Abstract PS5-01: Biomarkers of resistance to palbociclib in ER+ primary breast cancer in the PALLET trial

Abstract

Background: CDK4/6 inhibitors are being used in combination with aromatase inhibitors as therapy for advanced ER+ breast cancer (BC) and are being explored for use in primary BC. Few mechanisms driving resistance to added CDK4/6 inhibitors have been defined. The PALLET phase II randomized neoadjuvant trial of letrozole (LET) ± palbociclib (PALBO) in postmenopausal ER+HER2- primary BC reported that clinical response rate over 14wks was not significantly increased by adding PALBO to LET but suppression of Ki67 was significantly increased (Johnston et al, JCO 2018, 37, 178): after 14wks complete cell cycle arrest (CCCA, Ki67<2.7%) was present in 59% on LET and 90% on LET+PALBO. Here we sought to identify biomarkers of de novo resistance to allow selection of patients most likely to benefit from added PALBO. Methods: 307 patients were randomized to LET (n=103) or LET+PALBO (n=204) for 14 wks. The first 2wks of LET+PALBO patients were randomised to LET, PALBO, or LET+PALBO. Biopsies were taken at baseline, 2wks and 14wks. Biomarker data are presented here for baseline only, other than Ki67 at both baseline and 14wks. IHC analyses were conducted on FFPE biopsies for ER, PgR, RB1, cyclin-E1, and cyclin-D1 (also FISH). RNA-seq was performed on fresh frozen biopsies. Association of each biomarker with CCCA was determined by logistic regression. Differentially expressed genes (DEGs) were identified between patients sensitive (CCCA) (n=94) and resistant (non-CCCA) (n=10) to treatments with or without PALBO at 14wks by DESeq2. Fifty hallmark gene sets were tested for significant enrichment with DEGs and differential gene sets were identified by using Gene Set Enrichment Analysis (GSEA). Results: The association of IHC biomarkers with CCCA is shown in the table. Lower levels of ER, higher levels of cyclin-E1, and amplification of cyclin-D1 were each significantly associated with a greater chance of non-CCCA with LET+PALBO. High cyclin-E1 levels were also associated with reduced chance of CCCA with LET only. Patients with high baseline Ki67 also exhibited higher non-CCCA with LET+PALBO at 14wks (p=0.0002). In the RNAseq data we identified 1973 DEGs between the 14wk CCCA and non-CCCA patients for LET+PALBO. E2F and MYC targets, PI3K/AKT/MTOR signalling and interferon response gene sets were among the hallmark gene sets enriched for genes with higher expression in non-CCCA patients at 14wks for LET+PALBO (FDR<0.05). For LET-only, 311 DEGs were identified and the “Estrogen Response Early” gene set was significantly enriched in genes with higher expression in CCCA samples. At the individual gene level, genes significantly associated with non-CCCA after 14wks LET+PALBO included CCNE1, CDK2 and several E2F-related genes (p<0.05). Their expression was not significantly different between non-CCCA and CCCA patients with LET alone. Conclusion: Biomarkers associated with response/resistance to added PALBO were different from LET only. PALBO resistance was associated with higher baseline expression of cyclin-E1 (both IHC and RNA), CDK2, and genes related to E2F, MYC, interferon and MTOR signalling. These results suggest that multiple identifiable mechanisms of de novo resistance to PALBO are likely to exist in primary ER+ BC. On-going WES analyses will allow the significance of alterations at the DNA level to be presented. Table 1.Continuous measurement in a logistic regression for CCCA at 14 weeks; Oddsratio calculated separately for group A and groups B,C,D and were adjusted forregion (UK vs NA); * amplified vsnon-amplifiedContinuous measurement in a logistic regression for CCCA at 14 weeksBiomarkerLET LET+PALBOOdds ratio95% CIpOdds ratio95% CIpER1.120.36, 3.480.844.471.62, 12.380.004PgR4.381.03, 18.580.053.050.50, 18.560.23RB13.010.24, 38.560.400.420.05, 38.490.83Cyclin-E10.100.01, 0.840.030.020.00, 0.200.001Cyclin-D1 IHC3.090.51, 18.490.222.560.28, 23.330.40CyclinD1 FISH*1.470.43, 4.990.530.280.06, 0.860.03 Citation Format: Eugene F Schuster, Hui Xiao, Elena Lopez-Knowles, Lucy Kilburn, Mothaffar Rimawi, David A Wheeler, Katherine Pogue-Geile, Yuan Lui, Samuel A Jacobs, Chet Cornman, Shannon Puhalla, Maggie Cheang, Judith Bliss, Stephen Johnston, Mitch Dowsett, On behalf of the PALLET Trialists. Biomarkers of resistance to palbociclib in ER+ primary breast cancer in the PALLET trial [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS5-01.

Acquired resistance to CDK4/6 inhibition is associated with dysregulation of multiple pathways including cyclin D-CDK4/6-RB, EGFR/HER, AKT/mTOR and IFN signaling

“…Clinically, ISG signatures were significantly enriched in tumours intrinsically resistant to neoadjuvant palbociclib or abemaciclib in combination with anastrozole (NeoPalAna, Neo-MONARCH) 37 . Subsequent analysis of two separate trials suggests that ISGs are also enriched in tumours resistant to neoadjuvant palbociclib plus letrozole (PALLET) 51,52 , as well as palbociclib plus fulvestrant in metastatic disease after progression on endocrine therapy (PALOMA-3) 53 . Using independently derived models, our results provide further support that IFN signaling is associated with both de novo and acquired resistance to CDK4/6i invitro.…”