Reviewed 19 May 2026

Example research study

Low-Dose Osimertinib Paradoxically Activates ERK Signaling in EGFR-Mutant Lung Cancer Cells

Treating PC9 cells (EGFR exon 19 deletion) with 1–5 nM osimertinib increases ERK phosphorylation and cell viability above untreated controls—even as p-AKT declines at every dose. BioSkepsis traced this paradox through 16 verified PMIDs to a convergence of kinetic proofreading, feedback-loop disruption, and MIG6 rheostat failure that illuminates how sub-therapeutic TKI exposure can make tumour signaling worse, not better.

TL;DR Sub-stoichiometric osimertinib shortens the active lifetime of EGFR without fully suppressing it. This prevents the receptor from completing multi-site autophosphorylation needed to engage negative feedback regulators—MIG6 recruitment, receptor internalization, DUSP4/6 expression. With feedback disarmed but the kinase still partially active, ERK signaling becomes paradoxically sustained. The effect disappears above ~10 nM, where drug occupancy is sufficient for comprehensive pathway blockade. The same feedback architecture drives the adaptive ERK rebound observed at therapeutic doses within 16–24 hours and is a primary mechanism limiting single-agent TKI durability.

Kinetic Proofreading Explains How Partial EGFR Inhibition Sustains ERK Output

Full EGFR activation triggers a sequential autophosphorylation program that engages negative feedback: recruitment of protein tyrosine phosphatases, receptor internalization, and expression of inducible inhibitors such as MIG6. In PC12 and MCF7 cells, Kiyatkin et al. demonstrated that ERK dynamics faithfully follow receptor tyrosine kinase (RTK) activation kinetics—not downstream feedback topology (PMID: 32817373).

When sub-stoichiometric erlotinib (30–50 nM) was added to MCF7 cells, both EGFR and ERK phosphorylation increased at 60 minutes post-stimulation, despite partial suppression of the initial signal at 5 minutes. The drug reduces the average active-state lifetime of each kinase domain within a dimer, preventing the completion of the phosphorylation steps required to engage internalization and feedback. The result is a paradoxically more sustained signal (PMID: 32817373).

Osimertinib is a covalent, third-generation EGFR inhibitor—but at 1–5 nM it behaves as a sub-stoichiometric occupant. Only a fraction of EGFR molecules are inhibited at any instant, placing the system in the same kinetic proofreading regime described for reversible inhibitors.

MIG6 (ERRFI1): The Activity-Based Rheostat That Fails at Low Drug Doses

MIG6 is a direct, activity-based inhibitor of EGFR. Its segment 2 domain must be phosphorylated at Y394 and Y395—by EGFR itself and by Src—before it can fold into a hairpin that blocks the kinase's peptide-substrate binding site (PMID: 26280531). This design means MIG6 only engages when the receptor is robustly active; partial inhibition by a low-dose TKI prevents the receptor from catalysing MIG6 phosphorylation fast enough to close the loop.

Because ERRFI1 transcription is itself maintained by ERK pathway activity, initial partial inhibition of ERK reduces new MIG6 protein synthesis. Data from KRAS G12D-mutant colorectal cancer cells treated with MRTX1133 showed that ERRFI1 mRNA and protein fell in a time-dependent manner following pathway blockade (PMID: 37020035). The same principle applies to osimertinib at sub-therapeutic doses: ERK is dampened just enough to reduce MIG6 induction, but not enough to suppress the pathway—a rheostat failure that amplifies the paradoxical signal.

In gefitinib-sensitive PC9 cells, MIG6 overexpression was sufficient to reverse TKI resistance, confirming that the stoichiometric balance between EGFR and MIG6 determines the signaling threshold (PMID: 25400829). In glioblastoma models, the EGFR-to-MIG6 ratio likewise predicted sensitivity to feedback regulation (PMID: 39129344).

Relief of Negative Feedback: DUSP Depletion and Wild-Type RAS Reactivation

MIG6 is not the only feedback regulator disrupted at low doses. DUSP4 and DUSP6—dual-specificity phosphatases that tonically dephosphorylate and inactivate ERK—are direct transcriptional targets of the ERK pathway. Acute treatment of KRAS-mutant cells with AMG 510 (sotorasib) produced rapid, universal suppression of DUSP4 and DUSP6 protein expression (PMID: 35732135). When these phosphatases disappear, any residual ERK activity is amplified because its negative regulators are no longer present.

Concurrently, wild-type RAS isoforms (NRAS, HRAS) can be reactivated through RTK-mediated signaling once the oncogenic driver is partially blocked. In KRAS G12C-inhibited cells, feedback activation of wild-type RAS sustained ERK activity even while KRAS-GTP levels remained suppressed (PMID: 35732135). SHP2—a nonreceptor protein tyrosine phosphatase that transduces RTK signals to RAS—acts as the critical node connecting receptor-level feedback relief to downstream ERK rebound (PMID: 33046519).

Negative feedback regulators disrupted by sub-therapeutic EGFR/KRAS inhibition

Regulator	Function	Effect of inhibitor	Evidence
MIG6 (ERRFI1)	Activity-based EGFR kinase inhibitor	Transcription reduced; phosphorylation impaired	PMID: 26280531, 37020035
DUSP4	ERK phosphatase	Protein expression suppressed	PMID: 35732135
DUSP6	ERK phosphatase	Protein expression suppressed	PMID: 35732135
SPRY proteins	Inhibit Grb2–RTK recruitment and Raf activation	Transcription reduced (ERK-dependent)	PMID: 33046519
ErbB2 T677 phosphorylation	ERK-mediated negative feedback on ErbB2	Lost when ERK is suppressed	PMID: 29434871

Paradoxical Raf-1 Activation: Lessons from Sorafenib and Vemurafenib in Lung Cancer Biology

The ERK paradox has a structural parallel one level downstream. Low-dose sorafenib paradoxically activates Raf-1 in polycystin-2-defective hepatocytes by binding to one member of a B-Raf/Raf-1 heterodimer and transactivating the partner in a RAS-GTP-dependent manner (PMID: 22653837). The same mechanism operates in acute myeloid leukaemia cells lacking FLT3-ITD mutations, where sorafenib induced phosphorylation of Raf-1 and Erk2 at sub-therapeutic concentrations (PMID: 25665465).

Vemurafenib extends the principle to melanoma: in tissues with elevated wild-type RAS activity, the drug stimulates RAF dimerisation and downstream ERK signaling instead of blocking it (PMID: 28188228). Although osimertinib is not a Raf inhibitor, its partial suppression of EGFR at low doses creates the same upstream condition—elevated RAS-GTP from feedback relief—that licenses paradoxical Raf dimer transactivation.

From Bench Paradox to Bedside Resistance: The ERK Rebound in EGFR-Mutant NSCLC

The low-dose paradox is a compressed version of the adaptive resistance observed at therapeutic concentrations. Across tyrosine-kinase-driven leukaemias, breast cancers, and NSCLC, acute TKI treatment triggers ERK reactivation within 16–24 hours even while the primary target remains suppressed (PMID: 28923853). In HER2-positive breast cancer, chronic lapatinib treatment induces a FOXO-dependent feedback loop that reactivates the RAF/MEK/ERK cascade (PMID: 28744398).

Osimertinib-specific resistance pathways include MDM2 amplification, which targets the FBW7 tumour suppressor for K48-linked polyubiquitination, thereby stabilising the anti-apoptotic protein MCL-1 and blocking apoptosis (PMID: 39543744). ERK-dependent mechanisms also sustain the CDK4/6-RB axis under partial EGFR inhibition, maintaining cell-cycle progression even when AKT is suppressed (PMID: 38025818). Preclinical data show that acquired AZD9291 resistance in PC9 cells is dependent on ERK signaling rather than EGFR reactivation (PMID: 29641535).

BioSkepsis-synthesised evidence vs. general-purpose LLM output on osimertinib ERK paradox

Dimension	BioSkepsis	General-purpose LLM
Citation grounding	Every claim linked to a specific PMID with directness tier and confidence rating	Narrative plausible but PMIDs absent or unverified
Unverified citation flagging	Three-stage verification; failed citations quarantined with failure reasons	No verification pipeline; hallucinated citations indistinguishable from real ones
Hypothesis generation	Empirically testable hypothesis with falsification criteria and study design	Vague "future directions" without operationalisable predictions
Mechanistic links table	Structured graph of molecular actors, link types, targets, effects, and PMIDs	Prose-only; no machine-readable mechanistic map
Evidence tiering	Direct / Indirect / Derived labels with High / Medium / Low confidence	All claims presented at equal confidence

Clinical Context: Osimertinib in Adjuvant EGFR-Mutant NSCLC and Molecular Residual Disease

Despite the feedback-driven limitations at the molecular level, osimertinib delivers substantial clinical benefit when dosed at therapeutic concentrations. In the phase III ADAURA trial, adjuvant osimertinib in resected stage IB–IIIA EGFR-mutated NSCLC produced a disease-free survival hazard ratio of 0.17 (99.06% CI, 0.11–0.26; p < 0.001) in the stage II–IIIA population, maturing to 0.23 (95% CI, 0.18–0.30) at four years of follow-up (PMID: 32955177; PMID: 36720083).

Post hoc molecular residual disease (MRD) analysis showed that ctDNA detection preceded imaging-based recurrence by a median of 4.7 months (95% CI, 2.2–5.6), suggesting that the ERK-driven signaling adaptations observable at the bench eventually manifest as detectable clonal outgrowth in the clinic (PMID: 40097663).

These data underscore that the low-dose ERK paradox is not a clinical liability at standard dosing—but the same feedback architecture that produces it is the mechanistic root of the adaptive resistance that eventually limits osimertinib's durability.

BioSkepsis-Generated Testable Hypothesis: The MIG6 Rheostat Model

BioSkepsis synthesised the evidence into an empirically testable hypothesis: sub-therapeutic osimertinib amplifies ERK signaling in PC9 cells by inducing a critical imbalance in the EGFR–MIG6 rheostat, where partial inhibition is sufficient to transcriptionally downregulate ERRFI1 while failing to achieve activity-based engagement of existing MIG6 protein.

The hypothesis predicts that 1–5 nM osimertinib will reduce MIG6 protein levels within 8 hours; that p-EGFR intensity at 6 hours will be higher in 2.5 nM-treated cells than in untreated controls; that siRNA-mediated MIG6 knockdown will abolish the dose-dependent ERK surge by producing constitutively high p-ERK; and that the low-dose viability boost will be sensitive to MEK or wild-type RAS inhibition.

Falsification criteria: the hypothesis fails if the ERK surge occurs without a concomitant decrease in MIG6 protein, or if MIG6 overexpression does not dampen the surge. If p-ERK increases without a rise in p-EGFR, the mechanism is downstream of the receptor—pointing instead to paradoxical Raf-1 activation.

Who Benefits From Citation-Grounded Signaling Synthesis in Lung Cancer Research?

Frequently Asked Questions About Osimertinib ERK Paradox in EGFR-Mutant Lung Cancer

Why does low-dose osimertinib increase ERK phosphorylation in EGFR-mutant cells?

At sub-therapeutic concentrations (1–5 nM), osimertinib shortens the active lifetime of the EGFR kinase domain without fully suppressing it. This prevents the receptor from completing the multi-site autophosphorylation program required to engage negative feedback regulators such as MIG6 (ERRFI1) and receptor internalization. With feedback disarmed but the receptor still partially active, ERK signaling becomes paradoxically more sustained than in untreated cells (PMID: 32817373).

Is the paradoxical ERK surge specific to osimertinib or does it occur with other EGFR TKIs?

The phenomenon is not unique to osimertinib. Sub-stoichiometric erlotinib increases EGFR and ERK phosphorylation in MCF7 cells at 30–50 nM (PMID: 32817373). Low-dose sorafenib paradoxically activates Raf-1 and ERK in hepatocellular carcinoma and AML models (PMID: 22653837; PMID: 25665465). The principle extends to any reversible ATP-competitive kinase inhibitor used at concentrations that partially occupy the target.

What role does MIG6 (ERRFI1) play in the low-dose ERK paradox?

MIG6 is an activity-based inhibitor of EGFR that requires phosphorylation at Y394 and Y395 to block the receptor's peptide-substrate binding site (PMID: 26280531). Its expression is maintained by ERK pathway activity. At low inhibitor doses, both the activation of existing MIG6 protein and the transcription of new ERRFI1 mRNA are impaired, creating a window where the EGFR escapes negative regulation and ERK signaling rises (PMID: 37020035).

Could this paradoxical activation cause clinical resistance to osimertinib?

The in vitro paradox at 1–5 nM is below clinical plasma concentrations (Cmin ~160 nM). However, it illuminates the feedback architecture that drives adaptive resistance. The same feedback-relief mechanisms—DUSP4/6 downregulation, MIG6 depletion, wild-type RAS reactivation—are observed at therapeutic doses within 16–24 hours and contribute to the ERK rebound that limits single-agent TKI durability (PMID: 28923853; PMID: 35732135).

What experimental controls distinguish a real ERK rebound from a drug-degradation artefact?

Drug media should be refreshed every 24 hours and inhibitor stability confirmed by LC-MS/MS. Serum concentration must be standardised (10% FBS), because serum growth factors are required to drive the RTK-mediated ERK rebound. Cells should be low-passage, STR-authenticated, and mycoplasma-free. Including a MEK inhibitor arm confirms that the viability boost is ERK-dependent (PMID: 35732135).

How did BioSkepsis synthesise this evidence compared to a general-purpose LLM?

BioSkepsis grounded every claim to a specific PMID with directness tier and confidence rating, flagged unverified citations that failed independent three-stage checks, and generated a testable hypothesis with falsification criteria. A general-purpose LLM would produce a plausible narrative but cannot verify PMIDs, tier evidence, or distinguish direct from derived claims.

What is the ADAURA trial and how does osimertinib perform in early-stage EGFR-mutant NSCLC?

ADAURA is a phase III trial of adjuvant osimertinib in resected stage IB–IIIA EGFR-mutated NSCLC. In the stage II–IIIA population, the disease-free survival hazard ratio was 0.17 (p < 0.001), maturing to 0.23 after four years of follow-up (PMID: 32955177; PMID: 36720083). Post hoc MRD analysis showed ctDNA detection preceded imaging-based recurrence by a median of 4.7 months (PMID: 40097663).

Run Your Own PMID-Grounded Signaling Analysis on BioSkepsis

Paste your bench observation—unexpected dose-response, paradoxical phosphorylation, resistance phenotype—and get a citation-verified mechanistic synthesis with evidence tiering, testable hypotheses, and falsification criteria.

Start free

Sources & further reading

1. Kiyatkin A et al. Kinetics of receptor tyrosine kinase activation define ERK signaling dynamics. Sci Signal. 2020. PMID: 32817373
2. Park E et al. Structure and mechanism of activity-based inhibition of the EGF receptor by Mig6. Nat Struct Mol Biol. 2015. PMID: 26280531
3. Freed DM et al. KRAS-independent feedback activation of wild-type RAS constrains KRAS inhibitor efficacy. Cell Rep. 2022. PMID: 35732135
4. Nagaria TS et al. Kinase inhibitors of HER2/AKT pathway induce ERK phosphorylation via a FOXO-dependent feedback loop. Am J Cancer Res. 2017. PMID: 28744398
5. Wu Y-L et al. Osimertinib in resected EGFR-mutated non-small-cell lung cancer. N Engl J Med. 2020. PMID: 32955177
6. Herbst RS et al. Adjuvant osimertinib for resected EGFR-mutated stage IB–IIIA NSCLC: updated ADAURA results. J Clin Oncol. 2023. PMID: 36720083
7. John T et al. Molecular residual disease analysis of adjuvant osimertinib in resected EGFR-mutated NSCLC. Nat Med. 2025. PMID: 40097663
8. Zhang J et al. MDM2 drives resistance to osimertinib by contextually disrupting FBW7-mediated destruction of MCL-1 protein. J Exp Clin Cancer Res. 2024. PMID: 39543744
9. Liu C et al. Combinations with allosteric SHP2 inhibitor TNO155 to block receptor tyrosine kinase signaling. Clin Cancer Res. 2021. PMID: 33046519
10. Nukaga S et al. Acquired resistance to AZD9291 as an upfront treatment is dependent on ERK signaling. PLoS One. 2018. PMID: 29641535
11. Tian J et al. Feedback activation of EGFR/wild-type RAS signaling axis limits KRAS inhibitor efficacy in KRAS-mutated colorectal cancer. Oncogene. 2023. PMID: 37020035
12. Bhatt DL et al. Cyclic AMP/PKA-dependent paradoxical activation of Raf/MEK/ERK signaling in polycystin-2 defective mice treated with sorafenib. Hepatology. 2012. PMID: 22653837
13. Borthakur G et al. Sorafenib induces paradoxical phosphorylation of the extracellular signal-regulated kinase pathway in AML cells lacking FLT3-ITD. Leuk Lymphoma. 2015. PMID: 25665465
14. Bruner JK et al. Adaptation to TKI treatment reactivates ERK signaling in tyrosine kinase-driven leukemias and other malignancies. Cancer Res. 2017. PMID: 28923853
15. Romaniello D et al. A combination of approved antibodies overcomes resistance of lung cancer to osimertinib by blocking bypass pathways. Clin Cancer Res. 2018. PMID: 29967248
16. Hata AN et al. CDK4/6 signaling attenuates the effect of EGFR TKIs in EGFR-mutant NSCLC. Transl Lung Cancer Res. 2024. PMID: 38025818