Synaptic elimination and the complement system in Alzheimer’s disease
30
present study. However, the correlations of complement factors with age
were not straight-forward. When examining correlations in the four
subgroups separately, the correlations remained in controls and stable MCI
patients, but disappeared in the AD subgroups, in spite of a large age span in
those groups as well (Table 1, paper III).
As expected, T-tau and P-tau levels were correlated in both groups. T-tau and
P-tau were also correlated with age, Aβ42 and CR1 in the non-demented
subjects. In AD patients, on the other hand, we observed a weak correlation
between MMSE and Aβ42, MMSE and C4.
In conclusion, the results of this study speak for complement involvement in
the disease process occurring in patients with AD. None of the complement
proteins C3, C4 and CR1 are, however, suitable as CSF biomarkers for
clinical use.
4.4
Complement gene SNPs in AD
Clearly the genes of the complement system are interesting as candidate
susceptibility genes for AD. CR1 and CFH have been implicated by previous
studies (Harold et al. 2009; Lambert et al. 2009; Naj et al. 2011; Seshadri et
al. 2010; Zetterberg et al. 2008; Ferrari et al. 2012; Hu et al. 2011). For the
present gene association study, we selected genes that had previously been
found associated with the neurodegenerative disease age-related macular
degeneration (AMD) which in turn is epidemiologically linked to AD
(Kaarniranta et al. 2011; Kirby et al. 2010). These genes were: C2, C3, and
the gene encoding complement factor B (CFB). We also included CR1, since
this gene has been associated with AD in several genome wide association
studies (Harold et al. 2009; Lambert et al. 2009; Naj et al. 2011; Seshadri et
al. 2010; Ferrari et al. 2012; Hu et al. 2011).
No significant associations between the investigated SNPs and risk of AD
were detected in our material (Table 3, paper IV). C2/CFB, (these two genes
are tightly linked, and therefore inherited together) however, was associated
with MMSE and T-tau in AD subjects; since the ORs are in the same
direction as in previous studies on AMD (Sun et al. 2012) and the overlaps of
the confidence intervals are substantial, it is quite possible that we failed to
detect an actual association in this study. The same rational goes for the CR1
replication. Meta-analysis of the SNP investigated here ends up with an
OR=1.17 (www.alzgene.org) and is well within our estimated CI=(0.84 –
1.37), thus our result supports rather than refutes the previously found
associations.
Jonny Daborg
31
In conclusion, the study failed to detect any associations of the selected
complement SNPs and AD diagnosis. An association of the C2/CFB SNPs
with T-tau levels and MMSE was found. These two parameters reflect the
seriousness of the symptoms, and the association lends some support to the
hypothesis of aberrant complement regulation in AD.
Synaptic elimination and the complement system in Alzheimer’s disease
32
5
DISCUSSION
In the present thesis I have explored some predictions of the hypothesis that
AD is primarily a synaptic disorder. I have shown that a variant of the gene
encoding the RAGE protein, which has been linked to synaptic dysfunction
in animal models of AD, is associated with AD diagnosis. This study was
followed by animal experiments in which the mechanisms of synaptic
elimination in the hippocampus were investigated. Consistent with the results
from previous studies in other brain areas, we found that synapses were
eliminated in a complement mediated manner. This finding led us to explore
certain neurochemical and genetic aspects of the complement system in
patients with AD. The investigations showed a trend towards increased
complement levels in AD CSF, and an association of C2/CFB variants with
MMSE and tau levels in AD patients.
In the following section I will present a general discussion
the
pathogenesis and pathophysiology of AD. I will present a possible sequence
of events that lead to AD, and discuss the results presented in my thesis in
this hypothetical context. Finally, I will give my view on the ultimate reasons
as to why people develop AD.
5.1
The events leading to AD – a proposed
model
The common conception is that in AD, Aβ kills neurons. This is in principal
correct, but several crucial steps in the process have been overlooked.
Without doubt Aβ as well as tau are involved in AD. However, it is the
number of synapses that best predict the severity of the dementia (DeKosky
and Scheff 1990; Scheff and Price 2003; Terry et al. 1991). This comes as no
surprise since synapses constitute the physical locus of memory storage.
Lastly, AD is associated with a loss of neurons; this is, however, occurring in
the later stages of the disease.
5.1.1
It all starts with A
β
In FAD, the disease causing mutations all affect Aβ processing, either by
favouring the more toxic 42 amino acid long variant, or by simply promoting
production, with increased amounts as a consequence. In SAD, however, no
such linear relationship exists. A multitude of risk factors, foremost genetic
ones, are thought to cause the disease. Aside from ageing, APOE ε4 is the
most prominent risk factor, and is known to be involved in Aβ clearance.