CD86 (B7-2) curation notes

Deep research status

2026-10-05: just deep-research-falcon human CD86 was launched in the background.
The Edison/Falcon API returned repeated 429 Too Many Requests errors during the
session (shared rate limit with concurrent jobs). No perplexity key is available, so
no fallback was used. The review below was built from UniProt P42081, the cached
GOA-cited publications (publications/PMID_*.md, mostly abstract-only) and cached
Reactome entries. If a CD86-deep-research-falcon.md file appears later, it was
not used for this review. Final outcome: the falcon job exited without producing
an output file (Edison API 429 rate limiting); no deep-research file exists.
UPDATE 2026-10-05: the statement above is superseded. The falcon job did complete
late (end_time 01:34) and CD86-deep-research-falcon.md now exists; it has been
reconciled with the review (see "Falcon deep research reconciliation" below).

UniProt P42081 (CD86_HUMAN), "T-lymphocyte activation antigen CD86", alt names
B7-2, B70, CTLA-4 counter-receptor B7.2; 329 aa type I transmembrane protein of the
immunoglobulin superfamily (IgV + IgC extracellular domains, short cytoplasmic tail).
Six splice isoforms, including a soluble deltaTM form and deltaEC form; isoform 2
reported to interfere with CD86 clustering [file:human/CD86/CD86-uniprot.txt
"[Isoform 2]: Interferes with the formation of CD86 clusters, and thus acts as a
negative regulator of T-cell activation."].

Cloned as a second CD28/CTLA4 counter-receptor that costimulates T cells
PMID:7694363;
PMID:7694363.
ETC-1 (identical to B7-2) bound CD28-Ig and CTLA4-Ig and enhanced CD4 T cell
proliferation PMID:7513726.
CD80 and CD86 transfectants costimulate proliferation, IL-2 and IFN-gamma, and CTL
generation similarly PMID:7527824.
Biophysics: B7-2 binds both receptors more weakly than B7-1, is monomeric, and is less
biased toward CTLA4 PMID:12196291; PMID:12196291.
Crystal structure of CTLA4/B7-2 complex: alternating lattice of bivalent CTLA4 and
B7-2 dimers proposed as synapse organisation PMID:11279501.
IgC and intracellular domains needed for full co-signaling; CD86 monomeric in living
cells PMID:24845157.
Signalling differences vs CD80: CD86 does not induce detectable CD28 tyrosine
phosphorylation but shares CBL/VAV phosphorylation and NFAT activation
PMID:9915850.
CTLA4 acts as a decoy/competitive inhibitor for CD86 [Reactome:R-HSA-388808 "CTLA4
primarily affects CD28 costimulation by acting as a competitive inhibitor."].

Plasma membrane, single-pass type I; APC surface (B cells, monocytes, DCs)
[file:human/CD86/CD86-uniprot.txt "Expressed on the surfaces of antigen-presenting
cells."]. Also found in exosome proteomes (B-cell exosomes, plasma vesicles) – likely
passive cargo from APC membranes.
Regulated by MARCH8 (c-MIR) ubiquitination -> endocytosis/lysosomal degradation
(UniProt PTM).

UniProt (By similarity to mouse P42082): CD86 engagement on B cells regulates IgG1
production and activates NF-kappa-B via PLC/PKC. These ISS-based annotations
(B cell activation, Ig production, non-canonical NF-kB) are plausible but not core;
kept as non-core.

Receptor for adenovirus species B fiber (PMID:16920215, per UniProt); vaccinia M2
binds CD80/CD86 (PMID:30918073). Not annotated in GOA; not added.

Core MF: receptor ligand activity (GO:0048018) — CD28/CTLA4 ligand.
Core BP: T cell costimulation (GO:0031295), via MODIFY of the T cell activation IDA
(GO:0042110 is the wrong parent; GO:0031295 is_a positive regulation of T cell
activation and its definition is literally the provision of a surface-bound
ligand-receptor second signal). Comparator check: CD80 already carries GO:0031295.
protein binding IPI rows with CTLA4/CD28 -> MODIFY to signaling receptor binding.
coreceptor activity / signaling receptor activity -> MODIFY to receptor ligand
activity (CD86 is the ligand; CD28 is the (co)receptor).

CD86-deep-research-falcon.md completed after the review was written. It agrees with
the review's core picture (CD86 = APC-surface ligand for CD28 and CTLA4, not an enzyme;
signalling is downstream of CD28 in the T cell) and found no contradictions.
Material additions verified against primary literature and acted on:
MARCH1 (not only MARCH8/c-MIR, which UniProt cites) ubiquitinates CD86 via a
TMD recognition surface centred on Pro254 PMID:34157285. Description updated to
"MARCH-family E3 ligase (MARCH1 in APCs, also MARCH8)"; reference added.
CTLA4-mediated transendocytosis of CD86, with CTLA4 recycling after CD86 release
PMID:35999394. Reference
added; this is a CTLA4 process (CD86 is its input), so no new CD86 annotation.
Not acted on: soluble CD86 splice variant costimulation (Jeannin 2000; no PMID
resolved, already covered by an existing suggested question), 2023-2024 mouse/
translational studies (context-specific, no GO impact), mouse B-cell reverse
signalling (already kept as non-core ISS).
Deep-research file added to references and cited as support on the IBA receptor
ligand activity annotation. No annotation actions changed.