The plasma membrane - the one compartment this graph targets with drugs and had never written down
also called plasma membrane · cell surface · surface antigen compartment · surfaceome
Three of the things doctors actually aim at in this leukaemia - CD19, CLEC12A and NG2 - are proteins sitting on the outside of the cell, and a drug can only reach them there. This entry is the graph's record of that compartment, because until now all 46 localization entries described places inside the cell.
The same thing, for a scientist
Of the 46 localization-* entries in this graph before 2026-09-20, none described the plasma membrane, although every antibody, BiTE and CAR modality the clinical beats discuss requires the target epitope to be present on the outer leaflet of an intact cell. The compartment is therapeutically load-bearing in a way no nuclear compartment is: a nuclear protein only has to be inhibitable, a surface protein has to be topologically accessible. Three genes in this graph are read out in this compartment by unpermeabilised flow cytometry - CD19 (P15391, type I single-pass), CLEC12A (Q5QGZ9, type II single-pass) and CSPG4/NG2 (Q6UVK1, type I single-pass proteoglycan) - and for CD19 the documented failures of therapy include failures of localisation rather than of expression.
Why a surface compartment is not just another compartment
Unpermeabilised flow cytometry IS a localisation assay, and it is the commonest one in this graph
What this compartment does NOT cover
| compartment | plasma membrane (outer leaflet, accessible to an antibody on an intact cell) |
|---|---|
| assay that defines it | flow cytometry or imaging on UNPERMEABILISED cells; a permeabilised stain, a lysate blot or a transcript measurement does not bound this compartment |
| gap this entry closes | 46 localization-* entries existed on 2026-09-20 and every one described an intracellular or intranuclear compartment; the compartment every antibody-based therapy in this graph acts on had no entry |
| created because | the Q62 localization tranche needed an edge target for three surface antigens and found none |
genes in this graph localised here
[ "gene-cd19", "gene-clec12a", "gene-cspg4" ]
topology
{
"CD19": "type I single-pass, P15391",
"CLEC12A": "type II single-pass, Q5QGZ9",
"CSPG4": "type I single-pass proteoglycan, Q6UVK1"
}- Delivery, not potency, is the failure mode in this compartment. Both documented CD19 escape mechanisms leave the protein expressed and merely not where the drug is, so a resistance screen scored on transcript or on total protein will call the cell sensitive.
- Shed ectodomain as a sink: if CSPG4/NG2 is shed in leukaemia at any appreciable rate, circulating antigen will bind an anti-NG2 antibody or CAR before it reaches a blast. This has not been measured in KMT2A-rearranged B-ALL and would be a cheap plasma ELISA.
- Essential-function toxicity: CD19 is on every normal B cell and CLEC12A is on normal granulocytes and monocytes, so the on-target off-tumour cost of hitting this compartment is B-cell aplasia and myeloid toxicity respectively. CLEC12A's claimed sparing of normal haematopoietic stem cells is the entire basis for targeting it and is a claim about the surface compartment specifically.
- Antigen density is a surface property, not an expression property. A CAR needs a threshold number of epitopes per cell, and low surface density with high transcript is the configuration that produces early relapse - invisible to every non-surface assay.
- Has anyone measured the surface-to-total ratio of CD19 in KMT2A-rearranged infant blasts at diagnosis, before any antigen-directed pressure? If a fraction of CD19 is already intracellular in this subtype the baseline would predict escape risk, and no such measurement was located for this beat.
- Does menin inhibition change the surface density of any of these three antigens? CSPG4/NG2 is a direct KMT2A-fusion target gene, so a fusion-targeted drug should move it, and the combination of a menin inhibitor with an anti-NG2 or anti-CD19 modality is untested against that prediction.
- What fraction of NG2 scored positive by diagnostic flow in KMT2A-rearranged B-ALL is membrane-anchored versus shed, and does a shed fraction absorb a therapeutic antibody?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a patient with CD19-negative relapse after blinatumomab for acute lymphoblastic leukaemia, the molecular basis of escape was a disrupted CD19 membrane export in the post-endoplasmic-reticulum compartment rather than loss of the CD19 gene; the relapse immunophenotype was otherwise identical to diagnosis.
Resistance to anti-CD19/CD3 BiTE in acute lymphoblastic leukemia may be mediated by disrupted CD19 membrane trafficking. · Braig F, Brandt A, Goebeler M, Tony HP, Kurze AK, Nollau P, Bumm T, Böttcher S, Bargou RC, Binder M · Blood 2017What this evidence cannot show
Four relapse cases, with the trafficking mechanism worked up in ONE of them; this is a single-patient molecular result, not a frequency. The paper's own title says 'may be mediated by'. Lineage and fusion status of the index case are not resolvable from the abstract, so this must NOT be read as a KMT2A-rearranged result.
Evidence that argues against this (1)
- The competing and better-replicated mechanism is genetic rather than purely post-translational: acquired frameshift and missense mutations in CD19 together with alternative splicing that skips exon 2 are selected under CART-19 pressure. A localisation explanation and a splicing explanation are not exclusive - the exon-2 products are also mislocalised - but an entry citing only the trafficking paper overstates how much of CD19 escape is a pure trafficking defect. Sotillo E et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
CD19 alterations emerging after CD19-directed immunotherapy cause retention of the misfolded CD19 protein in the endoplasmic reticulum, so the escape variant is expressed but not surface-localised.
CD19 Alterations Emerging after CD19-Directed Immunotherapy Cause Retention of the Misfolded Protein in the Endoplasmic Reticulum. · Bagashev A, Sotillo E, Tang CH, Black KL, Perazzelli J, Seeholzer SH, Argon Y, Barrett DM, Grupp SA, Hu CC, Thomas-Tikhonenko A · Mol Cell Biol 2018What this evidence cannot show
This record carries a published correction (PMID 35976030, Mol Cell Biol 2022). The correction notice was seen in the PubMed listing for this beat but its content was NOT read, so it should be checked before this item supports a quantitative claim. The system is CD19-directed immunotherapy in B-ALL generally, not specifically KMT2A-rearranged infant disease.
Evidence that argues against this (1)
- Not every CD19-negative relapse is a trafficking or splicing event. Lineage switch to a myeloid phenotype - disproportionately a KMT2A-rearranged phenomenon - removes the antigen by removing the B-cell programme, and no localisation model applies. The Braig series argues against lineage switch in its own four cases precisely because the rest of the immunophenotype was unchanged, which is the control a lineage-switch case would fail. Braig F et al. 2017
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
NG2/CSPG4 is shed from the cell surface by MMP-13, so the proteoglycan exists as a membrane-anchored and a released form, and shedding attenuates anoikis.
Shedding of NG2 by MMP-13 attenuates anoikis. · Joo NE, Miao D, Bermúdez M, Stallcup WB, Kapila YL · DNA Cell Biol 2014What this evidence cannot show
Not a haematopoietic system. The relevance here is methodological rather than biological: it establishes that an anti-NG2 reagent can encounter non-membrane NG2, which bounds what a surface stain proves.
Evidence that argues against this (1)
- In KMT2A-rearranged B-ALL specifically the functional readouts are consistent with membrane-anchored NG2 - NG2 is described there as the membrane-bound proteoglycan and acts through an ITGA4 partnership that requires co-residence in the membrane - so the shed pool, whatever its size, is not obviously what carries the leukaemic phenotype in this disease. Rubio-Gayarre A et al. 2026
extends the graph's plasma-membrane antigen inventory from a curated list to the whole surfaceome
CD19 is one of the three surface antigens this graph targets in this compartment
- related to →CD19 - the molecule this disease is treated by, named in 114 entries and written in none of them until now
CD19 escape after blinatumomab and CAR-T is documented as a failure of localisation, not of expression
CLEC12A is one of the three surface antigens this graph targets in this compartment
NG2/CSPG4 is the one surface antigen in this tranche whose plasma-membrane localisation is measured in KMT2A-rearranged blasts
diagnostic NG2 flow cytometry is an unpermeabilised surface-localisation assay in the index disease
CD34 is a single-pass type I membrane glycoprotein - the compartment's defining case in this graph, and the only entry in this localisation tranche for which the subnuclear question is NOT APPLICABLE rather than unpublished
THE ENZYME IS NOT ITS PRODUCT: FUT4 is a single-pass type II Golgi stack membrane protein (UniProt P22083), while the Lewis-x/CD15 epitope it helps build is displayed at the plasma membrane, which is where flow cytometry and any antibody therapeutic act - two compartments, and conflating them is the localisation form of the rule-69 'FUT4 is CD15' error
NEGATIVE edge: DNTT is NUCLEAR (UniProt P04053, ECO:0000269) and is not a surface antigen at all; it appears beside the surface markers only because it is read on the same diagnostic flow panel, which requires permeabilisation
KRAS is the one protein in this tranche with no nuclear pool at all - lipid-anchored at the plasma membrane, so the subnuclear question is rejected rather than left empty
type I membrane protein of the secretory pathway; plasma membrane main, Golgi and ER en route
lipid-anchored at the plasma membrane and shuttling to the Golgi; the palmitoylation cycle, not prenylation alone, sets the address
B2M is the invariant light chain of the surface class I complex; its compartment is the secretory route and it has no nuclear pool, so subnuclear localisation is NOT APPLICABLE (Q87 2026-09-21)
PROM1/CD133 is a five-transmembrane apical membrane glycoprotein concentrated in membrane protrusions; subnuclear localisation is NOT APPLICABLE (Q87 2026-09-21)
CD33/Siglec-3 is a single-pass type I surface receptor (signal 1-17, TM 260-282), so subnuclear localisation is NOT APPLICABLE - but the exon-2-skipped CD33m isoform is PEROXISOMAL, not surface, and that is where the GO epitope goes (PMID 28747436) (Q89 2026-09-21)
NOT APPLICABLE for the subnuclear question, by TOPOLOGY: signal peptide 1-24 and a single helical transmembrane segment 310-333 of 362 residues commit HLA-B to the secretory pathway and the plasma membrane. Category exclusion, kept OUT of the denominator of undetermined subnuclear questions (Q93 2026-09-21)
NOT APPLICABLE for a subnuclear answer, by TOPOLOGY: type II signal-anchor transmembrane cytokine, TM 36-56, cytoplasmic 1-35, extracellular 57-233, shed as a soluble secreted form. The one escape route was checked and rejected - SPPL2a/SPPL2b intramembrane proteolysis in activated dendritic cells releases a fragment that triggers IL-12 from the cytosolic side, with no nuclear localisation and no DNA-binding domain (Q95 2026-09-22)