Papassotiropoulos et al. have described the association of an intronic polymorphism in the CYP46 gene with AD. They found this polymorphism to influence Aβ load, however, the authors failed to show an effect on cholesterol or 24S-hydroxycholesterol CSF levels, which may be important in this respect.

This study is a continuation of the work Andreas Papassotiropoulos performed in our team in Bonn. In 1999, we showed that plasma 24S-hydroxycholesterol (cerebrosterol) is increased in Alzheimer's disease and vascular dementia (Lutjohann et al., 1999), and that this metabolite acts as a neurotoxin on SH-SY5Y cells (Kolsch et al., 1999). Later, we found that neurotoxicity of 24S-hydroxycholesterol leading to apoptosis is mediated by the generation of free radicals (Kolsch et al., 2001), and reported that plasma 24S-hydroxycholesterol is a peripheral indicator of neuronal...  Read more

Papassotiropoulos et al. have described the association of an intronic polymorphism in the CYP46 gene with AD. They found this polymorphism to influence Aβ load, however, the authors failed to show an effect on cholesterol or 24S-hydroxycholesterol CSF levels, which may be important in this respect.

This study is a continuation of the work Andreas Papassotiropoulos performed in our team in Bonn. In 1999, we showed that plasma 24S-hydroxycholesterol (cerebrosterol) is increased in Alzheimer's disease and vascular dementia (Lutjohann et al., 1999), and that this metabolite acts as a neurotoxin on SH-SY5Y cells (Kolsch et al., 1999). Later, we found that neurotoxicity of 24S-hydroxycholesterol leading to apoptosis is mediated by the generation of free radicals (Kolsch et al., 2001), and reported that plasma 24S-hydroxycholesterol is a peripheral indicator of neuronal degeneration (Papassotiropoulos et al., 2000). Our observation of increased 24S-hydroxycholesterol in cerebrospinal fluid in early stages of dementia was published recently (Papassotiropoulos et al., 2002). In a continuation of this work, we found that another polymorphism in the CYP46 gene (IVS3+43Cy´T) was associated with an increased CSF ratio of 24S-hydroxycholesterol/cholesterol in carriers of the C-allele, as well as with a higher risk of AD (Kolsch et al., 2002). Supporting the observation of Papassotiropoulos et al. that genetic variants of the CYP46 gene influence the Aβ levels as well as the risk of AD, we also observed increased Aβ levels in homocygote carriers of the C allele (Kölsch et al., unpublished). In contrast to the present paper of Papassotiropoulos et al., but in agreement with Desai et al., 2002, we could not find an association of the rs755814 polymorphism described by Papassotiropoulos et al. with AD. It may be mentioned that this polmorphism is currently denoted as a Gy´A polymorphism, but was recently named a Cy´T transition in reverse antisense sequence of the CYP46 gene.

In contrast to Desai et al., both studies (Kolsch et al., 2002; Papassotiropoulos et al., 2003) support that Cyp46 polymorphism might influence the risk of AD by increasing β-amyloid levels. However, the relative relevance of the different polymorphisms in the CYP46 gene is not yet clear. It might even be speculated that a third, nearby polymorphism in linkage disequilibrium is the relevant actor. In contrast to the results of Papassotiropoulos, our study additionally indicates an important relevance of cholesterol metabolism, including the intracerebral metabolism of 24S-OH-cholesterol in AD. The relevance of cholesterol in AD is currently a major focus of theoretical and therapeutic studies.

In summary, we strongly support Benjamin Wolozin's conclusions (see editorial accompanying Papassotiropoulos et al., 2003) that cholesterol metabolism is of major relevance in AD and that CYP46 polymorphisms seem to play an important role. The results concerning CYP46 should be proven by additional association, as well as functional, studies. Until these studies are performed, the present data should be discussed carefully.

View all comments by Heike Kölsch

In our paper last year (Desai et al., 2002), we did not see any association of this intronic polymorphism with late-onset AD in our case-control cohort. One of the obvious differences between our study and this one is that our American White sample size was much larger (434 AD cases and 401 controls) than the one used in this study (201 AD cases and 248 controls). Since the question of power is a common concern in association studies, it is essential that such studies be performed on a large case-control sample having sufficient statistical power to avoid any false association. In our paper, we also examined the association of this polymorphism in a small sample of African Americans available in our center (61 cases and 54 controls). Although the distribution of this polymorphism was significantly different between white and blacks, no association was observed with AD in the black sample, either.

Our data, based on a relatively large sample size, suggest that this intronic CYP46 polymorphism is not...  Read more

In our paper last year (Desai et al., 2002), we did not see any association of this intronic polymorphism with late-onset AD in our case-control cohort. One of the obvious differences between our study and this one is that our American White sample size was much larger (434 AD cases and 401 controls) than the one used in this study (201 AD cases and 248 controls). Since the question of power is a common concern in association studies, it is essential that such studies be performed on a large case-control sample having sufficient statistical power to avoid any false association. In our paper, we also examined the association of this polymorphism in a small sample of African Americans available in our center (61 cases and 54 controls). Although the distribution of this polymorphism was significantly different between white and blacks, no association was observed with AD in the black sample, either.

Our data, based on a relatively large sample size, suggest that this intronic CYP46 polymorphism is not associated with the risk of AD. We cannot, however, rule out the possibility that other sequence variations in this gene may be associated with AD, but this awaits additional studies on relatively large case-control cohorts.

View all comments by M. Ilyas Kamboh