levels as do patients homozygous for the 235 T allele [14]. The role of slightly higher angiotensinogen levels for the regulation of blood pressure was supported by the observation that blood pressure rose with the number of Agt copies in these mice [14]. We recently confirmed these findings [16], using the same strains of mice generated by those authors. Moreover, we observed that the number of Agt copies had no effect in high-renin renovascular hypertension but exerted a marked effect in low-renin mineralocorticoid hypertension [16]. We did not observe malignant hypertension in this animal model but our findings emphasize that the effects of the Agt genotype may depend on the pathogenesis of hypertension. The rapid development of vascular damage in malignant hypertension bears similarities to the vascular pathology in pregnancy-induced hypertension. In that regard, it is of some interest to recall the evidence for a role of the AGT gene in the latter disease. A rare activating mutation of AGT was described as a cause of pre-eclampsia [17]. The 235T allele was also reported to increase the risk of pre-eclampsia in several populations [18] although this finding was not confirmed by some authors [19]. Interestingly, the 235T allele is expressed to a higher level in decidual spiral arteries than the 235M allele [20]. Tissue samples are rarely obtained from patients with malignant hypertension, but the findings from decidual arteries do support the notion that the 235T allele (or, more precisely, AGT promoter polymorphisms linked with the 235T variant) may be associated with an elevated expression of angiotensinogen in vascular tissue. One should not forget, however, that the potential role of the M235T variant for malignant hypertension may make a plausible story but that plausible stories are not necessarily true. Stefansson et al.[4] previously reported a lack of association between malignant hypertension and the M235T polymorphism in a cohort of similar size. Instead, these authors did report an association with the angiotensin-converting enzyme insertion/deletion polymorphism, a finding which in turn was not confirmed by the present study of van den Born et al.[1]. Making sense of the many controversial findings of association studies will remain a difficult task, especially with regard to relatively rare conditions, which do not readily permit studies on a larger scale. The data from van den Born et al.[1] point to the potential role of AGT gene variants for malignant hypertension. Experimental approaches beyond association studies, however, will be required to obtain a clearer understanding of the importance of angiotensinogen for hypertensive target organ damage and malignant hypertension.