Despite frequent and durable responses for androgen
receptor (AR)-targeted therapies in castration-resistant
prostate cancer (CRPC), resistance inevitably occurs. Resis-
tance is attributed to genomic and metabolic reactivation of
the AR supported by complex and context-dependent
activation of stress response, kinase signaling, and devel-
opmental pathways
[1]. Therefore, additional therapies are
needed to enhance the armamentarium of efficacious CRPC
drugs, and the current best treatment sequences need to be
identified
[2]. In parallel, newer and less expensive genomic
assays enable the evaluation of efficacy of new agents based
on the genomic status of individual patients in line with a
precision urologic oncology approach
[3] .A promising new agent for CRPC patients is the novel
AR inhibitor darolutamide (ODM-201). Darolutamide
inhibits AR nuclear translocation and has a distinctly
different chemical structure than the currently known AR
inhibitors enzalutamide (ENZ), bicalutamide, flutamide,
and apalutamide (ARN-509). It has higher binding affinity
to the AR compared with ENZ and apalutamide and does
not cross the blood-brain barrier which decreases the
likelihood of seizures
[4] .Results from the open-label
phase 1
–
2 dose escalation trial (ARADES) showed that
darolutamide had a favorable safety profile, was well
tolerated and provided anticancer activity comparable to
ENZ
[5,6] .Two randomized phase 3 trials evaluating
efficacy and safety of darolutamide versus placebo in
high-risk nonmetastatic CRPC (ARAMIS trial, Clinical-
Trials.gov identifier NCT02200614) and of darolutamide
in combination with standard androgen deprivation
therapy and docetaxel in patients with metastatic
hormone sensitive prostate cancer (PCa; ARASENS trial,
ClinicalTrials.gov identifier NCT02799602) are currently
recruiting.
With a distinctly different chemical structure compared
with currently approved AR antagonists, we hypothesized
that darolutamide might exhibit differential antagonist
versus agonist profiles in AR mutated PCa cells, and under
specific genomic states provide AR pathway suppression in
preclinical models of ENZ-resistant (ENZ-R) PCa.
To evaluate darolutamide in ENZ-R PCa, we used the
LNCaP-derived MR49F cell line that inherits the F877L
mutation, which has been described to confer resistance to
ENZ
[7]. Methodologies are described in Supplementary
data. In vitro, darolutamide significantly and dose-depen-
dently inhibited MR49F cell growth
( Fig. 1 A) and AR
transcriptional activity
( Fig. 1B) compared with ENZ, which
acted as an agonist at doses above 1
m
M. Fluorescence
microscopy demonstrated inhibition of AR nuclear translo-
cation by darolutamide in MR49F cells, but not by ENZ
(Supplementary Fig. 1A). Darolutamide significantly de-
creased both
prostate-specific antigen
(
PSA
) gene
( Fig. 1 C)
and protein expression
( Fig. 1 D) in ENZ-R MR49F cells,
where ENZ acted as an agonist. AR protein expression was
not affected by darolutamide (Supplementary Fig. 1B).
Based on these in vitro results, we compared activity of
darolutamide, ENZ, and a vehicle in subcutaneous MR49F
xenografts in mice. Darolutamide significantly inhibited
both tumor growth and serum PSA levels
( Fig. 1E), and
significantly prolonged survival of mice bearing ENZ-R
tumors, compared with ENZ or vehicle
( Fig. 1F). Waterfall
plots in
Fig. 1G illustrate individual tumor responses from
baseline between groups after 3 wk of treatment for both
tumor volume (left) and serum PSA (right). While apoptotic
rates were similar (data not shown), Ki67 staining of
representative samples of xenografts indicates in-vivo
antiproliferative effects of darolutamide compared with
ENZ or vehicle (Supplementary Fig. 1C).
We next tested the effects of darolutamide on the
transcriptional activity of a panel of mutated AR variants,
previously reported in literature or detected in a cohort of
62 CRPC patients at disease progression after treatment
with ENZ, abiraterone, bicalutamide, or docetaxel, using
plasma circulating tumor DNA (ctDNA) sequencing
[8,9]. We compared the activity of ENZ and darolutamide
on a panel of 25 ARmutants using a luciferase reporter assay
in AR-negative PC3 cells. Darolutamide behaved similarly to
ENZ in inhibiting most AR variants (Supplementary Fig. 2).
Importantly, darolutamide inhibited transcriptional activity
of three AR variants with partial agonism to ENZ: F877L,
F877L/T878A, and H875Y/T878A
( Fig. 2A). This data suggest
that none of these AR mutants induced under the selective
pressures of AR pathway inhibition are cross-resistant to
darolutamide.
To gain atomic insights into the mode of action of
darolutamide and ENZ in the setting of F877L AR mutation,
structure-based computer modeling was performed. ENZ
establishes a key hydrogen bond and a
p
–
p
stacking
interaction with L877 and F765 residues, respectively. In
contrast, darolutamide adopted a different binding pose in
the ligand-binding site of F877L AR mutant
( Fig. 2B). Instead
darolutamide maintained its binding conformation in the
F877L-mutated pocket through hydrogen and Van der
Waals interactions
( Fig. 2 B); the stability of this conforma-
tion was reflected in quantitative structure activity rela-
tionship models
( Fig. 2 C).
In addition, we recently reported a previously unchar-
acterized AR mutant T878G (substitution of threonine for
glycine at the 878 position), for which bicalutamide and
flutamide demonstrated agonist behavior
[10]. Notably,
darolutamide was the only drug that was modeled to bind
the pocket of T878G mutant
( Fig. 2D) without partially
activating it at high concentrations
( Fig. 2E).
Based on these preclinical findings of darolutamide in
ENZ-R PCa, a randomized phase 2 trial of darolutamide
versus ENZ in metastatic CRPC is planned through the
Canadian Cancer Trial Group using ctDNA assays to study
effects of AR mutations on response rates. Besides providing
support for a sequencing trial of darolutamide after ENZ, our
results show the first head-to-head comparison of darolu-
tamide and ENZ over a large range of ARmutants detected in
patients with CRPC and no clinical trial directly comparing
both drugs is underway. Therefore, our results are
contemporarily the most comprehensive genomic and
biologic comparison of darolutamide versys ENZ in the
preclinical setting.
The findings from this study have several implications
for clinical practice. First, preclinical anticancer activity of
E U R O P E A N U R O L O GY 7 3 ( 2 0 18 ) 4
–
8
5




