Acute promyelocytic leukemia (APL) is a distinct subtype of acute myeloid leukemia (AML) defined by a specific chromosomal accident: the translocation t(15;17), where pieces of chromosomes 15 and 17 break and swap. This fusion joins the PML (promyelocytic leukemia) gene on chromosome 15 with the RARa (retinoic acid receptor alpha) gene on chromosome 17, creating an abnormal fusion protein called PML/RARa.
PML/RARa is the central driver of APL. It is both necessary and sufficient to cause the disease - introducing PML/RARa into mice causes APL, and the leukemia cells depend on its continued expression for survival. It works by blocking normal white blood cell maturation, freezing immature cells in a proliferating state. Eliminating or degrading PML/RARa is therefore the primary therapeutic strategy in APL.
Standard treatment with ATRA (all-trans retinoic acid) and ATO (arsenic trioxide) has transformed APL from one of the most deadly leukemias into one of the most curable. Both drugs work by triggering the degradation of PML/RARa through the cell's protein disposal machinery. However, these treatments have significant limitations: toxic side effects and the emergence of drug resistance in some patients.
Drug resistance in APL arises when mutations develop in the PML or RARa portions of the fusion protein. These mutations disrupt the binding sites for ATRA and ATO, preventing the drugs from triggering degradation. Patients with these resistant mutations face very poor outcomes, making novel therapeutic strategies urgently needed.
Cells continuously produce and destroy proteins as needed. A key mechanism for controlled protein destruction is the ubiquitin-proteasome pathway: a cellular quality control and regulatory system where specific proteins are tagged with small molecules called ubiquitin and then degraded by a molecular shredder called the proteasome.
The ubiquitin tagging process can be reversed by enzymes called deubiquitinases (DUBs). DUBs remove ubiquitin tags from proteins, rescuing them from proteasomal destruction and thereby increasing their stability. When a DUB acts on an oncogenic protein - a cancer-promoting protein that should be eliminated - it inadvertently protects the cancer by extending the life of the harmful protein.
The researchers hypothesized that PML/RARa's unusual stability (it has a 24-hour half-life - much longer than typical regulatory proteins) might be partly explained by a specific DUB that keeps removing its ubiquitin tags, shielding it from degradation. Identifying and blocking this DUB could offer a completely new way to destroy PML/RARa - one that would work even when ATRA and ATO cannot.
DUBs are particularly attractive as drug targets because they have well-defined active sites (catalytic pockets) where small-molecule inhibitors can bind and block their activity. The human genome encodes about 100 different DUBs, most of which target specific substrates - making it possible to design inhibitors that selectively affect one DUB without disrupting the entire system.
To identify which of the approximately 98 known human DUBs was responsible for stabilizing PML/RARa, the researchers conducted a systematic siRNA library screen. They temporarily silenced each DUB one at a time using small RNA molecules (siRNA - short interfering RNA that blocks specific gene expression) and measured the effect on PML/RARa protein levels.
The screen used a clever dual-luciferase reporter system: PML/RARa was tagged with firefly luciferase (a light-producing enzyme), and a second luciferase from a different organism served as an internal control. By measuring how much light was produced, researchers could precisely quantify PML/RARa protein levels in thousands of cells simultaneously and systematically.
From 98 DUBs tested, three candidates reduced PML/RARa levels by approximately 50%: USP28, USP37, and YOD1. Follow-up validation using a more specific genetic silencing approach (shRNA) confirmed that only YOD1 consistently and substantially reduced PML/RARa protein levels. The effect was proteasome-dependent - blocking the proteasome reversed YOD1 silencing's effect - confirming that YOD1 protects PML/RARa specifically through the ubiquitin-proteasome pathway.
The researchers also confirmed that YOD1 does not affect PML/RARa at the RNA level (the gene is still transcribed normally) - only the protein level changes. This means YOD1 acts post-translationally, by removing ubiquitin tags from the PML/RARa protein and preventing its degradation, not by affecting gene expression.
The researchers confirmed through multiple approaches that YOD1 physically interacts with PML/RARa and directly removes its ubiquitin tags. Co-immunoprecipitation experiments showed that YOD1 and PML/RARa proteins bind to each other in cells. In biochemical assays, purified YOD1 protein efficiently stripped ubiquitin chains from PML/RARa in a cell-free system - and a catalytically dead version of YOD1 (YOD1-C160S, with its active site mutated) could not.
Critically, YOD1 also physically interacted with and deubiquitinated all four tested drug-resistant PML/RARa mutants - including the ATRA-resistant variants (DF286 and R276Q) and ATO-resistant variants (A216V and L218P). The drug-resistance mutations affect the drug-binding sites, not the YOD1 interaction site. This means YOD1 silencing can degrade PML/RARa regardless of which drug-resistance mutation is present.
This finding is of great clinical significance: while ATRA and ATO become ineffective when their binding sites on PML/RARa mutate, YOD1 continues to interact with and stabilize even these mutated proteins. Blocking YOD1 therefore provides a route to destroy drug-resistant forms of PML/RARa that cannot be targeted by existing therapies.
When YOD1 was overexpressed, both wild-type and all four drug-resistant PML/RARa mutants accumulated to higher levels - the opposite of what happens when YOD1 is silenced. This bidirectional relationship further confirms YOD1's central role as the guardian of PML/RARa stability in APL cells.
With the molecular mechanism established, the researchers tested whether YOD1 inhibition could actually kill APL cells. In laboratory cultures, silencing YOD1 with shRNA potently inhibited the proliferation and colony-forming ability of NB4 cells (a standard APL cell line), induced significant apoptosis (programmed cell death), and promoted cell differentiation - all of which represent restoration of normal cell biology that PML/RARa had blocked.
Importantly, silencing YOD1 had similar strong killing effects on both drug-resistant APL cell lines (NB4R1 and NB4R2, which carry ATRA-resistant PML/RARa mutations). These cells, which cannot be effectively treated with current drugs, lost their proliferative capacity and underwent apoptosis at similar rates to drug-sensitive APL cells when YOD1 was removed.
By contrast, YOD1 depletion caused very little toxicity in non-APL leukemia cell lines (U937 and HL60) and in normal human hematopoietic stem cells (HSCs) - the healthy bone marrow cells that should be preserved during leukemia treatment. This selectivity is highly encouraging for therapeutic development, suggesting that YOD1 inhibition could be targeted against APL cells without broadly damaging normal blood cell production.
In mouse models (xenograft models where human APL cells are transplanted into immunocompromised mice), YOD1 knockdown dramatically reduced leukemia burden in bone marrow and significantly prolonged survival compared to control mice. Mice with drug-resistant APL (NB4R1 cells) similarly showed reduced tumor burden and extended survival when YOD1 was silenced - proof that the therapeutic strategy works against resistant disease in a living organism.
Genetic silencing (shRNA) is valuable for proof-of-concept studies but cannot be used directly as a medicine. The researchers therefore searched for a small-molecule drug that could pharmacologically block YOD1's enzymatic activity. They screened a library of 31 known DUB inhibitors in cells expressing PML/RARa, measuring which compounds caused the greatest reduction in PML/RARa protein levels.
Three compounds reduced PML/RARa levels substantially. Using biochemical assays, only one - G5 (ubiquitin isopeptidase inhibitor I) - was confirmed to block YOD1's actual enzymatic activity (deubiquitination of PML/RARa). The other two compounds affected PML/RARa stability through different, YOD1-independent mechanisms. G5 is thus identified as the first pharmacological inhibitor of YOD1.
G5 inhibited YOD1's deubiquitination activity in a concentration-dependent manner: higher doses of G5 caused more complete YOD1 inhibition and greater PML/RARa accumulation of ubiquitin tags (making it more susceptible to proteasomal degradation). In APL cells, G5 potently degraded endogenous PML/RARa protein - including all four drug-resistant PML/RARa mutants in NB4R1 and NB4R2 cells.
The drug specificity is reflected in the mechanism: G5 degrades PML/RARa not by directly targeting the fusion protein, but by blocking the enzyme (YOD1) that protects it. This is conceptually similar to how proteolysis-targeting chimeras (PROTACs) work - an emerging drug strategy in oncology where the cancer cell's own degradation machinery is redirected against oncoproteins.
G5 showed potent anti-APL activity in cell culture: it inhibited proliferation, suppressed colony formation, and induced apoptosis in both ATRA-sensitive NB4 cells and ATRA-resistant NB4R1 and NB4R2 cells in a dose-dependent manner. Importantly, G5 was more effective against drug-resistant cells than ATRA itself - the cells that cannot be treated with the current standard drug responded strongly to G5.
In a mouse xenograft model using drug-resistant NB4R2 cells, G5 administered intravenously every other day at doses of 20 or 40 mg/kg significantly inhibited tumor growth and reduced tumor weight compared to vehicle-treated mice, while ATRA had minimal effect. The T/C ratio (a standard measure of tumor growth inhibition; values below 50% indicate effective treatment) confirmed G5's activity met the threshold for therapeutic efficacy.
Most compellingly, G5 was tested against primary APL cells directly isolated from 12 patients' bone marrow. These patient-derived blasts are the closest possible test of clinical relevance. G5 treatment caused significant apoptosis in primary APL cells from all 12 patients, including samples from pediatric patients, confirming that the therapeutic mechanism identified in cell lines and mice operates in actual human leukemia cells.
The combination of G5 with existing APL drugs (ATO or ATRA) was also explored, with results suggesting additive or potentially synergistic cell killing. If G5 can be combined with standard therapies, this could allow dose reductions (reducing toxicity) while maintaining or improving treatment efficacy - an important consideration for a patient population that already faces significant treatment-related side effects.
This study establishes YOD1 as a previously unrecognized therapeutic target in APL and identifies G5 as the first pharmacological inhibitor of YOD1. Together, these discoveries define a new molecular vulnerability in APL that is independent of the drug-binding sites targeted by ATRA and ATO - meaning this approach can overcome the mutations that make APL resistant to standard therapy.
The key mechanistic insight is that YOD1 maintains PML/RARa stability by continuously removing its ubiquitin degradation tags. When YOD1 is blocked, PML/RARa accumulates ubiquitin and is destroyed by the proteasome. Critically, this mechanism operates on all tested forms of PML/RARa - both wild-type and drug-resistant mutants - because the resistance mutations affect drug-binding regions, not the YOD1-interaction region.
The selective toxicity toward APL cells over normal hematopoietic stem cells is an important safety feature. Because normal blood cells do not express PML/RARa, they lack the specific vulnerability that YOD1 inhibition exploits. This PML/RARa dependency creates a therapeutic window - a dose range where G5 kills leukemia cells more effectively than it harms normal tissue.
Future work will need to optimize G5 as a drug candidate, including improving its pharmacokinetic properties (how the body absorbs and eliminates it) and conducting formal toxicity and efficacy studies. Clinical translation will require careful testing, but the convergence of a validated target, a novel mechanism of action, and demonstrated activity in patient-derived cells makes this a compelling foundation for developing a new APL therapy.