CTNI-23 - Single cell transcriptomics, pharmacokinetics, and pharmacodynamics of combined CDK4/6 and mTOR inhibition in a phase 0 trial of recurrent high-grade glioma

Friday, November 17, 2023

4:55 PM – 5:00 PM PT

Location: Room 109-110

Oral Abstract Presenter(s)

Kevin Johnson

Yale University School of Medicine
New Haven, Connecticut, United States

Author(s)

An-Chi Tien

Ivy Brain Tumor Center
Phoenix, Arizona, United States
JJ

Jun Jiang

Karmanos Cancer Institute
Detroit, Michigan, United States
James McNamara

Ivy Brain Tumor Center
Phoenix, Arizona, United States
YC

Yu-Wei Chang

Ivy Brain Tumor Center
Phoenix, Arizona, United States
CM

Chelsea Montgomery

Ivy Brain Tumor Center
Phoenix, Arizona, United States
Anita DeSantis

Ivy Brain Tumor Center
Phoenix, Arizona, United States
LE

Leonel Elena

Ivy Brain Tumor Center
Phoenix, Arizona, United States
YF

Yoko Fujita

Ivy Brain Tumor Center
Phoenix, Arizona, United States
SK

Seongho Kim

Karmanos Cancer Institute
Detroit, Michigan, United States
Artak Tovmasyan

Barrow Neurological Institute
P, Arizona, United States
Jing Li

Karmanos Cancer Institute, Wayne State University
Detroit, Michigan, United States
Shwetal Mehta

Ivy Brain Tumor Center
Phoenix, Arizona, United States
RV

Roel Verhaak

Yale University School of Medicine
New Haven, Connecticut, United States
Nader Sanai

Barrow Neurological Institute
Phoenix, Arizona, United States

Adult High-grade gliomas (HGGs) often exhibit deregulation of RB-CDK4/6 and mTOR signaling pathways. This study investigated tumor single cell transcriptomics, pharmacokinetics, and pharmacodynamics immediately following combined CDK4/6 (ribociclib) and mTOR (everolimus) inhibition in recurrent HGG. 24 patients with recurrent gliomas harboring RB+, CDKN2A/B deletion or CDK4/6 amplification, and PTEN loss or PIK3CA mutations were enrolled in a phase 0 trial. Patients received ribociclib treatment for 5 days before surgery, including six who received ribociclib (400mg QD) plus everolimus (2.5mg QD), and six dose-escalation cohorts (n=3 each) reaching ribociclib 600mg QD plus everolimus 70mg QW. No dose-limiting toxicities were observed. Validated LC-MS/MS methods determined total and unbound drug concentrations. Preoperative MRI revealed a marked decrease in gadolinium-enhancement in all patients. In gadolinium-nonenhancing tumor regions, the median unbound concentration of ribociclib was 561 nM, while no unbound everolimus concentrations were detected. Ribociclib’s pharmacodynamic effects in surgical specimens were compared with matched archival tissue using IHC. Across IDH-wild-type tumors, there was significant decrease in MIB1 levels between the matched archival and surgical specimens (p < 0.05) suggesting drug-associated cell cycle inhibition. Single nuclei RNAseq (snRNA) analyses were performed on 17 IDH-wild-type recurrent samples treated with ribociclib plus everolimus and 29 IDH-wild-type recurrences treated with standard therapies. Cellular state analysis based on snRNAseq data demonstrated a decreased neural progenitor-like (NPC-like) malignant cell state compared with standard therapies (p < 0.01), indicating inhibition of the CDK4-driven neural progenitor-like cell state. This observation was validated using snRNA and patient-derived glioma cell lines exposed to ribociclib, showing a consistent decrease in NPC-like proportions and cycling cells (p < 0.05). This study highlights the value of integrating pharmacokinetics/pharmacodynamics and snRNA data to assess treatment effects in phase 0 surgical tissues. These findings support further development of ribociclib, but not everolimus, for the treatment of glioma patients.