Jonny Daborg
35
trophic factors to the cell (Hardingham and Bading 2010). Moreover, LTD is
dependent on the apoptosis protein caspase-3 (Li et al. 2010b). It is
conceivable that excessive LTD in the absence of LTP drives the neuron to
apoptosis. In this respect it is interesting to note that overexpression of NP1
increases the levels of activated caspase-3, the number of apoptotic cells in
neuronal cultures (Abad et al. 2006) and that caspase-3 is increased in
synaptosomes from AD brains compared to controls (Louneva et al. 2008). In
addition, NP1 knockdown rescues these cells from Aβ induced toxicity
(Abad et al. 2006). Moreover, post-mortem analyses of AD brains have
shown that NP1 is associated with senile plaques and the synaptic protein
SNAP-25 (Abad et al. 2006).
5.1.6
Summary of the model
To sum up, Aβ inhibits LTP and promotes LTD. This leads to cleavage of
NPR, which enables AMPARs with associated NPs to escape the synaptic
cleft, as they get overrun by complement binding to the NPs they attract
microglia that engulf the synapse. The declining numbers of non-plastic
synapses manifest as anterograde amnesia in the patient; decreasing synapse
numbers and extensive LTD ultimately leads to accumulating apoptosis
promoting factors and the subsequent death of neurons, manifesting as
retrograde amnesia, and in the end, death of the patient. The model should be
regarded as a working hypothesis, and is in need of rigorous testing. Outlined
below is a discussion on how the results presented in this thesis render some
support for some of the steps in the model.
5.2
The present results in relation to the
model
RAGE has been implicated in the production of Aβ (Cho et al. 2009), as well
as the transport of Aβ from the periphery to the brain over the blood brain
barrier (Deane et al. 2003). Thus, the 82S variant of the RAGE receptor,
which has an increased ligand binding affinity (Hofmann et al. 2002; Osawa
et al. 2007), could lead to increased levels of Aβ. Although we did not
observe any associations of AGER genotypes with Aβ levels, it is possible
that such an association is visible only in the early stages of the disease, since
high affinity would be most relevant in a context of low ligand concentration.
Moreover, RAGE has been shown to mediate the Aβ-inhibition of LTP
(Arancio et al. 2004; Origlia et al. 2009; Origlia et al. 2008), and to activate
microglia (Bianchi et al. 2010). Aβ levels, LTP-inhibition, and activation of
microglia are all important features of the proposed model, and by linking the
Synaptic elimination and the complement system in Alzheimer’s disease
36
82S allele of
AGER to AD, we have provided circumstantial evidence in
support of the model.
Complement mediated synaptic elimination is a key feature of the proposed
model. Our results suggest that synapses in the hippocampus are eliminated
through the actions of the complement system. This is a basic requirement for
the model to hold up, since AD pathology primarily affects the hippocampus.
These results were obtained in animal experiments; mice, do not become
demented, however, and these results are in need of confirmation in humans.
If true, however, a prediction that follows is that CSF complement levels
would be altered in AD as a result of on-going synapse elimination in the
hippocampus. Therefore, we investigated the levels of complement proteins
in CSF from AD patients. Our results revealed a trend towards higher levels
of complement in CSF from AD subjects, thus, supporting the model by
suggesting increased complement activation in patients diagnosed with AD.
In light of the previous findings linking aberrant complement regulation to
AD, we hypothesised that some SNPs in genes encoding complement
proteins could be associated with AD diagnosis. The selected SNPs did,
however, not show any associations with diagnosis of AD. An association
would have had to be regarded as support for the model, but the fact that we
did not find any associations do not really speak against it. Moreover, the
results were likely a consequence of an overestimated effect size on our
behalf, to detect the differences with a satisfactory precision, a larger sample
size would have been needed. We did, however, find an association of
C2/CFB genotype with MMSE and t-tau in AD patients. This is interesting
since CFB is a part in the alternative pathway of the complement cascade;
this pathway mainly functions to amplify the cascade and would thus not
initiate it. This could mean that AD patients that carry the minor CFB allele
have a more intense complement cascade, and thereby more adverse
symptoms of the disease (i.e. lower MMSE and higher t-tau). This reasoning
also lends some support to the model, in which complement activation is one
of the key events.
A wide range of methods has been used during the work on this thesis, thus
approaching a holistic view on the subject. The downside of this, however, is
that it limits the depth of the investigation. My primary interest has been
disease mechanisms and pathogenesis, a subject that is often researched by
use of animal models. I have mostly studied human subjects, however. This
is of major importance since only humans develop AD. Regrettably, our
material was somewhat heterogeneous with respect to age between the