Discussion
This study leveraged genetic data to identify support for ACE in
prevention of AD, with no strong evidence identified supporting effects
of ACE on other neurodegenerative traits. While increased corticalACE expression associated with lower AD risk, there was no MR
evidence supporting that genetically predicted SBP affects risk of AD.
From a mechanistic perspective, ACE has been shown to breakdown
neurotoxic amyloid-beta isoform (Aβ42) to a less toxic form (Aβ40).
Administration of a clinical dose of ACE inhibitor to human amyloid
precursor protein transgenic mice was associated with increased brain
amyloid deposition4. In humans, patients with AD have
lower Aβ42-to-Aβ40–converting activity compared with sera from
age-matched healthy individuals4. Our current findings
support that ACE protects against AD, although further work is required
to investigate whether this is attributable to reduced amyloid
aggregation or other unrelated mechanisms.
An observational study among 406 participants with mild-to-moderate AD
demonstrated a reduction in cognitive decline for people receiving a
centrally-acting ACE inhibitor (perindopril) compared to
peripherally-acting ACE inhibitor18. Other studies
have shown increased risk of incident dementia and disability associated
with peripherally-acting ACE inhibitors compared to other
anti-hypertensive medication19. Conflicting findings
between genetic and observational studies could be explained by MR being
less liable to environmental confounding and reverse
causality20, due to the random allocation of genetic
variants at conception.
Our current work is consistent with other genetic studies supporting a
role of ACE in preventing AD and has several additional strengths.
Firstly, obtaining association estimates from the MetaBrain consortium
(n = 6,601 participants)10, we investigate corticalACE expression and AD risk. This dataset is significantly larger
than the GTEx resource (n = 205) that has been utilised in previous
work7. Secondly, we investigate whether SBP mediates
the relationship between ACE and AD risk, and do not find evidence that
supports this. Finally, we explore the associations of genetically
proxied cortical ACE expression with other traits and do not find
evidence to support that this association applies across other
neurodegenerative traits.
This work also has several limitations. Clinical diagnosis of AD is
challenging as there is significant overlap in symptoms with other forms
of dementia, limiting the specificity of case definitions in GWAS. To
explore for this, we also assessed several other neurocognitive traits.
Given the absence of strong evidence of ACE effects for these outcomes,
it seems likely that our findings are specific for AD risk, rather than
a generic effect on dementia or cognition. As with all studies
leveraging genetic data, there remains the possibility of biological
pleiotropy introducing confounding. It is also not possible to
extrapolate the magnitude of clinical effect or required drug exposure
for ACE inhibitors to represent a real-world risk for AD. Factors such
as the ability of an ACE inhibitor to cross the blood-brain-barrier may
also shape AD risk and should be further studied. Furthermore, our work
was based on data obtained from individuals of European genetic
ancestry, and it is unclear whether these findings extend to other
ethnic groups.
In summary, while ACE inhibitors have numerous indications and are the
cornerstone of hypertension, chronic kidney disease and heart failure
management, this study finds evidence for a beneficial effect of
cerebral cortex ACE in preventing AD. It would be premature to alter
current clinical practice based on this evidence, and rather these
findings should encourage further research into the effect of ACE
inhibitors on AD risk.