Lysosomal storage grouping review: completeness, members, literature and gaps

2026-08-20. Scope: the Lysosomal Storage Disorders grouping and its three nested groupings — Mucopolysaccharidoses, Mucolipidoses, Niemann-Pick Diseases. 48 disease entries reviewed. Literature scanned to August 2026.

Summary

The tree was structurally incomplete rather than merely under-populated. The LSD grouping carries a NECESSARY_AND_SUFFICIENT criterion, which makes an unlisted conforming disorder a contradiction, not a backlog item — and the audit reported 30 unlisted candidates against 14 listed members. Two further classes of defect sat underneath that count: four entries were invisible to the criteria machinery because they declared no conformance to the shared module, and one grouping criterion contradicted a member it should have admitted.

All of this is now fixed (commit cd53785, plus the joint-criterion refinement that follows it). The grouping audit reports 48/48 members SATISFIED, no contradictions, and an empty candidate list.

The content review is a different story. The structural tree is now sound; the entries it points at are, for the MPS / mucolipidosis / Niemann-Pick branches, substantially out of date with 2025–2026 literature, and 15 of 48 are stubs of under 450 lines. Details and a prioritized worklist below.

1. Completeness audit

1.1 What was missing

Grouping	Before	After	Added
Lysosomal Storage Disorders	14 immediate	36 immediate	1 nested grouping (NCL), 22 diseases, −1 redundant
Mucopolysaccharidoses	4	7	MPS VI, MPS VII, MPS IX
Mucolipidoses	3	4	ML III alpha/beta
Niemann-Pick Diseases	4	5	NPD type E

LSD grouping. The Neuronal Ceroid Lipofuscinoses grouping already existed with 7 members and was simply not wired in; nesting it removed 7 of the 30 candidates in one edit. The remaining 22 were added as direct DISEASE members, chosen to close a systematic bias — the original list was sphingolipidosis- and MPS-weighted, and carried almost nothing from the other arms of the LSD definition:

Arm of the definition	Previously	Now
Primary hydrolase deficiency	8 members	+ GM1 ×3, LIPA ×2, NAGA ×3, ACP2, Pompe subtypes ×2
Activator / cofactor deficiency	none	PSAP saposin disorders ×3, GM2A (Tay-Sachs AB variant)
Protective-protein deficiency	none	Galactosialidosis (CTSA)
Oligosaccharidosis / glycoproteinosis	sialidoses only	+ alpha-mannosidosis, aspartylglucosaminuria, glycoproteinoses
Trafficking failure, normal hydrolases	none	MPS-plus syndrome (VPS33A), HSP48 (AP5Z1)

The grouping's own description promises "hydrolase, activator, membrane transporter, or trafficking protein"; before this review only the first of those four was represented.

One redundancy removed. Niemann-Pick Disease Type C was listed as a direct DISEASE member and covered by the nested Niemann-Pick grouping, contradicting the grouping's stated rationale that Niemann-Pick subtypes are not mixed in at the LSD level. Dropped as a direct member.

Two stale notes corrected. The Niemann-Pick grouping recorded NPD-E as "intentionally not listed in this batch because an open PR already exists". That PR merged; the entry exists. Both the note and the MONDO consistency note were rewritten, and the consistency note now also records why NPD type D has no separate entry (an NPC1 Nova Scotia founder haplotype, covered by type C).

1.2 Entries invisible to the criteria machinery

Four entries could not satisfy criteria they were asserted against — the exact hazard CLAUDE.md warns about, where a drifted or absent conforms_to silently drops an entry out of a criterion it is supposed to satisfy.

Entry	Was	Now
Maroteaux-Lamy (MPS VI)	`mps_gag_storage` only	+ `lysosomal_substrate_accumulation#Lysosomal Hydrolase or Cofactor Deficiency`
Sly (MPS VII)	`mps_gag_storage` only	+ same
MPS IX	`mps_gag_storage` only	+ same
ML III alpha/beta	no `conforms_to` at all	+ hydrolase-deficiency and substrate-accumulation nodes, mirroring ML II

Each declaration restates what the node's own prose already says (ARSB, GUSB and HYAL1 deficiency are lysosomal hydrolase deficiencies), so this is wiring, not a new mechanistic claim.

1.3 A module hierarchy that was implicit

mps_gag_storage#Lysosomal GAG Accumulation is a specialization of lysosomal_substrate_accumulation#Lysosomal Substrate Accumulation and did not say so. It now declares conformance, following the existing module-to-module pattern (7 prior instances, e.g. er_protein_storage_disease → fibrotic_response).

Note for tooling, not fixed here: groupings.py matches a CONFORMS_TO_MODULE criterion on the module stem only — it ignores the #Node anchor, and it does not follow module→module conformance. So this declaration documents the hierarchy correctly but does not by itself make mps_gag_storage conformers satisfy a lysosomal_substrate_accumulation criterion. That is why the four entries above needed direct declarations. Worth an issue: transitive module conformance would make the MPS-branch wiring unnecessary and would prevent this class of silent drop-out recurring.

1.4 A criterion that contradicted a correct member

Adding MPS IX produced a NOT_SATISFIED contradiction. The cause was not the entry: MPS IX genuinely has neither coarse facies nor dysostosis multiplex. It stores hyaluronan rather than a sulfated GAG and presents with periarticular soft-tissue masses and short stature. The two-hallmark criterion encoded the classic MPS picture as necessary, and MPS IX is the counterexample.

The hallmark branch now carries two further operands — structural joint disease (HP:0001367) and restricted joint mobility (HP:0011729). These are kept separate deliberately: HPO splits joint structure from joint function, and the two MPS joint phenotypes fall on opposite sides of that split (MPS IX's joint swelling is morphological; the near-universal MPS joint stiffness is not). A single operand would have silently failed half the members.

2. Member-by-member review

Cited 2024+ counts evidence PMIDs ≥ 39000000 as a proxy for recency. ⚠️ marks entries under 450 lines.

Lysosomal Storage Disorders (direct members)

Entry	Lines	Evidence items	Cited 2024+	Phenotypes	Treatments	Trials	Models	Gaps
`Lysosomal_Acid_Phosphatase_Deficiency` ⚠️	175	9	0	1	0	0	1	0
`Hereditary_Spastic_Paraplegia_48` ⚠️	251	12	1	6	1	0	0	0
`Congenital_Sialidosis_Type_2` ⚠️	256	13	0	3	1	0	0	0
`Gaucher_Disease_Due_To_Saposin_C_Deficiency` ⚠️	256	13	0	2	1	0	0	0
`Krabbe_Disease_Due_To_Saposin_A_Deficiency` ⚠️	275	11	0	5	1	0	0	0
`Combined_Saposin_Deficiency` ⚠️	283	12	0	4	1	0	0	0
`GM1_Gangliosidosis_Type_3` ⚠️	287	14	0	4	1	0	0	0
`Schindler_Disease` ⚠️	307	20	1	7	1	0	0	0
`Glycoprotein_Storage_Disease` ⚠️	326	18	14	7	2	0	0	0
`Tay-Sachs_Disease_AB_Variant` ⚠️	326	15	0	6	1	0	0	0
`Cholesteryl_Ester_Storage_Disease` ⚠️	327	17	13	6	2	0	0	0
`Kanzaki_Disease` ⚠️	335	20	1	7	2	0	0	0
`NAGA_Deficiency_Type_3` ⚠️	349	15	1	10	1	0	0	0
`Spinal_Muscular_Atrophy_Progressive_Myoclonic_Epilepsy` ⚠️	358	21	9	8	1	0	0	0
`Juvenile_Sialidosis_Type_2` ⚠️	436	26	0	14	1	0	0	0
`GM1_Gangliosidosis_Type_2`	527	33	0	12	1	2	0	0
`Galactosialidosis`	547	16	0	12	1	0	0	0
`Metachromatic_Leukodystrophy`	649	40	21	6	2	0	0	0
`GM1_Gangliosidosis_Type_1`	984	61	2	19	4	4	0	0
`Pompe_Disease`	1362	85	13	17	8	0	0	0
`Krabbe_Disease`	1397	62	0	8	5	0	0	0
`Tay-Sachs_Disease`	1456	71	10	21	6	2	0	0
`Sandhoff_Disease`	1464	80	20	6	3	0	0	0
`Sialidosis_Type_1`	1468	97	13	20	8	0	0	0
`Wolman_Disease`	1542	119	42	13	5	0	2	2
`Infantile-Onset_Pompe_Disease`	1548	114	12	18	7	2	1	1
`Gaucher_Disease`	1641	95	6	19	6	0	0	0
`Aspartylglucosaminuria`	1827	58	0	41	4	1	0	0
`Late-Onset_Pompe_Disease`	1905	119	22	11	4	4	0	2
`Mucopolysaccharidosis-Plus_Syndrome`	2124	129	19	34	6	0	0	3
`Fabry_Disease`	2690	138	34	24	5	0	5	0
`Alpha_Mannosidosis`	2715	63	5	71	4	3	1	1

Mucopolysaccharidoses

Entry	Lines	Evidence items	Cited 2024+	Phenotypes	Treatments
`Mucopolysaccharidosis_type_IX` ⚠️	398	17	0	3	1
`Maroteaux-Lamy_syndrome`	615	37	0	11	3
`Sly_syndrome`	826	34	0	8	3
`Hurler_syndrome`	1242	66	0	16	3
`Morquio_syndrome`	1470	98	1	11	4
`Sanfilippo_syndrome`	1522	74	0	18	3
`Hunter_syndrome`	1646	80	3	13	4

Mucolipidoses

Entry	Lines	Evidence items	Phenotypes	Treatments
`Mucolipidosis_Type_IV` ⚠️	278	13	6	1
`GNPTG-Mucolipidosis` ⚠️	422	24	9	4
`Mucolipidosis_Type_II`	519	27	10	2
`Mucolipidosis_Type_III_Alpha_Beta`	1836	82	36	5

Niemann-Pick Diseases

Entry	Lines	Evidence items	Cited 2024+	Phenotypes	Treatments	Gaps
`Niemann-Pick_Disease_Type_E` ⚠️	182	8	0	2	0	1
`Chronic_Neurovisceral_Acid_Sphingomyelinase_Deficiency` ⚠️	363	20	0	4	2	0
`Niemann-Pick_Disease_Type_B` ⚠️	376	16	0	7	2	0
`Niemann-Pick_Disease_Type_A` ⚠️	397	20	0	6	2	0
`Niemann_Pick_Disease_Type_C`	1344	61	4	22	6	0

2.1 What the table shows

Recency is the dominant problem, and it is branch-specific. Every MPS entry except Hunter (3) and Morquio (1) cites nothing from 2024 onward. All four mucolipidosis entries cite nothing recent. Four of five Niemann-Pick entries cite nothing recent. The direct-LSD members are healthier — Fabry (34), Wolman (42), Late-Onset Pompe (22), MLD (21), Sandhoff (20) are current — but that is the exception across the tree, not the rule.

Clinical trials are almost entirely absent. All 7 MPS entries, all 4 mucolipidosis entries and all 5 Niemann-Pick entries carry zero clinical_trials records, despite these being among the most trial-active rare diseases in medicine. Section 3 lists specific trials that belong in them.

Knowledge gaps were essentially unrecorded. Before this review, 6 of 48 entries carried any KNOWLEDGE_GAP or HUMAN_MODEL_MISMATCH discussion (Alpha-mannosidosis 1, Wolman 2, IOPD 1, LOPD 2, MPS-plus 3, NPD-E 1). The other 42 carried none. Three cross-cutting gaps have now been added to the shared module (section 4); entry-specific gaps remain open work.

Models are nearly absent. 5 of 48 entries declare any animal or experimental model. For a disease group whose therapeutic pipeline is built on mouse, sheep, canine and iPSC models, and where ModelMechanismLink exists precisely to carry them, this is a large structural gap.

Two spot findings worth fixing directly:

Lysosomal_Acid_Phosphatase_Deficiency has zero treatments and one phenotype across 175 lines — the thinnest member of the tree.
Sly_syndrome carried no joint phenotype at all, though joint stiffness and contracture are cardinal MPS VII features. It satisfied the grouping criteria via dysostosis multiplex, so nothing flagged it; the criteria audit surfaced it incidentally. Fixed in this branch — HP:0001387 Joint stiffness added with GeneReviews evidence (PMID:38190471).

3. Latest publications (2025–2026)

Verified against PubMed; each PMID was resolved and its abstract read. Items marked ★ are, in my judgement, the ones that change what an entry should say.

3.1 The headline: brain-penetrant enzyme replacement is now approved

★ PMID:42313339 — Tividenofusp Alfa: First Approval (Mol Diagn Ther, Jun 2026). Tividenofusp alfa (AVLAYAH), iduronate-2-sulfatase fused to a transferrin-receptor-binding Fc domain, received US accelerated approval in March 2026 for the neurologic manifestations of MPS II. This is the first approved therapy that delivers a replacement lysosomal enzyme across the blood– brain barrier, and it directly addresses the defining therapeutic limitation of this entire disease group. Hunter_syndrome does not mention it.

The approval rests on reduction of CSF heparan sulphate, not on a neurocognitive outcome; continued approval is contingent on a confirmatory trial. That distinction is now recorded as a module-level knowledge gap (section 4.1) and must not be collapsed when curating.

Adjacent: PMID:42527655 (IGF2-tagged lentiviral gene therapy for Hunter syndrome), PMID:42496452 (combined intracerebroventricular ERT and cord blood transplantation in newborn-screened neuronopathic MPS II), PMID:42323476 (oral nanoparticle-encapsulated ERT for MPS I, proof of concept), PMID:42281208 (CSF-delivered bidirectional AAV9 improving optic nerve and retinal pathology in a sheep Tay-Sachs model).

3.2 Therapy: durability, discontinuation and long-term outcome

★ PMID:42047222 — ASCEND final results: adults with ASMD sustained improvement over up to 5 years of olipudase alfa; DLCO 50.1%→66.5% predicted, spleen volume −57.5%, liver −36.8%, plasma lyso-sphingomyelin −72%. None of the three SMPD1 entries cites olipudase long-term data.
★ PMID:42541184 — AAV9-GAA gene therapy after ERT discontinuation in six children with infantile-onset Pompe (ChiCTR2200065664): all remained ventilator-free at 12 months with reduced muscle glycogen, but lung and cardiac structure showed no further improvement. A carefully hedged result that belongs in Infantile-Onset_Pompe_Disease with its limitations intact.
PMID:42538546 — position statement defining the "therapeutic corridor of stability" for ERT in Pompe disease.
★ PMID:42551329 — 4-year real-world arimoclomol outcomes in NPC (US early access programme).
PMID:42421116 — idursulfase beta 2-year phase 3 extension in MPS II, including patients switched from idursulfase.
★ PMID:42617290 — International LAL-D Registry longitudinal lipid data on sebelipase alfa across paediatric and adult patients — directly relevant to both Wolman_Disease and Cholesteryl_Ester_Storage_Disease.
PMID:42504165 — early versus late ERT in Morquio A siblings: early treatment attenuated spinal and upper-body disease but did not prevent lower limb skeletal progression. A rare within-family natural-history control.
PMID:42547363, PMID:42526029 — HSC gene therapy and cross-correction in metachromatic leukodystrophy.

3.3 Mechanism

★ PMID:42587127 — targeting CD44 reverses sphingomyelin-induced oligodendrocyte maturation arrest in ASMD (EMBO Mol Med). A candidate mechanism for the neuronopathic arm that ERT does not reach.
★ PMID:42157966 — sinbaglustat is efficacious in GM2 gangliosidosis primarily through GBA2 inhibition rather than GCS. A target reassignment for a substrate-reduction agent; if the benefit is not glucosylceramide synthase inhibition, the substrate-reduction rationale for that class needs restating.
★ PMID:42467639 — levacetylleucine normalizes TFEB by reducing nuclear TFEB in an NPC cell model where it is already over-activated by lysosomal stress; the effect is stereospecific. See section 4.3.
PMID:42585307 — Mito-TEMPO improves survival in npc1-knockout zebrafish via Sod2-dependent mitophagy (Sci Adv).
PMID:42511762 — autophagy–lysosomal dysfunction as a converging cardiomyopathy mechanism across LSDs. A candidate for a new shared module, or for an additional node on the existing one.
PMID:42574251 — mucolipidosis II novel variants plus HAP1 cells as a disease model.

3.4 Nosology and phenotype expansion

★ PMID:42557761 — rare heterozygous MCOLN1 loss-of-function variants in two patients with α-synucleinopathies (MSA, early-onset PD), with reduced lysosomal currents on patch clamp. Extends the GBA1 carrier paradigm to the mucolipidosis IV gene. See section 4.2.
★ PMID:41830174, PMID:40081374, PMID:42568187 — biallelic AP5Z1/AP5B1 variants cause retinal degeneration and hereditary macular dystrophy. Hereditary_Spastic_Paraplegia_48 is curated as a pure spastic paraplegia and carries no retinal phenotype; this is a direct phenotype expansion for a member added in this review.
PMID:42351339 — biallelic HGSNAT variants causing autosomal recessive retinitis pigmentosa without overt Sanfilippo syndrome, with reduced enzyme activity. Relevant to Sanfilippo_syndrome as an attenuated non-syndromic allelic presentation.
PMID:42580759 — progressive myoclonic ataxia due to late-onset sialidosis.
PMID:42008923 — glycoproteinoses review covering clinical features, therapeutic landscape and regulatory pathways; the single best entry point for the six thin oligosaccharidosis members.
PMID:42482083 — systematic review of therapeutic approaches in alpha-mannosidosis.
PMID:42190545 — ClinGen lysosomal diseases VCEP ACMG/AMP specification for IDUA, with 131 variants classified. Authoritative variant-interpretation source for Hurler_syndrome.
PMID:42496454 — first-year results of MLD newborn screening in Lombardy.

4. Knowledge gaps

Three cross-cutting gaps were added to kb/modules/lysosomal_substrate_accumulation.yaml as KNOWLEDGE_GAP discussions with verified evidence (14/14 snippets verified). They are placed on the shared module rather than on individual entries because each applies across the whole tree, and every member conforms to it.

4.1 CNS delivery: corrected biomarker ≠ neurologic benefit

gap_cns_delivery_biomarker_versus_clinical_benefit, attached to Lysosomal Substrate Accumulation and Progressive Multisystem and Neurodegenerative Disease.

The first brain-penetrant ERT is approved on a CSF storage biomarker sitting at this module's own central effector node, with clinical benefit explicitly unverified. The same surrogate governs intrathecal and intracerebroventricular routes and CNS-directed gene therapy. The curation rule that follows: record the biomarker claim and the clinical-benefit claim as separate assertions, and never curate CSF substrate correction as evidence of neurologic efficacy. Evidence: PMID:42313339 (×2), PMID:42525369.

4.2 Heterozygous carriers and adult neurodegeneration

gap_heterozygous_lsd_carrier_neurodegeneration_risk, attached to Autophagic-Lysosomal Dysfunction and Storage-Cell Cytotoxicity.

GBA1 established that a single loss-of-function allele — carrier status for a member of this group — is among the strongest genetic risk factors for Parkinson disease without causing Gaucher disease. Whether that is a GBA1 peculiarity or a general property of partial lysosomal dysfunction is unresolved, and it determines whether carrier status belongs in the other members' entries at all. The proposed lesion is reduced lysosomal function without demonstrable storage, so it does not travel down this module's chain and must not be modelled as if it did. Evidence: PMID:42557761 (SUPPORT for the gap, PARTIAL for the association — the authors themselves call MCOLN1 a candidate needing cohort confirmation).

4.3 Which direction is TFEB dysregulated?

gap_tfeb_direction_of_dysregulation_in_storage, attached to Autophagic-Lysosomal Dysfunction and Secondary Cascade.

TFEB activation has been pursued as a substrate-clearing strategy across this group on the assumption that storage suppresses it. In an NPC cell model the baseline is the opposite — TFEB already over-activated and nuclear from chronic lysosomal stress — and levacetylleucine, an approved NPC therapy, works by lowering nuclear TFEB. If the baseline direction differs by disorder, cell type or stage, "TFEB activation" is not a coherent shared rationale, and a conforming entry should curate the measured direction in its own cells rather than inherit a directional assumption. Caveat preserved in the entry: the finding is from a HeLa model, not patient neurons. Evidence: PMID:42467639.

4.4 Gaps identified but not curated

These are real and evidenced but are entry-level rather than module-level, so they belong in the disorder entries and are left as work:

Substrate-reduction target ambiguity — if sinbaglustat acts via GBA2 rather than GCS (PMID:42157966), the mechanism of benefit for substrate-reduction therapy in the gangliosidoses is unsettled. Belongs on Tay-Sachs_Disease / Sandhoff_Disease.
Genotype–phenotype discontinuity in the allelic series — NAGA (Schindler / Kanzaki / type 3), LIPA (Wolman / CESD), GAA (infantile / late-onset) and GNPTAB (ML II / ML III alpha/beta) each span extreme phenotypic ranges from one locus. The residual-activity threshold model is assumed across the tree and directly tested nowhere in it.
Cardiomyopathy as a convergent LSD endpoint (PMID:42511762) — currently modelled, if at all, per-disease. Candidate for a shared node or module, alongside cardiomyopathy_maladaptive_remodeling.
Non-syndromic allelic presentations — HGSNAT retinitis pigmentosa without Sanfilippo (PMID:42351339), AP5Z1 macular dystrophy (PMID:41830174). Storage-gene variants presenting as isolated organ disease are a systematic ascertainment gap in a KB organized around syndromic entries.

5. Prioritized worklist

P1 — the entry is now wrong or materially incomplete

Hunter_syndrome: add tividenofusp alfa (PMID:42313339) with therapeutic_modality, target_mechanisms on the CNS storage node, and the accelerated-approval caveat. Add idursulfase beta extension (PMID:42421116).
Hereditary_Spastic_Paraplegia_48: add the AP5Z1 retinal/macular phenotype (PMID:41830174, PMID:40081374).
Three SMPD1 entries: add ASCEND 5-year olipudase outcomes (PMID:42047222).
Infantile-Onset_Pompe_Disease: add the AAV9-GAA post-ERT-discontinuation trial with its negative cardiopulmonary finding intact (PMID:42541184).
Wolman_Disease / Cholesteryl_Ester_Storage_Disease: LAL-D Registry longitudinal sebelipase data (PMID:42617290).
Lysosomal_Acid_Phosphatase_Deficiency: has no treatments and one phenotype.
~~Sly_syndrome: no joint phenotype annotated.~~ Done in this branch.

P2 — structural

Populate clinical_trials across the MPS, mucolipidosis and Niemann-Pick branches (all currently zero).
Open an issue for transitive module conformance in groupings.py (§1.3).
Resolve the umbrella Disease entries, of which this tree has two, both tracked on issue #4490:
- Mucopolysaccharidosis (MONDO:0019249) holds the same MONDO term the MPS grouping maps to with skos:exactMatch, so the concept exists twice in the KB, as a Disease and as a Grouping.
- Glycoprotein_Storage_Disease (MONDO:0009296) self-describes as "a group of autosomal recessive lysosomal storage disorders" and is now listed in the LSD grouping beside Aspartylglucosaminuria and Alpha-mannosidosis, two of the things it is an umbrella for — structurally the same duplication this review removed for Niemann-Pick type C. Caught in review of this work, not by the audit.
Both are grandfathered rather than fixed. The LSD criterion is NECESSARY_AND_SUFFICIENT and both the umbrella and its constituents conform to the module, so dropping either side would create a contradiction — worse than the redundancy. Note both have empty has_subtypes, so the duplication is invisible from the umbrella entry and only appears once a grouping lists umbrella and constituents as siblings; the has_subtypes check proposed on #4490 would catch neither. This is a lump/split decision for a curator, not a side-effect of a completeness pass. 11. Add modeled_mechanisms-linked models: 5 of 48 entries declare any.

P3 — content depth

The 15 sub-450-line entries, using PMID:42008923 (glycoproteinoses) and PMID:42482083 (alpha-mannosidosis) as entry points for the oligosaccharidosis cluster.
Entry-level knowledge gaps from §4.4.

6. Verification

Check	Result
`just check-groupings` (4 groupings)	48/48 members SATISFIED, 0 contradictions, 0 unlisted candidates
`just validate-grouping` ×4	passed
`just validate-disorders` (4 changed entries)	passed, 224/231 snippets verified (7 skipped by prefix)
`just validate` (module)	passed
`just count-verified-snippets` (module)	14/14 verified
`just check-duplicate-keys`	clean
`just validate-history-all`	5828 records, no issues
`pytest -k "conforms"`	2238 passed

Literature: PubMed E-utilities, pdat 2024-01-01 to 2026-12-31, ~25 queries across therapy, mechanism, delivery, screening and per-member axes. Every PMID cited above was resolved and its abstract read; the six used as evidence were cached via just fetch-reference and their snippets verified against the cache.