Comparative biomarker analysis of PALOMA-2/3 trials for palbociclib

Abstract

While cyclin-dependent kinase 4/6 (CDK4/6) inhibitors, including palbociclib, combined with endocrine therapy (ET), are becoming the standard-of-care for hormone receptor-positive/human epidermal growth factor receptor 2‒negative metastatic breast cancer, further mechanistic insights are needed to maximize benefit from the treatment regimen. Herein, we conducted a systematic comparative analysis of gene expression/progression-free survival relationship from two phase 3 trials (PALOMA-2 [first-line] and PALOMA-3 [≥second-line]). In the ET-only arm, there was no inter-therapy line correlation. However, adding palbociclib resulted in concordant biomarkers independent of initial ET responsiveness, with shared sensitivity genes enriched in estrogen response and resistance genes over-represented by mTORC1 signaling and G2/M checkpoint. Biomarker patterns from the combination arm resembled patterns observed in ET in advanced treatment-naive patients, especially patients likely to be endocrine-responsive. Our findings suggest palbociclib may recondition endocrine-resistant tumors to ET, and may guide optimal therapeutic sequencing by partnering CDK4/6 inhibitors with different ETs. Pfizer (NCT01740427; NCT01942135).

Fig. 1. Overview of comparative PALOMA-2/3 biomarker analysis described in this study.

ET endocrine therapy; GE gene expression; PAL palbociclib; PCA principal component analysis; PFS progression-free survival; MBC metastatic breast cancer; Tx treatment.

Fig. 2. Lack of correlation in predictive biomarker pattern between PALOMA-2 and PALOMA-3 appears driven by poor concordance in the ET arm.

Each data point corresponds to a gene on the EdgeSeq Oncology panel. R-value in the plots refers to Pearson correlation. a Comparison of gene expression/treatment effect interaction (dependency) in predicting PFS across the two trials. Plotted are interaction coefficient values (i.e. ln(Hazard Ratio)). b Comparison of gene expression/PFS association within each treatment arm (ET, left panel; ET plus palbociclib, right panel) across the two trials. ET endocrine therapy; PAL palbociclib; PFS progression-free survival.

Fig. 3. Relative consistency within the combination arm indicates a likely conserved mechanism in palbociclib response regardless of prior lines of treatment.

Shown are genes whose expression are associated with PFS from treatment with palbociclib plus ET in both PALOMA-2 and PALOMA-3 (nominal P ≤ 0.05, top panels) and their enriched hallmark pathways (FDR ≤ 0.25, bottom panels) where orange bars highlight those with FDR ≤ 0.05. a Sensitivity genes whose higher expression is associated with longer PFS. b Resistance genes whose higher expression is associated with shorter PFS. ET endocrine therapy; FDR false discovery rate; PFS progression-free survival.

Fig. 4. The addition of palbociclib may recondition endocrine responsiveness through concerted actions on CDK4 and ER signaling networks.

a Hierarchical clustering analysis of gene expression/PFS association across the four treatment arms from PALOMA-2 (II) and PALOMA-3 (III). Distance metric was defined using 1 minus correlation coefficient; average linkage was used for clustering. b PCA of global gene expression/PFS association by molecular subtype. LumA or LumB subtype classification is represented by shape, treatment arm represented by color, and cohort represented by number in label (“2” for PALOMA-2 and “3” for PALOMA-3). The percentage values on the axes refer to the proportion of total variance accounted for by the first two principal components respectively. c The expression (top panel) and activity (bottom panel) of ER after palbociclib treatment in ER + cells. Transcript-level change in ESR1 on vehicle control or palbociclib at Day 1 or Day 7. GSEA enrichment plots for ER regulon (estrogen response genes) from MCF7 at Day 1 or Day 7. Genes were rank ordered from most downregulated by palbociclib (vs vehicle control) on the left to most upregulated on the right; each vertical line corresponds to a signature gene. ER estrogen receptor; ET endocrine therapy; FDR false discovery rate; FUL fulvestrant; GSEA gene set enrichment analysis; LET letrozole; LumA luminal A; LumB luminal B; NES normalized enrichment score; PAL palbociclib; PC principal component; PCA principal component analysis; PFS progression-free survival.

Fig. 5. PALOMA-2–specific associations point to a potentially stronger role of the T-cell–inflamed TME in mediating palbociclib resistance in the front-line setting.

a Gene expression/PFS association result (numeric value and color intensity represent statistical significance of the association; red = resistance, green = sensitivity) of enriched interferon-gamma response and PD1 signaling genes from palbociclib plus ET treatment in PALOMA-2, PALOMA-3, and the postmenopausal patient/metastatic sample subset of PALOMA-3, respectively. b Relationship of PFS and the expression of a T-cell–inflamed TME signature that has been shown to be a pan-tumor predictive biomarker for immune checkpoint inhibitor response in the clinic in PALOMA-2 (top panel) and PALOMA-3 (bottom panel) cohorts from median (left panel) and quartile (right panel) analyses. c Association between PFS and the expression of various immune cell types from palbociclib plus ET treatment. Shown is statistical significance in the direction of resistance (top) or sensitivity (bottom); blue lines mark P-value of 0.05. ET endocrine therapy; MDSC myeloid-derived suppressor cell; Mets metastatic; PFS progression-free survival; TME tumor microenvironment.