3633 - Activity of the novel Aurora kinase B inhibitor AZD2811 in biomarker-defined models of small cell lung cancer

Background

Aurora Kinases (AURK) regulate mitosis and are often upregulated in cancer. SCLC is characterized by TP53 and RB1 loss and frequent MYC amplification. Our group, and others, have proved that MYC-driven SCLC are vulnerable to AURKA-inhibitors (AURKAi), but their use is limited by toxicity due to high AURKA expression. AURKB levels are variable in SCLC, making AURKB an attractive targeted therapeutic approach. A novel AURKBi, AZD2811NP (nanoparticle), is now being investigated in relapsed SCLC patients (NCT02579226). We hypothesize that molecularly defined subsets may be sensitive to AZD2811.

Methods

We tested 50 human-derived SCLC cell lines with AZD2811 in 96 hour proliferation assays. To identify translatable biomarkers of response, we correlated IC50 values with genomic (WES), transcriptomic (RNASeq), and proteomic profiling (RPPA).

Results

AZD2811 is active in a subset of SCLC cells: 13/30 (26%) have high sensitivity (IC50<30 nM, Cmax) and 7/30 (14%) intermediate (IC50=30-100nM). Comparing protein expression, we found that cMYC (Fold change, FC:2.5; P = 0.015) and Vimentin (FC:1.66; P = 0.022) were top biomarkers of sensitivity, while high E-cadherin (FC:1.94; P = 0.046) and BCL-2 (FC:1.86; P = 0.032) associated with resistance. Interestingly, among the 13 highly sensitive cells, only 5 overexpress cMYC, highlighting that activity of AURKBi is not limited to MYC-driven SCLC. Indeed, among recently-described transcription factor defined subsets of SCLC, a majority of the NEUROD1-defined subset (cMYC enriched) were sensitive to AZD2811, as 25% of ASCL1- and 67% of POU2F-3-defined ones. These data suggest that the ASCL1 subset includes heterogeneous sub-groups that require further analysis to dissect their molecular features. Mutations in EP300, a Histone acetyltransferase that controls chromatin modification and transcription, also predict sensitivity.

Conclusions

Our results show encouraging single agent in vitro activity of AZD2811 in SCLC cells, and suggest a novel biomarker-driven approach for combinations. Candidate biomarkers will be tested in samples from the ongoing clinical trial. The high BCL2 levels observed in resistant cells provide a rationale for exploring AURKB/BCL2-i combinations.

Clinical trial identification

NCT02579226.

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

AstraZeneca.

Disclosure

C.A. Adelman: Full / Part-time employment: AstraZeneca. J. Travers: Full / Part-time employment: AstraZeneca. J. Heymach: Advisory / Consultancy, Research grant / Funding (self): AstraZeneca; Advisory / Consultancy, Licensing / Royalties: BioTree; Advisory / Consultancy: Boehringer Ingelheim; Advisory / Consultancy: Exelixis; Advisory / Consultancy: Genentech; Advisory / Consultancy: GSK; Advisory / Consultancy: Guardant Health; Advisory / Consultancy: Hengrui; Advisory / Consultancy: Lilly; Advisory / Consultancy: Novartis; Advisory / Consultancy, Research grant / Funding (self), Licensing / Royalties: Spectrum; Advisory / Consultancy: EMD Serono; Advisory / Consultancy: Synta; Research grant / Funding (self): Bayer; Research grant / Funding (self): GlaxoSmithKline. L.A. Byers: Advisory / Consultancy, Research grant / Funding (self): AstraZeneca; Advisory / Consultancy, Research grant / Funding (self): AbbVie; Advisory / Consultancy, Research grant / Funding (self): GenMab; Advisory / Consultancy: BergenBio; Advisory / Consultancy: Pharma Mar; Advisory / Consultancy: SA; Advisory / Consultancy, Research grant / Funding (self): Sierra Oncology; Research grant / Funding (self): Tolero Pharmaceuticals. All other authors have declared no conflicts of interest.