🏷

Classifications

Harrison's Part

◈

Mechanistic Hypotheses

2

Cytotoxic T-cell attack on Purkinje cells presenting an intracellular onconeural antigen

pcd_t_cell_mediated_intracellular CANONICAL

Evidence balance 1 support

In anti-Yo and other intracellular-antigen forms, the effector is a cytotoxic CD8 T-cell response and the antibody is a diagnostic marker rather than the agent of injury. The strongest support is neuropathological - MHC class I upregulation on neurons, CD8-positive granzyme-B-positive T cells, and microglial nodules are the signature of a T cell-mediated attack, and an antibody cannot reach an intracellular antigen in a living cell.

Pathograph links

Antitumour Immune Response Cross-Reactive with Purkinje Cells -> Cytotoxic T Cell-Mediated Purkinje Cell Destruction DIRECT Where the antigen is intracellular, only the T-cell arm can engage it, and the neuropathology is that of a cytotoxic attack.

Cytotoxic T Cell-Mediated Purkinje Cell Destruction -> Irreversible Purkinje Cell Loss DIRECT Cytotoxic killing removes the cell, and Purkinje cells are not replaced.

Cytotoxic T Cell-Mediated Purkinje Cell Destruction -> Prodromal vertigo and imbalance DIRECT Ongoing cell death before the vulnerable population is exhausted. The edge is drawn from the destruction node rather than from Irreversible Purkinje Cell Loss on purpose: the prodrome is what partial loss looks like, and by the time the loss is complete the patient is no longer prodromal. This is the only clinical feature in the entry that reports on a still-treatable state.

Show evidence (1 reference)

PMID:36070310 SUPPORT Human Clinical

"Although the intracellular location of the target antigen in anti-Yo-PCD supports a T cell-mediated pathology, the immune mechanisms in anti-P/Q-VGCC-PCD remain unclear."

States the argument from antigen compartment for this hypothesis, and in the same sentence marks the surface-antigen arm as unresolved.

Direct antibody-mediated Purkinje cell dysfunction against a surface antigen

pcd_antibody_mediated_surface ALTERNATIVE

Evidence balance 3 support

In surface-antigen forms the autoantibody plausibly causes the disease directly, by acting on a target it can physically reach in a living cell. In anti-P/Q-VGCC disease that is altered synaptic transmission at the channel, producing calcium dysregulation and eventual Purkinje cell death. It is curated as ALTERNATIVE rather than as a competing account of the same disease, because it describes a different antigen class - the two hypotheses are not rivals for one mechanism so much as two mechanisms sharing a syndrome. The compartment rule is not built on the anti-P/Q-VGCC case. Anti-Tr/DNER and anti-mGluR1 are the other well-characterised surface-antigen cerebellar syndromes, and they behave as the rule predicts: anti-Tr binds the extracellular domain of DNER, and anti-mGluR1 disease responds to immunotherapy in most patients where anti-Yo disease does not. What rests on the single case is the neuropathological contrast - MHC class I, complement deposition, topography of cell loss - and that narrower claim is where the knowledge gap in this entry is scoped.

Pathograph links

Antitumour Immune Response Cross-Reactive with Purkinje Cells -> Antibody-Mediated Purkinje Cell Dysfunction DIRECT Where the antigen is on the cell surface, the antibody can bind it on a living neuron and interfere with its function directly.

Antibody-Mediated Purkinje Cell Dysfunction -> Irreversible Purkinje Cell Loss INDIRECT KNOWN INTERMEDIATES Sustained dysfunction and calcium dysregulation eventually kill the cell, but the intermediate dysfunctional state is what creates a therapeutic window.

Show evidence (3 references)

PMID:36070310 SUPPORT Human Clinical

"Our findings support a pathogenic role of anti-P/Q-VGCC autoantibodies in causing neuronal dysfunction, probably due to altered synaptic transmission resulting in calcium dysregulation and subsequent PC death."

The authors' statement of this hypothesis, with their own hedging ("probably") preserved. The evidence base is a single autopsy case, which is recorded in a discussion.

PMID:25745634 SUPPORT In Vitro

"Anti-Tr antibodies bind to the extracellular domain of DNER and can be detected by RC-IFA using HEK293 cells expressing the recombinant receptor."

Establishes that a second onconeural antibody in this syndrome targets an extracellular domain, which is the physical precondition this hypothesis rests on. Graded IN_VITRO because the localisation is established by a recombinant cell-based assay, not in patients.

PMID:41197574 SUPPORT INDIRECT Human Clinical

"Immunotherapy yields favorable responses in most patients, with generally good long-term outcomes."

A third surface-antigen syndrome, and the treatment-response prediction of this hypothesis borne out across 42 pooled cases. Graded INDIRECT because a good response to immunotherapy is consistent with an antibody-mediated mechanism without demonstrating one - the inference runs through the assumption that immunotherapy acts by removing the antibody.

?

Discussions and Knowledge Gaps

4

Is checkpoint-inhibitor-triggered cerebellar ataxia the same disease as spontaneous paraneoplastic cerebellar degeneration, or a distinct drug toxicity that resembles it?

INTERPRETATION pcd_ici_triggered_same_disease

Attached to

environmental#Immune checkpoint inhibitor therapy pathophysiology#Antitumour Immune Response Cross-Reactive with Purkinje Cells

The case for one disease is mechanistic and strong. Checkpoint inhibitors act by removing restraint from an antitumour T-cell response, which is exactly the response this entry's trigger node describes; about half of affected patients carry a neuronal antibody; and the syndrome is cerebellar. If the spontaneous disease is a failure of tolerance in an antitumour response, the drug version is that failure induced deliberately. The case against is empirical and also strong. The one study that compares the two directly finds them significantly different on four axes at once - sex, tumour type, whether the ataxia is isolated, and outcome. Male predominance and lung cancer are the inverse of the anti-Yo picture, which is female and gynaecological. A better outcome is the difference that matters clinically, and it is not a small one. The entry does not resolve this. It curates the exposure and its mechanism link, and does not create a subtype, because a subtype would assert sameness that the comparative data do not support. Nor is it excluded, because excluding it would assert difference on the strength of one retrospective comparison of 35 against 15 patients. Worth stating what would settle it: whether the antibody-defined subgroup of checkpoint-inhibitor cases behaves like its spontaneous counterpart. The comparison available pools antibody-positive and antibody-negative cases together, so the better outcome could belong entirely to the antibody-negative half and tell us nothing about whether anti-Yo disease triggered by a drug differs from anti-Yo disease arising on its own.

Show evidence (2 references)

PMID:39153058 SUPPORT Human Clinical

"When compared with a series of PCA (n = 15), the cerebellar irAE group was significantly more associated with male sex, lung cancer (rather than gynecological/breast cancers), isolated ataxia, and a better outcome."

The direct comparison, with its denominators. Four simultaneous differences in a 35-versus-15 retrospective comparison is the whole case against treating the two as one disease, and the sample sizes are why it is not decisive.

PMID:39153058 SUPPORT Human Clinical

"Clinical presentation-antibody-tumor triad in the ICI group only partially reflects the associations described in paraneoplastic disorders."

The authors' own summary of the relationship - partial correspondence rather than identity or independence, which is the position this discussion records.

In anti-Yo paraneoplastic cerebellar degeneration, is the autoantibody an effector of Purkinje cell death or a marker of a T cell-mediated attack?

CONTROVERSY pcd_antibody_versus_t_cell_effector

Attached to

pathophysiology#Antitumour Immune Response Cross-Reactive with Purkinje Cells pathophysiology#Cytotoxic T Cell-Mediated Purkinje Cell Destruction mechanistic_hypotheses#pcd_t_cell_mediated_intracellular

The review literature states plainly that the relative importance of autoantibodies and cytotoxic T lymphocytes in the neuronal loss is debated. This entry curates the T-cell reading as canonical, and the reasoning should be visible rather than assumed. The argument for it is not primarily empirical, it is mechanical: CDR2 is intracellular, and an antibody in the extracellular space cannot bind an intracellular antigen in a living cell. The neuropathology then supplies exactly what a cytotoxic attack should look like - MHC class I upregulated on neurons, CD8 granzyme-B T cells present, microglial nodules. What keeps it a controversy rather than a settled question is that cytoplasmic IgG staining is seen in some neurons, and that nobody has shown the T-cell response is sufficient. The practical stake is therapeutic: if antibodies are effectors, plasma exchange and B-cell depletion should work, and the published experience is that they mostly do not - which is itself weak evidence for the T-cell reading.

Proposed experiments

Comparative depletion of the antibody and T-cell arms in anti-Yo PCD

pcd_effector_depletion_comparison

Compare outcomes between antibody-directed treatment (plasma exchange, B-cell depletion) and T cell-directed treatment in anti-Yo patients treated at a comparable point in the clinical course, since the existing experience confounds treatment type with how late it was given.

Readouts

Cerebellar function after arm-specific immunotherapy

pathophysiology#Cytotoxic T Cell-Mediated Purkinje Cell Destruction

Direction: UNCHANGED

Interpretation: If antibody removal does not change the course while T cell-directed treatment does, the canonical hypothesis is supported over the antibody-effector one.

Should paraneoplastic cerebellar degeneration be treated as one disease, given that the antigen compartment appears to determine both the mechanism and whether treatment can work?

INTERPRETATION pcd_antigen_compartment_determines_window

Attached to

mechanistic_hypotheses#pcd_antibody_mediated_surface pathophysiology#Antibody-Mediated Purkinje Cell Dysfunction

The neuropathological comparison in this entry finds the intracellular-antigen and surface-antigen forms differ in nearly every measured feature: distribution of cell loss, MHC class I upregulation, complement deposition, antigen downregulation, and evidence of endoplasmic reticulum stress. That is a large difference to hold inside one disease name. The reason it is curated as one entry with two hypothesis groups rather than split is that the clinical syndrome, the tumour association and the diagnostic pathway are shared, and the antibody is often not known when the patient presents. A clinician meets one syndrome and finds out later which mechanism it was. The consequence worth flagging is prognostic. The uniformly poor outlook attached to this disease comes from anti-Yo series. It should not be assumed to apply to the surface-antigen forms, where the pathology suggests dysfunction preceding death and therefore a window in which early treatment might matter - a point the source paper makes explicitly. This is also the boundary question against the existing Paraneoplastic_Neurological_Syndromes entry, which splits by antigen compartment rather than by syndrome. The two entries are complementary and should stay consistent: that one covers the compartment distinction across syndromes, and this one covers what it means within the cerebellar syndrome.

How much confidence should the neuropathological contrast between the intracellular-antigen and surface-antigen forms carry, given that it rests on a single anti-P/Q-VGCC autopsy case?

KNOWLEDGE GAP pcd_surface_antigen_evidence_thin

Attached to

mechanistic_hypotheses#pcd_antibody_mediated_surface pathophysiology#Antibody-Mediated Purkinje Cell Dysfunction

The scope of this gap is narrower than it first appears, and getting the scope right is the point of recording it. What does not rest on one case: the compartment rule itself. That an antibody cannot reach an intracellular antigen in a living cell is mechanical, and the surface-antigen side of it is supported independently by anti-Tr/DNER, whose target is a demonstrated extracellular domain, and by anti-mGluR1, where 42 pooled cases respond to immunotherapy as an antibody-mediated mechanism predicts. The entry's two-hypothesis architecture is not fragile. What does rest on one case: the specific neuropathological contrast. The paper that establishes it is explicit about its own denominator - one anti-P/Q-VGCC case against two anti-Yo cases and controls - and everything the entry says about topography of cell loss, MHC class I upregulation, complement deposition, antigen downregulation and endoplasmic reticulum stress in the surface-antigen arm inherits that limitation. So does the prognostic implication drawn from it. This is recorded rather than hedged into the node descriptions because it is a property of the evidence base, not of the claim: no amount of careful wording in the pathophysiology nodes tells a reader which parts rest on a single case, and the alternative - omitting the arm - would lose a real and well-characterised finding. A separate and larger gap sits underneath all of this. Four independent attempts to reconstruct the disease in animals - passive transfer of patient IgG, active immunization across four MHC haplotypes, transfer of immunized spleen cells, and DNA immunization raising both antibodies and CDR2-specific cytotoxic T cells - have all failed to produce cerebellar degeneration. Neither hypothesis in this entry has been shown sufficient in vivo. The animal_models section records these individually; the point of noting it here is that the choice between the two hypotheses is being made on human neuropathology and antigen compartment, not on any positive experimental reconstruction.

Proposed experiments

Multi-case neuropathological series of surface-antigen paraneoplastic cerebellar degeneration

pcd_surface_antigen_autopsy_series

Assemble additional anti-P/Q-VGCC and other surface-antigen PCD autopsy cases and repeat the MHC class I, complement, CD8 and antigen-downregulation assessments, to establish whether the single-case findings generalise.

Readouts

MHC class I upregulation in surface-antigen PCD cerebellum

pathophysiology#Antibody-Mediated Purkinje Cell Dysfunction

Direction: UNCHANGED

Interpretation: Consistent absence of MHC class I upregulation across further cases would confirm that the surface-antigen arm is not a cytotoxic T-cell mechanism.

⚙

Pathophysiology

6

Ectopic Onconeural Antigen Expression by Tumour

The initiating event, and the reason the disease is paraneoplastic rather than primarily neurological. A tumour expresses a protein whose normal expression is restricted to the nervous system. For anti-Yo that protein is CDR2, normally found in the cerebellum and ectopically produced by tumour cells. The immune system meets a neural antigen in a context where it is not tolerated, inside a tumour, and responds to it.

Downstream

Antitumour Immune Response Cross-Reactive with Purkinje Cells The response raised against the tumour antigen is by construction also a response against the identical protein in the cerebellum.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"The underlying mechanism is believed to be an immunological reaction to cerebellar degeneration-related protein 2 (CDR2), a protein usually found in the cerebellum that is ectopically produced by tumor cells."

States the antigen, its normal restriction to the cerebellum, and its ectopic tumour expression - the three facts this node asserts. The source's "believed to be" is preserved.

Antitumour Immune Response Cross-Reactive with Purkinje Cells

The immune response against the onconeural antigen, comprising both a B-cell arm (detectable autoantibody) and a T-cell arm. This node is the branch point of the entry: which downstream arm carries the injury depends on whether the target antigen is intracellular or on the cell surface, and the two arms have different pathology and different reversibility.

Cytotoxic T-cell attack on Purkinje cells presenting an intracellular onconeural antigen Direct antibody-mediated Purkinje cell dysfunction against a surface antigen

CD8-positive cytotoxic T cell CL:0000625

Animal models

Anti-Yo IgG passive transfer to rodent brain and ventricles Mouse fails to recapitulate Anti-Yo IgG delivered directly to the target tissue, alone or with complement or activated mononuclear cells, does not produce Purkinje cell loss.

Active immunization with recombinant Yo protein across four MHC backgrounds Mouse fails to recapitulate A systemic anti-Yo immune response, raised in vivo and MHC-diverse, is not sufficient to produce the disease.

Downstream

Cytotoxic T Cell-Mediated Purkinje Cell Destruction Where the antigen is intracellular, only the T-cell arm can engage it, and the neuropathology is that of a cytotoxic attack.

Antibody-Mediated Purkinje Cell Dysfunction Where the antigen is on the cell surface, the antibody can bind it on a living neuron and interfere with its function directly.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"Although both B- and T-cell abnormalities are seen, there is debate about the relative importance of the autoantibodies and cytotoxic T lymphocytes in the neuronal loss."

Establishes that both arms of the response are present and that their relative contribution is contested, which is why this node branches rather than committing to one effector.

Cytotoxic T Cell-Mediated Purkinje Cell Destruction

The anti-Yo arm. Neuropathology shows the full signature of a cytotoxic T-cell attack: diffuse and widespread Purkinje cell loss, microglial nodules, CD8-positive granzyme-B-positive T cells, and upregulation of MHC class I on neurons, which is what makes a neuron visible to a CD8 T cell in the first place. This is cell killing, and it is why the deficit does not recover.

Cytotoxic T-cell attack on Purkinje cells presenting an intracellular onconeural antigen

Purkinje cell CL:0000121 CD8-positive cytotoxic T cell CL:0000625

cerebellar cortex UBERON:0002129

Animal models

DNA immunization raising both anti-Yo antibodies and CDR2-specific cytotoxic T cells Mouse fails to recapitulate A CDR2-specific cytotoxic T-cell response demonstrably capable of lysing peptide-pulsed targets does not, on its own, produce cerebellar degeneration in vivo.

Downstream

Irreversible Purkinje Cell Loss Cytotoxic killing removes the cell, and Purkinje cells are not replaced.

Prodromal vertigo and imbalance Ongoing cell death before the vulnerable population is exhausted. The edge is drawn from the destruction node rather than from Irreversible Purkinje Cell Loss on purpose: the prodrome is what partial loss looks like, and by the time the loss is complete the patient is no longer prodromal. This is the only clinical feature in the entry that reports on a still-treatable state.

Show evidence (1 reference)

PMID:36070310 SUPPORT Human Clinical

"Anti-Yo-PCD showed characteristic features of a T cell-mediated pathology, whereas this was not observed in 1 case of anti-P/Q-VGCC-PCD."

The direct neuropathological comparison. Note the source states its own denominator, one anti-P/Q-VGCC case, which is why the contrast is curated as a hypothesis rather than settled fact.

Antibody-Mediated Purkinje Cell Dysfunction

The surface-antigen arm, seen with anti-P/Q-VGCC. Purkinje cell loss is focal rather than diffuse, predominantly affecting the upper vermis while caudal regions and lateral hemispheres are spared. There is no MHC class I upregulation and no complement deposition, so this is not a cytotoxic attack. The channel itself is downregulated and synaptophysin reduced, and axonal spheroids with accumulated amyloid precursor protein and GRP78 indicate endoplasmic reticulum stress and failing axonal transport - a picture of neuronal dysfunction preceding neuronal death.

Direct antibody-mediated Purkinje cell dysfunction against a surface antigen

Purkinje cell CL:0000121

cerebellar cortex UBERON:0002129

Animal models

Anti-Yo IgG passive transfer to rodent brain and ventricles Mouse fails to recapitulate Cited as a bounding result for the antibody-effector mechanism generally. Fidelity is LOW because this node is about a surface antigen and the transferred antibody targets an intracellular one.

Downstream

Irreversible Purkinje Cell Loss Sustained dysfunction and calcium dysregulation eventually kill the cell, but the intermediate dysfunctional state is what creates a therapeutic window.

Show evidence (2 references)

PMID:36070310 SUPPORT Human Clinical

"In contrast, PC loss in anti-P/Q-VGCC-PCD was focal and predominantly affected the upper vermis, whereas caudal regions and lateral hemispheres were spared."

The topographical difference from the anti-Yo arm, one of the observations arguing that the two are different mechanisms rather than one mechanism of differing severity.

PMID:36070310 SUPPORT Human Clinical

"No complement deposition or MHC-I upregulation was detected. Moreover, synaptophysin was reduced, and neuronal P/Q-VGCC was downregulated."

The negative findings are the load-bearing part. Absence of MHC class I upregulation is what excludes a cytotoxic T-cell mechanism in this arm.

Irreversible Purkinje Cell Loss

Widespread loss of Purkinje cells, the sole output neuron of the cerebellar cortex. Calbindin expression is reduced or lost even in the Purkinje cells that remain, in both arms, so surviving cells are not necessarily functional cells. Cerebellar atrophy follows on imaging, though it is typically absent early, which is why a normal early MRI does not exclude the diagnosis.

Purkinje cell CL:0000121

cerebellum UBERON:0002037

Downstream

Pancerebellar syndrome Loss of the cerebellar cortex output neuron produces the full cerebellar syndrome.

Cerebellar atrophy Cell loss becomes visible as volume loss on imaging over time.

Gait ataxia Loss of vermal and paravermal Purkinje output produces ataxia of stance and gait.

Truncal ataxia The vermal component of the same cell loss.

Limb ataxia Loss of hemispheric Purkinje output produces appendicular incoordination.

Dysarthria Cerebellar control of speech articulation is lost with the same output neuron.

Nystagmus Loss of cerebellar control of eye movement.

Severe disability The accumulated cerebellar deficit is what leaves most patients unable to walk unassisted. The edge is indirect because disability is the summed consequence of the individual deficits rather than of cell loss acting on it directly.

Show evidence (2 references)

PMID:36070310 SUPPORT Human Clinical

"In both PCD types, calbindin expression was reduced or lost in the remaining PCs."

Records that surviving Purkinje cells are themselves abnormal, in both arms. It is the reason cell counts understate the functional deficit.

PMID:36070310 SUPPORT Human Clinical

"Paraneoplastic cerebellar degeneration (PCD) is characterized by a widespread loss of Purkinje cells (PCs)"

States Purkinje cell loss as the defining pathological feature of the syndrome.

Pancerebellar syndrome

The clinical endpoint - a subacute pancerebellar deficit that progresses and then plateaus, typically within six months. The plateau is itself informative, since it marks the point at which the vulnerable cell population has been exhausted rather than the point at which the immune response stops.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"The typical presentation involves the subacute development of pancerebellar deficits with a clinical plateau within 6 months."

States the tempo and the plateau, which together define the clinical course.

⬡

Pathograph

Use the checkboxes to hide or show graph categories. Hover nodes for evidence and cross-linked metadata.

●

Phenotypes

12

Ear 1

Prodromal vertigo and imbalance FREQUENT HP:0002321

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"Vertigo and imbalance can be present early in the disease course in about two thirds of patients, as a prodromal phase."

States both the frequency and, critically, that it is a prodrome. The review names this as one of its own novel contributions, so it is a pooled finding rather than a restatement of prior consensus.

Eye 1

Nystagmus FREQUENT HP:0000639

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were frequent during the disease course."

As for dysarthria, the oculomotor component is part of the reported pancerebellar pattern rather than a separately tabulated frequency in this review.

Metabolism 2

Increased CSF protein concentration FREQUENT HP:0002922

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and oligoclonal bands, are common in the early stages."

Names the three CSF findings and states that they are early. The same sentence supports the pleocytosis and oligoclonal-band phenotypes curated below; each is bound to its own HPO term rather than to this one.

CSF oligoclonal bands FREQUENT CSF oligoclonal immunoglobulin G bands HP:6000397

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and oligoclonal bands, are common in the early stages."

The source names oligoclonal bands explicitly. Intrathecal synthesis is what makes this distinct from the protein elevation above rather than a second measure of it.

Nervous System 8

Pancerebellar syndrome OBLIGATE Progressive cerebellar ataxia HP:0002073

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"The typical presentation involves the subacute development of pancerebellar deficits with a clinical plateau within 6 months."

States the phenotype and its time course.

Gait ataxia VERY_FREQUENT HP:0002066

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were frequent during the disease course."

Establishes pancerebellar ataxia - of which gait ataxia is the presenting component - as the main feature across 379 pooled patients. The same sentence is the basis for the truncal and limb components curated below, which is why they share it.

Truncal ataxia FREQUENT HP:0002078

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were frequent during the disease course."

The pooled review establishes the pancerebellar pattern. It does not report a separate figure for the truncal component, which is why the frequency is FREQUENT rather than a banded percentage.

Limb ataxia FREQUENT HP:0002070

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were frequent during the disease course."

Same pooled statement of the pancerebellar pattern. Recorded here because the appendicular component is what distinguishes this syndrome from a purely vermal one.

Dysarthria FREQUENT HP:0001260

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were frequent during the disease course."

Pancerebellar involvement by definition includes speech. The review does not report dysarthria as a separately tabulated frequency, so this is curated as a component of the reported pattern rather than as an independently quantified finding.

Cerebellar atrophy FREQUENT HP:0001272

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"Magnetic resonance imaging of the brain is often normal in the early stages, with cerebellar atrophy seen later."

States both the finding and its absence early, which is the clinically important half.

CSF pleocytosis FREQUENT HP:0012229

Show evidence (2 references)

PMID:27606347 SUPPORT Human Clinical

"Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and oligoclonal bands, are common in the early stages."

The source names lymphocytic pleocytosis explicitly, which is what this bound term records.

PMID:34817790 SUPPORT Human Clinical

"Cerebrospinal fluid was inflammatory in 64% of the patients."

Quantifies CSF inflammation, in the anti-Tr/DNER population rather than the anti-Yo one. Recorded because it is the only pooled figure available for this finding, and its population is stated so it is not read as an anti-Yo rate.

Severe disability FREQUENT Inability to walk HP:0002540

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"The prognosis for anti-Yo PCD is almost uniformly poor, with most patients left bedridden."

States the outcome for the anti-Yo form specifically. It should not be read as covering the surface-antigen arm, whose outlook may differ.

🧬

Genetic Associations

4

CDR2 (The onconeural antigen itself. CDR2 is a cytoplasmic Purkinje cell protein ectopically expressed by the tumour and is the target of the anti-Yo response.)

Gene: CDR2 hgnc:1799 relationship_type: BIOMARKER

Show evidence (1 reference)

PMID:11771954 SUPPORT INDIRECT Model Organism

"Paraneoplastic cerebellar degeneration associated with gynecological and breast malignancies (PCD) is known to develop autoantibodies and autoreactive cytotoxic T lymphocytes (CTLs) specific for a cytoplasmic protein of Purkinje cells PCD17/cdr2, in the blood of patients."

Names the antigen and states its cytoplasmic location, the fact the canonical hypothesis rests on, and states that both antibodies and cytotoxic T cells against it are present in patients - which is why the effector question is a controversy here rather than a settled point. Regraded MODEL_ORGANISM in review. The cited paper is a BALB/c mouse DNA-immunisation study, and evidence_source classifies the publication rather than the quoted sentence, even where that sentence is itself reporting a human observation. Graded INDIRECT for the same reason. The claim it supports is not species-contingent and is independently supported by PMID:27606347 elsewhere in this entry.

CDR2L (A CDR2 paralogue proposed as the major anti-Yo antigen in place of, or alongside, CDR2.)

Gene: CDR2L hgnc:29999 relationship_type: BIOMARKER

HLA-DRB1 (The DRB1*13:01~DQA1*01:03~DQB1*06:03 class II haplotype is increased in anti-Yo disease arising from ovarian cancer.)

Gene: HLA-DRB1 hgnc:4948 relationship_type: SUSCEPTIBILITY

Show evidence (2 references)

PMID:29306402 SUPPORT Human Clinical

"Increased DRB1*13:01~DQA1*01:03~DQB1*06:03 was also found in PCD ovarian cases (OR=5.4, p=0.0016)."

The risk association, quoted with its effect size and the tumour type it is specific to.

PMID:29306402 SUPPORT Human Clinical

"These results suggest differential genetic susceptibility to anti-Yo per cancer and with a primary HLA Class II involvement."

The authors' own conclusion, which states both the class II localisation and the per-cancer specificity. Their hedge ("suggest") is preserved.

HLA-DPB1 (The DPA1*01:03~DPB1*04:01 haplotype is protective against anti-Yo disease overall.)

Gene: HLA-DPB1 hgnc:4940 relationship_type: PROTECTIVE

Show evidence (1 reference)

PMID:29306402 SUPPORT Human Clinical

"We detected protective effects of DPA1*01:03~DPB1*04:01 (OR=0, p=0.0008), DRB1*04:01~DQA1*03:03(OR=0, p=0.0016) and DPA1*01:03~DPB1*04:01 (OR=0.35, p=0.0047) overall."

The protective associations with their effect sizes. The sentence carries the same haplotype twice with different odds ratios, which is quoted rather than tidied because it is what makes the zero cells interpretable.

💊

Medical Actions

7

Corticosteroid Therapy

Action: Pharmacotherapy NCIT:C15986

Agent: glucocorticoid CHEBI:24261

Platform: Small molecule

Steroids, usually intravenous methylprednisolone pulses followed by oral maintenance, as first-line immunosuppression. Used extensively and with limited success, and the mechanism explains why: in the anti-Yo form the Purkinje cells are already dead by the time the syndrome is recognised, and no immunosuppression replaces a Purkinje cell.

Mechanism Target:

Antitumour Immune Response Cross-Reactive with Purkinje Cells — Broad suppression of the immune response upstream of the cerebellar injury, which is the only point at which it can act.

Show evidence (2 references)

PMID:27606347 SUPPORT Human Clinical

"Immunotherapies, such as steroids, intravenous immune globulins, and plasma exchange, have been extensively used in managing this condition, with limited success."

States both the practice and its poor result, in the source's own words. The same sentence supports the IVIG and plasma exchange records below; each is curated separately because they act by different routes and are separately bindable.

PMID:41197574 SUPPORT Human Clinical

"Glucocorticoids constituted the most common immunotherapy (n = 29, 80.6 %)."

Quantifies steroid use, in the anti-mGluR1 population. Recorded here because the outcomes in that population are markedly better than in anti-Yo disease, which is the entry's central prognostic claim rather than an incidental difference.

Intravenous Immunoglobulin

Action: Intravenous Immunoglobulin Therapy NCIT:C121331

Platform: Other

Pooled immunoglobulin, used first line alongside or after steroids. Its mechanism in this disease is not established, which is one reason the record is bound to the immunoglobulin therapy action rather than to a named agent.

Mechanism Target:

Antitumour Immune Response Cross-Reactive with Purkinje Cells — Modulates the systemic immune response upstream of the cerebellar injury.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"Immunotherapies, such as steroids, intravenous immune globulins, and plasma exchange, have been extensively used in managing this condition, with limited success."

Names intravenous immune globulin among the treatments used and states the outcome.

Therapeutic Plasma Exchange

Action: Plasmapheresis NCIT:C15304

Platform: Other

Removal of circulating antibody. Its rationale is entirely conditional on the antibody-effector hypothesis, which makes it the treatment whose success or failure carries the most mechanistic information in this entry: if antibodies were the effectors in anti-Yo disease, this should work, and it mostly does not.

Mechanism Target:

Antibody-Mediated Purkinje Cell Dysfunction — Removes the circulating antibody, which is the effector in the surface-antigen arm and the point at which this treatment has a mechanistic rationale.

Show evidence (2 references)

PMID:27606347 SUPPORT Human Clinical

"Immunotherapies, such as steroids, intravenous immune globulins, and plasma exchange, have been extensively used in managing this condition, with limited success."

Names plasma exchange among the treatments used and states the outcome, which is the weak evidence against the antibody-effector reading discussed in this entry.

PMID:32872070 SUPPORT Human Clinical

"After four courses of plasma exchange, the patient could walk independently, the Romberg test was negative, and anti-Yo antibodies were undetectable."

A single case with an unusually good outcome, in an atypical patient - male, and with no tumour found. It is curated because it is a genuine counterexample to the uniformly poor prognosis, and its atypia is stated so it is not read as representative.

B-Cell Depletion

Action: Pharmacotherapy NCIT:C15986

Agent: ofatumumab NCIT:C66952 rituximab NCIT:C1702

Platform: Monoclonal antibody

Anti-CD20 monoclonal antibodies - rituximab, and more recently ofatumumab - as second-line treatment when first-line immunotherapy fails. It is the most mechanistically targeted option available, and like plasma exchange its rationale depends on the antibody arm mattering. The evidence available for this record is two single cases and it is deliberately not presented as more than that. Both are reported as favourable outcomes, and favourable single cases are the ones that get published. Note also that the mGluR1 case is a surface-antigen disease with a much better baseline prognosis, so it should not be read as evidence for B-cell depletion in anti-Yo disease.

Mechanism Target:

Antitumour Immune Response Cross-Reactive with Purkinje Cells — Depletes the B-cell compartment producing the onconeural antibody.

Show evidence (2 references)

PMID:39737186 SUPPORT Human Clinical

"When a second course of methylprednisolone pulse therapy proved ineffective, treatment was switched to ofatumumab. After two doses, the patient achieved partial symptomatic relief."

A single anti-Yo case, quoted with the outcome the authors actually report - partial relief - rather than the "successfully treated" of their title. The distinction matters because this entry's prognosis claims are about degree of recovery.

PMID:40760473 SUPPORT INDIRECT Human Clinical

"As the initial treatment was insufficient, rituximab was administered, leading to significant improvement, including the ability to walk unaided."

A single surface-antigen case. Graded INDIRECT because the inference to anti-Yo disease requires an extra step the case does not support - anti-mGluR1 disease is antibody-mediated against an accessible target and recovers far more often, so a good result there does not predict one here.

Tumour Resection

Action: Surgical Procedure NCIT:C15329

Platform: Surgery

Surgical removal of the underlying malignancy, on the rationale that removing the antigen source should shut down the immune response driving the neurological disease. The rationale is sound and the results are inconsistent, which is itself informative - it suggests the autoimmune response becomes self-sustaining, or that the damage is complete before the tumour is found.

Mechanism Target:

Ectopic Onconeural Antigen Expression by Tumour — Removes the source of the antigen that initiated the cross-reactive response, acting at the trigger node.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"Although some reports indicate benefit from antitumor therapies like surgery and chemotherapy, this has not been consistently observed."

States the inconsistency directly. The claim this record makes is that benefit is reported and unreliable, which is what the source supports.

Antitumour Chemotherapy

Action: Chemotherapy NCIT:C15632

Platform: Small molecule

Systemic treatment of the underlying malignancy, with the same antigen-removal rationale as resection and the same inconsistent neurological result. Curated separately from resection because the two are different modalities with different bindings, not because their evidence differs.

Mechanism Target:

Ectopic Onconeural Antigen Expression by Tumour — Reduces or removes the antigen-expressing tumour burden at the trigger node.

Show evidence (2 references)

PMID:27606347 SUPPORT Human Clinical

"Although some reports indicate benefit from antitumor therapies like surgery and chemotherapy, this has not been consistently observed."

The same sentence names chemotherapy alongside surgery, and states the same inconsistency.

PMID:34817790 SUPPORT Human Clinical

"Oncological response was complete in 88%, but neurological prognosis was poor with only 41% of the patients presenting significant neurological improvement at the last follow up."

The cleanest available dissociation between the two outcomes: in 85 anti-Tr/DNER patients the tumour is controlled in most and the ataxia is not. This is the strongest evidence in the entry that removing the antigen source does not reverse the neurological disease, and it is why the target_mechanisms edge points at the trigger node rather than at the cerebellar injury.

Symptomatic and Supportive Care

Action: Supportive Care NCIT:C15747

Platform: Behavioral / lifestyle

Rehabilitation, mobility aids and management of the established deficit. Curated because it is what most patients actually receive over the long term, and because the entry would otherwise imply that a disease in which most survivors cannot walk unassisted has nothing offered to it after immunotherapy fails. Recorded without an efficacy claim. No cited source in this entry reports a controlled outcome for supportive care in this disease, and none is asserted.

Show evidence (1 reference)

PMID:36334195 SUPPORT Human Clinical

"However, despite treatment, 84% of survivors were unable to walk unassisted on follow-up."

Establishes the residual disability that this record exists to address. It is also the number that makes supportive care the substantive part of long-term management rather than an afterthought.

🌍

Environmental Factors

1

Immune checkpoint inhibitor therapy

The exposure term is broader than the exposure. ECTO has no term for checkpoint inhibitor exposure specifically, and no term for pembrolizumab or nivolumab exposure; exposure to antineoplastic agent is the most specific correct ancestor, with the actual drug class kept in preferred_term. Both searches are recorded here so the binding is not re-derived. This matches the existing binding in Autoimmune_Encephalitis, which uses the same CURIE for the same exposure. just environmental-term-audit reports a name conflict against ECTO:0000509 (exposure to drug); that is a normalisation artifact rather than a disagreement - the other CURIE is bound to generic drug exposures in unrelated entries, not to checkpoint inhibitors. environmental_effect is TRIGGERS rather than the PREDISPOSES used in Autoimmune_Encephalitis. The difference is deliberate: the cited series gives a median of eleven weeks from drug start to neurological onset, which is a triggering interval, where the encephalitis entry records a predisposing relationship without that temporal anchor.

Anti-PD-1, anti-PD-L1 and anti-CTLA-4 therapy can trigger a cerebellar syndrome that is clinically and immunologically continuous with spontaneous paraneoplastic cerebellar degeneration. It is the one exposure in this disease with a mechanism that is not merely plausible but by design: checkpoint inhibitors work by removing the restraint on an antitumour T-cell response, and this entry's trigger node is an antitumour response that has become cross-reactive with Purkinje cells. Curated as an exposure rather than as a subtype because the resulting syndrome is not clearly the same disease - see the discussion pcd_ici_triggered_same_disease.

Show evidence (2 references)

PMID:39153058 SUPPORT Human Clinical

"Neuronal antibodies were detected in 15/31 patients tested (48%)."

Roughly half of checkpoint-inhibitor cerebellar cases carry a neuronal antibody, which is what places the syndrome on the same immunological spectrum as the spontaneous disease rather than making it a separate toxicity.

PMID:37151179 SUPPORT Human Clinical

"CNS manifestations included encephalitis (n = 16), meningoencephalitis (n = 8), cerebellar ataxia (n = 4), demyelinating syndrome (n = 2), and myelopathy (n = 1)."

Places the cerebellar syndrome in proportion within checkpoint-inhibitor CNS autoimmunity - it is a minority manifestation, four of thirty-one, which bears on how often a clinician should expect it.

Mechanism Target:

TRIGGERS Antitumour Immune Response Cross-Reactive with Purkinje Cells — Checkpoint blockade removes the inhibitory restraint on the antitumour T-cell response that this node describes, which is the same response that becomes cross-reactive with Purkinje cells. The exposure acts on the node by design rather than as a side effect of an unrelated mechanism.

Show evidence (1 reference)

PMID:39153058 SUPPORT Human Clinical

"Symptoms developed at a median of 11 weeks after ICI onset."

The temporal relationship between exposure and syndrome onset, across 35 pooled patients. A median of eleven weeks is what makes this a triggering relationship rather than a coincidental one in patients who have cancer anyway.

🔬

Diagnosis

4

Onconeural antibody detection with risk stratification

Detection of a characteristic antibody in serum or CSF. Current criteria classify antibodies by how strongly they predict an underlying cancer rather than treating them as a single category, which changes how a positive result should be acted on.

Show evidence (2 references)

PMID:34006622 SUPPORT Human Clinical

"The term "onconeural antibody" was replaced by "high risk" (>70% associated with cancer) and "intermediate risk" (30%-70% associated with cancer) antibodies."

Gives the quantitative threshold defining the high-risk antibody category that anti-Yo falls into, and the intermediate-risk band alongside it, so the reader can see that the classification is a graded scale rather than a binary.

PMID:34006622 SUPPORT Human Clinical

"we propose updated diagnostic criteria for PNS."

Establishes that this reference is the current criteria statement rather than a review of practice.

Search for an occult malignancy

Investigation for a gynaecological or breast primary in the typical anti-Yo presentation. The neurological syndrome usually precedes the cancer diagnosis, so a negative initial search does not end the question.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"The vast majority of cases have been reported in women with pelvic or breast tumors."

States where to look, and the demographic that makes the search targeted.

Antibody-guided imaging search for the primary tumour

Which antibody is found determines where and how hard to look. Current guidance is that the antibody, more than the clinical syndrome, sets the risk and type of the underlying malignancy, and prescribes a modality sequence per body region rather than a single scan. It also sets a duration: a negative first search is repeated at 3 to 6 months and then six-monthly for up to four years, because the neurological syndrome almost always precedes the cancer.

Show evidence (3 references)

PMID:20880069 SUPPORT Other

"the nature of antibody, and to a lesser extent the clinical syndrome, determines the risk and type of an underlying malignancy."

The principle behind antibody-guided rather than syndrome-guided screening. Graded OTHER because this is a task-force consensus recommendation, and the report states that no evidence-based grade A-C recommendations were possible.

PMID:20880069 SUPPORT Other

"For screening of the thoracic region, a CT-thorax is recommended, which if negative is followed by fluorodeoxyglucose-positron emission tomography (FDG-PET)."

Places FDG-PET as the second-line thoracic modality rather than the first, which is the practical content of the recommendation.

PMID:20880069 SUPPORT Other

"If primary screening is negative, repeat screening after 3-6 months and screen every 6 months up till 4 years."

The surveillance interval and its endpoint. This is the operational answer to the problem stated in the record above - that a negative initial search does not end the question.

FDG-PET/CT when no well-characterised antibody is found

Whole-body FDG-PET/CT in clinically suspected paraneoplastic disease, including patients whose antibody testing is negative. This is the case where antibody-guided screening has nothing to guide it, and it is also where the yield data are weakest.

Show evidence (2 references)

PMID:22157026 SUPPORT Human Clinical

"Of the 27 patients, 6 (22%) had an abnormal FDG uptake. Of these 6 patients, 5 (19%) were histologically confirmed as having a malignancy."

The actual yield, from a 27-patient retrospective single-centre series. A five-cancer denominator is small, and the record is written so the reader sees that rather than inheriting the recommendation without its evidence base.

PMID:22157026 SUPPORT Human Clinical

"PET/CT may be a useful screening tool for patients with clinically suspected PNS who do not exhibit well-characterized paraneoplastic antibodies."

The authors' recommendation with their own hedge ("may") preserved, and scoped to the antibody-negative group where the antibody-guided approach does not apply.

📊

Prevalence

3

Worldwide

Unknown Rare

No rate is established. The condition is described as having a low prevalence that has prevented randomised controlled trials, which is a statement about the size of the evidence base as much as about occurrence.

Show evidence (1 reference)

PMID:27606347 SUPPORT Human Clinical

"The low prevalence of this condition has not allowed for large-scale randomized controlled trials."

A qualitative rarity statement with no measure attached, which is why the record uses measure_type UNKNOWN. It also states the consequence for the treatment evidence in this entry.

Udine, Pordenone and Gorizia provinces, Friuli-Venezia Giulia, Northeastern Italy (2009-2017, 983,190 people)

Annual Incidence 0.89 per 100,000 person-years 1–9 per 100,000 per year

This figure is for paraneoplastic neurological syndromes as a class, not for cerebellar degeneration specifically. It is curated because it is the only population-based estimate available anywhere near this disease, and because a reader otherwise has nothing but the qualitative "rare" above. It must not be read as a PCD incidence. The study does report what share is cerebellar degeneration: 28% of its PNS cases, second to limbic encephalitis at 31%. Multiplying gives roughly 0.25 per 100,000 person-years, but that product is not curated as its own record because no source states it - the 28% is a share of cases, not of incidence, and treating the two as interchangeable assumes a uniform detection rate across syndromes that the study does not establish. The arithmetic is recorded here so a reader can weigh it without it being presented as a finding. The population string and the fraction claim on this record were both corrected in review; the history record for that round carries what changed and why.

Show evidence (1 reference)

PMID:31552550 SUPPORT INDIRECT Human Clinical

"The incidence of PNS was 0.89/100,000 person-years."

The study's own measured incidence, from RESULTS rather than the rounded figure its CONCLUSIONS give. Graded INDIRECT because the denominator is the syndrome class rather than this disease.

Udine, Pordenone and Gorizia provinces, Friuli-Venezia Giulia, Northeastern Italy (2009-2017, 983,190 people)

Point Prevalence 4.37 per 100,000 1–9 per 100,000

The prevalence half of the same population-based study, subject to the same class-level caveat and the same 28% share recorded on the incidence record above. Kept as a separate record rather than merged with the incidence, because prevalence and incidence are different measures and this entry should not invite the comparison.

Show evidence (2 references)

PMID:31552550 SUPPORT INDIRECT Human Clinical

"The prevalence of PNS was 4.37 per 100,000."

The measured prevalence this record carries. Graded INDIRECT because the denominator is the syndrome class rather than this disease.

PMID:31552550 SUPPORT INDIRECT Human Clinical

"Most common PNS were limbic encephalitis (31%), cerebellar degeneration (28%) and encephalomyelitis (20%)."

The syndrome breakdown behind the 28% share this record's notes rely on, so the caveat is supported on the record that states it rather than only on the incidence record above.

🐁

Animal Models

3

Anti-Yo IgG passive transfer to rodent brain and ventricles

Patient anti-Yo IgG injected into mouse brain or, with complement or activated mononuclear cells, into rat ventricles. The antibody is taken up by Purkinje cells and stays there, and the cells do not die. This is a negative result and it is curated as the load-bearing evidence for the entry's canonical hypothesis rather than as a footnote. The uptake half matters as much as the failure half: it rules out the trivial explanation that the antibody simply never reached its target.

Species

Mouse

Genotype

Wild-type, various strains; also rat recipients

Publication

PMID:7707074

Pathograph links

Antitumour Immune Response Cross-Reactive with Purkinje Cells fails to recapitulate fidelity: moderate observes: tissue Anti-Yo IgG delivered directly to the target tissue, alone or with complement or activated mononuclear cells, does not produce Purkinje cell loss.

Limitations: A negative transfer result is weaker than it looks, and the specific weaknesses are worth naming. Rodent Purkinje cells may not present the human epitope in the way human cells do; a single injection over 36 hours is a fraction of the exposure a patient has over months; and intracerebral or intraventricular delivery is not the route the antibody takes in disease. What the experiment establishes is that antibody plus effector is not sufficient on the timescale tested, not that antibody is irrelevant.

Divergences from the human mechanism

temporal scope qualifying

The observation window is 36 hours after a single injection. Human disease develops over weeks to months of continuous antibody exposure, so the model tests acute toxicity rather than the chronic process.

species mismatch qualifying

Rodent Purkinje cells are the recipients of a human autoantibody. The antibody does bind them, which limits how far this divergence can be pressed, but epitope presentation and downstream effector engagement need not match the human cell.

Readouts

Purkinje cell survival after intracerebral anti-Yo IgG

Direction: UNCHANGED

Interpretation: The measurement that carries the negative result. UNCHANGED here is the finding, not a missing datum.

Show evidence (1 reference)

PMID:7707074 SUPPORT Model Organism

"On a single injection of PCD IgG to mouse brain, IgG was taken into Purkinje cells and remained there for more than 36 h without Purkinje cell loss."

Records both halves - the antibody reached and entered the target cell, and the cell survived. The uptake is what makes the negative result interpretable.

Show evidence (1 reference)

PMID:7707074 REFUTE Model Organism

"The anti-Yo antibody, with or without complement or activated mononuclear cells, therefore could not be the sole cause of Purkinje cell loss."

The authors' conclusion, graded REFUTE against the antibody-effector reading of this node. Note the claim they make is bounded - "sole cause" - and the entry does not stretch it into a claim that the antibody plays no part.

Antibody-Mediated Purkinje Cell Dysfunction fails to recapitulate fidelity: low observes: tissue Cited as a bounding result for the antibody-effector mechanism generally. Fidelity is LOW because this node is about a surface antigen and the transferred antibody targets an intracellular one.

Limitations: The antibody transferred here is anti-Yo, against the intracellular CDR2. This node describes an antibody against a cell-surface channel, which is a different mechanism with a different physical basis. A failure to transfer intracellular-antigen disease says little about whether surface-antigen disease is antibody-mediated, and the link is recorded only so that a reader does not mistake the anti-Yo negative for a general refutation of antibody effectors.

Divergences from the human mechanism

proxy quantity invalidating

The quantity transferred is an antibody against an intracellular antigen; the quantity this node is about is an antibody against a surface antigen. The two differ in whether the target is physically accessible to the antibody in a living cell, which is the whole mechanistic distinction the entry draws.

Show evidence (1 reference)

PMID:7707074 NO_EVIDENCE Model Organism

"The anti-Yo antibody, with or without complement or activated mononuclear cells, therefore could not be the sole cause of Purkinje cell loss."

Graded NO_EVIDENCE against this node deliberately. The experiment is real and its conclusion is sound, and it does not bear on a surface-antigen mechanism at all. Recording that explicitly is the point of the link.

Active immunization with recombinant Yo protein across four MHC backgrounds

Mice immunized with recombinant Yo protein raise high anti-Yo titres and reactive spleen cells across four MHC backgrounds, and none develops ataxia or loses Purkinje cells. Using four haplotypes is what makes this more than one failed experiment - it rules out the explanation that a single strain simply lacked the right restriction element.

Species

Mouse

Genotype

Wild-type BALB/c (H-2d), C3H (H-2k), C57BL/6 (H-2b), SJL/J (H-2s); also C57BL/6 nu/nu

Background

Four inbred strains bearing different MHC haplotypes

Publication

PMID:7788981

Pathograph links

Antitumour Immune Response Cross-Reactive with Purkinje Cells fails to recapitulate fidelity: moderate observes: organism A systemic anti-Yo immune response, raised in vivo and MHC-diverse, is not sufficient to produce the disease.

Limitations: The immunization raises a peripheral response. Nothing in the protocol breaks the blood-brain barrier, provides the tumour that in patients precedes the neurological disease, or supplies whatever second event lets a peripheral response reach the cerebellum. The authors draw exactly this conclusion, so the model's failure is informative about sufficiency rather than about the mechanism being wrong.

Divergences from the human mechanism

cause unrepresented qualifying

There is no tumour in this model. In patients the antigen is presented by a malignancy, which is also the setting in which the response is shaped and, plausibly, the setting that licenses access to the CNS. Immunization with recombinant protein represents the antigen without representing that context.

Readouts

Cerebellar ataxia and Purkinje cell loss after Yo immunization

Direction: UNCHANGED

Interpretation: The negative outcome, measured across four MHC haplotypes rather than one.

Show evidence (1 reference)

PMID:7788981 SUPPORT Model Organism

"All the strains produced high anti-Yo antibody titer but none developed cerebellar ataxia or showed Purkinje cell loss."

The measurement, with the antibody response confirmed so the failure cannot be attributed to a failed immunization.

Show evidence (2 references)

PMID:7788981 REFUTE Model Organism

"Results suggest that antibody production in peripheral blood alone is not sufficient for the development of PCD and that MHC class II molecules function in the activation of T cells to help B cells to help B cells produce antibodies."

The authors' conclusion, graded REFUTE against a peripheral-antibody account of this node. Quoted as written, including the duplicated clause in the source.

+ 1 more reference

DNA immunization raising both anti-Yo antibodies and CDR2-specific cytotoxic T cells

The most demanding of the four transfer and immunization experiments, and the one this entry treats as most informative. DNA immunization generated both arms of the human immune response at once - anti-Purkinje cell autoantibodies and CDR2-specific cytotoxic T lymphocytes able to lyse peptide-pulsed targets - and still produced no clinical or pathological cerebellar degeneration. It is curated because it does not favour either hypothesis in this entry. The other three experiments fail to transfer antibody-mediated disease and so support the T-cell reading by elimination. This one raises the T cells too, and still fails.

Species

Mouse

Genotype

Wild-type BALB/c immunized with naked pcd17 (CDR2) cDNA

Publication

PMID:11771954

Pathograph links

Cytotoxic T Cell-Mediated Purkinje Cell Destruction fails to recapitulate fidelity: moderate observes: organism A CDR2-specific cytotoxic T-cell response demonstrably capable of lysing peptide-pulsed targets does not, on its own, produce cerebellar degeneration in vivo.

Limitations: The T cells were shown to lyse syngeneic myeloma cells pulsed with peptide - an artificial target presenting the epitope at high density. That is not evidence they could recognise a Purkinje cell, which presents endogenous antigen and which upregulates MHC class I in human disease but need not do so in an unmanipulated mouse. So the experiment shows that raising CDR2-specific CTLs is not sufficient; it does not show that the CTLs generated were the same effector population as in patients.

Divergences from the human mechanism

proxy quantity qualifying

Cytotoxic capability is measured against peptide-pulsed myeloma cells, not against Purkinje cells. The measured quantity is epitope-specific lysis in vitro; the quantity of interest is killing of an endogenously presenting neuron in vivo.

cause unrepresented qualifying

No tumour and no MHC class I upregulation on cerebellar neurons - the two features that in human neuropathology make the Purkinje cell a visible target for a CD8 T cell.

Readouts

Cerebellar degeneration after CDR2 DNA immunization

Direction: UNCHANGED

Interpretation: No degeneration, despite both effector arms being present. This is the strongest single argument in the entry that the mechanism is not yet fully accounted for by either hypothesis.

Show evidence (1 reference)

PMID:11771954 REFUTE Model Organism

"In spite of the generation of anti-Purkinje cell antibodies and PCD17-specific CTLs in vivo, neither clinical nor pathological changes consistent with significant cerebellar degeneration have been detected."

The negative result stated with both arms confirmed present. Graded REFUTE because it cuts against the sufficiency of the cytotoxic mechanism this node asserts, which is a different and weaker claim than refuting the mechanism itself.

Show evidence (1 reference)

PMID:11771954 SUPPORT Model Organism

"We show that DNA immunization with naked pcd17 cDNA could induce autoantibodies against the cytoplasmic protein of Purkinje cells and CTLs could lyse syngenic myeloma cells pulsed with H-2K-restricted PCD17 peptide."

Establishes that the model genuinely produced both effector arms, which is what makes the accompanying negative outcome worth citing at all. Graded SUPPORT because this sentence supports treating the model as informative for this node, separately from what the model then showed.

Notes

Read together with the two transfer experiments above, these four studies say something the entry's two hypotheses do not: no experimental reconstruction of either arm of the human immune response has produced the disease in an animal. That is recorded as a knowledge gap rather than resolved, and it is the reason the canonical hypothesis is argued from human neuropathology and from the antigen's compartment rather than from a model.

{ }

Source YAML

click to show

name: Paraneoplastic Cerebellar Degeneration
creation_date: '2026-09-06T08:15:00Z'
category: Complex
classifications:
  harrisons_chapter:
  - classification_value: NEUROLOGIC
    notes: The disease is a cerebellar degeneration, and everything that makes it clinically
      devastating is neurological. This is where a reader looking for the disease itself
      would arrive.
  - classification_value: ONCOLOGY_HEMATOLOGY
    notes: 'The second Part is assigned because this is a paraneoplastic syndrome, which the
      classification guidance treats as the case where a cancer entry also warrants an
      organ-system Part - here read in the other direction. It is not decorative: the
      neurological syndrome is frequently what leads to the cancer diagnosis rather than
      following it, so the oncological reader is not a secondary audience. The antibody
      profile also points at the tumour type, which is a diagnostic use rather than a
      mechanistic one.'
    evidence:
    - reference: PMID:29306402
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: Anti-Yo paraneoplastic cerebellar degeneration (PCD) is a devastating
        autoimmune complication of gynecological cancers.
      explanation: States the disease as a complication of an underlying malignancy, which is
        what makes the oncology Part apply as well as the neurological one. The quote is
        specific to the anti-Yo form, which is this entry's principal subject.
disease_term:
  preferred_term: paraneoplastic cerebellar degeneration
  term:
    id: MONDO:0044877
    label: paraneoplastic cerebellar degeneration
synonyms:
- PCD
- subacute cerebellar degeneration
- anti-Yo paraneoplastic cerebellar degeneration
- paraneoplastic cerebellar syndrome
description: 'A subacute pancerebellar syndrome in which an immune response raised against a tumour
  destroys Purkinje cells. The tumour ectopically expresses a protein that is normally restricted to the
  cerebellum, the immune system attacks it, and the cerebellum is collateral damage. The archetype is
  anti-Yo, directed against cerebellar degeneration-related protein 2 (CDR2), in women with
  gynaecological or breast cancer.

  The entry is built around a distinction that a purely clinical description would hide. Neuropathology
  shows that PCD is not one mechanism but at least two, and which one applies depends on where the
  target antigen sits.

  When the antigen is intracellular (anti-Yo), the pathology is that of a cytotoxic T-cell attack:
  diffuse and widespread Purkinje cell loss, microglial nodules, CD8-positive granzyme-B-positive T
  cells, and upregulation of MHC class I on neurons. An antibody cannot reach an intracellular antigen
  in a living cell, so on this arm the antibody is a marker of the response rather than its effector.

  When the antigen is on the cell surface (anti-P/Q-type voltage-gated calcium channel), the picture
  differs in nearly every respect: Purkinje cell loss is focal rather than diffuse, sparing caudal
  regions and lateral hemispheres, there is no MHC class I upregulation and no complement deposition,
  the channel itself is downregulated, and there is evidence of endoplasmic reticulum stress and
  impaired axonal transport. Here the antibody plausibly is the agent.

  That distinction has a direct clinical consequence, and it is the reason this entry exists separately
  from the general paraneoplastic entry. Cells killed by cytotoxic T cells do not come back, which is
  why anti-Yo PCD has an almost uniformly poor prognosis whatever is done. Cells made dysfunctional by
  antibody binding might recover if the antibody is removed early, which is why the surface-antigen arm
  has a therapeutic window that the intracellular arm does not.'
parents:
- Paraneoplastic Neurological Syndrome
- Acquired Cerebellar Ataxia
mechanistic_hypotheses:
- hypothesis_group_id: pcd_t_cell_mediated_intracellular
  hypothesis_label: Cytotoxic T-cell attack on Purkinje cells presenting an intracellular onconeural
    antigen
  status: CANONICAL
  description: In anti-Yo and other intracellular-antigen forms, the effector is a cytotoxic CD8 T-cell
    response and the antibody is a diagnostic marker rather than the agent of injury. The strongest
    support is neuropathological - MHC class I upregulation on neurons, CD8-positive granzyme-B-positive
    T cells, and microglial nodules are the signature of a T cell-mediated attack, and an antibody
    cannot reach an intracellular antigen in a living cell.
  evidence:
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Although the intracellular location of the target antigen in anti-Yo-PCD supports a T
      cell-mediated pathology, the immune mechanisms in anti-P/Q-VGCC-PCD remain unclear.
    explanation: States the argument from antigen compartment for this hypothesis, and in the same
      sentence marks the surface-antigen arm as unresolved.
- hypothesis_group_id: pcd_antibody_mediated_surface
  hypothesis_label: Direct antibody-mediated Purkinje cell dysfunction against a surface antigen
  status: ALTERNATIVE
  description: 'In surface-antigen forms the autoantibody plausibly causes the disease directly, by
    acting on a target it can physically reach in a living cell. In anti-P/Q-VGCC disease that is
    altered synaptic transmission at the channel, producing calcium dysregulation and eventual Purkinje
    cell death. It is curated as ALTERNATIVE rather than as a competing account of the same disease,
    because it describes a different antigen class - the two hypotheses are not rivals for one mechanism
    so much as two mechanisms sharing a syndrome.

    The compartment rule is not built on the anti-P/Q-VGCC case. Anti-Tr/DNER and anti-mGluR1 are the
    other well-characterised surface-antigen cerebellar syndromes, and they behave as the rule predicts:
    anti-Tr binds the extracellular domain of DNER, and anti-mGluR1 disease responds to immunotherapy in
    most patients where anti-Yo disease does not. What rests on the single case is the neuropathological
    contrast - MHC class I, complement deposition, topography of cell loss - and that narrower claim is
    where the knowledge gap in this entry is scoped.'
  evidence:
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Our findings support a pathogenic role of anti-P/Q-VGCC autoantibodies in causing neuronal
      dysfunction, probably due to altered synaptic transmission resulting in calcium dysregulation and
      subsequent PC death.
    explanation: The authors' statement of this hypothesis, with their own hedging ("probably")
      preserved. The evidence base is a single autopsy case, which is recorded in a discussion.
  - reference: PMID:25745634
    reference_title: Standardized test for anti-Tr/DNER in patients with paraneoplastic cerebellar
      degeneration.
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: Anti-Tr antibodies bind to the extracellular domain of DNER and can be detected by RC-IFA
      using HEK293 cells expressing the recombinant receptor.
    explanation: Establishes that a second onconeural antibody in this syndrome targets an
      extracellular domain, which is the physical precondition this hypothesis rests on. Graded
      IN_VITRO because the localisation is established by a recombinant cell-based assay, not in
      patients.
  - reference: PMID:41197574
    reference_title: 'Clinical characteristics, treatment, and outcomes in anti-metabotropic glutamate
      receptor-1 encephalitis: A systematic review.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: Immunotherapy yields favorable responses in most patients, with generally good long-term
      outcomes.
    explanation: A third surface-antigen syndrome, and the treatment-response prediction of this
      hypothesis borne out across 42 pooled cases. Graded INDIRECT because a good response to
      immunotherapy is consistent with an antibody-mediated mechanism without demonstrating one - the
      inference runs through the assumption that immunotherapy acts by removing the antibody.
prevalence:
- population: Worldwide
  measure_type: UNKNOWN
  prevalence_class: RARE
  notes: No rate is established. The condition is described as having a low prevalence that has
    prevented randomised controlled trials, which is a statement about the size of the evidence base as
    much as about occurrence.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The low prevalence of this condition has not allowed for large-scale randomized controlled
      trials.
    explanation: A qualitative rarity statement with no measure attached, which is why the record uses
      measure_type UNKNOWN. It also states the consequence for the treatment evidence in this entry.
- population: Udine, Pordenone and Gorizia provinces, Friuli-Venezia Giulia, Northeastern Italy
    (2009-2017, 983,190 people)
  measure_type: ANNUAL_INCIDENCE
  prevalence_class: BAND_1_9_PER_100000
  rate_per_100000: 0.89
  rate_denominator: PERSON_YEARS
  notes: 'This figure is for paraneoplastic neurological syndromes as a class, not for cerebellar
    degeneration specifically. It is curated because it is the only population-based estimate available
    anywhere near this disease, and because a reader otherwise has nothing but the qualitative "rare"
    above. It must not be read as a PCD incidence.

    The study does report what share is cerebellar degeneration: 28% of its PNS cases, second to limbic
    encephalitis at 31%. Multiplying gives roughly 0.25 per 100,000 person-years, but that product is
    not curated as its own record because no source states it - the 28% is a share of cases, not of
    incidence, and treating the two as interchangeable assumes a uniform detection rate across
    syndromes that the study does not establish. The arithmetic is recorded here so a reader can weigh
    it without it being presented as a finding.

    The population string and the fraction claim on this record were both corrected in review; the
    history record for that round carries what changed and why.'
  evidence:
  - reference: PMID:31552550
    reference_title: 'Epidemiology of paraneoplastic neurological syndromes: a population-based study.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: The incidence of PNS was 0.89/100,000 person-years.
    explanation: The study's own measured incidence, from RESULTS rather than the rounded figure its
      CONCLUSIONS give. Graded INDIRECT because the denominator is the syndrome class rather than this
      disease.
- population: Udine, Pordenone and Gorizia provinces, Friuli-Venezia Giulia, Northeastern Italy
    (2009-2017, 983,190 people)
  measure_type: POINT_PREVALENCE
  prevalence_class: BAND_1_9_PER_100000
  rate_per_100000: 4.37
  rate_denominator: POPULATION
  notes: The prevalence half of the same population-based study, subject to the same class-level caveat
    and the same 28% share recorded on the incidence record above. Kept as a separate record rather
    than merged with the incidence, because prevalence and incidence are different measures and this
    entry should not invite the comparison.
  evidence:
  - reference: PMID:31552550
    reference_title: 'Epidemiology of paraneoplastic neurological syndromes: a population-based study.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: The prevalence of PNS was 4.37 per 100,000.
    explanation: The measured prevalence this record carries. Graded INDIRECT because the denominator is
      the syndrome class rather than this disease.
  - reference: PMID:31552550
    reference_title: 'Epidemiology of paraneoplastic neurological syndromes: a population-based study.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: Most common PNS were limbic encephalitis (31%), cerebellar degeneration (28%) and
      encephalomyelitis (20%).
    explanation: The syndrome breakdown behind the 28% share this record's notes rely on, so the caveat
      is supported on the record that states it rather than only on the incidence record above.
pathophysiology:
- name: Ectopic Onconeural Antigen Expression by Tumour
  role: trigger
  biological_scale: MOLECULAR
  description: 'The initiating event, and the reason the disease is paraneoplastic rather than primarily
    neurological. A tumour expresses a protein whose normal expression is restricted to the nervous
    system. For anti-Yo that protein is CDR2, normally found in the cerebellum and ectopically produced
    by tumour cells. The immune system meets a neural antigen in a context where it is not tolerated,
    inside a tumour, and responds to it.'
  downstream:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    causal_link_type: DIRECT
    description: The response raised against the tumour antigen is by construction also a response
      against the identical protein in the cerebellum.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The underlying mechanism is believed to be an immunological reaction to cerebellar
      degeneration-related protein 2 (CDR2), a protein usually found in the cerebellum that is
      ectopically produced by tumor cells.
    explanation: States the antigen, its normal restriction to the cerebellum, and its ectopic tumour
      expression - the three facts this node asserts. The source's "believed to be" is preserved.
- name: Antitumour Immune Response Cross-Reactive with Purkinje Cells
  role: central_effector
  biological_scale: CELLULAR
  description: 'The immune response against the onconeural antigen, comprising both a B-cell arm
    (detectable autoantibody) and a T-cell arm. This node is the branch point of the entry: which
    downstream arm carries the injury depends on whether the target antigen is intracellular or on the
    cell surface, and the two arms have different pathology and different reversibility.'
  cell_types:
  - preferred_term: CD8-positive cytotoxic T cell
    term:
      id: CL:0000625
      label: CD8-positive, alpha-beta T cell
  downstream:
  - target: Cytotoxic T Cell-Mediated Purkinje Cell Destruction
    causal_link_type: DIRECT
    hypothesis_groups:
    - pcd_t_cell_mediated_intracellular
    description: Where the antigen is intracellular, only the T-cell arm can engage it, and the
      neuropathology is that of a cytotoxic attack.
  - target: Antibody-Mediated Purkinje Cell Dysfunction
    causal_link_type: DIRECT
    hypothesis_groups:
    - pcd_antibody_mediated_surface
    description: Where the antigen is on the cell surface, the antibody can bind it on a living neuron
      and interfere with its function directly.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Although both B- and T-cell abnormalities are seen, there is debate about the relative
      importance of the autoantibodies and cytotoxic T lymphocytes in the neuronal loss.
    explanation: Establishes that both arms of the response are present and that their relative
      contribution is contested, which is why this node branches rather than committing to one effector.
- name: Cytotoxic T Cell-Mediated Purkinje Cell Destruction
  role: central_effector
  biological_scale: TISSUE
  description: 'The anti-Yo arm. Neuropathology shows the full signature of a cytotoxic T-cell attack:
    diffuse and widespread Purkinje cell loss, microglial nodules, CD8-positive granzyme-B-positive T
    cells, and upregulation of MHC class I on neurons, which is what makes a neuron visible to a CD8 T
    cell in the first place. This is cell killing, and it is why the deficit does not recover.'
  cell_types:
  - preferred_term: Purkinje cell
    term:
      id: CL:0000121
      label: Purkinje cell
  - preferred_term: CD8-positive cytotoxic T cell
    term:
      id: CL:0000625
      label: CD8-positive, alpha-beta T cell
  locations:
  - preferred_term: cerebellar cortex
    term:
      id: UBERON:0002129
      label: cerebellar cortex
  downstream:
  - target: Irreversible Purkinje Cell Loss
    causal_link_type: DIRECT
    hypothesis_groups:
    - pcd_t_cell_mediated_intracellular
    description: Cytotoxic killing removes the cell, and Purkinje cells are not replaced.
  - target: Prodromal vertigo and imbalance
    causal_link_type: DIRECT
    hypothesis_groups:
    - pcd_t_cell_mediated_intracellular
    description: 'Ongoing cell death before the vulnerable population is exhausted. The edge is drawn
      from the destruction node rather than from Irreversible Purkinje Cell Loss on purpose: the prodrome
      is what partial loss looks like, and by the time the loss is complete the patient is no longer
      prodromal. This is the only clinical feature in the entry that reports on a still-treatable state.'
  evidence:
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Anti-Yo-PCD showed characteristic features of a T cell-mediated pathology, whereas this was
      not observed in 1 case of anti-P/Q-VGCC-PCD.
    explanation: The direct neuropathological comparison. Note the source states its own denominator, one
      anti-P/Q-VGCC case, which is why the contrast is curated as a hypothesis rather than settled fact.
- name: Antibody-Mediated Purkinje Cell Dysfunction
  role: central_effector
  biological_scale: TISSUE
  description: 'The surface-antigen arm, seen with anti-P/Q-VGCC. Purkinje cell loss is focal rather than
    diffuse, predominantly affecting the upper vermis while caudal regions and lateral hemispheres are
    spared. There is no MHC class I upregulation and no complement deposition, so this is not a cytotoxic
    attack. The channel itself is downregulated and synaptophysin reduced, and axonal spheroids with
    accumulated amyloid precursor protein and GRP78 indicate endoplasmic reticulum stress and failing
    axonal transport - a picture of neuronal dysfunction preceding neuronal death.'
  cell_types:
  - preferred_term: Purkinje cell
    term:
      id: CL:0000121
      label: Purkinje cell
  locations:
  - preferred_term: cerebellar cortex
    term:
      id: UBERON:0002129
      label: cerebellar cortex
  downstream:
  - target: Irreversible Purkinje Cell Loss
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    hypothesis_groups:
    - pcd_antibody_mediated_surface
    description: Sustained dysfunction and calcium dysregulation eventually kill the cell, but the
      intermediate dysfunctional state is what creates a therapeutic window.
  evidence:
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: In contrast, PC loss in anti-P/Q-VGCC-PCD was focal and predominantly affected the upper
      vermis, whereas caudal regions and lateral hemispheres were spared.
    explanation: The topographical difference from the anti-Yo arm, one of the observations arguing that
      the two are different mechanisms rather than one mechanism of differing severity.
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: No complement deposition or MHC-I upregulation was detected. Moreover, synaptophysin was
      reduced, and neuronal P/Q-VGCC was downregulated.
    explanation: The negative findings are the load-bearing part. Absence of MHC class I upregulation is
      what excludes a cytotoxic T-cell mechanism in this arm.
- name: Irreversible Purkinje Cell Loss
  role: consequence
  biological_scale: TISSUE
  description: 'Widespread loss of Purkinje cells, the sole output neuron of the cerebellar cortex.
    Calbindin expression is reduced or lost even in the Purkinje cells that remain, in both arms, so
    surviving cells are not necessarily functional cells. Cerebellar atrophy follows on imaging, though
    it is typically absent early, which is why a normal early MRI does not exclude the diagnosis.'
  cell_types:
  - preferred_term: Purkinje cell
    term:
      id: CL:0000121
      label: Purkinje cell
  locations:
  - preferred_term: cerebellum
    term:
      id: UBERON:0002037
      label: cerebellum
  downstream:
  - target: Pancerebellar syndrome
    causal_link_type: DIRECT
    description: Loss of the cerebellar cortex output neuron produces the full cerebellar syndrome.
  - target: Cerebellar atrophy
    causal_link_type: DIRECT
    description: Cell loss becomes visible as volume loss on imaging over time.
  - target: Gait ataxia
    causal_link_type: DIRECT
    description: Loss of vermal and paravermal Purkinje output produces ataxia of stance and gait.
  - target: Truncal ataxia
    causal_link_type: DIRECT
    description: The vermal component of the same cell loss.
  - target: Limb ataxia
    causal_link_type: DIRECT
    description: Loss of hemispheric Purkinje output produces appendicular incoordination.
  - target: Dysarthria
    causal_link_type: DIRECT
    description: Cerebellar control of speech articulation is lost with the same output neuron.
  - target: Nystagmus
    causal_link_type: DIRECT
    description: Loss of cerebellar control of eye movement.
  - target: Severe disability
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    description: The accumulated cerebellar deficit is what leaves most patients unable to walk
      unassisted. The edge is indirect because disability is the summed consequence of the individual
      deficits rather than of cell loss acting on it directly.
  evidence:
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: In both PCD types, calbindin expression was reduced or lost in the remaining PCs.
    explanation: Records that surviving Purkinje cells are themselves abnormal, in both arms. It is the
      reason cell counts understate the functional deficit.
  - reference: PMID:36070310
    reference_title: Paraneoplastic Cerebellar Degeneration With P/Q-VGCC vs Yo Autoantibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Paraneoplastic cerebellar degeneration (PCD) is characterized by a widespread loss of
      Purkinje cells (PCs)
    explanation: States Purkinje cell loss as the defining pathological feature of the syndrome.
- name: Pancerebellar syndrome
  role: outcome
  biological_scale: ORGANISM
  description: The clinical endpoint - a subacute pancerebellar deficit that progresses and then
    plateaus, typically within six months. The plateau is itself informative, since it marks the point at
    which the vulnerable cell population has been exhausted rather than the point at which the immune
    response stops.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The typical presentation involves the subacute development of pancerebellar deficits with a
      clinical plateau within 6 months.
    explanation: States the tempo and the plateau, which together define the clinical course.
phenotypes:
- category: Neurologic
  name: Pancerebellar syndrome
  frequency: OBLIGATE
  description: Subacute progressive ataxia affecting gait, limbs, speech and eye movements together,
    reaching a plateau within about six months.
  phenotype_term:
    preferred_term: Subacute pancerebellar syndrome
    term:
      id: HP:0002073
      label: Progressive cerebellar ataxia
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The typical presentation involves the subacute development of pancerebellar deficits with a
      clinical plateau within 6 months.
    explanation: States the phenotype and its time course.
- category: Neurologic
  name: Gait ataxia
  frequency: VERY_FREQUENT
  description: Ataxia of stance and gait, part of the pancerebellar deficit and usually the symptom that
    brings the patient in. Curated separately from the syndrome node because it is what the disability
    outcome is measured on.
  phenotype_term:
    preferred_term: Gait ataxia
    term:
      id: HP:0002066
      label: Gait ataxia
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were
      frequent during the disease course.
    explanation: Establishes pancerebellar ataxia - of which gait ataxia is the presenting component - as
      the main feature across 379 pooled patients. The same sentence is the basis for the truncal and
      limb components curated below, which is why they share it.
- category: Neurologic
  name: Truncal ataxia
  frequency: FREQUENT
  description: Ataxia of the trunk, reflecting vermal involvement. Part of the same pancerebellar
    deficit.
  phenotype_term:
    preferred_term: Truncal ataxia
    term:
      id: HP:0002078
      label: Truncal ataxia
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were
      frequent during the disease course.
    explanation: The pooled review establishes the pancerebellar pattern. It does not report a separate
      figure for the truncal component, which is why the frequency is FREQUENT rather than a banded
      percentage.
- category: Neurologic
  name: Limb ataxia
  frequency: FREQUENT
  description: Appendicular incoordination, reflecting hemispheric involvement. Its presence alongside
    truncal and gait ataxia is what makes the syndrome pancerebellar rather than focal.
  phenotype_term:
    preferred_term: Limb ataxia
    term:
      id: HP:0002070
      label: Limb ataxia
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were
      frequent during the disease course.
    explanation: Same pooled statement of the pancerebellar pattern. Recorded here because the
      appendicular component is what distinguishes this syndrome from a purely vermal one.
- category: Neurologic
  name: Dysarthria
  frequency: FREQUENT
  description: Cerebellar dysarthria, the speech component of the pancerebellar deficit.
  phenotype_term:
    preferred_term: Dysarthria
    term:
      id: HP:0001260
      label: Dysarthria
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were
      frequent during the disease course.
    explanation: Pancerebellar involvement by definition includes speech. The review does not report
      dysarthria as a separately tabulated frequency, so this is curated as a component of the reported
      pattern rather than as an independently quantified finding.
- category: Neurologic
  name: Nystagmus
  frequency: FREQUENT
  description: Oculomotor involvement, the eye-movement component of the pancerebellar deficit.
  phenotype_term:
    preferred_term: Nystagmus
    term:
      id: HP:0000639
      label: Nystagmus
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Pancerebellar ataxia was the main clinical feature, but extracerebellar findings were
      frequent during the disease course.
    explanation: As for dysarthria, the oculomotor component is part of the reported pancerebellar
      pattern rather than a separately tabulated frequency in this review.
- category: Neurologic
  name: Prodromal vertigo and imbalance
  frequency: FREQUENT
  description: 'Vertigo and imbalance preceding the established cerebellar syndrome in about two thirds
    of patients. This is the phenotype with the most clinical leverage in the entry: it is the only
    feature that appears while Purkinje cells may still be alive, and every treatment argument in this
    entry turns on whether the patient is seen before the cell population is exhausted. It is also the
    presentation most likely to be attributed to a peripheral vestibular cause and not investigated.'
  phenotype_term:
    preferred_term: Vertigo
    term:
      id: HP:0002321
      label: Vertigo
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Vertigo and imbalance can be present early in the disease course in about two thirds of
      patients, as a prodromal phase.
    explanation: States both the frequency and, critically, that it is a prodrome. The review names this
      as one of its own novel contributions, so it is a pooled finding rather than a restatement of prior
      consensus.
- category: Neurologic
  name: Cerebellar atrophy
  frequency: FREQUENT
  description: Volume loss on imaging, absent in the early stages and appearing later. The early normal
    scan is a diagnostic trap.
  phenotype_term:
    preferred_term: Cerebellar atrophy
    term:
      id: HP:0001272
      label: Cerebellar atrophy
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Magnetic resonance imaging of the brain is often normal in the early stages, with cerebellar
      atrophy seen later.
    explanation: States both the finding and its absence early, which is the clinically important half.
- category: Laboratory
  name: Increased CSF protein concentration
  frequency: FREQUENT
  description: Raised CSF protein, common early and tending to subside as the disease plateaus.
  phenotype_term:
    preferred_term: Increased CSF protein concentration
    term:
      id: HP:0002922
      label: Increased CSF protein concentration
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and
      oligoclonal bands, are common in the early stages.
    explanation: Names the three CSF findings and states that they are early. The same sentence supports
      the pleocytosis and oligoclonal-band phenotypes curated below; each is bound to its own HPO term
      rather than to this one.
- category: Laboratory
  name: CSF pleocytosis
  frequency: FREQUENT
  description: Lymphocytic pleocytosis in the CSF. It is the finding that most directly supports an
    inflammatory rather than degenerative process, which is why it is curated as its own phenotype rather
    than folded into a CSF summary.
  phenotype_term:
    preferred_term: CSF pleocytosis
    term:
      id: HP:0012229
      label: CSF pleocytosis
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and
      oligoclonal bands, are common in the early stages.
    explanation: The source names lymphocytic pleocytosis explicitly, which is what this bound term
      records.
  - reference: PMID:34817790
    reference_title: 'Anti-Tr/DNER Antibody-Associated Cerebellar Ataxia: a Systematic Review.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Cerebrospinal fluid was inflammatory in 64% of the patients.
    explanation: Quantifies CSF inflammation, in the anti-Tr/DNER population rather than the anti-Yo one.
      Recorded because it is the only pooled figure available for this finding, and its population is
      stated so it is not read as an anti-Yo rate.
- category: Laboratory
  name: CSF oligoclonal bands
  frequency: FREQUENT
  description: Oligoclonal IgG bands in CSF, indicating intrathecal immunoglobulin synthesis.
  phenotype_term:
    preferred_term: CSF oligoclonal immunoglobulin G bands
    term:
      id: HP:6000397
      label: CSF oligoclonal immunoglobulin G bands
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and
      oligoclonal bands, are common in the early stages.
    explanation: The source names oligoclonal bands explicitly. Intrathecal synthesis is what makes this
      distinct from the protein elevation above rather than a second measure of it.
- category: Neurologic
  name: Severe disability
  frequency: FREQUENT
  description: Most patients are left bedridden. This is the outcome that defines the clinical problem
    and the reason the therapeutic window matters.
  phenotype_term:
    preferred_term: Inability to walk
    term:
      id: HP:0002540
      label: Inability to walk
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The prognosis for anti-Yo PCD is almost uniformly poor, with most patients left bedridden.
    explanation: States the outcome for the anti-Yo form specifically. It should not be read as covering
      the surface-antigen arm, whose outlook may differ.
genetic:
- name: CDR2
  gene_term:
    preferred_term: CDR2
    term:
      id: hgnc:1799
      label: CDR2
  relationship_type: BIOMARKER
  association: The onconeural antigen itself. CDR2 is a cytoplasmic Purkinje cell protein ectopically
    expressed by the tumour and is the target of the anti-Yo response.
  notes: 'This is not a disease gene and is deliberately not typed as one. Nobody inherits paraneoplastic
    cerebellar degeneration, and no germline CDR2 variant causes it. CDR2 is curated here because the
    entry''s central mechanistic claim is about a specific antigen, and an entry that names an antigen in
    prose without giving it an identifier is not queryable by that antigen.

    relationship_type is BIOMARKER rather than CAUSATIVE, RISK_FACTOR or SUSCEPTIBILITY. None of those
    fits: the gene is not mutated, not a risk allele, and not a susceptibility locus. What CDR2 is, is
    the molecular identity of the thing the immune system attacks. BIOMARKER is the closest available
    value and it is still an imperfect fit - see the entry-level notes.

    The antigen''s intracellular location is the mechanical argument behind the CANONICAL hypothesis, so
    getting the identifier right matters. Note that the deep-research report and the round-1 review both
    proposed hgnc:1802 for CDR2; hgnc:1802 is CDSN (corneodesmosin), an unrelated skin gene. The correct
    identifier is hgnc:1799, confirmed against the ontology.'
  evidence:
  - reference: PMID:11771954
    reference_title: Induction of cytotoxic T lymphocytes specific for paraneoplastic cerebellar
      degeneration-associated antigen in vivo by DNA immunization.
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: MODEL_ORGANISM
    snippet: Paraneoplastic cerebellar degeneration associated with gynecological and breast
      malignancies (PCD) is known to develop autoantibodies and autoreactive cytotoxic T lymphocytes
      (CTLs) specific for a cytoplasmic protein of Purkinje cells PCD17/cdr2, in the blood of patients.
    explanation: 'Names the antigen and states its cytoplasmic location, the fact the canonical
      hypothesis rests on, and states that both antibodies and cytotoxic T cells against it are present
      in patients - which is why the effector question is a controversy here rather than a settled
      point.

      Regraded MODEL_ORGANISM in review. The cited paper is a BALB/c mouse DNA-immunisation study, and
      evidence_source classifies the publication rather than the quoted sentence, even where that
      sentence is itself reporting a human observation. Graded INDIRECT for the same reason. The claim
      it supports is not species-contingent and is independently supported by PMID:27606347 elsewhere
      in this entry.'
- name: CDR2L
  gene_term:
    preferred_term: CDR2L
    term:
      id: hgnc:29999
      label: CDR2L
  relationship_type: BIOMARKER
  association: A CDR2 paralogue proposed as the major anti-Yo antigen in place of, or alongside, CDR2.
  notes: 'Recorded with its identifier and deliberately without a supporting evidence item, because none
    of the references cached for this entry states the CDR2L claim in quotable form. The identifier is
    here so that a curator who later finds that literature has the anchor to attach it to, and so that
    the entry does not silently imply CDR2 is the only candidate antigen. This is an incomplete record
    rather than an unsupported claim - the distinction is that it asserts nothing beyond the gene''s
    existence and its relationship to CDR2.'
- name: HLA-DRB1
  gene_term:
    preferred_term: HLA-DRB1
    term:
      id: hgnc:4948
      label: HLA-DRB1
  relationship_type: SUSCEPTIBILITY
  association: The DRB1*13:01~DQA1*01:03~DQB1*06:03 class II haplotype is increased in anti-Yo disease
    arising from ovarian cancer.
  notes: 'The association is tumour-specific, which is the interesting part rather than an incidental
    qualifier: the same antibody arising from a different primary does not carry the same haplotype
    signal. That is consistent with the mechanism this entry curates, where the tumour is the site at
    which the antigen is presented and the anti-tumour response is shaped.

    Scale is the limit. Forty cases against ethnically matched controls is small for an HLA study, and
    the odds ratio of 5.4 is not accompanied by a confidence interval in the abstract. No replication
    cohort is cited. relationship_type is SUSCEPTIBILITY, which is the correct value for a common-variant
    risk haplotype and should not be read as a determinant.'
  evidence:
  - reference: PMID:29306402
    reference_title: Complex HLA association in paraneoplastic cerebellar ataxia with anti-Yo antibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Increased DRB1*13:01~DQA1*01:03~DQB1*06:03 was also found in PCD ovarian cases (OR=5.4,
      p=0.0016).
    explanation: The risk association, quoted with its effect size and the tumour type it is specific to.
  - reference: PMID:29306402
    reference_title: Complex HLA association in paraneoplastic cerebellar ataxia with anti-Yo antibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: These results suggest differential genetic susceptibility to anti-Yo per cancer and with a
      primary HLA Class II involvement.
    explanation: The authors' own conclusion, which states both the class II localisation and the
      per-cancer specificity. Their hedge ("suggest") is preserved.
- name: HLA-DPB1
  gene_term:
    preferred_term: HLA-DPB1
    term:
      id: hgnc:4940
      label: HLA-DPB1
  relationship_type: PROTECTIVE
  association: The DPA1*01:03~DPB1*04:01 haplotype is protective against anti-Yo disease overall.
  notes: 'Curated as a separate record from HLA-DRB1 rather than folded into a single HLA entry, because
    relationship_type is a per-record value and these two point in opposite directions. Reporting them
    together would have forced one of the two directions to be lost.

    The reported odds ratio of 0 is a zero cell, not a measured effect size - no case carried the
    haplotype. That is a real observation in a 40-case study and it is also exactly the situation in
    which an odds ratio is uninformative about magnitude. The second, non-zero estimate for the same
    haplotype (OR=0.35) is quoted alongside it for that reason.'
  evidence:
  - reference: PMID:29306402
    reference_title: Complex HLA association in paraneoplastic cerebellar ataxia with anti-Yo antibodies.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: We detected protective effects of DPA1*01:03~DPB1*04:01 (OR=0, p=0.0008),
      DRB1*04:01~DQA1*03:03(OR=0, p=0.0016) and DPA1*01:03~DPB1*04:01 (OR=0.35, p=0.0047) overall.
    explanation: The protective associations with their effect sizes. The sentence carries the same
      haplotype twice with different odds ratios, which is quoted rather than tidied because it is what
      makes the zero cells interpretable.
environmental:
- name: Immune checkpoint inhibitor therapy
  description: 'Anti-PD-1, anti-PD-L1 and anti-CTLA-4 therapy can trigger a cerebellar syndrome that is
    clinically and immunologically continuous with spontaneous paraneoplastic cerebellar degeneration.
    It is the one exposure in this disease with a mechanism that is not merely plausible but by design:
    checkpoint inhibitors work by removing the restraint on an antitumour T-cell response, and this
    entry''s trigger node is an antitumour response that has become cross-reactive with Purkinje cells.

    Curated as an exposure rather than as a subtype because the resulting syndrome is not clearly the
    same disease - see the discussion pcd_ici_triggered_same_disease.'
  exposure_term:
    preferred_term: exposure to an immune checkpoint inhibitor
    term:
      id: ECTO:9001737
      label: exposure to antineoplastic agent
  influences_mechanisms:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    environmental_effect: TRIGGERS
    causal_link_type: DIRECT
    description: Checkpoint blockade removes the inhibitory restraint on the antitumour T-cell response
      that this node describes, which is the same response that becomes cross-reactive with Purkinje
      cells. The exposure acts on the node by design rather than as a side effect of an unrelated
      mechanism.
    evidence:
    - reference: PMID:39153058
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: Symptoms developed at a median of 11 weeks after ICI onset.
      explanation: The temporal relationship between exposure and syndrome onset, across 35 pooled
        patients. A median of eleven weeks is what makes this a triggering relationship rather than a
        coincidental one in patients who have cancer anyway.
  evidence:
  - reference: PMID:39153058
    reference_title: 'Immune Checkpoint Inhibitor-Related Cerebellar Toxicity: Clinical Features and
      Comparison with Paraneoplastic Cerebellar Ataxia.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Neuronal antibodies were detected in 15/31 patients tested (48%).
    explanation: Roughly half of checkpoint-inhibitor cerebellar cases carry a neuronal antibody, which
      is what places the syndrome on the same immunological spectrum as the spontaneous disease rather
      than making it a separate toxicity.
  - reference: PMID:37151179
    reference_title: Immune checkpoint inhibitor-associated central nervous system autoimmunity.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: CNS manifestations included encephalitis (n = 16), meningoencephalitis (n = 8), cerebellar
      ataxia (n = 4), demyelinating syndrome (n = 2), and myelopathy (n = 1).
    explanation: Places the cerebellar syndrome in proportion within checkpoint-inhibitor CNS
      autoimmunity - it is a minority manifestation, four of thirty-one, which bears on how often a
      clinician should expect it.
  notes: 'The exposure term is broader than the exposure. ECTO has no term for checkpoint inhibitor
    exposure specifically, and no term for pembrolizumab or nivolumab exposure; exposure to antineoplastic
    agent is the most specific correct ancestor, with the actual drug class kept in preferred_term. Both
    searches are recorded here so the binding is not re-derived.

    This matches the existing binding in Autoimmune_Encephalitis, which uses the same CURIE for the same
    exposure. just environmental-term-audit reports a name conflict against ECTO:0000509 (exposure to
    drug); that is a normalisation artifact rather than a disagreement - the other CURIE is bound to
    generic drug exposures in unrelated entries, not to checkpoint inhibitors.

    environmental_effect is TRIGGERS rather than the PREDISPOSES used in Autoimmune_Encephalitis. The
    difference is deliberate: the cited series gives a median of eleven weeks from drug start to
    neurological onset, which is a triggering interval, where the encephalitis entry records a
    predisposing relationship without that temporal anchor.'
treatments:
- name: Corticosteroid Therapy
  therapeutic_modality: SMALL_MOLECULE
  description: 'Steroids, usually intravenous methylprednisolone pulses followed by oral maintenance, as
    first-line immunosuppression. Used extensively and with limited success, and the mechanism explains
    why: in the anti-Yo form the Purkinje cells are already dead by the time the syndrome is recognised,
    and no immunosuppression replaces a Purkinje cell.'
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: glucocorticoid
      term:
        id: CHEBI:24261
        label: glucocorticoid
  target_mechanisms:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    description: Broad suppression of the immune response upstream of the cerebellar injury, which is
      the only point at which it can act.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Immunotherapies, such as steroids, intravenous immune globulins, and plasma exchange, have
      been extensively used in managing this condition, with limited success.
    explanation: States both the practice and its poor result, in the source's own words. The same
      sentence supports the IVIG and plasma exchange records below; each is curated separately because
      they act by different routes and are separately bindable.
  - reference: PMID:41197574
    reference_title: 'Clinical characteristics, treatment, and outcomes in anti-metabotropic glutamate
      receptor-1 encephalitis: A systematic review.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Glucocorticoids constituted the most common immunotherapy (n = 29, 80.6 %).
    explanation: Quantifies steroid use, in the anti-mGluR1 population. Recorded here because the
      outcomes in that population are markedly better than in anti-Yo disease, which is the entry's
      central prognostic claim rather than an incidental difference.
- name: Intravenous Immunoglobulin
  therapeutic_modality: OTHER
  description: Pooled immunoglobulin, used first line alongside or after steroids. Its mechanism in this
    disease is not established, which is one reason the record is bound to the immunoglobulin therapy
    action rather than to a named agent.
  treatment_term:
    preferred_term: Intravenous Immunoglobulin Therapy
    term:
      id: NCIT:C121331
      label: Intravenous Immunoglobulin Therapy
  target_mechanisms:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    description: Modulates the systemic immune response upstream of the cerebellar injury.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Immunotherapies, such as steroids, intravenous immune globulins, and plasma exchange, have
      been extensively used in managing this condition, with limited success.
    explanation: Names intravenous immune globulin among the treatments used and states the outcome.
- name: Therapeutic Plasma Exchange
  therapeutic_modality: OTHER
  description: 'Removal of circulating antibody. Its rationale is entirely conditional on the
    antibody-effector hypothesis, which makes it the treatment whose success or failure carries the most
    mechanistic information in this entry: if antibodies were the effectors in anti-Yo disease, this
    should work, and it mostly does not.'
  treatment_term:
    preferred_term: Plasmapheresis
    term:
      id: NCIT:C15304
      label: Plasmapheresis
  target_mechanisms:
  - target: Antibody-Mediated Purkinje Cell Dysfunction
    description: Removes the circulating antibody, which is the effector in the surface-antigen arm and
      the point at which this treatment has a mechanistic rationale.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Immunotherapies, such as steroids, intravenous immune globulins, and plasma exchange, have
      been extensively used in managing this condition, with limited success.
    explanation: Names plasma exchange among the treatments used and states the outcome, which is the
      weak evidence against the antibody-effector reading discussed in this entry.
  - reference: PMID:32872070
    reference_title: 'Plasma exchange for treating anti-Yo-associated paraneoplastic cerebellar
      degeneration: Case report and literature review.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: After four courses of plasma exchange, the patient could walk independently, the Romberg
      test was negative, and anti-Yo antibodies were undetectable.
    explanation: A single case with an unusually good outcome, in an atypical patient - male, and with
      no tumour found. It is curated because it is a genuine counterexample to the uniformly poor
      prognosis, and its atypia is stated so it is not read as representative.
  notes: 'The plasma exchange case above is one patient. The authors themselves conclude that the
    approach needs more study, and the patient differed from the typical anti-Yo population in two ways
    that plausibly matter - male sex and the absence of a detected tumour. It is recorded as a
    counterexample, not as evidence of efficacy.'
- name: B-Cell Depletion
  therapeutic_modality: MONOCLONAL_ANTIBODY
  description: 'Anti-CD20 monoclonal antibodies - rituximab, and more recently ofatumumab - as
    second-line treatment when first-line immunotherapy fails. It is the most mechanistically targeted
    option available, and like plasma exchange its rationale depends on the antibody arm mattering.

    The evidence available for this record is two single cases and it is deliberately not presented as
    more than that. Both are reported as favourable outcomes, and favourable single cases are the ones
    that get published. Note also that the mGluR1 case is a surface-antigen disease with a much better
    baseline prognosis, so it should not be read as evidence for B-cell depletion in anti-Yo disease.'
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: ofatumumab
      term:
        id: NCIT:C66952
        label: Ofatumumab
    - preferred_term: rituximab
      term:
        id: NCIT:C1702
        label: Rituximab
  target_mechanisms:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    description: Depletes the B-cell compartment producing the onconeural antibody.
  evidence:
  - reference: PMID:39737186
    reference_title: 'Case report: paraneoplastic cerebellar degeneration associated with anti-Yo
      antibody successfully treated with ofatumumab.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: When a second course of methylprednisolone pulse therapy proved ineffective, treatment was
      switched to ofatumumab. After two doses, the patient achieved partial symptomatic relief.
    explanation: A single anti-Yo case, quoted with the outcome the authors actually report - partial
      relief - rather than the "successfully treated" of their title. The distinction matters because
      this entry's prognosis claims are about degree of recovery.
  - reference: PMID:40760473
    reference_title: 'Efficacy of early rituximab treatment in anti-mGluR1 encephalitis: a case report.'
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: As the initial treatment was insufficient, rituximab was administered, leading to
      significant improvement, including the ability to walk unaided.
    explanation: A single surface-antigen case. Graded INDIRECT because the inference to anti-Yo disease
      requires an extra step the case does not support - anti-mGluR1 disease is antibody-mediated against
      an accessible target and recovers far more often, so a good result there does not predict one here.
- name: Tumour Resection
  therapeutic_modality: SURGERY
  description: Surgical removal of the underlying malignancy, on the rationale that removing the antigen
    source should shut down the immune response driving the neurological disease. The rationale is sound
    and the results are inconsistent, which is itself informative - it suggests the autoimmune response
    becomes self-sustaining, or that the damage is complete before the tumour is found.
  treatment_term:
    preferred_term: Surgical Procedure
    term:
      id: NCIT:C15329
      label: Surgical Procedure
  target_mechanisms:
  - target: Ectopic Onconeural Antigen Expression by Tumour
    description: Removes the source of the antigen that initiated the cross-reactive response, acting at
      the trigger node.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Although some reports indicate benefit from antitumor therapies like surgery and
      chemotherapy, this has not been consistently observed.
    explanation: States the inconsistency directly. The claim this record makes is that benefit is
      reported and unreliable, which is what the source supports.
- name: Antitumour Chemotherapy
  therapeutic_modality: SMALL_MOLECULE
  description: Systemic treatment of the underlying malignancy, with the same antigen-removal rationale
    as resection and the same inconsistent neurological result. Curated separately from resection
    because the two are different modalities with different bindings, not because their evidence
    differs.
  treatment_term:
    preferred_term: Chemotherapy
    term:
      id: NCIT:C15632
      label: Chemotherapy
  target_mechanisms:
  - target: Ectopic Onconeural Antigen Expression by Tumour
    description: Reduces or removes the antigen-expressing tumour burden at the trigger node.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Although some reports indicate benefit from antitumor therapies like surgery and
      chemotherapy, this has not been consistently observed.
    explanation: The same sentence names chemotherapy alongside surgery, and states the same
      inconsistency.
  - reference: PMID:34817790
    reference_title: 'Anti-Tr/DNER Antibody-Associated Cerebellar Ataxia: a Systematic Review.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Oncological response was complete in 88%, but neurological prognosis was poor with only 41%
      of the patients presenting significant neurological improvement at the last follow up.
    explanation: 'The cleanest available dissociation between the two outcomes: in 85 anti-Tr/DNER
      patients the tumour is controlled in most and the ataxia is not. This is the strongest evidence in
      the entry that removing the antigen source does not reverse the neurological disease, and it is why
      the target_mechanisms edge points at the trigger node rather than at the cerebellar injury.'
  notes: Treating the tumour is not optional - it is the treatment for the cancer, whatever it does for
    the neurology. This record is about its neurological effect only, and should not be read as
    questioning oncological indication.
- name: Symptomatic and Supportive Care
  therapeutic_modality: BEHAVIORAL
  description: 'Rehabilitation, mobility aids and management of the established deficit. Curated because
    it is what most patients actually receive over the long term, and because the entry would otherwise
    imply that a disease in which most survivors cannot walk unassisted has nothing offered to it after
    immunotherapy fails.

    Recorded without an efficacy claim. No cited source in this entry reports a controlled outcome for
    supportive care in this disease, and none is asserted.'
  treatment_term:
    preferred_term: Supportive Care
    term:
      id: NCIT:C15747
      label: Supportive Care
  evidence:
  - reference: PMID:36334195
    reference_title: 'A Systematic Review on Anti-Yo/PCA-1 Antibody: Beyond Cerebellar Ataxia in
      Middle-Aged Women with Gynecologic Cancer.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: However, despite treatment, 84% of survivors were unable to walk unassisted on
      follow-up.
    explanation: Establishes the residual disability that this record exists to address. It is also the
      number that makes supportive care the substantive part of long-term management rather than an
      afterthought.
  notes: 'The 84% figure is the single most consequential number in this entry and it is pooled from
    case reports and small series, which are enriched for severe and unusual presentations. It should be
    read as the outcome among published anti-Yo patients, not as a population estimate.'
diagnosis:
- name: Onconeural antibody detection with risk stratification
  description: Detection of a characteristic antibody in serum or CSF. Current criteria classify
    antibodies by how strongly they predict an underlying cancer rather than treating them as a single
    category, which changes how a positive result should be acted on.
  evidence:
  - reference: PMID:34006622
    reference_title: Updated Diagnostic Criteria for Paraneoplastic Neurologic Syndromes.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The term "onconeural antibody" was replaced by "high risk" (>70% associated with cancer)
      and "intermediate risk" (30%-70% associated with cancer) antibodies.
    explanation: Gives the quantitative threshold defining the high-risk antibody category that anti-Yo
      falls into, and the intermediate-risk band alongside it, so the reader can see that the
      classification is a graded scale rather than a binary.
  - reference: PMID:34006622
    reference_title: Updated Diagnostic Criteria for Paraneoplastic Neurologic Syndromes.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: we propose updated diagnostic criteria for PNS.
    explanation: Establishes that this reference is the current criteria statement rather than a review
      of practice.
- name: Search for an occult malignancy
  description: Investigation for a gynaecological or breast primary in the typical anti-Yo presentation.
    The neurological syndrome usually precedes the cancer diagnosis, so a negative initial search does
    not end the question.
  evidence:
  - reference: PMID:27606347
    reference_title: Paraneoplastic cerebellar degeneration with anti-Yo antibodies - a review.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: The vast majority of cases have been reported in women with pelvic or breast tumors.
    explanation: States where to look, and the demographic that makes the search targeted.
- name: Antibody-guided imaging search for the primary tumour
  description: 'Which antibody is found determines where and how hard to look. Current guidance is that
    the antibody, more than the clinical syndrome, sets the risk and type of the underlying malignancy,
    and prescribes a modality sequence per body region rather than a single scan. It also sets a
    duration: a negative first search is repeated at 3 to 6 months and then six-monthly for up to four
    years, because the neurological syndrome almost always precedes the cancer.'
  evidence:
  - reference: PMID:20880069
    reference_title: 'Screening for tumours in paraneoplastic syndromes: report of an EFNS task force.'
    supports: SUPPORT
    evidence_source: OTHER
    snippet: the nature of antibody, and to a lesser extent the clinical syndrome, determines the risk
      and type of an underlying malignancy.
    explanation: The principle behind antibody-guided rather than syndrome-guided screening. Graded
      OTHER because this is a task-force consensus recommendation, and the report states that no
      evidence-based grade A-C recommendations were possible.
  - reference: PMID:20880069
    reference_title: 'Screening for tumours in paraneoplastic syndromes: report of an EFNS task force.'
    supports: SUPPORT
    evidence_source: OTHER
    snippet: For screening of the thoracic region, a CT-thorax is recommended, which if negative is
      followed by fluorodeoxyglucose-positron emission tomography (FDG-PET).
    explanation: Places FDG-PET as the second-line thoracic modality rather than the first, which is the
      practical content of the recommendation.
  - reference: PMID:20880069
    reference_title: 'Screening for tumours in paraneoplastic syndromes: report of an EFNS task force.'
    supports: SUPPORT
    evidence_source: OTHER
    snippet: If primary screening is negative, repeat screening after 3-6 months and screen every 6
      months up till 4 years.
    explanation: The surveillance interval and its endpoint. This is the operational answer to the
      problem stated in the record above - that a negative initial search does not end the question.
- name: FDG-PET/CT when no well-characterised antibody is found
  description: Whole-body FDG-PET/CT in clinically suspected paraneoplastic disease, including patients
    whose antibody testing is negative. This is the case where antibody-guided screening has nothing to
    guide it, and it is also where the yield data are weakest.
  evidence:
  - reference: PMID:22157026
    reference_title: Utility of F-18 FDG PET/CT in screening for paraneoplastic neurological syndromes.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Of the 27 patients, 6 (22%) had an abnormal FDG uptake. Of these 6 patients, 5 (19%) were
      histologically confirmed as having a malignancy.
    explanation: The actual yield, from a 27-patient retrospective single-centre series. A five-cancer
      denominator is small, and the record is written so the reader sees that rather than inheriting the
      recommendation without its evidence base.
  - reference: PMID:22157026
    reference_title: Utility of F-18 FDG PET/CT in screening for paraneoplastic neurological syndromes.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: PET/CT may be a useful screening tool for patients with clinically suspected PNS who do not
      exhibit well-characterized paraneoplastic antibodies.
    explanation: The authors' recommendation with their own hedge ("may") preserved, and scoped to the
      antibody-negative group where the antibody-guided approach does not apply.
animal_models:
- name: Anti-Yo IgG passive transfer to rodent brain and ventricles
  species: Mouse
  genotype: Wild-type, various strains; also rat recipients
  description: 'Patient anti-Yo IgG injected into mouse brain or, with complement or activated
    mononuclear cells, into rat ventricles. The antibody is taken up by Purkinje cells and stays there,
    and the cells do not die.

    This is a negative result and it is curated as the load-bearing evidence for the entry''s canonical
    hypothesis rather than as a footnote. The uptake half matters as much as the failure half: it rules
    out the trivial explanation that the antibody simply never reached its target.'
  publication: PMID:7707074
  modeled_mechanisms:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    relationship: FAILS_TO_RECAPITULATE
    fidelity: MODERATE
    model_scale: TISSUE
    description: Anti-Yo IgG delivered directly to the target tissue, alone or with complement or
      activated mononuclear cells, does not produce Purkinje cell loss.
    limitations: 'A negative transfer result is weaker than it looks, and the specific weaknesses are
      worth naming. Rodent Purkinje cells may not present the human epitope in the way human cells do;
      a single injection over 36 hours is a fraction of the exposure a patient has over months; and
      intracerebral or intraventricular delivery is not the route the antibody takes in disease. What
      the experiment establishes is that antibody plus effector is not sufficient on the timescale
      tested, not that antibody is irrelevant.'
    divergences:
    - divergence_type: TEMPORAL_SCOPE
      materiality: QUALIFYING
      description: The observation window is 36 hours after a single injection. Human disease develops
        over weeks to months of continuous antibody exposure, so the model tests acute toxicity rather
        than the chronic process.
    - divergence_type: SPECIES_MISMATCH
      materiality: QUALIFYING
      description: Rodent Purkinje cells are the recipients of a human autoantibody. The antibody does
        bind them, which limits how far this divergence can be pressed, but epitope presentation and
        downstream effector engagement need not match the human cell.
    readouts:
    - name: Purkinje cell survival after intracerebral anti-Yo IgG
      target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
      direction: UNCHANGED
      interpretation: The measurement that carries the negative result. UNCHANGED here is the finding,
        not a missing datum.
      evidence:
      - reference: PMID:7707074
        reference_title: Passive transfer and active immunization with the recombinant leucine-zipper
          (Yo) protein as an attempt to establish an animal model of paraneoplastic cerebellar
          degeneration.
        supports: SUPPORT
        evidence_source: MODEL_ORGANISM
        snippet: On a single injection of PCD IgG to mouse brain, IgG was taken into Purkinje cells and
          remained there for more than 36 h without Purkinje cell loss.
        explanation: Records both halves - the antibody reached and entered the target cell, and the
          cell survived. The uptake is what makes the negative result interpretable.
    evidence:
    - reference: PMID:7707074
      reference_title: Passive transfer and active immunization with the recombinant leucine-zipper (Yo)
        protein as an attempt to establish an animal model of paraneoplastic cerebellar degeneration.
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      snippet: The anti-Yo antibody, with or without complement or activated mononuclear cells,
        therefore could not be the sole cause of Purkinje cell loss.
      explanation: The authors' conclusion, graded REFUTE against the antibody-effector reading of this
        node. Note the claim they make is bounded - "sole cause" - and the entry does not stretch it
        into a claim that the antibody plays no part.
  - target: Antibody-Mediated Purkinje Cell Dysfunction
    relationship: FAILS_TO_RECAPITULATE
    fidelity: LOW
    model_scale: TISSUE
    description: Cited as a bounding result for the antibody-effector mechanism generally. Fidelity is
      LOW because this node is about a surface antigen and the transferred antibody targets an
      intracellular one.
    limitations: The antibody transferred here is anti-Yo, against the intracellular CDR2. This node
      describes an antibody against a cell-surface channel, which is a different mechanism with a
      different physical basis. A failure to transfer intracellular-antigen disease says little about
      whether surface-antigen disease is antibody-mediated, and the link is recorded only so that a
      reader does not mistake the anti-Yo negative for a general refutation of antibody effectors.
    divergences:
    - divergence_type: PROXY_QUANTITY
      materiality: INVALIDATING
      description: The quantity transferred is an antibody against an intracellular antigen; the
        quantity this node is about is an antibody against a surface antigen. The two differ in whether
        the target is physically accessible to the antibody in a living cell, which is the whole
        mechanistic distinction the entry draws.
    evidence:
    - reference: PMID:7707074
      reference_title: Passive transfer and active immunization with the recombinant leucine-zipper (Yo)
        protein as an attempt to establish an animal model of paraneoplastic cerebellar degeneration.
      supports: NO_EVIDENCE
      evidence_source: MODEL_ORGANISM
      snippet: The anti-Yo antibody, with or without complement or activated mononuclear cells,
        therefore could not be the sole cause of Purkinje cell loss.
      explanation: Graded NO_EVIDENCE against this node deliberately. The experiment is real and its
        conclusion is sound, and it does not bear on a surface-antigen mechanism at all. Recording that
        explicitly is the point of the link.
- name: Active immunization with recombinant Yo protein across four MHC backgrounds
  species: Mouse
  genotype: 'Wild-type BALB/c (H-2d), C3H (H-2k), C57BL/6 (H-2b), SJL/J (H-2s); also C57BL/6 nu/nu'
  background: Four inbred strains bearing different MHC haplotypes
  description: Mice immunized with recombinant Yo protein raise high anti-Yo titres and reactive spleen
    cells across four MHC backgrounds, and none develops ataxia or loses Purkinje cells. Using four
    haplotypes is what makes this more than one failed experiment - it rules out the explanation that a
    single strain simply lacked the right restriction element.
  publication: PMID:7788981
  modeled_mechanisms:
  - target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
    relationship: FAILS_TO_RECAPITULATE
    fidelity: MODERATE
    model_scale: ORGANISM
    description: A systemic anti-Yo immune response, raised in vivo and MHC-diverse, is not sufficient
      to produce the disease.
    limitations: The immunization raises a peripheral response. Nothing in the protocol breaks the
      blood-brain barrier, provides the tumour that in patients precedes the neurological disease, or
      supplies whatever second event lets a peripheral response reach the cerebellum. The authors draw
      exactly this conclusion, so the model's failure is informative about sufficiency rather than about
      the mechanism being wrong.
    divergences:
    - divergence_type: CAUSE_UNREPRESENTED
      materiality: QUALIFYING
      description: There is no tumour in this model. In patients the antigen is presented by a
        malignancy, which is also the setting in which the response is shaped and, plausibly, the setting
        that licenses access to the CNS. Immunization with recombinant protein represents the antigen
        without representing that context.
    readouts:
    - name: Cerebellar ataxia and Purkinje cell loss after Yo immunization
      target: Antitumour Immune Response Cross-Reactive with Purkinje Cells
      direction: UNCHANGED
      interpretation: The negative outcome, measured across four MHC haplotypes rather than one.
      evidence:
      - reference: PMID:7788981
        reference_title: 'Trial to establish an animal model of paraneoplastic cerebellar degeneration
          with anti-Yo antibody. 1. Mouse strains bearing different MHC molecules produce antibodies on
          immunization with recombinant Yo protein, but do not cause Purkinje cell loss.'
        supports: SUPPORT
        evidence_source: MODEL_ORGANISM
        snippet: All the strains produced high anti-Yo antibody titer but none developed cerebellar
          ataxia or showed Purkinje cell loss.
        explanation: The measurement, with the antibody response confirmed so the failure cannot be
          attributed to a failed immunization.
    evidence:
    - reference: PMID:7788981
      reference_title: 'Trial to establish an animal model of paraneoplastic cerebellar degeneration
        with anti-Yo antibody. 1. Mouse strains bearing different MHC molecules produce antibodies on
        immunization with recombinant Yo protein, but do not cause Purkinje cell loss.'
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      snippet: Results suggest that antibody production in peripheral blood alone is not sufficient for
        the development of PCD and that MHC class II molecules function in the activation of T cells to
        help B cells to help B cells produce antibodies.
      explanation: The authors' conclusion, graded REFUTE against a peripheral-antibody account of this
        node. Quoted as written, including the duplicated clause in the source.
    - reference: PMID:7788964
      reference_title: 'Trial to establish an animal model of paraneoplastic cerebellar degeneration
        with anti-Yo antibody. 2. Passive transfer of murine mononuclear cells activated with
        recombinant Yo protein to paraneoplastic cerebellar degeneration lymphocytes in severe combined
        immunodeficiency mice.'
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      snippet: Spleen cells from these Yo protein immunized mice were injected intravenously or
        intracerebrally into naive mice that subsequently showed no neurological symptoms or loss of
        Purkinje cells.
      explanation: The companion paper to the immunization study above. It transfers the cellular arm
        rather than the antibody, including directly into brain, and is the fourth of the four failed
        reconstructions this entry records. Graded REFUTE for the same reason - it bears against the
        sufficiency of the transferred response.
- name: DNA immunization raising both anti-Yo antibodies and CDR2-specific cytotoxic T cells
  species: Mouse
  genotype: Wild-type BALB/c immunized with naked pcd17 (CDR2) cDNA
  description: 'The most demanding of the four transfer and immunization experiments, and the one this
    entry treats as most informative. DNA immunization generated both arms of the human immune response
    at once - anti-Purkinje cell autoantibodies and CDR2-specific cytotoxic T lymphocytes able to lyse
    peptide-pulsed targets - and still produced no clinical or pathological cerebellar degeneration.

    It is curated because it does not favour either hypothesis in this entry. The other three
    experiments fail to transfer antibody-mediated disease and so support the T-cell reading by
    elimination. This one raises the T cells too, and still fails.'
  publication: PMID:11771954
  modeled_mechanisms:
  - target: Cytotoxic T Cell-Mediated Purkinje Cell Destruction
    relationship: FAILS_TO_RECAPITULATE
    fidelity: MODERATE
    model_scale: ORGANISM
    description: A CDR2-specific cytotoxic T-cell response demonstrably capable of lysing peptide-pulsed
      targets does not, on its own, produce cerebellar degeneration in vivo.
    limitations: 'The T cells were shown to lyse syngeneic myeloma cells pulsed with peptide - an
      artificial target presenting the epitope at high density. That is not evidence they could
      recognise a Purkinje cell, which presents endogenous antigen and which upregulates MHC class I in
      human disease but need not do so in an unmanipulated mouse. So the experiment shows that raising
      CDR2-specific CTLs is not sufficient; it does not show that the CTLs generated were the same
      effector population as in patients.'
    divergences:
    - divergence_type: PROXY_QUANTITY
      materiality: QUALIFYING
      description: Cytotoxic capability is measured against peptide-pulsed myeloma cells, not against
        Purkinje cells. The measured quantity is epitope-specific lysis in vitro; the quantity of
        interest is killing of an endogenously presenting neuron in vivo.
    - divergence_type: CAUSE_UNREPRESENTED
      materiality: QUALIFYING
      description: No tumour and no MHC class I upregulation on cerebellar neurons - the two features
        that in human neuropathology make the Purkinje cell a visible target for a CD8 T cell.
    readouts:
    - name: Cerebellar degeneration after CDR2 DNA immunization
      target: Cytotoxic T Cell-Mediated Purkinje Cell Destruction
      direction: UNCHANGED
      interpretation: No degeneration, despite both effector arms being present. This is the strongest
        single argument in the entry that the mechanism is not yet fully accounted for by either
        hypothesis.
      evidence:
      - reference: PMID:11771954
        reference_title: Induction of cytotoxic T lymphocytes specific for paraneoplastic cerebellar
          degeneration-associated antigen in vivo by DNA immunization.
        supports: REFUTE
        evidence_source: MODEL_ORGANISM
        snippet: In spite of the generation of anti-Purkinje cell antibodies and PCD17-specific CTLs in
          vivo, neither clinical nor pathological changes consistent with significant cerebellar
          degeneration have been detected.
        explanation: The negative result stated with both arms confirmed present. Graded REFUTE because
          it cuts against the sufficiency of the cytotoxic mechanism this node asserts, which is a
          different and weaker claim than refuting the mechanism itself.
    evidence:
    - reference: PMID:11771954
      reference_title: Induction of cytotoxic T lymphocytes specific for paraneoplastic cerebellar
        degeneration-associated antigen in vivo by DNA immunization.
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: We show that DNA immunization with naked pcd17 cDNA could induce autoantibodies against
        the cytoplasmic protein of Purkinje cells and CTLs could lyse syngenic myeloma cells pulsed with
        H-2K-restricted PCD17 peptide.
      explanation: Establishes that the model genuinely produced both effector arms, which is what makes
        the accompanying negative outcome worth citing at all. Graded SUPPORT because this sentence
        supports treating the model as informative for this node, separately from what the model then
        showed.
  notes: 'Read together with the two transfer experiments above, these four studies say something the
    entry''s two hypotheses do not: no experimental reconstruction of either arm of the human immune
    response has produced the disease in an animal. That is recorded as a knowledge gap rather than
    resolved, and it is the reason the canonical hypothesis is argued from human neuropathology and from
    the antigen''s compartment rather than from a model.'
discussions:
- discussion_id: pcd_ici_triggered_same_disease
  kind: INTERPRETATION
  prompt: Is checkpoint-inhibitor-triggered cerebellar ataxia the same disease as spontaneous
    paraneoplastic cerebellar degeneration, or a distinct drug toxicity that resembles it?
  attaches_to:
  - environmental#Immune checkpoint inhibitor therapy
  - pathophysiology#Antitumour Immune Response Cross-Reactive with Purkinje Cells
  rationale: 'The case for one disease is mechanistic and strong. Checkpoint inhibitors act by removing
    restraint from an antitumour T-cell response, which is exactly the response this entry''s trigger
    node describes; about half of affected patients carry a neuronal antibody; and the syndrome is
    cerebellar. If the spontaneous disease is a failure of tolerance in an antitumour response, the drug
    version is that failure induced deliberately.

    The case against is empirical and also strong. The one study that compares the two directly finds
    them significantly different on four axes at once - sex, tumour type, whether the ataxia is isolated,
    and outcome. Male predominance and lung cancer are the inverse of the anti-Yo picture, which is
    female and gynaecological. A better outcome is the difference that matters clinically, and it is not
    a small one.

    The entry does not resolve this. It curates the exposure and its mechanism link, and does not create
    a subtype, because a subtype would assert sameness that the comparative data do not support. Nor is
    it excluded, because excluding it would assert difference on the strength of one retrospective
    comparison of 35 against 15 patients.

    Worth stating what would settle it: whether the antibody-defined subgroup of checkpoint-inhibitor
    cases behaves like its spontaneous counterpart. The comparison available pools antibody-positive and
    antibody-negative cases together, so the better outcome could belong entirely to the antibody-negative
    half and tell us nothing about whether anti-Yo disease triggered by a drug differs from anti-Yo
    disease arising on its own.'
  evidence:
  - reference: PMID:39153058
    reference_title: 'Immune Checkpoint Inhibitor-Related Cerebellar Toxicity: Clinical Features and
      Comparison with Paraneoplastic Cerebellar Ataxia.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: When compared with a series of PCA (n = 15), the cerebellar irAE group was significantly
      more associated with male sex, lung cancer (rather than gynecological/breast cancers), isolated
      ataxia, and a better outcome.
    explanation: The direct comparison, with its denominators. Four simultaneous differences in a
      35-versus-15 retrospective comparison is the whole case against treating the two as one disease,
      and the sample sizes are why it is not decisive.
  - reference: PMID:39153058
    reference_title: 'Immune Checkpoint Inhibitor-Related Cerebellar Toxicity: Clinical Features and
      Comparison with Paraneoplastic Cerebellar Ataxia.'
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Clinical presentation-antibody-tumor triad in the ICI group only partially reflects the
      associations described in paraneoplastic disorders.
    explanation: The authors' own summary of the relationship - partial correspondence rather than
      identity or independence, which is the position this discussion records.
- discussion_id: pcd_antibody_versus_t_cell_effector
  kind: CONTROVERSY
  prompt: In anti-Yo paraneoplastic cerebellar degeneration, is the autoantibody an effector of Purkinje
    cell death or a marker of a T cell-mediated attack?
  attaches_to:
  - pathophysiology#Antitumour Immune Response Cross-Reactive with Purkinje Cells
  - pathophysiology#Cytotoxic T Cell-Mediated Purkinje Cell Destruction
  - mechanistic_hypotheses#pcd_t_cell_mediated_intracellular
  rationale: 'The review literature states plainly that the relative importance of autoantibodies and
    cytotoxic T lymphocytes in the neuronal loss is debated. This entry curates the T-cell reading as
    canonical, and the reasoning should be visible rather than assumed.

    The argument for it is not primarily empirical, it is mechanical: CDR2 is intracellular, and an
    antibody in the extracellular space cannot bind an intracellular antigen in a living cell. The
    neuropathology then supplies exactly what a cytotoxic attack should look like - MHC class I
    upregulated on neurons, CD8 granzyme-B T cells present, microglial nodules.

    What keeps it a controversy rather than a settled question is that cytoplasmic IgG staining is seen
    in some neurons, and that nobody has shown the T-cell response is sufficient. The practical stake is
    therapeutic: if antibodies are effectors, plasma exchange and B-cell depletion should work, and the
    published experience is that they mostly do not - which is itself weak evidence for the T-cell
    reading.'
  proposed_experiments:
  - experiment_id: pcd_effector_depletion_comparison
    name: Comparative depletion of the antibody and T-cell arms in anti-Yo PCD
    description: Compare outcomes between antibody-directed treatment (plasma exchange, B-cell depletion)
      and T cell-directed treatment in anti-Yo patients treated at a comparable point in the clinical
      course, since the existing experience confounds treatment type with how late it was given.
    readouts:
    - name: Cerebellar function after arm-specific immunotherapy
      target: pathophysiology#Cytotoxic T Cell-Mediated Purkinje Cell Destruction
      direction: UNCHANGED
      interpretation: If antibody removal does not change the course while T cell-directed treatment
        does, the canonical hypothesis is supported over the antibody-effector one.
- discussion_id: pcd_antigen_compartment_determines_window
  kind: INTERPRETATION
  prompt: Should paraneoplastic cerebellar degeneration be treated as one disease, given that the antigen
    compartment appears to determine both the mechanism and whether treatment can work?
  attaches_to:
  - mechanistic_hypotheses#pcd_antibody_mediated_surface
  - pathophysiology#Antibody-Mediated Purkinje Cell Dysfunction
  rationale: 'The neuropathological comparison in this entry finds the intracellular-antigen and
    surface-antigen forms differ in nearly every measured feature: distribution of cell loss, MHC class I
    upregulation, complement deposition, antigen downregulation, and evidence of endoplasmic reticulum
    stress. That is a large difference to hold inside one disease name.

    The reason it is curated as one entry with two hypothesis groups rather than split is that the
    clinical syndrome, the tumour association and the diagnostic pathway are shared, and the antibody is
    often not known when the patient presents. A clinician meets one syndrome and finds out later which
    mechanism it was.

    The consequence worth flagging is prognostic. The uniformly poor outlook attached to this disease
    comes from anti-Yo series. It should not be assumed to apply to the surface-antigen forms, where the
    pathology suggests dysfunction preceding death and therefore a window in which early treatment might
    matter - a point the source paper makes explicitly.

    This is also the boundary question against the existing Paraneoplastic_Neurological_Syndromes entry,
    which splits by antigen compartment rather than by syndrome. The two entries are complementary and
    should stay consistent: that one covers the compartment distinction across syndromes, and this one
    covers what it means within the cerebellar syndrome.'
- discussion_id: pcd_surface_antigen_evidence_thin
  kind: KNOWLEDGE_GAP
  prompt: How much confidence should the neuropathological contrast between the intracellular-antigen
    and surface-antigen forms carry, given that it rests on a single anti-P/Q-VGCC autopsy case?
  attaches_to:
  - mechanistic_hypotheses#pcd_antibody_mediated_surface
  - pathophysiology#Antibody-Mediated Purkinje Cell Dysfunction
  rationale: 'The scope of this gap is narrower than it first appears, and getting the scope right is
    the point of recording it.

    What does not rest on one case: the compartment rule itself. That an antibody cannot reach an
    intracellular antigen in a living cell is mechanical, and the surface-antigen side of it is
    supported independently by anti-Tr/DNER, whose target is a demonstrated extracellular domain, and by
    anti-mGluR1, where 42 pooled cases respond to immunotherapy as an antibody-mediated mechanism
    predicts. The entry''s two-hypothesis architecture is not fragile.

    What does rest on one case: the specific neuropathological contrast. The paper that establishes it
    is explicit about its own denominator - one anti-P/Q-VGCC case against two anti-Yo cases and
    controls - and everything the entry says about topography of cell loss, MHC class I upregulation,
    complement deposition, antigen downregulation and endoplasmic reticulum stress in the surface-antigen
    arm inherits that limitation. So does the prognostic implication drawn from it.

    This is recorded rather than hedged into the node descriptions because it is a property of the
    evidence base, not of the claim: no amount of careful wording in the pathophysiology nodes tells a
    reader which parts rest on a single case, and the alternative - omitting the arm - would lose a real
    and well-characterised finding.

    A separate and larger gap sits underneath all of this. Four independent attempts to reconstruct the
    disease in animals - passive transfer of patient IgG, active immunization across four MHC
    haplotypes, transfer of immunized spleen cells, and DNA immunization raising both antibodies and
    CDR2-specific cytotoxic T cells - have all failed to produce cerebellar degeneration. Neither
    hypothesis in this entry has been shown sufficient in vivo. The animal_models section records these
    individually; the point of noting it here is that the choice between the two hypotheses is being made
    on human neuropathology and antigen compartment, not on any positive experimental reconstruction.'
  proposed_experiments:
  - experiment_id: pcd_surface_antigen_autopsy_series
    name: Multi-case neuropathological series of surface-antigen paraneoplastic cerebellar degeneration
    description: Assemble additional anti-P/Q-VGCC and other surface-antigen PCD autopsy cases and repeat
      the MHC class I, complement, CD8 and antigen-downregulation assessments, to establish whether the
      single-case findings generalise.
    readouts:
    - name: MHC class I upregulation in surface-antigen PCD cerebellum
      target: pathophysiology#Antibody-Mediated Purkinje Cell Dysfunction
      direction: UNCHANGED
      interpretation: Consistent absence of MHC class I upregulation across further cases would confirm
        that the surface-antigen arm is not a cytotoxic T-cell mechanism.
notes: 'Boundary with Paraneoplastic_Neurological_Syndromes. That entry (MONDO:0018215) exists and
  subdivides by antigen compartment - Intracellular Antigen PNS and Surface Antigen PNS - across all
  paraneoplastic neurological syndromes. This entry is the cerebellar syndrome, and it adds what the
  compartment-level entry does not carry: the specific onconeural antigen (CDR2), the Purkinje cell as
  the vulnerable population, the topography of cell loss, and the syndrome-specific prognosis. The two
  are complementary rather than redundant, and the compartment distinction is deliberately expressed
  here in the same terms so they stay consistent.

  A curator who disagrees has a reasonable case that this should instead be a has_subtypes entry on
  Paraneoplastic_Neurological_Syndromes. The argument against is that MONDO carries a distinct term, that
  the Purkinje-specific mechanism is not derivable from the compartment classification, and that PCD is
  the best-characterised paraneoplastic syndrome in its own right. This is flagged rather than assumed.

  kb/groupings/Cerebellar_Ataxias.yaml mentions paraneoplastic cerebellar degeneration and may want this
  entry as a member. That is a grouping edit and is deliberately not made here, since it changes a
  different artifact than the one this PR curates.

  On the antigen compartments. The two mechanistic_hypotheses are not rival explanations of one
  mechanism. They describe different antigen classes, and both are probably true of their respective
  patients. They are modelled as hypothesis groups because the causal edges genuinely diverge and each
  edge belongs to one of them, which is what hypothesis_groups is for.

  On ontology bindings. One phenotype term is broader than the finding, with the specific concept kept
  in preferred_term: HPO has no term for a pancerebellar syndrome as distinct from progressive cerebellar
  ataxia.

  An earlier version of this entry bound the CSF findings to a single broad term and justified it here on
  the grounds that HPO has no term for the triad together. That was true but misleading, and it produced
  the wrong curation: HPO has perfectly good terms for each component - HP:0002922, HP:0012229 and
  HP:6000397 - and the right answer was three phenotypes, not one broad binding. The three are now
  curated separately. The general lesson is worth keeping: the absence of a term for a bundle is an
  argument for unbundling the claim, not for binding the bundle to whichever component has a term.

  On gene identifiers. CDR2 is hgnc:1799. Both the deep-research report and the first round of review
  proposed hgnc:1802, which is CDSN, an unrelated skin gene. Recorded because the two independent
  sources agreeing on a wrong identifier is exactly the situation in which it would have been accepted
  without checking.'

📚

References & Deep Research

Deep Research

1

Deep research results are used as seeds for research; they do not undergo the same validation as the main records and may contain errors. How we use deep research.

Evaluations and curation notes (3)

Record notes

Boundary with Paraneoplastic_Neurological_Syndromes. That entry (MONDO:0018215) exists and subdivides by antigen compartment - Intracellular Antigen PNS and Surface Antigen PNS - across all paraneoplastic neurological syndromes. This entry is the cerebellar syndrome, and it adds what the compartment-level entry does not carry: the specific onconeural antigen (CDR2), the Purkinje cell as the vulnerable population, the topography of cell loss, and the syndrome-specific prognosis. The two are complementary rather than redundant, and the compartment distinction is deliberately expressed here in the same terms so they stay consistent. A curator who disagrees has a reasonable case that this should instead be a has_subtypes entry on Paraneoplastic_Neurological_Syndromes. The argument against is that MONDO carries a distinct term, that the Purkinje-specific mechanism is not derivable from the compartment classification, and that PCD is the best-characterised paraneoplastic syndrome in its own right. This is flagged rather than assumed. kb/groupings/Cerebellar_Ataxias.yaml mentions paraneoplastic cerebellar degeneration and may want this entry as a member. That is a grouping edit and is deliberately not made here, since it changes a different artifact than the one this PR curates. On the antigen compartments. The two mechanistic_hypotheses are not rival explanations of one mechanism. They describe different antigen classes, and both are probably true of their respective patients. They are modelled as hypothesis groups because the causal edges genuinely diverge and each edge belongs to one of them, which is what hypothesis_groups is for. On ontology bindings. One phenotype term is broader than the finding, with the specific concept kept in preferred_term: HPO has no term for a pancerebellar syndrome as distinct from progressive cerebellar ataxia. An earlier version of this entry bound the CSF findings to a single broad term and justified it here on the grounds that HPO has no term for the triad together. That was true but misleading, and it produced the wrong curation: HPO has perfectly good terms for each component - HP:0002922, HP:0012229 and HP:6000397 - and the right answer was three phenotypes, not one broad binding. The three are now curated separately. The general lesson is worth keeping: the absence of a term for a bundle is an argument for unbundling the claim, not for binding the bundle to whichever component has a term. On gene identifiers. CDR2 is hgnc:1799. Both the deep-research report and the first round of review proposed hgnc:1802, which is CDSN, an unrelated skin gene. Recorded because the two independent sources agreeing on a wrong identifier is exactly the situation in which it would have been accepted without checking.

Review round 1: address all eight findings on PR #11225 · 2026-09-06T10:32:54Z · View source

Review round 1 on PR #11225. All eight findings addressed in one push, plus the three suggestions. Evidence base goes from 21 to 62 verified snippets and from 4 to 19 distinct references. 1. Phenotype under-coverage. Six cerebellar phenotypes added and bound: gait ataxia (HP:0002066), truncal ataxia (HP:0002078), limb ataxia (HP:0002070), dysarthria (HP:0001260), nystagmus (HP:0000639) and prodromal vertigo (HP:0002321). Each is wired into the pathograph from Irreversible Purkinje Cell Loss, except the prodrome, which is deliberately drawn from the ongoing destruction node instead: the prodrome is what partial cell loss looks like, and by the time loss is complete the patient is no longer prodromal. It is the only feature in the entry that reports on a still-treatable state. 2. CSF binding rationale. The reviewer was right that the justification was wrong. The single broad HP:0002922 binding is split into three bound phenotypes (HP:0002922, HP:0012229 CSF pleocytosis, HP:6000397 CSF oligoclonal immunoglobulin G bands). The entry-level notes now record what the original reasoning got wrong: the absence of a term for a bundle is an argument for unbundling the claim, not for binding the bundle to whichever component happens to have a term. 3. Genetic section added. CDR2, CDR2L and two HLA class II records. IMPORTANT: the CURIE the reviewer supplied for CDR2, hgnc:1802, is CDSN, an unrelated skin gene; the deep-research report proposed the same wrong ID. The correct identifiers are hgnc:1799 (CDR2) and hgnc:29999 (CDR2L), both already in the term cache. Recorded in the entry notes because two independent sources agreeing on a wrong identifier is exactly when it gets accepted unchecked. CDR2 is typed BIOMARKER, not CAUSATIVE or SUSCEPTIBILITY - nobody inherits this disease and no germline variant causes it; what CDR2 is, is the molecular identity of the attacked antigen, and no GeneDiseaseRelationshipEnum value expresses that. HLA is split into two records because HLA-DRB1 carries a risk haplotype and HLA-DPB1 a protective one, and relationship_type is per-record. 4. Failed antibody-transfer models added, as three animal_models entries covering all four studies (PMID:7707074, 7788981, 7788964, 11771954), each with typed divergences and readouts. These carry the negative results as FAILS_TO_RECAPITULATE links, which is what those slots are for. The DNA immunization study is the important one and it does not favour either hypothesis: it raised both anti-Yo antibodies and CDR2-specific cytotoxic T cells in vivo and still produced no degeneration. The knowledge gap now says so - no experimental reconstruction of either arm has produced the disease in an animal, and the canonical hypothesis is argued from human neuropathology and antigen compartment rather than from any positive model. 5. Surface-antigen arm broadened and the knowledge gap re-scoped. Anti-Tr/DNER (extracellular domain demonstrated, PMID:25745634) and anti-mGluR1 (42 pooled cases responding to immunotherapy, PMID:41197574) added to the ALTERNATIVE hypothesis. The gap is rewritten to separate what does not rest on one case (the compartment rule, now independently supported) from what does (the neuropathological contrast: topography, MHC class I, complement, ER stress). The reviewer's diagnosis of the overstatement was correct. 6. Treatments split from 2 bundled NCIT:C49236 records into 7 with precise bindings: Pharmacotherapy + glucocorticoid, Intravenous Immunoglobulin Therapy (NCIT:C121331), Plasmapheresis (NCIT:C15304), B-Cell Depletion (ofatumumab NCIT:C66952 / rituximab NCIT:C1702), Surgical Procedure, Chemotherapy, Supportive Care. All four new bindings confirmed reachable from NCIT:C25218. Cyclophosphamide and 4-aminopyridine are NOT added: no cached reference supports either in this disease, and inventing a treatment record without evidence is worse than omitting it. 7. Diagnosis expanded with antibody-guided imaging (EFNS task force, PMID:20880069, graded OTHER because the report states no evidence-based grade A-C recommendations were possible) and FDG-PET/CT for the antibody-negative case (PMID:22157026), quoted with its actual 27-patient, five-cancer yield rather than only its recommendation. 8. Checkpoint-inhibitor-triggered disease added as an environmental exposure with a TRIGGERS mechanism link, plus a new INTERPRETATION discussion (pcd_ici_triggered_same_disease) on whether it is the same disease. It is curated as an exposure and deliberately not as a subtype: the one direct comparison finds the two differ on sex, tumour type, isolated ataxia and outcome simultaneously, which is too much to assert sameness over, and too little (35 versus 15, retrospective) to assert difference. Self-caught errors during this round, recorded because they are the interesting part: (a) the two new diagnosis records were first inserted before `discussions:` and silently absorbed into `animal_models` - YAML parsed, schema passed, and only a structural readback caught it; (b) a snippet was truncated at "follow-" with a fabricated justification about the cache hard-wrapping the line, which it does not - the full sentence was substituted; (c) a snippet anglicised the source's "tumor" to "tumour", caught by the reference validator. Validated: just validate-disorders (62/62 snippets), just validate-terms, plus check-snippet-length, check-title-snippets, check-snippet-grading, check-folded-hyphens, check-environmental-evidence, check-enum-values, check-duplicate-keys, check-entity-refs, check-causal-targets, check-qualifier-terms, model-scale-audit --strict, check-not4curation. All green.

Create: Paraneoplastic Cerebellar Degeneration · 2026-09-06T09:08:38Z · View source

De novo curation of paraneoplastic cerebellar degeneration as a syndrome-level entry alongside the existing compartment-level Paraneoplastic_Neurological_Syndromes entry. One openscientist deep-research run was performed and is committed. The entry is organised around a two-mechanism structure taken from PMID:36070310, a neuropathological comparison of anti-Yo against anti-P/Q-VGCC cases: where the onconeural antigen is intracellular the pathology is cytotoxic T cell-mediated with MHC class I upregulation and CD8 granzyme B cells, and where it is on the cell surface there is no MHC class I upregulation or complement deposition and the picture is one of antibody-driven dysfunction preceding death. These are curated as two mechanistic_hypotheses (CANONICAL and ALTERNATIVE) with each causal edge opting into one via hypothesis_groups, rather than as rival accounts of one mechanism. The prognostic consequence is stated: the uniformly poor outlook attributed to this disease derives from anti-Yo series and should not be assumed to cover the surface-antigen arm. Three limitations are recorded as discussions rather than hedged into node text: the antibody-versus-T-cell effector debate, the compartment-determines-window interpretation, and the fact that the entire surface-antigen contrast rests on a single autopsy case. A plasma exchange case report with an unusually good outcome is curated as a genuine counterexample to the poor prognosis, with its atypia (male, no tumour found) stated so it is not read as representative. One snippet was initially written as a paraphrase and corrected against the cache after the snippet audit flagged it; the true wording was stronger than the paraphrase. Validated: schema, terms, 21/21 snippets, entity refs, causal targets.

OpenScientist ▸

Paraneoplastic Cerebellar Degeneration (PCD): A Comprehensive Disease Characteristics Report

openscientist-autonomous 36 citations 2026-09-06T09:05:01.422741

Paraneoplastic Cerebellar Degeneration (PCD): A Comprehensive Disease Characteristics Report

MONDO ID: MONDO:0044877 · Category: Complex (immune-mediated, non-Mendelian) Report type: Disease knowledge-base synthesis from literature (69 papers reviewed; 12 confirmed findings) Evidence base: Human clinical cohorts and case series, model-organism studies, in vitro immunology

Summary

Paraneoplastic cerebellar degeneration (PCD) is a rare, immune-mediated, subacute pancerebellar syndrome that arises as a remote effect of cancer. It is not a genetic disease in the Mendelian sense; instead, the initiating lesion is a malignancy that ectopically expresses a neuronal ("onconeural") antigen normally restricted to cerebellar Purkinje cells. This aberrant expression breaks immune tolerance in genetically predisposed (HLA Class II) hosts, triggering an anti-tumor immune response that cross-reacts with the cerebellum. The dominant effector mechanism is cytotoxic CD8+ T-cell-mediated destruction of Purkinje cells, with autoantibodies serving primarily as diagnostic biomarkers rather than the direct cause of neuronal death (for the intracellular-antigen subtypes). The clinical result is subacute (weeks-to-months) progressive gait and limb ataxia, dysarthria, nystagmus, and dizziness, typically plateauing within ~6 months, often leaving patients severely and irreversibly disabled.

The single most important organizing principle to emerge from this investigation is the antigen-location prognosis rule: antibodies against intracellular antigens (anti-Yo/CDR2, anti-Hu/ANNA-1, Ma2) mark a T-cell-driven, largely irreversible process with poor neurological outcomes, whereas antibodies against cell-surface antigens (anti-DNER/Tr, anti-mGluR1) mark a more antibody-mediated, treatment-responsive process. Anti-Yo (PCA-1), directed against CDR2/CDR2L, is the most common variant, occurring almost exclusively in women with gynecologic or breast cancer, and carries a dismal prognosis (≈84% of survivors unable to walk unassisted). Anti-Tr/DNER PCD, by contrast, occurs mostly in middle-aged men with Hodgkin lymphoma and is more treatable.

Management rests on two pillars applied urgently: prompt tumor detection and treatment (antibody-guided screening, with FDG-PET/CT central when conventional imaging is negative) and immunotherapy (corticosteroids, IVIG, plasma exchange, cyclophosphamide, rituximab). Good outcomes correlate with early diagnosis, low disability at presentation (mRS <3), absence of metastasis, and combined immunotherapy plus tumor-directed therapy. An emerging iatrogenic trigger is the class of immune checkpoint inhibitors (ICIs), which can unmask or amplify pre-existing onconeural immunity. Population data indicate PCD is the second most common paraneoplastic neurological syndrome (≈28% of PNS), against an overall PNS incidence of ~0.89/100,000 person-years.

Section 1 — Disease Information

Overview. PCD is an autoimmune cerebellar syndrome triggered by an underlying (often occult) malignancy. It belongs to the broader family of paraneoplastic neurological syndromes (PNS) and immune-mediated cerebellar ataxias (IMCAs). The typical presentation is "the subacute development of pancerebellar deficits with a clinical plateau within 6 months" (PMID: 27606347).

Key identifiers. - MONDO: MONDO:0044877 - MeSH: Paraneoplastic Cerebellar Degeneration - ICD-10: G13.1 (Systemic atrophy primarily affecting central nervous system in neoplastic disease) / D48.9 with paraneoplastic manifestation - Orphanet: Paraneoplastic cerebellar degeneration (rare neurological PNS) - OMIM: Not applicable — PCD is not a heritable Mendelian disorder; no OMIM disease entry (relevant onconeural genes have Gene OMIM entries, e.g., CDR2).

Synonyms / alternative names. Paraneoplastic cerebellar degeneration; subacute cerebellar degeneration (paraneoplastic); anti-Yo/PCA-1 cerebellar ataxia (for that subtype); paraneoplastic cerebellar ataxia; onconeural cerebellar syndrome. Antibody-defined subtypes: PCA-1 (anti-Yo), PCA-Tr (anti-Tr/DNER), ANNA-1 (anti-Hu), ANNA-2 (anti-Ri), PCA-2 (anti-MAP1B).

Information source type. This report is derived from aggregated disease-level resources — systematic reviews, laboratory serology cohorts (Mayo Clinic, Barcelona), population-based epidemiology, and case series — rather than from individual EHR patient records.

Section 2 — Etiology

Primary cause. The causal factor is an underlying malignancy that ectopically expresses a neuronal antigen. In anti-Yo disease, "the Yo autoantibodies are directed against the Yo antigens, aberrantly overexpressed by tumor cells with frequent somatic mutations and gene amplifications" (PMID: 38494293). The disease is therefore fundamentally a cancer-triggered autoimmune process, not genetic, environmental (toxic), or infectious in origin.

Tumor-type risk factors (antibody-dependent). - Anti-Yo → gynecologic (ovarian, endometrial) and breast carcinoma; 96% female, 82% gynecologic cancer (PMID: 36334195). - Anti-Tr/DNER → Hodgkin lymphoma, middle-aged males (PMID: 34817790). - Anti-Hu, CV2/CRMP5, PCA-2/MAP1B, ZIC4 → small-cell lung cancer (SCLC), often in smokers. - Anti-Ma2 → testicular germ-cell tumors (young men <50), also lung/breast.

Genetic risk factors. Host susceptibility is conferred by HLA Class II haplotypes rather than causal coding mutations. High-resolution typing of 40 anti-Yo cases identified protective haplotypes (DPA1*01:03~DPB1*04:01, OR=0, p=0.0008) and an ovarian-cancer-specific susceptibility haplotype (DRB1*13:01~DQA1*01:03~DQB1*06:03, OR=5.4, p=0.0016), indicating "differential genetic susceptibility to anti-Yo per cancer and with a primary HLA Class II involvement" (PMID: 29306402).

Environmental / lifestyle risk factors. Tobacco smoking is a strong indirect risk factor via SCLC-associated subtypes (anti-Hu, CRMP5). No direct toxic, occupational, or infectious cause of PCD itself is established.

Iatrogenic trigger. Immune checkpoint inhibitors (anti-PD-1/PD-L1/CTLA-4) are an emerging cause; they can "amplify pre-existing onconeural immunity" (PMID: 42442848).

Protective factors. The only clearly documented protective factors are the HLA Class II protective haplotypes above. No dietary or lifestyle protective factors are established.

Gene–environment interaction. The model is: a tumor (environmental/somatic event) expressing an onconeural antigen, in a host carrying a permissive/susceptible HLA Class II genotype, produces tolerance breakdown. Molecular mimicry directs the specificity, and co-signaling molecules (checkpoint pathways) modulate the strength — a concept reinforced by the ICI-triggered cases (PMID: 39052041).

Section 3 — Phenotypes

The core phenotype is a subacute, progressive pancerebellar syndrome (HPO: HP:0001251 Ataxia; HP:0002070 Limb ataxia; HP:0002066 Gait ataxia).

Phenotype	HPO term	Type	Frequency / notes
Gait ataxia	HP:0002066	Clinical sign	Near-universal; often presenting feature
Limb ataxia / dysmetria	HP:0002070 / HP:0001310	Clinical sign	Very common
Dysarthria	HP:0001260	Clinical sign	Common
Nystagmus (incl. downbeat)	HP:0000639	Clinical sign	Common; downbeat characteristic
Vertigo / dizziness	HP:0002321	Symptom	Prodromal in ~two-thirds of anti-Yo patients (PMID: 36334195)
Diplopia / oscillopsia	HP:0000651 / HP:0011495	Symptom	Frequent
Truncal instability	HP:0002078	Clinical sign	Common
CSF lymphocytic pleocytosis	HP:0012310	Lab abnormality	Frequent, inflammatory CSF
Oligoclonal bands / elevated IgG index	HP:0032101	Lab abnormality	Frequent

Characteristics. Onset is adult/late-adult (median age ~60 for anti-Yo, ~68 in the population PNS cohort). Course is subacute and progressive over weeks to months, with "a clinical plateau within 6 months" (PMID: 27606347). Severity is typically severe for intracellular-antigen subtypes: despite treatment, 84% of anti-Yo survivors are unable to walk unassisted at follow-up (PMID: 36334195). "Vertigo and imbalance can be present early in the disease course in about two thirds of patients, as a prodromal phase" (PMID: 36334195).

Quality-of-life impact. Profound — most patients with intracellular-antigen PCD become wheelchair-dependent, non-ambulatory, and dependent for activities of daily living, with dysarthria impairing communication. Surface-antigen subtypes may recover substantially.

Extracerebellar features. Rare in isolated anti-Tr/DNER PCD (8%) (PMID: 34817790); anti-Hu and anti-Ma2 more often present with multifocal encephalomyelitis, limbic/brainstem involvement, and peripheral neuropathy.

Section 4 — Genetic / Molecular Information

PCD has no causal germline disease gene. The relevant molecules are the onconeural target antigens and the host HLA Class II susceptibility loci.

Onconeural antigens (target autoantigens).

Antibody (alias)	Target antigen / gene	Antigen location	Typical tumor
Anti-Yo (PCA-1)	CDR2 / CDR2L	Intracellular (cytoplasm)	Ovarian, breast
Anti-Tr (PCA-Tr)	DNER (Delta/Notch-like EGF-related receptor)	Cell surface	Hodgkin lymphoma
Anti-Hu (ANNA-1)	HuD / ELAVL family	Intracellular (nuclear)	SCLC
Anti-Ri (ANNA-2)	NOVA1/2	Intracellular (nuclear)	Breast, lung
Anti-Ma2	PNMA2 (Ma2)	Intracellular	Testicular germ-cell
Anti-CV2/CRMP5	CRMP5/DPYSL5	Intracellular	SCLC, thymoma
Anti-PCA-2	MAP1B	Intracellular	SCLC
Anti-mGluR1	GRM1 (metabotropic glutamate receptor 1)	Cell surface	Often non-paraneoplastic / lymphoma
Anti-Homer-3	HOMER3	Intracellular/postsynaptic	Breast adenocarcinoma
Anti-ZIC4	ZIC4	Intracellular	SCLC

The anti-Yo antigen designation and target are established: anti-Yo is "directed against cerebellar degeneration-related protein 2 (CDR2) and CDR2L" and is "the most common variant of paraneoplastic cerebellar degeneration" (PMID: 27606347). Anti-Tr binds "the extracellular domain of DNER" (PMID: 25745634).

Somatic vs germline. The genetic aberrations are somatic in the tumor — "aberrantly overexpressed by tumor cells with frequent somatic mutations and gene amplifications" of the Yo antigen genes (PMID: 38494293). There is no germline pathogenic variant driving PCD.

Modifier genes / functional consequences. HLA Class II haplotypes modify susceptibility per cancer type (PMID: 29306402). Epigenetic and chromosomal abnormality data specific to PCD are not established beyond tumor-level somatic changes.

HGNC / Gene identifiers: CDR2 (HGNC:1802), CDR2L (HGNC:14002), ELAVL4/HuD, PNMA2, DPYSL5/CRMP5, MAP1B, GRM1, DNER, HOMER3, ZIC4.

Section 5 — Environmental Information

Environmental/toxic factors: No direct environmental toxin causes PCD. The relevant "environmental" exposure is tobacco smoke, acting indirectly by causing SCLC (the tumor substrate for anti-Hu, CRMP5, PCA-2, ZIC4 subtypes).
Lifestyle factors: Smoking (via SCLC). No dietary triggers established.
Infectious agents: None causative. PCD is a paraneoplastic (tumor-driven) autoimmune process, though molecular mimicry conceptually parallels post-infectious immune cerebellar ataxias in the broader IMCA family (PMID: 39052041).
Iatrogenic: Immune checkpoint inhibitor therapy (see Sections 2, 6, 12).

Section 6 — Mechanism / Pathophysiology

Ordered causal chain

A malignancy arises (e.g., ovarian/breast carcinoma, SCLC, Hodgkin lymphoma) and, through somatic mutation/gene amplification, ectopically expresses an onconeural antigen (CDR2/CDR2L, HuD, DNER, Ma2) normally restricted to neurons → results in presentation of a neural self-antigen to the immune system in an immunogenic (tumor) context (PMID: 38494293).
In a host carrying permissive HLA Class II haplotypes, this ectopic antigen presentation leads to breakdown of immune self-tolerance (PMID: 29306402).
Tolerance breakdown results in an anti-tumor adaptive immune response: generation of onconeural autoantibodies AND antigen-specific T cells. (For surface antigens, the branch is antibody-dominant; for intracellular antigens, the branch is T-cell-dominant — see branch below.)
Intracellular-antigen branch (Yo, Hu, Ma2): cross-reactive cytotoxic CD8+ T lymphocytes (CTLs) traffic across the blood–brain barrier into the CSF/cerebellum → leads to antigen recognition of Purkinje cells → results in CTL-mediated Purkinje cell apoptosis/death (PMID: 10632096; PMID: 9879687). Antibody alone is insufficient to cause degeneration (inferred from failed passive-transfer/immunization models) (PMID: 7707074; PMID: 7788981).
Surface-antigen branch (DNER, mGluR1): autoantibodies bind the extracellular domain of the target → results in receptor dysfunction/internalization that is potentially reversible → more treatment-responsive disease (PMID: 25745634; PMID: 41197574).
Purkinje cell loss leads to loss of the sole output neuron of the cerebellar cortex → results in pancerebellar dysfunction (gait/limb ataxia, dysarthria, nystagmus).
Progressive Purkinje depletion leads to cerebellar atrophy (radiographically visible late) and, for intracellular subtypes, irreversible clinical disability (PMID: 27606347).

 Tumor (ectopic onconeural antigen: CDR2/CDR2L, HuD, DNER, Ma2)
    │  (somatic mutation/amplification)
    ▼
 HLA Class II–permissive host  ──►  Tolerance breakdown
    │
   ┌────────┴─────────┐
   ▼                  ▼
 INTRACELLULAR Ag    SURFACE Ag
 (Yo, Hu, Ma2)       (DNER, mGluR1)
   │                  │
 CD8+ CTL response   Antibody binds
 (dominant)          extracellular domain
   │                  │
 Purkinje-cell        Receptor dysfunction
 cytotoxic killing    (± reversible)
   │                  │
 IRREVERSIBLE  ◄──── vs ────►  TREATMENT-RESPONSIVE
 disability                    recovery
   │
   ▼
 Purkinje cell loss → pancerebellar syndrome → cerebellar atrophy

Molecular pathways / cellular processes. Adaptive immunity: MHC Class I/II antigen presentation, T-cell receptor engagement, CTL granule-mediated cytotoxicity, apoptosis (GO:0006915), and antigen-specific B-cell/plasma-cell antibody production. In active PCD CSF, ">75% of cells were CD3+ alphabeta T cells and 20–40% were activated T cells," and "activated cdr2-specific CTLs in the CSF contribute to Purkinje degeneration in PCD" (PMID: 10632096). CASPR2-associated cerebellar ataxia similarly shows combined CD8+ T-cell and CD138+ plasma-cell CSF infiltration (PMID: 22759321).

Protein dysfunction. For intracellular antigens the antigen is a normal neuronal protein (loss of Purkinje cells eliminates its function); for surface antigens (DNER, a Notch-pathway EGF-repeat receptor; mGluR1, essential for motor coordination/learning), antibody binding impairs signaling. Notably Homer-3's partner mGluR1A "is predominantly expressed in Purkinje cells where its function is essential for motor coordination and motor learning" (PMID: 35871640).

Immune system involvement. Central and defining — a cell-mediated (CTL) plus humoral autoimmune attack. Autoantibodies of multiple immunoglobulin classes bind Purkinje cytoplasm across species (PMID: 3346369).

Suggested GO / CL terms. GO:0006915 (apoptotic process), GO:0002456 (T cell mediated immunity), GO:0001913 (T cell mediated cytotoxicity), GO:0002376 (immune system process), GO:0019882 (antigen processing and presentation). CL: CL:0000121 (Purkinje cell — primary target), CL:0000794 (CD8-positive, alpha-beta cytotoxic T cell — effector), CL:0000786 (plasma cell), CL:0000909 (CD8-positive alpha-beta memory T cell).

Section 7 — Anatomical Structures Affected

Primary organ / system: Central nervous system — the cerebellum (UBERON:0002037), specifically the cerebellar cortex (UBERON:0002129).
Primary cell target: Purkinje cells (CL:0000121; UBERON: Purkinje cell layer UBERON:0002974). Purkinje-cell loss is the pathological hallmark; molecular-layer interneurons are also stained by autoantibodies (PMID: 3346369).
Secondary involvement: Brainstem, limbic system, diencephalon, and peripheral nerves in overlap syndromes (anti-Hu encephalomyelitis, anti-Ma2 limbic/diencephalic/brainstem encephalitis, anti-Ri brainstem syndrome).
Tissue type: Nervous tissue.
Subcellular compartments: Depends on antigen — cytoplasm (CDR2/CDR2L; GO:0005737), nucleus (HuD, Ri; GO:0005634), plasma membrane / cell surface (DNER, mGluR1; GO:0005886), postsynaptic density (Homer-3; GO:0014069).
Localization / lateralization: Bilateral, symmetric cerebellar involvement is typical (pancerebellar). FDG-PET may show bilateral cerebellar hypometabolism (early) or hyperperfusion in some inflammatory cases.

Section 8 — Temporal Development

Onset: Adult to older-adult (median ~60–68 years). Onset pattern is subacute — evolving over weeks to a few months.
Prodrome: Vertigo/imbalance precedes the full syndrome in ~two-thirds of anti-Yo patients (PMID: 36334195).
Progression: Rapidly progressive during the active phase, then plateaus within ~6 months (PMID: 27606347), after which deficits are usually fixed (intracellular subtypes). Cerebellar atrophy appears late on MRI.
Course pattern: Monophasic progressive-then-plateau for classic onconeural PCD; surface-antigen cases may relapse or remit with immunotherapy.
Duration: Chronic/lifelong disability once the plateau is reached in intracellular-antigen disease.
Critical window: There is a narrow therapeutic window before irreversible Purkinje loss; "early tumor detection, diagnosis, and PCD treatment are essential because any delay can result in the progression of the disorder and irreversible neurological damage" (PMID: 35501715).
Tumor timing: The neurological syndrome usually precedes cancer diagnosis; screening should be repeated over time (tumors may surface years later) (PMID: 22157026).

Section 9 — Inheritance and Population

Epidemiology: In a population-based 9-year Italian study (983,190 people), overall PNS incidence was 0.89/100,000 person-years and prevalence 4.37/100,000; "PNS developed in 1 in every 334 cancers" (PMID: 31552550). Cerebellar degeneration was the second most common PNS at 28% (after limbic encephalitis 31%): "Most common PNS were limbic encephalitis (31%), cerebellar degeneration (28%) and encephalomyelitis (20%)" (PMID: 31552550). Antibody specificities in that cohort: Yo 30%, Hu 26%, Ma2 22%; associated tumors lung 17%, breast 16%, lymphoma 12%.
Inheritance: Not heritable — PCD is acquired/autoimmune. The only genetic contribution is HLA Class II susceptibility (polygenic/multifactorial predisposition), not a Mendelian inheritance pattern.
Penetrance / expressivity / anticipation / mosaicism / founder effects / consanguinity / carrier frequency: Not applicable (no causal germline mutation).
Sex ratio: Strongly subtype-dependent. Anti-Yo: 96% female (PMID: 36334195). Anti-Tr/DNER: predominantly middle-aged males (PMID: 34817790). Anti-Ma2: predominantly young males (testicular tumors). Overall PNS cohort ~52% female, median age 68 (PMID: 31552550).
Age distribution: Adult/older adult, tracking the age of the underlying cancers.

Section 10 — Diagnostics

Clinical / CSF. CSF is frequently inflammatory: "lymphocytic pleocytosis, elevated protein, elevated IgG index, and oligoclonal bands" (PMID: 27606347; PMID: 35871640).

Antibody testing (the diagnostic cornerstone). Serum and CSF onconeural antibody panels: anti-Yo/CDR2/CDR2L, anti-Hu, anti-Ri, anti-Tr/DNER, anti-Ma2, anti-CV2/CRMP5, anti-PCA-2/MAP1B, anti-mGluR1, anti-Homer-3, anti-ZIC4. The anti-Tr/DNER recombinant cell-based assay reaches 100% sensitivity/specificity (PMID: 25745634). Rodent cerebellar tissue reliably reproduces the diagnostic Purkinje-cytoplasm immunostaining (PMID: 3346369).

Imaging. "Magnetic resonance imaging of the brain is often normal in the early stages, with cerebellar atrophy seen later" (PMID: 27606347). FDG-PET can show cerebellar hypometabolism.

Tumor screening (antibody-guided, time-critical). Per the EFNS task force, "the nature of antibody, and to a lesser extent the clinical syndrome, determines the risk and type of an underlying malignancy," and "for screening of the thoracic region, a CT-thorax is recommended, which if negative is followed by fluorodeoxyglucose-positron emission tomography (FDG-PET)" (PMID: 20880069). FDG-PET/CT detected malignancy in ~19% of suspected PNS patients and is recommended regardless of antibody status (PMID: 22157026). Screening should be repeated if initially negative, because the tumor usually postdates neurological onset.

Diagnostic criteria. The 2021 updated PNS diagnostic criteria replaced "classical syndromes" with "high-risk phenotypes," reclassified antibodies as high-risk (>70% cancer association) vs intermediate-risk (30–70%), and introduced the PNS-Care Score combining phenotype, antibody, cancer presence, and follow-up to grade diagnoses as definite/probable/possible (PMID: 34006622).

Differential diagnosis. Hereditary/degenerative ataxias, multiple system atrophy (cerebellar type), toxic/metabolic cerebellar injury, gluten ataxia, anti-GAD ataxia, post-infectious cerebellitis, primary autoimmune cerebellar ataxia (PACA), and metastatic/leptomeningeal disease (PMID: 39052041; PMID: 35618871).

Section 11 — Outcome / Prognosis

Antigen-location prognosis rule (central finding). Intracellular-antigen antibodies (Yo, Hu) predict irreversible disability; surface-antigen antibodies (DNER, mGluR1) predict a treatable course. In anti-Yo disease, despite treatment, 84% of survivors are unable to walk unassisted at follow-up (PMID: 36334195). In anti-mGluR1 cerebellar syndrome, "the majority of patients showed clinical improvement (n = 31)" of 42 (PMID: 41197574).

Predictors of good outcome. In a 97-patient PNS cohort with good outcome in 54.6%, "factors associated with good outcome were: early diagnosis, mRS <3 at presentation, absence of metastatic disease, and adjuvant immunotherapy" (PMID: 33911377).

Oncologic paradox. PNS can prompt earlier cancer detection, conferring a survival advantage despite neurological morbidity: ANNA1/Hu-IgG PNS + SCLC patients had "a 41% lower hazard of death" than SCLC-only patients (HR=0.59, 95% CI 0.37–0.96) (PMID: 42233990).

Subtype-specific prognosis. Anti-Ma2-only patients fare significantly better than anti-Ma (Ma1+Ma2) patients — "the clinical outcome was significantly better in the anti-Ma2 group" (PMID: 27460184). ICI-related cerebellar ataxia: 46% improved but with residual disability (PMID: 39153058). In ICI-related PNS more broadly, risk-antibody positivity carried 29% mortality vs 10% in antibody-negative patients (P=0.012) (PMID: 41488641).

Morbidity / QoL. High disability burden: wheelchair dependence, dysarthria, and loss of independence dominate the intracellular-antigen subtypes.

Section 12 — Treatment

Treatment rests on two urgent pillars: (1) prompt tumor removal/therapy and (2) immunotherapy.

Tumor-directed therapy. Surgery, chemotherapy, or radiotherapy of the underlying malignancy is the mainstay and can stabilize or improve neurological symptoms.

Immunotherapy (NCIT: Immunotherapy). - Corticosteroids / IV methylprednisolone pulses (NCIT: Methylprednisolone) - Intravenous immunoglobulin (NCIT: Intravenous Immunoglobulin Therapy) - Plasma exchange (NCIT: Plasmapheresis) — can give repeated benefit in some cases (PMID: 31142706) - Cyclophosphamide (NCIT: Cyclophosphamide) - Rituximab / anti-CD20 (NCIT: Rituximab); ofatumumab reported in refractory anti-Yo (PMID: 39737186)

Response by subtype. "Patients with surface receptor autoimmunity … usually show a good response to treatment," in contrast to classical (intracellular) onconeural PNS (PMID: 29327271). Anti-Yo/intracellular cases often deteriorate despite aggressive therapy (PMID: 35501715). Early rituximab benefited non-tumor anti-DNER (PMID: 37991702) and anti-mGluR1 (PMID: 40760473) cases.

ICI-related PCD management. Discontinue the checkpoint inhibitor and start immunosuppression; outcomes are variable, and rechallenge can provoke relapse (PMID: 37151179).

Supportive / rehabilitative. Symptomatic agents (e.g., 4-aminopyridine for downbeat nystagmus/oscillopsia) plus physical, occupational, and speech therapy.

Overall. Good outcome in ~55% when combined immunotherapy + tumor treatment is applied early (PMID: 33911377).

Section 13 — Prevention

Primary prevention: Reducing cancer risk (e.g., smoking cessation to reduce SCLC-associated subtypes) is the only meaningful primary preventive lever; there is no vaccine or specific prophylaxis.
Secondary prevention (most actionable): Early recognition of a subacute cerebellar syndrome, prompt antibody testing, and antibody-guided tumor screening with FDG-PET to detect and treat the malignancy before irreversible Purkinje loss (PMID: 20880069; PMID: 22157026). In known cancer survivors, PCD can herald relapse (e.g., Hodgkin lymphoma) — vigilance enables early re-treatment.
Tertiary prevention: Immunotherapy plus rehabilitation to limit disability once disease is established.
Pre-ICI risk stratification: Pre-treatment screening for onconeural antibodies before checkpoint-inhibitor therapy may identify high-risk patients (PMID: 41488641).
Genetic counseling / carrier screening / immunization: Not applicable (non-heritable, non-infectious).

Section 14 — Other Species / Natural Disease

Taxonomy: Naturally occurring PCD is essentially a human (Homo sapiens, NCBI:txid9606) disease. No well-characterized spontaneous paraneoplastic cerebellar degeneration is documented in companion animals as a defined entity.
Related veterinary autoimmune cerebellar/encephalitic disease: Neuronal-autoantibody encephalitis is increasingly recognized in cats — including a reported case of anti-mGluR1 antibodies in a cat (PMID: 42656301) — but this was not clearly paraneoplastic.
Orthologous genes: CDR2/CDR2L, ELAVL4/HuD, Grm1, Dner, Pnma2, Homer3 have rodent orthologs (mouse Cdr2, Grm1, Dner), enabling model work.
Comparative pathology: Autoantibodies from human PCD sera bind Purkinje-cell cytoplasm in human, rat, and mouse cerebellum, indicating cross-species antigen conservation (PMID: 3346369).
Zoonotic potential: None (autoimmune, non-transmissible).

Section 15 — Model Organisms

Attempts to build a faithful animal model have repeatedly failed to reproduce degeneration by antibody alone, which is itself the key mechanistic evidence for T-cell causation.

Model approach	Result	Interpretation
Active immunization of mice (BALB/c, C3H, C57BL/6, SJL/J) with recombinant Yo	"All the strains produced high anti-Yo antibody titer but none developed cerebellar ataxia or showed Purkinje cell loss" (PMID: 7788981)	Antibody insufficient
Passive transfer of PCD patient IgG ± complement/macrophages into rodent brain	IgG taken up by Purkinje cells >36 h without cell loss; "could not be the sole cause of Purkinje cell loss" (PMID: 7707074, PMID: 7788964)	Antibody insufficient
DNA immunization against pcd17/cdr2	Induced antibodies AND CTLs, but "neither clinical nor pathological changes consistent with significant cerebellar degeneration" (PMID: 11771954)	Immunity inducible; degeneration not recapitulated — model limitation
In vitro human CTLs (HLA-A24) + recombinant Yo on autologous dendritic cells	CTLs reacted with Yo; "cytotoxic T cells are involved in Purkinje cell loss in PCD" (PMID: 9879687)	Positive evidence for CTL effector

Model type: Mammalian (mouse, rat, SCID mouse) and in vitro human cellular immunology. Genetic models: immunization/transgene (DNA immunization) rather than knockout disease models. Phenotype recapitulation: Poor — antibodies and even antigen-specific CTLs can be generated without overt cerebellar degeneration, a major limitation likely reflecting incomplete CNS T-cell trafficking or additional required signals. Applications: Established that antibody is not the sole effector and pointed to CTLs; useful for diagnostic reagent validation (rodent cerebellum for immunostaining).

Mechanistic Model / Interpretation

PCD is best understood as collateral autoimmune damage from an anti-tumor immune response. The unifying model is antigenic: a tumor ectopically expresses a neuronal protein → tolerance breaks in an HLA-Class-II-permissive host → adaptive immunity attacks both tumor and cerebellum. The antigen's subcellular location dictates the effector mechanism and therefore the prognosis:

Intracellular antigens (CDR2/CDR2L, HuD, Ma2) can only be seen by the immune system as MHC-presented peptides → CD8+ CTL cytotoxicity → Purkinje-cell death that is rapid, irreversible, and poorly antibody-treatable. The antibodies are bystander biomarkers.
Surface antigens (DNER, mGluR1) are directly accessible to antibodies → antibody-mediated receptor dysfunction that is often reversible → treatment-responsive disease.

This single axis explains the epidemiology (tumor associations by antibody), the diagnostics (antibody-guided screening), the treatment response gradient, and the prognosis. The failed animal models are not a gap but positive evidence: they demonstrate that circulating antibody, complement, and macrophages cannot kill Purkinje cells, forcing the conclusion that the effector is the cytotoxic T cell.

Evidence Base (key literature)

PMID	Contribution	Type
27606347	Anti-Yo most common PCD variant; CDR2/CDR2L; subacute pancerebellar course, MRI evolution	Human review
36334195	379-patient anti-Yo systematic review: 96% female, 82% gynecologic, 84% non-ambulatory	Human systematic review
38494293	Ectopic tumor overexpression of Yo antigens with somatic mutations/amplifications	Human review
34817790	85-patient anti-Tr/DNER review: middle-aged males, 91% tumor, Hodgkin lymphoma	Human systematic review
25745634	Anti-Tr binds extracellular DNER; 100% sensitive/specific CBA	Human/in vitro
10632096	Activated cdr2-specific CTLs in CSF drive degeneration	Human immunology
9879687	Patient CTLs react with recombinant Yo — CTL effector role	In vitro
11771954	DNA immunization induces antibody+CTL but no degeneration	Mouse model
7788981 / 7707074 / 7788964	Antibody alone insufficient for Purkinje loss	Mouse/rat model
29306402	HLA Class II susceptibility/protective haplotypes in anti-Yo	Human genetics
31552550	Population-based PNS incidence 0.89/100k; PCD = 28% of PNS	Human epidemiology
42233990	Anti-Hu survival paradox (41% lower death hazard w/ SCLC)	Human cohort
41197574	Anti-mGluR1: majority improve with immunotherapy	Human systematic review
33911377	Predictors of good outcome; 54.6% good outcome	Human cohort
20880069 / 22157026	Antibody-guided FDG-PET tumor screening	Guideline/human
34006622	2021 PNS diagnostic criteria & PNS-Care Score	Consensus guideline
27460184	Anti-Ma2 testicular tumors 40%; Ma2-only better outcome	Human series
39153058 / 42442848	ICI-related cerebellar ataxia; amplification of onconeural immunity	Human cohort/case
3346369	Cross-species Purkinje-cytoplasm antibody binding; diagnostic substrate	Human/animal

Limitations and Knowledge Gaps

No faithful animal model of the degeneration exists; the CTL effector mechanism is inferred from failed antibody-transfer experiments plus in vitro human CTL reactivity, not from a reproducible in vivo lesion.
Rarity and heterogeneity limit prospective, controlled treatment data; most evidence is from case series, retrospective cohorts, and systematic reviews of case reports — susceptible to publication and referral bias (serology-lab cohorts over-represent antibody-positive cases).
Molecular detail of tolerance breakdown (why specific tumors over-express onconeural antigens, and the exact epitope-spreading/mimicry events) is incompletely defined.
Prognostic biomarkers beyond antibody class/antigen location are lacking; no validated molecular predictor of immunotherapy response.
Epigenetic, transcriptomic, proteomic, and single-cell profiling of PCD cerebellum are largely absent from the literature reviewed — a genuine data gap.
HLA association data derive from a single modest cohort (n=40); replication across ancestries is needed.

Proposed Follow-up Experiments / Actions

Single-cell / spatial profiling of PCD-affected cerebellum and matched CSF (CITE-seq, TCR sequencing) to define the clonality and antigen-specificity of infiltrating CD8+ T cells and confirm the CTL model in situ.
Humanized HLA-transgenic mouse models expressing onconeural antigens in a tumor context, with adoptive transfer of antigen-specific CD8+ T cells, to finally recapitulate Purkinje-cell loss.
Multi-center prospective registry with standardized PNS-Care scoring, antibody subtyping, and mRS trajectories to quantify outcomes and treatment effects by antigen location.
Replication of HLA Class II associations in larger, ancestrally diverse cohorts, extended to non-Yo subtypes.
Trials of T-cell-directed immunotherapy (e.g., agents targeting CD8+ CTLs or trafficking) for intracellular-antigen PCD, where B-cell-directed therapy underperforms.
Pre-ICI onconeural antibody screening protocols with prospective evaluation of whether screening plus surveillance reduces severe checkpoint-inhibitor cerebellar toxicity.
Biomarker discovery (CSF proteomics/neurofilament light) to identify early, treatable-window markers before irreversible atrophy.

Report compiled from 69 reviewed publications and 12 confirmed findings across 5 investigative iterations. Evidence types are labeled per claim; PMIDs link to primary sources.

Artifacts

Reference Validation

Checked with linkml-reference-validator 0.2.1.

Outcome	Count
References checked	36
Resolved	36
Unresolved (possible confabulation)	0
Unverifiable	0
Quoted claims checked	24
Quoted claims found in source	21
Quoted claims not found in source	3
References weighed for topical relevance	36
On topic	25
Off topic	0

Quotes not found in the cited source

Searched the abstract, any retrieved full text, and the title. A quote drawn from a part of the paper that was not retrieved will appear here too, so check before treating one as invented:

Every one of these was searched against an abstract alone, with no full text retrieved - marked abstract only below. Where full text can be fetched, re-running with it will settle them; where the source publishes only a summary to PubMed, as GeneReviews chapters do, it will not, and the quote has to be checked by hand against the chapter itself.

PMID:27606347 (abstract only): "lymphocytic pleocytosis, elevated protein, elevated IgG index, and oligoclonal bands"
closest text in source: "Cerebrospinal fluid abnormalities, primarily elevated protein, lymphocytic pleocytosis, and oligoclonal bands, are common in the early stages"
PMID:35871640 (abstract only): "lymphocytic pleocytosis, elevated protein, elevated IgG index, and oligoclonal bands"
closest text in source: "Examination of CSF showed a lymphocytic pleocytosis of 11 cells/µl and an intrathecal IgG synthesis of 26%"
PMID:7788964 (abstract only): "could not be the sole cause of Purkinje cell loss"
closest text in source: "We conclude that the anti-Yo antibody, either in combination with or without complement or activated mononuclear cells, cannot be the sole cause of Purkinje cell loss."

Term Validation

Checked with linkml-term-validator 0.4.5, through the ols: adapter.

Outcome	Count
Terms checked	31
Resolved	29
Unresolved (possible confabulation)	0
Obsolete	0
Unverifiable	2
Terms whose name was checked	17
Terms named correctly	10
Terms named as a different term	2
Terms whose name is worth a second look	5

Terms the report names something else

These identifiers resolve, so nothing about them looks wrong, and the ontology calls them something unrelated to what the report calls them. That usually means the identifier is not the one the sentence needs:

HP:0012310 (1 mention) - the report calls it "CSF lymphocytic pleocytosis"; HP calls it Abnormal monocyte count
HP:0032101 (1 mention) - the report calls it "Oligoclonal bands / elevated IgG index"; HP calls it Unusual infection

Terms whose name is worth a second look

The report's name for these is recognisably related to the term's own name without being one of them. A loose paraphrase reads the same way as a citation of the wrong sibling term - and so does a related synonym, which the ontology records precisely because it names something adjacent rather than the same thing - so these are listed rather than judged:

HP:0000639 (1 mention) - the report calls it "Nystagmus (incl. downbeat)"; HP calls it Nystagmus
HP:0002321 (1 mention) - the report calls it "Vertigo / dizziness"; HP calls it Vertigo, and lists "Dizziness" among its other names
HP:0002078 (1 mention) - the report calls it "Truncal instability"; HP calls it Truncal ataxia
CL:0000121 (2 mentions) - the report calls it "Purkinje cell — primary target"; CL calls it Purkinje cell
CL:0000794 (1 mention) - the report calls it "CD8-positive, alpha-beta cytotoxic T cell — effector"; CL calls it CD8-positive, alpha-beta cytotoxic T cell

Ask OpenScientist

OpenScientist Report

Classifications

Mechanistic Hypotheses

Discussions and Knowledge Gaps

Pathophysiology

Pathograph

Phenotypes

Genetic Associations

Medical Actions

Environmental Factors

Diagnosis

Prevalence

Animal Models

Source YAML

References & Deep Research

Deep Research

Paraneoplastic Cerebellar Degeneration (PCD): A Comprehensive Disease Characteristics Report

Summary

Section 1 — Disease Information

Section 2 — Etiology

Section 3 — Phenotypes

Section 4 — Genetic / Molecular Information

Section 5 — Environmental Information

Section 6 — Mechanism / Pathophysiology

Ordered causal chain

Section 7 — Anatomical Structures Affected

Section 8 — Temporal Development

Section 9 — Inheritance and Population

Section 10 — Diagnostics

Section 11 — Outcome / Prognosis

Section 12 — Treatment

Section 13 — Prevention

Section 14 — Other Species / Natural Disease

Section 15 — Model Organisms

Mechanistic Model / Interpretation

Evidence Base (key literature)

Limitations and Knowledge Gaps

Proposed Follow-up Experiments / Actions

Artifacts

Reference Validation

Quotes not found in the cited source

Term Validation

Terms the report names something else

Terms whose name is worth a second look