Received: 4 November 2017 Revised: 6 February 2018 Accepted: 7 February 2018

DOI: 10.1002/pbc.27035

Pediatric
RESEARCH ARTICLE Blood &
The American Society of
Cancer Pediatric Hematology/Oncology

Phase 1 trial, pharmacokinetics, and pharmacodynamics of

dasatinib combined with crizotinib in children with recurrent
or progressive high-grade and diffuse intrinsic pontine glioma
Alberto Broniscer1,2 Sujuan Jia3 Belinda Mandrell4 Dima Hamideh1
Jie Huang5 Arzu Onar-Thomas5 Amar Gajjar1,2 Susana C. Raimondi3
Ruth G. Tatevossian3 Clinton F. Stewart6

1 Department of Oncology, St. Jude Children's

Research Hospital, Memphis, Tennessee Abstract

2 Department of Pediatrics, University of Background: Progressive/recurrent high-grade and diffuse intrinsic pontine gliomas (DIPGs) are
Tennessee Health Science Center, Memphis, fatal. Treatments targeting molecular pathways critical for these cancers are needed.
Tennesssee
3 Department of Pathology, St. Jude Children's Methods: We conducted a phase 1 study (rolling-six design) to establish the safety and maximum
Research Hospital, Memphis, Tennessee tolerated dose (MTD) of dasatinib, an oral platelet-derived growth factor receptor A (PDGFRA)
4 Department of Pediatric Medicine, St. Jude inhibitor, and crizotinib, an oral c-Met inhibitor, in such patients. Pharmacokinetics of both agents
Children's Research Hospital, Memphis, were performed. Biomarkers of cellular pathway activation in peripheral-blood mononuclear cells
Tennessee
(PBMC) were evaluated before and after administration of dasatinib. PDGFRA and MET amplifica-
5 Department of Biostatistics, St. Jude Children's
tion, and PDGFRA mutations were studied in tumor samples.
Research Hospital, Memphis, Tennessee
6 Department of Pharmaceutical Sciences, St.
Results: Twenty-five patients were enrolled in this study (median age: 11.9 years). Eleven patients
Jude Children's Research Hospital, Memphis,
had DIPG. Glioblastoma accounted for 40% of cases. Dasatinib at 50 mg/m2 and crizotinib at
Tennessee
130 mg/m2 or 100 mg/m2 were poorly tolerated when administered twice daily. Drug adminis-
Correspondence
Alberto Broniscer, Department of Pediatrics, tration was then switched to once daily. Dasatinib administered at 50 mg/m2 and crizotinib at
Children's Hospital of Pittsburgh, 4401 Penn 215 mg/m2 once daily was the MTD. Dose-limiting toxicities consisted of diarrhea, fatigue, pro-
Avenue, Pittsburgh, PA 15224. teinuria, hyponatremia, rash, and grade 4 neutropenia. Only two patients received therapy for at
Email: [email protected]
least 6 months. No objective radiologic responses were observed. Pharmacokinetics of dasatinib
Funding information
Grant sponsor: United States National Insti-
and crizotinib were comparable to previous studies. A statistically significant decrease in the ratio
tutes of Health Cancer Center Support (CORE), of p-AKT/total AKT in PBMC occurred after dasatinib administration. PDGFRA and MET amplifi-
Grant number: P30 CA21765, Grant sponsor: cation were found in four and two cases, respectively. Only one of 10 tumors harbored a PDGFRA
American Lebanese Syrian Associated Charities
mutation.
(ALSAC)

Conclusions: This drug combination was poorly tolerated and its activity was minimal. We do not
recommend further testing of this combination in children.

KEYWORDS
c-Met, children, crizotinib, dasatinib, diffuse intrinsic pontine glioma, high-grade glioma, PDGFRA

1 INTRODUCTION nervous system (CNS) tumors in children.1 The prognosis of affected

patients remains dismal despite treatment with maximum safe surgical
High-grade gliomas (HGGs), including diffuse intrinsic pontine gliomas resection, radiation therapy (RT), and chemotherapy.1 Children with
(DIPGs), are among the most aggressive and lethal primary central recurrent or progressive HGGs and DIPGs are ideal candidates for
investigational clinical trials since there are no standard chemothera-
peutic regimens for these tumors.
Abbreviations: CNS, central nervous system; DIPG, diffuse intrinsic pontine glioma; DLT, Two major developments have stimulated the design of more
dose-limiting toxicity; HGF, hepatocyte growth factor; HGG, high-grade glioma; HRQoL,
rational therapies for children with HGGs and DIPGs. First, genome-
health-related quality of life; MTD, maximum tolerated dose; PBMC, peripheral-blood
mononuclear cells; PDGFR, platelet-derived growth factor receptor; RT, radiation therapy wide studies have uncovered the molecular characteristics of these

Pediatr Blood Cancer. 2018;e27035. wileyonlinelibrary.com/journal/pbc 

c 2018 Wiley Periodicals, Inc. 1 of 8
https://doi.org/10.1002/pbc.27035
2 of 8 BRONISCER ET AL .

cancers, which are mostly different from adult tumors.2–6 Second, the ies; (9) interval ≥ 3 months from previous high-dose chemotherapy
establishment of preclinical models, including cell lines and human with stem-cell rescue; (10) interval ≥ 1 week and ≥ 2 weeks from the
xenografts, has allowed for the high-throughput screening of new use of filgrastim and pegfilgrastim, respectively; and (11) effective con-
promising agents.7 Therapies targeting abnormally activated cellular traception for females of childbearing age and males of child fathering
pathways have shown promising activity against preclinical models potential. Exclusion criteria consisted of the following: (1) use of other
of pediatric HGGs and DIPGs including inhibitors of platelet-derived anticancer therapies; (2) use of enzyme-inducing anticonvulsants less
growth factor receptor (PDGFR) and c-Met.7–9 than 10 days before start of protocol therapy; (3) previous treatment
The PDGF pathway influences tumor formation by stimulating with a PDGFR or c-Met inhibitor; (4) presence of other medical con-
mitogenesis, dedifferentiation, and increased angiogenesis, and its ditions that could interfere with the study procedures or tolerance to
modulation in vivo in combination with other oncogenic abnormalities therapy; and (5) pregnant or lactating patients.
led to the formation of HGGs and DIPGs in preclinical models.10–13 Our institutional review board approved this protocol before initial
While PDGFRA amplification can be found in one-third and 19% of patient enrollment, and continuing approval was maintained through-
DIPGs and pediatric nonbrainstem HGGs, respectively,2,6 PDGFRA out the study. Written informed consent for participation was obtained
mutations occurred in approximately 5% and 14% of such tumors, from patients’ parents or legal guardians, and assents were obtained
respectively.12 when appropriate.
Activation of the c-Met pathway has been recognized as a con-
tributing mechanism for the formation of HGGs.14 MET is the sec-
2.1 Study design
ond most common amplified oncogene in DIPG.2 MET fusion tran-
scripts have recently been identified in approximately 10% of pediatric This single-institution clinical trial followed the rolling-six design. The
glioblastomas (WHO grade IV), particularly in those originating from MTD was defined as the highest dosage at which no more than one
the cerebral hemispheres.9 MET fusion transcripts in combination with of six evaluable patients experienced a dose-limiting toxicity (DLT).
other oncogenic mechanisms also led to the formation of HGGs in an Two regimens of this combination were tested. Dasatinib and crizotinib
in vivo preclinical model.9 Finally, cell lines and xenografts containing were initially administered twice daily and the DLT evaluation period
MET fusion transcripts were responsive to c-Met inhibition.9 lasted for 6 weeks. Drug administration was switched to once daily
Dasatinib (Sprycel, Bristol–Myers Squibb) is a potent oral inhibitor and the DLT-evaluation period to 4 weeks as this initial regimen was
of PDGFRA and B, Src, and c-Kit, for which a phase 2 single-agent not tolerable. Dose-limiting toxicities consisted of the following side
recommended dose has already been established in children.15 Crizo- effects attributable to study drugs: (1) grade 4 neutropenia; (2) grade
tinib (Xalkori, Pfizer) is a potent inhibitor of ALK and c-Met that has 3 or 4 thrombocytopenia; (3) any grade 3 or 4 nonhematologic toxi-
already undergone phase 1 testing in children.16 Based on previously city except for grade 3 weight change, grade 3 elevation in transam-
described data, we initiated a phase 1 clinical trial to test the tolerabil- inases that returned to baseline or ≤grade 1 within 7 days of drug
ity and safety, and to establish the maximum tolerated dose (MTD) of interruption, grade 3 or 4 electrolyte abnormalities that returned to
the combination of dasatinib and crizotinib in pediatric patients with ≤grade 2 within 7 days, and grade 3 fever or infection lasting < 5 days;
recurrent or progressive HGGs and DIPGs. and (4) any grade 2 nonhematologic toxicity lasting > 7 days and caus-
ing significant clinical repercussion. Toxicities were graded based on
the National Cancer Institute Common Toxicity Criteria for Adverse
Events (version 4.0).
2 PATIENTS AND METHODS For the twice-daily regimen, the starting doses of dasatinib and
crizotinib were 50 mg/m2 and 130 mg/m2 , respectively. For the once-
Patients between the ages of 2 and 21 years old with clinically and/or daily regimen, the starting doses of dasatinib and crizotinib were
radiologically recurrent or progressive HGG or DIPG were eligible for 50 mg/m2 and 165 mg/m2 , respectively (Table 1). In both regimens,
this study. Other inclusion criteria consisted of the following: (1) per- patients received the first dose of crizotinib on day 1 and then treat-
formance score ≥ 50; (2) adequate hematologic (hemoglobin ≥ 8 g/dl ment was held for 48 hr to allow for completion of pharmacokinetic
[irrespective of previous transfusions], absolute neutrophil count ≥ studies. The administration of dasatinib was delayed for 48 hr as well.
1,000/mm3 , and platelet count ≥ 100,000/mm3 [transfusion indepen- Treatment with both drugs continued uninterrupted on day 3 of ther-
dent]), renal (normal serum creatinine for age), and hepatic (transam- apy except for the omission of the evening dose of crizotinib on day 14
inases and bilirubin <3× and <1.5× the institutional upper limit of in the twice-daily regimen so that extended pharmacokinetic studies
normal, respectively, and albumin ≥ 2 g/dl) functions; (3) stable neu- could be completed.
rologic deficits on a fixed or decreasing dose of dexamethasone for Treatment was divided into 28-day cycles except for the first two
≥1 week; (4) no more than a grade 1 toxicity attributed to previous cycles on the twice-daily regimen that comprised 6 weeks. The maxi-
therapies; (5) interval from previous local, craniospinal, and palliative mum planned treatment duration was 2 years.
RT of ≥3 months, ≥6 months, and ≥2 weeks, respectively; (6) interval Dasatinib was administered with or without food as 20, 50, or 70
≥ 4 weeks from previous chemotherapy (6 weeks if nitrosourea); (7) mg tablets. The tablets of dasatinib could be cut in half or crushed to
interval ≥ 1 week from previous small-molecule inhibitors with short ease administration. Crizotinib was administered with or without food
half-lives; (8) interval ≥ 3 half-lives from previous monoclonal antibod- as 200 mg and 250 mg capsules or as a solution (10 mg/ml) depending
BRONISCER ET AL . 3 of 8

TA B L E 1 Proposed dose escalation schema pharmacokinetic studies of dasatinib, serial samples were obtained
Dasatinib (mg/m 2 2
Crizotinib (mg/m before and at 1, 2, 4, 8, and 24 hr after their morning dose. Other details
Dosage level per dose) per dose) are provided in Supplementary Appendix S1.
Twice-daily 0 50 100
regimen
1a 50 130
2.3 Pharmacodynamic studies
2 65 130
Optional studies consisted of serial evaluation of the plasma concen-
3a 65 165
tration of the hepatocyte growth factor (HGF) obtained concurrently
3b 85 130
with brain MRIs before the start and during therapy, and analysis of
4 85 165
total and phosphorylated AKT and ERK in peripheral-blood mononu-
5 85 215
clear cells (PBMCs) obtained before and 3 hr after the administration
Once-daily 0 50 130
of dasatinib on day 14 of therapy and at the end of the DLT-evaluation
regimen
period. Details about sample processing, collection, and analysis are
1a 50 165
provided in Supplementary Appendix S1.
2a 50 215
2b 65 165
3 65 215
2.4 Molecular studies
4a 65 280
Analysis of amplification of PDGFRA and MET by FISH was performed
4b 85 280
as previously described.17 The methods used for PDGFRA sequencing
a
Starting dosage level.
are provided in Supplementary Appendix S1.

on the planned dose and the ability to swallow capsules. The dose of
both medications was rounded to the nearest 10 mg. The actual dose of 2.5 Assessment of quality of life
both medications was calculated based on the body surface area (BSA)
Health-related quality of life (HRQoL) was assessed through patient
at the start of therapy. Changes in dose based on variations of BSA
and parent-proxy report using the Pediatric Quality of Life Inven-
were allowed after 6 months of therapy. When applicable, an inter-
tory (PedsQL 4.0) and the Pediatric Quality of Life Brain Tumor Mod-
val of at least 4 hr was recommended between the administration of
ule (PedsQL BT module).18,19 Patients and parent-proxy reported on
dasatinib and histamine receptor 2 antagonists. Use of septra was flexi-
HRQoL and symptoms over the previous 7 days. The HRQoL assess-
ble, depending on the physician's preferences and patients’ needs. Lop-
ments were obtained at the initiation of therapy, at weeks 2, 4, and 6 of
eramide was strongly recommended if patients developed diarrhea.
therapy, and before every other cycle of therapy beginning at cycle 3.
A clinical assessment, complete blood counts with differential,
chemistry panel, urinalysis, and electrocardiogram were obtained
before the start of therapy and weekly during cycle 1, every 2 weeks
2.6 Statistical analysis
during cycle 2, and every 4 weeks thereafter. Magnetic resonance
imaging (MRI) scans for tumor evaluation were obtained before the Descriptive statistics were provided. Random coefficient model was
start of therapy, before cycle 2, and after every other cycle thereafter. used to test whether HGF concentration changed with time. The
An echocardiogram was obtained before the start of therapy, at the Spearman correlation was used to estimate the association between
completion of cycle 1, and every 6 months thereafter during treat- HGF plasma concentrations and length of therapy. Sign test, a nonpara-
ment. Chest X-ray was obtained before the start of therapy and every metric analog of the t-test for paired samples, was used to compare p-
6 months thereafter. ERK/total ERK (or p-AKT/total AKT) before and after administration of
dasatinib, and to compare data obtained after administration of dasa-
tinib on different days. A significance threshold of 0.05 was used with-
2.2 Pharmacokinetic studies
out adjusting for multiplicity.
For the twice-daily regimen, mandatory pharmacokinetic studies of The Wilcoxon signed-rank test was used to compare the change in
crizotinib were scheduled on days 1 and 14, and studies of crizotinib HRQoL from baseline to week 2, week 2 to week 4, and week 4 to
and dasatinib were to occur approximately at the end of the DLT- week 6 of therapy for patients and parents. Patient (≥5 years of age)
evaluation period (day 42 ± 3 days). In the once-daily regimen, the self-reported and parent-proxy reported HRQoL domain scores were
pharmacokinetic studies of crizotinib were scheduled on days 1 and 14, compared at each therapy time point using the Wilcoxon signed-rank
and the pharmacokinetic studies of dasatinib occurred on day 14. test, with data only reported for patients and parents with compar-
Day 1 serial blood samples (3 mL) for crizotinib were collected ative data. The false discovery rate adjustment was used to account
before and at 1, 2, 4, 8, 24, and 48 hr after the administration of crizo- for multiple testing when appropriate. A two-sided significance level
tinib. On day 14, serial samples of crizotinib were obtained before of P < 0.05 was used for all statistical tests. Statistical analyses were
and at 1, 2, 4, 8, and 24 hr after the morning dose of crizotinib. For conducted using SAS Version 9.4 (SAS Institute, Cary, NC).
4 of 8 BRONISCER ET AL .

TA B L E 2 Characteristics of 25 patients enrolled in this study 2a. Only one of six evaluable patients at this dosage level had a DLT
Number of consisting of grade 3 diarrhea and rash. Hence, three patients were
Characteristics patients (%) treated at dosage level 2b, two of whom experienced DLTs consisting
Median age (in years) at 11.9 of grade 3 rash (n = 1) and grade 4 neutropenia (n = 1). Therefore, the
study enrollment (range: 2.9–21.3)
MTD was established as dasatinib 50 mg/m2 and crizotinib 215 mg/m2
Gender administered once daily. Table 4 provides a summary of the most signif-
Male 10 (40) icant toxicities attributed to the study agents. Diarrhea was common
Female 15 (60) during and after the completion of the DLT-evaluation period despite
Race/ethnicity use of loperamide. Fatigue related to study agents was often difficult to
Caucasian 17 (68) distinguish from clinical progression and was suspected based on rapid
African-American 3 (12) onset without other signs of clinical deterioration and improvement
Hispanic 5 (20) by holding the dose of dasatinib, the medication with the shortest

Histologic diagnoses
half-life. Only one patient experienced QTc prolongation grade 1
at dosage level 2a. No pleural effusions related to dasatinib were
Anaplastic astrocytoma 3 (12)
observed.
Glioblastoma 10 (40)
High-grade glioma 4 (16)
No histologic confirmation 8 (32)
Previous RT 25 (100)
3.1 Pharmacokinetic studies
Previous regimens of chemotherapy
Pharmacokinetic studies of crizotinib were collected from 25 patients
0 2 (8)
on day 1 of therapy. The median (range) Tmax , apparent oral clear-
1 16 (64)
ance, and elimination half-life of crizotinib were 4.1 hr (1.0–23.9),
2 5 (20)
56.5 l/hr/m2 (17.7–113.2), and 13.1 hr (9.8–34.7), respectively. Nine-
≥3 2 (8)
teen patients had pharmacokinetic studies of crizotinib performed
Of note, 11 patients had a radiologic diagnosis of diffuse intrinsic pontine on day 14 of therapy. The median (range) apparent oral clearance
glioma.
Three patients with diffuse intrinsic pontine glioma had histologic confir-
and elimination half-life were 50.3 l/hr/m2 (12.0–126.9) and 12.8 hr
mation at biopsy (n = 2) or at autopsy (n = 1). (8.0–36.5), respectively.
Pharmacokinetic studies of dasatinib were performed in 15
patients. However, only those studies obtained for 12 patients on
3 RESULTS
day 14 of therapy were deemed appropriate for analysis. The median
(range) Tmax , apparent oral clearance, and elimination half-life of dasa-
Twenty-five patients were enrolled in this study from November 2012
tinib were 2.1 hr (1.1–4.0), 1,328 mL/min/m2 (532–2,125), and 4.4 hr
until September 2016. Tables 2 and 3 list their characteristics at study
(2.9–7.9), respectively. The results of the Cmax and AUC0→t of crizotinib
enrollment. All patients except one had measurable radiographic dis-
and dasatinib are summarized in Table 5.
ease at the start of this therapy. Three (12%) of 25 patients had lep-
tomeningeal metastases at the start of therapy although complete
metastatic work-up was not mandatory.
Of five patients enrolled on dosage level 1 of the twice-daily regi-
3.2 Pharmacodynamic studies
men, three of four evaluable patients experienced DLTs consisting of
grade 3 diarrhea (n = 1) and fatigue (n = 2). Therefore, three patients Twenty-one patients had 46 serial plasma samples analyzed for the
started treatment on dosage level 0, two of whom developed DLTs. One concentration of HGF (median = 198.3 pg/ml; range: 88.1–3,925.8).
patient experienced grade 3 hyponatremia concurrently with disease There was no change in the log-transformed concentration of HGF
progression; further evaluation showed that hyponatremia was at least in plasma over time for patients with serially collected samples
partially related to inappropriate secretion of antidiuretic hormone. (P = 0.087). We found a negative correlation between HGF plasma con-
Another patient with normal urinalysis at the start of therapy devel- centration at baseline and the duration of therapy (Spearman correla-
oped grade 3 proteinuria within the nephrotic range which improved tion = -0.59; P = 0.0048)
with discontinuation of treatment. Twenty paired samples in 14 patients were available for analysis
Since the twice-daily regimen was poorly tolerated even at the low- of the changes produced by dasatinib in markers of cellular pathway
est planned dosage level, the clinical trial was amended to test doses activation in PBMC. When all paired samples and only those collected
of the study drugs closer to their single-agent MTD administered once on day 14 of therapy (n = 13) were analyzed, there was a statistically
daily. Six patients were enrolled on dosage level 1 of the once-daily significant decrease in the ratio of p-AKT/total AKT (P = 0.041 and
regimen, three of whom were unevaluable due to early disease pro- 0.003, respectively) 3 hr after the administration of dasatinib. No
gression. None of the three evaluable patients experienced significant statistically significant changes in the ratio of p-ERK/total ERK were
toxicities. Therefore, eight patients started treatment on dosage level found after dosing.
BRONISCER ET AL . 5 of 8

TA B L E 3 Clinical and molecular data for all 25 patients

Age at study Duration of

Gender/race or ethnicity enrollment (years) Histologic diagnosis Molecular abnormalities therapy (weeks)
1 F/Caucasian 11.9 Anaplastic astrocytoma No PDGFRA or MET ampl 6
2 F/African-American 15.5 Not available Not available 11
3 F/Caucasian 8 Glioblastoma No PDGFRA or MET ampl 12
4 M/Caucasian 14.9 Not available Not available 6
5 F/Caucasian 16.8 Glioblastoma No PDGFRA or MET ampl <1
6 F/Caucasian 5 Not available Not available 6
7 F/African-American 7.7 Anaplastic astrocytoma No PDGFRA or MET ampl 5
8 F/Hispanic 5.9 Not available Not available 2
9 M/Caucasian 21.3 Glioblastoma MET ampl; no PDGFRA ampl 3
10 M/Caucasian 5.3 Glioblastoma No PDGFRA or MET ampl <1
11 M/Caucasian 6.5 Not available Not available 4
12 F/Hispanic 10.1 High-grade glioma No PDGFRA or MET ampl 16
13 M/Caucasian 7.4 Not available Not available 2
14 M/Caucasian 16.1 Glioblastoma No PDGFRA or MET ampl 48
15 F/Hispanic 16.9 High-grade glioma Not available 1
16 F/Caucasian 5.7 Not available Not available 4
17 M/Caucasian 13.1 Glioblastoma PDGFRA ampl; no MET ampl 8
18 F/Caucasian 7.1 Not available Not available 4
19 M/African-American 19.1 High-grade glioma No PDGFRA or MET ampl 4.5
20 F/Hispanic 13.1 Glioblastoma No PDGFRA or MET ampla 3
21 F/Caucasian 13.8 Anaplastic astrocytoma No PDGFRA or MET ampl 16
22 F/Caucasian 13 Glioblastoma PDGFRA and MET ampl 9
23 M/Hispanic 14.9 High-grade glioma Focal PDGFRA ampl; no MET ampl 25
24 M/Caucasian 2.9 Glioblastoma PDGFRA ampl; no MET ampl 8
25 F/Caucasian 9.3 Glioblastoma No PDGFRA or MET ampl 4
a A PDGFRA nonsynonymous mutation was only identified in one of 10 tumors tested (patient 20).

Eleven patients (patients 2–6, 8, 10, 11, 13, 16, and 18) had a radiological diagnosis of diffuse intrinsic pontine glioma.
F, female; M, male; ampl, amplification.

3.3 Molecular studies 3.5 Quality of life

Seventeen patients underwent histologic confirmation of their tumors. HRQoL measures were collected from 23 patients. Only median total
Of 10 (59%) tumors that could undergo sequencing, only one harbored scores were reported due to the small sample size. We included evalua-
a nonsynonymous PDGFRA mutation (c.1432T > C p.S478P). tions up to week 6 of therapy because by that time point only nine par-
FISH analysis of PDGFRA and MET amplifications were avail- ticipants remained in the study. Only the physical, emotional, and social
able in 16 (94%) cases. While four tumors harbored PDGFRA ampli- domains were reported for the PedsQL 4.0, as these patients were not
fication, one of them focal, MET amplification occurred in two attending school.
cases. Only one tumor harbored co-amplification of both oncogenes Patients reported physical function decline starting at week 6 of
(Table 3). treatment. Patients did not report a significant increase in brain tumor-
specific symptom scores through the first 6 weeks of therapy. Parent-
proxy reports were then compared to those obtained from patients.
3.4 Outcome
Parent-proxy reported their child as having significantly lower physical
The median number of cycles of therapy received was one (range: function at week 6 and significantly lower emotional and social func-
0–12). Seven patients experienced clinical progression before com- tion at baseline and at weeks 2 and 4 compared to patients’ report.
pletion of cycle 1 of therapy. Only two patients received at least Parent-proxy reported their child as experiencing poorer cognitive
six cycles of therapy (Table 3). No objective radiologic responses function at weeks 2 and 4, more pain and hurt at week 2, poorer
were observed with this therapy. Only three patients remain movement and balance at baseline and at weeks 2 and 4, more nau-
alive. sea at week 4, more worry at baseline and at week 2, and overall lower
6 of 8 BRONISCER ET AL .

TA B L E 4 Most significant toxicities attributed to study agents during and after DLT-evaluation period

Twice-daily regimen Once-daily regimen

0 1 1 2a 2b
Dosage level n=3 n=5 n=6 n=8 n=3
Grade 1/2 Grade 3/4 Grade 1/2 Grade 3/4 Grade 1/2 Grade 3/4 Grade 1/2 Grade 3/4 Grade 1/2 Grade 3/4
Anemia 2 (67%) 0 3 (60%) 0 3 (50%) 0 5 (62%) 0 3 0
Neutropenia 0 0 0 0 1 (16%) 0 1 (12%) 0 1 (33%) 1 (33%)a
Thrombocytopenia 0 0 0 0 1 (16%) 0 1 (12%) 0 0 0
Diarrhea 2 (67%) 0 3 (60%) 1 (20%)a 4 (67%) 0 7 (87%) 1 (12%)a 3 (100%) 0
Nausea/vomiting 1 (33%) 0 2 (40%) 0 5 (83%) 0 6 (75%) 0 3 (100%) 0
Increase in 2 (67%) 0 3 (60%) 0 1 0 1 (12%) 0 2 (67%) 0
transaminases
Hypoalbuminemia 3 0 4 (80%) 0 1 (16%) 0 6 (75%) 0 3 (100%) 0
Hyponatremia 2 (67%) 1 (33%)a 2 (40%) 1 (20%) 1 (16%) 0 0 1 (12%) 0 0
Hypokalemia 1 (33%) 0 2 (40%) 2 (40%) 2 (33%) 0 4 (50%) 0 0 0
Hypophosphatemia 2 (66%) 1 (33%) 3 (60%) 1 (20%) 0 1 (16%) 5 (62%) 1 (12%) 2 (67%) 0
a
Proteinuria 2 (66%) 1 (33%) 3 (60%) 0 0 0 3 (37%) 0 3 (100%) 0
Rash 0 0 0 0 1 (16%) 0 6 (75%) 1 (12%)a 1 1 (33%)a
a
Fatigue 3 0 0 2 (40%) 4 0 2 (25%) 0 1 (33%) 0
a
Dose-limiting toxicity.

TA B L E 5 Summary of dasatinib and crizotinib pharmacokinetic parameters

Drug Dose (mg/m2 ) Day of therapy Number of patients Cmax (ng/ml) AUC0→t (ng/ml/hr)
Dasatinib 50 14 10 119.8 (54.1–385.8) 593.8 (388.1–1,521.1)
65 14 2 209.5 (95.7–323.2) 863.9 (721.9–1,006)
Crizotinib 100 1 3 49.1 (42–97.8) 1,288.9
100 14 3 392.6 (383–487.4) 5,706 (4,740.9–6,671.9)
130 1 5 95.7 (57.5–155.5) 2,189.4 (1,261.5–2,668.5)
130 14 5 443.9 (254.4–793.7) 5,486.1 (4,286.6–10,869)
165 1 9 155.5 (55.8–324.4) 3,360.1 (1,457.5–4,939.1)
165 14 7 155.9 (95.9–343.1) 1,933.8 (1,678.1–6,210.4)
215 1 8 278.8 (123.1–690.2) 5,326.3 (2,068.3–12,182.9)
215 14 4 261.9 (93.8–356.2) 3,118.9 (1,693.8–4,278.5)

Of note, results are shown as median (ranges in parentheses).

AUC0→t of crizotinib were t = ∞ for day 1 and t = 24 hr for day 14.

HRQoL from baseline until week 4 when compared to the patients’ Unfortunately, this combination was poorly tolerated and the recom-
reports. mended daily phase 2 dose represented 29% and 38% of the MTD of
dasatinib and crizotinib, respectively, when used as single agents in
children with solid tumors.15,16 Most of the toxicities observed, includ-
ing diarrhea, nausea and vomiting, electrolyte abnormalities, rash, and
4 DISCUSSION
hypoalbuminemia were predictable based on the adverse events pre-

The rational design of innovative clinical trials using small-molecule viously reported with these medications.15,16 However, the observa-

inhibitors for the treatment of children with HGGs and DIPGs should tion of severe proteinuria in the nephrotic range and hyponatremia was

be ideally based on multiple premises including selection of agents that unexpected. No objective radiologic responses were observed.

inhibit a biologically relevant target, use of medications with sufficient Although we initially designed our pharmacokinetic studies to eval-

CNS penetration, and if possible, proven activity against relevant pre- uate potential interactions between dasatinib and crizotinib, the lim-

clinical models. Since HGGs and DIPGs are highly refractory to therapy, ited number of patients evaluated prevented us from drawing firm con-

there is great enthusiasm for the use of promising treatment combina- clusions. The Cmax and AUC of dasatinib in our patients showed simi-

tions. lar results to those previously reported in children with DIPG.20 Like-

We report detailed clinical, pharmacokinetic, and pharmaco- wise, several pharmacokinetic parameters of crizotinib in our patients

dynamic data about the combination of dasatinib and crizotinib. (for example, Cmax , Tmax , AUC0→t , and apparent oral clearance)
BRONISCER ET AL . 7 of 8

were comparable to those reported recently in children with solid CONFLICT OF INTEREST
tumors despite significant methodologic differences between both The authors have no conflicts of interest to disclose in association with
studies (for example, drug formulation and pharmacokinetic sampling the work presented in this manuscript.
strategy).21
Our results demonstrating a decrease in p-AKT in PBMC after
ACKNOWLEDGMENTS
administration of dasatinib are interesting, although significant con-
cern remains that dasatinib does not reach clinically effective con- We thank Terri O'Neill for her expertise in performing FISH studies and
centrations in the CNS.20,22 The longitudinal analysis of plasma HGF Robbin Christensen for her help in producing the liquid formulation of
was performed based on the report that its expression was predic- crizotinib.
tive of response to c-Met inhibition in HGG preclinical models.23
Future studies using c-Met inhibitors in the treatment of patients
ORCID
with CNS cancers should further explore the association between con-
Alberto Broniscer http://orcid.org/0000-0001-9646-0463
centrations of HGF in plasma and cerebrospinal fluid and treatment
response.
The results of this clinical trial need to be interpreted in the context REFERENCES

of data available at its design. Dasatinib is one of the most potent com- 1. Ostrom Q, Gittleman H, Xu J, et al. CBTRUS statistical report: pri-
mary brain and other central nervous system tumors diagnosed
mercially available PDGFRA inhibitors. Multiple in vitro studies in HGG
in the United States in 2009–2013. Neuro Oncol. 2016;18:v1–v75.
or DIPG-derived cell lines with or without PDGFRA abnormalities have https://doi.org/10.1093/neuonc/now207.
shown activity of this agent, which seems to be mostly cytostatic.8,12,24 2. Paugh BS, Broniscer A, Qu C, et al. Genome-wide analyses iden-
However, the promiscuity of dasatinib in inhibiting multiple targets tify recurrent amplifications of receptor tyrosine kinases and cell-
probably accounts for a large share of its toxicities. Furthermore, cycle regulatory genes in diffuse intrinsic pontine glioma. J Clin Oncol.
2011;29:3999–4006. https://doi.org/10.1200/JCO.2011.35.5677.
as a substrate of p-glycoprotein and BCRP,22 it is now recognized
that its CNS penetration is more limited than initially recognized.20 3. Wu G, Broniscer A, McEachron TA, et al. Somatic histone H3
alterations in pediatric diffuse intrinsic pontine gliomas and non-
Finally, the recommended phase 2 dose reached in this clinical trial
brainstem glioblastomas. Nat Genet. 2012;44:251–253. https://doi.
provided short exposure to this agent considering that its t1/2 is quite org/10.1038/ng.1102.
short.20 4. Schwartzentruber J, Korshunov A, Liu XY, et al. Driver mutations in his-
Likewise, crizotinib was the only c-Met inhibitor commercially avail- tone H3.3 and chromatin remodeling genes in paediatric glioblastoma.
able at the start of this study. Although there was anecdotal evidence Nature. 2012;482:226–231. https://doi.org/10.1038/nature10833.
of its activity against MET-amplified glioblastoma,25 it is now clear 5. Sturm D, Witt H, Hovestadt V, et al. Hotspot mutations in H3F3A
that the CNS penetration of crizotinib is minimal and probably inad- and IDH1 define distinct epigenetic and biological subgroups
of glioblastoma. Cancer Cell. 2012;22:425–437. https://doi.org/
equate for the treatment of primary CNS cancers.25,26 One study
10.1016/j.ccr.2012.08.024.
recently reported promising activity of the combination of dasatinib
6. Korshunov A, Ryzhova M, Hovestadt V, et al. Integrated analysis
and cabozantinib, another c-Met inhibitor, against DIPG-derived cell of pediatric glioblastoma reveals a subset of biologically favor-
lines.8 able tumors with associated molecular prognostic markers. Acta
We believe that specific and better CNS-penetrant PDGFRA and Neuropathol. 2015;129:669–678. https://doi.org/10.1007/s00401-
015-1405-4.
c-Met inhibitors warrant further testing in children with CNS can-
cers. It would also be beneficial if preclinical studies could determine 7. Grasso CS, Tang Y, Truffaux N, et al. Functionally defined therapeutic
targets in diffuse intrinsic pontine glioma. Nat Med. 2015;21:555–559.
the subgroup of patients whose tumors are more likely to respond
https://doi.org/10.1038/nm.3855.
to PDGFRA and/or c-Met inhibition (for example, those harboring
8. Truffaux N, Philippe C, Paulsson J, et al. Preclinical evaluation of dasa-
PDGFRA and/or MET amplification). Although we performed detailed tinib alone and in combination with cabozantinib for the treatment
molecular analyses of available tumors in the context of a phase 1 of diffuse intrinsic pontine glioma. Neuro Oncol. 2015;17:953–964.
study, we did not select patients based on the presence of specific https://doi.org/10.1093/neuonc/nou330.
genetic abnormalities within their tumors. Of note, a focal amplifica- 9. Bender S, Gronych J, Warnatz HJ, et al. Recurrent MET fusion
tion of PDGFRA was found in one of the patients who received this genes represent a drug target in pediatric glioblastoma. Mat Med.
2016;22:1314–1320. https://doi.org/10.1038/nm.4204.
therapy for 6 months. Although no PDGFRA or MET amplifications
10. Shih AH, Holland EC. Platelet-derived growth factor (PDGF) and glial
were found in the tumor of the patient who received therapy for 12
tumorigenesis. Cancer Lett. 2006;232:139–147.
cycles, we were unable to sequence PDGFRA in his tumor. We were
11. Becher OJ, Hambardzumyan D, Walker TR, et al. Preclinical evalu-
unable to evaluate tumors for the recently described MET fusions
ation of radiation and perifosine in a genetically and histologically
either.9 accurate model of brainstem glioma. Cancer Res. 2010;70:2548–2557.
Based on the data provided, particularly the significant toxicity and https://doi.org/10.1158/0008-5472.CAN-09-2503.
low recommended phase 2 dose, we do not feel that further testing of 12. Paugh BS, Zhu X, Qu C, et al. Novel oncogenic mutations in
the combination of dasatinib and crizotinib in children with CNS can- pediatric high-grade gliomas. Cancer Res. 2013;73:6219–6229.
https://doi.org/10.1158/0008-5472.CAN-13-1491.
cers is warranted.
8 of 8 BRONISCER ET AL .

13. Funato K, Major T, Lewis PW, Allis CD, Tabar V. Use of 22. Chen Y, Agarwal S, Shaik NM, Chen C, Yang Z, Elmquist WF. P-
human embryonic stem cells to model pediatric gliomas with glycoprotein and breast cancer resistance protein influence brain
H3.3K27M histone mutation. Science. 2014;346:1529–1533. distribution of dasatinib. J Pharmacol Exp Ther. 2009;330:956–963.
https://doi.org/10.1126/science.1253799. https://doi.org/10.1124/jpet.109.154781.
14. Awad AJ, Burns TC, Zhang Y, Abounader R. Targeting MET 23. Zhang Y, Farenholtz KE, Yang Y, et al. Hepatocyte growth factor
for glioma therapy. Neurosurg Focus. 2014;37:E10. https://doi. sensitizes brain tumors to c-MET kinase inhibition. Clin Cancer
org/10.3171/2014.9.FOCUS14520. Res. 2013;19:1433–1444. https://doi.org/10.1158/1078-0432.
15. Aplenc R, Blaney SM, Strauss LC, et al. Pediatric phase I trial and CCR-12-2832.
pharmacokinetic study of dasatinib: a report from the children's 24. Koschmann C, Zamler D, MacKay A, et al. Characterizing and target-
oncology group phase I consortium. J Clin Oncol. 2011;29:839–844. ing PDGFRA alterations in pediatric high-grade glioma. Oncotarget.
https://doi.org/10.1200/JCO.2010.30.7231. 2016;7:65696–65706.
16. Mossé YP, Lim MS, Voss SD, et al. Safety and activity of crizo- 25. Chi AS, Batchelor TT, Kwak EL, et al. Rapid radiographic and clin-
tinib for paediatric patients with refractory solid tumours or ical improvement after treatment of a MET-amplified recurrent
anaplastic large-cell lymphoma: a Children's Oncology Group glioblastoma with a mesenchymal-epithelial transition inhibitor. J
phase 1 consortium study. Lancet Oncol. 2013;14:472–480. Clin Oncol. 2012;30:e30–e33. https://doi.org/10.1200/JCO.2011.38.
https://doi.org/10.1016/S1470-2045(13)70095-0. 4586.
17. Broniscer A, Tatevossian RG, Sabin ND, et al. Clinical, radiologi- 26. Metro G, Lunardi G, Floridi P, et al. CSF concentration of crizo-
cal, histological and molecular characteristics of paediatric epithe- tinib in two ALK-positive non-small-cell lung cancer patients with
lioid glioblastoma. Neuropathol Appl Neurobiol. 2014;40:327–336. CNS metastases deriving clinical benefit from treatment. J Thor
https://doi.org/10.1111/nan.12093. Oncol. 2015;10:e26–e27. https://doi.org/10.1097/JTO.000000
18. Varni JW, Burwinkle TM, Katz ER, Meeske K, Dickinson P. The 0000000468.
PedsQL in pediatric cancer: reliability and validity of the pedi-
atric quality of life inventory generic core scales, multidimen-
sional fatigue scale, and cancer module. Cancer. 2002;94:2090–2106.
https://doi.org/10.1002/cncr.10428.
SUPPORTING INFORMATION
19. Palmer SN, Meeske KA, Katz ER, Burwinkle TM, Varni JW. The PedsQL
brain tumor module: initial reliability and validity. Pediatr Blood Cancer. Additional Supporting Information may be found online in the support-
2007;49:287–293. https://doi.org/10.1002/pbc.21026. ing information tab for this article.
20. Broniscer A, Baker SD, Wetmore C, et al. Phase I trial, phar-
macokinetics, and pharmacodynamics of vandetanib and dasatinib
in children with newly diagnosed diffuse intrinsic pontine glioma. How to cite this article: Broniscer A, Jia S, Mandrell B,
Clin Cancer Res. 2013;19:305–308. https://doi.org/10.1158/1078- et al. Phase 1 trial, pharmacokinetics, and pharmacodynam-
0432.CCR-13-0306.
ics of dasatinib combined with crizotinib in children with
21. Balis FM, Thompson PA, Mosse YP, et al. First-dose and steady- recurrent or progressive high-grade and diffuse intrinsic pon-
state pharmacokinetics of orally administered crizotinib in children
tine glioma. Pediatr Blood Cancer. 2018;e27035. https://doi.
with solid tumors: a report on ADVL0912 from the Children's Oncol-
ogy Group Phase 1/Pilot Consortium. Cancer Chemother Pharmacol. org/10.1002/pbc.27035
2017;79:181–187. https://doi.org/10.1007/s00280-016-3220-6.