ITIH3 review notes

Review scope and provenance

Target: human ITIH3, UniProtKB Q06033 (inter-alpha-trypsin inhibitor heavy chain H3/HC3).
Review date: 2026-07-18.
The GOA stub contained nine source annotations, and every annotation was reviewed manually.
just deep-research-falcon human ITIH3 --fallback perplexity-lite was attempted. Falcon/Edison returned HTTP 402 and Perplexity-lite returned HTTP 401, so no provider-named deep-research file was created. The synthesis below instead uses the reviewed UniProt record, cached primary literature, and the current Reactome participant record.

Protein identity, processing, and complex assembly

ITIH3 is a secreted precursor processed to a mature HC3 chain. UniProt maps a signal peptide (residues 1-20), an N-terminal propeptide (21-34), the mature heavy chain (35-651), and a C-terminal propeptide (652-890), together with VIT and VWFA domains. The mature C-terminal Asp651 is ester-linked to the chondroitin-4-sulfate chain of bikunin after processing [UniProtKB:Q06033, "Heavy chains are linked to bikunin via chondroitin 4-sulfate"].

Human pre-alpha-inhibitor is composed of HC3 and bikunin. Direct structural analysis showed that the heavy and light chains are joined by a chondroitin-sulfate protein-glycosaminoglycan-protein cross-link [PMID:1898736, "The human blood protein pre-alpha-inhibitor is composed of one heavy and one light protein chain"]. Critically, bikunin is the inhibitory component: purified human pre-alpha-inhibitor "comprises bikunin (BK) responsible for antiprotease activity, covalently linked to a heavy chain H3" PMID:9188816. The older chain-stoichiometry study likewise found a single 30-kDa trypsin-inhibitory chain and a distinct 90-kDa heavy chain in pre-alpha-inhibitor [PMID:2476436, "pre-alpha-trypsin inhibitor contains a heavy chain of 90,000 Da"]. ITIH3/HC3 must therefore not inherit the protease-inhibitor activity of bikunin or of the named complex.

Direct hyaluronan and extracellular-matrix evidence

Purified human plasma HC3 contains an N-terminal BX7B hyaluronan-binding site, but the activity was conditional: a thermolysin-generated N-terminal fragment strongly bound HA, whereas intact mature HC3 did not [PMID:11168393, "a thermolysin-generated, N-terminal fragment of this H3 chain strongly bound HA whereas the intact H3 chain did not"]. This supports hyaluronic acid binding for a processed HC3 species, not constitutive noncovalent binding by every intact plasma molecule.
Purified human plasma experiments directly established transfer of HC3 from pre-alpha-inhibitor through a covalent HC3-TSG-6 intermediate to an HC3-hyaluronan product. HC2 is an essential cofactor: "Proteins containing HC2 including IalphaI, HC2.bikunin, and free HC2 promoted the formation of HC3.TSG-6 and subsequently HC3.hyaluronan complexes" PMID:18448433. ITIH3 is the transferred substrate/structural heavy chain; TSG-6 and HC2 provide the transfer machinery.
Mouse biochemical studies found HC3 in physiological and pathological SHAP-hyaluronan complexes as well as in complexes formed in vitro [PMID:18293178, "All three HC isoforms are found in the SHAP-hyaluronan complexes of physiological or pathological origins as well as that formed in vitro"]. Together with the direct human transfer experiment, this supports an extracellular-matrix location for human HC3 by orthology.
In human airway studies, TSG-6 mediates transfer of heavy chains from inter-alpha-inhibitor/pre-alpha-inhibitor onto HA, while TSG-6-driven release of bikunin increases antiprotease activity [PMID:16873769, "TSG-6 mediates the transfer of the heavy chains from IαI/PαI onto HA"; "bikunin in its free state, but not when associated with HCs, is a relevant protease inhibitor"]. The paper also directly detected HC3 expression in human airway epithelial cells, but this inflammatory context does not justify a broad human airway-disease process annotation for ITIH3.

Annotation decisions

Source term	Decision	Rationale
serine-type endopeptidase inhibitor activity	REMOVE	InterPro transfers the activity of bikunin/the named complex to HC3; direct chain analysis assigns inhibition to bikunin.
extracellular region (IEA, TAS, NAS)	MODIFY to extracellular space	Correct but overly broad for a soluble secreted plasma protein. All three duplicate terms receive the same action.
hyaluronan metabolic process	ACCEPT	Human HC3 is directly transferred covalently to HA and therefore participates in HA remodeling.
protein binding	MARK_AS_OVER_ANNOTATED	The HC3-bikunin/TSG-6 relationships are real but covalent glycosaminoglycan assembly and transfer are not usefully represented by generic protein binding.
platelet dense granule lumen	KEEP_AS_NON_CORE	Reactome's current participant set explicitly contains Q06033/ITIH3 in the platelet dense-granule-content set, but the underlying 1992 source is not gene-specific in the accessible record and the location is not a defining HC3 function.
extracellular exosome	KEEP_AS_NON_CORE	Retain the HDA detection; a secreted plasma protein can co-purify with vesicle preparations, so this is not a core location.
endopeptidase inhibitor activity	REMOVE	The cited cDNA paper does not establish intrinsic HC3 inhibition; independent direct biochemistry assigns inhibition to bikunin.
hyaluronic acid binding	NEW	Direct human HC3-fragment binding plus covalent human HC3-HA formation supports the term, with the processing caveat recorded.
extracellular matrix	NEW	Direct human HC3-HA formation and mouse HC3 recovery in physiological SHAP-HA matrices support this location conservatively by ISS.

Curation boundaries

HC3 is a transferred substrate/structural matrix component, not the TSG-6 heavy-chain transferase and not the HC2 cofactor that makes HC3 transfer possible.
The direct noncovalent HA-binding assay is conditional: intact mature human HC3 did not bind in that experiment, whereas a proteolytically generated N-terminal fragment did. The core description must retain this limitation.
The mouse SHAP-HA work supports a conserved matrix role, but mouse reproductive or inflammatory phenotypes are not sufficient for new human disease, fertility, or inflammatory-process annotations specific to ITIH3.
The Reactome record is target-specific at the curated participant level: the current R-HSA-481009 participant set explicitly lists UniProt Q06033/ITIH3. The only linked publication is PMID:1435334, whose cached record has no abstract and no independently checkable ITIH3 text; the localization is therefore retained as non-core rather than promoted into the core function.
Exosome proteomics establishes recovery from a preparation, not vesicle incorporation or an exosomal activity.

Open questions

Which endogenous proteases expose the N-terminal HC3 hyaluronan-binding site, and in which tissues or inflammatory states does that processing occur?
How does HC2 enable TSG-6-mediated transfer of HC3 from pre-alpha-inhibitor, and does HC2 remain associated with the transfer machinery in vivo?
Does HC3 confer matrix properties distinct from HC1- or HC2-containing HC-HA complexes?
Is ITIH3 genuinely stored within platelet dense granules, or does the Reactome participant assignment reflect adsorption or co-isolation of an abundant plasma protein?