XPO1 mRNA goes UP when you inhibit the protein - the pharmacodynamic marker that made the paediatric trial interpretable, and what it cannot tell you
also called XPO1 mRNA induction · SINE pharmacodynamic biomarker · XPO1 autoregulation
When a drug blocks XPO1, the cell notices and makes more XPO1 messenger RNA. That rise is easy to measure in a blood sample and proves the drug reached its target - which is why a small children's trial could say with confidence that the dose was doing something.
The same thing, for a scientist
**The observation.** XPO1 inhibition induces a compensatory rise in **XPO1 mRNA**. In the paediatric SELHEM phase 1 (NCT02212561), a **2-fold or greater increase in XPO1 mRNA was demonstrated in 15 of 16 patients**, and **all patients treated at 40 mg/m2 or above showed inhibition** [E1]. That is an unusually clean pharmacodynamic result for a paediatric phase 1 and it is why this graph's `drug-selinexor` entry treats target engagement in children as established while treating efficacy as unproven. **Why it happens, at the level this beat can defend.** XPO1 exports its own regulators; blocking export traps transcription factors in the nucleus, and the cell's response includes up-regulating the transporter. The precise circuit was not resolved in this beat and should not be asserted - what is established is the **direction and the magnitude in patients**, not the mechanism. **What it measures and what it does not.** | the marker reports | the marker does NOT report | |---|---| | that the drug reached XPO1 in a cell you can sample | that the drug reached chromatin-bound XPO1 | | that the dose is pharmacologically active | that the fusion or mutant has left its target loci | | a cellular response to inhibition | anything about duration between doses | **That third row is the one that matters here.** The NPM1c schedule work showed that weekly selinexor disrupts the XPO1-NPM1c interaction only **temporarily** [E2] - and an XPO1 mRNA rise measured at a single post-dose timepoint would look identical whether the interaction stayed disrupted or not. **A marker that confirms engagement at a timepoint cannot bound engagement over an interval**, and the interval is what the NPM1c result says decides efficacy.
The biomarker this axis actually needs and does not have
hyp-xpo1-is-a-shared-address-for-hoxa-driven-leukaemia proposes selecting patients by address usage - whether their lesion is docking onto chromatin-bound XPO1 - rather than by fusion identity. XPO1 mRNA induction is the wrong assay for that, because it measures the drug's effect on the cell, not the lesion's use of the platform.
Candidate assays, none of which exists as a clinical test:
- Fusion or mutant occupancy at HOXA before and after a dose - the direct readout, needs a marrow sample and a CUT&RUN, feasible in a trial and never done.
- HOXA9/MEIS1 transcript fall - a downstream surrogate, easy, and already used in the preclinical literature; it would at least be specific to the pathway rather than to the drug.
- XPO1 chromatin occupancy itself - the ideal and the least practical.
The cheapest useful addition to any future trial is the second one, and a trial that measures only XPO1 mRNA will again be unable to say whether the pathway moved.
| mechanism class | pharmacodynamic feedback |
|---|---|
| marker | XPO1 mRNA fold-change from baseline |
| threshold used clinically | 2-fold or greater increase |
| clinical performance | 15 of 16 paediatric patients; all patients at or above 40 mg/m2 |
| trial | NCT02212561 |
| mechanism of the feedback | not resolved in this beat; direction and magnitude are established in patients, the circuit is not |
what it reports
[ "drug reached XPO1 in a samplable cell", "dose is pharmacologically active" ]
what it does not report
[ "engagement of the chromatin-bound XPO1 pool", "displacement of a fusion or mutant from its target loci", "duration of engagement across a dosing interval - which the NPM1c work identifies as what decides efficacy" ]
missing assays for address usage
[ "fusion or mutant occupancy at HOXA before and after a dose (CUT&RUN on a marrow sample)", "HOXA9 and MEIS1 transcript fall as a pathway-specific surrogate", "XPO1 chromatin occupancy itself" ]
- A compensatory rise in the target's mRNA is a plausible route to reduced effect on the next dose, and the marker used to prove engagement is measuring the beginning of the escape.
- Confirming target engagement in a trial that cannot show pathway modulation gives false reassurance: the paediatric trial established the drug hit XPO1 and remains unable to say whether it hit the leukaemia.
- Any biomarker strategy built on XPO1 mRNA will look identical in a patient whose lesion uses the chromatin address and one whose does not.
- What is the circuit? XPO1 mRNA induction is used as a standard pharmacodynamic marker across the SINE programme and the transcriptional mechanism was not established in anything read here.
- In which cells was the paediatric measurement made - blasts, or whole blood? A rise in normal leucocytes proves systemic exposure and says nothing about the leukaemia.
- Does the mRNA rise translate into a protein rise, and could that blunt a subsequent dose? A compensatory increase in the target is a resistance mechanism in waiting and nobody appears to have measured XPO1 protein across a dosing cycle.
- Would adding a HOXA9/MEIS1 transcript endpoint to a future paediatric trial be free? It is a qPCR on samples already being drawn, and its absence is why no existing trial can say whether the pathway moved.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In the paediatric SELHEM phase 1 of selinexor with fludarabine and cytarabine (NCT02212561), XPO1 target inhibition measured as a two-fold or greater increase in XPO1 mRNA was demonstrated in 15 of 16 patients, and all patients treated at 40 mg/m2 or above showed inhibition.
Phase I Study of Selinexor, a Selective Inhibitor of Nuclear Export, in Combination With Fludarabine and Cytarabine, in Pediatric Relapsed or Refractory Acute Leukemia. · Alexander TB, Lacayo NJ, Choi JK, et al. · Journal of Clinical Oncology 2016What this evidence cannot show
n=18 enrolled, 16 evaluable for this endpoint, single-arm phase 1. The cell type in which the mRNA was measured and the post-dose timing are not stated in what was read here, and both matter for interpreting the marker. Given with fludarabine and cytarabine, so the measured cells were under combination pressure. Recorded in full in `drug-selinexor`.
Evidence that argues against this (1)
- Target engagement was demonstrated in 15 of 16 children and the trial still could not attribute its 47% CR/CRi rate to selinexor, because fludarabine and cytarabine alone produce remissions in this setting. A pharmacodynamic marker that is clean while the efficacy question stays wide open is a demonstration that engagement and benefit are different quantities, not a validation of the marker's usefulness. Alexander TB et al. 2016
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Selinexor administered once or twice per week results in only a temporary disruption of the XPO1-NPM1c interaction, which the authors give as the reason it did not translate into clinical benefit in NPM1-mutated patients, whereas the more frequently dosed eltanexor caused irreversible HOX down-regulation and terminal differentiation.
Prolonged XPO1 inhibition is essential for optimal antileukemic activity in NPM1-mutated AML. · Pianigiani G, Gagliardi A, Mezzasoma F, et al. · Blood Advances 2022What this evidence cannot show
Abstract-level; NPM1-mutated AML models. Cited here for the logical point that a single-timepoint engagement marker cannot bound engagement over a dosing interval - which is an inference this entry draws, not a claim the paper makes.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT against the inference this item is cited to license. The logical point drawn here is that a single-timepoint engagement marker cannot bound engagement over a dosing interval, and that intermittent selinexor therefore under-engages. But a twice-weekly selinexor schedule combined with daunorubicin and cytarabine produced CR/CRi in 8 of 19 evaluable previously untreated poor-risk AML patients with the MTD not reached. Whatever the XPO1-mRNA marker is or is not bounding, intermittent dosing in AML is not reliably sub-therapeutic, so the marker's inadequacy and the schedule's inadequacy are separate questions and this entry currently runs them together. Sweet K et al. 2020
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
XPO1 is annotated by UniProt as the receptor whose export of NES cargoes includes transcriptional regulators, providing the substrate for a compensatory transcriptional response to its own inhibition.
UniProtKB - O14980 (XPO1_HUMAN), Exportin-1 · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Cited only to establish the plausibility of a transcriptional feedback loop. The specific circuit driving XPO1 mRNA up under SINE treatment was NOT resolved in this beat and is deliberately not asserted.
Evidence that argues against this (1)
- NAMED METHODOLOGICAL DEFEATER for treating an XPO1 mRNA rise as attributable to the SINE. XPO1 protein levels move under compounds annotated to entirely different targets: Chen et al. screened 3,200 bioactive small molecules and found that C646 - the canonical p300 HAT inhibitor - directly targets and DEGRADES XPO1 through neddylation-dependent proteolysis, with multiple C646 phenotypes abrogated by the XPO1-C528S allele. If a widely used epigenetic tool compound degrades XPO1 as an unrecognised molecular glue, then XPO1 protein and the transcriptional feedback that follows it are movable by agents nobody is attributing to XPO1. In SELHEM the children received fludarabine and cytarabine alongside selinexor, and a 2-fold XPO1 mRNA rise was not shown to be selinexor-specific by any comparator arm. Chen YF et al. 2024
the pharmacodynamic marker used in the paediatric phase 1
XPO1 mRNA rises when XPO1 protein is inhibited
it measures target engagement and NOT address usage - the missing biomarker
XPO1 protein is degraded by C646, a compound universally annotated to p300, via a neddylation-dependent molecular-glue mechanism (PMID 38917791) - so XPO1 levels and the feedback that follows them move under agents nobody attributes to XPO1, and the SELHEM children received fludarabine and cytarabine alongside selinexor with no comparator arm