Preclinical patient‐derived modeling of castration‐resistant prostate cancer facilitates individualized assessment of homologous recombination repair deficient disease

Abstract

The use of mutation analysis of homologous recombination repair (HRR) genes to estimate PARP‐inhibition response may miss a larger proportion of responding patients. Here, we provide preclinical models for castration‐resistant prostate cancer (CRPC) that can be used to functionally predict HRR defects. In vitro, CRPC LNCaP sublines revealed a HRR defect and enhanced sensitivity to olaparib and cisplatin due to impaired RAD51 expression and recruitment. Ex vivo‐induced castration‐resistant tumor slice cultures or tumor slice cultures derived directly from CRPC patients showed increased olaparib‐ or cisplatin‐associated enhancement of residual radiation‐induced γH2AX/53BP1 foci. We established patient‐derived tumor organoids (PDOs) from CRPC patients. These PDOs are (i) morphologically similar to their primary tumors and (ii) genetically clustered with prostate cancer but not with normal prostate or other tumor entities. Using these PDOs, we functionally confirmed the enhanced sensitivity of CRPC patients to olaparib and cisplatin. Moreover, olaparib but not cisplatin significantly decreased the migration rate in CRPC cells. Collectively, we present robust patient‐derived preclinical models for CRPC that recapitulate the features of their primary tumors and enable individualized drug screening, allowing translation of treatment sensitivities into tailored clinical therapy recommendations.

Full text

Preclinical patient-derived modeling of castration-resistant
prostate cancer facilitates individualized assessment of
homologous recombination repair deficient disease
Mohamed E. Elsesy
1,2
, Su Jung Oh-Hohenhorst
3,4
, Christoph Oing
1,5
, Alicia Eckhardt
1,6,7
,
Susanne Burdak-Rothkamm
1,8
, Malik Alawi
9
, Christian M€
uller
1,9
, Ulrich Sch€
uller
6,10
,
Tobias Maurer
3,11
, Gunhild von Amsberg
3,12
, Cordula Petersen
1
, Kai Rothkamm
1
and
Wael Y. Mansour
1,5
1 Department of Radiotherapy and Radiooncology, University Medical Center Hamburg-Eppendorf, Germany
2 Department of Tumor Biology, National Cancer Institute, Cairo University, Giza, Egypt
3 Martini-Klinik Prostate Cancer Center, University Medical Center Hamburg-Eppendorf, Germany
4 Centre de Recherche du Centre Hospitalier de l’Universit
e de Montr
eal (CRCHUM), QC, Canada
5 Mildred Scheel Cancer Career Center HaTriCS4, University Medical Center Hamburg-Eppendorf, Germany
6 Department of Pediatric Hematology and Oncology, University Medical Center Hamburg-Eppendorf, Germany
7 Research Institute Children’s Cancer Center Hamburg, Germany
8 Department of Molecular & Clinical Cancer Medicine, University of Liverpool, UK
9 Bioinformatics Core, University Medical Center Hamburg-Eppendorf, Germany
10 Institute of Neuropathology, University Medical Center Hamburg-Eppendorf, Germany
11 Department of Urology, University Medical Center Hamburg-Eppendorf, Germany
12 Department of Oncology, University Cancer Center Hamburg Eppendorf, University Medical Center Hamburg-Eppendorf, Germany
Keywords
castration-resistant prostate cancer; ex vivo
tumor cultures; homologous recombination;
PARP inhibition; patient-derived organoids
Correspondence
W. Y. Mansour, Laboratory of Radiobiology
and Experimental Radiooncology,
Department of Radiotherapy and Radiation
Oncology, University Medical Center
Hamburg-Eppendorf, Martinistr. 52, 20251,
Hamburg, Germany
Fax: +49 40 7410 55139
Tel: +49 40 7410 53831
E-mail: w.mansour@uke.de
(Received 10 August 2022, revised 24
October 2022, accepted 23 January 2023,
available online 16 March 2023)
doi:10.1002/1878-0261.13382
The use of mutation analysis of homologous recombination repair (HRR)
genes to estimate PARP-inhibition response may miss a larger proportion
of responding patients. Here, we provide preclinical models for castration-
resistant prostate cancer (CRPC) that can be used to functionally predict
HRR defects. In vitro, CRPC LNCaP sublines revealed an HRR defect
and enhanced sensitivity to olaparib and cisplatin due to impaired RAD51
expression and recruitment. Ex vivo-induced castration-resistant tumor slice
cultures or tumor slice cultures derived directly from CRPC patients
showed increased olaparib- or cisplatin-associated enhancement of residual
radiation-induced cH2AX/53BP1 foci. We established patient-derived
tumor organoids (PDOs) from CRPC patients. These PDOs are morpho-
logically similar to their primary tumors and genetically clustered with
prostate cancer but not with normal prostate or other tumor entities. Using
these PDOs, we functionally confirmed the enhanced sensitivity of CRPC
patients to olaparib and cisplatin. Moreover, olaparib but not cisplatin sig-
nificantly decreased the migration rate in CRPC cells. Collectively, we pre-
sent robust patient-derived preclinical models for CRPC that recapitulate
the features of their primary tumors and enable individualized drug
Abbreviations
Abi, abiraterone; ADT, androgen deprivation therapy; ARN, apalutamide; Bic, bicalutamide; CFA, colony formation assay; CisER, cisplatin-
induced enhancement ratio; COSMIC, catalog of somatic mutations in cancer; CR, castration-resistant; CRPC, castration-resistant prostate
cancer; CSS, charcoal-stripped serum; DAPI, 40-6-diamidino-2-phenylindole; DMEM, Dulbecco’s modified Eagle’s medium; DSB, double
strand break; DT, doubling time; EdU, 5-ethynyl-20-deoxyuridine; ENZA, enzalutamide; H&E, hematoxylin and eosin; HP, hormone-proficient;
HRR, homologous recombination repair; mCRPC, metastatic castration-resistant prostate cancer; mHSPC, metastatic hormone-sensitive
prostate cancer; MSI, microsatellite instability; MTT, 3-(4, 5-dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide; PARPi, PARP inhibitors;
PCa, prostate cancer; PDO, patient-derived tumor organoids; PiER, PARPi-induced enhancement ratio; RNA-SEQ, RNA-sequencing; SBS,
single base substitutions; SF, surviving fraction; WGS, whole genome sequencing.
1129Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.

screening, allowing translation of treatment sensitivities into tailored clini-
cal therapy recommendations.
1. Introduction
Prostate cancer (PCa) is the most frequent malignant
tumor in males; approximately one in six men will be
diagnosed within their lifetime [1]. PCa is clinically vari-
able, with often indolent and low-risk disease that will
not pose a health threat over one’s lifetime, but also
aggressive phenotypes with rapid disease progression
and treatment resistance. In the metastatic stage of dis-
ease, patients have no option for cure despite significant
progress with new therapeutic treatment strategies.
Androgen deprivation therapy (ADT) induces tumor
regressions in the vast majority of patients with meta-
static hormone-sensitive prostate cancer (mHSPC) and
treatment intensification with docetaxel and/or andro-
gen receptor (AR) pathway targeting agents has
improved overall survival in randomized phase 3 clini-
cal trials [2]. However, even at this early stage, the het-
erogeneity of the disease becomes obvious, with a
nearly doubling of life expectancy in patients with meta-
chronous development of metastases compared with
patients with primary metastatic disease [3]. After sys-
temic treatment initiation for mHSPC, virtually all
patients progress to castration-resistant prostate cancer
(CRPC) as a result of selection and/or acquired resis-
tance [4]. CRPC carries a worse prognosis, with esti-
mated median survival times of 16–18 months from the
onsetofCRPCprogression[5–7]. Although additional
treatments exist for such patients, including docetaxel,
enzalutamide, abiraterone, radium-223, sipuleucel-T and
cabazitaxel, these treatment successes are not long-
lasting with only a modest overall survival benefit [8].
Comprehensive genomic characterization of CRPC
identified frequent mutations in DNA repair genes, spe-
cifically those involved in homologous recombination
repair (HRR) [9,10], with a frequency reaching in some
studies to approximately 40% among patients with met-
astatic CRPC (mCRPC) [11]. Consequently, the para-
digm of PARP inhibitors (PARPi)-mediated synthetic
lethality or any other chemotherapy that targets HRR-
deficient tumors expands the management options for
mCRPC. However, not all HRR defects respond
equally to PARP inhibition. This led, for example, to a
restricted approval of olaparib in Europe to patients
with BRCA1/2 alterations, since the pivotal PRO-
FOUND trial revealed the highest PARP inhibitor effi-
cacy for patients carrying these respective mutations. In
contrast, for alterations in other HRR genes, e.g.
ATM, the effects were less convincing [12,13]. The story
gets even more complicated when comparing two recent
phase 3 trials on the efficacy of a combination of abira-
terone/prednisone and olaparib (PROpel) or niraparib
(MAGNITUDE). Whereas in the PROpel trial, pre-
sumably due to a BRCAness effect of abiraterone, a
prolonged progression-free survival was also observed
for the combination treatment in HRR wild-type
patients, the MAGNITUDE trial did not show such an
advantage [14,15]. This illustrates that we clearly do not
yet comprehensively understand the molecular patho-
logical processes of HRR in PCa. Furthermore, to date,
HRR defects are basically detected by large-scale
sequencing analysis, a costly process. Despite these
comprehensive analyses, information on the actual
effect of these alterations on HRR function in the
tumor cell is limited, which highlights the need for valid
assays to functionally detect HRR defects.
Preclinical and translational research into novel syn-
thetic lethality concepts is, however, hampered by the
lack of appropriate preclinical models for such disease.
Various preclinical models have been introduced to
advance CRPC research. Most studies relied on using
immortalized cell lines grown in two-dimensional (2D)
cultures or xenografts of such cell lines in immuno-
compromised animals. While these PCa cell lines are
readily available and simple to use, only a limited
number of cell lines are available and they are far
from being authentic exemplars of CRPC due to their
prolonged time in culture. In addition, the available
cell lines fail to capture the various aspects of hetero-
geneity of PCa. Further, commonly used in vitro
CRPC models, such as DU145 and PC3 cells, neither
reflect the diversity of this disease, nor do they accu-
rately predict patient sensitivity to treatment [16,17].
Although several genetically engineered mouse models
exist [18], they fail to generally model clinical CRPC,
as castration-resistant tumors in mice do not depend
upon AR signaling mechanisms [19]. Therefore, three-
dimensional (3D) culture models of PCa are currently
gaining increasing attention as preclinical models that
better mimic the in vivo tumor biology and microenvi-
ronment. Ex vivo culturing of freshly collected tumor
slices as well as patient-derived organoids (PDOs) are
considered promising 3D models. We and others have
shown previously that tissue slice cultures show
1130 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

comparability with the original tumor, preserving the
tumor morphology and its microenvironment [20–23].
However, although the ex vivo assay allows a variety
of functional analyses and biological readouts such as
DSB repair and apoptosis, these do not include robust
analysis of survival rates, one of the most important
endpoints of drug sensitivity analysis.
Patient-derived tumor organoids are 3D tissue cul-
tures that promise to enable the validation of preclinical
drug testing in precision medicine and co-clinical trials
by modeling tumors for predicting therapeutic responses
with more reliable efficacy. Although there has also
been a significant improvement in the generation of
PDOs from PCa patients, their long-term propagation
in culture has remained challenging. To the best of our
knowledge, there are only a few studies reporting suc-
cessful establishment of PDOs from PCa specimens
[24–29],withasuccessrateof<20%. Furthermore, the
capacity for long-term maintenance of these PDOs is
variable and limited [30]. In the current study, we pro-
vide different preclinical models derived from CRPC
in vivo and the rationale for using these models to reca-
pitulate and predict the response of individual CRPCs
to PARP inhibitors, that is, olaparib and cisplatin.
2. Materials and methods
2.1. Cell culture, drugs and X-irradiation
LNCaP prostate cancer cells (ATCC, Manassas, VA,
USA) were grown in Dulbecco’s modified Eagle’s
medium (DMEM, Gibco, Paisley, UK) supplemented
with 10% fetal calf serum (FCS), 100 UmL
1
penicil-
lin and 100 mgmL
1
streptomycin at 37 °C with 10%
CO
2
. Novel antiandrogen-resistant sublines LNCaP-
ARN509 and LNCaP-abi were generated by long-term
treatment with apalutamide (ARN-509) (up to 40 lM)
and abiraterone acetate (up to 10 lM), respectively,
until acquiring androgen-independent growth feature
[31]. C4-2B-Enza cells (kindly provided by Prof. C. P.
Evans, UC Davis School of Medicine, Sacramento,
CA, USA) were maintained in medium containing
20 lM enzalutamide. Abiraterone acetate, apalutamide
and enzalutamide were kindly provided by Janssen
Cilag GmbH, Neuss, Germany. LNCaP-abl cells (a
gift from Prof. Z. Culig, Medical University Inns-
bruck, Austria) were grown in DMEM supplemented
with 10% Charcoal Stripped FBS (Sigma-Aldrich,
Taufkirchen, Germany) [32]. All cell lines tested nega-
tive for mycoplasma contamination. Irradiation was
performed as previously described (200 kV, 15 mA,
additional 0.5-mm Cu filter at a dose rate of
0.8 Gymin
1
)[15].
2.2. Cell lines authentication
All cell lines used in the current study have been authen-
ticated before executing the experiments. Authentication
of cell lines used in the current study was performed in
our laboratory. The profiles for all cell lines have been
compared and matched with the STR profile database.
2.3. Proliferation assay
Proliferation assay was performed as previously
described [31]. Briefly, cells were cultured in triplicate
in 6-well plates before treatment. For LNCaP-abl sub-
line, cells were seeded in CS-FCS-supplemented
medium for 18–24 h before changing to the FBS full
medium for the treatment. To assess the effect of any
treatment regimes, the cell number was determined
with a Beckman Coulter cell counter (Life Science,
Beckman Coulter cell counter, Krefeld, Germany) at
3-, 6- and 10-days post-treatment. In all experiments,
media with or without drugs were changed twice dur-
ing the 10-day treatment course.
2.4. 3D colony formation assay
3D colony formation assay (CFA) for cell lines was per-
formed by mixing cells in cold-reduced growth factor
basement membrane extract (RGF BME) type 2 (R&D
Systems, Minneapolis, MN, USA) and platted at 2000
cells per dome onto 24-well plates. For tumor organoids
3D CFA, tumor organoids were first harvested and
sheared into single cells before being mixed in cold BME
and plated at 4000 cells per dome. Upon completed
gelation, different concentrations of olaparib or cisplatin
as well as DMSO controls were added in triplicate
in 500 lL of corresponding medium. After 3–4weeks,
colonies (tumoroids or 3D cell cultures) were
stained with 0.5 mgmL
1
3-(4,5-dimethylthiazolyl-2)-
2, 5-diphenyltetrazolium bromide (MTT) for 1.5 h. For
organoid colonies, medium was removed and BME
domes were dissolved using Cultrex
TM
Organoid Harvest-
ing Solution. MTT-stained 3D cellular colonies and
tumoroids were photographed using REBEL Microscopy
(ECHO, San Diego, CA, USA) and analyzed using
IMAGE-J. Surviving fractions (SFs) were calculated by nor-
malization to the plating efficiency of the untreated con-
trol. DMSO was used as a control at the same
concentration.
2.5. Immunofluorescence
After treatment, cells cultured on coverslips were washed
and fixed with 4% paraformaldehyde/PBS for 10 min.
1131Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

Fixed cells were permeabilized with 0.2% Triton X-100/
PBSonicefor5minandincubatedfor1hatroomtem-
perature with primary antibodies: mouse monoclonal
anti-phospho-S139-H2AX antibody (Millipore, Darm-
stadt, Geramay) at a dilution of 1 : 500 and rabbit poly-
clonal anti- 53BP1 antibody (Novus, Braunschweig,
Germany) at a dilution of 1 : 500 or rabbit polyclonal
anti-RAD51 (Sigma Aldrich, Taufkirchen, Germany ) at
a dilution of 1 : 500. After being washed three times with
cold PBS, the cells were incubated for 1 h with secondary
anti-mouse Alexa Fluor 594 (Invitrogen) at a dilution of
1 : 500, and anti-rabbit Alexa Fluor 488 (Invitrogen) at a
dilution of 1 : 600. The nuclei were counterstained
with 40-6-diamidino-2-phenylindole (DAPI, 10 ngmL
1
).
Slides were mounted in Vectashield mounting medium
(Vector Laboratories, Newark, CA, USA). Immunofluo-
rescence of cultured tumor tissue was performed as previ-
ously described [21]. Fluorescence microscopy was
performed using the Zeiss AxioObserver.Z1 microscope
(objectives: 920, resolution 0.44 lm; Plan Apo 63/1.4 Oil
DICII, resolution 0.24 lm; and filters: Zeiss 43, Zeiss 38,
Zeiss 49, G€
ottingen, Germany). Z-stacks of semi-confocal
images were obtained using the ZEISS APOTOME,ZEISS AXIO-
CAM MRM and ZEISS AXIOVISION software. For DSB analy-
sis, fields of view were taken per time point or treatment
with a minimum of 100 cells (cell lines) or 50 cells (tumor
tissue). All staining was performed in duplicate. DSBs
were analyzed using IMAGE-JandDAPI-basedimage
masks and normalized to single nucleus values [21].
2.6. Western blot
Whole cell lysates were subjected to western blot as
previously described [33,34]. RAD51 immunoblot anal-
ysis was performed with the rabbit anti-RAD51
(Merck, Darmstadt, Germany, Cat#PC130). Beta-actin
was immunoblotted by mouse anti-beta-actin
(Sigma Aldrich, Taufkirchen, Germany, Cat#A1978)
and used as a loading control. Goat-anti-mouse IgG-
Alexa Fluor 594 (Molecular Probes, Sigma Aldrich,
Taufkirchen, Germany, Cat#A11005) and goat-anti-
rabbit IgG-AlexaFluor 488 (Molecular Probes,
Cat#A11008) secondary antibodies were used. Mem-
branes were developed and analyzed using LiCor Bio-
sciences (Lincoln, NE, USA) at room temperature.
2.7. Patient sample collection
Fresh PCa tissue was obtained from patients with high-
risk PCa according to D’Amico risk stratification, who
underwent radical prostatectomy at Martini-Klinik,
Prostate Cancer Center Hamburg, Germany between
2019 and 2022. Immediately after resection, one to two
punch biopsies were taken by the surgeon in palpable
tumor areas. The biopsies were collected in culture
media and immediately taken to the laboratory. Pseudo-
nymized biopsies were processed within 30 min after
resection. The laboratory received a final pathology
report containing the Gleason score, PSA status and age
of each pseudonymized patient for clinical analysis. The
project was approved by the local ethics committee
[Ethik-Kommission der
€
Arztekammer Hamburg] with
the project number PV7007. The study methodologies
conformed to the standards set by the Declaration of
Helsinki. All experiments were undertaken with the
understanding and written consent of each subject.
2.8. Tissue slice cultures
Ex vivo tissue slice cultures were prepared as previ-
ously described [20]. Briefly, 300-lm slices were cut
using the MacIllvine tissue chopper and placed on
MillicellÒcell culture inserts (0.4 lm, 30 mm diame-
ter, Merck), which were inserted in 6-well dishes con-
taining 1 mL DMEM supplemented with 10% FCS
and incubated at 37 °C. Prior to ex vivo treatment, the
tissues slices were incubated for 1 day for recovery
and re-oxygenation. To monitor proliferation, un-
irradiated slice cultures were incubated with 5-ethynyl-
20-deoxyuridine (EdU, 1 : 1000; Click-iT Assay Kit,
Invitrogen) overnight for 16 h. All slices were addi-
tionally treated with pimonidazole (200 lM,
Hypoxyprobe, Burlington, MA, USA) 2 h before fixa-
tion to monitor tissue hypoxia.
2.9. Histology and imaging
PCa tumor tissues and PDOs were prepared as previ-
ously described [21]. Briefly, either tissues or tumoroids
were fixed using 4% PFA (Merck) followed by wash-
ing in 25% sucrose twice each for 1 h. The samples
were then frozen in TissueTekÒ(Sakura Finetek,
Alphen aan den Rijin, Netherlands) and stored at
80 °C. Using the Cryo Star NX70 Microtome
(Thermo Scientific) sectioning was performed to pre-
pare cryoslices (5 lm). Histological analysis was per-
formed by standard hematoxylin and eosin (H&E)
staining and percentage of cancer cells and Gleason
score were determined by an experienced PCa patholo-
gist. Immunohistochemistry was performed using
antibodies against AMACR (Thermo Scientific,
Regensburg, Germany, PA5-82739, 1 : 250), and Ki67
(Abcam, Cambridge, UK, ab15580, 1 : 250). Images
were acquired using ZEISS Axio Scan.Z1 Slide Scan-
ner and photos were then processed using netScopeÒ
Viewer.
1132 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

2.10. Prostate tumor tissue processing and
organoid establishment
Organoid establishment and culture were adapted
from Drost et al. [25] and Gao et al. [30]. Briefly,
prostate tumor specimens from patients who under-
went radical prostatectomy were received in
adDMEM F12+++ [advanced DMEM F12 (Thermo
Scientific) supplemented with 10 mMHEPES (Thermo
Scientific), GlutaMAX (Thermo Scientific), and peni-
cillin/streptomycin (Thermo Scientific)]. Tumor tissue
samples were first washed three times with PBS and
then placed in 3.5-cm culture dish where they under-
went mechanical dissection into small pieces, which
were then placed in 5 mgmL
1
Collagenase type II
(Sigma-Aldrich) in adDMEM +++ with 10 lM ROCK
inhibitor and incubated in a 37 °C shaker for 30–
90 min for digestion. After tissue digestion, the
suspension was passed through a 50-lm cell strainer
(Sysmex) before being washed with adDMEM F12
+++ and finally suspended in cold BME. Drops of
BME cell suspension each 40 lL were allowed to
solidify for 30 min at 37 °C onto a pre-warmed 24-
well culture plate. After stabilization of Matrigel-
containing cells, 500 lL of complete organoid medium
(Table S1) was added. Fresh medium was replenished
every 3–4 days during organoid growth. Organoids
were passaged every 4–6 weeks. During passaging, the
organoid droplets were mechanically sheared through
P1000 pipet tip and incubated with TrypLE Express
containing 10 lM ROCK inhibitor for a maximum of
5 min at 37 °C. The resulting cell clusters and single
cells were washed and re-plated following the protocol
described above.
2.11. Ex vivo induction of castration-resistant
status
To induce castration-resistant phenotype ex-vivo, PCa
tissues derived from na
€
ıve PCa patients were cultured
in hormone-depleted condition (DMEM supplemented
with CS-FBS containing 10 lM abiraterone). Each
sample was cultured for up to 6 weeks in either
hormone-proficient or -deficient conditions and the
culture medium was refreshed every 3 days. The prolif-
erative marker Ki67 was used to prove the attained
castration-resistant phenotype through proliferative
activity. The tumor tissue cultured in hormone-
deficient condition that still showed Ki67 proliferative
index similar to its counterpart slice cultured in hor-
mone proficient medium, was considered a castration-
resistant sample.
2.12. Migration assay
Chemotaxis assay was performed in 24-well Transwell
plate using 8-lm pore-size (Corning
Ò
BioCoat
Ò
,
354578). Either LNCaP or LNCaP-ARN509 cells were
harvested and re-suspended in FBS-free DMEM
medium at concentrations of 2 910
5
cells in 0.2 mL,
and then seeded into the upper chamber of a 24-well
plate. The lower chambers were filled with 0.7 mL
DMEM containing 10% FBS to act as an attractant.
Cells were incubated for 36 h. At the end of the exper-
iment, cells that migrated into the reverse side of the
Transwell membrane were fixed with 70% ethanol,
stained with 0.2% crystal violet, and then photo-
graphed using REBEL Microscopy and analyzed using
IMAGE-J.
2.13. DNA methylation profiling
Total DNA was isolated from PCa cell lines, tissues or
organoid cultures using Qiagen
Kit (Qiagen, Hilden, Germany). The Illumina Human-
Methylation450 BeadChip (450 K) arrays were used to
analyze genome-wide DNA methylation patterns of
tissues or organoids. Only sites covered by at least
three reads were considered for analysis. For each
sample, the percentage of methylation per site (beta
value) was computed. Average hierarchical clustering
of samples was performed by ‘1-Pearson’s correlation
coefficient’ as distance measured on the n=10 000
CpG sites showing the highest standard deviation
across the cohort. Several samples of the following
datasets were included as a reference set: TCGA-
BRCA, TCGA-Lung, TCGA-GBM, TCGA-PRAD,
GSE112047,GSE38240,GSE83842 and glioblastoma
samples (tissue and cell lines) and lung carcinoma sam-
ples were from UKE.
2.14. Whole genome sequencing (WGS)
WGS was performed by Novogene (Sacramento, CA,
USA). WGS data analysis was performed by the Bio-
informatics core facility at University Medical Center
Hamburg-Eppendorf (UKE), Hamburg, Germany.
Reads were aligned to the human genome assembly
GRCh38 using bwa mem [35] and structural and short
somatic mutations were labeled using Manta [36] and
Strelka2 [37], respectively. Variants with a depth below
60 or presence in the Genome aggregation database
(https://gnomad.broadinstitute.org/) were removed.
Single base substitutions (SBS) –as defined by the
Catalog of Somatic Mutations in Cancer (COSMIC) –
1133Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

were identified using sigProfiler [38] and putative
microsatellite instability (MSI) was determined using
MANTIS [39]. The method HRDetect described by
Davies et al. [40] for the identification of homologous
recombination deficiency was applied using the R
package signature.tools.lib [41].
2.15. RNA-sequencing (RNA-SEQ)
Total RNA was extracted from PCa cells using RNeasy
Mini Kit (Qiagen, Hilden, Germany). RNA were then
sent to Novogene for RNA-SEQ libraries preparation
and sequencing. RNA-SEQ data analysis was per-
formed by Novogene. Briefly, reads were aligned to the
human reference genome GRCh37 using STAR [42] and
differential expression analysis between LNCaP cells
and their castration-resistant sublines was performed
using Rpackage EDGER [43]. Genes with a false discovery
rate <0.005 and an absolute log2-fold-change >1 were
considered significant. Enrichment analysis of Gene
ontology terms and pathway categories was carried out
using Rpackage CLUSTERPROFILER [44].
2.16. Graphs and statistics
Statistical analyses, data fitting and graphics were per-
formed with the GRAPHPAD PRISM 9.0 program (Graph-
Pad Software, Boston, MA, USA). The IDAT files of
the samples were loaded, filtered and normalized with
the package LIMMA (version 3.40.0) in R(version 3.6.0).
By using multiple datasets containing different numbers
of CpG sites, our samples were reduced to 450 k sites.
In addition, a correction was made for possible batch
effects related to chip size using the LIMMA package.
3. Results
3.1. Castration-resistant cells are more
radiosensitive than hormone-sensitive cells due
to impaired DSB repair
Previously, we reported that DSB repair in CRPC cells is
less efficient than in hormone-sensitive cells [31].Since
DSB repair capacity is a determinant factor for cellular
radiosensitivity [45], we sought to analyze radiosensitivity
in CRPC cells. To that end, LNCaP cells and their
castration-resistant sublines (LNCaP-abl, LNCaP-abi,
C4-2B, and C4-2B-ENZA) were treated with different IR
doses and the effect on cell growth was monitored by cell
counting at 3-, 6- and 10-days post-irradiation. A
remarkable irradiation-related decrease in cell growth
was observed in the resistant clones compared with the
parental cells (Fig. 1A). Consistently, IR resulted in a
significant increase in doubling times (DT) in resistant
clones compared with their parental cell lines (Fig. 1B),
indicating an enhanced radiosensitivity in CRPC sub-
lines. In keeping with this idea, using Matrigel-based 3D-
culturing, we identified a significant radiosensitivity in all
CRPC clones compared with their sensitive parental cells
using colony forming assay (Fig. 1C). To analyze DSB
repair efficiency, androgen-sensitive LNCaP and
castration-resistant sublines were exposed to IR with a
dose of 2 Gy, and cH2AX and 53BP1 foci were quanti-
fied after 1 and 24 h post-irradiation (Fig. 1D). Although
we observed no difference in the number of cH2AX/
53BP1 between sensitive and resistant cells at the 1-h time
point post-2 Gy, the exposure to 2 Gy significantly
increased the number of residual cH2AX/53BP1 foci
(threefold) at 24 h in all resistant sublines compared with
sensitive LNCaP cells, pointing at impaired double-
strand break repair capacity (Fig. 1E).
3.2. Impaired HR in castration-resistant cells due
to lower RAD51 expression and loading
To unveil the mechanism underlying the impaired DSB
repair in castration-resistant cells, we compared the
transcription profile of LNCaP cells and their
castration-resistant sublines using RNA-SEQ in biologi-
cal duplicates. We then pooled all resistant clones and
compared the commonly expressed genes with those in
their parental hormone-sensitive LNCaP cells. More
than 4500 genes were found to be significantly differen-
tially expressed in the resistant clones, 2413 genes of
which were downregulated in pooled resistant clones
(Fig. 2A). Interestingly, gene ontology analysis
(Fig. 2B) revealed that the most differentially repressed
molecular pathways were DNA damage-response path-
ways including DNA replication, cell cycle and HRR.
Among the HR repressed genes, RAD51 was signifi-
cantly downregulated in resistant sublines (https://www.
ebi.ac.uk/ena/browser/view/PRJEB55017). Given that
the level of RNA is not necessarily always correlated
with protein levels, we analyzed RAD51 protein levels
in LNCaP cells and their castration-resistant sublines as
well as in 22-RV1, DU145 and PC3 cell lines, which
have been established from xenografts or metastatic
lesions of patients with CRPC. Except for the 22-RV1
cells, twofold lower RAD51 protein levels were detected
in resistant LNCaP sublines as well as in other
castration-resistant cells than in hormone-sensitive
LNCaP cells (Fig. 2C), indicating impaired HR in
castration-resistant cells. To recapitulate this, LNCaP
cells and their resistant sublines were irradiated with 2
Gy and RAD51 colocalized with cH2AX foci were
monitored at 3 and 24 h post-irradiation (Fig. 2D,
1134 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

upper panel). As illustrated in the lower panel of
Fig. 2D, the number of RAD51/cH2AX foci was signif-
icantly lower (twofold) in resistant clones than in
sensitive LNCaP cells (P=0.003). Together, these data
indicate that HRR may be impaired during transition
to a castration-resistant phenotype.
0510
0
50
100
Time, Days
0Gy
1Gy
2Gy
4Gy
5Gy
10Gy
LNCaP
0510
0
10
20
30
Time, Days
Increase in cell number, relative
0Gy
2Gy
5Gy
10Gy
LNCaP-abi
0510
0
10
20
30
40
50
Time, Days
Increase in cell number, relative
0Gy
2Gy
5Gy
10Gy
LNCaP-abl
0510
0
50
100
Time, Days
Increase in cell number, relative
0Gy
1Gy
2Gy
4Gy
5Gy
10Gy
C4-2B
0510
0
10
20
30
40
Time, Days
Increase in cell number, relative
0Gy
2Gy
5Gy
10Gy
C4-2B-ENZA
0Gy 1Gy 2Gy 4Gy 5Gy 10Gy
LNCaP 1,48 1.53 1.66 2.32 5,50 18.20
LNCaP-abl 2.03 ND 2.15 ND 111.4 112.5
LNCaP-abi 2.22 ND 3.49 ND 104.6 112.3
C4-2B 1.46 1.48 1.57 2.10 3.10 20.25
C4-2B-ENZA 1.94 ND 2.18 ND 87.41 98.21
0510
0.00001
0.0001
0.001
0.01
0.1
1
IR Doses, Gy
LNCaP
LNCaP-ARN509
LNCaP-abi
LNCaP-abl
LNCaP-Bic
0
10
20
30
40
Hγ2AX/53BP1 foci/DAPI area
0Gy 2Gy
1h 24h
LNCaP
LNCaP-abl
LNCaP-abi
LNCaP-ARN509
LNCaP-Bic
**
**
**
**
(A)
(B)
(C)
(D) (E)
Relave increase in cell number
Survival Fracon (SF)
Fig. 1. Castration-resistant cells are more radiosensitive than the parental hormone-sensitive cells. (A) Cell number was determined in
LNCaP, LNCaP-abl, LNCaP-abi, C4-2B and C4-2B-ENZA cells on days 0, 3, 6 and 10 post-irradiation with the indicated doses. (B) Cell dou-
bling time in days was calculated for each treatment by fitting exponential growth curves using GRAPHPAD PRISM 9. Shown are means SEM
of three independent experiments. (C) Radiosensitivity of the indicated cells was analyzed by colony formation assay to calculate survival
fractions after different irradiation doses. Shown are means SEM of four independent experiments. (D) Representative micrographs of
cH2AX/53BP1 foci (Scale bar: 100 lm) in the indicated cells 24 h after irradiation with 2 Gy and (E) quantifications of cH2AX/53BP1 foci in
the indicated cells before and 1 and 24 h after irradiation with 2 Gy. Shown are means SEM of three independent experiments. Signifi-
cance was calculated using the Mann–Whitney U-test: **P<0.01 vs. control.
1135Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

-20000
0
20000
40000
60000
80000
Band intensity, AU
HSPC CRPC
RAD51 expression
(P=0.06)
0
5
10
15
20
RAD51, 2Gy/3h
Foci per DAPI area
HSPC CRPC
(P=0.003)
Downregulated pathways in CRPC vs HNPC
(C)
(A) (B)
(D)
DNA replicaon (n=24)
Cell cycle (n=48)
Mismatch repair (n=11)
Purine metabolism (n=37)
Neuroacve ligand-receptor interacon (n=51)
P53 signaling pathway (n=18)
Homologous recombinaon (n=10)
Maturity onset diabetes of the young (n=9)
Oocyte meiosis (n=25)
Progesterone-mediated oocyte maturaon (n=20)
Nucleode excision repair (n=12)
Metabolism of xenobiocs by cytochrome P450 (n=16)
Vascular smooth muscle contracon (n=23)
TGF-beta signaling pathway (n=18)
Calcium signaling pathway (n=32)
Base excision repair (n=9)
Gastric acid secreon (n=16)
Pyrimidine metabolism (n=20)
Small cell lung cancer (n=17)
-Log 10 (padj)
02 4 6 81012
*
*
*
*
*
*
*
*
*
Fig. 2. Castration-resistant cells show a HRR deficiency due to RAD51 downregulation. (A) Volcano plot showing differentially expressed
genes in polled castration-resistant sublines (n=4) vs. hormone-sensitive LNCaP cells (n=2), measured by RNA-SEQ. (B) Gene ontology
(GO) enrichment analysis of the significantly downregulated pathways [log10 (adjusted Pvalue)] in castration-resistant cells. Significance:
*P<0.05. (C) Upper panel: Western blotting showing the expression of RAD51 protein in LNCaP cells (WT), castration-induced LNCaP sub-
lines (abl, abi, ARN509, Bic), in vivo induced castration-resistant C4-2B cells and their enzalutamide-resistant subline (Enza), as well as 3
castration-resistant cell lines (22-RV1, DU145, and PC3). HSC70 was used as a loading control. Lower panel: band intensities were calcu-
lated from three independent blots. (D) Upper panel: Representative immunofluorescence images of cH2AX and RAD51 foci (Scale bar:
100 lm) detected after 3 and 24 h post-irradiation with 2 Gy. DAPI counterstain was used to visualize nuclei. Lower panel: Quantification of
RAD51 foci numbers induced at 3 h post irradiation with 2 Gy. Shown are means SEM of three independent experiments. Significance
was calculated using the Mann–Whitney U-test vs control.
1136 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

3.3. Castration-resistant cells are more sensitive
to olaparib or cisplatin
A direct association between the response to PARPi or
cisplatin and non-functional HRR pathway was reported
[19,32]. We therefore investigated the sensitivity of
castration-resistant cells to PARP inhibition with ola-
parib. To this end, the effect of 1 lMolaparibonthe
proliferation of castration-resistant cells was measured.
Although olaparib did not reduce the cell growth in the
hormone-sensitive LNCaP cells (Fig. 3A), it significantly
decreased the proliferation rate in LNCaP-abl (Fig. 3B),
LNCaP-abi (Fig. 3C) and LNCaP-ARN509 (Fig. 3D), as
exemplified by an increase in DT (1.3-, 2.6- and 2.3-fold,
respectively; Fig. 3E). To further verify this, LNCaP and
resistant sublines were treated with different concentra-
tions of olaparib (0, 0.5, 1, 2, 5 and 10 lM) and effects on
survival were analyzed using 3D Matrigel CFA (Fig. 3F).
To ensure better visualization and counting of living cells,
MTT was used to stain living cells within 3D cultures pre-
harvesting. Again, a significantly enhanced sensitivity to
olaparib was observed in castration-resistant sublines
compared with parental LNCaP cells (Fig. 3G). Similar
results were obtained for cisplatin (Fig. 3G).
Since CRPC is a very heterogeneous disease, we
sought to analyze the genome profile of the CR
LNCaP sublines to elucidate whether they carry the
alterations found in CRPC in vivo. WGS data reported
no big structural differences in CRPC sublines com-
pared with their hormone na€
ıve parental cells
(Fig. S1A). Mutational signatures analysis revealed
that single base substitutions (SBS) –as defined by the
Catalog of Somatic Mutations in Cancer (COSMIC) –
detected in the CR samples included age-related signa-
tures SBS1 and SBS5 but also SBS44, which is associ-
ated with defective DNA mismatch repair (Fig. S1B).
In fact, all CR samples had high MSI scores (LNCaP-
ARN509: 0.69, LNCaP-abi: 0.72, LNCaP-abl: 0.69,
LNCaP-bic: 0.73), indicating genomic instability in the
derived CRPC sublines. No mutation signature for
HRR defect was detected in CR cells, despite showing
a functional HRR deficiency associated with lower
RAD51 expression at the transcriptional level. Possi-
bly, CRPC sublines did not have enough time during
establishment of resistance phenotype to accumulate
genetic aberrations to show the HRR defect signature.
3.4. Ex vivo induction of castration resistance as
an approach to study the sensitivity of CRPC to
olaparib or cisplatin
The above results may imply that HRR is compro-
mised during the development of castration resistance.
To further confirm this hypothesis, we used an
approach to ex vivo inducement of a castration-
resistant phenotype in primary hormone na
€
ıve prostate
tumor specimens (Fig. 4A). Briefly, tumor slices from
three patients with hormone na€
ıve prostate cancer
were cultured ex vivo for up to 6 weeks in androgen-
depleted medium supplemented with charcoal-stripped
serum (CSS) in the presence of abiraterone to induce a
CRPC state. Ki67 IHC staining was used to confirm
the development of the androgen-independency of the
cultured primary tumor samples. Tumor slices showing
no change in Ki67 index after culturing under castration-
resistant (CR) conditions were considered to be CRPC.
As a control, tumor slices from the same three patients
were cultured under hormone-proficient (HP) conditions,
i.e. in normal medium containing FCS plus DHT for the
same period. We assessed this approach with 12 PCa
samples from individual eight PCa patients, but only
three samples from three patients showed no difference in
the Ki67 index upon hormone-depleted culturing condi-
tions (Fig. 4B). Next, we compared the radiosensitization
effect mediated by either olaparib or cisplatin in PCa tis-
sue slices cultured under either HP or CR conditions. To
that end, slice cultures from more than six tumor punch
biopsies collected from 3 PCa patients were irradiated ex
vivo with 3 Gy in the presence or absence of either ola-
parib or cisplatin. cH2AX and 53BP1 foci were then ana-
lyzed 1 and 24 h later (Fig. 4C). No change was
observed in the number of foci at the 1-h time point
between the slices cultured under either condition.
However, tumor slices cultured in CR conditions
demonstrated an increased number of residual cH2AX
and 53BP1 at 24 h post-IR (Fig. S2A,B). Further, the
radiosensitization enhancement ratio mediated by
olaparib (PiER) or cisplatin (CisER) was evaluated using
the mean standard error of all samples set as a threshold
for each DSB marker. Data clearly showed that the
PiER of tumor slices cultured under CR but not HP
rose above the threshold, with minor alterations between
both markers (Fig. 4D). This confirms the findings in
Fig. 3that HRR is compromised in the induced CR
models.
3.5. Olaparib increases the cytotoxicity of
ionizing radiation in castration-resistant but not
hormone na€
ıve prostate cancer tissues
To further validate the above findings and the applica-
bility of the presented preclinical CRPC models, we
sought to recapitulate the data from Fig. 4, using
tumor biopsies collected from CRPC patients. A total
of 13 tumor biopsies from nine PCa patients (CRPC;
n=5 and HSPC; n=4) were collected and irradiated
1137Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

ex vivo with 3 Gy after treatment with either olaparib
or cisplatin. The impact of irradiation on the number
of cH2AX/53BP1 foci was then analyzed 1 and 24 h
later (Fig. 5A) and the PiER and CisER were assessed.
Ultimately, 10 punch biopsies from seven patients ful-
filled all the previously described requirements [21] and
were therefore used for further analysis. Again, no dif-
ference was observed in the number of cH2AX/53BP1
foci at 1 h post-3 Gy between CRPC and HSPC
biopsies; however, tumor biopsies from CRPC patients
showed a distinct increase at 24 h for both cH2AX
and 53BP1 markers upon pretreatment with olaparib
(Fig. S3A,B). Consistently, PiER of all biopsies from
CRPC clearly increased above the threshold, again
with minor alterations between both DSB markers
(Fig. 5B). Similar results were obtained using cisplatin
(Fig. S3A,B, Fig. 5C). Together these findings reveal
the plausibility of the preclinical models of CRPC to
0510
0
2×10
6
4×10
6
6×10
6
8×10
6
Days post treatment
Cell number
CTR
+1μMOlap
LNCaP
0510
0
5×10
5
1×10
6
1.5×10
6
2×10
6
Days post treatment
Cell number
CTR
+1μMOlap
LNCaP-abl
0510
0
1×106
2×106
3×106
4×106
5×106
Days post treatment
Cell number
CTR
+1μMOlap
LNCaP-abi
0510
0
5×105
1×106
1.5×106
2×106
Days post treatment
Cell number
CTR
+1μMOlap
LNCaP-ARN509
CTR Olap
LNCaP 1,92 2.06
LNCaP-abl 3.31 8.45
LNCaP-abi 3.51 9.32
LNCaP-ARN509 3.54 8.31
(E)
(A) (B) (C) (D)
0246810
0.0001
0.001
0.01
0.1
1
Olaparib, μM
LNCaP
LNCaP-ARN509
LNCaP-abi
LNCaP-abl
0510
0.001
0.01
0.1
1
Cisplatin, μM
Survival Fraction,SF
LNCaP
LNCaP-ARN509
LNCaP-abi
LNCaP-abl
(F)
(G)
Survival Fracon (SF)
Fig. 3. Olaparib is more toxic for castration-resistant cells than for the parental hormone-sensitive cells. Cell number was determined in
LNCaP (A), LNCaP-abl (B), LNCaP-abi (C) and LNCaP-ARN509 (D) cells on days 0, 3, 6 and 10 post treatment with 1 lM of the PARP inhibi-
tor olaparib. (E) Cell doubling time in days was calculated for each treatment by fitting exponential growth curves using GRAPHPAD PRISM 9.
Shown are means SEM of four independent experiments. (F) Representative images of 3D-cultures of the LNCaP and LNCaP-abi cells
treated with the indicated concentrations of olaparib before (upper panel) and after (lower panel) harvesting and staining with MTT. (G) Sur-
vival fractions measured by colony forming assay after treating the indicated cells with different concentrations of olaparib (left panel) or cis-
platin (right panel). Shown are means SEM of three independent experiments.
1138 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

detect DSB repair defect, sensitivity to cisplatin and
radiosensitization effect of olaparib.
3.6. Establishment and characterization of
patient-derived organoids cultures from PCa
patients
Twelve long-term PDO cultures from four na
€
ıve and
four CRPC patients were established using a modified
protocol from Gao et al. [30]. We obtained a success
rate of more than 60%. All established PDOs were suc-
cessfully expanded and maintained under the same
culturing conditions for >12 passages with no obvious
morphological changes (data not shown) and were fro-
zen down to create an organoid biobank. Clinical infor-
mation and pathological parameters showed similarities
between the established patient-derived organoids and
their donors (Table S2). The histological features of
each of the established PDOs were of similar appear-
ance to their matched primary tumors (Fig. 6A), with a
strong AMCAR immunohistochemistry staining
(Fig. 6A). We also evaluated CpG-rich methylation in
PDOs on a genome-wide scale using Illumina Human-
Methylation450 Bead Chip (450 K) arrays [46].The
01234
0
1
2
3
4
PiER
γH2AX
53BP1
D93-CR
D93-HP
D94-CR
D94-HP
A82a-HP
A82a-CR
(C)
(A)
(D)
(B)
01234
0
1
2
3
4
CisER
γH2AX
53BP1
D93-CR
D93-HP
D94-CR
D94-HP A82a-HP
A82a-CR
HPCR
D94D93
Fig. 4. Ex vivo induction of castration resistance in prostate cancer increases cytotoxicity to olaparib and cisplatin via impairing double
strand break repair efficiency. (A) Schematic representation of castration resistance induction in ex vivo cultures of prostate cancer (PCa)
specimens which had shown no signs of castration resistance in vivo. Briefly, PCa slices were cultured for up to 6 weeks in either
androgen-stripped serum containing medium in the presence of abiraterone (10 lM) or in full serum containing medium. Ki67 expression
was monitored by immunohistochemical staining. Only tumor slices showing a similar Ki67 index under castration resistance culturing condi-
tions were included in the next experiments. (B) Representative images of Ki67 staining (Scale bar: 100 lm) in two tumor slices from two
PCa patients (D93 and D94) cultured under hormone-proficient (HP) or castration resistance-inducing (CR) conditions. High magnification
images (magnification, 809) of ROI are shown. (C) Representative micrographs of cH2AX (red) and 53BP1 (green) foci 24 h after treatment
with olaparib plus 3 Gy or cisplatin alone in PCa slices from patient #D93. Scale bar: 100 lm. (D) Plots showing the correlation between
PARP inhibitor (PARPi) enhancement ratio (PiER, left panel) or cisplatin enhancement ratio (CisER, right panel) of residual cH2AX (X-axis)
and 53BP1 (Y-axis) at 24 h post treatment; three independent experiments for each tumor slice. Black and red dots represent tumor slices
cultured under hormone-proficient and castration resistance conditions, respectively.
1139Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

PDOs models were found to cluster with PCa but not
with normal prostate or other tumor entities based on
DNA methylation profiling (Fig. 6B). Interestingly, the
methylation profile of the PDOs did not cluster with
that of the established cell lines, indicating that the
PDOs more appropriately represent the patients’ in vivo
tumor. Altogether, this confirms that the established
PDOs represent the PCa in vivo and can therefore serve
as preclinical models for PCa research.
3.7. Olaparib and cisplatin are more toxic for
organoids established from CRPC than hormone
na€
ıve patients
The advantage of using PDOs as preclinical models is
that they not only allow DSB repair monitoring but
also enable the analysis of the effect of olaparib or cis-
platin on clonogenic survival by colony formation
assay. Briefly, PDOs were treated with different
01234
0
1
2
3
4CisER
γH2AX
53BP1
A81
C80
D99
B80
C82a
C82b
C84a
C84b
D109a
D109b
012
0.8
1.0
1.2
1.4
1.6 PiER
γH2AX
53BP1
A81
C80
D99
B80
C82a
C82b
C84a
C84b
D109a
D109b
(B)(A) (C)
3Gy /24h 3Gy+Olap /24h
A79B80
Fig. 5. Olaparib or cisplatin increases cytotoxic effects of IR in castration-resistant PCa. (A) Representative micrographs of cH2AX (red) and
53BP1 (green) foci 24 h after treatment with olaparib plus 3 Gy in tumor biopsies from a castration-resistant PCa patient #B80. Scale bar:
100 lm. (B) PiER of residual cH2AX (X-axis) and 53BP1 (Y-axis) foci at 24 h post 3 Gy. (C) Cisplatin-enhancement ratio (CisER) of residual
cH2AX (X-axis) and 53BP1 (Y-axis) at 24 h. Shown are mean SEM of three independent experiments. Black and red dots represent data
from hormone na
€
ıve and castration-resistant PCa, respectively.
(A)
(B)
Material
Type
Entity
Healthy Prostate (n=20)
Prostate Ca (n=30)
Mamma Ca (n=20)
Glioblastoma(n=20)
Bronchial Ca (n=20)
Cell line (n=10)
Organoid(n=5)
Tissue (n=95)
FFPE (n=16)
Frozen(n=94)
Hamburg (n=38)
TCGA (n=40)
GEO (n=32)
Origin
0.8
0.6
0.4
0.2
Methylation
Entity
Type
Material
Origin
Tumor Organoid AMACR
A81
B80
Fig. 6. Establishment and characterization of patient-derived organoid cultures from prostate cancer (PCa) patients. (A) Representative
images of the established organoids and their corresponding primary tumors with H&E and immunohistochemical staining for AMACR. Scale
bar: 50 lm. (B) Genome-wide DNA methylation cluster analysis of the established organoids showing clustering with PCa but not with nor-
mal prostate or other tumor entities using a cohort of 110 samples.
1140 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

concentrations of either olaparib (0, 0.5, 1, 2, 5 and
10 lM) or cisplatin (0, 2, 5 10 and 20 lM) and CFA
was used to quantify the survival fractions in 3D set-
tings. To ensure the better visualization and counting
of living cells, MTT was used to stain living cells
within individual organoids pre-harvesting (Fig. 7A).
Compared with hormone-sensitive ones, CRPC orga-
noids were found to be clearly more sensitive to ola-
parib (Fig. 7B) and cisplatin (Fig. 7C), which indeed
rationalizes the use of either olaparib or cisplatin as
effective drugs in CRPC patients.
3.8. Enhanced pro-metastatic signaling and
migration in castration-resistant cells
To date, metastatic CRPC remains incurable and the
prognosis for these patients is poor. Therefore, it is
important to have preclinical models to facilitate the
identification of other treatment options for this disease
setting. RNA-SEQ analysis revealed more than 2000
upregulated genes in pooled castration-resistant LNCaP
sublines compared with the parental LNCaP cells
(Fig. 2A). Interestingly, when gene ontology analysis
was performed (Fig. 8A), we found that despite the fact
that LNCaP cells were originally established from meta-
static PCa, the most differentially upregulated molecu-
lar pathways in the resistant sublines were those related
to metastasic progress, dissemination, including ECM
receptor interaction, focal and cell adhesion molecules.
This indicates that the metastatic potential might be
further stimulated in the CR sublines. To investigate
this issue, we monitored the ability of castration-
resistant clones to scatter outside the clones using a cell
scattering assay. As illustrated in Fig. 8B, compared
with the parental LNCaP cells, LNCaP-ARN509,
LNCaP-abi and LNCaP-abl cells displayed the typical
scattering phenotype characterized by the loss of cell-to-
cell contacts and drastic cellular elongation in both 2D
and 3D culture settings. In contrast, DU145 cells, which
are known to have a lower metastatic potential, showed
no signs of cell scattering. Analysis of cell migration –
an integral part of the metastatic cascade –using a
chamber assay confirmed the enhanced invasive proper-
ties in the castration-resistant LNCaP-ARN509 cells
compared with their hormone-sensitive parental LNCaP
cells as evidenced by a significantly higher number (two-
fold) of migrating cells in LNCaP-ARN509 (346 50.7
vs. 724.3 188.4 migrated cells per field, P=0.01).
Interestingly, pretreatment with 1 lM olaparib signifi-
cantly decreased the migrating cells both in LNCaP and
LNCaP-ARN509 castration-resistant cells (P=0.001)
(Fig. 8C,D). In contrast, pretreatment with 2 lM cis-
platin failed to reduce migration in LNCaP-ARN509
cells. Notably, no change was seen in the proliferation
or growth rate upon treatment with either olaparib or
cisplatin for the entire 36 h of this experiment in either
LNCaP or LNCaP-ARN509 cells (Fig. S4). Together,
these data reflect the ability of olaparib but not cis-
platin to inhibit the metastatic behavior of CRPC cells.
4. Discussion
Compared with other tumor entities, translational
research in PCa has lagged behind due to a lack of
0510
0.001
0.01
0.1
1
Olaparib, μM
Survival Fraction, SF
A79
B80
D109a
D109b
D93
A81
D111a
D111b
A89a
A89b
D110a
D110b
0 5 10 15 20
0.01
0.1
1
Cisplatin, μM
Survival Fraction, SF
A79
B80
D109a
D109b
D93
A81
D111a
D111b
A89a
A89b
D110a
D110b
(B)(A) (C)
Fig. 7. Olaparib is more toxic for castration-resistant organoids than for the hormone-sensitive PCa organoids. (A) Representative images of
organoid cultures of a hormone na
€
ıve (A79) and castration-resistant (B80) PCa organoid treated with the indicated concentrations of olaparib
before (upper panel) and after (lower panel) harvesting and staining with MTT. (B,C) Survival fractions measured by colony forming assay
after treatment of organoids established from hormone na
€
ıve (black) or castration-resistant (red) PCa patients with different concentrations
of olaparib (B) or cisplatin (C). Shown are means SEM of four independent experiments (except for hormone n€
aive patients, where n=3).
1141Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

appropriate preclinical models. Few preclinical models
accurately reflect the clinical and molecular variability
seen in PCa patients, impeding the rational develop-
ment of molecularly derived tailored treatment
options. The techniques and models described in the
current study are essential tools not only for bolstering
the understanding of the drivers behind oncogenesis
and how this affects the clinical course, but also to
provide a rationale for alternative therapeutic targets
for individual PCa patients.
Among the commonly used human PCa cancer
models are cancer cell lines that are established from
cancer patients. However, these do not represent PCa
in vivo. Here we employed our previously reported
CRPC-induced sublines from the hormone-sensitive
PCa cell line LNCaP [31] and revealed compromised
DSB repair efficiency and increased radiosensitivity in
the hormone-resistant clones than in the parental cells.
WGS analysis did not show any evidence of genetic
mutations in DSB genes. RNA-SEQ demonstrated a
different transcriptome in CRPC sublines with
decreased expression of several HRR-related genes
such as RAD51. A lower RAD51 protein expression
was also reported in the CRPC cells, resulting in HRR
deficiency. In line with this finding, RAD51 foci
recruitment at IR-induced DSB sites was decreased
twofold in CR sublines. These data pinpoint the HRR
deregulation in CRPC and rationalize the use of
PARPi for this subtype of the disease. In fact, the effi-
cacy of different PARPi has been or is being tested for
Upregulated pathways in CRPC vs HNPC
0
500
1000
Migrated cells/field
DMSO
Olap
Cis
DMSO
Olap
Cis
**
LNCaP LNCaP-ARN509
*
ns
**
ns
(B)
(A)
(D)
(C)
LNCaP
LNCaP-ARN509
UT 1 μM Olap 2 μM Cis
Metabolism of xenobiocs by cytochrome P450 (n=22)
ECM-receptor interacon (n=25)
Renol metabolism (n=19)
Drug metabolism - cytochrome P450 (n=20)
Steroid hormone biosynthesis (n=16)
Linoleic acid metabolism (n=11)
Protein digeson and absorpon (n=19)
Amoebiasis (n=22)
Bile secreon (n=17)
Focal adhesion (n=34)
Cell adhesion molecules (CAMs) (n=25)
Pathways in cancer(n=48)
Systemic lupus erythematosus (n=19)
Toxoplasmosis (n=23)
Ascorbate and aldarate metabolism (n=8)
ABC transporters (n=11)
Arachidonic acid metabolism (n=13)
Drug metabolism – other enzymes (n=12)
Type II diabetes mellitus (n=11)
Aldosterone-regulated sodium reabsorpon (n=10)
- Log 10 (padj)
012345
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
Fig. 8. Pro-metastatic pathways and migration are enhanced in castration-resistant cells. (A) Gene ontology enrichment analysis of the signif-
icantly upregulated pathways [log10 (adjusted P-value)] in castration-resistant cells showing upregulation in pro-metastatic pathways. Signif-
icance: *P<0.05. (B) Cell colony scattering assays were performed with the indicated cells by seeding cells at a low density and allowing
them to form colonies in 2D or 3D cultures. Light microscopy images of the colonies were taken at random for each cell line. (C) Transwell
migration assay: representative microscopic images of the indicated cells that migrated through the Transwell in the migration assay after
treatment with 1 lM olaparib or 2 lM cisplatin for 36 h. UT, untreated control. Crystal violet was used to visualize cells. Scale bar: 200 lm.
(D) Quantification of the experiments performed in (C). Shown are means SEM of four independent experiments (n=3 for cis). Signifi-
cance was calculated using the Mann–Whitney U-test: *P<0.05, **P<0.01 vs. control. ns, not significant.
1142 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

the treatment of mCRPC patients in several clinical
trials. Most of these studies revealed that PARPi sig-
nificantly improve tumor response in terms of disease
control and overall survival for mCRPC patients with
HRR mutations. Recently, nine clinical trials using
PARPi in mCRPC have been analyzed in a meta-
analysis to test the benefit of PARPi in mCRPC
patients [47]. The trials demonstrated that the magni-
tude of benefits from PARPi varies greatly between
different HR-defect subgroups, showing the most vig-
orous efficacy for PARPi in BRCA-mutation carriers
compared with patients who harbored no BRCA
mutations. Furthermore, this analysis reported that
BRCA2 mutations are likely the most effective muta-
tions that predict the response to PARPi in PCa.
Interestingly, a significant benefit in BRCA wild-type
tumors was observed, supporting the view that besides
BRCA mutations, other non-BRCA HRR-related gene
aberrations may also be used to predict the antitumor
activity of PARPi. Hence, only using BRCA muta-
tional status as a marker for PARPi sensitivity is inad-
equate, and it may miss a potentially larger proportion
of responding patients. Following large-scale cancer
sequence analysis, mutations in other HR-related genes
such as CDK12, ATM and PALB2 were commonly
found in mCRPC [11,48,49], and these non-BRCA
DNA repair genes could be used as alternative
biomarkers to predict the sensitivity of PARPi.
TOPARP-A and B clinical studies by Mateo et al.
[50,51] provided evidence that mCRPC patients with
other mutations in genes related to the HRR machin-
ery also appear to benefit from PARP inhibitors.
A deep sequencing of all PCa patients would facili-
tate identifying the common genetic alterations with
HRR to predict the benefit from PARPi. Despite the
latest advances in the field of large-scale sequencing
analysis, it is still difficult to apply it in regular routine
clinical work, as it is very costly and more importantly
requires a previous knowledge about the role of each
gene in HRR. In fact, there is only limited knowledge
of the functional consequences of these mutations on
HRR. Moreover, it is ultimately unclear why alter-
ations in the same gene lead to a therapy response in
one patient but not in another.
Here we present, in addition to the aforementioned
in vitro cell lines, ex vivo preclinical models (ex vivo
tumor slice and organoid cultures) that may help in
detecting functional HRR defects to predict the
response to PARPi. Furthermore, we present the ratio-
nale for the use of platinum-based therapy such as cis-
platin, which so far has not been routinely used in the
treatment of CRPC but which has been reported to
have some activity, especially in patients harboring
HRR defects. This is in line with the previously pub-
lished multicenter retrospective analysis showing anti-
tumor activity for treatment with platinum-based
therapies in the cohort of CRPC patients with tumors
harboring DNA repair gene aberrations [52]. Impor-
tantly, we believe that this is the first work that shows
that olaparib but not cisplatin is able to impair the
metastatic potential of PCa. Previously, we reported a
functional ex vivo assay that enables the analysis of
DSB formation and repair directly in tumor slice cul-
tures from individual PCa patients [21]. Compared
with tumor slices from HSPC patients, we clearly dem-
onstrated here, in the CRPC slice cultures, an increase
in the number of residual and thus unrepaired IR-
induced cH2AX and 53BP1 foci upon pretreatment
with olaparib or cisplatin as evidenced by increased
PiER and CisER indices, respectively. This was further
confirmed using a modified approach to induce a CR
phenotype through growing hormone-sensitive tumor
sample slices for several weeks in androgen-depleted
medium supplemented with CSS and abiraterone. Of
note, there are some concerns about the applicability
of the CR-induction ex vivo approach in the clinical
settings because of the relatively limited success rate in
inducing CR (<40% in our study) and the uncertainty
as to whether the CR induction process represents the
in vivo situation. Despite these concerns, this model
confirmed the benefits of olaparib in patients pre-
treated with anti-hormone therapy, especially abirater-
one, which is in line with the results from the PROpel
phase III study showing prolonged progression-free
survival for olaparib +abiraterone compared with
abiraterone alone, irrespective of HRR status [53].As
a patient-derived preclinical model, we further present
here a very robust protocol for establishing organoids
from PCa patients. Attempts to establish PCa orga-
noids have been performed by other laboratories, but
with lower overall success rates for longer/indefinite
propagation and expansion. We could increase the suc-
cess rate for the establishment of PCa organoids to
60–70% irrespective of hormone sensitivity state
(HSPC and CRPC). This increased rate could be
attributed to several factors such as ROCK inhibitor
and epithelial growth factors, which enable the cells to
adapt very quickly/more efficiently from tissue to cul-
ture conditions without inducing senescence as previ-
ously described in many studies [25,26,30]. Another
explanation for our high success rate might be the effi-
cient logistics, which enables the timely transport of
freshly collected samples immediately from the operat-
ing theater to the lab, avoiding delays which may
affect the efficiency of the lab-based organoid forma-
tion. Tumor cell content has been confirmed in our
1143Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

established PCa PDO cultures by histopathological
analysis. Also, PDOs have consistent IHC-positive
staining of the well-established tumor marker
AMACR and show histological similarities to their
original primary tumors. In addition, evaluation of
CpG-rich global methylation revealed clustering of the
organoids with other in vivo PCa datasets but not with
other tumor entities or with normal prostate tissues.
Given that PDOs enable the measurement of the direct
effect on survival fractions, we demonstrated the
increased sensitivity of CRPC organoids to olaparib or
cisplatin compared with the organoids established
from HSPC patients. Together, these data confirm a
HRR-deficient state of CRPC patients irrespective of
distinct genomic alternations in known key players of
HRR. It is important to note, however, that there is
still space for future improvement for the PDO system
presented here. For example, it is important to estab-
lish conditions that allow the establishment of micro-
environmental elements such as blood vessels, immune
cells and other stroma cells.
5. Conclusion
In conclusion, we present reliable preclinical models
that allow for rapid functional testing and comparison
of multiple individual drugs prior to in vivo analysis for
example testing the presence of HRR deficiency in
CRPC and response prediction to olaparib or cisplatin.
This individual assessment of HRR functional capacity
will enable us to improve future patient selection for
personalized treatment approaches and thus increase
the likelihood of response to PARP inhibitor therapy.
Acknowledgements
This work was supported by BMBF grants
02NUK032 & 02NUK035B, the German Academic
Exchange Service (DAAD) and Milderd-Scheel Cancer
Career Center, HaTRiCs4 program.
Conflict of interest
The authors declare no conflict of interest.
Author contributions
ME and WM designed the study concept and the
experiments and wrote the paper. GvA, SJO-H, CO,
TM, AE, US, CP and KR collected and analyzed the
clinical data. AE and US examined the methylome
and performed the corresponding analysis; CM and
MA performed the bioinformatics for WGS and
RNA-SEQ. SB-R performed the pathological analyses
of tumor and organoid samples SJO-H provided drug-
resistant cell lines. ME and WM did artwork. All
authors reviewed and edited the paper and have
approved its final version.
Peer review
The peer review history for this article is available at
https://publons.com/publon/10.1002/1878-0261.13382.
Data accessibility
Sequence data have been deposited in the European
Nucleotide Archive (ENA) at EMBL-EBI under acces-
sion number PRJEB55017 (https://www.ebi.ac.uk/ena/
browser/view/PRJEB55017). Generated raw data sup-
porting the findings of this study are available from
the corresponding author (WM) on request.
References
1 Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA,
Jemal A. Global cancer statistics 2018: GLOBOCAN
estimates of incidence and mortality worldwide for 36
cancers in 185 countries. CA Cancer J Clin. 2018;68:394–
424. https://doi.org/10.3322/caac.21492
2 Sayegh N, Swami U, Agarwal N. Recent advances in
the management of metastatic prostate cancer. JCO
Oncol Pract. 2022;18:45–55. https://doi.org/10.1200/op.
21.00206
3 Finianos A, Gupta K, Clark B, Simmens SJ, Aragon-
Ching JB. Characterization of differences between
prostate cancer patients presenting with de novo versus
primary progressive metastatic disease. Clin Genitourin
Cancer. 2017;16:85–9. https://doi.org/10.1016/j.clgc.
2017.08.006
4 Grasso CS, Wu YM, Robinson DR, Cao X,
Dhanasekaran SM, Khan AP, et al. The mutational
landscape of lethal castration-resistant prostate cancer.
Nature. 2012;487:239–43. https://doi.org/10.1038/
nature11125
5 Harris WP, Mostaghel EA, Nelson PS, Montgomery B.
Androgen deprivation therapy: progress in
understanding mechanisms of resistance and optimizing
androgen depletion. Nat Clin Pract Urol. 2009;6:76–85.
https://doi.org/10.1038/ncpuro1296
6 Marques RB, Dits NF, Erkens-Schulze S, van Weerden
WM, Jenster G. Bypass mechanisms of the androgen
receptor pathway in therapy-resistant prostate cancer
cell models. PLoS One. 2010;5:e13500. https://doi.org/
10.1371/journal.pone.0013500
7 Sun S, Sprenger CC, Vessella RL, Haugk K, Soriano
K, Mostaghel EA, et al. Castration resistance in human
1144 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

prostate cancer is conferred by a frequently occurring
androgen receptor splice variant. J Clin Invest.
2010;120:2715–30. https://doi.org/10.1172/jci41824
8 Ruch JM, Hussain MH. Evolving therapeutic
paradigms for advanced prostate cancer. Oncolo Gy
(Williston Park). 2011;25(496–504):508.
9 Beltran H, Yelensky R, Frampton GM, Park K,
Downing SR, MacDonald TY, et al. Targeted next-
generation sequencing of advanced prostate cancer
identifies potential therapeutic targets and disease
heterogeneity. Eur Urol. 2013;63:920–6. https://doi.org/
10.1016/j.eururo.2012.08.053
10 Leongamornlert D, Mahmud N, Tymrakiewicz M,
Saunders E, Dadaev T, Castro E, et al. Germline
BRCA1 mutations increase prostate cancer risk. Br J
Cancer. 2012;106:1697–701. https://doi.org/10.1038/bjc.
2012.146
11 Robinson D, Van Allen EM, Wu YM, Schultz N,
Lonigro RJ, Mosquera JM, et al. Integrative clinical
genomics of advanced prostate cancer. Cell.
2015;161:1215–28. https://doi.org/10.1016/j.cell.2015.05.
001
12 De Bono JS, Matsubara N, Penel N, Mehra N,
Kolinsky MP, Bompas E, et al. Exploratory gene-by-
gene analysis of olaparib in patients (pts) with
metastatic castration-resistant prostate cancer
(mCRPC): PROfound. J Clin Oncol. 2021;39:126–6.
https://doi.org/10.1200/JCO.2021.39.6_suppl.126
13 Hussain M, Mateo J, Fizazi K, Saad F, Shore N,
Sandhu S, et al. Survival with olaparib in metastatic
castration-resistant prostate cancer. N Engl J Med.
2020;383:2345–57. https://doi.org/10.1056/
NEJMoa2022485
14 Chi KN, Rathkopf DE, Smith MR, Efstathiou E,
Attard G, Olmos D, et al. Phase 3 MAGNITUDE
study: first results of niraparib (NIRA) with abiraterone
acetate and prednisone (AAP) as first-line therapy in
patients (pts) with metastatic castration-resistant
prostate cancer (mCRPC) with and without
homologous recombination repair (HRR) gene
alterations. J Clin Oncol. 2022;40:12–2. https://doi.org/
10.1200/JCO.2022.40.6_suppl.012
15 Clarke Noel W, Armstrong Andrew J, Thiery-Vuillemin
A, Oya M, Shore N, Loredo E, et al. Abiraterone and
olaparib for metastatic castration-resistant prostate
cancer. NEJM Evid. 2022;1:EVIDoa2200043. https://
doi.org/10.1056/EVIDoa2200043
16 Sobel RE, Sadar MD. Cell lines used in prostate cancer
research: a compendium of old and new lines--part 1. J
Urol. 2005;173:342–59. https://doi.org/10.1097/01.ju.
0000141580.30910.57
17 Sobel RE, Sadar MD. Cell lines used in prostate cancer
research: a compendium of old and new lines--part 2. J
Urol. 2005;173:360–72. https://doi.org/10.1097/01.ju.
0000149989.01263.dc
18 Arriaga JM, Abate-Shen C. Genetically engineered
mouse models of prostate cancer in the postgenomic
era. Cold Spring Harb Perspect Med. 2019;9:a030528.
https://doi.org/10.1101/cshperspect.a030528
19 Zhang M, Lin YH, Sun YJ, Zhu S, Zheng J, Liu K, et
al. Pharmacological reprogramming of fibroblasts into
neural stem cells by signaling-directed transcriptional
activation. Cell Stem Cell. 2016;18:653–67. https://doi.
org/10.1016/j.stem.2016.03.020
20 De-Colle C, Yaromina A, Hennenlotter J, Thames H,
Mueller AC, Neumann T, et al. Ex vivo gammaH2AX
radiation sensitivity assay in prostate cancer: inter-
patient and intra-patient heterogeneity. Radiother
Oncol. 2017;124:386–94. https://doi.org/10.1016/j.
radonc.2017.08.020
21 Kocher S, Beyer B, Lange T, Nordquist L,
Volquardsen J, Burdak-Rothkamm S, et al. A
functional ex vivo assay to detect PARP1-EJ repair
and radiosensitization by PARP-inhibitor in prostate
cancer. Int J Cancer. 2019;144:1685–96. https://doi.
org/10.1002/ijc.32018
22 Menegakis A, De Colle C, Yaromina A, Hennenlotter
J, Stenzl A, Scharpf M, et al. Residual gammaH2AX
foci after ex vivo irradiation of patient samples with
known tumour-type specific differences in radio-
responsiveness. Radiother Oncol. 2015;116:480–5.
https://doi.org/10.1016/j.radonc.2015.08.006
23 Menegakis A, von Neubeck C, Yaromina A, Thames
H, Hering S, Hennenlotter J, et al. gammaH2AX assay
in ex vivo irradiated tumour specimens: a novel method
to determine tumour radiation sensitivity in patient-
derived material. Radiother Oncol. 2015;116:473–9.
https://doi.org/10.1016/j.radonc.2015.03.026
24 Drost J, Clevers H. Organoids in cancer research. Nat
Rev Cancer. 2018;18:407–18. https://doi.org/10.1038/
s41568-018-0007-6
25 Drost J, Karthaus WR, Gao D, Driehuis E, Sawyers
CL, Chen Y, et al. Organoid culture systems for
prostate epithelial and cancer tissue. Nat Protoc.
2016;11:347–58. https://doi.org/10.1038/nprot.2016.006
26 Elbadawy M, Abugomaa A, Yamawaki H, Usui T,
Sasaki K. Development of prostate cancer organoid
culture models in basic medicine and translational
research. Cancers (Basel). 2020;12:777. https://doi.org/
10.3390/cancers12040777
27 Gleave AM, Ci X, Lin D, Wang Y. A synopsis of
prostate organoid methodologies, applications, and
limitations. Prostate. 2020;80:518–26. https://doi.org/10.
1002/pros.23966
28 Puca L, Bareja R, Prandi D, Shaw R, Benelli M,
Karthaus WR, et al. Patient derived organoids to
model rare prostate cancer phenotypes. Nat Commun.
2018;9:2404. https://doi.org/10.1038/s41467-018-04495-z
29 Richards Z, McCray T, Marsili J, Zenner ML,
Manlucu JT, Garcia J, et al. Prostate stroma increases
1145Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models

the viability and maintains the branching phenotype of
human prostate organoids. iScience. 2019;12:304–17.
https://doi.org/10.1016/j.isci.2019.01.028
30 Gao D, Vela I, Sboner A, Iaquinta PJ, Karthaus WR,
Gopalan A, et al. Organoid cultures derived from
patients with advanced prostate cancer. Cell.
2014;159:176–87. https://doi.org/10.1016/j.cell.2014.08.
016
31 Elsesy ME, Oh-Hohenhorst SJ, L€
oser A, Oing C,
Mutiara S, K€
ocher S, et al. Second-generation
antiandrogen therapy radiosensitizes prostate cancer
regardless of castration state through inhibition of
DNA double Strand break repair. Cancers (Basel).
2020;12:2467. https://doi.org/10.3390/cancers12092467
32 Culig Z, Hoffmann J, Erdel M, Eder IE, Hobisch A,
Hittmair A, et al. Switch from antagonist to agonist of
the androgen receptor bicalutamide is associated with
prostate tumour progression in a new model system. Br
J Cancer. 1999;81:242–51. https://doi.org/10.1038/sj.bjc.
6690684
33 Bakr A, Oing C, Kocher S, Borgmann K, Dornreiter I,
Petersen C, et al. Involvement of ATM in homologous
recombination after end resection and RAD51
nucleofilament formation. Nucleic Acids Res.
2015;43:3154–66. https://doi.org/10.1093/nar/gkv160
34 Kasten U, Borgmann K, Burgmann P, Li G, Dikomey
E. Overexpression of human Ku70/Ku80 in rat cells
resulting in reduced DSB repair capacity with
appropriate increase in cell radiosensitivity but with no
effect on cell recovery. Radiat Res. 1999;151:532–9.
35 Li H, Durbin R. Fast and accurate long-read alignment
with burrows-wheeler transform. Bioinformatics.
2010;26:589–95. https://doi.org/10.1093/bioinformatics/
btp698
36 Chen X, Schulz-Trieglaff O, Shaw R, Barnes B,
Schlesinger F, K€
allberg M, et al. Manta: rapid
detection of structural variants and indels for germline
and cancer sequencing applications. Bioinformatics.
2016;32:1220–2. https://doi.org/10.1093/bioinformatics/
btv710
37 Kim S, Scheffler K, Halpern AL, Bekritsky MA, Noh
E, K€
allberg M, et al. Strelka2: fast and accurate calling
of germline and somatic variants. Nat Methods.
2018;15:591–4. https://doi.org/10.1038/s41592-018-0051-
x
38 Bergstrom EN, Huang MN, Mahto U, Barnes M,
Stratton MR, Rozen SG, et al.
SigProfilerMatrixGenerator: a tool for visualizing and
exploring patterns of small mutational events. BMC
Genomics. 2019;20:685. https://doi.org/10.1186/s12864-
019-6041-2
39 Kautto EA, Bonneville R, Miya J, Yu L, Krook MA,
Reeser JW, et al. Performance evaluation for rapid
detection of pan-cancer microsatellite instability with
MANTIS. Oncotarget. 2017;8:7452–63. https://doi.org/
10.18632/oncotarget.13918
40 Davies H, Glodzik D, Morganella S, Yates LR, Staaf
J, Zou X, et al. HRDetect is a predictor of BRCA1
and BRCA2 deficiency based on mutational
signatures. Nat Med. 2017;23:517–25. https://doi.org/
10.1038/nm.4292
41 Degasperi A, Amarante TD, Czarnecki J, Shooter S,
Zou X, Glodzik D, et al. A practical framework and
online tool for mutational signature analyses show
inter-tissue variation and driver dependencies. Nat
Cancer. 2020;1:249–63. https://doi.org/10.1038/s43018-
020-0027-5
42 Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski
C, Jha S, et al. STAR: ultrafast universal RNA-seq
aligner. Bioinformatics. 2013;29:15–21. https://doi.org/
10.1093/bioinformatics/bts635
43 Robinson MD, McCarthy DJ, Smyth GK. edgeR: a
Bioconductor package for differential expression
analysis of digital gene expression data. Bioinformatics.
2010;26:139–40. https://doi.org/10.1093/bioinformatics/
btp616
44 Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R
package for comparing biological themes among gene
clusters. Omics. 2012;16:284–7. https://doi.org/10.1089/
omi.2011.0118
45 Dikomey E, Borgmann K, Brammer I, Kasten-Pisula
U. Molecular mechanisms of individual radiosensitivity
studied in normal diploid human fibroblasts. Toxicolo
Gy. 2003;193:125–35.
46 Capper D, Jones DTW, Sill M, Hovestadt V, Schrimpf
D, Sturm D, et al. DNA methylation-based
classification of central nervous system tumours.
Nature. 2018;555:469–74. https://doi.org/10.1038/
nature26000
47 Wu K, Liang J, Shao Y, Xiong S, Feng S, Li X.
Evaluation of the efficacy of PARP inhibitors in
metastatic castration-resistant prostate cancer: a
systematic review and meta-analysis. Front Pharmacol.
2021;12:777663. https://doi.org/10.3389/fphar.2021.
777663
48 The molecular taxonomy of primary prostate cancer.
Cell. 2015;163:1011–25. https://doi.org/10.1016/j.cell.
2015.10.025
49 Pritchard CC, Mateo J, Walsh MF, De Sarkar N,
Abida W, Beltran H, et al. Inherited DNA-repair gene
mutations in men with metastatic prostate cancer. N
Engl J Med. 2016;375:443–53. https://doi.org/10.1056/
NEJMoa1603144
50 Mateo J, Carreira S, Sandhu S, Miranda S, Mossop H,
Perez-Lopez R, et al. DNA-repair defects and olaparib
in metastatic prostate cancer. N Engl J Med.
2015;373:1697–708. https://doi.org/10.1056/
NEJMoa1506859
1146 Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
Functional detection of HRR defects in PCa models M. E. Elsesy et al.

51 Mateo J, Porta N, Bianchini D, McGovern U, Elliott
T, Jones R, et al. olaparib in patients with metastatic
castration-resistant prostate cancer with DNA repair
gene aberrations (TOPARP-B): a multicentre, open-
label, randomised, phase 2 trial. Lancet Oncol.
2020;21:162–74. https://doi.org/10.1016/s1470-2045(19)
30684-9
52 Schmid S, Omlin A, Higano C, Sweeney C, Martinez
Chanza N, Mehra N, et al. Activity of platinum-based
chemotherapy in patients with advanced prostate cancer
with and without DNA repair gene aberrations. JAMA
Netw Open. 2020;3:e2021692. https://doi.org/10.1001/
jamanetworkopen.2020.21692
53 Saad F, Armstrong AJ, Thiery-Vuillemin A,
Oya M, Loredo E, Procopio G, et al. PROpel:
phase III trial of olaparib (Ola) and abiraterone (abi)
versus placebo (pbo) and abi as first-line (1L)
therapy for patients (pts) with metastatic castration-
resistant prostate cancer (mCRPC). J Clin Oncol.
2022;40:11–1. https://doi.org/10.1200/JCO.2022.40.6_
suppl.011
Supporting information
Additional supporting information may be found
online in the Supporting Information section at the end
of the article.
Fig. S1. Whole genome profile differences between cas-
tration-resistant cells and their parental sensitive coun-
terpart LNCaP.
Fig. S2. Effect of olaparib or cisplatin on DSB-repair
in CR-induced ex vivo PCa cultures.
Fig. S3. Effect of olaparib or cisplatin on DSB-repair
in freshly collected tumor tissues from hormone na
€
ıve
or castration-resistant PCa patients.
Fig. S4. Effect of 1 lM olaparib or 2 lM cisplatin on
cell survival.
Table S1. Culture medium components for human
prostate organoids.
Table S2. Clinical information and pathological
parameters of donors for the established patient-
derived organoids.
1147Molecular Oncology 17 (2023) 1129–1147 Ó2023 The Authors. Molecular Oncology published by John Wiley & Sons Ltd on behalf of
Federation of European Biochemical Societies.
M. E. Elsesy et al. Functional detection of HRR defects in PCa models