7 Progressive supranuclear palsy (or Steele–Richardson–Olszewski disease)

In 1963, Dr J. Clifford Richardson presented the first clinical report of this disease (Richardson et al. 1963) (Fig. 7.1).

‘In recent years in Toronto we have seen these few cases of progressive degenerative cerebral disease with a common syndrome of ocular, motor, and mental symptoms which we had not previously recognized which seems to have escaped clear identification and description in the medical literature’. Richardson, Steele, and Olszewski described eight cases, six of whom had died and had been studied neuropathologically (Olszewski et al. 1963). The illness began between the ages of 52 and 62, and the six fatal cases steadily deteriorated, leading to death within 7 years. The illness began with clumsiness, mild forgetfulness, a complaint of visual disturbance, or difficulty with walking. As the condition progressed, speech became slurred, swallowing became difficult, and equilibrium became impaired. A highly distinctive opthalmolplegia

developed in all cases affecting voluntary vertical gaze, especially downgaze. A pseudobulbar palsy appeared, although forced laughter and crying were not encountered. The neck was noted to be extended and rigid. A degree of rigidity was noticed in the limbs in some cases, but none developed typical features of parkinsonism, and none showed tremor. Most of the patients had presented with symptoms of personality change and impaired intellectual ability, and dementia occurred to some degree in all patients but remained mild.

In a companion report (Olszewski et al. 1963) the neuropathological findings were described. There was nerve cell loss and gliosis in the globus pallidus, subthalamic, and red nuclei, the substantia nigra and the locus coeruleus, superior colliculi, periaqueductal grey matter, and pretectal regions, the vestibular nuclei, various nuclei in the reticular formation, and the dentate nuclei. In affected regions, surviving neurons contained neurofibrillary tangles (NFT) without senile plaques. The definitive report of this condition, which Dr Richardson chose to call progressive supranuclear palsy (PSP), was published in 1964 (Steele et al. 1964). No aetiology was discovered, but the authors suggested that it might be due to a degenerative process or a chronic and late viral infection. The resemblance of the condition to postencephalitic parkinsonism and the parkinsonism–dementia complex of Guam was noted.

In retrospect, the earliest clinical accounts of this condition may well date back to Posey (1904) and Spiller (1905).

Others who may have described the condition were Chavagny et al. (1951) and Brusa (1961). However, few clinicians at the time recognized the condition. Nevertheless, following the report of Steele et al. (1964), many patients with PSP were identified. By 1972, Steele was able to review 73 such patients reported between 1951 and that date. Fifty-one of these patients were men, and all were sporadic cases.

PSP or Steele–Richardson–Olszewski disease (SROD) is an idiopathic, sporadic disease, with onset in middle or late life. It is characterized clinically by a supranuclear gaze palsy, especially for vertical gaze and particularly downgaze, neck rigidity, and dystonia, a predominantly axial akinetic–rigid parkinsonian syndrome, a pseudobulbar palsy, loss of equilibrium with frequent falls, and a frontal lobe-type dementia. The condition is progressive and leads to death on average about 6 years from the onset.

Neuropathologically, the condition is characterized by widespread neuronal loss, glial changes, and the presence of characteristic NFT, mainly in diencephalic and brainstem nuclei and tau-positive filamentous inclusions in both neurons and glia.

On gross examination the brain usually shows only minor abnormalities or may appear normal. On slicing, the ventricles may be slightly enlarged. The midbrain is shrunken, particularly the superior colliculi, the mesencephalic tegmentum, and the periaqueductal grey matter. The substantia nigra usually is pale and the red nucleus may occasionally be discoloured. The pontine tegmentum is atrophied and the locus coeruleus is pale. The subthalamic regions and inner segment of the globus pallidus are atrophied and discoloured. The superior cerebellar peduncles and the dentate nuclei are wasted, while the dentate hilum appears discoloured.

The characteristic histological abnormality is the presence of NFT in surviving neurons. NFT with neuronal loss and glial change are found throughout the diencephalon, brainstem, and cerebellum (Jellinger and Bancher 1993; De Bruin and Lees 1994; Daniel et al. 1995). Granulovacuolar degeneration, myelin loss, and perivascular cuffing are inconstant findings (Jellinger and Bancher 1993; Daniel et al. 1995). Table 7.1 lists the major sites of pathology with NFT in PSP. De Bruin and Lees (1994) reviewed 90 cases from the literature with histopathological features compatible with SROD/PSP and gave more extensive information on the different brain regions involved. Areas always affected were the globus pallidus, the subthalamus, and the substantia nigra.

NFT are easily seen on routine haematoxylin and eosin-stained sections, but are dramatically shown in silver-impregnated

preparations (Figs 7.2 and 7.3). Two types are recognized: the globose type is more common than the flame-shaped form. On electronmicroscopy, NFT are seen to consist of straight filaments of indeterminate length, about 12–15 nm in diameter (Tellez-Nagel and Wisniewski 1973) arranged in circling or interlacing bundles. These straight filaments are composed of six or more protofilaments of 2–5 nm (Montpetit et al. 1985). Twisted or paired filaments also have been described, and it has been suggested that the straight filaments may represent a stage in the formation of the latter (Tomonaga 1977; Yagishita et al. 1979). Nevertheless, most observers have concluded that the straight tangles seen in PSP are ultrastructurally different from tangles found in any other disease, and they are considered the hallmark of the condition (Tellez-Nagel and Wisniewski 1973; Roy et al. 1974; Bugiani et al. 1979; Jelllinger et al. 1980; Joachim et al. 1985; Jellinger and Bancher 1993) (Fig. 7.4).

The number of neurons containing NFT was thought to be proportional to the extent of cell loss in different structures (Steele et al. 1964; Kristensen 1985). This was not always the case. In some areas severe loss of neurons was reported with only slight neurofibrillary degeneration or, conversely, many tangles were found where neuronal drop-out was slight or absent. Anzil (1969) suggested that the extent of neuronal loss in relation to that of tangles depends on the duration of the disease before death.

Although the straight NFT characteristic of PSP are morphologically different from the paired helical filaments (PHF) characteristic of Alzheimer's change, initially there was some confusion as immunohistochemistry suggested that there were some similar elements (Probst et al. 1988). For example, Yen et al. (1983) found that both the tangles of Alzheimer's change and those in PSP stained with anti-serum raised against microtubules. Both structures also give positive reactions with antibodies to tau protein (Pollock et al. 1986; Bancher et al. 1987; Probst et al. 1988; Flament et al. 1991; Shin et al. 1991) (Fig. 7.5). However, Galloway (1988) and Flament et al. (1991) found that the abnormal tau species in PSP was different from that in Alzheimer's change. Probst et al. (1988) found that antibodies to phosphorylated epitopes on high and medium molecular weight neurofilament subunits do not, or only weakly, stain tangles in PSP. Schmidt et al. (1988) also found that 12 anti-neurofilament monoclonal antibodies specific for phosphorylated domains did not bind to tangles in PSP. Some ubiquitin epitopes have been described in the tangles of PSP (Mannetto et al. 1988), but most studies found either negative or weakly positive reactions (Love et al. 1988; Leigh et al. 1989). On the other hand, tangles in Alzheimer's disease (AD) readily stain with ubiquitin.

These immunohistochemical investigations revealed differences as well as similarities in the antigenicity of tangles in AD and PSP. Some concluded that the antigenetic profile of tangles in PSP was similar to that of early tangles in AD (Bancher et al. 1987), while others maintained that the mechanisms leading to tangle formation in the two conditions may be different (Shin et al. 1991).

It was also recognized that the presence of NFT was not a specific abnormality and these could be found in many neurological disorders. The characteristic globose NFT in PSP disease were typically found in the brainstem (Steele 1972). Other disorders with tangles involving the brainstem included presenile AD (Hunter 1985; Shortridge et al. 1985; Tabaton et al. 1985), with similar globose tangles described in brainstem neurons, and Arima et al. (1922b) also found such tangles in brainstem neurons in corticobasal degeneration (CBD). The histopathological features of PSP also closely resemble those described in postencephalitic parkinsonism (Ishii and Nakamura 1981) and in the Parkinson–dementia

complex of Guam (Hirano et al. 1961) with tangles present in both disorders, a point emphasized by De Bruin and Lees (1994). The similarities of the neuropathological findings in these three conditions led Daniel et al. (1995) to speculate on a common cause for them all. However, there are some differences. Compared with the findings in PSP, the globus pallidus and subthalamus tend to be spared in postencephalitic parkinsonism, and the cerebral cortex seems more involved in the Parkinson–dementia complex of Guam (Daniel et al. 1995). Accordingly, despite their characteristic presence in PSP, these tangles were felt to represent a form of cytoskeletal disorganization which was not entirely restricted to a single disease. Thus, although there was confusion between these conditions there has been some clarification with further development of anti-tau immunohistochemistry and ultrastructural and genetic studies (see later) (Feany and Dickson 1996).

Immunostaining and silver impregnation techniques (especially the Gallyas stain) showed that in addition to tau-positive inclusions in the neuronal somata and cellular processes (neuropil threads and grains) numerous tau-positive inclusions also occur in glial or astrocytic cells in PSP and related conditions such as CBD (Abe et al. 1992; Nishimura et al. 1992; Yamada et al. 1992, 1993; Komori et al. 1999). One type of astrocytic inclusion in PSP has been described under several terms including ‘tuft-shaped’ or ‘tufted’ astrocytes or ‘glial fibrillary tangles’ (Nishimura et al. 1992; Yamada et al. 1992 and 1993) and another as ‘thorn- shaped’ astrocytes (Ikeda et al. 1995). The ‘tuft-shaped’ astrocytes are typical of PSP and are present in the precentral and premotor cortex but are rare in the temporal and limbic areas. In the subcortical nuclei they preferentially occur in the putamen but may also be present in other brain areas (Fig. 7.6).

Tau-positive intracytoplasmic oligodendroglial inclusions (‘coiled bodies’) are also found in the cortex and underlying white matter as well as several subcortical structures including substantia nigra, pallidum, thalamus, and pons (Yamada et al. 1993; Ikeda et al. 1994).

Development of anti-tau immunohistochemistry and ultrastructural studies has helped clarify the composition of neuronal inclusions in these disorders. It was clear that the microtubule-associated protein tau forms the basic subunit of the inclusions. Tau is a binding protein which serves to promote and stabilize the polymerization of monomeric tubulin into microtubules. The tau protein itself is alternatively spliced from the tau gene with six different types of proteins (isoforms) appearing in normal adult human

brain (Morris et al. 1999). These isoforms differ in the presence of three or four repeated microtubule binding domains (three or four repeat tau) and the extra microtubule binding domain is determined by the inclusion of exon 10 of the tau gene. (also see Chapter 8, ‘Molecular pathology and etiology’) Ultrastructural studies show that the NFT seen in PSP and CBD consist of straight filaments that contain predominantly 4-repeat tau protein isoforms whereas in AD the NFT consist of paired helical filaments containing six major tau isoforms (4- repeat and 3-repeat). On electrophoresis, in contrast to the three major bands and a fourth minor band in AD, abnormal tau in PSP (and CBD) has an electrophoretic migration pattern of two strong bands at 68 and 64 kilodalton (kDa), whereas the minor band at 72 kDa is only variably detected. Tau deposition or dysfunction in these disorders is not simply an incidental finding but is suggested to be the pathogenic cause of neurodegeneration (Hutton et al. 1998; Goedert and Spillantini 2001). Amyloid A4 protein may occur in PSP (Tan et al. 1988; Hauw et al. 1990; Mann and Jones 1990), but plaque formation as seen in AD usually does not occur.

The extent of neuronal loss in PSP varies from structure to structure. In the substantia nigra, Fearnley and Lees (1991) found that the ventromedial portion was most affected. In 14 cases of PSP there was an 82% loss of neurons in the whole substantia nigra pars compacta. This involved both ventral and dorsal tiers about equally. Nerve cell loss has also been reported in the locus coeruleus (Mann et al. 1983), but this has not been confirmed (Tomonaga 1983).

In the nucleus basalis of Meynert, neuronal loss varied from 12.6 to 54.1% compared to controls (Tagliavini et al. 1984; Brandel et al. 1991). Brandel et al. (1991) found loss of cholinergic neurons not only in substantia innominata, but also in the laterodorsal tegmental nucleus, both of which project to the mediodorsal nucleus of the thalamus. In the cholinergic pedunculopontine nucleus Jellinger (1988) noted a loss of 60% of neurons in the pars compacta, associated with tangle formation in 40–64% of surviving neurons. Neuronal loss in this nucleus has also been found by others (Hirsch et al. 1987; Zweig et al. 1987), along with tangle formation. There is also loss of cholinergic neurons in the interstitial nucleus of Cajal (Juncos et al. 1991) and pontine reticular formation (Malessa et al. 1991).

Other brainstem nuclei involved include the cuneiform and subcuneiform nuclei and the ventral tegmental area in the midbrain. Except for the oculomotor nucleus, cranial nerve nuclei mostly are spared, although lesions have been described in the trochlear nucleus, vestibular nucleus, dorsal nucleus of the vagus, and hypoglossal nucleus (Jellinger 1971). Of significance to abnormalities of eye movements, neurofibrillary degeneration has been described in the rostral interstitial nucleus of the medial longitudinal fasciculus, the nucleus of Darkshewitsch (Constantinidis et al. 1970; Davis et al. 1985), the interstitial nucleus of Cajal (Davis et al. 1985), and the nucleus of the posterior commissure (Davis et al. 1985) (see also Daniel et al. 1995).

Oyanagi et al. (1988) have shown a 30–40% loss of large cholinergic interneurons in the head of the caudate nucleus and the putamen, but small neurons were well preserved. However, Levy et al. (1995) reported a 50–60% loss of glutamic acid decarboxylase containing neurons in the caudate nucleus, ventral striatum, and the external and internal globus pallidus in PSP. The medius globus pallidus is said to be more affected than the lateral segment (Jellinger 1971). Ito et al. (1992) reported loss of calbindin-D_28k immunoreactivity in medium-sized neurons of the globus pallidus. Thus, there appears to be major destruction of the striatopallidal complex in this condition.

There is severe neuronal loss in the dentate nucleus, sometimes associated with grumose degeneration (Arai 1989). The lesion of the dentate nucleus is associated with degeneration of the central tegmental tract, resulting in hypertrophy of the olivary nucleus in some cases (see also Mizusawa et al. 1989). The cerebellar cortex is spared or only mildly affected. The thalamus may be involved (David et al. 1968; Gomori et al. 1984).

Kato et al. (1986) also described pathology in the spinal cord, with tangles in anterior, posterior, and lateral horns in Clarke's column and the intermediate grey matter. The posterior horns were most severely affected.

There had been debate over the extent to which the cerebral cortex was involved in PSP. Although the original report suggested that it was spared, it was realized that the cortex may be affected with NFT and neuropil threads (Ishino and Otskuki 1976; Takahashi et al. 1989; Hauw et al. 1990; Hof et al. 1992). Hauw et al. (1990) found neurofilament pathology in frontal cortex and hippocampus, and mentioned it was different in character from that in AD. Amano et al. (1989) and Braak et al. (1992) showed that NFT in the cerebral cortex were largely confined to the hippocampal formation, especially in the parahippocampal gyri and other limbic areas. There was also involvement of the prefrontal and inferior temporal cortex, particularly in layers II and III, and moderate tangle deposition in the primary motor cortex (Hof et al. 1992). The tangles in cortical neurons were straight tubules of 15 nm diameter (Takahashi et al. 1989). Silver impregnation by the method of Gallyas had also shown neuropil threads in affected cortical areas, subcortical grey matter, neocortex, hippocampus, and white matter tracts (Probst et al. 1988; Hauw et al. 1990; Daniel et al. 1995). Gliosis is part of the histological picture and is particularly evident in areas of severe neuronal loss.

De Bruin and Lees (1994), reviewing 90 pathologically verified cases, reported that the cerebral cortex, most commonly the frontal cortex, was described as abnormal (NFT, senile plaques, neuronal loss, and gliosis) in 40% of cases. Later, Daniel et al. (1995) found NFT in the cerebral cortex in all of 17 cases with the condition. These always involved the hippocampus and parahippocampus, and also usually the frontal cortex. Primary motor cortex was more severely involved than association cortices. Their density varied from infrequent to numerous. Mostly they involved small neurons of laminae 3 to 6. In this material, neocortical senile plaques were infrequent and below the density accepted for the diagnosis of AD.

Based upon neuropathological criteria, Lantos (1994) suggested that there were three types of PSP. The typical or type 1 case conforms to the original definition. Type 2 atypical variants differ in that the severity or the distribution of abnormalities does not conform to the typical pattern. Type 3 cases exhibit the typical pathology accompanied by lesions characteristic of other neurodegenerative or vascular disease (for example, Lewy bodies of Alzheimer-type change).

Lewy bodies have been described in brainstem nuclei and in the cerebral cortex in some cases of PSP (Mori et al. 1986; D’Amato et al. 1991, 1992). Swollen cortical neurons similar to those seen in CBD also have been reported (D’Amato et al. 1992; Watts et al. 1994). Even the concomitant neuropathological findings of Pick's disease have been found in occasional cases (Wilson et al. 1989; Arima et al. 1992a). Elderly patients with PSP may exhibit diffuse plaques in the cerebral cortex (Mann and Jones 1990; Sasaki et al. 1991).

Gearing et al. (1994) reviewed 13 patients with pathologically diagnosed PSP, and emphasized both neuropathological and clinical heterogeneity. Five had been diagnosed as having AD in life. Seven had concomitant pathological changes of Alzheimer's, Parkinson's disease (PD), or both. Collins et al. (1995) reported that a definite clinical diagnosis had only been made in 8 out of 12 patients with pathological changes diagnostic of the disease. Jellinger et al. (1995) also found that only 12 out of 24 pathologically proven cases of PSP had been diagnosed in life.

Masliah et al. (1991) described five late onset dementia patients whose brains showed no evidence of AD or any other specific pathological diagnosis. They found argyrophilic grains, NFT, and neuropil threads in the hippocampus, entorhinal cortex, locus coeruleus, substantia nigra, and inferior olives. Electronmicroscopy showed the structures to consist of straight tubulofilamentous profiles, 25 nm in diameter, similar to those found in patients with PSP, but different from the paired helical filaments of patients with AD. They comment that these findings may represent an atypical form of PSP, as may be the entity described by Braak and Braak (1987, 1989) as ‘adult onset dementia with argyrophilic grains’.

Because of these difficulties Hauw et al. (1994) proposed preliminary National Institute of Neurological Disorders and Stroke (NINDS) criteria for the neuropathological diagnosis of PSP. This distinguishes typical, atypical, and combined cases and employs a semiquantitative distribution of NFT. In typical cases there are numerous NFT or neuropil threads, or both, in the basal ganglia and brainstem. Such abnormalities must be present at least in three of the following areas: globus pallidus, subthalamus, substantia nigra, and pons. In addition, one or more neurons containing tangles or neuropil threads must be evident in three or more of the following areas: striatum, oculomotor complex, medulla, and dentate nucleus. In atypical cases tangles and neuropil threads must be present in basal ganglia and brainstem, involving at least five of the following areas: globus pallidus, subthalamic nucleus, substantia nigra, pons, medulla, and dentate nucleus. In combined PSP cases there are the neuropathological changes of atypical cases, with additional findings diagnostic of other neurological diseases. The presence of tau-positive astrocytes or processes (or the presence of tangles in astrocytes) are considered to support the diagnosis. The extent of neuronal loss and gliosis may be variable.

Exclusion criteria for typical and atypical cases include the presence of Lewy bodies and extensive Alzheimer's pathology.

Litvan et al. (1996) investigated the validity and reliability of these preliminary NINDS neuropathologic diagnostic criteria for PSP and related disorders. The specific disorders were typical, atypical, and combined PSP, postencephalitic parkinsonism, CBD, and Pick's disease. The results showed that with routine sampling and staining methods, neuropathologic examination alone was not fully adequate for differentiating the disorders. The main difficulties were discriminating the subtypes of PSP and separating postencephalitic parkinsonism from PSP. CBD and Pick's disease were less difficult to distinguish from PSP. The addition of minimal clinical information contributed to the accuracy of the diagnosis and thereafter the criteria for clinical diagnosis were also defined (see later) (Litvan et al. 1996b).

A number of studies have investigated changes in brain monoamines (Table 7.2), acetylcholine (Fig. 7.8), peptides, and various receptors (Table 7.3) in autopsy material of PSP (Young 1985; Ruberg et al. 1993).

There is a marked loss of dopamine in the caudate nucleus, putamen, and substantia nigra of some 87–92% (Kish et al. 1985; Ruberg et al. 1985, 1993). The level of noradrenaline in the caudate nucleus and putamen also is reduced by around 70%, but there is only a small (20–30%) and insignificant reduction in serotonin (5-HT) concentrations in these regions. Levels of dopamine, noradrenaline, and 5-HT were normal in the nucleus accumbens, but only dopamine was reduced in the parolfactory gyrus (Brodmann area 25) by about 60% (Kish et al. 1985; Ruberg et al. 1985). No significant changes in monoamines were found in other cortical areas.

Table 7.2 summarizes the brain monoamine changes in PSP in comparison to those in idiopathic Parkinson's disease (Hornykiewicz and Shannak 1994). The two conditions are similar in that there is extensive loss of striatal dopamine associated with degeneration of the nigrostriatal dopaminergic pathway, and loss of dopamine in the parolfactory cortex. They differ, however, in that the nucleus accumbens seems relatively spared with regard to dopamine content in PSP. On more detailed analysis, dopamine loss in the striatum in PSP was as severe in caudate nucleus (92%) as in putamen (88%), whereas in Parkinson's disease the putamen is more affected than the caudate. Another difference between the two conditions was the more severe loss of noradrenaline and 5-HT in idiopathic Parkinson's disease compared with PSP.

There is widespread damage to the cholinergic systems in PSP. Choline acetyltransferase (ChAT) activity, an index of cholinergic innervation, was decreased by 40–70% in the caudate nucleus, the putamen, the nucleus accumbens, the internal pallidum, and the subthalamic nucleus (Ruberg et al. 1985; Agid et al. 1986). This probably represents loss of cholinergic neurons in these structures. The loss of cholinergic interneurons in the striatum of patients with PSP contrasts with preservation of such neurons in Parkinson's disease (Hirsch et al. 1989).

ChAT activity in the cerebral cortex is decreased in PSP to a small degree in some areas: 20–40% in the frontal cortex, the cingulate cortex, and the hippocampus (Kish et al. 1985; Ruberg et al. 1985; Agid et al. 1986; Whitehouse et al. 1988). There is a considerable decrease (some 70%) of ChAT activity in the substantia innominata (Ruberg et al. 1985). This was considered to represent the combined effects of a modest loss of neurons in the substantia innominata, with an additional loss of cholinergic afferents from other brainstem nuclei. There also is loss of ChAT in certain nuclei of the thalamus in PSP (Brandel et al. 1990, 1991), particularly in the mediodorsal nucleus. Suzuki et al. (2002) autoradiographically determined the regional expression of acetylcholine vesicular transporter (VAChT) and monoamine vesicular transporter type 2 (VMAT2) binding sites in postmortem basal ganglia samples from subjects with PSP. VAChT expressions and ChAT activities in caudate nucleus and putamen were markedly decreased in PSP, whereas benzodiazepine binding was unaffected, consistent with selective losses of striatal cholinergic interneurons.

Immunocytochemical identification of ChAT in small brainstem regions has shown a loss of cholinergic neurons in many areas, including the Edinger–Westphal nucleus (69%), the rostral interstitial nucleus of the medial longitudinal fasciculus (97%), and the interstitial nucleus of Cajal (78%), as well as in the deep layers of superior colliculus (93%) (Juncos et al. 1991). There is also significant loss of cholinergic neurons in the nucleus pedunculopontinus (Zweig et al. 1985; Hirsch et al. 1987) and in the lower pontine reticular formation (Malessa et al. 1991). A comparative study of cholinergic neuronal changes in the nucleus basalis of Meynert (NBM) and the laterodorsal tegmental and pedunculopontine tegmental nuclei (LdtgN and PptgN) of brains obtained at autopsy suggested that in PSP, cholinergic neurons in the LdtgN are likely to be more vulnerable than PptgN and NBM (Kasashima, and Oda 2003). Javoy-Agid (1994) summed up the above findings as follows: PSP is associated with a widespread cholinergic deficit likely corresponding to a loss in cholinergic neurons. The cholinergic damage dramatically affects the basal ganglia and specific cell groups of the mesencephalon and pons, thus providing an anatomically defined basis for motor and supranuclear oculomotor syndromes characteristic of PSP. Unlike AD and Parkinson's disease with dementia, however, PSP is not associated with a marked cholinergic deficiency in the cerebral cortex. These findings have also now been corroborated using positron emission tomography (PET) and the ligand N-methyl-4-[11C]piperidyl acetate to calculate an index of acetylcholinesterase activity. Results comparing Parkinson's disease and PSP suggest that there is a loss of cholinergic innervation to the cerebral cortex in association with cholinergic innervation to the thalamus in Parkinson's disease, whereas there appears to be a preferential loss of cholinergic innervation to the thalamus in PSP (Shinotoh et al. 1999).

Initial studies suggested normal levels of alpha-aminobutyric acid (GABA) and reduced activity of glutamic acid decarboxylase (GAD) in striatum and pallidum in PSP disease (Kish et al. 1985; Agid et al. 1986). Levy et al. (1995) examined the mRNA for GAD₆₇ in the striatum and pallidum of three patients using in situ hybridization. They reported a 50–60% decrease in the number of neurons expressing GAD₆₇ in the caudate nucleus, putamen, ventral striatum, and the external and internal pallidum, with reduced expression in individual neurons. These authors attributed such changes to marked loss of small striatal and pallidal GABA neurons. For GABA receptors see later.

The concentrations of several peptide neurotransmitters, methionine–enkephalin, leucine–enkephalin, dynorphin, substance P, and cholicystokinin-8, are normal in striatum, substantia nigra, and pallidum in PSP (Taquet et al. 1987). This is another difference from PD, in which significant changes in these various peptides in the basal ganglia have been demonstrated. There was no significant change in the concentration of somatostatin in the cerebral cortex in PSP (Applebaum 1987).

The density of various neurotransmitter receptors and transporters in different brain regions in PSP has been reported by Landwehrmeyer and Palacios (1994) and by Pascual et al. (1994).

There was a profound reduction of dopamine transporter uptake sites in the caudate nucleus (82%) and putamen (83%), with a lesser reduction in the nucleus accumbens (66%) (Chinaglia et al. 1992). This reflects the considerable loss of the dopaminergic nigrostriatal pathway particularly to the caudate and putamen.

Dopamine D2 receptors have been reported to be unchanged or decreased in density in the striatum, depending on the radioactive ligand used (Bokobza et al. 1984; Ruberg et al. 1985, 1993; Pierot et al. 1988). There is general agreement that D2 receptors labelled with ³H-spiperone are considerably reduced in the striatum, by about 50% (Landwehrmeyer and Palacios 1994) to more than 90% (Pascual et al. 1992, 1994). However, studies using other dopamine D2 selective compounds, such as iodosulpiride, have shown no significant difference, and levels of D2 mRNA message in the striatum in PSP have been found to be normal (Landwehrmeyer and Palacios 1994).

Dopamine D1 receptor density in the striatum also appears unchanged in PSP (Pierot et al. 1988; Pascual et al. 1992, 1994; Landwehrmeyer and Palacios 1994). These findings suggest that only one subtype of dopamine D2-like receptors, identified by spiperone, is affected.

5-HT uptake sites and 5-HT receptors were found to be unaltered throughout the basal ganglia (Chinaglia et al. 1993; Landwehrmeyer and Palacios 1994; Pascual et al. 1994). However, 5-HT 1 receptors were reduced in the substantia nigra and there was also a profound loss of nigral D1 dopamine receptors in that structure (Pascual et al. 1994).

Muscarinic receptors identified by N-methyl-scopolamine, which labels all muscarinic receptor subtypes, were reduced in the striatum in PSP by about 30% (Landwehrmeyer and Palacios 1994; Pascual et al. 1994). In contrast, using radioactive quinuclidinyl benzilate (QNB) as a ligand, no loss of striatal muscarinic binding sites was found (Ruberg et al. 1985). This suggests that there may be a loss of muscarinic receptors of the M2 subtype thought to be expressed on cholinergic interneurons. Muscarinic receptors in the cerebral cortex and hippocampus, however, appeared to be normal in density in PSP (Pascual et al. 1994).

Noradrenergic alpha-2 receptors were decreased in frontal cortex, hippocampus, and striatum (Pascual et al. 1994). GABA_A and benzodiazapine receptors were significantly reduced in the globus pallidus in PSP (Landwehrmeyer and Palacios 1994). There was an extensive loss of neurotensin binding in the substantia nigra (Landwehrmeyer and Palacios 1994). However, there was no significant reduction in substance P, although there was a trend to a slightly decreased density (Landwehrmeyer and Palacios 1994).

Dexter et al. (1991) described normal levels of total iron and ferritin in cerebral cortex, caudate nucleus, and cerebellum in the brains of 11 patients with PSP. There was a 70% increase of total iron in substantia nigra pars compacta, with a corresponding 70% increase in ferritin; iron levels in putamen were increased by 36% in the putamen, but there was no change in ferritin. Total manganese content was normal in all these brain areas, except for a 37% increase in the cerebral cortex. Total copper levels were normal, except for a reduction in cerebellum. Total zinc levels also were normal except for a reduction in the caudate nucleus.

Sian et al. (1994) reported decreased levels of reduced glutathione (GSH) in caudate nucleus, but not in cerebral cortex, putamen, or substantia nigra, with no change in oxidized glutathione (GSSG). Perry et al. (1988) reported increased levels of glutathione in the substantia nigra, contrasting with reduced levels in Parkinson's disease.

Zemaitaitis et al. (2003) measured transglutaminase enzyme activity and mRNA and protein levels of three transglutaminase isoforms that are expressed in human brain. Transglutaminases catalyze the covalent cross-linking of substrate proteins to form insoluble protein complexes that are resistant to degradation. Overall, transglutaminase 1 and 2 and mRNA activity was significantly increased in the globus pallidus and pons in PSP. Tissues with more transglutaminase activity had more NFT. These findings suggested that transglutaminase enzymes could be involved in the formation and/or stabilization of NFT in selectively vulnerable brain regions in PSP and these transglutaminases could be potential targets for future therapeutic interventions.

Di Monte et al. (1993) reported decreased rates of ATP production in muscle, suggesting impaired mitochondrial respiratory chain activity. In fact, a whole body of literature has accumulated implicating mitochondrial dysfunction and oxidative stress mechanisms in the pathophysiology of PSP based primarily on biochemical studies in post-mortem PSP brain tissue and PSP cybrids (Sverdlow et al. 2000; Albers and Beal 2002). These studies support a contributory role of bio-energetic defects in the pathogenesis of PSP (for review see Albers and Augood 2001). In this regard, decreased activities of a mitochondrial enzyme complex, alpha-ketoglutarate dehydrogenase complex (KGDHC), and marked increases in tissue malondialdehyde and 4-hydroxynonenal levels (both markers of lipid perioxidation) have been reported in post-mortem superior frontal cortex, subthalamic nucleus, midbrain, and cerebellum from patients with PSP compared with controls (Albers and Augood 2001; Park et al. 2001). In addition, an increase in inducible nitric oxide synthase protein has been reported in tufted astrocytes typical of PSP, suggesting that they may be at least one source of free radicals in the PSP brain (Komori et al. 1998; Albers and Augood 2001).

Ishizawa and Dickson (2001) examined the role of microglia in PSP brains versus CBD. Microglial and tau burdens were determined with image analysis on brain sections that had been immunostained with monoclonal antibodies to HLA-DR and phospho-tau. They found that microglial activation was greater in PSP and CBD than in normal controls, and that the microglial burden correlated with the tau burden in most areas, with PSP showing more pathology in infratentorial structures and CBD showing more pathology in supratentorial structures. The authors concluded that microglia may play a role in disease pathogenesis.

Although PSP is uncommon, Kristensen (1985) was able to collect 325 cases from the literature in 1984, two decades after the disease had been defined. It was estimated that about 4–6% of patients with akinetic–rigid parkinsonian syndromes will have PSP (Jackson et al. 1993). With regard to incidence, Mastaglia et al. (1973) estimated that the incidence of PSP in Western Australia was four cases per million of the population per year. With regard to prevalence, Golbe et al. (1988) surveyed two counties in Central New Jersey in 1986 with a population of 800 000 and found 11 cases. They calculated an age-adjusted prevalence rate in the United States of 1.39 per 100 000. The prevalence rate in those over the age of 65 was 7 per 100 000 of the population. Golbe et al. considered these figures to be a minimum estimate of prevalence in view of ascertainment difficulties. This prevalence rate of 1.39 per 100 000 of the population for PSP was about 1% of that for Parkinson's disease in Western countries, which approximates to 100–170 per 100 000.

Golbe et al. (1988) also calculated that the incidence of PSP was about 2.9 cases per million population per year. Rajput et al. (1984) found two cases over a 13 year period in a population of about 50 000, given a crude prevalence rate of 3.1 per million per year.

However, subsequent analyses have suggested a higher incidence. An incidence study of PSP and all types of parkinsonism in Olmsted County, Minnesota, from 1976 to 1990, found a crude incidence rate for PSP of 1.1 per 100 000 per year, which increased exponentially from 1.7 cases per 100 000 per year at ages 50–59 years to 14.7 per 100 000 per year at ages 80–99 years (Bower et al. 1997, 1999). The incidence of PSP was therefore estimated to be close to 10% of the incidence of Parkinson's disease (Bower et al. 1999), making it much more common than previously recognized. It is thought that these figures are higher because of better methodological case findings and increased recognition of the disorder by neurologists and non-neurologists. Similarly, the earlier underestimated prevalence of 1.39 per 100 000 (Golbe et al. 1988) has been found to be 6.0 and 6.4/100 000 in two population-based prevalence studies conducted in the United Kingdom using currently accepted diagnostic criteria for PSP (Schrag et al. 1999; Nath et al. 2001) Of note, most of the PSP cases identified in the London area had not been diagnosed as PSP until the study, suggesting that PSP continues to be under-diagnosed (Schrag et al. 1999).

In most series [(except for that of Maher and Lees (1986)], there is a male preponderance, with about 60% of patients being men and about 40% being women (Kristensen 1985; SantaCruz et al. 1998). Maher and Lees (1986) found that women predominated in their series of 52 cases; 22 were men and 30 were women. The male predominance was, however, clear in 90 pathologically verified cases reviewed by De Bruin and Lees (1995), of whom 51 were male and 34 were female.

The mean age of onset is in the early 60s, being about 62 years (Kristensen 1985; Maher and Lees 1986; Golbe et al. 1988; De Bruin and Lees 1995). Seventy-one per cent of patients noticed the first symptoms of their illness between the ages of 50 and 65 years (Kristensen 1985). In the series reported by Golbe et al. (1988), 30% had onset before the age of 60 and 14% after the age of 69 years. Maher and Lees (1986) found that the median delay from symptom onset to diagnosis was 3 years.

In the series of Maher and Lees (1986), the 33 patients who had died had a median adjusted survival of 5.9 years (range 1.2–10.3 years). Amongst the 15 deceased patients of Golbe et al. (1988), the median duration was 6.9 years (2–17 years). Kristensen (1985) calculated an average survival time in 73 cases as 5.7 years (range 1–23 years). In the series of de Bruin and Lees (1995), the mean age at death was 67 years.

There appears to be no obvious connection between the age of onset and the duration of the disease (Hind et al. 1982; Maher and Lees 1986). Most patients die of intercurrent infections, usually bronchopneumonia, aspiration, or inanition.

One study retrospectively evaluating the medical records of 24 autopsy-confirmed cases with regard to the natural history of PSP (Litvan et al. 1996a) found that onset of falls during the first year, early dysphagia, and incontinence predicted a shorter survival time. As mentioned in previous studies (Hind et al. 1982), age at onset had no effect on survival and neither had gender, early onset of dementia, vertical supranuclear gaze palsy, or axial rigidity. Similar predictors of survival were identified in a case record-based study of 187 PSP patients (Nath et al. 2001). These investigators found that classification as probable PSP according to the NINDS-SPSP criteria was associated with a poorer survival. In particular, they found that onset of falls, speech problems, or diplopia within 1 year and swallowing problems within 2 years of onset were associated with a worse prognosis. It was also noted that PSP was diagnosed later in men than women and that, once diagnosed, men died earlier than women.

The cause of PSP is unknown. There have been few analytical epidemiological studies of the condition world-wide. However, the disease has been described in most parts of the developing world and in many ethnic groups (Hind et al. 1982).

With regard to case-control studies to ascertain risk factors for developing PSP (Davis et al. 1988; Golbe et al. 1996) there was no increased risk of PSP as a result of living in rural areas or in small towns or on farms. There was also no relationship with alcohol, smoking, contact sports, head trauma, exposure to occupational toxins, maternal age, or a history of familial neurological disease. One statistically significant finding in the first study was that patients, relative to controls, were about three times more likely to have completed High School education and to have been to college (Davis et al. 1988). This may have been due to ascertainment bias. A follow-up study (Golbe et al. 1996) found in fact (contrary to the previous report) that PSP patients were significantly less likely to have completed at least 12 years of school. This the authors hypothesized could be a proxy for poor early-life nutrition or for occupational or residential exposure to an as-yet unsuspected toxin (Golbe et al. 1996). There has been one report of four patients who experienced unusual exposure to organic solvents (McCrank and Rabheru 1989).

One issue in the literature has been the debate about whether presymptomatic hypertension is associated with the occurrence of PSP, with some reports mentioning that nearly 80% of patients with clinically diagnosed PSP had a history of hypertension. These authors felt that as adrenergic nuclei of the brainstem are severely affected, hypertension could be the first symptom arising from this involvement and could also explain the features of small vessel disease seen on computed tomography or magnetic resonance imaging (MRI) in 50% of their patients (Ghika and Bogousslavsky 1997). Such MRI changes, however, could be non-specific and age related. Other workers have been unable to confirm the association with hypertension in the setting of a multicentre case-control study of patients affected by PSP (Fabbrini et al. 1998). Because previous studies were performed on patients with presumed diagnosis of PSP, Colismo et al. (2003) looked at the issue of antemortem hypertension in 73 pathologically proven PSP cases. The prevalence of hypertension (> 140/90 mm Hg) in PSP patients at the first and at the last visit during their neurological disease was compared with normal controls; 29 of 73 (39.7%) of the patients were recorded as having hypertension at the first visit during the disease course, and this ratio increased to 42 of 73 (57.5%) at the last visit before death. When these figures were compared with the 21 normal controls (11 of 21 with hypertension, 52.4%), they did not find an increased prevalence of hypertension in PSP.

Repeated head injury can lead to accumulation of NFT. In one report the authors examined the brains of four young men and a frontal lobectomy specimen from a fifth, age range 23–28 years, all of whom had suffered mild chronic head injury. There were two boxers, a footballer, a mentally subnormal man with a long history of head banging, and an epileptic patient who repeatedly hit his head during seizures. Pathological findings in all five cases were of neocortical NFT and neuropil threads, with groups of tangles consistently situated around blood vessels in the worst affected regions. No a-beta immunoreactivity was detected. It appears that repetitive head injury in young adults is initially associated with neocortical NFT formation in the absence of a-beta deposition. The distribution of the tau pathology suggests that the pathogenesis of cytoskeletal abnormalities may involve damage to blood vessels or perivascular elements (Geddes et al. 1999). Similarly, Irving et al. (1996) found tau-positive oligodendrocytes in head-injured but not in control cases. Epidemiological studies have identified a history of head injury as a risk factor for AD (Smith et al. 2003). However, as mentioned above, in the few case control studies so far no clear association has yet been described between incidence of PSP with head trauma.

There are limited data regarding smoking and PSP. However, it appears that the inverse association with smoking found previously in Parkinson's disease is shared by multiple system atrophy but not by PSP, suggesting, according to the authors of one report, that different smoking habits are associated with different groups of neurodegenerative disease (Vanacore et al. 2000).

In general, PSP is a sporadic illness and there is no family history. However, PSP is associated with a specific form of the tau gene (H1 tau haplotype) (Baker et al. 1999; Di Maria et al. 2000; Houlden et al. 2001), as shown in Table 7.4.

However, because the H1/H1 genotype is present in approximately 90% of patients with this disorder, but also in approximately 60% of healthy Caucasians, it is unclear whether inheritance of the H1/H1 tau genotype represents a predisposition to develop PSP (requiring other environmental or genetic factors) or whether a relatively rare mutation with low penetrance (rather than an inherited susceptibility variant) could contribute to the abnormal tau aggregation present in this disorder. It is likely that differential environmental or genetic factors may lead to different cell vulnerability, phenotypes, and rate of disease progression observed in PSP patients. CBD is also associated with the inheritance of the H1 haplotype (Di Maria et al. 2000; Houlden et al. 2001). A pilot study showed that H1 haplotype dosage does not influence age at onset, severity, or survival of PSP patients (Litvan et al. 2001), in contrast to what is observed with APOEε4 in AD.

The observation that coding and splice-site mutations in the tau gene cause frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) demonstrates that tau dysfunction is sufficient to induce neurodegeneration (Hutton et al. 1998; Dsouza et al. 1999). The parallels between FTDP-17 and PSP/CBD also include that there are FTDP-17 patients with defined tau mutations who share many clinical and pathological features with PSP and CBD (Poorkaj et al. 1998; Bird et al. 1999; Bugiani et al. 1999; van Swieten et al. 1999; Morris et al. 2002) such as selective deposition of 4-R tau. However, because highly penetrant tau mutations are not found in PSP and CBD, it is likely that other genetic or environmental factors contribute to the development of these disorders.

A few familial cases of PSP have been reported, but it remains doubtful whether they have been true PSP or may have been PSP-like conditions due to a tau gene mutation, for example, as seen in familial cases with the FTDP-17 mutations of the tau gene. Many of the case reports of familial PSP were described before the advent of tau genetics. However, even in those reports, close inspection often suggests some discrepancies compared with PSP. For example, Mata et al. (1983) reported three siblings who in their 20s developed a PSP-like syndrome, but with more pyramidal signs and dementia. First, the onset was very early, dementia was prominent, and, neuropathologically, the distribution of NFT was highly atypical for PSP, and the authors rightly concluded that their patients had an undescribed but different illness. Ohara et al. (1992) reported two siblings from a consanguineous mating with an illness resembling PSP both clinically and neuropathologically. However, again the distribution of pathology had certain features different from those characteristic of typical PSP. A case reported by David et al. (1968) has been quoted to indicate familial occurrence of the illness, as the mother of one of their patients was said to have a parkinsonian syndrome by history. No further information on this family is available. Brown et al. (1993) described ‘familial progressive supranuclear palsy’ in three family members in two generations. They developed progressive dementia, an akinetic–rigid syndrome, gait disturbance, and dysarthria in their 60s. Neuropathological examination of the brain of one of their cases revealed changes compatible with PSP. Rojo et al. (1999) reported 12 pedigrees with familial PSP, confirmed in four probands by internationally agreed pathological criteria. The spectrum of the clinical phenotypes in these families was variable, including 34 typical cases of PSP (12 probands plus 22 secondary cases), while others had an atypical phenotype. The inheritance was of an autosomal dominant transmission with incomplete penetrance. No tau genetics was carried out and some of these families at least could well have FTDP-17 or another tau gene mutation as the cause of their disorder. However, Morris et al. (2002) analysed the tau sequence in two small families with PSP, and a number of clinically typical and atypical sporadic cases with pathological confirmation of the diagnosis. The tau mutations described in FTDP-17 were not found in most clinically diagnosed patients with PSP. The authors concluded that usually FTDP-17 and PSP, including the rare familial form of PSP, are likely to be separate conditions and that usually PSP and typical PSP-like syndromes are not due to mutations in tau. The overall conclusion, however, is that few really convincing examples of inheritance of PSP have been described.

Lees (1987) (Table 7.5) and Tolosa et al. (1994) (Table 7.6) proposed criteria for the clinical diagnosis of PSP, but the criteria generally accepted currently, particularly for research, are the NINDS-SPSP mentioned above (Table 7.7) (Litvan et al. 1996b).

Duvoisin (1994) also has given the relative value of different clinical manifestations for the diagnosis of the condition (Table 7.8). Collins et al. (1995) described 12 cases with the pathology of PSP

and drew up neuropathologically based diagnostic clinical criteria for the diagnosis (Fig. 7.9). These authors noted atypical findings in a number of their cases, including absence of supranuclear gaze palsy (two cases), prominent asymmetry (two cases), arm dystonia (two cases), arm apraxia (two cases), myoclonus (two cases), chorea (one case), and respiratory disturbance (one case).

De Bruin and Lees (1994) have reviewed the clinical findings in 90 cases of PSP, pathologically verified, described in the literature from 1951 to 1992. The frequency of symptoms and signs is give in Table 7.9 and the frequency of eye movement and eyelid abnormalities in Table 7.10.

The earliest complaints are often vague and non-specific. Unsteadiness with poor balance and unexplained falls are the commonest initial features, occurring in about two-thirds of patients (Lees 1987; Jankovic et al. 1990; Nath et al. 2003). Dysarthria is another common early symptom. Often this has a characteristic strained, spastic, or strangled quality. There also may be stuttering, palilalia, or even echolalia. Behavioural disturbances are evident early in the illness in at least half the patients and include depression, irritability, and uncharacteristic aggressiveness, emotional lability, apathy, and slowness of thinking, with complaints of memory impairment (Lees 1987). Many patients are misdiagnosed as having a dementia or a psychotic illness. Visual disturbances also may occur early, with complaints of blurred vision, difficulty reading, problems going downstairs or while eating, diplopia, and dry eyes. Other early symptoms include difficulty swallowing, with explosive coughing, and a general slowing up or clumsiness of all movements with stiffness of the limbs.

About a third of patients with PSP complain of blurred vision, diplopia, and eye discomfort, and many eventually lose their ability to read or maintain eye contact (Friedman et al. 1992).

The most characteristic and diagnostic abnormality on examination is supranuclear gaze palsy (Fig. 7.10). Inability to move the eyes, especially in the vertical plane, accompanied by lid retraction, often with furrowing of the forehead, and by an extended neck, gives a quite characteristic facial appearance.

Voluntary saccadic eye movements are affected first, the velocity declining before their magnitude. Upward gaze and convergence usually are impaired initially, but it can be difficult to separate this from the effects of normal ageing. As the conditions progresses, however, downgaze is impaired. Indeed, many consider that the diagnosis cannot be made until an impairment of downgaze appears. Paralysis of vertical saccades differentiates PSP from other parkinsonian disorders (Vidailhet et al. 1994).

It should be emphasized that a supranuclear palsy for vertical gaze, sometimes including downgaze, is not specific to PSP. Table 7.11 summarizes the degenerative disorders in which such a supranuclear opthalmoplegia has been reported (Lees 1987).

In the early stages of development of the supranuclear vertical gaze palsy, there may be merely hesitancy on attempting to look downwards or impersistence of gaze. At this stage there may be defective suppression of the vestibulo-ocular reflex and loss of the fast component of optokinetic nystagmus in the vertical plane. Failure to suppress the vestibulo-ocular reflex is not specific to PSP, but also occurs in Parkinson's disease (Rascol et al. 1989), and loss of the optokinetic nystagmus and failure to suppress the vestibulo-ocular reflex have been described in olive-ponto-cerebellar atrophy (Duvoisin 1987).

Initial slowing of saccadic velocity gives way to hypometric saccades, sometimes with square wave jerks (Troost and Daroff 1977; Rivaus-Pechoux et al. 2000). Pursuit at this stage is normal. However, as the disease progresses vertical smooth pursuit becomes impaired and saccadic. By definition, reflex eye movements to oculocephalic testing (Doll's eye movements) are preserved. However, it may be difficult to extend and flex the very rigid neck. With time, horizontal eye movements become involved, with saccades becoming abnormal before pursuit, and with preserved oculocephalic reflexes (Dix et al. 1971; Rivaus-Pechoux et al. 2000).

The nystagmus elicited by caloric testing is replaced by a tonic conjugate ocular deviation in the direction of the anticipated slow component (Dix et al. 1971). Vestibulo-ocular reflexes may be lost, suggesting additional nuclear involvement (Ishino et al. 1973). A few patients develop an inter-nuclear opthalmoplegia (Mastaglia and Grainger 1975), and periodic alternating nystagmus also has been reported (Perkin et al. 1978). Other oculomotor and eyelid abnormalities have been described (Troost and Daroff 1977; Kristensen 1985; Lepore et al. 1988), and these are summarized in Table 7.12.

Because of the critical significance of a supranuclear vertical gaze palsy for the diagnosis of PSP, strict criteria have been proposed (Golbe and Davis 1993). These authors suggest that voluntary downward gaze should be less that 15% with preservation of oculocephalic reflexes (except in the terminal phases of the illness), or the following abnormalities should be present: (1) hesitancy of voluntary vertical gaze, (2) impaired optokinetic nystagmus with the stimulus moving downwards, and (3) poor suppression of vertical vestibulo-ocular reflexes.

However, eye movement abnormalities may appear only late in the course of the illness (Pfaffenbach et al. 1972; Perkin et al. 1978; Dubas et al. 1983; Davis et al. 1985; Kleinschmidt-DeMasters 1989). Indeed, a few cases have been reported with a pathological diagnosis of PSP who had no eye movement abnormality in life (Probst 1977; Jellinger et al. 1980; Dubas et al. 1983; Kida et al. 1992; Collins et al. 1995; Daniel et al. 1995). Daniel et al. (1995) described an unselected series of 17 cases of pathologically verified cases only seven of whom had a typical supranuclear gaze palsy. Nine of the other 10 cases had been misdiagnosed as having Parkinson's disease in life; the remaining case was diagnosed as having AD.

Nevertheless, a complete supranuclear opthalmoplegia with eventual involvement of lateral gaze appears in at least half of patients. A small number of cases finally develop a nuclear ophthalmoplegia with loss of oculocephalic responses (Blumental and Miller 1969).

Disorders of the eyelid control are common in PSP (Grandas and Esteban 1994). Lid retraction has been mentioned, and blink rate is reduced. Globe et al. (1989) found that blink rates were decreased more in PSP than in Parkinson's disease. A number of patients develop levator inhibition (Lepore and Duvoisin 1985) (apraxia of eyelid opening) with vigorous frontalis contraction (Jankovic 1984). Others develop apraxia of eyelid closure with complete inability to close the eyes to command, but preservation of reflex closure on blinking or sleeping. Blepharospasm has been reported to occur in as many as 10–28% of patients with this condition (Brusa et al. 1980; Jankovic et al. 1990). Friedman et al. (1992) noted that 29% of patients with PSP have blepharospasm, and over a third have apraxia of eyelid opening, eye closure, or both.

A pseudobulbar palsy develops in almost all patients and may be the earliest symptom. The dysarthria characteristically combines features of that seen in parkinsonism with that associated with spasticity. This give a characteristic strained, strangled quality along with hypophonia (Kluin et al. 1993). Speech worsens and may become unintelligible. Tongue movements are slow and facial and jaw jerks are pathologically brisk. Dysphagia also can occur early on and tends to progress. Drooling can become prominent and in the advanced stage of the disease patients frequently choke on their own saliva and develop explosive coughing. Sonies (1992) investigated swallowing in 22 patients with PSP. All had problems with both the oral and pharyngeal phases of swallowing. Criocopharyngeal dysfunction and recurrent pneumonia may occur (Schleider and Nagurney 1977), as may abnormalities of respiratory rhythm (Perkin et al. 1978; Maher and Lees 1986). Many patients eventually require a semi-liquid diet, often administered by a nasogastric tube or a gastrostomy.

Uncontrollable crying and laughter are common, as part of the pseudobulbar palsy, and such emotional incontinence may add to the impression of cognitive impairment. Stuttering, palilalia, and echolalia have also been mentioned, and forced singing, humming, groaning, grunting, or crying occasionally occur (Jankovic et al. 1990).

The extent and distribution of the akinetic–rigid features of PSP are variable. At the one extreme, a minority of patients with the condition can appear very similar to patients with Parkinson's disease. Agid et al. (1986) reported that 7% of patients with a diagnosis of parkinsonism turned out to have PSP. Likewise, Duvoisin et al. (1987) said that about 12% of patients diagnosed on clinical grounds as having parkinsonism eventually turned out to have PSP. Duvoisin et al. (1987) considered the presence of bradykinesia as essential to the diagnosis of the latter condition, but Brusa et al. (1979), in a review of the clinical features of 75 published pathologically proven cases, mentioned bradykinesia in only 22% of patients in the first year from the onset of symptoms.

Nevertheless, most patients have some elements of the akinetic–rigid syndrome. In particular, rigidity in the neck, trunk, and around the shoulders is common, although the distal limbs may be relatively spared. Akinesia of whole-body movement with an expressionless face is typical. However, akinesia of distal hand and foot movements may be minimal or even absent. Frequently the body is not flexed as in true parkinsonism and an extended neck is typical. The gait often is not short-stepped and associated movements of the arms may be present. However, equilibrium is severely impaired, often early in the illness, with a marked tendency to unexplained falls without protective or rescue reactions. Many patients therefore injure themselves. Patients frequently fall backwards. Locomotion often is on a broad base. Although the gait in PSP may not resemble that of typical parkinsonism early in the illness, later on shortness of stride, shuffling, festination, and freezing may occur.

Lees (1987) highlighted an isolated akinetic–rigid syndrome with limited or no response to levodopa. Imai et al. (1986), Yamamoto et al. (1991), and Matsuo et al. (1991) described pathologically proven cases with ‘pure akinesia’, at least early in the course of the illness. Riley et al. (1994) described five patients with a stereotyped syndrome of progressive akinesia of gait, speech, and handwriting without rigidity, tremor, or dementia, which did not respond to levodopa. Although four of the five patients had abnormalities of eyelid control (blepharospasm or freezing of eyelid motion), none had a gaze palsy on clinical examination. However, two cases with the longest duration of disease had slow or small vertical saccades on eye movement recordings.

Classical parkinsonian rest tremor is extremely uncommon in PSP, but occasionally occurs (Gearing et al. 1994). A postural tremor of the outstretched hands is more frequent (Masucci and Kurtzke 1989), and was described in nearly half of the cases reported by Daniel et al. (1995).

Besides the extended neck, dystonia may occur elsewhere. Anzil (1969) described prominent trismus. Rafal and Friedman (1987) reported limb dystonia in eight out of 30 patients with PSP. Leger et al. (1987) described limb dystonia as an early sign of the illness. Rivest et al. (1990) noted the presence of dystonia in only 15 of 118 (13%) pathologically proven cases of the disease.

Muscle power and bulk usually are normal. The tendon reflexes often are pathologically brisk, and the plantar responses frequently become extensor later in the illness. Sensation is normal and there is no evidence of a peripheral neuropathy, except in rare cases. Apart from disequilibrium and a broad-based gait, there is little evidence of cerebellar ataxia. Ideomotor apraxia, when tested for, may be found in some patients (Collins et al. 1995). Myoclonus and chorea have been noted in occasional cases (Collins et al. 1995). Sense of smell has been reported to be normal (Doty et al. 1993).

Patients with PSP frequently have behavioural and cognitive problems. They complain of forgetfulness and their families often state that they have become inattentive, disinterested, and indifferent to their surroundings. Mental responses may seem slow, and the patient often appears to have lost the capacity to take initiative or make decisions. This constellation of behavioural changes is very suggestive of the abulic syndrome seen with frontal lobe or basal ganglia damage. It may give a superficial impression of dementia, but while cognitive processes may be slow they often are accurate given sufficient time. Dementia was described in seven of the original patients reported by Steele et al. (1964) but is now considered to be less common. However, occasionally patients may exhibit early severe dementia (Davis et al. 1985). Amongst the seven cases with a supranuclear gaze palsy described by Daniel et al. (1995), all were said to become demented.

In 1974, Albert et al. (1974) introduced the term subcortical dementia to describe this characteristic profile of the mental status in PSP. They emphasized the slowness of thought (bradyphrenia), forgetfulness, changes in personality, with apathy or depression, and an impaired ability to manipulate acquired knowledge. They attributed this to the subcortical lesions of PSP, and drew attention to the difference between this clinical syndrome and the cortical dementia characteristic of conditions such as AD.

In an extensive neuropsychological assessment of 27 patients with PSP, Maher et al. (1985) emphasized that most of the patients exhibited abnormalities on tests relatively specific for frontal lobe function, such as verbal fluency and the Weigl test. Of the 27 patients, only nine were considered to have severe dementia at the time of first referral, nine others had mild to moderate dementia, but the remainder appeared unimpaired. These authors also emphasized other frontal lobe signs such as forced grasping, utilization behaviour, and marked motor perseveration. The latter can be demonstrated by the ‘applause’ sign reported by Dubois et al. (2005). The patient is asked to clap three times; however, those with PSP carry on clapping more then three times and sometimes indefinitely due to motor perseveration characteristic of the disorder. This test according to the report helps differentiate PSP from Parkinson's disease and FTDP (Dubois et al. 2005). Cambier et al. (1985) noted frontal lobe abnormalities in 10 patients, with mental slowing, impaired attention, poor abstract thinking and reasoning, reduced verbal fluency, mild to moderate memory loss, and impaired linguistic abilities. Pillon et al. (1986) compared the neuropsychological abnormalities in PSP with those in Parkinson's disease and AD. Tests sensitive to frontal lobe dysfunction were more abnormal in the PSP patients, whereas tests of verbal and logical memory were more affected in AD. No differences were detected between the three groups on language tests, calculation, apraxia, and visuomotor activity (Doty et al. 1993).

The overall picture of the cognitive impairments in PSP is that these most commonly involve cognitive slowing, with particular difficulties on tests sensitive to frontal lobe dysfunction; there is relative preservation of memory, language, and praxis. These cognitive deficits may progress to a more pervasive multifocal dementia syndrome in a minority of cases. There are, however, occasional patients with PSP, some of whom with pathological proof, who present with a true dementia and are diagnosed in life as AD or some similar condition (Davis et al. 1985; Masliah et al. 1991; Olsen et al. 1992).

Morris et al. (2002) identified two groups of patients with clinically typical and atypical pathologically diagnosed PSP and investigated their genetic and molecular pathological characteristics. The clinically atypical PSP group contained a number of different clinical syndromes, including an l-dopa unresponsive bradykinetic syndrome and a clinical syndrome closely resembling idiopathic Parkinson's disease. Williams et al. (2005) further refined this concept by describing two broad clinical variants of PSP, one of which resembled Parkinson's disease at onset with much more involvement of the limbs without falls early in the first year of the disease onset. They called this type PSP-P (parkinsonian type of PSP). The other variety was called the Richardson type and had the more typical presentation of PSP with early falls, a supranuclear gaze palsy, and axial symptoms. For instance, speech difficulty, axial rigidity, and gait and postural problems were more prominent in the latter (Williams et al. 2005). These reports may explain the clinical discrepancies, mentioned earlier, noted by various observers regarding the different manifestations of PSP, and the clinical overlap with Parkinson's disease in some cases with regard to the initial presentation. There also appears to be genotypic and pathological differences between the atypical and typical presentation. Those with clinically typical PSP were more likely to have the PSP susceptibility genotype and to have the deposition of PSP-type hyperphosphorylated tau protein (Morris et al. 2002). The clinically atypical PSP group were less likely to have the PSP susceptibility genotype.

Many of the characteristic clinical features of PSP can be correlated with the known anatomical and biochemical pathology of the illness. For instance, loss of cholinergic neurons in the rostral interstitial nucleus of the medial longitudinal fasciculus, which projects to the third nerve complex, disrupts premotor control particularly for vertical eye movements (Buttner-Ennever and Buttner 1982). Lesions in the lower pontine reticular formation likewise may disrupt eye movement. Damage to the nucleus papillioformis disrupts communication from the frontal eye fields, superior colliculus, and pretectum with the cerebellum. Lesions in the nucleus pontis centralis caudalis are associated with defective premotor control of horizontal saccades (Pierrot-Deseilligny et al. 1989). Lesions of cholinergic neurons in the interstitial nucleus of Cajal and in the pedunculopontine nucleus might contribute to the abnormalities of posture, gait, and equilibrium in PSP. Loss of cholinergic neurons in the deep layers of the superior colliculus might affect co-ordination of simultaneous neck and gaze-related movements, altering axial tone and causing gaze abnormalities (Troost and Daroff 1977). Degeneration of the nigrostriatal dopaminergic pathway with loss of striatal dopaimine content is thought to be the substrate of parkinsonism in PSP, with additional contributions from damage to the globus pallidus and subthalamic nucleus. The cognitive and neuro-behavioural changes may be due to the combination of cortical tangle formation along with damage to the substantia innominata with cortical acetylcholine depletion.

Tau pathology is common in the olfactory bulb of AD and Lewy body disease but is minimal or absent in PSP and CBD (Tsuboi et al. 2003). Not surprisingly, marked olfactory dysfunction occurs in Parkinson's disease and dementia with Lewy bodies (and occurs in AD) but is not found in PSP (and CBD) (Katzenschlager and Lees 2004).

Finally, to summarize, Table 7.13 outlines the major clinical differences between PSP and idiopathic PD (Tolosa et al. 1994).

Caparros-Lefebvre and Elbaz (1999) reported an excess of cases of atypical parkinsonism and those with a PSP phenotype in Guadeloupe. Of the 87 patients seen by them at one centre within 2 years, 22 were classified as having Parkinson's disease, 31 were said to have PSP and 30 other atypical parkinsonism, and four had atypical parkinsonism associated with motor neuron disease. Forty four of the patients with PSP or atypical parkinsonism were male. Those with atypical parkinsonism (apart from PSP) were said to have a ‘symmetrical rigidity and bradykinesia and no levodopa peak-dose dyskinesias’. Patients with PSP had typical supranuclear vertical down-gaze palsy, severe gait and balance problems, and frontal-lobe syndrome. They found a statistically significant association with consumption of tropical fruits (pawpaw) or herbal teas (boldo) in 31 patients with PSP and 30 patients with atypical parkinsonism compared with controls. It has been suggested that chronic exposure to the tetrahydroisoquinolones (TIQs), present in these herbal teas and fruits could be linked to developing PSP and atypical parkinsonism (Caparros-Lefebvre and Elbaz 1999; Caparros-Lefebvre 2001). Biochemical analysis of the central nervous system of three of those atypical parkinsonian patients with a PSP phenotype who died showed a 4-R tauopathy as seen in classical PSP.

Dubinsky and Jankovic (1987) put forward the view that up to a third of patients with clinically diagnosed PSP could have a vascular aetiology. The same group (Winikatis and Jankovic 1994) subsequently analysed 128 consecutive patients with a clinical diagnosis of PSP. The diagnosis was based on onset after the age of 40 years, bradykinesia, axial rigidity, gait disorder, postural instability, and little or no tremor; pseudo bulbar signs including dysarthria, dysphagia, and emotional lability, and marked vertical supranuclear gaze palsy with inability to move the eyes more than 50% below the horizontal (Jankovic et al. 1990). The presence of vascular disease was analysed by a rating scale. Two points were given for pathologically or angiographically proven vascular disease, one point for the combination of both vascular disease and neurodegenerative disease, one point each for onset of symptoms within 1 month of a clinical stroke, a history of two or more risk factors for stroke, and neuro-imaging evidence of diffuse vascular disease or vascular disease in two or more vascular territories. This yielded a vascular score ranging from 0 to a maximum of 6. The authors considered that a vascular score of two or more placed the patient in the vascular group. Of their 121 patients, 30 (23.3%) satisfied these criteria for vascular PSP. The remainder were considered to be idiopathic PSP. The only two clinical features distinguishing the vascular group from the idiopathic group were asymmetry of clinical signs and predominance of lower body involvement in the former. Two additional features, presence of pyramidal signs and pseudobulbar signs, were also nearly twice as common in the vascular group compared with the idiopathic group, but these differences did not reach statistical significance. Both groups were of similar age, duration of symptoms, and sex ratio. The authors concluded that whilst the majority of patients with PSP have a neurodegenerative disease, a significant percentage have clinical PSP secondary to cerebrovascular disease.

This study emphasizes that a clinical picture of a PSP involving vertical gaze associated with pseudobulbar and extrapyramidal signs can be produced by cerebrovascular disease. While we would agree that this clinical phenotype can be produced by vascular disease, the clinical history usually differs from that of PSP. The latter is an insidious and slowly progressive condition, while those with a similar phenotype due to stroke often have an acute or stuttering onset. It is well known that strokes or Binswangers subcortical arteriosclerotic encephalopathy may cause pseudobulbar palsy, and lower-half parkinsonism, and impaired eye movements, including vertical gaze, if critical brainstem areas are involved. On these grounds, we believe that most cases of vascular PSP can be distinguished from typical PSP.

To improve the specificity and sensitivity of the clinical diagnosis of PSP SROD, the NINDS, and the Society for PSP, Inc. (SPSP) sponsored an international workshop to develop an accurate and universally accepted set of criteria for this disorder. The NINDS-SPSP criteria, which were formulated from an extensive review of the literature, comparison with other previously published sets of criteria, and the consensus of experts, were validated on a clinical data set from autopsy-confirmed cases of PSP. The criteria specify three degrees of diagnostic certainty: possible PSP, probable PSP, and definite PSP. Possible PSP requires the presence of a gradually progressive disorder with onset at age 40 or later, either vertical supranuclear gaze palsy, or both slowing of vertical saccades and prominent postural instability with falls in the first year of onset, as well as no evidence of other diseases that could explain these features. Probable PSP requires vertical supranuclear gaze palsy, prominent postural instability, and falls in the first year of onset, as well as the other features of possible PSP (Tables 7.6. and 7.7). Definite PSP requires a history of probable or possible PSP and histopathologic evidence of typical PSP. Validity of these clinical diagnostic criteria has been demonstrated. The criteria for probable PSP are highly specific (100%), making them suitable for therapeutic, analytic epidemiologic, and biologic studies, but not very sensitive (83%). The criteria for possible PSP are substantially sensitive, making them suitable for descriptive epidemiologic studies, but less specific (Litvan et al. 1996b).

Despite the good clinical criteria for PSP, the literature continues to grow with case reports of alternative pathologic diagnoses in patients considered to have a clinical diagnosis of PSP. Some of the disorders misdiagnosed as PSP include: vascular disease (Winakates and Jankovic 1994; Josephs et al. 2002), familial frontotemporal dementia with parkinsonism (Miyamoto et al. 2001; Morris et al. 2003), progressive multifocal leukoencephalopathy (Alafuzoff et al. 1999), CBD (Shiozawa et al. 2000), and postencephalitic parkinsonism (Pramstaller et al. 1996), as well as diffuse Lewy body disease (De Bruin et al. 1992), PD with argyrophilic grains (Seno et al. 2000), and Creutzfeldt–Jacob Disease (CJD) (Josephs et al. 2004). Others include Whipple's disease (Averbuch-Heller et al. 2001), neurosyphilis (Murialdo et al. 2001), and progressive subcortical gliosis (Will et al. 1988) (also see Table 7.11). Difficulty in clinical diagnosis of PSP can occur also when PSP coexists with another neurodegenerative disorder, such as AD.

Josephs and Dickson (2003) compared the clinical and genetic features of pathologically confirmed cases of PSP with misdiagnosed cases to determine ways to improve diagnostic accuracy. They reviewed the medical records of their 180 brain bank cases diagnosed as PSP, all of whom had standardized neuropathologic evaluations as well as determination of apolipoprotein E and tau genotypes. Of the 180 cases studied, 137 had PSP and 43 had other pathologic diagnoses. CBD, multiple system atrophy, and diffuse Lewy body disease accounted for 70% of the misdiagnosed cases. History of tremor, psychosis, dementia, and asymmetric findings were more frequent in misdiagnosed cases. The frequencies of the H1 tau haplotype (93 vs. 80%) and H1H1 genotype (86 vs. 66%) were significantly greater and APOE epsilon 4 carrier state was significantly less (17 vs. 41%) in PSP compared with misdiagnosed cases. They concluded that tremor, psychosis, early dementia, asymmetric findings, absence of H1 haplotype, and presence of APOE epsilon 4 should raise questions about a diagnosis of PSP. However, they pointed out that genetic testing, although useful as a research tool, is not available in clinical practice. They also cautioned that genetic testing for the tau haplotype is not diagnostic by itself as this finding that the H1 tau haplotype is increased not only in PSP but also in CBD precludes this genetic test as a way to differentiate PSP from CBD. It may also not even be useful in differentiating PSP from Parkinson's disease, because some reports have suggested also that the H1 tau haplotype frequency may be increased in Parkinson's disease. However, if tau genotyping showed an H2/H2 genotype, the diagnosis of PSP would be very unlikely.

The usual battery of screening blood tests and chemistry are normal in PSP. So too are routine the cerebrospinal fluid markers. An electroencephalogram may show non-specific abnormalities, and occasionally epileptic seizure activity has been recorded (Su and Goldensohn 1973; Nygaard et al. 1989). Seizures were reported in 7 out of the 62 patients described by Nygaard et al. (1989).

Sleep problems in PSP have been documented by polysomnography (Aldrich et al. 1989). Insomnia is common and worsens with increasing dementia. Abnormalities of motor and sensory evoked potentials have been described by Abbruzzese et al. (1991). Hearing impairment may occur (Dix et al. 1971; Daniel et al. 1995); brainstem auditory evoked responses, however, are usually normal (Tolosa and Zeese 1979). The auditory startle response has been investigated in PSP, because of the profound brainstem pathology involving components of the normal startle reflex. The startle reflex was absent or considerably reduced in such patients, consistent with pathology in the lower pontine reticular formation (Vidialhet et al. 1992). Gironell et al. (2003) assessed the diagnostic potential of acoustic startle reflex (ASR), acoustic blink reflex (ABR), and electro-oculography (EOG) in early stages of atypical parkinsonian syndrome. The three neurophysiological tests investigated provided sensitive and specific measures, with predictive value in early stages of atypical parkinsonian syndrome. Nerve conduction studies usually are normal, although a peripheral neuropathy has been described in two cases (Weinmann 1964, Daniel et al. 1995).

Cutaneous sympathetic function and cardiovascular function have been studied in patients with PSP and Parkinson's disease. The former may be abnormal in PSP, but cardiovascular function is well preserved (Kikkawa et al. 2003).

Brain imaging with CT or MRI can identify characteristic patterns of atrophy in PSP (Figs 7.11 and 7.12), but these are neither specific nor sensitive.

Thus, in earlier studies atrophy of the midbrain was recorded on CT (Masucci et al. 1985; Schonfield et al. 1987) and later on MRI brain scans (Savoiardo 1989; Savoiardo et al. 1994). The third ventricle often was noted to be dilated, and thinning of the quadrigeminal plate could be seen. In some patients there was diffuse supratentorial atrophy. There was no evidence of increased low signal intensity in the lentiform nucleus. MRI studies have revealed increased signal intensity on proton density in the superior cerebellar peduncle (Oka et al. 2001) and confirmed atrophy of the midbrain, as well as the basal ganglia and frontal cortex (Asato et al. 2000; Cordato et al. 2000; Schrag et al. 2000; Warmuth-Metz et al. 2001). Of these, the study by Warmuth-Metz et al. found that patients with PSP had significantly lower midbrain anteroposterior diameter (mean 13.4 mm) than patients with Parkinson's disease (mean 18.5 mm) and control subjects (mean 18.2 mm). They concluded that measurement of anteroposterior diameter of the midbrain on axial T2-weighted MRIs could be used to help differentiate patients with PSP from those with Parkinson's disease (but not from multiple system atrophy patients in the same study). This pattern due to midbrain atrophy has been likened to a hummingbird (Kato et al. 2003) (Fig. 7.13). However, only about half of patients with the diagnosis of PSP exhibit abnormalities on MRI brain scanning in the early stages of the disorder.

A number of studies of cerebral metabolism employing positron emission tomography (PET) and single photon emission computed tomography (SPECT) have been reported in PSP. D’Antona et al. (1985), utilizing ¹⁸F-deoxyglucose, found a reduction of medial and lateral frontal glucose metabolism by around 25–30%. Leenders et al. (1988) measured regional cerebral blood flow using ¹⁵-oxygen and also found a reduction in frontal cerebral metabolism. Since these original studies a number of authors have reported similar findings (Foster et al. 1988; Goffinet et al. 1989; Blin et al. 1990; Johnson et al. 1992; Karbe et al. 1992). The results of these studies are shown in Table 7.14.

All such studies have shown a diffuse fall in cerebral metabolism in PSP disease, with most demonstrating a significant reduction in frontal metabolism. Striatal, thalamic, and cerebellar metabolism are also generally reduced, but occipital metabolism is relatively preserved. The most likely explanation for the hypofrontality in PSP is loss of pallidal input, via thalamus to frontal cortex (Brooks 1994). Most authors have concentrated on the relative lack of marked pathology in the cerebral cortex (Jellinger and Bancher 1992) which tends to be localized to motor areas (Hauw et al. 1990). Blin et al. (1995) re-analyzed their data obtained from 20 patients and found an abnormal linkage between frontal cortex hypometabolism (using ¹⁸F-deoxyglucose) and hypometabolism in the thalamus, the latter being partly coupled to caudate nucleus hypometabolism. They concluded that frontal cortex hypometabolism is linked to thalamic dysfunction.

Striatal dopamine function measured using ¹⁸F-dopa, has consistently shown a significant reduction in tracer uptake in putamen (around 63%) and caudate (around 52% of normal) (Fig. 7.14) (Leenders et al. 1988; Brooks et al. 1990; Bhatt et al. 1991; Taniwaki et al. 1992).

This reflects damage to the dopaminergic nigrostriatal system. The considerable reduction of caudate tracer uptake in PSP contrasts with the relative preservation seen in Parkinson's disease.

Baron et al. (1986) studied striatal D2 receptor density in PSP using ⁷⁶Br-bromospirone. They found a reduction of 24% in the ratio of activity in the striatum compared with the cerebellum. Wienhard et al. (1990) used ¹⁸F-fluoroethylspirone and found a 17% fall in tracer uptake in the caudate. Brooks et al. (1992) used ¹¹C-raclopride and found a reduction in the striatal to cerebellum ratio of uptake of the tracer of 24% in the caudate nucleus and 9% in the putamen. Although mean striatal D2 receptor binding appears to be moderately reduced in PSP, individual patients may show normal levels of uptake of D2 tracers. Similar findings on striatal D2 receptor density have been obtained using ¹²³I-iodobenzamide and SPECT (Arnold et al. 1994). The cerebral distribution of [11C]flumazenil (FMZ), a ligand that binds to the gamma-aminobutyric acid A (GABA_A) receptor, was found to be globally reduced on PET scans in PSP patients, with a significant reduction only found in the anterior cingulate gyrus. Ligands for the cholinergic system have also been investigated. Analogues of vesamicol, an inhibitor of the acetylcholine vesicular transporter, can demonstrate loss of intrinsic striatal cholinergic neurons (Suzuki et al. 2002). Use of carbon-11-labelled N-methyl-4-piperidyl acetate and PET to measure acetylcholinesterase activity showed a preferential loss of cholinergic innervation to the thalamus in PSP compared with Parkinson's disease and controls (Shintoh et al. 1999).

A number of neuropsychological studies have been undertaken in PSP. Specifically, short-term memory has been found to be normal (Litvan et al. 1989). In contrast, the same authors found that long-term memory for verbal material was impaired compared to controls. PSP patients also performed poorly on a visuomotor procedural learning task. In contrast, Kimura et al. (1981), Maher and Lees (1985), and Milberg and Albert (1989) found no significant memory impairment, but others have disputed this conclusion. Pillon et al. (1986, 1991, 1992) compared the performance of patients with PSP with those with AD on several measures of memory. In general, explicit episodic and remote memory were more impaired in AD than in PSP. Patients with PSP also have problems in forming concepts, such as the ability to interpret proverbs (Albert et al. 1974; Cambier et al. 1985; Grafman et al. 1990a). They have difficulties shifting conceptual set in their performance on the Wisconsin Card Sorting and Trail Making Tests (Pillon et al. 1986; Grafman et al. 1990a). They have apparent cognitive slowing when comparing response time in simple reaction time tests with that in complex reaction time paradigms (Dubois et al. 1988). This central cognitive slowing has been confirmed by recording event-related potentials (Johnson et al. 1991, 1992) using an Oddball task. The early components were normal in amplitude and latency, but the P2 and P300 components had considerably increased latencies and decreased amplitude. The remarkably delayed latency of the P300 has not been reported in any other type of dementia.

Pillon et al. (1991, 1992) compared cognitive function in 44 patients with AD, 45 patients with PSP, 35 patients with Huntington's disease, and 164 patients with Parkinson's disease. They defined dementia as a global intellectual performance two standard deviations below mean control values. By this definition, 93% of the Alzheimer patients were demented, compared to 66% of Huntington's disease, 58% of the PSP, but only 18% of the Parkinson's disease patients.

Later, Pillon et al. (1994) compared explicit memory disorders in 15 patients with PSP with those in 15 cases each of Parkinson's disease, AD, and Huntington's disease, in relation to matched controls. Those with PSP showed memory deficits characterized by impaired immediate memory span, disturbed learning and consistency of recall, and an abnormal number of false alarms at recognition, which was dramatically alleviated by controlled cued recall. This pattern was similar to that seen in Parkinson's disease and Huntington's disease, but very different from that in AD, which was characterized by more rapid forgetting and less improvement in the controlled situation. These authors considered the profile of memory disturbance in PSP to be compatible with fronto-striatal dysfunction, rather than to the mild–moderate pathologic lesions in cortex and hippocampus.

Robbins et al. (1994) compared cognitive performance on computerized testing in three paradigms sensitive to frontal lobe dysfunction in 18 patients with PSP, 16 patients with multiple system atrophy, and 24 patients with Parkinson's disease. All showed deficits in planning the Tower of London task, in a test of spatial working memory, and in a task of attentional set shifting. Those with PSP exhibited the greatest deficits. However, there were differences in the exact nature of their impairments, one from another and from others with frontal lobe damage and AD (Table 7.15). These authors also concluded that the cognitive abnormalities of PSP were due to fronto-caudate dysfunction, in contrast to the multifocal dementia characteristic of AD.

Unfortunately there is no specific treatment for PSP. The akinesia and rigidity usually do not respond to levodopa, and it is generally held that a good therapeutic response to levodopa calls into question the diagnosis of the condition. However, some patients may improve early in the course of the illness, although the benefit usually is not sustained (Kiawans and Ringel 1971; Donaldson 1973; Jackson et al. 1983; Agid 1986; Litvan and Chase 1993). Accordingly, a trial of levodopa is worthwhile, although without great expectation of sustained benefit. There also is little response to treatment with directly acting dopamine agonist drugs, although the occasional patient may appear to benefit, usually transiently (Jackson et al. 1983). The failure of dopamine agonist therapy in PSP may be attributed to reduced striatal D2 receptors associated with loss of striatal neurons, and to degeneration of the globus pallidus and output pathways from the basal ganglia. However, there are hardly any formal trials looking at dopaminergic drug treatment in PSP. Kompoliti et al. (1998) reviewed pharmacological response in 12 patients with autopsy-confirmed cases of PSP and found that dopaminergic drugs were not useful and gave side effects, although an initial positive clinical response was detected in nearly two thirds (7/12) of the patients receiving these drugs. Adverse effects included orthostatic hypotension (six patients), hallucinations and delusions (three patients), gastrointestinal complaints (three patients), and dizziness (one patient), but dyskinesias were uncommon and seen in only one patient. Weiner et al. (1999) reported the results of pramipexole treatment (4.5 mg daily) in six patients with PSP (average disease duration 4.4 years). Patients were treated for 2 months. Pramipexole was not efficacious for the symptoms of PSP. Amantidine can be beneficial in some cases (Nieforth and Golbe 1993) and is worth trying, although the response may not be sustained. Because of the considerable damage to cholinergic neurons in PSP, a number of attempts have been made to increase cholinergic function. Treatment with cholinergic agents, such as the muscarinic agonist RS-86 (Foster et al. 1989) or oral physostigmine (Litvan et al. 1989) produced no significant functional improvement in memory and attention or in motor performance. However, motor function has been reported to improve with anticholinergic medication in some patients (Litvan and Chase 1992). Litvan et al. (1994) assessed the effect of increasing doses of an anticholinergic (scopolamine) and a cholinergic (physostigmine) on cognition and motor performance in nine patients. Physostigmine had no neurobehavioural effects at any dose. Low and medium dose scopolamine impaired memory and worsened the gait of some patients (Fig. 7.15); high doses could not be tolerated. These authors concluded that anticholinergic drugs should be avoided in this condition, and that cholinergic agents were ineffective.

There have been suggestions that there may be differential effects of physostigmine on some functions. For example, Blin et al. (1995) reported that following physostigmine administration in PSP patients, errors in antisaccades during ocular movement testing were significantly reduced, and a significant reduction in errors or performance was found in four out of seven neuropsychological tests, although motor disability was not significantly altered. In this regard Frattali et al. (1999) studied the effect of physostigmine on oral function and swallowing and found it had no effect in a double blind placebo controlled study. Overall, cholinergic drugs do not seem to have any significant effect on motor function in PSP. Studies have looked at the efficacy of donepezil, a centrally acting cholinesterase inhibitor in PSP. In one study of six patients no benefit was found with 10 mg donepezil daily on cognitive performance, motor function, and daily living activities (ADL) (Fabbrini et al. 2001). In another double blind placebo controlled study, although memory test scores improved with donepezil, ADL and mobility scores significantly worsened (Litvan et al. 2001). In light of its deleterious effects on ADL/mobility, the authors concluded that donepezil was not recommended for this patient population (Litvan et al. 2001). There have been anecdotal reports of benefit from treatment with tricyclic antidepressants (Newman 1985; Nieforth and Golbe 1993), but therapy with the 5-HT agonist lisuride (Neophytides 1982) and the 5-HT antagonist methysergide (Raphael and Grimm 1971; Paulson 1981) have generally been unsuccessful, although an occasional patient has benefited (Di Trappani et al. 1991). Zolpidem has been reported to produce motor benefit as per a double blind placebo controlled study in 10 PSP cases (Daniele et al. 1999) and improved eye movements and parkinsonism in another separate report of a single case (Mayr et al. 2002).