Access to this full-text is provided by Springer Nature.
Content available from Nature Communications
This content is subject to copyright. Terms and conditions apply.
ARTICLE
Homologous recombination DNA repair deficiency
and PARP inhibition activity in primary triple
negative breast cancer
Neha Chopra1, Holly Tovey2, Alex Pearson 1, Ros Cutts1, Christy Toms2, Paula Proszek3, Michael Hubank3,
Mitch Dowsett1,4, Andrew Dodson4, Frances Daley1, Divya Kriplani1, Heidi Gevensleben1, Helen Ruth Davies5,6,
Andrea Degasperi 5,6, Rebecca Roylance7, Stephen Chan8, Andrew Tutt1,9, Anthony Skene10, Abigail Evans11,
Judith M. Bliss2, Serena Nik-Zainal 5,6 & Nicholas C. Turner1,12 ✉
Triple negative breast cancer (TNBC) encompasses molecularly different subgroups, with a
subgroup harboring evidence of defective homologous recombination (HR) DNA repair. Here,
within a phase 2 window clinical trial, RIO trial (EudraCT 2014-003319-12), we investigate
the activity of PARP inhibitors in 43 patients with untreated TNBC. The primary end point,
decreased Ki67, occured in 12% of TNBC. In secondary end point analyses, HR deficiency was
identified in 69% of TNBC with the mutational-signature-based HRDetect assay. Cancers
with HRDetect mutational signatures of HR deficiency had a functional defect in HR, assessed
by impaired RAD51 foci formation on end of treatment biopsy. Following rucaparib treatment
there was no association of Ki67 change with HR deficiency. In contrast, early circulating
tumor DNA dynamics identified activity of rucaparib, with end of treatment ctDNA levels
suppressed by rucaparib in mutation-signature HR-deficient cancers. In ad hoc analysis,
rucaparib induced expression of interferon response genes in HR-deficient cancers. The
majority of TNBCs have a defect in DNA repair, identifiable by mutational signature analysis,
that may be targetable with PARP inhibitors.
https://doi.org/10.1038/s41467-020-16142-7 OPEN
1Breast Cancer Now Toby Robins Research Centre, The Institute of Cancer Research, London CB2 0XZ, United Kingdom. 2Clinical Trials and Statistics Unit,
The Institute of Cancer Research, London, United Kingdom. 3The Centre for Molecular Pathology, The Royal Marsden Hospital, 15 Cotswold Road, Sutton SM2
5NG Surrey, United Kingdom. 4Ralph Lauren Centre for Breast Cancer Research, Royal Marsden Hospital, London, United Kingdom. 5Department of Medical
Genetics, The Clinical School, Box 238, Level 6 Addenbrooke’s Treatment Centre, Cambridge Biomedical Campus, Cambridge CB2 0QQ, United Kingdom.
6MRC Cancer Unit, Hutchison/MRC Research Centre, University of Cambridge, Box 197, Cambridge Biomedical Campus, Cambridge CB2 0XZ, United
Kingdom. 7University College London Hospitals NHS Foundation Trust, NIHR University College London Hospitals Biomedical Research Centre, London, United
Kingdom. 8Nottingham University Hospital Trust (City Campus), Nottingham, United Kingdom. 9Breast Cancer Now Research Unit, Cancer Centre, Guy’s
Hospital, King’s College London, London, United Kingdom. 10 Royal Bournemouth Hospital, Bournemouth, United Kingdom. 11 Poole Hospital NHS Foundation
Trust, Poole, United Kingdom. 12Breast Unit, The Royal Marsden Hospital, Fulham Road, London, United Kingdom. ✉email: nick.turner@icr.ac.uk
NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org/10.1038 /s41467-020-16142-7 | www.nature.com/naturecommunications 1
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Triple negative breast cancer (TNBC) may have diverse
defects in HR DNA repair, through germline mutations in
BRCA1,BRCA2 and PALB2, somatic mutations in BRCA1
and BRCA2, promoter methylation of BRCA1 and RAD51C, and
other as yet to be identified mechanisms1–4. Over the last decade,
advances in whole-genome sequencing (WGS) have lead to the
identification of mutational processes that leave a characteristic
imprint, a mutational signature on the cancer genome. These
have revolutionalised our understanding of cancer and has the
capability to improve diagnosis and treatment of cancer5,6.
Cancers with defects in HR-based DNA repair have character-
istic chromosomal changes reflecting the use of alternative error-
prone repair pathways, including measures of genomic instability;
loss of heterozygosity, telomeric allelic imbalance and large-scale
state transitions to accurately identify BRCA1/2 tumours7–9,and
their combination to form the HRD Score has allowed identifi-
cation of HR-deficient tumours (HRD Score >42), independent of
BRCA1/2 deficiency within a sporadic TNBC population10.Recent
work has identified WGS signatures of HR deficiency with BRCA1/
2deficient tumours associated with distinct mutational signatures.
The mutational signatures and chromosomal instability markers of
HR deficiency have been aggregated into the HRDetect score,
robustly identifying BRCA1/2 tumours with potential greater
accuracy than indexes such as HRD-score11,12.
Whether mutational signature-based scores such as HRDetect,
can be used to direct therapy in the clinic is unknown, in part as
there is limited direct evidence that cancers classified as HR
deficient by these scores have a functional defect in HR. Breast
cancers with BRCA1 and BRCA2 germline mutations are highly
sensitive to PARP inhibitors13,14, which target the underlying HR
DNA repair defect in these cancers. However, no activity was
observed with PARP inhibitors in the treatment of heavily pre-
treated un-selected advanced TNBC15. The extent to which this
PARP inhibitor efficacy may translate to sporadic TNBC is
unknown, as is the best way to identify HR-deficient TNBC. To
address these questions, we designed a translational clinical trial,
the RIO trial (EudraCT 2014-003319-12), with the objective of
identifying biomarkers of PARP inhibitor activity in sporadic
TNBC.
Results
Biomarkers of HR deficiency in primary TNBC. Patients with
newly diagnosed, treatment naïve TNBC were treated with the
PARP inhibitor rucaparib for 2 weeks prior to surgery or
neoadjuvant chemotherapy. A total of 43 patients were entered
into the trial between August 2015 and August 2017. Blood and
tissue biopsies were taken prior to, and at the end of treatment,
for molecular analysis (Fig. 1a). Within the trial, a subset of
germline BRCA1/2 patients were recruited as a control popula-
tion. The trial prospectively examined three potential biomarkers
of PARP inhibitor activity, a molecular signature of HR deficiency
aDid not start treatment (n = 1)
Did not complete 7 days of
treatment (n = 4)
Consent for RIO (n = 43)
Patients with samples
for analysis of
endpoints (n = 38)
Baseline ctDNA level
too low (n = 11)
ctDNA assay failed
(n = 1)
Patients with
paired biopsies (n = 37)
Paired biopsies for Ki67
analysis (n = 26)
Patients with
paired plasma (n = 31)
Assessed for ctDNA
dynamics (n = 19)
bc
Ki67 immunohistochemistry cleaved PARP immunohistochemistry
100 0.4
0.3
0.2
0.1
–0.1
–0.2
–0.3
–0.4
0
80
60
40
20
0
–20
–40
–60
% change from baseline
Logfold change from baseline
–80
–100
BRCA mutation
BRCA mutated
BRCA wild-type
BRCAWildtype
50% reduction 30% reduction
Biological population
Biological population
Fig. 1 RIO study CONSORT diagram and HRDetect analysis. a RIO study CONSORT diagram. bEffect of rucaparib on Ki67 expression assessed by
immunohistochemistry (IHC). The change in proportion of tumor cells expressing Ki67 between baseline and EOT, in patients that had assessable pairs of
baseline and EOT samples. BRCA mutation cancers had no evidence of decreased Ki67. cEffect of rucaparib on cleaved PARP expression assessed by
immunohistochemistry, as a marker of apoptosis. The change in proportion of tumor cells expressing cleaved PARP between baseline and EOT, in patients
that had assessable pairs of baseline and EOT samples. BRCA mutation cancers had no evidence of increased cleaved PARP expression. Grey bars, BRCA
wild type patients; Blue bars, BRCA germline mutant patients. Orange line, >30% but <50% reduction; Red line, >50% reduction.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7
2NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org /10.1038/s41467-020-16142-7 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
using HRDetect, RAD51 focus formation in a tumor biopsy at the
end of treatment, and BRCA1 methylation. The primary activity
end point was a fall in Ki67 on the end of treatment biopsy, with
circulating tumor DNA dynamics as a prospectively planned
exploratory end point of activity. Patient demographics were as
expected for this population (Table 1). Rucaparib was well tol-
erated with adverse effect profile similar to previous clinical
studies16,17 (Supplementary Table 1).
We assessed rucaparib activity and the relationship with
prospectively planned biomarkers. The primary end point was
assessed in tissue samples, using Ki67 suppression after two weeks
assessed by immunohistochemistry, as a potential biomarker
(Fig. 1b). A drop in Ki67 by 50% in the triple negative patients
without a known BRCA1/2 mutation at trial entry was seen in
12% tumors (95% CI: 2.5–31.2; n=3/25). In secondary end point
analysis, one additional patient with known BRCA mutation at
trial entry was assessed and did not have a 50% drop in Ki67
(Fig. 1b). No association was observed between Ki67 change with
BRCA1/2 mutated cancers (Fig. 1b). Similarly, no association was
observed with cleaved PARP levels as a marker of apoptosis with
BRCA1/2 mutated cancers (Fig. 1c).
In secondary end point analysis, baseline biopsies of patients
entering the trial, we interrogated the prevalence of HR deficiency
in primary TNBC (Fig. 2and Supplementary Table 2). We
performed whole-genome sequencing analysed with the HRDetect
assay11, identifying mutational processes characteristic of HR
deficiency in 69% locally assessed TNBC (18/26 with score >0.70,
herewith called HRDetect+ve; Fig. 2a,b and Supplementary Fig. 1a)
as well as an additional control ER-positive cancer with a known
germline BRCA2 mutation. In ad hoc analysis, we individually
determined the mutational status of BRCA1, BRCA2 and PALB2,
and promoter methylation of BRCA1 and RAD51C (Fig. 2a). Of the
HRDetect+ve cancers, 74%(14/19) had a detectable underlying
mutation of BRCA1/2 and PALB2 or gene promoter hypermethy-
lation of BRCA1 or RAD51C. None of the eight HRDetect−ve
cancers had an underlying genetic/epigenetic defect (p=0.0005,
Fisher’s exact test; Fig. 2a). A loss of heterozygosity (LOH)/copy-
number-based HRD score was positive in more cancers than
HRDetect, with the HRD score identifying cancers with genomic
instability but without rearrangement signatures and indels at
microhomology. None of the HRD score high but HRDetect low
tumours had detectable pathway aberrations (Supplementary
Fig. 1b), suggesting that HRDetect was more specific.
HRDetect identified all cancers with known HR pathway
defects, as well as additional sporadic cancers with no single
detectable defect (Fig. 2a). In secondary end point analysis, we
next addressed whether HRDetect+ve cancers had an underlying
functional defect in HR DNA repair, using RAD51 focus
formation in the end of treatment (EOT) biopsy. When cells
are exposed to genotoxic agents such as PARP inhibition, RAD51
is recruited to sites of DNA damage and stalled replication forks,
mediating the search for a homologous sequence during HR18,
with RAD51 nuclear foci visible at sites of repair as a hallmark
for HR-mediated repair19. The impaired ability to form RAD51
foci after DNA damage may identify cancers with defective
HR20. We developed a novel immunohistochemistry assay to
assess RAD51 foci, co-staining with geminin (GMNN) to
identify cells in S/G2 phase of the cell cycle and after cytotoxic
treatment; RAD51 score <20% (less than 20% geminin positive
cells having RAD51 foci, RAD51 foci deficiency) was assessed to
indicate HR deficiency in an independent sample set (Supple-
mentary Fig. 2). Within the RIO trial, RAD51 IHC scores
increased significantly from baseline to EOT (p=0.0016,
Wilcoxon test), reflecting rucaparib induced DNA damage and
RAD51-mediated repair (Fig. 3a). In EOT biopsies, RAD51 foci
deficiency was identified in 77% (17/22) locally assessed TNBC,
as well as an ER-positive BRCA2 mutant control cancer (Fig. 3b).
Of the RAD51-deficient cancers, 61% (11/18) had an underlying
detectable HR defect compared to none (0/5) of RAD51 foci
proficient cancers (p=0.037 Fisher’s exact test; Fig. 3b). Cancers
with RAD51 foci deficiency had significantly higher HRDetect
scores than tumour samples that were RAD51 foci proficient
(n=18, p=0.0146 Mann–Whitney test; Fig. 3c). HRDetect
therefore identified cancers with a functional defect in HR-based
DNA repair, with functional HR deficiency occurring in the
majority of TNBC.
Rucaparib activity assessed by ctDNA dynamics. Circulating
tumour DNA (ctDNA) is released from the tumor, allowing for
serial sampling through the course of treatment21,22. Early
changes in ctDNA dynamics represent an early biomarker of drug
Table 1 RIO study patient demographics.
n%
Age (mean (standard deviation)) 54.6 (13.9)
Age group (years) <40 7 16.3
40-49 13 30.2
50-59 9 20.9
60-69 5 11.6
70+9 20.9
BRCA status Triple negative, known BRCA1/2
carrier at registration
2 4.7
Not TN, known BRCA1/2
mutation carrier at registration
1 2.3
Triple neg, no BRCA mutation 35 81.4
Triple negative, BRCA1/2
mutation identified while on trial
5 11.6
Planned standard
treatment after
rucaparib
Neoadjuvant chemotherapy 32 74.4
Surgical resection 11 25.6
Hormone
receptor status
ER & PR negativea42 97.7
ER positive & PR negative 1 2.3
Tumour grade
(diagnostic sample)
G1 0 0
G2 12 27.9
G3 24 55.8
Not known 7 16.3
Histological type Infiltrating ductal 38 88.4
Infiltrating lobular 4 9.3
Mixed ductal & lobular 1 2.3
DCIS present Yes 7 16.3
No 35 81.4
Not known 1 2.3
Tumour size by
ultrasound
<1.5 2 4.7
1.5 1 2.3
>1.5 & ≤2 10 23.3
>2 & ≤5 26 60.5
>5 4 9.3
Lymph node
involvement
Yes 16 37.2
No 27 62.8
Side of tumour Left 18 41.9
Right 25 58.1
Evidence of metastatic
disease
Yes 0 0
No 43 100
aOne patient was locally assessed as triple negative but central assessment noted weak PgR
score of 3/8 by Allred.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org/10.1038 /s41467-020-16142-7 | www.nature.com/naturecommunications 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
activity, as cancers that respond to treatment rapidly suppress the
level of ctDNA in plasma23–25. Analysis of ctDNA was pro-
spectively planned as an exploratory end point of rucaparib
activity, in part, as Ki67 change has only been validated as an
activity end point in endocrine based therapies26–28. To assess
ctDNA in RIO, the primary tumor was sequenced in 35 patients,
with somatic mutations identified in 31 patients, and personalised
digital PCR used to track changes in ctDNA levels between
baseline and end of treatment plasma (EOT). Change in ctDNA
was assessable in the 19 patients with sufficiently high baseline
Patient #4, HRDetect <0.7
b
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
21
Patient #27, HRDetect >0.7
aHRDetect positive n = 19
HRDetect negative n = 8
p = 0.0005 Fishers’s exact test
BRCA1/2 germline mutation
BRCA 1 methylation
PALB2 germline mutation
RAD51C methylation
None/Unknown
None/unknown
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Patient
HRDetect score
***
1 2 3 4 5 6 7 8 9 11 12 14 15 16 17 18 19 20 21 22 23 24 25 26 2713 10
22
X
Y
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
21
22
X
Y
C>A Sig26
4841 substitutions
348 deletions and insertions
14,000 substitutions
1302 deletions and insertions
Copy number
51 rearrangements
Complex
Insertion
Deletion other
M-homology
Deletion repeat
deletion
Sig20
Sig18
Sig17
Sig13
RS1
RS2
RS3
RS4
RS5
RS6
0 5 10 15 20 25 30 35
370 rearrangements
RS1
RS2
RS3
RS4
RS5
RS6
050 100 150 200
Sig8
Sig6
Sig5
Sig3
Sig2
Sig1
Sig26
02000 4000 6000 8000 10,000
Sig20
Sig18
Sig17
Sig13
Sig8
Sig6
Sig5
Sig3
Sig2
Sig1
0500
0
LOH Gain
Copy number
LOH Gain
50 100 150
Complex
Insertion
Deletion other
M-homology
Deletion repeat
deletion
0
100
200
300
400
500
600
700
1000 1500 2000 2500
C>G
C>T
T>A
T>C
T>G
C>A
C>G
C>T
T>A
T>C
T>G
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7
4NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org /10.1038/s41467-020-16142-7 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Fig. 2 HRDetect analysis. a HRDetect scores were established with whole-genome sequencing in baseline biopsies of 26 patients with untreated primary
triple negative breast cancer (locally assessed) entering the RIO window clinical trial. HRDetect positive cancers (HRDetect score >0.7) were enriched for
inactivating mutations and promoter methylation of HR genes compared to HRDetect negative cancers (p=0.0005, Fisher’s exact test). Star, an additional
patient with ER-positive breast cancer and BRCA2 germline mutation is shown. 2 stars, locally assessed as TNBC but centrally assessed as PR-positive
breast cancer with BRCA1 methylation. Magenta bars, BRCA1/2 germline mutation; Blue bars, BRCA1 methylation; Yellow bars, PALB2 germline mutation;
Turquoise bars, RAD51C methylation; Grey bar, None/unknown. bExamples of genome plots for a sample with low (top) and high (bottom) HRDetect
scores. The histograms associated with each circos plot show mutation counts for each mutation class: the topmost histogram shows the number of
mutations contributing to each substitution signature; the middle histogram represents indel patterns; and the bottom histogram shows the number of
rearrangements contributing to each rearrangement signature.
a
b
c
RAD51 foci proficient
patient #5
RAD51 foci deficient
patient #14
Baseline EOT
0
10
20
30
40
50
RAD51 score
p = 0.0016, n = 21
p = 0.0146, n = 18
HR Detect score
RAD51 foci
deficient
RAD51 foci
proficient
0.0
0.2
0.4
0.6
0.8
1.0
RAD51 foci deficient n = 18
p = 0.037, Fisher’s exact test
RAD51 foci proficient n = 5
BRCA1/2 germline mutation
BRCA 1 methylation
PALB2 germline mutation
RAD51C methylation
None/Unknown
None/Unknown
0
10
20
30
40
50
RAD51 IHC score
020
40 µm
020
40 µm
6 8 15 27 14 21 25 22 17 31 18 33 28 39 2 12 24 26 5 4 29 1 9
Patient
Undetected
***
Fig. 3 Biomarkers of homologous recombination (HR) repair deficiency in TNBC. a RAD51 focus assessment in paired baseline and on-treatment
biopsies, p=0.0016 Wilcoxon test. Inset right, example immunohistochemistry images of two cancers, one deficient and one proficient in RAD51 focus
formation (brown), in nuclei stained for geminin (blue). Scale bars show 40 μm. bRAD51 immunohistochemistry score (>20% RAD51 foci proficient), as a
functional assessment of HR proficiency, was assessed in end of treatment biopsies after 2 weeks of rucaparib in 25 patients. RAD51 foci deficient cancers
are enriched for inactivating mutations and promoter methylation of HR genes compared to RAD51 proficient cancers (p=0.037 Fisher’s exact test). *ER-
positive breast cancer and BRCA2 germline mutation is shown. **Locally assessed TNBC but centrally assessed as PR-positive breast cancer with BRCA1
methylation. Magenta bars, BRCA1/2 germline mutation; Blue bars, BRCA1 methylation; Yellow bars, PALB2 germline mutation; Turquoise bars, RAD51C
methylation; Grey bars, None/unknown. cAssociation between RAD51 foci proficiency and HRDetect scores in 18 patients, p=0.0146 Mann–Whitney
UTest. Line indicates median level.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org/10.1038 /s41467-020-16142-7 | www.nature.com/naturecommunications 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
ctDNA to assess change (Figs. 1a and 4a and Supplementary
Tables 4 and 5). In contrast to the tumor biopsy-based data, a
substantial proportion of patients suppressed ctDNA after ruca-
parib treatment (Fig. 4a).
Patients with germline BRCA1/2 mutations did not suppress
Ki67, nor induce PARP cleavage, in the end of treatment biopsy
(Fig. 1). In contrast, patients with germline BRCA1/2 mutations
had a greater suppression of ctDNA than patients without
germline mutations (n=19, p=0.021, Mann–Whitney; Fig. 4b),
validating ctDNA dynamics as a marker of rucaparib activity.
Cancers with deficient RAD51 foci formation (n=12, p=0.033,
Mann–Whitney) and HRDetect+ve cancers had greater ctDNA
suppression (n=15, p=0.027, Mann–Whitney; Fig. 4b). In ad
hoc analysis patients with suppressed ctDNA after two weeks
rucaparib (ctDNA ratio (CDR) <0.25, methods) were enriched for
germline mutations of BRCA1/2 and PALB2 and gene promoter
methylation of BRCA1 and RAD51C (Fig. 4b). These data
illustrate the potential of ctDNA analysis to transform window
trials, presenting a simple and robust assay of drug activity,
without the potential sampling challenges involved with repeat
biopsies. However, analytical challenges associated with low
plasma DNA levels and low purity tumor samples, will benefit
from further technological development.
PARP inhibition induces an interferon response in HR-
deficient cancers. Having demonstrated that HRDetect identi-
fies sporadic TNBC in which PARP inhibitors have activity, in ad
BRCA1/2 germline mutation
BRCA 1 methylation
PALB2 germline mutation
RAD51C methylation
None/Unknown
Normalised mutant copies /mL
0.0
0.03125
0.0625
0.125
0.25
0.5
1
2
4
Baseline EOT
121523
34
7
17
18
24
2822
19
35
11
16
31
142
5, PD within 6 months
aPatient 15
BRCA2, c.36_41TGAAATdelAGAAAins
Baseline
2000
4000
6000
8000
10,000
12,000
14,000
VIC Fluorescence
FAM Fluorescence
BRCA2 wt
BRCA2 ins
8000
6000
4000
End of treatment
2000
4000
6000
8000
10,000
12,000
14,000
VIC Fluorescence
FAM Fluorescence
BRCA2 wt
8000
6000
4000
b
RAD51
deficient
RAD51
proficient
p = 0.027 (Mann–Whitney), n = 12
CDR 15 Ratio
RAD51 score vs. CDR15 HR Detect vs. CDR15
p = 0.028 (Mann–Whitney), n = 15
HR Detect
>0.7
HR Detect
<0.7
1.5
2.0
2.5
0.0
0.5
1.0
CDR 15 Ratio
p = 0.021 (Mann–Whitney), n = 19
BRCA status vs. CDR15
0.0
0.5
1.0
1.5
2.0
2.5
CDR 15 Ratio
BRCA
germline
Non
BRCA
PALB2
germline
RAD51C
methylated
BRCA2
methylated
0.0
0.5
1.0
1.5
2.0
2.5
Fig. 4 ctDNA dynamics reveals activity of rucaparib in primary triple negative homologous recombination repair deficient cancers. a Change in
circulating tumour DNA (ctDNA) copies/ml between baseline and end of treatment (EOT) after two weeks of rucaparib. The relative change of on-
treatment ctDNA levels (Circulating tumor DNA ratio, CDR15) with HR pathway defects indicated. Right, example digital PCR ctDNA analysis plots.
Magenta lines, BRCA1/2 germline mutation; Blue lines, BRCA1 methylation; Yellow lines, PALB2 germline mutation; Turquoise lines, RAD51C methylation;
Grey lines, None/unknown. bAssociations of ctDNA change on rucaparib at day 15 (CDR15) with left, BRCA1/2 germline mutations (n=19 patients),
middle, RAD51 focus proficiency (n=12 patients) and right, HRDetect score (n=15 patients). Centre line, mean; error bars, standard deviation. pvalues
Mann–Whitney Utest.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7
6NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org /10.1038/s41467-020-16142-7 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
hoc analysis we investigated the mechanisms of PARP inhibition
in these cancers with RNA exome sequencing in 20 paired
tumour samples, baseline and end of treatment. The majority of
tumours were basal-like by PAM50 (75%, 15/20), with HRDetect
scores >0.7 found in 80% (12/15) basal-like TNBC and none (0/4)
of the non-basal TNBC, with one non-basal BRCA1 mutant ER+
tumour having an HRDscore >0.7 (p=0.031 Fisher’s exact test,
Fig. 5a). Analysis of the immune micro-environment of baseline
samples with CIBERSORT29 suggested higher levels of follicular
helper T cells (p=0.0126, n=20 patients) in HRDetect+ve
cancers along with higher levels of activated macrophages (Sup-
plementary Fig. 3a and b).
a
c
b
PAM50
Basal n = 15
HR Detect >0.7
HR Detect <0.7
PAM50
Non-Basal n = 5
ER+ve BRCA1/2
germline
p = 0.031 Fishers’s exact test
d
Differentially expressed genes —
Good vs Poor response
Differentially expressed genes —
Low CDR15 vs High CDR15
Interferon-α response
Enrichment score
G2M Checkpoint
Enrichment score
Interferon gamma response
Allograft rejection
Interferon alpha response
Inflammatory response
NFkB signalling
IL6 JAK STAT3 signalling
Complement
IL2 STAT5 signalling
Coagulation
Epithelial mesenchymal transition
KRas Signalling upregulation
p53 pathway
Apoptosis
Apical junction
Xenobiotic metabolism
Spermatogenesis
KRas Signalling downregulation
Reactive oxygen species pathway
Adipogenesis
Fatty acid metabolism
UV response upregulation
Unfolded protein response
MTORC1 signalling
G2M checkpoint
Cholesterol homeostasis
Protein secretion
MYC targets v2
E2F targets
Oxidative phosphorylation
MYC targets v1
Pathway
Pathways NES from GSEA
Corrected
p < 0.05
False
True
Normalized enrichment score
None/Unknown
BRCA1/2 germline mutation
BRCA 1 methylation
RAD51C methylation
PALB2 germline mutation
Baseline
EOT
13
14
15
16
17
Gene expression Log2
Gene expression Log2
Gene expression Log2
Gene expression Log2
PARP1
q = 0.023
9
11
13
15
17
19
MKI67
q = 0.027
Baseline
EOT
10
11
12
13
14
15
CDKN1A
q = 0.002
Baseline
EOT
8
9
10
11
12
13
TMEM173
q = 0.05
Baseline
EOT
–2 –1 0 1 2 3
0.0
0.6
0.4
0.2
0.0
–0.1
–0.2
–0.3
15,00010,000
rank
50000 15,00010,000
rank
50000
415
1
2091
Good responders
Poor responders
13
Low CDR
37
High CDR
2
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org/10.1038 /s41467-020-16142-7 | www.nature.com/naturecommunications 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Cancers with ctDNA suppression (CDR < 0.25) had more
substantial changes in gene expression than cancers without ctDNA
suppression (CDR > 0.25), with 415 compared to 37 significantly
differentially expressed genes respectively (Fig. 5b), demonstrating
stability of gene expression through rucaparib treatment in HR-
proficient cancers. For individual genes, a significant decrease in
gene expression was noted in PARP1 mRNA (q=0.023, n=20,
DESeq2 with false discovery rate (FDR) correction), MKI67 mRNA
(Ki67, q=0.027, n=20, DESeq2 FDR correction), as well as
induction in gene expression of both CDKN1A mRNA (p21, q=
0.002, n=20, DESeq2 FDR correction) and TMEM173 mRNA
(STING, q=0.05, n=20, DESeq2 FDR correction; Fig. 5c). In
HRDetect+ve cancers (Supplementary Fig. 4) and cancers with
associated ctDNA suppression (Fig. 5d) pathway analysis demon-
strated significant gene expression changes in pathways that regulate
proliferation, apoptosis and immune function. There was decreased
expression of G2M checkpoint genes, reflecting cell cycle arrest with
rucaparib (Fig. 5d). There was substantial increase in interferon
response genes, which along with TMEM173 mRNA expression,
suggested PARP inhibition activated the cGAS–cGAMP–STING
pathway in HR-deficient cancers (Fig. 5d and Supplementary Fig. 4).
In contrast, HRDetect-ve cancers did not induce interferon response
genes, nor had expression changes of cell cycle arrest, on rucaparib
(Supplementary Fig. 4).
Discussion
Results from the RIO trial demonstrate that a majority of primary
TNBC have defects in HR-based DNA repair, reinforcing the
companion population-based TNBC study by Staaf et al4. The
primary end point of RIO, suppression of Ki67 in on-treatment
biopsies, was infrequent. In pre-planned secondary analyses, we
show that cancers with HR deficiency can be robustly identified
with the mutational-signatures based classifier HRDetect (Fig. 2
and Supplementary Fig. 1), which identifies cancers with a
functional deficiency in HR (Fig. 3b), and with evidence of
activity of PARP inhibitors restricted to these cancers using
ctDNA analysis (Fig. 3c). In ad hoc analysis, HRDetect was more
specific to underlying HR deficiency than HRD scores (Supple-
mentary Fig. 1b), suggesting that mutational signature assessment
might be more accurate in identifying cancers that would benefit
from platinum chemotherapy or PARP inhibition4,30–32.
Induction of RAD51 nuclear foci after neoadjuvant chemotherapy
and PARP inhibition can measure the homologous recombination
functionality in breast cancer biopsies20,33–36, with an association to
loss of heterozygosity measures of HR deficiency34,37. Studies have
shown that cells with deficient BRCA1/2 or other HR proteins, do
not efficiently form RAD51 foci which could be used as a marker for
PARP inhibitor sensitivity using FFPE tumour samples35,38,39.The
dynamics of DNA repair alter throughout tumour evolution and a
functional RAD51 assay can be used as a dynamic readout of
tumour HR status at the specific time for treatment decision-
making33,40,41. Current studies are using immunofluorescence (IF)
on FFPE samples which can be labour intensive, and here we
develop a RAD51 foci immunohistochemistry (IHC). Our cutoff
<20% for HR deficiency is consistent with the RECAP test which has
recently shown to be effective in ascertaining HR deficiency in
metastatic breast tumours treated with ionising radiation36.Further
validation of this novel HR deficient biomarker is required and if
clinically validated could be a useful tool in the clinic.
One of the main limitations of the study was lower recruitment
into the trial than was anticipated. This was possibly due to
patient preference to start treatment without the possibility of
delaying for short-term trial therapy. Additionally, failure of tis-
sue biopsies containing enough tumour content impacted end
point analysis. We therefore demonstrate the advantage of non-
invasive analysis, and the use of ctDNA as a potentially reliable
and effective surrogate end point to assess response. This will
require further validation.
We demonstrate that PARP inhibitors induce a pro-inflam-
matory/interferon response in HR-deficient TNBC, likely though
the cGAS–cGAMP–STING pathway. Consistent with our find-
ings, PARP inhibition has previously been shown to induce T-cell
recruitment through activation of the cGAS-cGAMP-STING
pathway in a BRCA1-null mouse model of TNBC42. Similarly,
PARP inhibitors have been shown to upregulate interferon
response in TNBC cell lines with BRCA2 depletion43 or mutant
BRCA144. Furthermore, Sceneay et al45 recently demonstrated
using mouse models of TNBC, that immune dysfunction char-
acterised by decreased interferon signalling and decreased antigen
presentation was abrogated by a STING agonists. Together these
findings underline the potential of exploiting immune dysfunc-
tion in the context of HR deficiency, notably in BRCA1/2 mutant
tumours, and TNBC more generally.
Our findings illustrate the potential of using whole-genome
sequencing mutational signatures to guide cancer treatment,
advocating for clinical trials of PARP inhibitors, potentially in
combination with PD(L)1 targeting immune checkpoint anti-
bodies, in sporadic TNBC with HR deficiency.
Methods
Study Design. Conducted in 10 hospitals throughout the United Kingdom, the
window study of the PARP inhibitor rucaparib in patients with primary triple
negative or BRCA1/2 related breast cancer (The RIO study; EudraCT 2014-003319-
12, Cancer Research UK trial CRUK/12/034) was a single-group, open-label, phase
II window of opportunity trial assessing rucaparib efficacy in patients with primary
triple negative or BRCA1/2 mutant breast cancer prior to commencing primary
treatment (neoadjuvant chemotherapy or surgery). The trial was co-sponsored by
the Institute of Cancer Research and the Royal Marsden Hospital NHS Foundation
Trust. Ethical approval for The RIO trial was given by the NRES Committee
London - Fulham Research Ethics Committee (REC ID: 14/LO/2181) and
informed consent was obtained from all patients enrolled in the study.
Key eligibility criteria include breast tumour size ≥2 or <2 cm with cytologically/
histologically confirmed axillary lymph nodes, WHO performance status 0–2, no
prior history of ipsilateral breast cancer within 5 years and no prior treatment with
PARP inhibitors. Patients received rucaparib 600 mg twice daily for 12 –14 days.
Baseline bloods (EDTA and STRECK) and core biopsies (FFPE and RNAlater™)
were collected at time of diagnostic biopsy or following trial entry. End of
Fig. 5 Expression analysis of primary triple negative homologous recombination repair deficient cancers. a Association between basal-like and non-
basal-like triple negative subtypes, assessed by PAM50, and HRDetect score. pvalue Fisher’s exact test. n=20 paired tumour samples. Purple, HRDetect
score >0.7; Grey, HRDetect score <0.7. bChange in gene expression on paired tumor biopsies between baseline and end of treatment on rucaparib.
Number of genes with a significant change in gene expression (Log fold change >0.5 and false discovery corrected qvalue < 0.1). Left, categorised by
ctDNA suppression or not (CDR15 response <0.25 vs ≥1, n=8 and n=3 paired tumour samples respectively). Right, by CDR15 response and HRD score
(CDR15 < 0.25 and HRD > 0.7 vs CDR15 ≥1 and HRD < 0.7, n=11 and n=6 paired tumour samples respectively). cGene expression changes of PARP1,
MKI67, CDKN1A and TMEM173 through treatment, from DESeq2 with false discovery rate (FDR) corrected qvalue for change. n=20 paired tumour
samples. Magenta lines, BRCA1/2 germline mutation; Blue lines, BRCA1 methylation; Yellow lines, PALB2 germline mutation; Turquoise lines, RAD51C
methylation; Grey lines, None/unknown. dLeft, Gene set enrichment pathway analysis (GSEA) for gene expression changes through treatment in patients
with ctDNA suppression (CDR15 < 0.25, n=8 paired tumour samples.) on rucaparib. Centre, suppression of G2M checkpoint genes on PARP inhibition,
q=0.006, and right, increased expression of interferon pathway genes on PARP inhibition, q=0.001. False discovery rate corrected qvalue for change.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7
8NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org /10.1038/s41467-020-16142-7 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
treatment bloods (STRECK) and biopsies (FFPE and RNAlater™) were taken at
surgery or prior to neo-adjuvant chemotherapy within 24–48 h of the last rucaparib
dose. The primary end point was Ki67 response from baseline to end of treatment
defined as a ≥50% decrease. Secondary endpoints (pre-specified endpoints in the
trial protocol) were association between baseline biomarkers of BRCA1 methylation
and a genomic predictor of HR deficiency (HRdetect) with Ki67 response to
rucaparib, apoptosis induction following 12–14 days of rucaparib, the proportion
of sporadic TNBC that fail to induce RAD51 foci on end of treatment biopsy
(RAD51 score), safety and tolerability of rucaparib, association of biomarkers with
RAD51 score, the proportion of patients with a change in circulating tumour DNA
levels in response to rucaparib, association between change in circulating tumour
DNA levels with biomarkers, and proportion of patients with BRCA1 and BRCA2
germline mutation related cancers with a Ki67 response to rucaparib and the
proportion that have reduced RAD51 score and increased apoptosis induction.
Pathological complete response was not assessed as an end point, as the two weeks
window was considered to be too short for this end point. Where molecular
analysis was not pre-specified in the trial protocol, this is highlighted as being ad
hoc analsyis.
Ki67 IHC and Scoring. Immunohistochemistry for Ki67 was performed and
scored according to Leung et al46. Tissue sections were deparafinised and rehy-
drated prior to antigen retrieval using low pH (pH 6.0) Target Antigen Retrieval
Solution (K8005, Dako UK Ltd). Tissue sections were stained for Ki67 using mouse
monoclonal anti-Ki67 antibody (MIB-1 clone, Dako, M7240) diluted 1:50 in
EnVision Antibody Diluent (K8006, Dako UK Ltd). Sections were washed using
wash buffer (Dako, S3006) before colour development using REAL Kit (Dako,
K5001). Tissues sections were counter stained with hematoxylin and coverslips
mounted using DPX. The percentage of Ki67 positive cells was calculated from
between 200-400 invasive tumour cells according to the method of Leung et al46.A
Ki67 response was determined as a >50% fall in Ki67 at EOT compared to baseline.
Cleaved PARP IHC and scoring. IHC for cleaved PARP (cPARP) was performed
using a rabbit monoclonal antibody specific for PARP cleaved at Asp214. Tissue
sections were deparafinised and rehydrated prior to antigen retrieval using High
pH (pH9.0) Target Antigen Retrieval Solution (K8004, Dako UK Ltd), in PT-LINK
(PT101, Dako UK Ltd). Tissue sections were stained for cPARP using rabbit
monoclonal anti-cPARP antibody (Asp214, clone D64E10, Cell Signaling Tech-
nology, #5625) diluted 1:100 in EnVision Antibody Diluent (K8006, Dako UK Ltd).
Sections were washed using wash buffer (Dako,S3006) before colour development
using anti-rabbit link reagent EnVision FLEX LINKER (K8019, Dako UK Ltd).
Tissues sections were counter stained with hematoxylin and coverslips mounted
using DPX. The percentage of cPARP positive cells was calculated from a mini-
mum of 500 invasive tumour cells. If fewer than 500 invasive tumour cells were
present the sample was recorded as insufficient invasive tumour (IIT).
Sample processing. All samples were processed by the central laboratory as part
of the RIO trial. Fresh tumour samples were collected in RNAlater™tubes, pro-
cessed within 24 h and stored at -80oC until required for extraction. Baseline
biopsies were sectioned using a cryostat. One section was cut for H&E and
16 sections were cut and stained with Nuclear Fast Red (NFR). A second H&E
section was cut at the end of the series. H&E sections were reviewed and marked by
a pathologist for macro-dissection. If the baseline biopsy did not have tumour the
EOT biopsy (day 12–14) was used for sequencing. DNA was extracted using the
Qiagen DNeasy Blood and Tissue kit according to the manufacturer’s instructions.
DNA was eluted into 200 μl buffer ATE and stored at -20oC before quantification.
DNA was quantified on the Bio-Rad QX-200 ddPCR system running Quantasoft
v1.7, using the RPPH1 (RNaseP, cat# 4403328 ThermoFisher) reference assay to
calculate copies/well and multiplying by the cvalue (3.3 pg), an estima te of the
mass of a single haploid human genome.
Blood collected in STRECK preservation tubes at baseline (day 1 prior to
treatment) and end of treatment were processed within 24 h of sample collection .
Plasma and buffy coat was separated by centrifugation 1600g for 20 min and stored
individually at −80oC until DNA extraction. For plasma extraction, up to 4mls of
archived plasma was extracted using the automated MagMax Cell-Free DNA
Isolation Kit (Thermo Cat # A29319) and ThermoScientific KingFisher Flex
Purification System. DNA was eluted into 100 μl buffer AVE and stored at −20oC.
Buffy coat extraction was performed using the Qiagen DNeasy blood and tissue kit
as per manufacturer’s instructions. DNA was eluted into 100 μl buffer AE and
stored at −20 °C. Plasma DNA was quantified on the Bio-Rad QX-200 ddPCR
system using the RPPH1 reference assay to calculate copies/well and multiplying by
the cvalue (3.3 pg), an estimate of the mass of a single haploid human genome.
HRDetect Assay. Extracted DNA from fresh tissue with >20% tumour content
and >200 ng quantifiable DNA, along with paired buffy coat germline DNA, were
subject to whole-genome sequencing at the Sanger Institute, Cambridge, UK.
A 500-bp insert genomic libraries were constructed according to Illumina
library protocols and 150 bp paired-end sequencing performed on an Illumina
HiSeq X Ten using HCS (v3.5.0) for HiSeq X systems, to an average sequence depth
of 38.5× for both tumour and normal. The resulting reads were aligned to the
reference human genome (GRCh37) using Burrows-Wheeler Aligner (BWA)
(0.7.16a (r1181)). Mutation calling was performed as described previously12.
CaVEMan (Cancer Variants Through Expectation Maximization: http://cancerit.
github.io/CaVEMan/) was used for calling somatic substitutions. Indels in the
tumour and normal genomes were called using a modified Pindel version 2.0.
(http://cancerit.github.io/cgpPindel/). Structural variants were discovered using a
bespoke algorithm, BRASS (BReakpoint AnalySiS) (https://github.com/cancerit/
BRASS). All annotation was to Ensembl build 75. Allele-specific copy number
analysis of tumours was performed using ASCAT (v2.1.1) applied to next-
generation whole-genome sequencing data as described previously11,12. Copy
number values and estimates of aberrant tumour cell provided by ASCAT were
input into the CaVEMan substitution algorithm. In addition, ASCAT segmentation
profiles were used to establish the presence of copy number changes and loss of
heterozygosity across the BRCA1,BRCA2 and PALB2 genes.
The predominant mutational signatures present in breast cancer have been
identified in a large WGS study involving 560 breast cancers. These comprise
12 substitution signatures and 6 structural rearrangement signatures. The
contributions of these consensus mutational signatures were estimated in the 27
RIO trial WGS samples as described previously12,47. In addition, the contribution
of small insertions and deletion at regions of micro-homology or repeats and HRD
LOH index were estimated7.
Mutational signature contributions for substitution signatures 3 and 8,
rearrangement signatures 3 and 5, deletions at microhomology and HRD LOH
index were calculated for each sample as input in to the weighted model, HRDetect.
The HRDetect algorithm was run as described previously, using the previously
described weights11.
Targeted tissue sequencing. Paired tissue and buffy coat DNA were sent to the
Centre of Molecular Pathology at The Royal Marsden Hospital for sequencing
using a targeted capture-based approach designed to detect mutations and
amplifications frequently seen in breast cancer. The targeted panel, “ABC-Bio”
panel, has been validated in the ABC-Bio (molecular screening for patients with
advanced breast cancer) trial and comprises of 41 genes commonly mutated in
breast cancer48. Libraries were run on a MiSeq (Illumina) using MiSeq Reporter
(MSR v2.5.1). All internally developed code is accessible on request.
AVENIO sequencing. ctDNA samples from six patients from whom adequate
tumour samples were not obtained were sent to Roche for sequencing using the
AVENIO ctDNA targeted tumour profiling kit (Roche Sequencing). A total of
4 samples were sequenced on the AVENIO ctDNA targeted kit (17 genes) and
2 samples were run on the AVENIO ctDNA expanded kit (77 genes), using
HighOutput 300 cyc kit on a NextSeq 500 (Illumina) using NextSeq system suite
(v2.2.0). Data was analysed using the AVENIO Oncology analysis software (v1.0.0
and v1.1) available from Roche.
BRCA1 and Rad51C methylation, bisulfite sequencing. The promoter region of
BRCA1 and RAD51C was identified using the Eukaryotic Promoter Database
(http://epd.vital-it.ch/index.php). BRCA1 promoter was amplified with forward-
TATTTTGAGAGGTTGCTGTTTAG and reverse-CTAAAAAACCCCACAACC
TATCCC primers. Analysis of BRCA1 methyaltion was pre-planned, and RAD51C
methylation was added ad hoc to BRCA1 methylation prior to association with
activity surrogates, as the potential importance of RAD51C methylation in TNBC
was only recognised after the start of the trial2. The RAD51C promoter was
amplified with forward-TGGTAATTGGTTAGTGTGTGT and reverse-TCCTCA
TCAAATATACACCCTAACT primers. BRCA1 and RAD51C PCR conditions
were optimised for multiplex assay using ThermoFisher Scientific AccuPrime Hi-
fidelity Taq. Human methylated and unmethylated DNA (Zymo Research, Human
HCT116DKO non-methylated DNA and HCT116DKO methylated DNA) primers
were used as a control.
Extracted DNA from RIO RNA Later samples were subjected to bisulfite
sequencing (Zymo Research Methylation Gold spin column kit D5005). Total DNA
input ranged from 10 to 500 ng. Samples were quantified post bisulfite sequencing
using Qubit 3.0 fluorimeter and subsequently subjected to PCR using
ThermoFisher Scientific AccuPrime Hi-fidelity Taq at 60 °C for 34 cycles. Samples
were cleaned using Qiagen QIAquick PCR purification kit (ID:28104) and
quantified using Qubit 3.0 fluorimeter. Samples were subjected to Illumina
NebNext Ultra II library preparation. Total library input ranged from 10-116 ng
and PCR cycles were adjusted accordingly, and sequenced on the Illumina MiniSeq
platform with Miniseq system suite v1.1. Mean number of reads for BRCA1
amplicon was 36909 (range 9654-60084) and RAD51C amplicon 48879 (range
28404- 71129) with a mean 47% of reads on target for the methylation sequencing
run (range 37-50%).
Bioinformatics analysis of methylation followed a similar workflow to previous
studies32. Paired overlapping reads were merged into a single sequence using
flash49 after adaptor trimming using trim-galore (https://www.bioinformatics.
babraham.ac.uk/projects/trim_galore/). Each read was aligned using pairwise
alignment to the BRCA1 or RAD51C amplicons using Biostrings R50 package with
90% identity. Reads with more than 1 mismatch in alignment were additionally
removed. Reads with incomplete bisulphite conversion were removed by
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org/10.1038 /s41467-020-16142-7 | www.nature.com/naturecommunications 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
calculating the unconverted cytosine count at non CG sites as well as reads where
all CG sites in the read were not C or T. Reads were assessed as being methylated
when >90% of CpG sites in the amplicon were methylated. For RAD51C
methylation, 2 sites in the RAD51C amplicon were removed as these were found to
be consistently methylated in all samples.
Immunohistochemistry for RAD51/Geminin (GMNN) double staining. FFPE
samples taken 24–48 h after the last dose of rucaparib were cut, deparafinised,
rehydrated and stained with hematoxylin and eosin (H&E) and double stained with
GMNN and RAD51. Pre- and post-radiotherapy-induced squamous cell carcinoma
were used as negative and positive controls.
Antigen retrieval was performed and RAD51 primary antibody (mouse
monoclonal, Genetex, GTX70230) was diluted 1/200 in Dako antibody diluent
(K8006) and applied. Slides were incubated with Dako Envision Flex HRP (K8002).
Geminin (GMNN) antibody (rabbit polyclonal, Proteintech 10802-1-AP) diluted
1/1500 in Dako antibody diluent was applied and incubated with Dako Envision
Flex HRP (K8002). Sections incubated with Vector TMB blue (SK4400) and
counterstained with Gills 1 hematoxylin, air dried and dehydrated in xylene before
mounted and cover slipped with Vectamount.
Five random fields at 40× magnification were identified and marked in PathXL.
GMNN staining was identified with blue/green staining and RAD51 was identified
by the presence of brown nuclear foci. Scoring was done by 2 scorers blinded to
each other, time point and clinical details.
The number of tumour cells, GMNN-positive cells and RAD51-positive cells
were counted. Cells with 5 or more RAD51 foci were classified as RAD51 positive.
A minimum of 300 tumour cells and 30 GMNN-positive cells were a minimum
requirement for inclusion. Raw data were collected and the proliferation fraction
((no. of GMNN-positive cells / total number of tumour cells) × 100) and RAD51
score ((no. of RAD51-positive cells / no. of GMNN-positive cells) × 100) was
calculated.
Circulating tumor DNA analysis. For each trial subject, dPCR assays were
designed for the mutations identified by tissue sequencing according to the method
of Garcia-Murillas et al22. Assays were optimised with temperature gradients and
patients with more than one mutation had multiplex assays optimised. If optimi-
sation was not achieved with multiplexes, a singleplex assay was used.
Mutation analysis was done using ddPCR assays specific for each patient’s
mutation(s) on a QX200 system (Bio-Rad) running Quantasoft v1.7. 50 μl DNA
(2 ml plasma equivalent) was used divided equally into 2 wells from each time
point. DNA was dried at 60 degrees for 100 min before preparing the PCR
reactions to a volume of 20 μl. Three NTCs and a negative control of the patient’s
buffy coat DNA were included for each dPCR assay. Only paired samples with at
least 4 positive droplets in baseline samples were analysed for change in ctDNA
levels. The circulating DNA ratio at day 15 (CDR15) was assessed as a ratio of the
ctDNA copies/ml at EOT copies/ml compared to ctDNA copies/ml at baseline.
Where more than one mutation was tracked a weighted mean of ctDNA change
was calculated.
The CDR15 cutoff <0.25 for ctDNA suppression, was pre-specified determined
in a separate study in metastatic breast cancer, that validated this cutoff to predict
progression free survival on cytotoxic paclitaxel therapy25.
RNA sequencing. Paired RNAlater™samples were identified and sectioned using a
cryostat. One section was cut for H&E and 10 sections were cut and stained with
nuclease-free Nuclear Fast Red. A second H&E section was cut at the end of the
series. H&E sections were reviewed and marked by a pathologist for tumour and
assessed for tumour content. NFR stained sections were micro-dissected and RNA
was extracted using the Qiagen RNeasy Mini kit according to the manufacturer’s
instructions. RNA was eluted twice into separate 50 μl RNA free water and stored
at -80oC before quantification. RNA was quantified using Qubit 3.0 fluorimeter
using the Qubit™RNA HS Assay Kit (Q32852, ThermoFisher Scientific).
Extracted RNA (~1 μg) was sent to Eurofins Genetic Services Limited for RNA
Exome sequencing. Total RNA was subjected to RiboZero depletion and Illumina
TruSeq RNA Exome library preparation. Libraries were pooled and sequenced on
an Illumina HiSeq 2500 (v4 chemistry) running HiSeq Control Software (HCS)
v2.2.68. Samples were aligned to the GChr37 genome using STAR aligner (https://
github.com/alexdobin/STAR)51 with a mean of 44,117,169 reads per sample (range
14010752-91,470,225). Gene counts were established using htseq (https://github.
com/simon-anders/htseq)52. DeSeq2 (https://doi.org/10.18129/B9.bioc.DESeq2)53
was used to establish gene-wise normalisation and to look for differential
expression between different sample groups. Gene set enrichment analysis was
carried out using the R package fgsea (https://doi.org/10.18129/B9.bioc.fgsea)54.
PAM50 and TNBC subtypes were established using AIMS (https://doi.org/
10.18129/B9.bioc.AIMS)55 and TNBCtype (http://cbc.mc.vanderbilt.edu/tnbc/)56.
Cibersort (https://cibersort.stanford.edu/index.php)29 was run in absolute mode
using normalized gene counts.
Statistics. The study size was determined using a Simon two stage Minimax design
on Ki67 response in patients with sporadic TNBC, with p0 =10% and p1 =25%
Ki67 response rate. With a two-sided alpha 1.6% and 90% power, four Ki67
responders were required in the first 41 assessable patients to proceed to a full 73
patients. The study would declare inefficacy if <4/41 or <14/73 responses were
observed. An initial futility assessment was also planned after 20 evaluable patients
had completed rucaparib treatment and consideration would be given to stopping
the trial if 0 responses were observed. Additionally, 5, 1.6 and 1.6% two-sided
alphas were allocated to assess rucaparib activity within BRCA1 methylated
tumours, RAD51 foci formation tumours, and genomic classifier HRDetect
tumours respectively, for a total study two-sided alpha of 10%. Up to 20 patients
with known BRCA1 or BRCA2 pathogenic germline mutations at the time of trial
entry were recruited as controls for exploratory determination of biomarker end-
points. The study closed to new patients after 43 patients had been recruited on
advice of the IDMC (Fig. 1) due to low recruitment.
Analysis of response data was performed on patients who had taken rucaparib
for 7 days or more. Safety and tolerability are assessed in all patients who received
at least one dose of rucaparib. Response rates and proportions are reported with
95% confidence intervals. Associations between biomarkers are analysed using
Fisher’s exact test or Mann–Whitney as appropriate. Change in biomarkers
between baseline and day12–14 samples are analysed using the Wilcoxon signed
rank test. Analyses were conducted in Stata v13 and GraphPad Prism, with all
analyses reported two-sided.
Reporting summary. Further information on research design is available in
the Nature Research Reporting Summary linked to this article.
Data availability
Sequencing data from whole-genome sequencing, exome RNA seq, and targeted
sequencing from tumour samples that support the findings of this study (Figs. 1–5and
Supplementary Fig. 1, 3 and 4) is deposited in the European Genome-phenome Archive
(EGA), reference EGAS00001004405 https://ega-archive.org.
Code availability
The code used for the analysis of sequencing data is detailed here with links and
references where appropriate.
Whole-genome sequencing: CaVEMan (Cancer Variants Through Expectation
Maximization: http://cancerit.github.io/CaVEMan/), Pindel version 2.0. (http://cancerit.
github.io/cgpPindel/), BRASS (BReakpoint AnalySiS) (https://github.com/cancerit/
BRASS), and ASCAT (v2.1.1). Annotation was to Ensembl build 75. Code for targetted
tumour sequencing is available upon request. AVENIO ctDNA sequencing: AVENIO
Oncology analysis software (v1.0.0 and v1.1) available from Roche. Bisulfite sequencing:
Eukaryotic Promoter Database (http://epd.vital-it.ch/index.php), trim-galore (https://
www.bioinformatics.babraham.ac.uk/projects/trim_galore/), Biostrings R package. RNA
sequencing: STAR aligner (https://github.com/alexdobin/STAR), htseq (https://github.
com/simon-anders/htseq), DeSeq2 (https://doi.org/10.18129/B9.bioc.DESeq2), fgsea
(https://doi.org/10.18129/B9.bioc.fgsea), AIMS (https://doi.org/10.18129/B9.bioc.AIMS),
TNBCtype (http://cbc.mc.vanderbilt.edu/tnbc/) and Cibersort (https://cibersort.stanford.
edu/index.php).
Received: 18 July 2019; Accepted: 3 April 2020;
References
1. Nik-Zainal, S. & Morganella, S. Mutational signatures in breast cancer: the
problem at the DNA level. Clin. Cancer Res. 23, 2617–2629 (2017).
2. Polak, P. et al. A mutational signature reveals alterations underlying deficient
homologous recombination repair in breast cancer. Nat. Genet. 49, 1476–1486
(2017).
3. Turner, N. C. Signatures of DNA-repair deficiencies in breast cancer. The. N.
Engl. J. Med. 377, 2490–2492 (2017).
4. Staaf, J. et al. Whole-genome sequencing of triple-negative breast cancers in a
population-based clinical study. Nat. Med. 25, 1526–1533 (2019).
5. Nik-Zainal, S. et al. Mutational processes molding the genomes of 21 breast
cancers. Cell 149, 979–993 (2012).
6. Alexandrov, L. B. et al. Signatures of mutational processes in human cancer.
Nature 500, 415–421 (2013).
7. Abkevich, V. et al. Patterns of genomic loss of heterozygosity predict
homologous recombination repair defects in epithelial ovarian cancer. Br. J.
Cancer 107, 1776–1782 (2012).
8. Birkbak, N. J. et al. Telomeric allelic imbalance indicates defective DNA
repair and sensitivity to DNA-damaging agents. Cancer Discov. 2, 366–375
(2012).
9. Popova, T. et al. Ploidy and large-scale genomic instability consistently
identify basal-like breast carcinomas with BRCA1/2 inactivation. Cancer Res.
72, 5454–5462 (2012).
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7
10 NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org /10.1038/s41467-020-16142-7 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
10. Telli, M. L. et al. Homologous recombination deficiency (hrd) score
predicts response to platinum-containing neoadjuvant chemotherapy in
patients with triple-negative breast cancer. Clin. Cancer Res. 22,3764–3773
(2016).
11. Davies, H. et al. HRDetect is a predictor of BRCA1 and BRCA2 deficiency
based on mutational signatures. Nat. Med. 23, 517–525 (2017).
12. Nik-Zainal, S. et al. Landscape of somatic mutations in 560 breast cancer
whole-genome sequences. Nature 534,47–54 (2016).
13. Robson, M., Goessl, C. & Domchek, S. Olaparib for metastatic germline
BRCA-mutated breast cancer. N. Engl. J. Med. 377, 1792–1793 (2017).
14. Litton, J. K. et al. Talazoparib in patients with advanced breast cancer and a
germline BRCA Mutation. N. Engl. J. Med 379, 753–763 (2018).
15. Gelmon, K. A. et al. Olaparib in patients with recurrent high-grade serous or
poorly differentiated ovarian carcinoma or triple-negative breast cancer: a
phase 2, multicentre, open-label, non-randomised study. Lancet Oncol. 12,
852–861 (2011).
16. Swisher, E. M. et al. Rucaparib in relapsed, platinum-sensitive high-grade
ovarian carcinoma (ARIEL2 Part 1): an international, multicentre, open-label,
phase 2 trial. Lancet Oncol. 18,75–87 (2017).
17. Wilson, R. H. et al. A phase I study of intravenous and oral rucaparib in
combination with chemotherapy in patients with advanced solid tumours. Br.
J. cancer 116, 884–892 (2017).
18. Vispe, S., Cazaux, C., Lesca, C. & Defais, M. Overexpression of Rad51 protein
stimulates homologous recombination and increases resistance of mammalian
cells to ionizing radiation. Nucleic Acids Res. 26, 2859–2864 (1998).
19. Tashiro, S. et al. S phase specific formation of the human Rad51 protein
nuclear foci in lymphocytes. Oncogene 12, 2165–2170 (1996).
20. Graeser, M. et al. A marker of homologous recombination predicts pathologic
complete response to neoadjuvant chemotherapy in primary breast cancer.
Clin. Cancer Res. 16, 6159–6168 (2010). [pii].
21. Olsson, E. et al. Serial monitoring of circulating tumor DNA in patients with
primary breast cancer for detection of occult metastatic disease. EMBO Mol.
Med. 7, 1034–1047 (2015).
22. Garcia-Murillas, I. et al. Mutation tracking in circulating tumor DNA predicts
relapse in early breast cancer. Sci. Transl. Med. 7, 302ra133, https://doi.org/
10.1126/scitranslmed.aab0021 (2015).
23. Fribbens, C. et al. Tracking evolution of aromatase inhibitor resistance with
circulating tumour DNA analysis in metastatic breast cancer. Ann. Oncol. 29,
145–153 (2018).
24. O’Leary, B. et al. Early circulating tumor DNA dynamics and clonal selection
with palbociclib and fulvestrant for breast cancer. Nat. Commun. 9, 896,
https://doi.org/10.1038/s41467-018-03215-x (2018).
25. Hrebien, S. et al. Early ctDNA dynamics as a surrogate for progression-free
survival in advanced breast cancer in the BEECH trial. Ann. Oncol. 30,
945–952 (2019).
26. Chang, J. et al. Biologic markers as predictors of clinical outcome from
systemic therapy for primary operable breast cancer. J. Clin. Oncol. 17,
3058–3063 (1999).
27. Dowsett, M. et al. Proliferation and apoptosis as markers of benefitin
neoadjuvant endocrine therapy of breast cancer. Clin. Cancer Res. 12,
1024s–1030s (2006).
28. Harper-Wynne, C. L. et al. Comparison of the systemic and intratumoral
effects of tamoxifen and the aromatase inhibitor vorozole in postmenopausal
patients with primary breast cancer. J. Clin. Oncol. 20, 1026–1035 (2002).
29. Newman, A. M. et al. Robust enumeration of cell subsets from tissue
expression profiles. Nat. Methods 12, 453–457 (2015).
30. Telli, M. L. et al. Phase II study of gemcitabine, carboplatin, and iniparib as
neoadjuvant therapy for triple-negative and BRCA1/2 mutation-associated
breast cancer with assessment of a tumor-based measure of genomic
instability: PrECOG 0105. J. Clin. Oncol. 33, 1895–1901 (2015).
31. Isakoff, S. J. et al. TBCRC009: a Multicenter Phase II Clinical Trial of Platinum
Monotherapy With Biomarker Assessment in Metastatic Triple-Negative
Breast Cancer. J. Clin. Oncol. 33, 1902–1909 (2015).
32. Tutt, A. et al. Carboplatin in BRCA1/2-mutated and triple-negative breast
cancer BRCAness subgroups: the TNT Trial. Nat. Med. 24, 628–637
(2018).
33. Naipal, K. A. et al. Functional ex vivo assay to select homologous
recombination-deficient breast tumors for PARP inhibitor treatment. Clin.
Cancer Res. 20, 4816–4826 (2014).
34. Castroviejo-Bermejo, M. et al. A RAD51 assay feasible in routine tumor
samples calls PARP inhibitor response beyond BRCA mutation. EMBO Mol.
Med.https://doi.org/10.15252/emmm.201809172 (2018).
35. Cruz, C. et al. RAD51 foci as a functional biomarker of homologous
recombination repair and PARP inhibitor resistance in germline BRCA-
mutated breast cancer. Ann. Oncol. 29, 1203–1210 (2018).
36. Meijer, T. G. et al. Direct ex vivo observation of homologous recombination
defect reversal after dna-damaging chemotherapy in patients with metastatic
breast cancer. JCO Precis. Oncol. 3, 112, https://doi.org/10.1200/Po.18.00268
(2019).
37. Mutter, R. W. et al. Bi-allelic alterations in DNA repair genes underpin
homologous recombination DNA repair defects in breast cancer. J. Pathol.
242, 165–177 (2017).
38. Farmer, H. et al. Targeting the DNA repair defect in BRCA mutant cells as a
therapeutic strategy. Nature 434, 917–921 (2005).
39. Bryant, H. E. et al. Specific killing of BRCA2-deficient tumours with inhibitors
of poly(ADP-ribose) polymerase. Nature 434, 913–917 (2005).
40. Watkins, J. A., Irshad, S., Grigoriadis, A. & Tutt, A. N. Genomic scars as
biomarkers of homologous recombination deficiency and drug response in
breast and ovarian cancers. Breast Cancer Res. 16, 211 (2014).
41. Li, X. & Heyer, W. D. Homologous recombination in DNA repair and DNA
damage tolerance. Cell Res. 18,99–113 (2008).
42. Pantelidou, C. et al. PARP Inhibitor Efficacy Depends on CD8(+) T-cell
recruitment via intratumoral STING pathway activation in BRCA-deficient
models of triple-negative breast cancer. Cancer Discov. 9, 722–737 (2019).
43. Reislander, T. et al. BRCA2 abrogation triggers innate immune responses
potentiated by treatment with PARP inhibitors. Nat. Commun. 10, 3143
(2019).
44. Chabanon, R. M. et al. PARP inhibition enhances tumor cell-intrinsic
immunity in ERCC1-deficient non-small cell lung cancer. The. J. Clin. Invest.
129, 1211–1228 (2019).
45. Sceneay, J. et al. Interferon signaling is diminished with age and is associated
with immune checkpoint blockade efficacy in triple-negative breast. Cancer
Cancer Discov. 9, 1208–1227 (2019).
46. Leung, S. C. Y. et al. Analytical validation of a standardized scoring protocol
for Ki67: phase 3 of an international multicenter collaboration. NPJ Breast
Cancer 2, 16014, https://doi.org/10.1038/npjbcancer.2016.14 (2016).
47. Davies, H. et al. Whole-genome sequencing reveals breast cancers with
mismatch repair deficiency. Cancer Res. 77, 4755–4762 (2017).
48. Pearson, A. et al. Inactivating NF1 mutations are enriched in advanced breast
cancer and contribute to endocrine therapy resistance. Clin.Cancer Res.
https://doi.org/10.1158/1078-0432.CCR-18-4044 (2019).
49. Magoc, T. & Salzberg, S. L. FLASH: fast length adjustment of short reads to
improve genome assemblies. Bioinformatics 27, 2957–2963 (2011).
50. Pagès, H., Aboyoun, P., Gentleman, R. & DebRoy, S. Biostrings:
efficient manipulation of biological strings. https://rdrr.io/bioc/Biostrings/
(2019).
51. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29,
15–21 (2013).
52. Anders, S., Pyl, P. T. & Huber, W. HTSeq-a Python framework to work
with high-throughput sequencing data. Bioinformatics 31,166–169
(2015).
53. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and
dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550, https://doi.
org/10.1186/s13059-014-0550-8 (2014).
54. Sergushichev, A. An algorithm for fast preranked gene set enrichment analysis
using cumulative statistic calculation. BioRxiv 060012 (2016).
55. Paquet, E. R. & Hallett, M. T. Absolute assignment of breast cancer intrinsic
molecular subtype. J. Natl Cancer Inst. 107, 357, https://doi.org/10.1093/jnci/
dju357 (2015).
56. Chen, X. et al. TNBCtype: A Subtyping Tool for Triple-Negative Breast
Cancer. Cancer Inform.11, 147–156 (2012).
Acknowledgements
This study was funded by the Cridlan Trust, Clovis Oncology Inc., Cancer Research UK
(CRUK; CRUK Ref A15777) core funding to the ICR Clinical Trials and Statistics Unit,
Breast Cancer Now to the Breast Cancer Now Research Centre at The Institute of Cancer
Research, and NHS funding to the NIHR Royal Marsden Biomedical Research Centre.
HRD is funded by a CRUK Grand Challenge Award (C38317/A24043). AD is funded by
a CRUK Pioneer Award. We acknowledge the assistance Katy Jarman, Lynsey Houlton,
Vera Martins, Arjun Naginlal Modi. Sequencing, genomic analyses and personal funding
of SNZ was supported by the Wellcome Trust Intermediate Clinical Fellowship
(WT100183MA), Wellcome Beit Prize, CRUK Advanced Clinician Scientist award
(C60100/A23916) and CRUK Pioneer Award.
Author contributions
N.T. conceived the RIO trial and was the chief investigator. N.T., J.B., C.T., M.D. A.T.
and A.S. formed the RIO protocol development group refining the design. JB oversaw
trial conduct in ICR-CTSU, statistical analysis and data interpretation. N.C. and A.P.
designed and performed translation studies. H.T. and R.C. performed statistical and
bioinformatics analyses. C.T. was RIO study senior trial manager. P.P. and M.H. per-
formed targeted next-generation sequencing experiments. A.D. and F.D. performed
immunohistochemistry experiments. D.K. and H.G. provided pathology support and
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7 ARTICLE
NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org/10.1038 /s41467-020-16142-7 | www.nature.com/naturecommunications 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
analysed immunohistochemistry. H.R.D., A.D. and S.N.Z. performed whole-genome
sequencing and HRDetect analysis. R.R., S.C., A.T., A.S., and A.E. were principal
investigators at trial sites. N.T., N.C. and A.P. wrote the manuscript and all authors
contributed comments and corrections.
Competing interests
N.T., J.M.B., H.T. and C.T. report research grants and non-financial support in the form
of study drug provision and distribution from Clovis Oncology Inc., during the conduct
of the study. N.T. reports advisory board Honoraria from AstraZeneca, Bristol-Myers
Squibb, Lilly, Merck Sharpe and Dohme, Novartis, Pfizer, Roche/Genentech, Tesaro,
Bicycle Therapeutics and research funding from Astra Zeneca, BioRad, Pfizer, Roche/
Genentech and Guardant Health, outside the submitted work. J.M.B. reports grants and
non-financial support from AstraZeneca, Merck Sharpe & Dohme, Puma Biotechnology
and Janssen-Cilag, grants, non-financial support and travel support from Pfizer and
grants from Medivation, Novartis and Roche outside the submitted work. M.H. reports
Honoraria or research funding from Boehringer Ingelheim, Roche Diagnostics, Bristol
Myers Squibb, Guardant Health, Celgene, Eli Lilley outside of the submitted work. R.R.
reports Honorarium from Pfizer. A.T. reports benefits from ICR’s Inventors Scheme
associated with patents for one of the PARP inhibitors in BRCA1/2 associated cancers.
A.T. also reports Honoraria from Pfizer, Vertex, Prime Oncology, Artios, honoraria and
stock in InBiomotion, honoraria and financial support from AstraZeneca, Medivation,
Myriad Genetics, Merck Serono. S.N.Z. and H.D. are inventors on a patent application
(WO2017191074A1) for HRDetect. S.N.Z. reports advisory board honoraria from Astra
Zeneca and Artios Pharma. M.D. reports advisory board Honoraria from GTx, Radius,
Orion and G1therapeutics, lectures fees from Myriad and Nanostring, research funding
from Pfizer and Radius and income from the Institute of Cancer Research’s Rewards for
Inventors Scheme (Abiraterone) outside the submitted work. All other authors declare no
competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
020-16142-7.
Correspondence and requests for materials should be addressed to N.C.T.
Peer review information Nature Communications thanks the anonymous reviewers for
their contribution to the peer review of this work
Reprints and permission information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2020
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16142-7
12 NATURE COMMUNICATIONS | (2020) 11:2662 | https://doi.org /10.1038/s41467-020-16142-7 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Available via license: CC BY 4.0
Content may be subject to copyright.
