Jonny Daborg
37
diagnostic groups; on the other hand we have used a well characterised
material so the diagnoses are quite certain
5.3
Final speculations on AD pathogenesis
The proposed sequence of events leading to AD could explain how people
get sick, but not why. The initial piece of the puzzle, the ultimate explanation,
is still missing in the proposed model.
A common conception is that given enough time everyone will develop AD.
Although not proven for a fact, this notion has something to it. The question
should perhaps not be if a person will develop AD; rather, the right question
would be when. Aging is the major risk factor for AD. I will argue that
genetics set the stage, and that over time the brain will do the rest. The neural
network of the brain is a complex system. As such it potentially ends up in an
attractor. Perhaps dementia is an inevitable point-attractor, and our brains
have evolved to handle just so many iterations, i.e. the passing of information
through the system. The system constantly changes in response to the input it
receives, with synapses growing stronger and weaker, but perhaps, given
enough time, many synapses will have become so strong that they are no
longer significantly weakened, with a non-plastic network as a result. This
notion is supported by findings showing that senescent rats have larger
unitary EPSPs, as well as a lowered threshold for LTD and a partially
impaired LTP (Burke and Barnes 2010). Aβ is released in a stimulus
dependent manner (Cirrito et al. 2005), and it might be that very strong
synapses release more Aβ, thus dooming the surrounding synapses, and
thereby eventually dooming themselves. In this context it is interesting to
note that APOE ε4 transgenic mice show a loss of synapses with aging, and
that this is accompanied with larger synapses (Cambon et al. 2000).
Genetic background would predispose different individuals to different ages
of onset. APOE genotype, for instance, is associated with age of onset
(Corder et al. 1993; Kurz et al. 1996; Poirier et al. 1993). A simple life
without intellectual challenges would also be a risk factor, since exposing the
network to the same kind of input over and over again, with little variation
would bring it to a non-plastic state much faster. Indeed, participation in
cognitively stimulating activities is associated with decreased risk of AD
(Wilson et al. 2002). In terms of the “point-attractor hypothesis” sketched
above, genetic background and types of input would determine the number of
iterations the neuronal network of an individual can withstand.
Synaptic elimination and the complement system in Alzheimer’s disease
38
Probably the setup
of genes of a person, should also be regarded as a complex
system, meaning that the relation between susceptibility genes and risk is not
linear, and that clinical symptoms should be regarded as emergent properties
of this complex system (Khachaturian 2000). Perhaps more or less specific
combinations of genes work in synergy, though in different ways,
nevertheless producing the same or at least similar outcome. In genetic
studies of AD we look for alleles that increase or decrease the risk for the
disease by comparing AD subjects with normal controls. Perhaps this is not
an optimal approach. AD is a relatively common disease, and some say that
eventually everyone will get it. Could it be that we fail to detect the SAD
causing genes because they are too common, and thus get lost in a sort of
genetic noise? This is a conceivable scenario if these genes do indeed work in
synergy, by themselves posing a very small risk, but in specific clusters
causing a great risk. This highlights the level on which future genetic
association studies should be made, and also the importance of control
subjects. Some people grow very old without developing dementia. Although
it seems as if Aβ is sufficient to initiate the events leading to AD, non-
demented elderly people can have substantial plaque pathology (Bennett et al.
2006; White 2009; O'Brien et al. 2009; Iacono et al. 2009). Something is
clearly missing in our understanding of AD - perhaps these extraordinary
non-demented people should be more extensively studied?
Jonny Daborg
39
6
CONCLUSIONS
I.
In paper I we identified AGER, the gene encoding
RAGE, as a probable susceptibility gene for SAD.
II.
In paper II we show that the key complement protein C3
is involved in elimination of synapses in the
hippocampus of mice, and that the subsequent increase
of synapses that follow deletion of C3 is partially
compensated by a reduced release probability in
glutamate synapses.
III.
In paper III we concluded that C3, C4 and CR1 are not
suitable as CSF biomarkers for AD. There was,
however, a trend towards increased levels of these
proteins in AD, and we interpret this trend as support for
the hypothesis of complement involvement in AD
pathogenesis.
IV.
In the final paper of this thesis we concluded that there
was no association of C2/CFB, C3 or CR1 with AD in
the investigated population. This does not mean that no
such association exists; it should rather be interpreted as
if the association exists, the effect size is very small.
Finally, we concluded that there was an association of
C2/CFB with MMSE and high tau levels in AD patients.
Thus, in the present thesis I have presented evidence in favour of viewing
AD as a disease primarily affecting synapses.