Of every antigen this project names, only NG2, LAMP5 and CD123 are present on all three KMT2A::AFF1 cell lines - CD19 is at 0.2 nTPM in MV4-11 and CD33 at 2.3 in RS4;11
also called the three lines share no lineage antigen
This leukaemia can appear as a B-cell disease or a myeloid one, and the three laboratory lines that carry its fusion gene span both. No lineage marker is present on all three: CD19 is absent from one, CD33 from another. The only surface proteins present on all three are the two that the fusion gene itself switches on.
The same thing, for a scientist
The three Cellosaurus-confirmed KMT2A::AFF1 lines are SEM and RS4;11 (B-ALL) and MV4-11 (monocytic AML with t(4;11)). Measured in one atlas at one time, HPA nTPM [E1]: | antigen | SEM | RS4;11 | MV4-11 | on all three above 10 nTPM? | |---|---:|---:|---:|---| | **CSPG4 / NG2** | 24.2 | 37.6 | 13.9 | **yes** | | **LAMP5** | 120.1 | 284.4 | 372.4 | **yes** | | **IL3RA / CD123** | 23.6 | 20.3 | 67.4 | **yes** | | CD19 | 229.6 | 400.9 | **0.2** | no | | CD22 | 108.8 | 156.8 | **0.2** | no | | CD33 | 14.7 | **2.3** | 146.9 | no | | CLEC12A | **0.2** | **0.1** | 7.3 | no | | CD38 | **0.8** | 15.2 | 7.8 | no | | PROM1 / CD133 | 280.0 | 25.6 | **0.2** | no | | MME / CD10 | **0.2** | **0.2** | **0.1** | no - absent from all three | **The only antigens shared across the fusion's full lineage range are the two the fusion drives, plus CD123.** Every lineage or maturation antigen in this project's inventory fails on at least one line. This is the quantitative form of a claim the corpus has made in prose - that KMT2A-rearranged leukaemia is treated with antigens it can abandon by changing lineage - and it is now a table rather than an argument. The CD10 row is the internal control: **MME is at 0.1-0.2 nTPM in all three**, which is the CD10-negative pro-B phenotype that defines KMT2A-rearranged infant ALL and that this graph already records. An atlas that recovers that without being asked is behaving.
Why this is a statement about the disease and not about three cell lines
A sceptical reading is that SEM, RS4;11 and MV4-11 simply are different cell types, so of course their lineage antigens differ - which is true and is exactly the point. KMT2A::AFF1 causes disease across that same range in patients: B-ALL, mixed-phenotype acute leukaemia, and lineage switch from lymphoid to myeloid under CD19-directed pressure. The three lines are not an artificial spread; they are a model of the range the fusion actually produces.
An antigen that a therapy relies on must therefore either be present across the range, or the therapy must accept that escape by lineage switch is available. CD19, CD22, CD33 and CLEC12A are all in the second category by this measurement 12.
What CD123 is doing in the shared column, and why it does not solve the problem
IL3RA/CD123 is present on all three lines (23.6, 20.3, 67.4 nTPM) - it is the one lineage-independent antigen in the inventory, which is consistent with its use across AML and B-ALL. It is nonetheless not fusion-enriched (0.46x the KMT2A-other-partner arm) and it is killed by the normal-tissue filter, at 26.8 nTPM in adipose tissue 1. CD123 is a stem-and-progenitor antigen, and this graph already holds the corresponding myeloid-ablation risk for CLEC12A.
So the shared column contains one antigen that is broad but unsafe and two that are fusion-driven and unsafe. There is no cell in the table that is both.
Limits
Three cell lines, one atlas, one value per line, no replicates, none infant-derived. Expression is not surface presentation. Every expression number in this entry is mRNA. It says nothing about how much protein reaches the membrane, whether the epitope is accessible to an antibody, or whether the protein is shed. A cell line is not a patient, and the Human Protein Atlas consensus tissue panel is adult tissue while the population of concern is infants. A 0.2 nTPM value is at the floor of the assay and means 'not detected', not 'measured to be absent'. Most importantly, a cell line's immunophenotype drifts in culture, and MV4-11's CD19 absence is a property of a line maintained for decades, not a measurement on a patient with t(4;11) AML.
| script | analyses/kmt2a-aff1-surface-target-screen/run.py |
|---|---|
| status | run |
| internal control | MME/CD10 at 0.1-0.2 nTPM in all three, recovering the CD10-negative pro-B phenotype without being asked |
| result summary | No lineage or maturation antigen in this project's inventory is present on all three KMT2A::AFF1 lines; the only shared antigens are the two fusion-driven ones and CD123. |
per line nTPM — the numbers behind the chart
{
"CSPG4": {
"SEM": 24.2,
"RS4;11": 37.6,
"MV4-11": 13.9
},
"LAMP5": {
"SEM": 120.1,
"RS4;11": 284.4,
"MV4-11": 372.4
},
"IL3RA": {
"SEM": 23.6,
"RS4;11": 20.3,
"MV4-11": 67.4
},
"CD19": {
"SEM": 229.6,
"RS4;11": 400.9,
"MV4-11": 0.2
},
"CD22": {
"SEM": 108.8,
"RS4;11": 156.8,
"MV4-11": 0.2
},
"CD33": {
"SEM": 14.7,
"RS4;11": 2.3,
"MV4-11": 146.9
},
"CLEC12A": {
"SEM": 0.2,
"RS4;11": 0.1,
"MV4-11": 7.3
},
"CD38": {
"SEM": 0.8,
"RS4;11": 15.2,
"MV4-11": 7.8
},
"PROM1": {
"SEM": 280,
"RS4;11": 25.6,
"MV4-11": 0.2
},
"MME": {
"SEM": 0.2,
"RS4;11": 0.2,
"MV4-11": 0.1
}
}shared above 10 nTPM
[ "CSPG4", "LAMP5", "IL3RA" ]
- A therapy chosen because an antigen is present on all three lines would choose NG2, LAMP5 or CD123 - and all three are killed by the normal-tissue filter. Breadth across lineage and safety across tissue pull in opposite directions in this table.
- MV4-11's CD19 of 0.2 nTPM is the strongest single row here, and it is one cell line that has been in culture since 1988. If it has drifted, the central column of this table is wrong.
- Presenting fusion-driven antigens as escape-proof invites a trial design with no antigen-loss monitoring. NG2 is bimodal in patients and absent from candidate stem-cell populations, so escape routes exist for it too - they are just different ones.
- Does the same table hold in primary KMT2A::AFF1 blasts spanning B-ALL, MPAL and post-switch myeloid disease? Flow data for all three states exists in diagnostic archives and has never been assembled as one antigen-by-lineage-state matrix.
- Is NG2 retained through a lineage switch in a model? That is the direct test of whether a fusion-driven antigen is escape-resistant, and this graph's analogy entry names it as the experiment that would settle the principle.
- CD123 is present on all three lines and is the only lineage-independent antigen here. Is its normal-tissue profile genuinely disqualifying, or is 26.8 nTPM in adipose tissue a tolerable cost that the 5 nTPM threshold has over-punished?
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
antigen-by-lineage matrix across the fusion's three cell lines
the quantitative form of that entry's lineage-escape argument, and the correction that two fusion-driven antigens do exist
CD123 is the only lineage-independent antigen in the inventory present on all three lines, and it fails the tissue filter at 26.8 nTPM in adipose
CLEC12A is 0.2 and 0.1 nTPM in the two B-ALL index lines
the lineage range the three lines span is the range this fusion produces in patients