Lung Cancer Organoids Link a Basal-Lineage Shift to EGFR Resistance and Context-Dependent CDK4/6 Sensitivity

A Clinically Annotated Resource for Resistant Lung Cancer

Keio University researchers and collaborators established the EGFR-mutant Lung Cancer Organoid Library, or ELCOL, to investigate resistance to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors. The collection contains 39 organoid lines from 31 patients, including 28 patients previously treated with an EGFR inhibitor. Three patients supplied serial samples, and the collection included several forms of treatment-associated histological transformation.

All lines propagated for more than six months. Exome analysis confirmed their parental EGFR mutations, while histology and xenografts supported retention of major tumor identities. Seven-day ATP-based drug testing also distinguished clinically resistant and responsive sources; a pretreatment/post-treatment organoid pair illustrated acquired osimertinib resistance within one patient.

Known secondary mutations or amplifications explained only part of the resistant collection. The study therefore examined gene expression, chromatin accessibility and experimentally altered lineage regulators to investigate resistance beyond conventional mutation-based explanations.

Basal-Shift Tumors Combine Lineage Features

Bulk transcriptomes identified an intermediate population between conventional lung adenocarcinoma and squamous carcinoma. These resistant organoids had acquired expression of basal or squamous-associated genes such as KRT5, KRT14 and TP63 while retaining some alveolar or adenocarcinoma identity. The researchers termed this phenotype basal-shift adenocarcinoma.

Single-cell analysis of eight selected lines and staining of single-cell-derived organoids refined that description. Intermediate cultures contained variable expression of the two lineage programs, and simultaneous expression of selected markers was rare. A basal-shift organoid therefore need not be a uniform population in which every cell expresses both identities. Related marker patterns were retained in xenografts and observed in patient tissue.

The basal-shift group frequently had CDKN2A/B loss, but that alteration alone was insufficient: some organoids carrying the loss remained sensitive to osimertinib, and engineered TP53/CDKN2A-deficient alveolar organoids did not acquire basal-shift identity. Increasing TP63 alone also failed to confer osimertinib resistance in the tested tumor lines.

NKX2-1 Loss Supplies Functional Evidence

Basal-shift cultures showed reduced activity of NKX2-1, a regulator of alveolar identity. Knocking out NKX2-1 in three treatment-naïve EGFR-mutant organoid lines, followed by one month of osimertinib exposure, increased basal-associated expression and drug resistance. Additional NKX2-1 knockout in TP53/CDKN2A-deficient normal alveolar organoids from two donors also shifted their identity and growth-factor dependence.

The resulting cultures could grow with reduced dependence on EGF, IGF-1 and FGF-2. In resistant basal-shift organoids, osimertinib still reduced EGFR and ERK phosphorylation without suppressing growth. Those results distinguish inhibition of the intended signaling pathway from dependence on that pathway for proliferation.

A Drug Vulnerability Depends on Genetic Background

A seven-day screen of 54 compounds compared three basal-shift and two non-basal-shift lines. CDK4/6 inhibitors, including palbociclib and abemaciclib, preferentially slowed growth in the basal-shift group. Secondary testing connected sensitivity more closely to CDKN2A/B loss within EGFR-mutant cultures than to basal-shift identity alone. The response was largely cytostatic.

The same rule did not transfer uniformly to KRAS-mutant lung adenocarcinoma: two of three KRAS-mutant lines with CDKN2A/B loss resisted palbociclib. In four selected xenograft models, three-week palbociclib treatment suppressed the tested CDKN2A/B-loss tumors, including basal-shift and squamous-transformed models, while the non-basal-shift wild-type comparison did not respond.

The clinical marker definition required fewer than 5% NKX2-1-positive cells, TAp63 or ΔNp63 expression, and retained AQP5. Re-biopsy evaluation included 32 samples obtained after resistance to first- or second-generation EGFR inhibitors and 21 after osimertinib resistance. Four basal-shift cases were identified across the combined 53-sample cohort.

The Discussion separately reports basal-shift in 9.5% of the osimertinib-resistant clinical subset, compared with five of 19 osimertinib-resistant organoid lines. These are distinct denominators from the combined re-biopsy cohort. The higher frequency in the organoid subset raises possible culture-selection effects and does not establish population prevalence. The study supports functional modeling and a preclinical drug hypothesis; it does not validate CDKN2A/B loss for clinical treatment selection or demonstrate patient benefit from CDK4/6 inhibition.