A Phase 0 Trial of Ribociclib in Recurrent Glioblastoma Patients

Published OnlineFirst July 8, 2019; DOI: 10.1158/1078-0432.CCR-19-0133

Clinical Trials: Targeted Therapy Clinical Cancer Research A Phase 0 Trial of Ribociclib in Recurrent Glioblastoma Patients Incorporating a Tumor Pharmacodynamic- and Pharmacokinetic-Guided Expansion Cohort An-Chi Tien1, Jing Li1,2, Xun Bao2, Alanna Derogatis1, Seongho Kim2, Shwetal Mehta1, and Nader Sanai1

Abstract

Background: CDK4/6-dependent cell-cycle regulation is and enhancing tumor regions were 0.374 mmol/L, 0.560, disrupted in most glioblastomas. This study assesses the and 2.152 mmol/kg, respectively, which were more than central nervous system (CNS) pharmacokinetics and tumor 5-fold the in vitro IC50 for inhibition of CDK4/6 pharmacodynamics of ribociclib, a highly selective CDK4/6 (0.04 mmol/L). G1-to-S phase suppression was inferred inhibitor, in patients with recurrent glioblastoma. by decreases in phosphorylation of RB (P < 0.01) and Methods: Patients with recurrent glioblastoma with intact cellular proliferation (P < 0.05). Six of 12 patients were retinoblastoma protein (RB) expression and CDKN2A dele- enrolled into the pharmacokinetic/pharmacodynamic- tion or CDK4/6 amplification were treated with ribociclib guided expansion cohort and demonstrated a median daily (900 mg) for 5 days before tumor resection. Blood, PFS of 9.7 weeks. Examination of recurrent tumors follow- tumor, and cerebrospinal fluid (CSF) samples were collected, ing monotherapy indicated upregulation of the PI3K/ and total and unbound ribociclib concentrations were deter- mTOR pathway. mined. Pharmacodynamic effects, assessed by RB and FOXM1 Conclusions: Ribociclib exhibited good CNS penetra- phosphorylation, were compared with matched archival tis- tion, and target modulation was indicated by inhibition of sue. Patients with positive pharmacokinetic and pharmaco- RB phosphorylation and tumor proliferation. Ribociclib dynamic effects were enrolled into the expansion cohort for monotherapy showed limited clinical efficacy in patients preliminary assessment of progression-free survival (PFS). with recurrent glioblastoma. Combination therapy with Results: Twelve patients were enrolled. The mean CDK4/6 and PI3K/mTOR inhibitors may be explored for unbound ribociclib concentrations in CSF, nonenhancing, treating recurrent glioblastoma.

Introduction amplification of CDK6 (2). Overall, the CDK4/6–Rb–E2F axis is deregulated in approximately 80% of glioblastomas, under- A hallmark of human glioblastoma is aberrant cell-cycle con- scoring the clinical potential of cell-cycle targeting in this patient trol, resulting in unlimited cell-cycle reentry and progression population. Preclinical studies demonstrated that tumor cells (1–3). The CDK4/6–Rb–E2F axis controls the cell cycle and is treated with CDK4/6 inhibitors exhibited reduction of phos- tightly regulated by several factors such as cyclin D, INK4 family phorylation of RB, G -phase arrest, and decreased proliferation, proteins, p21CIP1, and p27KIP1. According to The Cancer Genome 1 and moreover, the inhibition of CDK4/6 activity slowed glioma Atlas database, homozygous deletion of the p16INK4a gene progression (4). occurs in 50% of glioblastomas (1). Other common cell-cycle– Ribociclib (LEE011) is a highly specific, orally bioavailable, related mutations in glioblastomas include amplification and small-molecule inhibitor of CDK4 and CDK6 (5, 6). In preclinical overexpression of CDK4 and homozygous deletion/mutation of and clinical studies, ribociclib demonstrated a manageable tol- RB, followed in frequency by overexpression of cyclin D1 and erability and therapeutic potential for a variety of cancer types (5). Ribociclib was approved by the FDA in 2017 for use in combination with an aromatase inhibitor (e.g., letrozole) for the treat- 1Ivy Brain Tumor Center, Barrow Neurological Institute, Phoenix, Arizona. 2Karmanos Cancer Institute, Wayne State University School of Medicine, Detroit, ment of postmenopausal women with advanced or metastatic Michigan. breast cancer that is hormone receptor (HR) positive and HER2 negative (5, 6). The efficacy and safety of single-agent ribociclib in Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). patients with neuroblastoma was evaluated as part of the phase I study in pediatric patients, where ribociclib demonstrated an Corresponding Author: Nader Sanai, Barrow Neurological Institute, acceptable toxicity profile and promising antitumor activity (7). 2910 N. 3rd Avenue, Phoenix, AZ 85013. Phone: 602-708-7728; E-mail: – [email protected] Preclinical data suggest that ribociclib can cross the blood brain barrier (BBB), supporting further clinical development for the Clin Cancer Res 2019;XX:XX–XX treatment of central nervous system (CNS) tumors (8). doi: 10.1158/1078-0432.CCR-19-0133 Phase 0 clinical trials are commonly defined as first-in-human 2019 American Association for Cancer Research. studies with no therapeutic or diagnostic intent, a limited number

www.aacrjournals.org OF1

Tien et al.

Table 1. Patient demographics and clinical characteristics Translational Relevance Characteristics The RB–CDK4/6–E2F signaling axis is frequently deregu- Sex (male/female) 8/4 – lated in glioblastoma, making it an ideal pathway to block Age (years) 49 (31 66) Weight (kg) 85 (63–109) tumor growth. In this phase 0 study, we assessed the CNS Height (cm) 174 (154–193) penetration and tumor pharmacodynamics of ribociclib, a ECOG/Zubrod performance status, n (%) CDK4/6 inhibitor, in patients with recurrent glioblastoma. 0, n (%) 3 (25%) Using an integrated approach involving quantitative analysis 1, n (%) 7 (58%) of drug penetration in the enhancing and nonenhancing 2, n (%) 2 (17%) tumor regions and downstream target inhibition, we provide Extent of resection fi GTR, n (%) 8 (67%) the rst clinical evidence for good CNS penetrance of riboci- STR, n (%) 4 (33%) clib. The pharmacodynamic- and pharmacokinetic-guided Unknown, n (%) 0 (0%) expansion cohort study indicates limited clinical efficacy of Prior bevacizumab, n (%) 0 (0%) ribociclib monotherapy. Longitudinal analysis suggests com- Timing of ribociclib, n (%) bination therapy with ribociclib and an mTOR inhibitor may First progression 8 (67%) provide therapeutic benefit in patients with recurrent Second progression 1 (8%) Third progression 2 (17%) glioblastoma. Fourth progression 1 (8%) Alanine aminotransferase (IU/L) 17.5 (10–39) Aspartate aminotransferase (IU/L) 16 (10–115) Total bilirubin (mg/dL) 0.6 (0.3–1) of patients, and microdosing of the experimental agent (9, 10). Serum albumin 4.2 (3.6–4.8) – For brain tumor patients, however, the utility of these design Serum creatinine 0.93 (0.67 3.8) elements is hampered by the BBB, a lack of predictive animal models, and the significant risks of tumor tissue acquisition. Here, we adapt the phase 0 strategy (11, 12) by using "limited thera- Study design and drug administration peutic" dosing (5 days of the maximally tolerated dose) instead of This investigational, phase 0, open-label, nonrandomized trial microdosing and uses matched archival controls (instead of pre- with a pharmacokinetic/pharmacodynamic-guided expansion and posttreatment biopsies) to assess pharmacodynamic effects. cohort was conducted at the Ivy Brain Tumor Center at the Barrow To mitigate the ethical and accrual challenges of a nontherapeutic Neurological Institute in Phoenix, Arizona, in partnership with study, we incorporate a pharmacokinetic- and pharmacodynam- Karmanos Cancer Institute. The study (NCT02933736) received ic-guided "trigger" that graduates phase 0 patients into a thera- approval from the institutional review board and was conducted peutic expansion cohort. in accordance with the Declaration of Helsinki and Good Clinical Here, to explore the utility of ribociclib in treating patients Practice. Patients were accrued at the Barrow Neurological Insti- with recurrent glioblastoma, we completed a phase 0 clinical tute and informed written consent was obtained from each trial with an expansion cohort for patients with recurrent glio- patient in the study. blastoma to examine pharmacokinetic and pharmacodynamic Enrolled phase 0 patients were administered 900 mg/day of endpoints following a 5-day exposure to ribociclib. This study ribociclib for 5 days before planned cranial tumor resection. was designed to (i) provide in vivo pharmacokinetic data for Patients were assigned to 3 time-escalation arms in which tumor nonenhancing and enhancing tumor tissue, (ii) identify the resection was performed at 2, 8, or 24 hours, respectively, fol- molecular effects of ribociclib in glioblastoma patients, and (iii) lowing the final dose of ribociclib. Intensive blood sampling was interrogate putative mechanisms of resistance in tumor recur- performed on the fourth day of drug administration. During rences. Positive results from the trial would accelerate our devel- tumor resection, blood, cerebrospinal fluid (CSF), and tumor opment of this combinatorial drug regimen and enable a larger- samples from contrast-enhancing and nonenhancing regions scale efficacy study, whereas negative results would enhance our (based on preoperative MRI and intraoperative neuronavigation) understanding of tumor resistance and potentially identify an were collected for pharmacokinetic and pharmacodynamic anal- additional agent for combination drug therapy. yses. We determined the maximum tolerated dose of ribociclib as 900 mg daily. To ensure tolerability of continuous treatment of ribociclib, patients enrolled into the expansion cohort were Materials and Methods treated with a lower dose at 600 mg daily for 21 days on and Patient selection 7 days off until disease progression was noted based on Response Study patients were older than 18 years with a recurrent WHO Assessment in Neuro-Oncology (RANO) criteria. grade IV glioma. Prior to accrual, all patients had undergone resection of their primary tumor, followed by standard-of-care Statistical methods Stupp regimen with temozolomide and fractionated radiothera- The primary objectives of this trial were to evaluate phar- py. The median time from initial diagnosis to recurrence was macokinetic and pharmacodynamic endpoints. The secondary 271 days, whereas the median time from completion of radiation objectives were to explore progression-free survival (PFS) and to recurrence was 200 days. Based on radiographic and/or clinical overall survival (OS) among patients who exhibited positive evidence of disease progression, all patients were previously pharmacokinetic and pharmacodynamic responses. Phase 0 scheduled for surgical re-resection, had an Eastern Cooperative patients were determined to be responders if both the un- Oncology Group performances status of <2, and exhibited ade- bound ribociclib concentration in the nonenhancing tumor quate organ function (Table 1). exceeded 5-fold of the in vitro IC50 for CDK4/6 kinase activity

OF2 Clin Cancer Res; 2019 Clinical Cancer Research

Phase 0 of Ribociclib for Glioblastoma

(IC50 ¼ 0.04 mmol/L; ref. 6) and if we observed a >20% Pharmacokinetic evaluation decrease in pRBþ cells compared with matched archival tissue. Blood pharmacokinetic samples were collected from each Only phase 0 responders were eligible to enroll into the patient at predosing, 0.5, 1, 2, 4, 6, 8, and 24 hours after the pharmacokinetic/pharmacodynamic-guided expansion cohort administration of the fourth presurgical dose of ribociclib. Plasma of the trial. Pharmacokinetic/pharmacodynamic-guided ex- was separated from whole blood by centrifugation (at 4C, pansion cohort patients were treated with ribociclib until 1,500 g for 10 minutes), and plasma samples were stored at disease progression. PFS was defined as the duration from the 80C until analysis. Tumor resection was performed after the date of recurrent tumor resection to the date of disease pro- administration of the fifth presurgical dose of ribociclib. Blood, gression or death from any cause, whichever occurred first. tumor (including contrast-enhancing and nonenhancing OS was defined as the duration from the date of recurrent regions), and CSF samples were collected intraoperatively at tumor resection to the date of death. The pharmacokinetic/ predefined time points (i.e., 2, 8, or 24 hours after the fifth pharmacodynamic-guided expansion cohort study was explor- presurgical dose). Two tumor samples (0.5 cm3), corresponding atory for efficacy, and no formal statistical comparison was to enhancing- and nonenhancing glioblastoma, were collected planned. Thus, the sample size was justified based on feasibi- intraoperatively from each patient and divided into 2 equal lity. For each of 3 time-escalation arms, 4 patients were allo- portions. Specimen locations were recorded with intraoperative cated and, if no pharmacodynamic effects were observed in the MRI neuronavigation system. Each tumor sample was immedi- setting of adequate drug penetration, time-escalation arms ately rinsed with ice-cold PBS to remove residual blood, then couldbeexpandedto8patientstoaccommodategreater dried and snap frozen in liquid nitrogen. The total concentrations molecular diversity. Thus, the planned sample size for phase of ribociclib in plasma, tumor, and CSF samples were determined 0 was 12 to 24 patients. Comparisons between 2 paired data using a validated liquid chromatography with tandem mass and 2 unpaired data were carried out using a paired t test spectrometry method (13). The fraction of unbound ribociclib and an unpaired t test, respectively, after data transformation. in plasma and tumor tissues were determined by equilibrium The distributions of survival outcomes (PFS and OS) were dialysis, and unbound drug concentration was calculated as the graphically summarized using Kaplan–Meier curves, and their product of total concentration and fraction unbound (13). median times and associated 95% confidence intervals were estimated using Kaplan–Meier estimates. Pharmacokinetic data analysis Plasma pharmacokinetic parameters of total and unbound Safety and assessments ribociclib were estimated from individual plasma concentra- Demographic data and medical history were recorded for all tion–time profiles using the noncompartmental analysis with study patients. Physical examination, vital signs, organ functions, Phoenix WinNonlin software (Certara). Estimated plasma phar- and other safety assessments (Eastern Cooperative Oncology macokinetic parameters, including the maximum plasma con- Group performance status, registration of concomitant medica- centration (Cmax), time to reach Cmax (Tmax), area under the curve tion, hematology, biochemistry, and urine analysis) were per- during one dosing interval (AUCt), and apparent oral clearance formed at baseline. Common Toxicity Criteria Adverse Event (CL/F, estimated from dose/AUCt) were summarized using (CTC AE) 4.0 criteria were used to document adverse events. descriptive statistics as the mean, median, range, and coefficient variation. Patient enrollment criteria with RB status and genomic Because unbound drug concentrations are pharmaco- screening logically relevant (14), CNS pharmacokinetics were evaluated Specimen samples from prior tumor resections were examined based on unbound ribociclib concentrations in CSF, none- with immunohistochemistry and array comparative genomic nhancing, and enhancing tumor regions. The extents of drug hybridization (aCGH). IHC staining for all samples was com- penetration into the CNS and tumors were assessed by the pleted on the Leica Bond, a fully automated platform using CSF-to-plasma unbound drug concentration ratio (CSF/Pu), optimized conditions that were standardized with archival glio- the tumor-to-plasma unbound drug concentration ratio blastoma tissue. Briefly, archival FFPE slides was stained with anti- (Tu/Pu), and the tumor-to-CSF unbound drug concentration RB (Cell Signaling, #9309, 1:100) for assessing the percentage of ratio (Tu/CSF). RBþ cells. The stained slides were imaged using Aperio Versa System (Leica) and analyzed using ImageScope software. In Pharmacodynamic controls parallel, the slides were also analyzed by a board-certified neu- To test the stability of proposed pharmacodynamic biomarkers ropathologist. aCGH was conducted within a CLIA-certified lab- (pRB, pFOXM1, MIB1, and cleaved caspase-3), we analyzed a oratory (MOgene LC). Genomic DNA from archival formalin- historical cohort of 10 matched primary and recurrent glioblas- fixed paraffin-embedded (FFPE) sections was extracted with toma patients who received standard-of-care Stupp regimen and DNeasy Blood and Tissue Kit (Qiagen). Quality control analysis were not enrolled in the study. FFPE tissues were stained with anti- and DNA quantification were determined using Tapestation pRB (Cell Signaling Technology; #8516, 1:400), anti-pFOXM1 (Agilent Technologies) and a Qubit 2.0 fluorometer (Thermo (Cell Signaling Technology; #14655, 1:200), anti-MIB1 (DAKO; Fisher Scientific). Genomic DNA was then labeled and hybridized M724029, 1:100), and anti-cleaved caspase-3 (Cell Signaling to a SurePrint G3 Human CGH microarray slide (Agilent Technology; #9661, 1:300) using our standardized IHC protocol Technologies). The array was then scanned using Agilent C with the BOND RX automated system (Leica Biosystems). The Scanner and the image analyzed with Feature Extraction software. stained slides were imaged and analyzed by a board-certified Patients who met all entry criteria, including positive RB status pathologist, and Aperio Image analysis software (Leica Biosys- (>20% RBþ cells) and loss of CDKN2A or amplification of tems) was used to assess differences in positivity for pRB, MIB1, CDK4 or CDK6 or CCND1/D2/D3 were eligible for enrollment. and cleaved caspase-3 in primary versus recurrent tumors. There

www.aacrjournals.org Clin Cancer Res; 2019 OF3

Tien et al.

were no significant differences between primary and recurrent prior to tumor recurrence, and no patients had received any other tissues in the levels of the tested biomarkers. adjuvant chemotherapy prior to enrollment. Similarly, no patients had been treated with tumor-treating fields technology. Pharmacodynamic assessment Twelve adult patients with recurrent glioblastoma were enrolled fi FFPE tissue from the patient's rst tumor resection (at the time in the Phase 0 component of the study and received 900 mg/day € of initial diagnosis) was used as "archival" or treatment-na ve of ribociclib for 5 days prior to the scheduled brain tumor control tissue for comparison. Both archival FFPE tumor tissue resection. Presurgical ribociclib was well-tolerated, and no and study specimens collected at the time of resection were patients were removed due to drug-related toxicities. All planned assayed simultaneously using our standardized IHC protocol surgical resections occurred within the designated interval fol- with the Leica BOND RX automated system. For each run, we lowing the last presurgical dose of ribociclib (median error, 19.7 included positive (historical glioblastoma tissue) and negative minutes). controls (no primary antibody). The following antibodies were utilized in the study: anti-RB (Cell Signaling Technology; #9309, Plasma and CNS pharmacokinetics 1:100), anti-pRB (Cell Signaling Technology; #8516, 1:400), anti- Fig. 1A illustrates the observed plasma concentration–time pFOXM1 (Cell Signaling Technology; #14655, 1:200), anti- profiles of total and unbound ribociclib in individual patients FOXM1 (Sigma; HPA029974, 1:500), anti-MIB-1 (DAKO; with glioblastoma following daily oral administration (900 mg) M724029, 1:100), anti-activated caspase-3 (Cell Signaling Tech- for 4 days. The plasma pharmacokinetics of total and unbound nology; #9661, 1:300), anti-pS6 ribosomal protein (Cell Signal- ribociclib are summarized (Fig. 1B). Ribociclib achieved the peak ing Technology; #4858, 1:100), and anti-4EBP1 (Cell Signaling plasma concentration at an average of 3.2 hours after oral admin- Technology; #2855, 1:500). Stained FFPE slides were imaged istration. It exhibited a mean elimination half-life of 13.5 hours, using a Leica DM55500 microscope and analyzed using Aperio based on which the steady-state drug exposure was considered Image analysis software. The function of "nucleus" in the software achieved after 4 or 5 days of daily treatment. A large inter- was used to systematically quantify the percentage of positive cells individual pharmacokinetic variability was observed, as demon- in a given area. strated by about 50% coefficient variation in the drug exposure C C Once the percentage of positive cells for a given protein was ( max, trough, and AUC0-24h) among 12 patients. Ribociclib was quantified, a paired t test was used to compare log-transformed modestly bound to plasma proteins, with the mean fraction percent positive cells between archival (predose) and phase 0 unbound of 0.12 in 12 patients with glioblastoma. At the surgical tissue (postdose) at a 2-sided 5% level. The comparisons steady-state, the unbound drug concentrations fluctuated were further carried out with the normalized percent positive cells 3.4-fold, from 0.13 mmol/L (mean trough level) to 0.44 mmol/L to adjust the baseline (i.e., historical primary and recurrent con- (mean peak level; Fig. 1B). trols) percent positive cells for both archival and surgical tissues. For CNS pharmacokinetics, as assessed by unbound ribociclib normalization, all percent positive cells of the treated tissues were concentrations in CSF, nonenhancing, and enhancing tumor divided by the median of the percent positive cells of the untreated regions (Fig. 2A) after oral administration of the fifth presurgical tissues (historical controls). The overall comparisons between predose are presented in Fig. 2B and summarized (Fig. 2C). Twelve and postdose were performed regardless of specific time-escalation patients were enrolled into 3 arms (with 4 patients per arm) for arms whereas the same comparisons were carried out for each of the assessment of CNS penetration and tumor pharmacodynam- 3 time-escalation arms (at 2, 8, and 24 hours postdose). ics at 2, 8, or 24 hours following the last dose of ribociclib. CSF drug concentrations are commonly used as a surrogate for Longitudinal tissue analysis unbound drug concentrations in the normal brain with intact For patients in the expansion cohort, evidence of tumor recur- BBB (15, 16). Ribociclib reached maximum CNS exposure at 2 to rence led to routine clinical consideration of re-resection as part of 8 hours following the oral administration, which was in-line with the patient's neurosurgical care. In cases where a re-resection was the Tmax in plasma (Fig. 2B). Drug elimination from the CNS planned, enhancing tumor tissue, in addition to blood and CSF, (including CSF, nonenhancing, and enhancing tumor regions) were collected intraoperatively. IHC analysis for phospho-S6, appeared relatively slower than that from plasma, as indicated pFOXM1, pS6, and p4EBP1 (as listed above) was performed on by the time-dependent increase of the CNS-to-plasma unbound archival-, surgical-, and posttreatment-derived tissue from each drug concentration ratios (Fig. 2C). At 2 to 24 hours after dosing, patient. Results were quantified using Aperio software, as the unbound ribociclib concentrations (mean SD) in CSF, described above. FFPE tumor tissue was analyzed using Vantage nonenhancing, and enhancing tumor regions were 0.374 3D solid tumor panel (28-plex protein expression; NanoString 0.272 mmol/L, 0.560 0.364 and 2.152 1.559 mmol/kg, Technologies). The quantification of each protein was normalized respectively. In 11 of 12 patients, the measured unbound drug by internal Histone-3 expression levels. The ratio of phospho- concentrations in CSF, nonenhancing, and enhancing tumor form of protein to total protein was calculated for key compo- regions at 2 to 24 hours post dosing were all 5-fold of the nents of the PI3K and MAPK pathways, including AKT, S6, 4EBP1, in vitro IC50 for inhibiting CDK4/6 kinase activity (0.04 mmol/L; and ERK1/2. Finally, RNA-sequencing was performed on FFPE ref. 6). tissue from all patients to characterize expression patterns following long-term ribociclib monotherapy. Tumor pharmacodynamics To assess the pharmacodynamic effects of ribociclib in human glioblastoma, phosphorylation of RB was selected as the primary Results determinant. Other assayed biomarkers included phosphoryla- Patient population tion of FOXM1, the proliferative marker MIB-1, and the apoptotic Patient demographics and clinical characteristics are summa- marker cleaved caspase-3. Among the 12 phase 0 patients, 9 rized in Table 1. All patients had completed the Stupp regimen (75%) exhibited a >20% of decrease in pRB levels (Fig. 3A). On

OF4 Clin Cancer Res; 2019 Clinical Cancer Research

Phase 0 of Ribociclib for Glioblastoma

A Total plasma 5 Unbound plasma 1 0.1 Ribociclib concentration ( m mol/L) Ribociclib

051015 20 25 B Sampling time (hours) Total ribociclib Unbound ribociclib

Tmax Cmax C trough AUC0−24 h CL/F C max C trough AUC0−24 h (hours) (µmol/L) (µmol/L) (h*µmol/L) T1/2 (hours) (L/h) Fu (µmol/L) (µmol/L) (h*µmol/L) CL/F (L/h) Mean±SD 3.2 ± 2.2 3.75 ± 1.74 1.16 ± 0.73 43.7 ± 21.7 13.5 ± 4.6 61.6 ± 35.4 0.12 ± 0.03 0.44 ± 0.20 0.13 ± 0.07 5.0 ± 2.1 503.6 ± 258.2 CV% 69 46 63 50 34 57 26 46 52 41 51 Median 3.0 3.72 1.05 40.2 12.5 51.5 0.12 0.43 0.16 5.5 379.5 Range 0.5−7.1 1.05−6.24 0.25−2.52 15.9−84.5 7.8−21.6 24.5−129.6 0.07−0.19 0.20−0.86 0.03−0.22 1.9−8.6 241.8−1110.6

Figure 1. Plasma pharmacokinetics of ribociclib in patients with glioblastoma. A, Observed plasma concentration–time profiles of total and unbound ribociclib in 12 patients with glioblastoma following daily oral administration (900 mg) for 4 days. Symbols represent observed plasma concentrations, and dash lines represent the mean plasma concentration–time profiles fitted with the one-compartment model. B, Plasma pharmacokinetic summary. Detailed summary of the steady-state plasma pharmacokinetic parameters of total and unbound ribociclib in patients with glioblastoma is listed.

AUC0–24h, area under the concentration–time curve during one dosing interval (24 hours); CL/F, apparent oral clearance; Cmax,maximumplasma concentration; Ctrough, trough plasma concentration (at 24 hours); T1/2, elimination half-life; Tmax,timetoachievetheCmax. average, there was a 50% decrease in pRB levels among the 12 Clinical response and longitudinal tissue analysis treated patients, which was significantly higher than paired Of the 12 phase 0 study patients, 4 demonstrated histological control specimens (Fig. 3B). With detailed quantification, both evidence of pseudo-progression, rendering them ineligible for the pRB and MIB-1 were significantly reduced following phase 0 expansion cohort study. Accordingly, acquired specimens from ribociclib exposure (P ¼ 0.001 and 0.039, respectively), and these 4 phase 0 patients were used for pharmacokinetics analysis cleaved caspase-3 also trended towards a concomitant increase only. Of the 8 phase 0 patients with histologically evident glio- (P ¼ 0.19; Fig. 3B). Observed declines in pRB and MIB-1 were blastoma recurrence, 6 (75%) were identified as responders based most prominent in the 8-hour postdose cohort (P ¼ 0.005 and on their pharmacokinetic and pharmacodynamic results and were 0.121, respectively). As a control, matched cohort analysis of subsequently enrolled into the expansion cohort study to receive 8 paired primary and recurrent glioblastoma samples without ribociclib continuously until recurrence is observed according to ribociclib exposure demonstrated no significant changes in RANO criteria. No patients were removed from the expansion any of the studied biomarkers (Fig. 3). Interestingly, levels of cohort due to drug-related toxicity. All patients ultimately pro- pFOXM1 had a downward trend (P ¼ 0.142) after normali- gressed to tumor recurrence, with a PFS6 of 16.7% and a median zation with historical control samples, primarily at the 8-hour OS of 7.8 months (Fig. 4A). The clinical course of these 6 patients, postdose interval (Fig. 3B). Both pRB and MIB-1 are also as well as associated genomic characterizations, are summarized significantly decreased after normalization. Among the 12 in Fig. 3B. We observed no patterns of genomic alterations phase 0 patients, 4 were diagnosed with histological pseu- statistically associated with better clinical response. do-progression. After removing these 4 patients, the pharma- Three of 6 patients underwent planned re-resection based on codynamic effects of ribociclib (n ¼ 8) remain evident with a evidence of tumor recurrence following ribociclib monotherapy significant decrease of pRB and MIB-1 (P ¼ 0.005 and 0.017, (termed posttreatment in figures). In all 3 patients, a significant respectively; Supplementary Fig. S1). Together, our pharma- increase in pRB was observed, as compared with matched speci- codynamic analysis suggests that the targeted inhibition of mens collected during their phase 0 operation (7.4% to 19.96%, CDK4/6 kinases is achieved in the recurrent tumor tissue from P ¼ 0.02; Fig. 5A and B). For these patients, we also observed short-term ribociclib treatment. a trend towards increase in the MIB-1 index (2.3% to 7.55%,

www.aacrjournals.org Clin Cancer Res; 2019 OF5

Tien et al.

A B

Unbound plasma CSF Unbound NE tumor Unbound EN tumor Median

Enhancing tumor

Nonenhancing tumor Ribociclib concentration ( m mol/L) 01234 2-4 hours 6-8 hours 23-25 hours C Sampling time (hours)

Sampling Unbound plasma CSF conc. Unbound NE Unbound EN µ µ time (hours) conc. (µmol/L) (µmol/L) tumor ( mol/kg) tumor ( mol/kg) CSF/Pu NETu/Pu ENTu/Pu NETu/CSF ENTu/CSF 0.354 0.650 0.511 2.660 1.5 1.3 7.0 0.9 4.0 2 hours (0.189−0.720) (0.316−0.838) (0.417−0.562) (0.829−4.743) (0.7−4.4) (0.8−3.0) (2.4−14.2) (0.6−1.3) (1.0−15.0)

0.211 0.340 0.597 1.412 1.4 3.1 7.7 1.8 6.6 8 hours (0.173−0.286) (0.212−0.656) (0.327−1.295) (1.115−4.322) (0.9−3.7) (1.8−4.5) (5.2−15.1) (1.2−2.8) (1.7−9.4)

0.077 0.122 0.351 1.215 2.0 4.6 13.8 3.0 12.4 24 hours (0.020−0.172) (0.043−0.200) (0.015−1.190) (0.307−3.928) (0.6−2.9) (0.7−11.3) (10.6−37.4) (0.4−6.0) (4.3−19.6)

0.184 0.267 0.458 1.644 1.8 2.0 10.1 1.4 6.6 All patients (0.020−0.720) (0.043−0.838) (0.015−1.295) (0.307−4.743) (0.6−4.4) (0.7−11.3) (2.4−37.4) (0.4−6.0) (1.0−19.6)

Central nervous system penetration of ribociclib in patients with glioblastoma.

Figure 2. CNS pharmacokinetics of ribociclib in patients with glioblastoma. A, Representative imaging indicating enhancing and nonenhancing region of the tumor. B, Observed unbound ribociclib concentrations in plasma, CSF, nonenhancing, and enhancing glioblastoma tumor regions at 2 to 24 hours after oral administration of the ﬁfth dose of ribociclib. Patients were treated with a daily oral dose of 900 mg for 5 days, and all samples were collected on day 5. Symbols represent observed concentrations, and dash lines represent the median concentration–time proﬁles. C, Summary of

pharmacokinetic characteristics of CNS ribociclib in patient with glioblastoma. Data are presented as the median (range). CSF/Pu,CSF-to-plasma unbound drug concentration ratio; EN, enhancing; ENTu/CSF, enhancing tumor-to–CSF unbound drug concentration ratio; ENTu/Pu, enhancing tumor-to–plasma unbound drug concentration ratio; NE, nonenhancing; NETu/CSF, nonenhancing tumor-to-CSF unbound drug concentration ratio; NETu/Pu, nonenhancing tumor-to-plasma unbound drug concentration ratio.

P ¼ 0.23; Fig. 5A and B). Interestingly, we detected a significant recurrent glioblastoma. Ribociclib achieves therapeutic concen- increase in the proportion of cells with mTOR/PI3K signaling trations not only in contrast-enhancing tumor regions with a pathway activity (Fig. 5A and B). Specifically, as compared with disrupted BBB but also in nonenhancing tumor tissues with matched Phase 0 samples, the posttreatment specimens exhibited largely intact BBB. Because nonenhancing regions of glioblas- significant increases in pS6 (21.3% to 36.7%, P ¼ 0.009) and toma are often unresectable, good penetration of ribociclib into p4EBP1 levels (2.8% to 31.2%, P ¼ 0.02; Fig. 5A and B). To further these regions would represent a significant therapeutic advan- examine mitogenic pathways in ribociclib-treated patients, we tage for treating glioblastoma. Following the administration of performed semiquantitative protein analyses using the nCounter the last dose (900 mg), the median unbound concentrations Vantage 3D protein solid tumor panel (NanoString Technolo- of ribociclib in nonenhancing region were 0.51, 0.6, and gies). The ratios of p4EBP-1/4EBP-1 in posttreatment samples 0.35 mmol/kg (or mmol/L) at 2, 8, and 24 hours, respective- were significantly higher compared with historical controls and ly—all exceeding the predefined threshold of 0.2 mmol/L (i.e., accompanied by an upward trend in the ratio of pAKT/AKT 5-fold of the in vitro IC50 for inhibition of CDK4/6). Given the (Fig. 5C), suggesting that ribociclib therapy results in the upre- linear pharmacokinetics of ribociclib, the median unbound gulation of the mTOR/PI3K pathway. Furthermore, for the 3 post- drug concentrations in nonenhancing tumor regions following treatment samples, we continued to observe an upward trend in the therapeutic dose of 600 mg would be approximately two- p4EBP1 and pAKT levels (Fig. 5D), indicating that the mTOR/ third of those observed at the 900-mg dose, which could be PI3K pathway may serve as a resistance mechanism following adequate for target inhibition (>0.2 mmol/L). Despite this CDK4/6 inhibition. No significant changes were observed in predicted correlation, the direct pharmacokinetic and pharma- EGFR or MEK1/2 signaling pathways (Fig. 5C and D). codynamic effects of 600-mg dosing were not measured in this study, and our understanding of this dose-level will be enhanced by future tissue-based study. Nevertheless, evidence Discussion of CDK4/6 pathway inhibition after 5 days of drug exposure This study provides the first clinical evidence of CNS pen- at 900 mg was suggested by significant reduction of RB etration of ribociclib and target modulation in patients with phosphorylation and tumor proliferation. The reliability of

OF6 Clin Cancer Res; 2019 Clinical Cancer Research

Phase 0 of Ribociclib for Glioblastoma

Figure 3. Pharmacodynamic analysis of surgical glioblastoma tissue after short-term ribociclib treatment. A, Representative pRB immunohistochemistry staining for matched archival and surgical specimens from ribociclib treatment patients and historical control patients are shown. Scale bar, 100 mm. B, Quantification of pRB, pFOXM1, MIB-1, and activated caspase-3 positive cells by Aperio imaging system from 8 matched archival and 12 surgical specimens is shown. Three time-escalating cohorts (2-, 8-, and 24-hour) after the last dose of ribociclib are indicated. Nontreatment control samples are in the left column, and the treatment group results and the normalized results are shown in the middle and right columns. P value of each marker is indicated. this observation is tempered by the study's need to employ cohort study suggest a limited clinical impact of ribociclib as archival tissues as the comparator for phase 0-collected specimens. monotherapy. Nevertheless, the accompanying longitudinal tis- Three orally bioavailable, highly-selective competitive inhibi- sue analysis raises the possibility that ribociclib-treated glioblas- tors of CDK4/6, including palbociclib, abemaciclib, and riboci- tomas utilize the PI3K/mTOR pathway for cellular resistance to clib, have been FDA-approved for the treatment of breast cancer CDK4/6 inhibition. Interestingly, previous work using glioma and are currently under clinical development for a variety of mouse models (18, 21), as well as reports from non-CNS can- cancers including glioblastomas (17, 18). In preclinical models, cers (18, 22), echo this observation. Taken together, these data all 3 CDK4/6 inhibitors demonstrated brain penetration and identify ribociclib as a potential component of a combinatorial antitumor activity against intracranial tumors (8, 19). Palbociclib drug strategy for recurrent glioblastoma and suggests an mTOR has been evaluated in 22 patients with recurrent RB-positive inhibitor as a candidate pairing. Although a regimen of ribociclib glioblastoma in a phase II trial in which, despite evidence of plus an mTOR inhibitor has yet to be tested in patients with tumor penetration (assessed by total drug tumor concentrations), glioblastoma, a phase I clinical trial for ribociclib plus everolimus it showed limited clinical efficacy in terms of prolonging PFS (20). plus exemestane in patients with endocrine-resistance advanced Abemaciclib is currently under investigation for the treatment breast cancer was reported in 2017 (23). In that study, the of patients with first recurrent glioblastoma [NCT02981940, maximally tolerated dose was not reached, and ribo- NCT03220646], and data are yet to be reported. Our present ciclib 300 mg/day þ everolimus 2.5 mg/day þ exemestane study on ribociclib complements previous knowledge on the 25 mg/day was declared the recommended phase II dose. therapeutic implications of modulating CDK4/6 signaling in Importantly, the ribociclib Ctrough was similar to expected glioblastomas. values, whereas the everolimus Ctrough was2-to3-foldgreater Compared with historical controls for recurrent glioblastoma, than expected, resulting in everolimus blood exposure levels the pharmacokinetic/pharmacodynamic-guided expansion equivalent to 5 to 10 mg dosing of everolimus (23). Based on

www.aacrjournals.org Clin Cancer Res; 2019 OF7

Tien et al.

Figure 4. PFS and OS of the expansion cohort. A, Kaplan–Meier plot of the PFS and the OS. B, Swimmer plot of the PFS for the 6 patients. Pharmacokinetic (mmol/L) and pharmacodynamic characteristics of the patients are shown.

these findings, a combinatorial ribociclib plus everolimus patients into a therapeutic continuous treatment cohort is phase 0/II study for recurrent glioblastoma is underway more compelling for potential study candidates as it provides (NCT03834740). direct biological evidence supporting the decision for thera- The study design used in this trial is an adaptation of peutic dosing. Although the expansion cohort study is not a conventional phase 0 paradigms (10) to accommodate chal- conventional phase II trial (26), it nevertheless provides a lenges associated with brain tumor patients, including the view of preliminary clinical responses and contextualizes high degree of risk surrounding tumor acquisition, the unsuit- these responses to therapy against the backdrop of detailed ability of microdosing when crossing the BBB, and the low molecular and pharmacological analyses. Embracing and accrual rates of nontherapeutic trials. Nevertheless, a number expanding this strategy for brain tumor patients may alleviate of important study design limitations exist. Conventional some of the challenges associated with conventional phase II phase 0 studies for non-CNS cancers often employ multiple studies. biopsies before and after drug exposure. In contrast, phase 0 trials such as this rely on archival tissue, instead, to serve as the baseline comparator. Additionally, inadequate penetra- Conclusion tion of therapeutic agents across the BBB remains a central We conducted a phase 0 clinical trial with a pharmacokinet- obstacle for CNS drug development (24, 25). For brain tumor ic/pharmacodynamic-guided expansion cohort to assess plas- patients, the utility of a microdosing strategy is limited by the ma and tumor pharmacokinetic and pharmacodynamic end- BBB. Specifically, drug microdoses (typically <1% of the points in recurrent glioblastoma patients. Ribociclib penetrated therapeutic dose) are often undetectable within the CNS, even well into not only human glioblastoma regions with disrupted when using advanced bioanalytical methods. Consequently, BBB, but also tumor regions with largely intact BBB. Drug brain tumor phase 0 studies necessitate higher systemic drug exposure was accompanied by a pharmacodynamic response, concentrations and higher risks of associated side effects. as indicated by significant reduction in RB phosphorylation Here, we used the maximally tolerated dose for a brief win- and tumor proliferation. However, preliminary analysis from dow of exposure prior to the planned resection. A final the expansion cohort suggests that ribociclib monotherapy challenge for brain tumor phase 0 studies is the dampening showed limited clinical efficacy in recurrent glioblastoma. Our effect that nontherapeutic trials have on patient accrual. A data support further investigation of combination therapy phase 0 trial that incorporates a pharmacokinetic- and phar- with CDK4/6 and PI3K/mTOR inhibitors for treating recurrent macodynamic-guided expansion cohort, graduating phase 0 glioblastoma.

OF8 Clin Cancer Res; 2019 Clinical Cancer Research

Phase 0 of Ribociclib for Glioblastoma

Figure 5. Longitudinal pharmacodynamic analysis of re-recurrent tumor specimens after continuous ribociclib treatment (posttreatment). A, Representative immunohistochemistry images of pRB and pS6 are shown from matched tumor specimens (archival, surgical, and posttreatment). B, Quantiﬁcation and statistical analysis of percentage of pRB, MIB-1, pS6, and p4EBP1 are shown for 3 patients with re-recurrent surgery tumor specimens. C and D, Protein array analysis was performed across study specimens and nonstudy control specimens. Quantiﬁcation of the ratio of p4EBP-1/4EBP-1, pAKT/ATK, pERK/ERK, pS6/S6, and pEGFR/EGFR is shown with nontreatment historical control (n ¼ 4) and ribociclib treatment cohort (n ¼ 7) (C) and posttreatment re-recurrent cohort (n ¼ 3; D). P value of each protein between samples is indicated.

Disclosure of Potential Conflicts of Interest Acknowledgments No potential conflicts of interest were disclosed. The study was supported by a grant from the Ben and Catherine Ivy Foundation as well as the United States National Institute of Health Cancer Authors' Contributions Center Support Grant (P30 CA022453). We thank staff and research nurses who participated in this work at the Ivy Brain Tumor Center as well as Novartis for Conception and design: A.-C. Tien, J. Li, S. Mehta, N. Sanai providing the study drug (ribociclib). We are indebted to the patients who Development of methodology: A.-C. Tien, J. Li, X. Bao, S. Mehta, N. Sanai enrolled in this study. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.-C.Tien,J.Li,X.Bao,A.Derogatis, N. Sanai The costs of publication of this article were defrayed in part by the Analysis and interpretation of data (e.g., statistical analysis, biostatistics, payment of page charges. This article must therefore be hereby marked advertisement computational analysis): A.-C. Tien, J. Li, A. Derogatis, S. Kim, S. Mehta, in accordance with 18 U.S.C. Section 1734 solely to indicate N. Sanai this fact. Writing, review, and/or revision of the manuscript: A.-C. Tien, J. Li, S. Kim, S. Mehta, N. Sanai Received January 15, 2019; revised April 14, 2019; accepted July 2, 2019; Study supervision: A.-C. Tien, J. Li, S. Mehta, N. Sanai published first July 8, 2019.

References 1. Cancer Genome Atlas Research Network. Comprehensive genomic char- 2. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, et al. An acterization deﬁnes human glioblastoma genes and core pathways. Nature integrated genomic analysis of human glioblastoma multiforme. Science 2008;455:1061–8. 2008;321:1807–12.

www.aacrjournals.org Clin Cancer Res; 2019 OF9

Tien et al.

3. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell last step cross metathesis diversification strategy. Mol Pharm 2008;5: 2011;144:646–74. 829–38. 4. Michaud K, Solomon DA, Oermann E, Kim JS, Zhong WZ, Prados MD, et al. 16. Liu X, Smith BJ, Chen C, Callegari E, Becker SL, Chen X, et al. Evaluation of Pharmacologic inhibition of cyclin-dependent kinases 4 and 6 arrests the cerebrospinal fluid concentration and plasma free concentration as a growth of glioblastoma multiforme intracranial xenografts. Cancer Res surrogate measurement for brain free concentration. Drug Metab Dispos 2010;70:3228–38. 2006;34:1443–7. 5. Edessa D, Sisay M. Recent advances of cyclin-dependent kinases as 17. Sherr CJ, Beach D, Shapiro GI. Targeting CDK4 and CDK6: from discovery potential therapeutic targets in HRþ/HER2- metastatic breast can- to therapy. Cancer Discov 2016;6:353–67. cer: a focus on ribociclib. Breast Cancer (Dove Med Press) 2017;9: 18. Knudsen ES, Witkiewicz AK. The strange case of CDK4/6 inhibitors: 567–79. mechanisms, resistance, and combination strategies. Trends Cancer 6. Tripathy D, Bardia A, Sellers WR. Ribociclib (LEE011): mechanism 2017;3:39–55. of action and clinical impact of this selective cyclin-dependent 19. Patel Y. PDTB-12. CNS Penetration of the CDK4/6 inhibitor ribociclib kinase 4/6 inhibitor in various solid tumors. Clin Cancer Res (LEE011) in non-tumor bearing mice and mice bearing orthotopic pedi- 2017;23:3251–62. atric brain tumors. Neuro Oncol 2016;18(suppl_6):vi152. 7. Geoerger B, Bourdeaut F, DuBois SG, Fischer M, Geller JI, Gottardo NG, 20. Taylor JW, Parikh M, Phillips JJ, James CD, Molinaro AM, Butowski NA, et al. A phase I study of the CDK4/6 inhibitor ribociclib (LEE011) in et al. Phase-2 trial of palbociclib in adult patients with recurrent RB1- pediatric patients with malignant rhabdoid tumors, neuroblastoma, and positive glioblastoma. J Neurooncol 2018;140:477–83. other solid tumors. Clin Cancer Res 2017;23:2433–41. 21. Olmez I, Brenneman B, Xiao A, Serbulea V, Benamar M, Zhang Y, et al. 8. Raub TJ, Wishart GN, Kulanthaivel P, Staton BA, Ajamie RT, Sawada GA, Combined CDK4/6 and mTOR inhibition is synergistic against et al. Brain exposure of two selective dual CDK4 and CDK6 inhibitors and glioblastoma via multiple mechanisms. Clin Cancer Res 2017;23: the antitumor activity of CDK4 and CDK6 inhibition in combination with 6958–68. temozolomide in an intracranial glioblastoma xenograft. Drug Metab 22. El-Naggar S, Liu Y, Dean DC. Mutation of the Rb1 pathway leads to Dispos 2015;43:1360–71. overexpression of mTor, constitutive phosphorylation of Akt on serine 9. Kola I, Landis J. Can the pharmaceutical industry reduce attrition rates? 473, resistance to anoikis, and a block in c-Raf activation. Mol Cell Biol Nat Rev Drug Discov 2004;3:711–5. 2009;29:5710–7. 10. Kummar S, Kinders R, Rubinstein L, Parchment RE, Murgo AJ, Collins J, 23. Hurvitz S, Yardley D, Zelnak A, DeMichele A, Tan-Chiu E, Ma C, et al. et al. Compressing drug development timelines in oncology using phase '0' Ribociclib in combination with everolimus and exemestane in men and trials. Nat Rev Cancer 2007;7:131–9. postmenopausal women with HRþ/HER2-advanced breast cancer follow- 11.SanaiN,LiJ,BoernerJ,StarkK,WuJ,KimS,etal.Phase0trialof ing progression on a CDK4/6 inhibitor: safety, tolerability and pharma- AZD1775 in first-recurrence glioblastoma patients. Clin Cancer Res cokinetic results from phase 1 of TRINITI-1 study [abstract]. In: Proceed- 2018;24:3820–8. ings of the AACR Annual Meeting 2017 Apr 1–5; Washington, DC. 12. Xu R, Shimizu F, Hovinga K, Beal K, Karimi S, Droms L, et al. Molecular and Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr clinical effects of notch inhibition in glioma patients: a phase 0/I trial. CT110. Clin Cancer Res 2016;22:4786–96. 24. Bicker J, Alves G, Fortuna A, Falcao A. Blood-brain barrier models and their 13. Bao X, Wu J, Sanai N, Li J. Determination of total and unbound ribociclib in relevance for a successful development of CNS drug delivery systems: a human plasma and brain tumor tissues using liquid chromatography review. Eur J Pharm Biopharm 2014;87:409–32. coupled with tandem mass spectrometry. J Pharm Biomed Anal 2019; 25. Parrish KE, Sarkaria JN, Elmquist WF. Improving drug delivery to primary 166:197–204. and metastatic brain tumors: strategies to overcome the blood-brain 14. Waters NJ, Obach RS, Di L. Consideration of the unbound drug concen- barrier. Clin Pharmacol Ther 2015;97:336–46. tration in enzyme kinetics. Methods Mol Biol 2014;1113:119–45. 26. Galanis E, Wu W, Cloughesy T, Lamborn K, Mann B, Wen PY, et al. Phase 2 15. Mooberry SL, Hilinski MK, Clark EA, Wender PA. Function-oriented trial design in neuro-oncology revisited: a report from the RANO group. synthesis: biological evaluation of laulimalide analogues derived from a Lancet Oncol 2012;13:e196–204.

OF10 Clin Cancer Res; 2019 Clinical Cancer Research

A Phase 0 Trial of Ribociclib in Recurrent Glioblastoma Patients Incorporating a Tumor Pharmacodynamic- and Pharmacokinetic-Guided Expansion Cohort

An-Chi Tien, Jing Li, Xun Bao, et al.

Clin Cancer Res Published OnlineFirst July 8, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-19-0133

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2019/07/04/1078-0432.CCR-19-0133.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2019/08/16/1078-0432.CCR-19-0133. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.