Introduction

In 1892, Arnold Pick described progressively aphasic patients with behavioral disturbances and frontal lobar atrophy [1]. Subsequent neuropathologic studies emphasized the evidence of focal frontal atrophy along with the presence of round silver-staining inclusions (Pick bodies) and ballooned neurons (Pick cells). Pick’s disease was subsequently used to classify patients with those findings at autopsy, which turned out to be observed in only some cases with clinical frontotemporal dementia (FTD). The paradox of FTD without Pick bodies gave rise to one of the first controversies about the nosology of this disease, in which the gross pattern of lobar atrophy was the main characteristic, but with a wide variety of clinical and pathologic features.

Almost 40 years later, the recognition of an extrapyramidal involvement in FTD patients further complicated the classification of these diseases. Indeed, after the initial description of Pick’s disease, von Braünmuhl et al. reported cases with atrophy extending from frontal lobe to basal ganglia, in particular the caudate nucleus [2]. A few years later, in 1944, Akelaitis described a patient with an early parkinsonian syndrome associated with confusion and apathy, who showed a progressive deterioration of motor status, behaviour, and language, dying demented approximately five years after disease onset [3]. At pathologic examination, atrophy was detected not exclusively in the frontal cortex, but also in the subcortical gray regions, with a severe involvement of the caudate nucleus. For some years, the “Akelaitis variant” was applied to patients who were recognized as having prominent extrapyramidal symptoms associated with FTD. The occurrence of parkinsonism along with an involvement of basal ganglia and nigrostriatal degeneration in Pick’s disease gave rise to a scientific debate, but at the same time promoted the concept of inclusion of different clinical and pathologic entities within the FTD spectrum.

Twenty years after the Akelaitis report, Steele and colleagues provided the description of a clinical syndrome defined by supranuclear ophthalmoplegia, mainly affecting vertical gaze, pseudobulbar palsy, dysarthria, dystonic axial rigidity along with variable cerebellar and pyramidal symptoms, mildly progressive toward dementia, and involving the brainstem [4]. The characterization of progressive supranuclear palsy (PSP) was soon followed by the report of another extrapyramidal condition affecting three patients who had been evaluated by Rebeiz and colleagues at the Massachusetts General Hospital [5]. The “strange” disorder that “defied all attempts at exact diagnosis and effective treatment” was named as corticodentatonigral degeneration with neuronal achromasia, later known as corticobasal degeneration syndrome (CBDS) [6]. The first reports of CBDS patients outlined the combination of a severe motor impairment with cerebral cortical deficits. In particular, the initial asymmetric awkwardness and involuntary movements, variably involving rigidity, tremor, and dystonia, also included language deficits, apraxia, alien limb, and cortical sensory loss. Autopsy examination revealed strikingly asymmetric frontal and parietal atrophy with involvement of white matter, as well as basal ganglia, including the substantia nigra and the subthalamic nucleus. In the Rebeiz et al. paper [7], PSP was suggested to share some clinical features with CBDS cases: the similar age and the slowly progressive course of a movement disorder with cognitive impairment. However, CBDS exhibited neuropathology primarily in the cerebral frontal and parietal cortices, while PSP showed major brainstem involvement with a lesser degree of frontal cortical involvement.

Some early investigators included extrapyramidal cases within the spectrum of FTD. Costantinidis and colleagues proposed that the “variant B” with ballooned neurons had the most frequent extrapyramidal symptoms [8]. In a comparative clinical and pathologic investigation between cases with classic Pick’s disease and a “generalized” variant, which included cases with cortical as well as subcortical atrophy, Munoz-Garcia and Ludwin distinguished the two series on the basis of inclusion type, and age of onset, being younger in the “generalized” variant [9]. Although Neary et al. [10] and the Lund and Manchester Groups [11] did not consider extrapyramidal variants in their original criteria, the increasing evidence that PSP, CBDS, and extrapyramidal syndromes shared clinical and pathologic overlap with FTD suggested a more integrated approach, eventually prompting their inclusion in consensus papers [10, 12].

An extensive clinicopathologic study of a prospective, clinic-based large cohort of patients with autopsy examination demonstrated how the first most common clinical presentation, in particular the behavioral variant of FTD (bvFTD), could be followed during disease progression by other syndromes, variably represented by primary progressive aphasias (PPAs), motor neuron disease (MND), or CBDS [13]. In analogy to bvFTD, also PPAs, CBDS, or PSP could evolve towards different clinical conditions [13] (see Chapter 2 for more detail). Indeed, the clinical features poorly predicted the histopathologic characteristics, supporting the view of a wide heterogeneity in the FTD spectrum.

The great advances in genetic investigations further contributed in the identification of the common determinants among PSP, CBDS, and extrapyramidal components in FTD [14]. Concomitantly, the growing knowledge of neuropathologic hallmarks has allowed the scientific community to highlight the specific features of these disorders, and to refer to PSP and corticobasal degeneration (CBD) as distinctive neuropathologic entities within the frontotemporal lobar degeneration (FTLD) spectrum.

Neuropathology and genetics of PSP and CBD

The original description by Steele et al. reported as the main pathologic features of PSP to be the presence of cell loss, gliosis, neurofibrillary tangles, granulovacuolar degeneration, and demyelination in different regions of basal ganglia, brainstem, and cerebellum [4]. The most severely affected regions were the globus pallidus, subthalamic nucleus, red nucleus, substantia nigra, superior colliculi, nuclei cuneiformis and subcuneiformis, periaqueductal gray matter, pontine tegmentum, and dentate nucleus, with a general preservation of cerebral cortex. On the contrary, the relevant findings by Rebeiz et al. in the cases of CBD included convolutional atrophy of frontal and parietal regions, with a pronounced asymmetric involvement, and consistently affecting some subcortical nuclei, including the substantia nigra, the subthalamic nucleus, and the dentate and roof nuclei of the cerebellum [7]. In CBD, astrocytic gliosis accompanied the loss of neurons; some astrocytes exhibited a peculiar pallor of staining named achromasia.

At present, the distinctive neuropathologic features of PSP include globose neurofibrillary tangles, oligodendroglial coiled bodies, tuft-shaped astrocytes, and thread-like processes, while CBD includes pre-tangle lesions, astrocytic plaques, and ballooned neurons [15–17]. The structural elements of these different morphologic entities are known to be determined by deposits of insoluble and hyperphosphorylated tau protein, which is invariably detected in PSP as well as CBD. Because of the sharing of common deposition of tau in PSP and CBD, the two pathologies are grouped within FTD-tau or tauopathies [14].

Tau is a protein specific to human brain tissue, present in neurons as well as glia, and known to promoting the assembly, spatial organization, and disassembly of microtubules in axons of neurons [18]. It is composed of 352–441 amino acids, and contains a characteristic tandem repeat region in its carboxyl-terminal half [19]. Tau is developmentally regulated, with six isoforms that differ by a 29–58-amino acid insert at the N-terminus, and a 31-amino acid repeat located in the C-terminus. The tandem repeat region constitutes the microtubule-binding domains of tau. The inclusion of the C-terminus repeat generates isoforms with 4 tandem repeats (4R), the others having 3 tandem repeats (3R). In Alzheimer’s disease (AD), tau constitutes the core of paired helical filaments, the major components of neurofibrillary tangle. In tauopathies, 3R tau predominates in Pick’s disease, whereas 4R tau is most prevalent in PSP and CBD. As one pathogenic mechanism, the 4R isoforms have a lower ability to promote microtubule assembly than 3R tau, thus directly affecting tau/tubulin interaction [20]. Although PSP and CBD pathologies are commonly sustained by an abnormal deposition of tau, the different anatomical distribution of 4R deposits in PSP and CBD is thought to relate to the wide and overlapping spectrum of clinical presentations. Interestingly, a subtle alternative biochemical proteolytic process of tau, differently affecting PSP and CBD, has been claimed as a possible mechanism to distinguish the two pathologies [21]. Genetic analyses further contributed to demonstrate the presence of common determinants in PSP and CBD, in particular the discovery of missense and 5′ splice-site mutations in the MAPT gene causing genetic cases of PSP and more rarely CBD (see Alzheimer Disease & Frontotemporal Dementia Database http://www.molgen.ua.ac.be/FTDmutations/ and Table 7.1). Moreover, the MAPT H1 haplotype is a common genetic risk factor for PSP and CBD [22, 23], with specific polymorphisms associated to the diseases [24, 25]. Also from a genome-wide association study, MAPT was confirmed as strictly influencing both pathologies [26].

Although PSP and CBD are usually sporadic, occasional reports of familial aggregation, some with an autosomal dominant pattern of inheritance, may provide clues relevant to understanding the pathogenesis of these diseases [27, 28].

Clinical phenotypes associated with PSP and CBD

PSP and CBD are considered rare neurodegenerative disorders [29], and for this reason for many years only anecdotal cases describing clinical features have been available. However, the knowledge about the clinical features associated with PSP and CBD neuropathologic entities is rapidly expanding, and in the last years new diagnostic criteria have been proposed to highlight distinctive hallmarks able to predict the neuropathology [30, 31].

PSP and CBD were originally defined as atypical extrapyramidal syndromes with spared cortical functions. The most recent literature contradicted this point and has demonstrated that cognitive disturbances, and deficits of executive functions in particular, are more common than previously thought. These aspects have supported the inclusion of these conditions in the FTLD spectrum, and have further complicated the diagnostic process [13].

Indeed, one of the most intriguing issues at present is to define the genotype–phenotype relationship, as mismatch between clinical features and neuropathologic characteristics exist.

Recent clinicopathologic studies have led to the appreciation of several clinical syndromes associated with the pathology of PSP and CBD with initial clinical presentations more consistent with PPA or FTD [32]. Conversely, clinical features resembling PSP or CBD may be the initial presentation of other neuropathologic diagnoses, such as FTD with tau or transactive response DNA-binding protein 43 (TDP-43) inclusions, AD, α-synucleinopathies, and spongiform encephalopathy [33–35].

The early clinical features of either PSP or CBD are often subtle and can be difficult to discern [36]. To increase diagnostic accuracy, a careful characterization of the different clinical syndromes associated with PSP and CBD pathology has been proposed.

PSP current criteria and clinical phenotypes. In PSP, early diagnosis is a challenge, with fewer than half of patients receiving an accurate diagnosis initially; as many as 20% of patients will have a different diagnosis at the time of death [37–39].

There are many proposed clinical criteria for PSP, compiled to identify PSP-tau pathology, but the majority of them have a similar characteristic: high specificity and low sensitivity [31]. The National Institute of Neurological Disorders and Stroke/Society for Progressive Supranuclear Palsy (NINDS-SPSP) criteria state that early falls due to postural instability and supranuclear gaze palsy or slowed vertical saccades are the most helpful defining clinical features [31]. Along with current diagnostic criteria, a more detailed description of clinical phenotypes associated with PSP-tau has been proposed, to increase diagnostic accuracy.

The classical and more common phenotype has been termed Richardson’s syndrome (PSP-RS), after J. C. Richardson’s original description along with J. Steele and J. Olszewski [4]. It is characterized by lurching gait and unexplained falls backwards without loss of consciousness. Patients usually present with axial rigidity, and retrocollis is a common form of dystonia. Ocular symptoms give the syndrome its name, as vertical supranuclear gaze palsy is the definitive diagnostic feature even though this commonly develops many years after disease onset. Slow saccades and involuntary eyelid closure due to blepharospasm may be adjunctive features. In PSP-RS, cognitive and behavioral deficits are often observed, including impairment of executive functions, aphasia or apraxia of speech, or apathy. On exam, it is often possible to elicit the “applause sign,” an indicator of impaired motor control leading to perseveration (repeated clapping despite instruction to clap only once) due to frontal dysfunction. Apraxia, even if less common than in CBD cases, and deficits of visuospatial functions are frequently observed. Apathy is the most common behavioral abnormality. Most patients become dependent on others for care within three to four years of diagnosis and the common causes of death are aspiration pneumonia due to dysphagia, primary neurogenic respiratory failure, and complications due to hip fractures.

PSP-parkinsonism (PSP-P), the second most common clinical phenotype, is characterized by bradykinesia, rigidity, and sometimes tremor, which tends to be asymmetric and at least modestly responsive to levodopa therapy. Parkinsonism dominates the early clinical picture and a jerky postural tremor and even a 4–6 Hz rest tremor are common in PSP-P patients. Falls and cognitive impairment occur later in PSP-P than in PSP-RS. However, after six years from the onset, PSP-RS and PSP-P tend to overlap and to be indistinguishable.

Another variant, known as pure akinesia with gait freezing (PSP-PAGF), has hypophonia and micrographia along with gait disturbance and later gait freezing. Usually, neither tremor, rigidity, dementia, nor eye movement abnormality are observed during the first five years of the disease.

Progressive apraxia of speech typically evolves into progressive non-fluent aphasia (PNFA), and this PSP-PNFA variant may have no aspects of classic PSP, at least at presentation. The clinical presentation is consistent with non-fluent spontaneous speech, hesitancy, phonemic errors, and agrammatism. Motor features of PSP may develop later.

The classic corticobasal syndrome (CBS) has also been described in association with PSP pathology, and is termed PSP-CBS.

CBD current criteria and clinical phenotypes. The relationship between CBD neuropathology and clinical phenotype is one of the most challenging issues in the neurodegenerative disease field, since there is such variable correspondence between autopsy findings and the clinical picture. Early literature on CBD suggested that it was a distinct clinic pathologic entity; however, over the last years there is a strong body of literature indicating considerable heterogeneity [40]. Hitherto, we currently referred to the classical clinical picture as CBS, and to the pathologic entity as CBD [41].

Previously suggested criteria for CBD [42–47] have outlined the clinical features now labeled CBS, reflecting an asymmetric movement disorder presentation combined with lateralized higher cortical features. The core features include basal ganglia dysfunction as reflected by a varying combination of stiffness, clumsiness, and dystonia, with lack of sustained levodopa response, and cortical dysfunction characterized by ideomotor apraxia, alien limb phenomenon, cortical sensory loss, visual or sensory hemineglect, myoclonus, and language deficits. However, recent observations have highlighted that cognitive impairment is a common presenting clinical feature [47].

Different neuropathologic series have demonstrated that a relatively small proportion of cases with a CBS presentation turn out to be CBD at autopsy [41, 48–50], suggesting that current criteria are not able to identify CBD with high accuracy. To address this problem, a recent consensus conference has developed a set of new research criteria [30]. On the bases of data derived from brain-bank cases and a critical literature review, five phenotypes have been described as being most frequently associated with CBD pathology: classical CBS, PSP syndrome (PSPS, also called Richardson’s syndrome), FTD, AD-like dementia, and PNFA [30].

In this view, in order to maximize the chance of diagnosing classic CBD without contamination from other pathologies, the new criteria included cases with insidious onset and gradual progression lasting more than one year in people with symptom onset occurring after the age of 50, presenting as classical CBS but including features resembling frontal behavioral-spatial syndromes and non-fluent/agrammatic variant of primary progressive aphasia. Possible CBD criteria allowed the inclusion of PSPS phenotype.

However, it has also become evident that a substantial number of cases presenting as CBS do not have CBD pathology at autopsy. Because of this, stringent exclusion criteria have been developed to exclude Lewy body disease (i.e., classic 4 Hz Parkinson’s disease resting tremor, excellent and sustained levodopa response, or hallucinations), multiple system atrophy (i.e., dysautonomia or prominent cerebellar signs), amyotrophic lateral sclerosis (i.e., presence of both upper and lower motor neuron signs), non-tau pathology (i.e., granulin mutation or reduced plasma progranulin levels, TDP-43 mutations, FUS mutations), and AD (i.e., low cerebrospinal fluid [CSF] amyloid-β (Aβ) and high CSF tau levels or positive amyloid PET scan, even though this will exclude some cases of CBD with coexisting amyloid) [30].

The proposed criteria will obviously need prospective study and continued revision and the development of new biomarkers will likely be key to the identification of CBD patients during life.

Brain imaging in PSP and CBD

Conventional structural neuroimaging has been proposed as an adjunctive tool to corroborate diagnosis in PSP and in CBS/CBD. However, the main limitation of many of these studies is the lack of autopsy confirmation.

In PSP, it is indeed widely established that the classical finding is represented by the “penguin/hummingbird sign,” given by the atrophy of the rostral midbrain tegmentum, the pontine base, and the cerebellum [51]. This finding appears highly specific for PSP, when disease is overt, present in 75% of cases, even if not pathognomonic for the disorder [51]. The areas of midbrain, pons, and cerebellar peduncles have also been used for quantitative analyses of regional atrophy, showing high sensitivity and specificity [52–55]. More recently, an automated computer classification based on MRI pattern analysis in parkinsonian syndromes was carried out, reporting 91% accuracy, 88% specificity, and 93% sensitivity in diagnosing PSP [56].

Advanced structural neuroimaging techniques, such as voxel-based morphometry and diffusion tensor imaging, have been used at group level and have helped in elucidating the clinical aspects associated with PSP. PSP is indeed characterized by involvement of brainstem and basal ganglia regions, but a reduced gray matter density in fronto-insular cortices has been reported, explaining the overlap with FTD symptoms in these patients [57–59]. Interestingly, these quantitative approaches can support a better definition of the anatomical correlates of the different entities in the clinical spectrum of PSP. PSP-RS is defined by more severe involvement of infratentorial white matter tracts and thalamic radiations, which are relatively spared in PSP-P [60, 61].

One of the ultimate goals of advanced neuroimaging techniques is their application at the single-subject level. A first study conducted using a support vector machine approach demonstrated 90% sensitivity and 100% specificity in the diagnosis of PSP compared with Parkinson’s disease and appears to be a very promising tool for future evaluation [62, 63].

The neuroimaging study of CBD and CBS is still challenging, since clinicopathologic correspondence is so poor [49, 64]. An autopsy study evaluated the neuroimaging features of either neuropathologically confirmed CBD patients or clinically diagnosed CBS patients with known histopathology [65]. CBD patients showed a common pattern of atrophy, involving left perirolandic and dorsal prefrontal cortices and striatum. However, even patients with CBS showed perirolandic atrophy irrespective of underlying pathology. From this perspective, perirolandic dysfunction, even if not specific for a single underlying pathology, is the fingerprint of both CBS and CBD [65, 66]. In addition, the analysis of white matter bundles in CBS supported this concept, showing a predominant involvement of long fronto-parietal associative fibers as well as sensorimotor projections of the cortical hand areas [67, 68].

Furthermore, an interesting study reported that the involvement of brain regions beyond the perirolandic region may predict the underlying pathology in CBS patients. In clinically defined CBS with CBD at autopsy, atrophy included also prefrontal cortex, whereas when brainstem and subcortical atrophy is more evident CBS-PSP could be suspected; finally, the extension of the pathology to posterior regions such as precuneus and temporo-parietal lobes are indicative of atypical presentation of AD pathology [65, 69, 70].

Along with structural neuroimaging, functional neuroimaging has been used to try to increase diagnostic accuracy. Dopamine receptor (DAT) imaging and [¹⁸F]FDG-PET (fluorodeoxyglucose positron emission tomography) are promising tools in distinguishing PSP among parkinsonian disorders. Compared with Parkinson’s disease, PSP patients showed more pronounced DAT losses in the anterior caudate, and anterior caudate/ventral striatum ratio may help in differentiating PSP from Parkinson’s disease [71]. Furthermore, [¹⁸F]FDG-PET demonstrated metabolic differences within the clinical spectrum of PSP, with more evident fronto-thalamic impairment in PSP-RS than PSP-P [72].

When considering CBS/CBD, data are more heterogeneous, as once again the lack of autopsy confirmation of the CBS patients. Indeed, some CBS patients had DAT-positive, others DAT-negative findings [73]; thus this tool is not obviously useful in differential diagnosis.