The canonical symbol is ANGPTL3 (HGNC:491; NCBI Gene:27329; UniProt:Q9Y5C1). NCBI lists ANL3, ANG-5, FHBL2 and ANGPT5 as aliases. Fresh GitHub queries at main e6c9e64f7791cfb14ea00f4e1e9301c5afd5afa0 found no canonical/alias gene directory or history, no task branch, and no matching open PR titles. NCBI Gene
One ordinary just fetch-gene human ANGPTL3 attempt failed resolving rest.uniprot.org, before producing UniProt, GOA or review YAML. The first normal Falcon invocation failed because uvx used a protected global tools directory. A changed-configuration attempt used the installed matching deep-research-client 0.2.7rc1, workspace UV paths, and the ordinary Falcon wrapper with perplexity-lite fallback; Falcon then failed before provider execution because the missing UniProt seed left template variables unset; the Perplexity-lite fallback failed DNS. Neither attempt produced a research report. These notes are manual research, not provider output. No annotation rows or source records have been manufactured.
The files under tmp/ANGPTL3-initial/ retain the actual command logs, bounded attempt receipts and preflight. At this initial checkpoint, annotation review, source identity preservation, independent consultation, core synthesis and validation awaited a genuine deterministic seed. The later imported seed and completed decisions are documented below. The five additional genes checked for seed retrieval are not being biologically reviewed here.
ANGPTL3 is a secreted hepatic regulator of extracellular lipid metabolism. Its N-terminal region inhibits lipases; association with ANGPTL8 is important for potent physiological inhibition of intravascular lipoprotein lipase (LPL). Endothelial lipase (EL) inhibition does not require ANGPTL8. The separate C-terminal fibrinogen-like domain can bind integrin alpha-v/beta-3 and elicit endothelial responses. The latter evidence should be retained in its experimental context rather than making every developmental or vascular phenotype a core molecular function. PMID:28413163(https://pubmed.ncbi.nlm.nih.gov/28413163/); PMID:41148215(https://insight.jci.org/articles/view/197827); PMID:11877390(https://pubmed.ncbi.nlm.nih.gov/11877390/)
Primary PMC Results and Methods are now accessible. Intact human ANGPTL3 and ANGPTL3/8 were expressed in HEK293 cells. Human LPL residues 1–448 carried R297A to remove furin cleavage and was produced in Drosophila S2 cells; bovine milk LPL was also used. Thus the inhibitor identity is human, but the substrate species/construct must still be matched to each assay. Mass photometry supports a 2:1 ANGPTL3:ANGPTL8 heterotrimer and an ANGPTL3 homotrimer. HDX-MS and thermal measurements support ATP-independent unfolding of LPL's catalytic domain; Figure 1 tests human LPL at nanomolar concentrations. GPIHBP1 and heparan sulfate protect LPL, especially against ANGPTL3 alone. Similar behavior at high concentrations does not erase differing low-concentration potency. This is not an ATP-dependent chaperone function or an experimentally solved AlphaFold structure; recombinant oligomer stoichiometry does not establish every circulating molecule's state. The furin-resistant LPL construct also helps distinguish unfolding from proteolytic cleavage in that assay. Primary article
The publisher Results, Discussion and Methods were read. The main double-alanine scan used mouse ANGPTL3 residues 17–221, expressed in HEK293T cells; human host cells do not make that a human-protein scan. Separate experiments examined endogenous ANGPTL3 from human liver-derived cells and normal human serum/plasma. Disrupting recombinant trimerization impaired EL inhibition. Serum fractionation, immunoassays and native gels identified abundant human monomeric ANGPTL3 alongside trimers and ANGPTL3/8; human recombinant L124A was also examined. The authors explicitly state that monomeric mutants were not tested in an animal model and population representation is uncertain. The proposed EL interface is an AlphaFold prediction. In vitro cleavage-resistant ANGPTL3 remained inhibitory, so proteolytic cleavage is not universally required for EL inhibition. Primary article
The unchanged local cache is abstract-only, despite a PMC link. Its abstract reports that ANGPTL8 alone is inactive, that coexpression with ANGPTL3 strongly enhances inhibition, and that an ANGPTL3 mutant lacking its own inhibitory activity can still activate ANGPTL8. The increase in circulating triglycerides was tested in mice. This supports a cooperative complex while leaving the detailed allocation of inhibitory contacts to the original constructs and the later biophysical work. Do not transfer every property of ANGPTL8 to ANGPTL3. PubMed
The unchanged local cache is abstract-only. GPIHBP1 protected LPL from ANGPTL3 and ANGPTL4 inhibition; heparin altered sensitivity differently. Double-knockout and antibody experiments were in mice. This is evidence for environmental modulation of an inhibitor, not evidence that ANGPTL3 lacks inhibitory activity. PubMed
The local cache is marked full-text, but its actual extracted body contains Abstract, Introduction and Discussion, without the Methods/Results sections. The authors report cell-based enhancement of LPL cleavage by furin and PACE4, but not PCSK5, with the ANGPTL3 N-terminal region sufficient. Their proposed physiological dominance of this indirect route is explicitly a hypothesis requiring further testing. Retain the observed substrate-specific assistance while considering the later unfolding mechanism. ANGPTL3 is not the protease that cleaves LPL. The data do not establish indiscriminate activation of all convertase substrates or all three convertases. PubMed
The unchanged local cache is abstract-only. Hepatocyte-specific knockout experiments and primary hepatocytes were from mice. Furin dominated intracellular processing and circulating cleaved ANGPTL3, while PACE4 contributed extracellular processing. ANGPTL3 is a substrate in these experiments; cleavage of ANGPTL3 does not by itself make ANGPTL3 a participant in proteolysis. PubMed
The primary PubMed abstract explicitly reports administration of human ANGPTL3 in a mouse study and increased plasma free fatty acids and glycerol, plus ANGPTL3 binding to adipose cells and stimulation of release from adipocytes. This is distinct from inhibition of vascular LPL. A positive lipid-catabolism annotation cannot be removed merely because the best-established endocrine function inhibits intravascular lipolysis. Precise assay cell/construct details and physiological scope require the paper's full text; the abstract does not establish a specific adipocyte receptor. PubMed
The primary indexed PubMed abstract reports binding of recombinant ANGPTL3 to alpha-v/beta-3 integrin, integrin-dependent endothelial adhesion/migration and downstream phosphorylation. A rat corneal assay showed angiogenesis; the C-terminal fibrinogen-like domain was sufficient. Co-immunoprecipitation did not detect Tie2 binding. Preserve the distinction between a direct receptor-binding/signaling activity and broad organismal angiogenesis. Do not turn absence of detectable Tie2 binding in this assay into an unlimited claim about all possible settings. PubMed
The primary indexed PubMed abstract reports association with both human HDL and LDL, with about three-quarters of lipoprotein-associated ANGPTL3 in HDL fractions of healthy volunteers. This is a fraction of the lipoprotein-associated pool, not necessarily all plasma ANGPTL3. A patient deficient in HDL had more LDL-associated protein. The functional assay used brown adipocyte T37i cells. Association with an extracellular particle should not be rewritten as localization inside the producing hepatocyte or constitutive membership of every HDL particle. PubMed
The primary indexed PubMed abstract describes three people with complete ANGPTL3 deficiency and seven controls. Tracer kinetics implicated both reduced liver-derived triglyceride-rich particle secretion and increased clearance. This supports a broader organismal phenotype; it does not identify a direct intracellular VLDL-assembly reaction performed by ANGPTL3. PubMed
The four actually read canonical records are PMIDs 19542565, 20581395, 23918928 and 28413163. Primary external content was substantively read for 11877390, 12565906, 37889183, 40112106, 40446802 and 41148215. Initial primary discovery/metadata also identified 10644446 (protein architecture), 17110602 (EL inhibition), 20942659 (familial hypolipidemia), 28385496/28538136 (human genetic/therapeutic context), 32646941 (EL-dependent remnant clearance), and 38462608 (HDL function); these are candidates to revisit after the seed, not fully adjudicated supporting references. The 17110602 primary PDF opened as a 20-page document but the attempted screenshots failed to expose usable text. No full-text reading is claimed from it.
Incidental local search hits 41384245, 25349778, 30559189, 39188988 and 27929370 were identified by title/cache metadata only, and are not evidence citations in the proposed review. In particular, a paper centered on ANGPTL4 must not substitute for direct ANGPTL3 evidence. No literal quote has yet been attached to a review row. Canonical source caching will be attempted once in a finite batch after the actual seed reveals all required references.
GO:0005178 integrin binding was checked in the official AmiGO record: molecular function, binding to an integrin, under signaling receptor binding, cell adhesion molecule binding and protein-containing complex binding. A QuickGO multi-term request failed; official pages for other candidate lipase-regulatory terms were not reliably accessible. Do not infer term obsoletion or manufacture a replacement ID from those access failures. All final term choices and redundancy checks remain pending. Official GO term
The local gocams/index.tsv search found no ANGPTL3/Q9Y5C1 match; that is only a result about this cached index. No new process term is proposed. Review every source annotation from the genuine YAML, preserve all original source objects/qualifiers, distinguish inhibition from catalysis and upstream necessity from process participation, and obtain the required independent annotation consultation before freezing the draft.
The ordinary hosted seed7 fetch produced the genuine Q9Y5C1 UniProt record, 55-row GOA and 55-annotation review seed. Root validated and imported the exact files. The baseline under tmp/ANGPTL3-initial/baseline/ records the seed SHA256 a872b5f9f6c658cbd1fa1f5f068fbbacc0f70c476a292f7d2edf5191ba8b7237; all 55 source objects, source references, qualifiers, identifiers and source lists remain unchanged. The raw UniProt record has no alternative-products section; proteolytic chains are not manufactured splice isoforms. Six absent normal seed auxiliaries were independently read, then strictly imported by root; the already present abstract-only PMID:19542565 cache was preserved despite the artifact containing a longer normal record.
With real UniProt inputs now available, a single further ordinary Falcon/fallback attempt used the installed client and workspace UV directories. Falcon reached authentication but a DNS connection failure culminated in Failed to authenticate; Perplexity-lite also failed DNS. Exit 1 after 6.752 seconds; no provider report was produced. The actual log/receipt are deep-research-seeded.log and deep-research-seeded-receipt.json under the working receipt directory. These manual notes never substitute for a provider-named file.
The independent annotation-reviewer consultation considered all 55 rows and primary scope. Current decisions are 29 ACCEPT, 8 KEEP_AS_NON_CORE, 10 MODIFY and 8 UNDECIDED. There are no new annotation rows and no removals. Molecular inhibition and integrin binding are distinguished from downstream readouts, while source-specific unresolved decisions preserve the original assertions. In particular, an unresolved donor chain does not by itself make an otherwise corroborated core lipid function uncertain. PAINT self-inclusion is legitimate descendant grounding, not circularity; no node placement was rejected from donor count.
The independent reviewer found the author-uploaded original JBC paper for PMID:11877390, DOI 10.1074/jbc.M109768200. I independently read its Methods and targeted Results/Discussion. Human ANGPTL3 was cloned into eukaryotic/baculovirus vectors; the human and mouse preparations and rat-cornea host are kept distinct. The FBN-domain experiment used an oligomerizing leucine-zipper fusion, so it is not an unmodified isolated native fragment. The direct binding assay used purified alpha-v/beta-3; human endothelial adhesion and haptotactic migration were positive. In contrast, the Results states that ANGPTL3 failed to elicit proliferation of HUVECs/HMVECs in 1% FCS, with the negative data not shown. This justifies refining this paper's growth-factor annotation to the already seeded integrin-binding activity, not inferring universal absence of growth effects in every setting. The normal cache remains abstract-only, and its attached quote supports the positive binding result rather than pretending the negative proliferation sentence is cached. Author-uploaded primary paper
PMID:17110602 was now read as a complete normal abstract: mouse Angptl3 deficiency and add-back affect HDL cholesterol/phospholipid; direct biochemical inhibition affects endothelial-lipase phospholipase activity; human plasma results are correlations. The construct species are not fully resolved from this abstract. The inhibitor role motivates regulation of phospholipid catabolism rather than assigning the hydrolysis reaction to ANGPTL3. PMID:17110602(https://pubmed.ncbi.nlm.nih.gov/17110602/)
The seed artifact also permitted actual Methods/Results/Discussion reading of PMID:19542565 without overwriting the existing canonical abstract. Human ANGPTL3 residues 17–221 were expressed in bacteria; experiments used human and bovine LPL plus soluble mouse GPIHBP1. DGGR was the fluorogenic lipase substrate. GPIHBP1/heparin protection was concentration dependent. The ELISA measures GPIHBP1–LPL binding, not ANGPTL3–LPL binding. This paper's inhibition measurements and mouse triglycerides do not become invented direct VLDL/chylomicron-particle remodeling or lipid-transport assays. The relevant regulatory functions remain supported by the broader established mechanism. Historical dimer descriptions are not adopted as current structural conclusions. PMID:19542565(https://pubmed.ncbi.nlm.nih.gov/19542565/)
PMID:23150577's ANGPTL3-relevant main Methods/Results/Discussion were read from the normal full cache. Human ANGPTL3/8 was expressed in mouse liver, and plasma complex formation and Angptl3 dependence were tested. The separate cleavage experiment used recombinant mouse ANGPTL3. Human ANGPTL3 residues 13–200 produced as an antigen must not be confused with the principal expressed experimental construct. The authors discuss a cleavage model but also other modes of cooperation; later direct biophysics does not require universal cleavage. Not every supplement was read. PMID:23150577(https://pubmed.ncbi.nlm.nih.gov/23150577/)
PMID:37634084 was read through the principal Results in PMC and the publisher article. Endogenous human knockdown/knockout and HepG2 co-immunoprecipitation are distinct from tagged ANGPTL3 imaging in HEK293T cells. The latter reports colocalization with ER, Golgi, late-endosome and multivesicular-body markers; the endosome marker is Rab7, not an early-endosomal marker. The accessible material did not fully resolve the overexpressed ANGPTL3 construct species, so it is not yet used to close the mouse-derived Golgi transfer. The study's intracellular IL1R1/IL1RAP-associated effects do not justify making every intracellular compartment or inflammatory readout a new core function. Publisher Methods fetches redirected to inaccessible content. Primary paper
Official GO definitions/parents were checked for GO:0005178, GO:0008083, GO:0055102, GO:0004859, GO:0019915, GO:0060696 and GO:1903726. Phospholipase inhibitor activity is a descendant of lipase inhibitor activity. The single lipid-regulatory core uses the broader lipase function to include LPL and endothelial lipase while retaining the specific evidence rows. It lists parallel triglyceride, cholesterol and phospholipid homeostasis, without repeating broad lipid/acylglycerol ancestors. Regulation of phospholipid catabolism is the common replacement for the metabolism/catabolism overextensions, with negative direction in prose. Integrin binding and endothelial responses remain non-core contextual biology, not discarded findings. No NEW process is proposed. The local GO-CAM index had no matching ANGPTL3 entry, and no absent term was inferred from pathway incompleteness.
Eight unresolved rows cover early endosome, Golgi, hormone response, lipid storage, lamellipodium, extracellular matrix and two artery-morphogenesis source rows. Source tracing distinguished mouse Q9R182, rat F7FHP0, Ensembl identifiers, UniProt subcellular mappings and PAINT identifiers, but did not recover all original donor experiments or the PAINT tree/MSA. Their uncertainty is source-specific; secretion alone was never used to refute a surface or intracellular pool.
The five newly consulted primary records still needing normal cache closure are PMIDs 37634084, 37889183, 40112106, 40446802 and 41148215. One finite ordinary fetch-pmid batch requested exactly those five into a temporary output directory, timed out after 45 seconds before producing records, and made no shared-cache writes. Root has the receipt for hosted source recovery. This draft remains pending exact-source attachment and final validation; every actually consulted paper will be represented in the final census. Earlier bibliographic/search-only candidates remain explicitly unconsulted, not silently counted as evidence.
The five normal source records were recovered, strictly verified, independently assessed from the actual archive and matched byte-for-byte to the staged copies before root imported them. The existing PMID:19542565 abstract was not replaced. All 13 substantively consulted PMIDs now have normal canonical caches and explicit reference assessments. The five added records contain substantial original article text; the scopes below describe what was actually read, rather than treating a full-text flag as proof that every supplement was examined.
For PMID:37634084, the actual interaction/localization Results, Figure 4/5 legends and construct/confocal Methods resolve part of the earlier access gap. Figure 5 labels the shared ANGPTL3 construct series human, and human-sequence shRNA/sgRNA experiments also suppress the ANGPTL3-Flag expression plasmid. Taken together, these support interpreting the tagged Golgi experiment as using human ANGPTL3. This is an inference from the paper's internally connected construct descriptions, not from the human host alone. Row 11 is therefore KEEP_AS_NON_CORE for a transient tagged secretory-pathway Golgi pool; endogenous Golgi residence is not established. The original mouse/Ensembl donor chain remains unreconstructed. Row 10 remains UNDECIDED because Rab7 marks late rather than early endosomes. Supplement image pixels were not independently inspected. A second reviewer independently read these passages and agreed with this bounded change. Primary paper
PMID:40112106 now directly supports row 31: HDX-MS maps the proximal ANGPTL3 helix to the lipase interface. The actual protein Methods distinguish human ANGPTL3/3–8 and human R297A LPL activity experiments from bovine LPL interface measurements. The original PMID:19542565 binding ELISA remains correctly identified as GPIHBP1–LPL. The newer unfolding evidence enriches the inhibitor mechanism without manufacturing ATPase or chaperone activity. The relevant Results, legends and protein Methods were read; not every supplemental dataset was inspected. Primary paper
PMID:37889183's actual human fractionation/binding Results and Methods confirm HDL/LDL association. Its activity experiments combine recombinant human ANGPTL3 and ANGPTL8 in T37i cells, which should not be described as an ANGPTL3-only human physiological assay. PMID:40446802's actual participant/kinetic Methods and targeted Results/Discussion confirm three lifelong-deficient participants versus seven controls and distinguish measured clearance/secretion from the explicitly unknown hepatic mechanism. These later papers corroborate the physiological synthesis without pretending that the original 19542565 study performed particle-turnover assays. PMID:37889183(https://pubmed.ncbi.nlm.nih.gov/37889183/); PMID:40446802(https://pubmed.ncbi.nlm.nih.gov/40446802/)
PMID:41148215's actual construct and assay Methods confirm that its extensive alanine scan is mouse ANGPTL3 17–221, with human EL as the inhibited enzyme. Human plasma, Huh7 secretion and human L124A spike-in experiments are separately identified. The mutant-inhibition, oligomerization and human-plasma Results were read, but not every supplement. Its oligomer-dependent inhibition supports the mechanism with those species and construct limits preserved. Primary paper
The final decisions are 29 ACCEPT, 9 KEEP_AS_NON_CORE, 10 MODIFY and 7 UNDECIDED, with no NEW or removed rows and one molecular-function core. All 55 original source objects and both raw seed files remain unchanged. The remaining uncertainties concern early endosome, hormone response, lipid storage, lamellipodium, extracellular matrix and two artery-morphogenesis source rows. COMPLETE records that every annotation was assessed and source caching is closed; it does not imply resolution of those scientific uncertainties.
Final validation, rendering and history validation passed. Strict GOA tuple checking also passed; all eight attached snippets match their canonical sources, all 66 supporting-cache availability flags match the records, and the four pre-existing publication files retain their original Git object hashes. The final source census contains 13 PMIDs and two Reactome records with no missing normal caches. The append-only history record is history/genes/human/ANGPTL3/2026-09-27T213151Z-codex-4334d1.yaml; independent consultation and byte-level source receipts are retained under tmp/ANGPTL3-initial/.
The review correctly identified a traceability problem: the normal caches for PMID:19542565 and PMID:11877390 contain abstracts, while some rationales use independently read original-paper body passages. These are two different access scopes. The immutable normal publication records are preserved under repository instructions, including the abstract-only flags; no source text has been edited or replaced. The source-specific conclusions below rely on the original experiments, not on silence in the abstracts.
The original JBC article is available as an author-uploaded full paper; its header DOI is 10.1074/jbc.M109768200. In Results, immediately before “Analysis of Signal Transduction Pathways Engaged by ANGPTL3 in Endothelial Cells”, the paper states: “ANGPTL3 failed to induce proliferation of primary human endothelial cells”. The following clause identifies HUVECs and HMVECs cultured with 1% FCS and marks the result as data not shown. The nearby positive assays measure adhesion and migration, and Figure 5C tests recombinant human ANGPTL3 binding to integrin alpha-v/beta-3. These are distinct from cell growth or proliferation. The Methods/targeted Results/Discussion were read previously; this exact negative passage, the article header and positive binding legend were reopened for the follow-up. No missing image is claimed to have been inspected.
The original endothelial proliferation result is negative under the stated culture conditions, while direct integrin binding is positive.
The growth-factor refinement therefore remains source-specific: it does not deny all possible growth-related effects of ANGPTL3, and it is not inferred merely because the abstract omits a proliferation claim. The attached PMID snippet reports the positive binding observation; the separate notes attachment records the externally inspected result and its locator.
The official PMC original article and publisher version provide public locators independent of the unpublished auxiliary extraction. The indexed original body was rechecked at “Production of recombinant human ANGPTL3 and ANGPTL4”, “Properties of recombinant soluble mouse GPIHBP1”, Figure1, and the LPL-stabilization Results/Figure2. The human inhibitor construct contains residues 17–221; the fluorescent substrate is DGGR. Figure1 measures soluble mouse GPIHBP1 binding to bovine LPL. The separate kinetic assay uses recombinant human LPL. Those statements describe the experiments actually read; they are not additional claims about human lipoprotein-particle remodeling. The existing abstract itself establishes LPL inhibition and mouse serum-triglyceride responses.
The original Figure1 binding assay tests GPIHBP1 with LPL; the later PMID:40112106 experiments independently establish an ANGPTL3–LPL interface.
The concise curated rationales now use the cache-supported inhibitory result and explicitly distinguish biochemical activity from the physiological interpretation. Detailed construct and reagent boundaries remain here, linked to the original sections. The later human kinetic paper supports physiological VLDL/chylomicron/triglyceride handling; it does not turn the earlier enzyme assay into a particle assay. The original source assertions remain unchanged.
is_active_in assertion. No donor-count or domain-only criticism is substituted for a node analysis.The follow-up changes evidence traceability and explanation, with all 55 original source objects, action decisions, immutable raw records and normal source caches preserved. No new biological-process or molecular-function annotation is introduced.