Protracted course progressive supranuclear palsy

Abstract

Background and purpose:

Progressive supranuclear palsy (PSP) encompasses a broader range of disease courses than previously appreciated. The most frequent clinical presentations of PSP are Richardson syndrome (RS) and PSP with a predominant Parkinsonism phenotype (PSP-P). Time to reach gait dependence and cognitive impairment have been proposed as prognostic disease milestones. Genetic polymorphisms in TRIM11 and SLC2A13 genes have been associated with longer disease duration (DD).

Methods:

Methods used include retrospective chart review, genetic single nucleotide polymorphism analyses (in three cases), and neuropathology.

Results:

We identified four cases with long (>10–15 years) or very long (>15 years) DD. Stage 1 PSP tau pathology was present in two cases (one PSP-P and one undifferentiated phenotype), whereas pallidonigroluysian atrophy (PSP-RS) and Stage 4/6 (PSP-P) PSP pathology were found in the other two cases. Three cases were homozygous for the rs564309-C allele of the TRIM11 gene and the H1 MAPT haplotype. Two were heterozygous for rs2242367 (G/A) in SLC2A13, whereas the third was homozygous for the G-allele.

Conclusions:

We propose a protracted course subtype of PSP (PC-PSP) based on clinical or neuropathological criteria in two cases with anatomically restricted PSP pathology, and very long DD and slower clinical progression in the other two cases. The presence of the rs564309-C allele may influence the protracted disease course. Crystallizing the concept of PC-PSP is important to further understand the pathobiology of tauopathies in line with current hypotheses of protein misfolding, seeding activity, and propagation.

INTRODUCTION

Progressive supranuclear palsy (PSP) is a pathologically well-defined idiopathic 4-repeat tauopathy characterized by the presence of numerous neurofibrillary tangles (NFTs) or neuropil threads in the brainstem and basal ganglia, variable neuronal loss, and gliosis associated with tau-positive astrocytes or tangles in astrocytes [1]. The highly variable clinical presentation includes diverse phenotypes [2,3] outlined in the most recent Movement Disorder Society (MDS) diagnostic criteria [4]. The most frequent phenotype, Richardson syndrome (PSP-RS), is associated with early postural instability and supranuclear gaze paresis, subcortical dementia, and bulbar dysfunction. The second most common presentation is a Parkinsonism-predominant phenotype (PSP-P), which comprises asymmetric bradykinesia and tremor with a partial response to levodopa, which makes its early clinical diagnosis challenging and difficult to distinguish from idiopathic Parkinson disease (PD).

PSP-P progresses more slowly, with a longer disease duration (DD) than PSP-RS, despite similar disease severity at baseline [5]. In non-RS PSP and cases with predominant subcortical pathology, longer DD has recently been associated with two single nucleotide polymorphisms (SNPs), one in the TRIM11 gene (TRIM11/17 locus) [6–8] and the other in intron 3 of SLC2A13 in the vicinity of the LRRK2 gene [9].

Restricted distribution of PSP neuropathology has seldom been described, and PSP cases with long DD corresponding with PSP-P presentation show moderate tau pathology when compared to PSP-RS [10] corresponding to Stage 4/6 according to a recent staging scheme [11]. Additionally, pallidonigroluysian atrophy (PNLA) shares the neuropathology of PSP, with a restricted anatomical distribution of neurodegeneration. Although the typical mean age at onset of PNLA is 59 years and DD approximately 8 years [12], rare cases have been reported with longer DD (11.5 years) that initially had good mobility with more delayed imbalance and falls than typical PSP [13–15].

Here, we report four cases fulfilling the neuropathological diagnostic criteria of PSP [1,16] presenting a combination of a long DD with longer latency to disease milestones, or long DD and a low burden of tau pathology. We propose a conceptual definition of a “protracted course PSP” (PC-PSP) subtype for these presentations that may be influenced by specific SNP genotypes.

MATERIALS AND METHODS

A retrospective chart review was performed to collect clinical data on the four patients. When available, time to onset of specific milestones was recorded, and the clinical features at symptom onset or initial examination were retrospectively assigned to the clinical phenotypes defined by the recent MDS PSP diagnostic criteria [4]. Long DD was defined as 10–15 years and very long DD as >15 years.

Neuropathology

Neuropathologic examination was performed using formalin-fixed paraffin-embedded tissue blocks of several cortical, subcortical, brainstem, and cerebellar regions. The following monoclonal antibodies were used for immunohistochemistry: anti-phospho-tau (clone AT8, 1:200, Thermo Fisher Scientific), anti-4R tau (RD4, 1:200, Millipore), anti-3R tau (RD3, 1:5000, Millipore), anti-phospho-TDP-43 (pS409/410, 1:2000, Cosmo Bio), anti-a-synuclein (5G4, 1:4000, Analytik Jena), anti-Abeta (6F/3D, 1:50, Dako), anti-p62 (clone 3/p62 lck ligand, 1:800, BD Transduction), and microglia (human leukocyte antigen–DR isotype [HLA-DR], 1:400, Dako). Nerve cell loss, the load of each tau cytopathology (oligodendrocytic, astrocytic, and neuronal), and the total tau burden were evaluated semiquantitatively (none, mild, moderate, severe). The scores were used to prepare anatomic heatmaps of regional involvement.

Genetic polymorphism analyses

Genomic DNA for Case 1 was extracted from frozen brain and Sanger sequenced using specific primers (for details see Appendix S1). Cases 3 and 4 were genotyped for rs2242367 and rs564309 SNPs using the Illumina GSA array. The SNPs had a (i) genotyping success rate ≥ 98% per SNP or per individual, (ii) minor allele frequency ≥ 0.01, and (iii) Hardy–Weinberg equilibrium (p ≥ 1 × 10⁻⁶); or were imputed using TOPMed (NHLBI Trans-Omics for Precision Medicine), and finally extracted from the Plink file.

This study was undertaken with the understanding and written consent of relatives of the patients and approved by the research ethics boards of University Health Network (20–5258) and University of Toronto (39459) and was performed per the ethical standards established in the 1964 Declaration of Helsinki and updated in 2008.

RESULTS

Case reports

Patient 1

This left-handed female presented at age 64 years with 2 years of slower and less-coordinated movement and intermittent rest tremor of her left arm. She later developed dragging of her left foot. She had no relevant family history, toxin, or neuroleptic exposure. Additional medical history included type 2 diabetes, ischaemic cardiomyopathy, meningitis at age 19 years, acromegaly diagnosed at age 28 years treated with pituitary irradiation, spinal and knee osteoarthritis, and shingles in her left flank. On initial examination, there was a normal cranial nerve examination, rest tremor, mild–moderate bradykinesia but no rigidity, and decreased arm swing and stride length. Her tremor, slowness, and pain improved with levodopa, but after 2 years she developed motor and non-motor fluctuations. Her maximal dose was 800 mg/day of levodopa. Eight years after symptom onset, she developed peak-dose dyskinesias, which improved after adjustment of dopaminergic medication and the addition of amantadine 300 mg/day. She did not develop signs of oculomotor, bulbar, or autonomic dysfunction over 7 years of clinical follow up. Her brain magnetic resonance imaging done at age 67 years showed mild T2 hyperintensities scattered throughout the white matter but no signs of regional midbrain atrophy. As per the MDS PSP clinical criteria [4], her asymmetric presentation, with tremor and levodopa responsiveness (functional domain: akinesia; level 3), together with the tendency to fall on the pull-test fulfills criteria for suggestive of PSP-P. At age 73 years, she was diagnosed with breast cancer and died that same year. In summary, this woman had PSP-P with long DD (11 years) and response to levodopa.

Patient 2

This 82-year-old man was diagnosed with PD at age 70 years. He had steady progression of motor symptoms. He had a past medical history of coronary artery disease. He took levodopa/carbidopa 250/25 mg four times daily and trihexyphenidyl 2 mg once daily. In the 6 months prior to death, he was living in a nursing home requiring full care and died of aspiration pneumonia secondary to dysphagia. There are no clinical details available regarding his mental state and extraocular movements. The duration of illness was 12 years. In summary, this man manifested undifferentiated parkinsonism and terminal dementia with long DD.

Patient 3

This 65-year-old left-handed man presented for evaluation of frequent falls. His symptoms started at age 56 years, with slow walking and hunched posture. At age 61 years, he developed micrographia, changes in dexterity, and emotional lability. Between the ages of 62 and 63 years, his voice became softer, and he started to fall backward, followed later by the development of dysphagia and freezing of gait on initiation, turning, and at doorways. At the time of evaluation, he had three to four falls per week. In addition to these symptoms, he had intermittent vertical diplopia and mild constipation, but no urinary symptoms. There was no family history of neurological problems. On the initial examination, he had hypophonia, restricted upward gaze but normal downgaze, saccadic pursuit, symmetric rigidity and bradykinesia in the four limbs, small steps, decreased arm swing, en bloc turning with freezing of gait, and a positive pull test. Levodopa/carbidopa was titrated up to 1125 mg daily, with subjective gait improvement but no clear change in motor examination. He reported motor fluctuations but never developed dyskinesia. The subsequent development of sedation required a decrease in levodopa dose. Trials of pramipexole and amantadine provided no clear motor benefit. By age 69 years, he required a wheelchair and had severe dysphagia. He developed blepharospasm, treated with botulinum toxin. His wife also reported symptoms suggestive of rapid eye movement sleep behavioral disorder. On assessment at age 72 years, he had clear dementia, with prominent short-term memory difficulties, marked apathy and akinesia, marked slowness in responding to questions or commands, severe hypophonia, and complete upgaze restriction. His clinical diagnosis at that time was PSP. He died at age 74 years of unknown cause. In summary, this male patient suffered from a very long DD. The initial manifestations were of undifferentiated parkinsonism, and only 4–5 years later did he develop mild gait freezing. Although examination done 9 years after symptom onset (moderate disease stage) was consistent with PSP-RS, development of moderate cognitive impairment occurred only 16 years after onset.

Patient 4

This 77-year-old right-handed man was evaluated for left leg tremor. He initially developed dexterity problems (difficulty shuffling cards), slow gait, and micrographia at age 67 years. A year later, he noted intermittent right hand action tremor. At age 70 years, he had left leg tremor. At age 71 years, his voice became softer. By age 74 years, he had a sense of retropulsion. Falls started by age 76 years. He was treated with levodopa/carbidopa and amantadine with clear benefit but subsequently developed wearing off. On initial examination, he had hypophonic speech, full extraocular movements with limited convergence, mild rigidity in the neck and in the left extremities, and symmetric bradykinesia. He was diagnosed with PD. Ropinirole was added and levodopa increased to 1000 mg/day with moderate benefit and no dyskinesia. Over the next 10 years, he developed increased freezing, with falls, anxiety, dysphagia, and progressive cognitive decline eventually with marked apathy and bradyphrenia. At age 85 years, he began using a walker due to worsening of balance. By age 86 years, he was using a wheelchair much of the time. He had one syncopal spell, thought to be due to orthostasis. Over the last several months of his life, he experienced visual hallucinations and increasing daytime sleepiness. He was placed in a hospice at age 87 years and died from inanition. In summary, this patient’s presentation and initial assessment 10 years after symptom onset are consistent with a PSP-P phenotype.

Neuropathology

Patients 1 and 2

In Patient 1, histological signs of neuronal degeneration and loss were moderate in the subthalamic nucleus (STN) and substantia nigra (SN), and mild in the midbrain tegmentum and globus pallidus (GP), whereas the striatum was relatively preserved (Figure 1a,c–f). In Patient 2, neuronal loss was severe in the STN, SN, and GP, and moderate in the midbrain tegmentum, whereas the striatum was mildly affected (Figure 1k,m–p). In both cases, tau pathology comprised globose NFTs together with neuropil threads in the SN, STN, GP, and locus coeruleus and isolated NFTs in the caudate nucleus together with oligodendrocyte coiled bodies and tufted astrocytes (Figure 1b,g–j,l,q–t). The total tau load was less in Patient 1 than Patient 2. Based on the recently proposed staging of PSP [11], both cases fulfilled Stage 1, but due to the occasional neuronal tau pathology in the dentate nucleus, a transition to Stage 2 was assigned. In both cases, flame-shaped NFTs and neuropil threads were seen in the entorhinal cortex and hippocampus (HC), compatible with Stage II according to Braak et al. [17] and primary age-related tauopathy (PART) [18]; no Aβ-positive plaques were present.

FIGURE 1.

Neuropathological findings in Patients 1 (a–j) and 2 (k–t). Heat mapping of neuronal loss (a, k) and tau pathology (b, l) is shown. Representative images show neuronal loss (hematoxylin and Luxol staining; c–f and m–p) and tau pathology (AT8; g–j and q–t) of the subthalamic nucleus (c, g, m, q), substantia nigra (d, h, n, r), globus pallidus (e, i, o, s), and caudate nucleus (f, j, p, t). The bar in c represents 150 μm for all histological images

Patients 3 and 4

Neuropathological alterations were more severe in Patients 3 and 4 (Figure 2). The neuronal degeneration pattern in Patient 3 was compatible with PNLA. Tau pathology in Patient 3 was prominent in the tegmentum of the brainstem, with isolated tufted-like astrocytes resembling the “equivocal” tufted astrocytes in PNLA described by Yokoyama et al. in PNLA [14] and mild in the frontal and motor cortex. Because tau pathology was also mild in the GP and striatum and moderate in the STN but the frontal cortex was also involved, we could not provide a clear stage for the distribution of tau pathology in this case. Finally, Patient 4 showed more tau pathology than the other three patients that included tufted astrocytes in the striatum and involvement of the cortex, consistent with Stage 4 PSP pathology [11].

FIGURE 2.

Neuropathological findings in Patients 3 (a–j) and 4 (k–t). Heat mapping of neuronal loss (a, k) and tau pathology (b, l) is shown. Representative image show neuronal loss (hematoxylin and Luxol staining; c–f and m–p) and tau pathology (AT8; g–j and q–t) of the subthalamic nucleus (c, g, m, q), substantia nigra (d, h, n, r), globus pallidus (e, i, o, s), and caudate nucleus (f, j, p, t). The bar in c represents 150 μm for all histological images except for the right upper inset of j, for which it represents 25 μm

Tau pathology in all cases consisted of 4R tau isoform (Figure 3), with the 3R tau isoform detected only in the entorhinal cortex and HC, consistent with PART. Immunostaining for microglia (HLA-DR) revealed mild labeling in the STN, pars reticulata of the SN, and GP in Patient 1 (Figure 4a–f); these same regions were moderately affected in Patient 2 (Figure 4g–l), and more severely in Patients 3 and 4 (Figure 4m–r,s–x, respectively). Alpha-synuclein pathology was not observed in any of the cases; Aβ-positive plaques were seen in one (Patient 3, Thal Phase 1), and Patients 1 and 3 showed cerebral amyloid angiopathy type 2 [19,20]. Furthermore, Patient 2 showed phTDP-43 pathology in the amygdala and HC, interpreted as limbic-predominant age-related TDP-43 encephalopathy (LATE), Stage 2 [21] (Table 1). Low stage argyrophilic grain disease (AGD) was seen in Cases 2 and 3. Ageing-related tau astrogliopathy (ARTAG) was seen in Cases 1, 2, and 4.

FIGURE 3.

Immunostaining for 4R (a–p) and 3R tau (all upper right insets) isoforms in the subthalamic nucleus (a, e, i, m), substantia nigra (b, f, j, n), globus pallidus (c, g, k, o), and caudate nucleus (d, h, l, p) in Patients 1 (a–d), 2 (e–h), 3 (i–l), and 4 (m–p)

FIGURE 4.

Immunostaining for microglia activation (human leukocyte antigen–DR isotype) in the subthalamic nucleus (a, d, g, j, m, p, s, v), substantia nigra (b, e, h, k, n, q, t, w), and globus pallidus (c, f, i, l, o, r, u, x) as seen in Patients 1 (a–f), 2 (g–l), 3 (m–r), and 4 (s–x). The bar in a represents 100 μm for a–c, g–i, m–o, and s–u and 40 μm for d–f, j–l, p–r, and v–x

TABLE 1.

Demographic, clinical, and neuropathology data of the four cases

Characteristic	Patient 1	Patient 2	Patient 3		Patient 4
Sex	Female	Male	Female		Male
Age at symptom onset	62	70	56		67
Disease duration, years	11	12	18		20
Age at death, years	73	82	74		87
PSP clinical featuresa	O = 0, P = 2, A = 3, C = 0	O = x, P = 3, A = x, C = x	O = 1, P = 2, A = 2, C = x		O = 0, P = 2, A = 3, C = x
Tremor	++	+	−		++
Asymmetry of parkinsonism	++	+	−		±
Response to levodopa	Exquisite, developed dyskinesia	Not reported	+, short lived		++
Cognitive decline	Absent	Dementia late in the disease course	Dementia late in the disease course, memory, word finding	Mild
Additional symptoms	Myoclonus, pain, neuropathy	–	RBD, emotional lability, constipation	Visual hallucination, dysautonomia
Other conditions	Diabetes, ischemic heart disease, acromegaly, pituitary irradiation, osteoarthritis, breast cancer	Coronary artery disease	−	Anxiety
rs564309 genotype	C/C	n/a	C/C		C/C
rs2242367 genotype	G/A	n/a	G/A		G/G
MAPThaplotype	H1/H1	n/a	H1/H1		H1/H1
PSP stageb	1 to 2	1 to 2	Atypical pattern		4
Braak NFT stage	II	II	II		II
Thal phase	−	−	1		−
NIA-AA	A0B1C0	A0B1C1	A1B1C0		A0B1C0
Braak Lewy stage	−	−	−		−
ARTAG	BFB/AMY: SP	BFB/AMY: SP, WM	−		AMY/HI/MO: SP; AMY: WM, SE
CAA	Type 2	−	Type 2		−
LATE	−	2	−		−
Vascular lesion	−	Microinfarct in GP	−		−
Other	PART	PART & AGD (I)	AGD (I)		PART

Abbreviations: −, absent; +, moderate; ++, severe; ±, mild; A, Akinesia; AGD (I), argyrophilic grain disease (Stage I) [33]; AMY, amygdala; ARTAG, ageing-related tau astrogliopathy; BFB, basal forebrain; C, Cognitive dysfunction; CAA, cerebral amyloid angiopathy; GP, globus pallidus; HI, hippocampus; LATE, limbic-predominant age-related TDP-43 encephalopathy; MO, medulla oblongata; n/a, not available; NFT, neurofibrillary tangle; NIA-AA, National Institute on Aging and Alzheimer’s Association; O, Ocular motor dysfunction; P, Postural instability; PART, primary age-related tauopathy; PSP, progressive supranuclear palsy; RBD, rapid eye movement sleep behavior disorder; SE, subependymal; SP, subpial; WM, white matter; x, unknown.

^aHöglinger et al. [4]

^bKovacs et al. [11]

Genetic testing (Patients 1, 3, and 4)

For SNP rs2242367, Cases 1 and 3 showed inconclusive results, as both were heterozygous for the risk A-allele and protective G-allele (G/A). Case 4 was homozygous for the G-allele of rs2242367. Regarding rs564309, all three patients were homozygous for the C-allele (C/C genotype). MAPT haplotype was H1/H1 for all three cases (see Table 1).

DISCUSSION

We report four patients designated as having PC-PSP who manifested long or very long duration atypical parkinsonism. In retrospect, two of them met the clinical criteria for suggestive of PSP-Pwith absent or mild cognitive dysfunction; the other two cases developed dementia only very late in their clinical course; one of them developed PSP-RS after 9 years, and the other remained an undifferentiated phenotype. We propose two approaches to define PC-PSP; both include long DD, one defined neuropathologically and the other defined clinically. Based on these categories, two cases (Patient 1, PSP-P; Patient 2, undifferentiated) showed a low amount and stage [11] of tau-associated pathological inclusions after a long DD and are considered here neuropathologically defined PC-PSP. The other two cases with clinically protracted PSP had very long DD, one (Case 3, PSP-RS) with neuropathology compatible with PNLA and the other (Case 4, PSP-P) with Stage 4/6 PSP neuropathology (arguably considerably milder than expected for the extreme DD of 20 years). Shared features among these four cases include slow progression to specific clinical disease milestones. Disease milestones were recently proposed as a prognostication tool in PSP [22], where gait assistance develops approximately 1.2 years after symptom onset and moderate cognitive impairment takes 4.1 years. All our cases reached those milestones after a considerably longer time than expected for PSP-RS. In one of our cases with long DD, the cognitive impairment milestone occurred 11 years (3% of total DD) after onset, whereas our cases with very long DD (PSP-P, PSP-RS) developed cognitive impairment 16 years after onset (11%–20% of the total DD) and wheelchair dependency after 13–18 years (10%–28% of the total DD). Thus, it seems that cases with restricted PSP pathology progress to moderate clinical impairment quite late in their disease course. However, it should be noted that this proposed milestone-related clinical staging tool [22] is based on the Progressive Supranuclear Palsy Rating Scale, and the linear estimation of disease course was tested only in patients with RS, which was present in only one of our four cases.

A restricted distribution of PSP neuropathology was previously reported in a case with parkinsonism and dementia (Clinical Dementia Rating Scale = 3), but DD was not reported, so it is not possible to confirm PC-PSP [23]. Zhang et al. [24] reported five cases with restricted PSP neuropathology; however, none had the long DD that defines cases as PC-PSP. In the 20 cases classified as PSP-P described by Kovacs et al. [11] there was mostly moderate (4+/6) stage tau pathology. Astrocyte inclusions were seen in the striatum, and to a lesser extent in frontal and occipital cortices, and oligodendroglial coiled bodies in the pallidum and dentate. The same findings, with a low density of coiled bodies and threads in the striatum and dentate nucleus, and no pathology in the cortex, were previously described by Williams et al. [10]. Additionally, PSP-P has shown higher tau grains in the HC and small vessel pathology (46%). In contrast to the cases reported in these studies, our neuropathological PC-PSP category is defined on the basis of a restricted distribution and a low stage of tau pathology associated with long DD. The restricted tau pathology in these cases (Cases 1 and 2) was associated with a lower degree of microglial activation when compared with the other two cases of very long DD that had more severe tau pathology (Cases 3 and 4; Figure 4). Of note, these cases showed additional but mild degree copathologies (PART, LATE, AGD, and ARTAG). The present study cannot establish the contribution of diverse restricted tau pathologies to clinical presentation or disease course, but hopefully our report, drawing attention to PC-PSP, will encourage further studies in larger cohorts.

Compared to typical PSP-RS-related neuropathology, PC-PSP can be associated with three constellations of neuropathology: (i) low level of tau pathology and restricted (early stage) anatomical involvement (“minimal change PSP”), (ii) higher level of tau pathology and moderate stage anatomical involvement (represented mostly by long-duration PSP-P), and (iii) high level of tau pathology with restricted anatomical involvement representing atypical forms of PSP such as PNLA. Compared to PSP-RS, these constellations are associated with a longer course of disease and later wheelchair and cognitive milestones. We speculate that the “minimal change PSP” cases might be driven by a slower process of tau protein aggregation. Further studies are warranted to establish whether this relates to the presence of tau strains with less aggressive seeding activity, and what the role of microglial activation and genetic predisposition is in PC-PSP. Interestingly, contrasting our experience with PC-PSP, cases of rapidly progressive PSP (2–4 years) have also been described [25,26], as well as fulminant forms of another 4R tauopathy, corticobasal degeneration (CBD) [27]. These combined observations support the notion that the propagation and potential neurotoxicity of tau aggregates in 4R tauopathies (i.e., PSP or CBD) might differ between cases currently classified exclusively based on their common neuropathological features.

The pathological category of PC-PSP can be compared to the proposed concept of “minimal change” multiple system atrophy (MC-MSA). In MC-MSA, neuronal loss is localized to the SN and locus coeruleus, with an otherwise typical distribution of glial cytoplasmic inclusions. A recent report defined the neuronal loss as no more than +1 in the Purkinje cell layer plus one other olivopontocerebellar or striatonigral structure [28]. However, in contrast to our two cases with neuropathologically defined PC-PSP, patients with MC-MSA can show an aggressive disease course, with >50% having young age at onset (<40 years) and early autonomic changes leading to 50% sudden unexpected death [28,29].

Two of our cases were classified as clinically PC-PSP, with very long DD and remarkably late development of cognitive and wheelchair milestones (PSP-P and PSP-RS). Although DD is longer in PSP-P than PSP-RS [3,10], the concept of PC-PSP adds to the longer DD a longer latency to achieve disease milestones. Furthermore, in earlier reports, PSP-P patients without dementia and supranuclear gaze palsy had a DD of 12 years [30], compared to our PC-PSP median DD of 15 years, as well as a later age at onset (67–75.5 years vs 56–70 years) [10,30].

Among 20 cases with PSP-P described by Kovacs et al. [11], DD ranged between 8 and 19 years, and tau pathology stage was moderate. In agreement with this, Case 4 of the present series had very long DD and moderate tau pathology, lending further support to the concept of PC-PSP and suggesting that PSP-P might be overrepresented in the population of PC-PSP. Reports of other cases with a similar degree of pathology and disease course support this, such as Case 17 in Daniel et al. [30], a woman with a 12-year history of levodopa-responsive bradykinesia, no dementia, moderate cell loss in SN, mild cell loss in locus coeruleus, frequent NFTs in basal ganglia and parahippocampus, and mild NFTs in oculomotor complex, periaqueductal gray, raphe, and HC. In summary, the neuropathology of clinically defined PC-PSP cases may either be low-stage PSP [11], mostly when presenting as PSP-P, or show distinct neuropathological features overlapping with PSP, such as PNLA.

Finally, the limited neuropathology and longer DD of our patients (Patients 1, 3, and 4) might in part be related to the homozygosity for the rs564309-C allele at TRIM11. The presence of the rs564309-C allele has been reported to be associated with later onset [7], whereas a higher degree of neurofibrillary tangle pathology was associated with the rs564309-A allele in a study of 797 PSP cases described as subcortical phenotype [8]. Although there is no conclusive evidence to suggest that the rs564309-C allele is protective, our three patients’ homozygosity for the rs564309-C allele supports this possibility [8]. The SNP rs2242367 was reported to be associated with PSP survival probability [9], and mapped to intron 3 of the SLC2A13 gene 190 kb from the LRRK2 locus, which is a common risk factor for PD. In three different cohorts of PSP [9], both RS and non-RS, rs2242367 showed an additive effect on DD, where the rs2242367-G allele is associated with longer survival in PSP. In contrast, analysis of rs2242367 under an additive model showed that carrying an A-allele was significantly associated with decreased survival. Our Case 4 was homozygous for the G-allele, which may have contributed to his very long DD (20 years). However, Cases 1 and 3 were heterozygous for this protective G-allele, which renders interpretation of any role of this SNP in their survival inconclusive. All three cases were homozygous for the MAPT H1 haplotype, which has been shown to be more prevalent in PSP compared to PD and Alzheimer disease and is associated with pathological severity [31].

CONCLUSIONS

In summary, the concept of PC-PSP includes cases with long DD combined with slow clinical progression and either a low burden of anatomically restricted tau pathology or atypical forms of PSP. Clinical presentation is not restricted to PSP-P but includes slow progression to PSP-RS and undifferentiated parkinsonism with dementia. The concept of PC-PSP operationalizes in two categories, clinical and neuropathological: 1) the presence of long DD associated with slow progression, classified as “clinically defined PC-PSP”, associated with a range of PSP-related tau pathologies; 2) an unexpected mild degree and stage of PSP-type tau pathology and neurodegenerative features despite a long DD classified as “mild PSP neuropathology associated with PC-PSP”. Crystalizing this concept provides a framework to investigate the biology underlying the long DD of atypical neurodegenerative parkinsonism, including its pathological underpinnings, and the influence of genotype. Designating patients as having PC-PSP will be important in future evaluations of seeding assays of misfolded protein and cryo-electron microscopy. This subclassification of atypical PSP is in line with a new molecular classification of tauopathies [32] and will add theoretically to the ongoing discussion of the varied nature, propagation, and potential management of misfolded tau protein. As more reliable diagnostic tests become available for PSP (e.g., seeding amplification assays), the recognition and future study of PC-PSP will be crucial, not only in providing patients and families more accurate prognostic information but also for the conduct (e.g., design, duration, outcome measures) of future trials of experimental disease-modifying therapies.

DATA AVAILABILITY STATEMENT

Data are available on request due to privacy/ethical restrictions.