Phase I/II Study of the Aurora Kinase A Inhibitor Alisertib and Pembrolizumab in Refractory, Rb-Deficient Head and Neck Squamous Cell Carcinomas

Abstract

Purpose: Effective therapy for recurrent head and neck squamous cell carcinoma (HNSCC) that is refractory to chemotherapy and immunotherapy is a considerable need. Aurora kinase A inhibition leads to apoptosis and immunogenic cell death in preclinical models of human papilloma virus (HPV)-driven cancers.

Patients and methods: Alisertib was administered orally twice daily on days 1-7 and pembrolizumab on day 1 of a 21-day cycle to adults with advanced solid tumors (phase I) or with immunotherapy- and platinum-resistant, HPV-positive HNSCC (phase II).

Results: The recommended phase II alisertib dose was 40 mg, which had only the expected toxicity including cytopenia that led to dose reductions in two phase II patients at cycles 13 and 16. We saw no objective responses, but the combination led to prolonged stable disease (SD) in several patients, including two of 10 phase I patients (8 and 27 months). Eight of the 15 HPV-positive patients had SD, of which four (heavily pretreated) had ≥6 months, with median overall and progression-free survival durations of 16.8 and 1.4 months, respectively. In circulating immune cells and plasma, patients with SD had markedly higher levels of HLA de novo resistance-expressing NK cells than did progressive disease patients who demonstrated a more immunosuppressive and inflammatory profile. Pharmacokinetics did not indicate any significant drug-drug interactions between pembrolizumab and alisertib.

Conclusions: The combination of alisertib and pembrolizumab was well tolerated and led to prolonged SD in some immunotherapy-resistant patients, supporting our hypothesis that Aurora kinase A inhibition can reverse immunotherapy resistance of retinoblastoma protein-deficient HNSCC.

Figure 1.

Best percent change in tumor size from baseline in patients enrolled in the phase 1 (A) and phase 2 (B) studies. Patient 3 met all the eligibility criteria for the phase 2 study and was included in the efficacy analysis for the phase 2 study. (C-D) Kaplan-Meier PFS (C) and OS (D) curves for all patients enrolled in the phase 2 study.

Figure 2.

Sagittal (A) and axial (B) baseline postcontrast neck computed tomography scans of a patient with recurrent squamous cell carcinoma in the anterior midline residual oral cavity in the setting of pre-existing extensive posttreatment anatomic distortion. (C and D) Follow-up images obtained after 1 year of therapy with alisertib and pembrolizumab demonstrating a qualitative reduction in the size of the enhancing mass. The tumor is denoted by the red arrows.

Figure 3.

HPV cfDNA levels and tumors size in nine patients that were measured longitudinally. The first column includes patients with a PFS of ≥ 4months and the second column includes those with a PFS , 4 months. HPV cfDNA levels were measured using ddPCR and normalized to baseline values. Tumor sizes were measured using RECIST and graphed as percent change from baseline.

Figure 4.

Flow cytometry analyses for the frequencies of circulating lymphocytes expressing the activation marker HLA-DR. PBMC isolated from the blood samples collected at the baseline from the different patients with progressive disease (PD) or stable disease (SD) were analyzed by flow cytometry for Differences in the frequencies of total CD8 T cells (A) and functional subsets expressing granzyme B (GrnzB+), IFNγ+ or both GrnzB+ and IFNγ+ (B-D); total natural killer (NK) cells and functional subsets expressing granzyme B (GrnzB+), IFNγ+ or both GrnzB+ and IFNγ+ (E-H); along with frequencies of the immunosuppressive populations, including HLA-DR+ Tregs (I) and MDSC (J). Sample size, PD: n=9, SD: n=3. Significance was determined using an unpaired t-test with Welch’s correction, **P<0.005, ****P<0.0001. ns, not significant.

Figure 5.

Flow cytometry analyses of HLA-DR expressing effector and inhibitory suppressor immune subsets. The combined frequencies of CD8 T and NK effector cells expressing HLA-DR along with the frequencies of the immune suppressive populations of MDSC and HLA-DR expressing Tregs were compared between the PD and SD groups of patients (A). Additionally, the frequencies of the HLA-DR expressing functional subsets of NK cells and CD8 T cells were compared between the PD and SD groups of patients (B). Sample size, PD: n=9, SD: n=3. significance of differences was evaluated using a paired T test or Hotelling’s T test.