Batch 1 of the INNATE_IMMUNITY project (Toll/TLR axis). Reviewed immediately after
TLR7 (Q9NYK1), with which TLR8 shares its PAINT node, most of its electronic
annotations, and several references.
Deep research: TLR8-deep-research-falcon.md (Edison/Falcon) was produced and
read. Its citation keys are internal document handles rather than PMIDs, so it is
used here only for orientation; every claim below is anchored to a cached
publication in publications/.
Endosomal single-pass type I membrane pattern-recognition receptor of the same
subfamily as TLR7 and TLR9: LRR ectodomain, one transmembrane helix, cytoplasmic
TIR domain. It is the most highly expressed nucleic-acid-sensing TLR in the human
myeloid compartment.
Ligands: uridine plus a short oligonucleotide, at two sites.
- PMID:25599397
- PMID:25599397
- PMID:25599397
Both ligands are generated in situ by RNase T2:
- PMID:31778653
- PMID:31778653
Note the contrast with TLR7, which matters for the ligand-specific pathway terms:
TLR7's site 1 takes guanosine, TLR8's takes uridine, and TLR7 has a second
site absent from TLR8 PMID:27742543. The two receptors nonetheless share
synthetic agonists (R-848) PMID:12032557.
Viral ssRNA.
- PMID:14976262
- PMID:33718825
Activation mechanism: Z-loop cleavage then ligand-induced dimer reorganisation.
- PMID:26929371
- PMID:23520111
- PMID:25297876
Localisation: early endosome and ER — explicitly not late endosome, lysosome or
Golgi. This is the sharpest difference from TLR7/TLR9 and the main reason several
location rows were demoted.
- PMID:22164301
- PMID:22164301
- PMID:22164301
- PMID:22164301
Disease. X-linked gain-of-function variants cause IMD98 (neutropenia,
lymphoproliferation, autoinflammation, bone-marrow failure), mostly as somatic
mosaicism.
- PMID:33512449
Localisation rows contradicted by PMID:22164301. The two GO:0000139 Golgi
membrane Reactome TAS rows and the two GO:0036020 endolysosome membrane rows
(ARBA IEA + Reactome TAS) assert compartments in which endogenous TLR8 was
explicitly not found in human monocytes. They are marked
MARK_AS_OVER_ANNOTATED, with GO:0031901 early endosome membrane proposed for
the endolysosome rows. GO:0010008 endosome membrane (EXP, IEA, two Reactome
rows) and the ER membrane Reactome rows are accepted.
GO:0009897 external side of plasma membrane (IDA, PMID:24942581) is
contradicted by the paper it cites. The cached record is abstract-only, so I
retrieved the open PMC full text (PMC4136363). TLR8 is assayed there — 26
mentions — which is exactly why the row cannot be dismissed as a wrong
identifier. But the measurement is explicitly intracellular: "Intracellular TLR8
staining was performed using the Cytofix/Cytoperm Fixation/Permeabilization
Solution Kit", and the discussion speaks of "the increase in expression of
pattern recognition receptors such as intracellular TLR8". The same paper's two
other TLR8 rows (GO:0034158, GO:0003727) are supported by its ssRNA40
stimulation experiments and are accepted; the surface-location row is REMOVEd.
GO:0003725 double-stranded RNA binding, two IDA rows. PMID:33718825 used
GU-rich single-stranded RNA throughout (its abstract says so), and PMID:16111635
is the nucleoside-modification study of RNA sensing. TLR8's structurally defined
sites bind uridine and a short oligonucleotide PMID:25599397. Both rows MODIFY
to GO:0003727 single-stranded RNA binding, which TLR8 already carries.
GO:0003677 DNA binding (IDA, PMID:16123302). The Treg-reversal paper used
"synthetic and natural ligands for human TLR8"; DNA sensing is TLR9's function,
and the TLR8 ligand-binding sites are defined for uridine and short RNA
oligonucleotides. Cached record is abstract-only, so the row is not removed on
an unverifiable mis-attribution claim; it is marked MARK_AS_OVER_ANNOTATED.
GO:0003723 RNA binding (TAS, same paper) MODIFYs to the informative
GO:0003727.
GO:0016064 immunoglobulin mediated immune response (TAS, PMID:16123302).
The cited paper concerns reversal of CD4+ regulatory T cell suppression; nothing
in it bears on immunoglobulin-mediated immunity, and TAS is not experimental
evidence to defer to. REMOVE.
Type II interferon. As for TLR7, the IDA PMID:16286015 row rests on a paper
measuring IFN-alpha/beta and IFN-lambda only, so it MODIFYs to GO:0034346
positive regulation of type III interferon production (the decision already
taken in the TLR9 review). Unlike TLR7 — where PMID:32706371 shows IFN-gamma
loss in TLR7-deficient patients — there is no TLR8-specific IFN-gamma evidence
in the reviewed literature, so the ARBA IEA row is MARK_AS_OVER_ANNOTATED,
matching TLR9.
GO:0071260 cellular response to mechanical stimulus (IEP, PMID:19593445).
Same wrong-reference problem as TLR7 and TLR3: the identifier resolves to
"Expression of the Bcl-2 protein BAD promotes prostate cancer growth", whose
cached full text has no occurrence of "TLR" or "toll". UNDECIDED, with
reference_review.correctness: WRONG_IDENTIFIER.
GO:0005515 protein binding, two IPI rows (UNC93B1). Uninformative. TLR8 is
the client of UNC93B1 and the biology is trafficking; the BioPlex row
(PMID:33961781) is a proteome-scale interactome screen with no functional
assignment. REMOVE both, mirroring the TLR3/TLR7 handling. Removal does not
dispute either interaction — PMID:22164301 establishes the UNC93B1 association
directly.
Plasma membrane IBA. The is_active_in GO:0005886 row sits at the TLR7/8/9
node. For TLR8 the endosomal restriction is not merely where it signals but a
prerequisite: the Z-loop must be cleaved by endosomal proteases before the
receptor can dimerise at all [PMID:26929371, PMID:25297876]. KEEP_AS_NON_CORE,
with the node-level question raised.
GO:0038023 signaling receptor activity and GO:0045087 (NAS,
PMID:12032557). Author-statement rows from the R-848 correspondence. Both are
true but generic; the receptor-activity row MODIFYs to GO:0038187 pattern
recognition receptor activity, which TLR8 already carries by IBA and which is
what the structures establish.
The Falcon report had already been read for orientation when the review was
written; this pass re-compared it against each decision.
referencesGO:0038187 IBA row.