USP1: an inhibitor, activity in B-ALL, and zero papers in the disease we care about
also called USP1 · ubiquitin-specific protease 1 · USP1 inhibitor · SJB3-019A · pimozide
The one part of the repair module our scan implicated that a drug can actually reach is USP1. Drugs against it exist, and killing leukaemia cells with them has been published. Nobody has ever tried it in the leukaemia this project is about.
The same thing, for a scientist
The FA enrichment test located the KMT2A-rearranged selectivity signal in the ID2 monoubiquitination module, where **USP1 (rank 129 of 17,728) is the one conventionally druggable member**. Small-molecule USP1 inhibitors exist with high-nanomolar IC50, USP1 is overexpressed in B-ALL patient marrow, and both knockdown and pharmacological inhibition kill B-ALL cells. **`USP1[tiab] AND (KMT2A[tiab] OR MLL[tiab])` returns zero records.** The complication is the substrate: the published leukaemia mechanism runs through **ID1, not FANCD2**.
The substrate problem, which cuts against the reason we came here
USP1 has two well-known substrate axes: it deubiquitinates FANCD2/FANCI, reversing the FA monoubiquitination step, and it deubiquitinates ID1, rescuing it from proteasomal degradation.
This project arrived at USP1 through the FA axis — it sits in the ID2 ubiquitination module that carries the enrichment signal. But every published leukaemia result for USP1 runs through ID1: inhibition downregulates ID1 and phospho-AKT in B-ALL, and over-expressing ID1 reverses the effect; the 2013 inhibitor paper is titled for ID1 degradation.
So USP1's high selectivity rank in a KMT2A-rearranged panel may not be an FA signal at all. If it is an ID1 effect, USP1's presence in the FA gene set is a coincidence of annotation — the same protein in two pathways — and using it as evidence for the FA module would be circular. That is a real weakness in the enrichment reading and it is recorded in the enrichment entry's counter-evidence too.
The test that separates them is cheap and stated below.
What exists, and the gap
Chemistry exists. A ubiquitin-rhodamine high-throughput screen produced small-molecule USP1 inhibitors with dose-dependent activity and IC50 in the high nanomolar range; pimozide is a known USP1 inhibitor; SJB3-019A is the tool compound used in the B-ALL work.
Activity in the right lineage exists. USP1 is upregulated in bone marrow cells from 30 newly diagnosed B-ALL patients; siRNA knockdown decreases growth and induces apoptosis; SJB3-019A represses proliferation and triggers apoptosis.
The intersection is empty. USP1[tiab] AND (KMT2A[tiab] OR MLL[tiab]) returns 0 records, counted 2026-09-20. USP1[tiab] AND (leukemia OR leukaemia)[tiab] returns 11, of which the B-ALL and pan-leukaemic ones are above; none is KMT2A-rearranged.
This is the cleanest shape a lead can have in this project: a target our own unbiased scan ranks 129th of 17,728, with published chemistry, published activity in the right lineage, and no publication at all in the index disease.
- Nothing here has been tested in a KMT2A-rearranged cell, an infant cell, or any KMT2A::AFF1 material. The selectivity rank comes from a panel of nine patients, seven of them myeloid.
- The published leukaemia mechanism is ID1, not FANCD2. If USP1's rank is an ID1 effect then the FA-module reasoning that led here does not apply, and the lead should be pursued on its own terms rather than as FA-pathway evidence.
- Pimozide is in this chemical literature and is an antipsychotic with hERG liability. Do not treat 'USP1 inhibitors exist' as 'a tolerable USP1 inhibitor exists'.
- USP1 reverses FANCD2 monoubiquitination, so inhibiting it does not simply switch the FA module off - it locks it in the ubiquitinated state. The direction of the intended effect needs stating before anyone designs an experiment.
- The separating experiment: does USP1 inhibition in a KMT2A-rearranged line phenocopy FANCL or UBE2T loss, or ID1 over-expression rescue? One answers whether the FA reading or the ID1 reading is right, and the reagents for both already exist.
- Is USP1 in the DepMap panel selective for the same lines as FANCL and UBE2T? A within-panel correlation across the eleven lines would say whether it is behaving as part of that module at all, and the data are already fetched.
- Read the full text of PMID 33390793 for the cell lines used. If any is KMT2A-rearranged, the zero-record absence above is wrong in the most embarrassing way available.
- Do the clinical-stage USP1 inhibitors developed for BRCA-mutant cancers have any published leukaemia data? The three-record query for named compounds plus leukaemia was not read.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
USP1 is generally upregulated in bone marrow cells from 30 newly diagnosed B-ALL patients by real-time PCR and western blot; siRNA knockdown of USP1 decreases B-ALL cell growth and induces apoptosis; pharmacological inhibition with SJB3-019A represses proliferation and triggers apoptosis; and the effect runs through ID1, since inhibition downregulates ID1 and phospho-AKT and ID1 over-expression reverses the suppression.
Inhibition of USP1 induces apoptosis via ID1/AKT pathway in B-cell acute lymphoblastic leukemia cells · International Journal of Medical Sciences 2021What this evidence cannot show
Abstract fetched from NCBI efetch and read by the coordinator on 2026-09-20; full text not read, so cell lines, concentrations and the number of replicates are not recorded and must not be quoted from this item. **None of the 30 patients is described as KMT2A-rearranged** — the abstract does not report cytogenetics at all. SJB3-019A is a tool compound, not a clinical agent. **The substrate question has been tested and answered with a third option.** Co-dependency across 1,208 lines puts USP1 at r = +0.098 to +0.207 against the FA module (internal median 0.576) and −0.022 against ID1. Neither axis explains its selectivity, so this lead should be pursued on its own terms and **not** as FA-pathway evidence. See `ana-usp1-is-in-neither-home`.
Evidence that argues against this (3)
- The mechanism reported is ID1/AKT, which is **not** the axis this project arrived by. USP1's rank in the KMT2A-rearranged selectivity scan was interpreted as part of an FA ubiquitination module; if its leukaemia role is ID1 stabilisation then its FA-set membership is annotation overlap rather than evidence, and the enrichment reading loses one of its more striking members. Oncofusion knowledge graph (internal analysis) 2026
- Pimozide, one of the 'known USP1 inhibitors' in this literature, is an antipsychotic with substantial off-target pharmacology including dopamine D2 antagonism and hERG liability. Growth inhibition by pimozide is weak evidence of on-target USP1 dependence, and its presence in the chemical series is a caution about the whole class's selectivity. Molecular Cancer Therapeutics 2013
- Tested directly rather than argued: USP1's cross-line dependency profile matches neither the FA module nor ID1, and the negative control ACTB (-0.127) exceeds several of the module correlations in magnitude. Correlations of that size in this matrix are what unrelated genes look like. Oncofusion knowledge graph (internal analysis) 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Small-molecule USP1 inhibitors were identified by ubiquitin-rhodamine-based high-throughput screening, block USP1 deubiquitinase activity in vitro in a dose-dependent manner with IC50 in the high nanomolar range, promote ID1 degradation and inhibit the growth of leukaemic cell lines dose-dependently.
Small-molecule inhibitors of USP1 target ID1 degradation in leukemic cells · Molecular Cancer Therapeutics 2013What this evidence cannot show
Abstract read 2026-09-20; full text not read. A 2013 screening paper: the compounds are tool molecules and the abstract reports no selectivity data against other deubiquitinases, no cellular target engagement assay and no in vivo work. 'Leukaemic cell lines' is unspecified in the abstract and no KMT2A-rearranged line is named.
Evidence that argues against this (1)
- Deubiquitinase inhibitors as a class have a poor track record of on-target selectivity, and high-nanomolar biochemical IC50 with no cellular engagement data leaves growth inhibition attributable to almost anything. This entry should not be read as saying a usable USP1 inhibitor exists — only that chemistry has been attempted and published. Molecular Cancer Therapeutics 2013
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
USP1 has never been studied in KMT2A-rearranged leukaemia in the indexed literature: `USP1[tiab] AND (KMT2A[tiab] OR MLL[tiab])` returns 0 records, counted 2026-09-20, against 11 records for USP1 in leukaemia generally and 38 for 'USP1 inhibitor'.
What this evidence cannot show
A counted absence over a `[tiab]` query, which cannot see a paper that discusses USP1 only in its body text or supplementary tables — the exact failure that hid AFF1 from a proximity map earlier in this project. Per the standing rule, an absence of testing is not a negative result: nothing here says USP1 is unimportant in KMT2A-rearranged disease.
Evidence that argues against this (1)
- Zero records may reflect that nobody had a reason to look rather than that the question is open and valuable. USP1 is a broadly studied DUB and KMT2A-rearranged leukaemia is a heavily studied disease; an empty intersection between two well-populated literatures is at least as likely to mean the combination was tried and unremarkable as that it is a gap. NCBI 2026