Progressive Supranuclear Palsy (PSP) and Corticobasal Syndrome (CBS) are both classified as 4R-tauopathies, diseases characterized by the predominant accumulation of tau isoforms containing four microtubule-binding repeats. Despite this shared molecular pathology — and considerable overlap in their genetic architecture — the two conditions produce strikingly different clinical syndromes. Understanding the mechanistic basis for this phenotypic divergence is central to developing targeted therapies for each condition.
PSP-RS presents with early postural instability, vertical supranuclear gaze palsy, and axial akinesia, reflecting vulnerability of brainstem and basal ganglia structures. CBS, by contrast, is defined by asymmetric cortical signs — apraxia, alien limb phenomenon, cortical sensory loss — reflecting damage to motor and parietal cortical regions. The divergence emerges from differences in: (1) genetic risk profiles, (2) regional patterns of neuronal vulnerability, (3) tau filament conformations, (4) neuroimmune contributions, and (5) spreading mechanisms.
PSP-RS is the most common PSP phenotype, accounting for 50-55% of pathologically confirmed cases. The cardinal features are:
- Vertical supranuclear gaze palsy (VSGP): Impaired downward gaze is the most characteristic early sign. Patients lose the ability to look down voluntarily before upward gaze is affected. Horizontal saccades are also progressively impaired.
- Postural instability with early falls: Backward falls within the first year are a defining feature, reflecting damage to the midbrain pedunculopontine nucleus and vestibular integration centers.
- Axial akinesia and rigidity: "Axial predominant" motor pattern with early neck extension (retrocollis) and generalized akinesia out of proportion to limb rigidity.
- Pseudobulbar features: Dysarthria, dysphagia, and emotional incontinence occur early.
- Frontal cognitive dysfunction: Executive impairment, slowed processing speed, and behavioral changes mirror frontostriatal circuit disruption.
CBS is defined by asymmetric cortical signs combined with extrapyramidal features:
- Asymmetric limb apraxia: Inability to perform learned motor tasks on command, affecting the dominant hand first. Distal > proximal, left more common than right.
- Alien limb phenomenon: Patient's limb appears to act autonomously, with grasping, levitation, and resistive behaviors the patient cannot suppress.
- Cortical sensory loss: Impaired two-point discrimination, stereognosis, and graphesthesia despite intact primary sensation.
- Myoclonus: Action myoclonus, often stimulus-sensitive, in the affected limb.
- Asymmetric dystonia: Typically affects the hand and arm in an akinetic-rigid pattern.
- Cognitive features: Visuospatial dysfunction, non-fluent aphasia, and executive impairment reflecting asymmetric frontoparietal involvement.
The clinical features map to distinct neuroanatomical substrates:
| Clinical Feature |
PSP-RS Anatomical Basis |
CBS Anatomical Basis |
| Vertical gaze palsy |
Midbrain superior colliculus, rostral interstitial nucleus of MLF |
Parietal eye fields, frontal eye fields |
| Early falls |
Subthalamic nucleus, pedunculopontine nucleus |
Posterior parietal cortex, supplementary motor area |
| Asymmetric apraxia |
Bilateral basal ganglia (less affected) |
Asymmetric motor cortex, premotor cortex, corpus callosum |
| Alien limb |
Minimal involvement |
Primary somatosensory cortex, superior parietal lobule |
| Axial vs limb rigidity |
Midbrain reticular formation |
Asymmetric basal ganglia output |
The MAPT gene on chromosome 17q21 encodes tau protein. Two major haplotypes — H1 and H2 — exist due to a 900 kb inversion polymorphism. The H1 haplotype is the major risk factor for both PSP and CBS, but with important quantitative differences:
- PSP: Strong association with H1/H1 genotype. OR approximately 5-8 for H1/H1 vs H2/H2. Multiple H1-specific variants contribute, including rs242557 (MAPT cis-regulatory element).
- CBS/CBD: H1/H1 association is present but weaker (OR ~2-4). H2 haplotype may be slightly protective, similar to PSP.
The H1 haplotype spans a region including MAPT and several flanking genes (ARAP1, STX6, KANSL1) — the inversion creates distinct H1 and H2 linkage disequilibrium blocks. Both diseases share the H1 haplotype but differ in associated SNPs within the H1 block:
- PSP-specific: rs242557 (A allele), rs4792891, rs7216058
- Shared or less specific: rs1052553 (exon 3 skip variant, minor allele A associated with reduced 3R tau)
TMEM106B polymorphisms influence lysosomal function and have been associated with TDP-43 proteinopathies. Recent evidence suggests:
- TMEM106B rs3173615 (T185S) affects progranulin levels and may modulate neuroinflammation
- The protective allele (T185) is less common in both PSP and CBS
- TMEM106B effects appear more pronounced in PSP than CBS, potentially via microglial modulation of tau spreading
¶ 2.3 LRRK2 and CBS-Specific Genetics
LRRK2 G2019S mutations — more commonly associated with Parkinson's disease — are found in a subset of CBS cases, particularly those with asymmetric presentations:
- LRRK2-associated CBS cases often show PSD-95 positive inclusions alongside tau pathology
- This suggests LRRK2 mutations may drive CBS through distinct pathways involving kinase dysregulation
- The mechanistic link between LRRK2 and 4R-tau pathology remains under investigation
Genome-wide association studies reveal substantial genetic overlap between PSP and CBD:
- Shared risk loci include MAPT H1 region and chromosome 2p13 (EPS8L3)
- PSP-specific signal near STX6 (involved in vesicular trafficking)
- CBD-specific signal near ARAP1
This pattern suggests the same H1 haplotype predisposes to both conditions, but additional genetic modifiers determine which phenotypic expression emerges.
¶ 3.1 PSP: Midbrain and Basal Ganglia Predominance
PSP pathology preferentially targets structures with dense tau expression under normal conditions:
- Subthalamic nucleus (STN): Among the highest tau-expressing regions in the human brain. Early pretangle and coiled body formation drives the characteristic early falls and axial symptoms.
- Globus pallidus interna (GPi): Tau accumulation in the output nucleus of the basal ganglia drives the axial akinesia and rigidity characteristic of PSP-RS.
- Substantia nigra pars compacta: Neuronal loss contributes to the parkinsonian features, but unlike PD, dopamine replacement therapy is ineffective.
- Midbrain tectum: Involvement of the superior colliculus explains the vertical gaze palsy.
- Dentate nucleus of cerebellum: Contributes to the PSP phenotype through disruption of cerebellar-thalamic loops.
¶ 3.2 CBS: Asymmetric Cortical and Basal Ganglia Involvement
CBS targets lateralized cortical circuits:
- Motor cortex (Brodmann areas 4, 6): Asymmetric involvement produces the characteristic apraxia and alien limb. Layer V pyramidal neurons are particularly vulnerable.
- Parietal cortex (BA 5, 7): Loss of somatosensory association cortex drives cortical sensory loss and visuospatial dysfunction.
- Basal ganglia (putamen > globus pallidus): Asymmetric striatal involvement contributes to the limb dystonia and rigidity.
- Corpus callosum: Transcallosal degeneration allows interhemispheric spread and may contribute to the alien limb phenomenon.
- Thalamus: Ventral posterior and ventral lateral nuclei show involvement correlating with sensory and motor deficits.
Both conditions show particular vulnerability of specific neuronal populations:
| Cell Type |
PSP Vulnerability |
CBS Vulnerability |
| Pyramidal neurons (cortical layer V) |
Moderate |
High (asymmetric) |
| Striatal medium spiny neurons |
High |
Moderate |
| Subthalamic nucleus neurons |
Very high |
Low |
| Midbrain oculomotor neurons |
High |
Low |
| Pontine nuclei |
High |
Low |
| Cerebellar Purkinje cells |
Moderate |
Low |
The molecular basis of this differential vulnerability involves region-specific tau expression levels, neuronal morphology, electrophysiological properties, and local immune microenvironment.
Both PSP and CBS are 4R-tauopathies, meaning that the accumulated tau filaments contain predominantly the 4-repeat isoforms (3R:4R ratio approaches 0:1 in most cases, compared to 1:1 in Alzheimer's disease where both 3R and 4R tangles are present). The exclusive use of 4R tau distinguishes both from Pick disease (3R tauopathy) and AD (mixed 3R+4R).
The critical distinction between PSP and CBS is not the 4R:3R ratio itself, but rather:
- The conformational variant of the 4R tau filament
- The ultrastructural morphology (different filament structures)
- The post-translational modifications on the tau protein
Cryo-EM studies have revealed that different tauopathies produce distinct filament structures:
- PSP filaments: The PSP-type fold shows a compact C-shaped conformation with specific residues (Lys-234, Val-306) involved in the core. The filaments show a "pillown" appearance on EM.
- CBD filaments: Corticobasal degeneration produces a distinct filament fold with different protease-resistant core boundaries. The CBD fold includes residues from the repeat domain that are excluded from PSP filaments.
- These structural differences constitute "strains" — distinct conformers that propagate templating of identical conformation, analogous to prion strains.
The strain hypothesis explains the phenotypic divergence: the same amino acid sequence (4R tau) can adopt different conformations that target different neuronal populations. PSP strains preferentially propagate in subthalamic and brainstem neurons; CBD strains propagate in cortical pyramidal neurons.
Key structural differences between PSP and CBD tau filaments:
- N-terminal region involvement: CBD filaments include more N-terminal residues in the core than PSP filaments.
- R2/R3 repeat domain boundary: The precise location of the protease-resistant core differs.
- Phosphorylation patterns: Both conditions show hyperphosphorylation at S202/T205 (AT8 epitope), but CBS cases may show additional phosphorylation at T217 (a marker more associated with AD pathology).
Neuroinflammation plays a differential role in PSP vs CBS:
PSP:
- Prominent activation of Iba1+ microglia in the basal ganglia and midbrain
- TREM2 expression is increased in PSP brains — the TREM2 Q317 variant shows nominal association with PSP risk
- Microglial morphology in PSP suggests a surveillance (M0) to neurotoxic (M1) shift in affected regions
- TSPO PET imaging shows increased binding in brainstem and basal ganglia in PSP
CBS:
- Asymmetric microglial activation mirrors the cortical pattern of atrophy
- CD33 (sialic acid-binding Ig-like lectin 3) polymorphisms affect microglial phagocytosis and have been associated with CBS/CBD risk
- TREM2 variants show weaker association with CBS than PSP
- Neuroinflammation may be more pronounced in CBS cases with concomitant AD pathology
Astrocyte reactivity differs between the conditions:
- PSP: Predominantly subcortical astrocyte involvement with GFAP+ astrocytes in basal ganglia. Cortical astrocyte pathology is less prominent.
- CBS: Asymmetric cortical astrocyte involvement, with many cases showing "astrocytic plaques" (the CBD-specific astrocytic tau pathology). This distinctive pattern — tau accumulation in astrocytic processes — is highly specific for CBD neuropathology.
¶ 5.3 Inflammatory Cytokines and Biomarkers
CSF and plasma studies reveal:
- NFL (neurofilament light chain): Elevated in both conditions, higher in PSP-RS than CBS
- Tau (total and phosphorylated): Both conditions show elevated p-tau181, but CBS cases with underlying AD show higher p-tau217
- GFAP: May be more elevated in CBS with significant cortical involvement
- IL-6, TNF-alpha: Variable elevations in both conditions; TSPO PET shows more inflammation in PSP brainstem
¶ 6. Clinical Trial Overlap and Divergence
Both conditions target common pathways:
- Tau aggregation inhibition: Lithium, methylene blue derivatives, and small molecule aggrega-tion inhibitors have been tested in both conditions
- Tau antisense oligonucleotides (ASOs): BIIB080 (MAPT-targeting ASO) is in trials for PSP and CBD as part of the TauNex program
- Immunotherapy: Aducanumab, semorinemab, and other anti-tau antibodies have been tested in PSP (NCT02880956) and CBD (various)
| Agent |
Target |
PSP Trial |
CBS Trial |
| Tideglusib (GSK-3 inhibitor) |
GSK3B |
Phase II (negative) |
Phase II (negative) |
| Davunetide (granin peptide) |
Neuroprotection |
Phase II/III (negative) |
No |
| BIIB080 (MAPT ASO) |
Tau mRNA |
Phase I/II recruiting |
Phase I/II recruiting |
| Bepranserastat (TDF-853) |
CK1delta |
Phase I (PSP) |
No |
| Semorinemab (anti-tau mAb) |
Tau extracellular |
Phase II (negative) |
Phase II (ongoing) |
- Outcome measures differ: PSP trials use PSP-Scale (PSPS), vertical saccade velocity; CBS trials use CBS-CGI-C, Jebsen Taylor test
- Population heterogeneity: CBS trials must account for the fact that ~50% of CBS patients have underlying AD pathology (ADNC), which confounds tau-targeting approaches
- Biomarker stratification: CSF p-tau217 and tau PET can distinguish AD-cBBS from primary 4R-tau CBS, allowing enriched trial designs
The following diagram summarizes the convergent and divergent pathways underlying the phenotypic split between PSP and CBS:
flowchart TD
A["4R-Tau\nOverexpression\n(MAPT H1 haplotype)"] --> B1["Common 4R-Tauopathy\nFoundation"]
B1 --> C1["PSP Convergent\nPathway"]
B1 --> C2["CBS Convergent\nPathway"]
C1 --> D1["Subthalamic Nucleus\nTau Seeding"]
C1 --> D2["Midbrain Vulnerability\n(SC, RiMLF)"]
C1 --> D3["Brainstem Propagation\nPattern"]
C1 --> D4["PSP-Specific Tau\nFilament Strain"]
C2 --> E1["Asymmetric Motor\nCortex Seeding"]
C2 --> E2["Parietal Cortex\nVulnerability"]
C2 --> E3["Corpus Callosum\nSpread"]
C2 --> E4["CBD-Specific Tau\nFilament Strain"]
D1 --> F1["Early Falls\nPostural Instability"]
D2 --> F2["Vertical Gaze Palsy\nAxial Akinesia"]
D3 --> F3["Pseudobulbar Syndrome\nFrontal Dysfunction"]
D4 --> F4["Brainstem-Centered\nClinical Phenotype"]
E1 --> G1["Limb Apraxia\nAlien Limb"]
E2 --> G2["Cortical Sensory Loss\nVisuospatial Deficits"]
E3 --> G3["Asymmetric Dystonia\nMyoclonus"]
E4 --> G4["Cortical-Centered\nClinical Phenotype"]
style A fill:#e1f5fe,stroke:#333
style B1 fill:#e1f5fe,stroke:#333
style C1 fill:#fff9c4,stroke:#333
style C2 fill:#fff9c4,stroke:#333
style D1 fill:#ffcdd2,stroke:#333
style E1 fill:#ffcdd2,stroke:#333
style F1 fill:#c8e6c9,stroke:#333
style F2 fill:#c8e6c9,stroke:#333
style F3 fill:#c8e6c9,stroke:#333
style F4 fill:#c8e6c9,stroke:#333
style G1 fill:#c8e6c9,stroke:#333
style G2 fill:#c8e6c9,stroke:#333
style G3 fill:#c8e6c9,stroke:#333
style G4 fill:#c8e6c9,stroke:#333
The model illustrates how a common H1-driven 4R-tauopathy foundation diverges into two clinically distinct syndromes through: (1) initial seeding site (brainstem vs. cortex), (2) region-specific neuronal vulnerability patterns, (3) distinct tau filament strains that preferentially propagate in specific neuronal populations, and (4) differential neuroimmune contributions.