**2.8.2 Plaques and tangles**

The neuropathology of AD is characterised by two pathognomonic entities; the intraneuronal neurofibrillary tangle (NFT) and the extracellular plaque. NFTs are chiefly composed of hyperphosphorylated fibrillar forms of a protein called microtubule associated protein tau (MAPT or tau), while plaques are predominantly composed of fibrils of peptides collectively termed beta-amyloid (Aβ) (Fig. 1 (c-d)).

For most individuals, there is a good correlation between spread of NFT pathology from the medial temporal lobe to the association and finally primary cortices, and the probability of dementia (Newell *et al.,* 1999). In comparison the regional distribution of plaques tends to be highly variable between patients (Braak and Braak, 1991). Similarly plaques vary more in their cortical laminar pattern whereas NFTs are found mainly in layers III and V coinciding with the corticocortical projection (pyramidal) neurons (Lewis *et al.,* 1987). Plaques also vary in their structure and are referred to as 'diffuse' (lacking associated inflammatory cells and dystrophic neurites) and 'cored' or 'neuritic' plaques. It is generally considered that diffuse plaques eventually become cored but an alternative explanation suggests that the amyloid of diffuse and cored plaques are derived from blood vesssels and neurons respectively (D'Andrea and Nagele, 2010).

Extracellular A deposition is considered to be the most likely precipitating event in the disease (Hardy and Selkoe, 2002). In comparison, NFT formation is a subsequent but necessary step in neurodegeneration of AD with intracellular tangles becoming 'ghost tangles' when the neuron eventually succumbs and the insoluble tau protein is left behind in the parenchyma. Presumably the brain macrophages find these extracellular tangles too difficult to phagocytose or they are relatively inert.

As discussed in the section above, a definitive diagnosis of AD can only be made at postmortem examination. Unfortunately the number of autopsies has declined markedly in the last three decades depriving clinicians and researchers alike of an important source of medical knowledge. Neuropathological examinations are now largely restricted to a research environment with the potential issue that clinically confirmed cases are not be

As with the clinical diagnosis, there are established criteria for the pathological diagnosis of AD (Hyman and Trojanowski, 1997) although these too are currently being revised. Interestingly, although postmortem examination is required for a definitive diagnosis of AD the pathological criteria again, only suggests that AD is 'probable' or 'possible'. In order to understand how these criteria have been derived it is useful to give a brief overview of the

The gross pathology of the end-stage AD brain is characterised by widespread atrophy due to the extensive neuronal loss. There is narrowing of gyri and widening of sulci with atrophy of the temporal cortices and a disproportionate atrophy of the entorhinal cortex, amygdala and hippocampus in particular. The remaining cortical regions are generally atrophic, but to a lesser extent than the temporal lobe, although there are also regions seemingly spared from

The neuropathology of AD is characterised by two pathognomonic entities; the intraneuronal neurofibrillary tangle (NFT) and the extracellular plaque. NFTs are chiefly composed of hyperphosphorylated fibrillar forms of a protein called microtubule associated protein tau (MAPT or tau), while plaques are predominantly composed of fibrils of peptides

For most individuals, there is a good correlation between spread of NFT pathology from the medial temporal lobe to the association and finally primary cortices, and the probability of dementia (Newell *et al.,* 1999). In comparison the regional distribution of plaques tends to be highly variable between patients (Braak and Braak, 1991). Similarly plaques vary more in their cortical laminar pattern whereas NFTs are found mainly in layers III and V coinciding with the corticocortical projection (pyramidal) neurons (Lewis *et al.,* 1987). Plaques also vary in their structure and are referred to as 'diffuse' (lacking associated inflammatory cells and dystrophic neurites) and 'cored' or 'neuritic' plaques. It is generally considered that diffuse plaques eventually become cored but an alternative explanation suggests that the amyloid of diffuse and cored plaques are derived from blood vesssels and neurons respectively

Extracellular A deposition is considered to be the most likely precipitating event in the disease (Hardy and Selkoe, 2002). In comparison, NFT formation is a subsequent but necessary step in neurodegeneration of AD with intracellular tangles becoming 'ghost tangles' when the neuron eventually succumbs and the insoluble tau protein is left behind in the parenchyma. Presumably the brain macrophages find these extracellular tangles too

neuronal loss such as the inferior frontal cortex (Halliday *et al.,* 2003) (Fig. 1(a-b)).

representative of the AD spectrum in the greater population.

characteristic or pathognomonic aspects of AD.

collectively termed beta-amyloid (Aβ) (Fig. 1 (c-d)).

difficult to phagocytose or they are relatively inert.

**2.8 Neuropathology** 

**2.8.1 Gross pathology** 

**2.8.2 Plaques and tangles** 

(D'Andrea and Nagele, 2010).

Fig. 1. The neuropathological features of Alzheimer's disease. A series of macro-and microscopic images show the pathological features of Alzheimer's disease (a) A lateral view of the right cerebral hemisphere shows extensive atrophy of the temporal and frontal lobes. There are also regions spared in this case including the precentral (Pr) and and postcentral (Po) gyri and the occipital pole (O), size bar = 2cm (b) A coronal view at the level of the lateral geniculate body demonstrates the severe atrophy in the temporal lobe including the hippocampus (h), along with enlargement of the Sylvian fissure, and third and lateral ventricles, size bar = 2cm (c) A modified Bielschowsky's silver stain shows neurofibrillary tangles in two cortical pyramidal neurons, magnification = 400x and (d) a cored plaque, with its dense amyloid core surrounded by diffuse amyloid, dystrophic neurites and cellular debris, magnification = 200x.

Alzheimer's Disease: Approaches to Pathogenesis in the Genomic Age 397

The current pathological criteria are based on the quantity of both plaques and NFTs (Hyman and Trojanowski, 1997) and incorporate a staging scheme for the neuropathological progression of NFTs ('Braak staging scheme') (Braak and Braak, 1991). The likelihood of AD is high if there are frequent neocortical plaques and NFTs consistent with Braak stage V/VI; is intermediate if plaques are moderate with Braak stage III/VI and low if neocortical

One of the reasons that the AD neuropathological diagnostic criteria are a probability scale is the occurrence of both plaques and NFTs, either independently or together, in a proportion of aged subjects without dementia. Diffuse plaques are seen in as many as 30% of all brains examined (Braak and Braak, 1997; Davis *et al.*, 1999). A similar percentage of cognitively normal individuals have tau-positive neuritic pathology (Knopman *et al.,* 2003) although tau pathology confined to the hippocampal formation appears to be seen in most, if not all, aged brains (Troncoso *et al.,* 1996). As discussed above the quantity of AD pathology tends to correlate relatively closely to the level of dementia and many researchers consider that the presence of mild pathology is representative of preclinical AD (Price *et al.,* 2009). Dementia then occurs when the quantity of AD pathology crosses a particular threshold. Two issues with this idea are that some individuals with A will not develop dementia prior to death (incidental AD pathology) and these A clinicopathological

correlations tend to breakdown in the oldest old (>90 years of age) (Kril, 2009).

reproducibility seen in AD association studies to date (Sutherland *et al.*, 2011b).

However, it was human genome project that provided the real impetus.

**3. The genomic era and its new technologies** 

**3.1 The human genome project** 

As will be discussed further below, the common occurrence of incidental AD and AD-type pathology in nondemented controls could also be a factor in the small effect sizes and lack of

The last decade has seen an incredible advance in the biologist's armamentarium for investigating complex diseases. This biological revolution has its roots in the development of molecular biology techniques such as gene cloning using restriction enzymes (Nathans and Smith, 1975), DNA hybridisation (Southern, 1975), Sanger sequencing (Sanger *et al.,* 1977) and the polymerase chain reaction (PCR) (Saiki *et al.*, 1985; Mullis and Faloona, 1987).

The human genome project began in 1989 and was initially headed by the co-founder of DNA's structure and Nobel laureate, James D Watson. He was succeeded by Francis Collins who headed the public sequencing effort while a private consortium (Celera) also undertook the challenge of sequencing the entire (haploid) human genome, an estimated three billion base pairs. The twin consortia announced the completion of the draft human genome in late 2000 (public (Lander *et al.,* 2001) and Celera (Venter *et al.,* 2001)) although a so-called completed version was only announced in 2003. The templates for these human genomes were pooled DNA samples while the entire genome of a single individual was only

**2.8.5 Diagnostic criteria** 

**2.9 The ageing brain** 

plaques are sparse with Braak stage is I/II.

#### **2.8.3 Neuronal loss**

The hippocampal CA1 (1st cornus ammonis or Ammon's horn) region experiences the greatest loss of neurons in AD of approximately 70%. The CA1 is anatomically and functional connected to the entorhinal cortex where the earliest development of NFTs and neuronal loss in AD is thought to occur. The locus coeruleus (noradrenalin production), nucleus basalis of Meynert and the Raphe nuclei can also experience losses of greater than 50% as does the majority of the temporal lobe (Kril and Halliday, 2001). The loss of neurons in the nucleus basalis of Meynert is similar to the number of ghost tangles but in the hippocampus (Kril *et al.,* 2002) and entorhinal cortex (Gomez-Isla *et al.,* 1996) neuron loss actually exceeds the number of ghost tangle suggesting other neurodegenerative mechanisms are at work. The spread of pathology and associated neuronal loss in AD dictates the symptomology. The reader will note the early involvement of the hippocampus, a key region in both memory generation and consolidation. Interestingly the eventual spread of AD pathology mirrors the anatomical boundaries of the default network – the "brain system active when individuals are not focused on their external environment" (Buckner *et al.,* 2008). The default network is involves with the process of internal mentation or self-relevant mental simulation. This peculiarly human activity could explain the regional selectivity of neuronal loss in AD while deficits in the network are certainly consistent with the loss of self-awareness in moderate to severe stages of the disease (Buckner *et al.,* 2008).

#### **2.8.4 Chronic inflammation**

AD is also characterised by a chronic inflammatory process that is commonly called reactive gliosis. Markers of inflammation such as MHC II expression are higher in demented patients than those in nondemented individuals with AD pathology and may be better correlated with synaptic dysfunction than either plaques or NFTs (Lue *et al.,* 1996). Essentially when we refer to reactive gliosis we mean reactive microgliosis, a hyperplastic and hypertrophic response of the resident macrophages in the brain. There is also astrocyte pathology but this remains less well defined in AD (Beach and McGeer, 1988). The increased expression of the cytosketal protein glial fibrillary acidic protein (GFAP) that characterised 'reactive' astrocytes has been described in AD brains although this was not correlated with either plaques or NFTs (Simpson *et al.*). Microglia are originally derived from the haemopoietic system and migrate to the brain during the early embryological period. Reactive microglia in AD are closely associated with plaques and studies with immunomodulatory therapies suggest that they are relatively effective, at least early in the disease process, at phagocytosing A plaques (Perlmutter *et al.,* 1990; Edison *et al.,* 2008). However neuroinflammation appears to exacerbate AD pathogenesis (Krause and Muller, 2010) and anti-inflammatory medication use is associated with a reduced risk of AD (Vlad *et al.,* 2008). This apparent paradox might be explained by microglia initially serving a protective role but eventually becoming overstimulated and producing excessive reactive oxygen species that lead to neurotoxicity (Innamorato *et al.,* 2009). Nevertheless anti-inflammatory medications do not reduce the progression of AD pathology (Halliday *et al.*, 2000). As microglia are now known to have physiological functions in the brain such the maintenance of synaptic plasticity and it has been suggested that they may be more 'victim' than 'villain' in AD (Graeber and Streit, 2010).
