Abstract 4524: The effect of acute intraoperative hypoxia in breast cancer

Abstract

Background: Hypoxia is defined as oxygen levels in tumor microenvironment of less than that in blood (90-100 mm Hg) and influences many aspects of tumour biology. During surgery tumour vasculature is cut off gradually leading to induction of acute hypoxia.The present study aims to experimentally test the genotypic and phenotypic effects of surgically induced acute hypoxia in breast cancer tumor samples and cell lines.
Methodology: Core biopsy samples were collected from breast tumors (N=8 patients) at three time points during their curative surgery: prior (pre), mid-way (intra) and at the end (post). The samples were subjected to RNA-Seq and a list of differentially expressed genes (DEG) was prepared. A set of 26 DEG (‘pre’ Vs ‘intra’ Vs ‘post’) obtained from RNA-Seq analysis and additional 17 genes involved in inflammation, EMT and hypoxia pathways were chosen for validation in tumor samples. These genes were validated using a customized qPCR Array (Qiagen). A gene was considered validated if it was significantly deregulated in at least 4 out of 8 patients. In another experiment, MCF-7 cells were exposed to varying levels of oxygen concentrations (0.1-20%) for varying time periods ranging from 30 minutes to 72 hours, to study time and dose dependent effects of hypoxia on following functional characteristics: proliferation, invasion and cell cycle changes.
Results: Concordant, statistically significant up-regulation of FOS, DUSP1, JUNB, FOSB, ZFP36, RGS1, S100A4, CXCL8 and CCL2 were observed in RNA-Seq and qPCR experiments while MMP13, HIF1A and VEGFA were up-regulated only in qPCR. However, 7 protein markers of inflammation, EMT and hypoxia did not show any significant change between pre, intra and post-operative samples. In MCF-7 cells, a dose and time dependent decrease in cell viability was observed with increasing severity of hypoxia as well as decrease in invasiveness, but there was no significant impact on cell cycle phases. When hypoxic cells were re-incubated under normoxic conditions an increase in cell proliferation and accumulation of cells in S phase (with a reduction in G2-M fraction) were observed, compared to cells grown under only normoxic conditions.
Conclusion: Acute intra-operative hypoxia up-regulates expression of genes related to cell survival, chemoresistance, invasiveness, inflammation and angiogenesis in breast tumors. Breast cancer cells exposed to acute severe hypoxia followed by normoxia show increased proliferation. These effects may have implications for tumor cells that disseminate during surgery.
Citation Format: Dimple R. Bhatia, Shalaka Joshi, Rohan Chaubal, Poonam Gera, Prajakta Kalkar, Farheen Naeem, Nisanth Mathew Raju, Nilesh Gardi, Nita Nair, Vaibhav Vanmali, Rohini W. Hawaldar, Amit Dutt, Rajendra A. Badwe, Sudeep Gupta. The effect of acute intraoperative hypoxia in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4524. doi:10.1158/1538-7445.AM2017-4524

Full text

Vol.:(0123456789)
1 3
Breast Cancer Research and Treatment
https://doi.org/10.1007/s10549-018-4749-3
BRIEF REPORT
Pre‑operative progesterone benets operable breast cancer patients
bymodulating surgical stress
ShataksheeChatterjee1· RohanChaubal2· ArindamMaitra1· NileshGardi2· AmitDutt2· SudeepGupta2,3·
RajendraA.Badwe2,4· ParthaP.Majumder1· PriyankaPandey1
Received: 12 January 2018 / Accepted: 8 March 2018
© Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
Purpose We have reported a survival benefit of single injection of hydroxyprogesterone prior to surgery for primary tumour
in patients with node-positive operable breast cancer. Hydroxyprogesterone was meant to recapitulate the luteal phase of
menstrual cycle in these women. We wanted to understand the molecular basis of action of hydroxyprogesterone on primary
breast tumours in a peri-operative setting.
Methods We performed whole transcriptome sequencing (RNA-Seq) of primary breast tumour samples collected from
patients before and after hydroxyprogesterone exposure and controls. Paired breast cancer samples were obtained from
patients who were given hydroxyprogesterone before surgery and a group of patients who were subjected to only surgery.
Results A test of significance between the two groups revealed 207 significantly altered genes, after correction for multi-
ple hypothesis testing. We found significantly contrasting gene expression patterns in exposed versus unexposed groups;
142 genes were up-regulated post-surgery among exposed patients, and down-regulated post-surgery among unexposed
patients. Significantly enriched pathways included genes that respond to progesterone, cellular stress, nonsense-mediated
decay of proteins and negative regulation of inflammatory response. These results suggest that cellular stress is modulated
by hydroxyprogesterone. Network analysis revealed that UBC, a mediator of stress response, to be a major node to which
many of the significantly altered genes connect.
Conclusions Our study suggests that pre-operative exposure to progesterone favourably modulates the effect of surgical
stress, and this might underlie its beneficial effect when administered prior to surgery.
Keywords Operable breast cancer· Hydroxyprogesterone· RNA-Seq· Network analysis
Introduction
Breast cancer is the most commonly diagnosed cancer in
women and is also a major cause of cancer-related deaths
worldwide [1]. Surgical resection of primary tumour is the
gold standard of treatment in early-stage or operable breast
cancer patients [2]. Women in whom surgical resection was
performed in their luteal phase of menstrual cycle showed
better prognosis than those on whom surgery was performed
in their follicular phase [3]. Therefore it was hypothesised
that the observed difference in prognosis may be due to
the effects of progesterone, the dominant female steroidal
hormone during luteal phase, which is also known to cause
Electronic supplementary material The online version of this
article (https ://doi.org/10.1007/s1054 9-018-4749-3) contains
supplementary material, which is available to authorized users.
* Sudeep Gupta
sudeep.gupta@actrec.gov.in
* Rajendra A. Badwe
badwera@tmc.gov.in
* Partha P. Majumder
ppm1@nibmg.ac.in
* Priyanka Pandey
pp1@nibmg.ac.in
1 National Institute ofBiomedical Genomics, P.O.: N.S.S.,
Kalyani741251, WestBengal, India
2 Tata Memorial Centre/Hospital, Parel, Mumbai400012,
India
3 Department ofMedical Oncology, Tata Memorial Centre,
Mumbai400012, India
4 Department ofSurgical Oncology, Tata Memorial Centre,
Mumbai400012, India

Breast Cancer Research and Treatment
1 3
terminal differentiation of mammary ducts and facilitate
breast development during puberty [4].
To test this hypothesis, Badwe etal. [5] conducted a ran-
domised clinical trial wherein they reported that a single
dose of 500mg depot hydroxyprogesterone (progesterone
equivalent; HP) injected within two weeks before surgery
increased disease-free and overall survival in women with
axillary lymph node-positive breast cancer. We undertook
this study to identify the biological correlates of this major
finding of the clinical trial and thereby gain an understand-
ing of the underlying mechanisms. We sought this under-
standing by studying the impact of HP exposure on levels of
expression of genes in breast tumours of patients undergoing
surgery after accounting for the effect of surgical procedure
on gene expression.
Materials andmethods
Recruitment ofpatients andsample collection
Thirty-one women diagnosed with operable breast cancer
were recruited in this study. Eighteen patients were admin-
istered a single dose of 500mg of HP within 15days prior
to surgery (Exposed Group), while 13 patients were not sub-
jected to this intervention (Unexposed Group). From each
patient, multiple core biopsies were collected immediately
before administration of HP (for patients in exposed group)
or at the start of surgery (for patients in unexposed group)
as pre-surgery samples. Multiple cores obtained from the
fully resected tumour constituted the post-surgery sample
in Exposed and Unexposed groups.
Total RNA extraction
Tissue samples were stored in RNA later immediately
after collection. Trizol® (Invitrogen)-based total RNA
was extracted and RNA purification was performed using
PureLink® RNA Mini Kit (Invitrogen). Total RNA integ-
rity was checked with Bioanalyzer 2100 (Agilent) using
Total RNA Nano Kit (Agilent) followed by ribosomal RNA
(rRNA) depletion using Ribo-Zero rRNA Removal Magnetic
Kit (Human/Mouse/Rat, Epicentre, Illumina).
RNA sequencing
Libraries for RNA Sequencing were prepared following
manufacturer’s instructions (Illumina True-Seq RNA Sam-
ple Preparation Kit). Briefly, the steps included random frag-
mentation of RNA, conversion to double stranded cDNA,
repair of the overhang ends resulting from fragmentation
into blunt ends and ligation of adapters. Libraries were
amplified using PCR and fragment sizes were validated
with Bioanalyzer (Agilent) using High Sensitivity Kit.
KAPPA SyBr Green Assays for qPCR (ABI 7900HT) were
used for library quantification. Paired-end sequencing on
HiSeq2000 (Illumina) platform with read length as 100bp
was performed. The quality of the raw sequence reads was
determined using FastQC.
The TUXEDO pipeline [6] was used to analyse the RNA-
Seq data. The steps included (1) alignment of the reads to
the transcriptome and human reference genome (hg19) with
TopHat [7] v2.0.8b; (2) assembly and classification of the
transcripts with Cufflinks v2.0.2 [8] and Cuffcompare [8],
respectively; (3) merging of the transcripts belonging to
paired samples collected pre- and post-exposure to hydroxy-
progesterone with Cuffmerge (Cufflinks v2.0.2); and, (4) cal-
culation of differential gene expression with Cuffdiff (Cuf-
flinks v2.0.2). The transcript abundance was calculated by
estimating the fragments per kilobase of exon per million
mapped fragments (FPKM) [8]. The transcripts were binned
according to their abundance, and the overall relative abun-
dances of the transcripts that were expressed in all breast
cancer samples on each chromosome were compared. Align-
ment metrics were generated using picard-tools-1.92 (http://
broad insti tute.githu b.io/picar d). The correlation between the
samples was determined using unsupervised hierarchical
clustering and Spearman’s correlation.
Identifying “signicantly modulated genes” due
tohydroxyprogesterone exposure
As mentioned earlier, from each of 31 patients, two sam-
ples of tumour tissue were collected: one pre-surgery (for
women in the HP-Exposed Group this sample was collected
immediately prior to HP administration; for women in the
Unexposed Group this sample was collected at the start of
surgery) and another post-surgery (i.e. immediately prior
to the end of the surgical procedure). We sought to iden-
tify changes in the levels of expression of genes attribut-
able to HP exposure, after accounting for changes due to
surgery. For this, we only considered such genes that were
substantially expressed (FPKM ≥ 1) in both tissues collected
post- and pre-surgery. Among such genes, we only consid-
ered a gene as a “candidate modulator” of response if the
direction of fold change (Post/Pre) of expression level was
the same in at least 50% (i.e. 9 out of 18) of the women in
the HP-exposed group. Although somewhat arbitrary, this
was a conservative approach to minimise the play of chance
due to random fluctuations of gene expression in a limited
sample size. We then computed the mean fold change over
all such patients for whom the direction of fold change was
concordant (up or down). For further analysis, we retained
only those genes that showed at least a doubling or halving
(i.e. ≥ 2 or ≤ 0.5) of mean fold change of expression levels
post-surgery compared to pre-surgery. We designate this

Breast Cancer Research and Treatment
1 3
subset of genes as {GE}. It is obvious that this subset of
genes {GE} are those whose expression levels are highly
modulated in the Exposed Group by the combined effect of
HP exposure and the act of surgery. To identify only those
genes whose expression levels are significantly modulated
by HP exposure alone, we first computed the post:pre fold
change of expression level for each gene in {GE}, in each
patient in the Unexposed Group, after removing those genes
for which the FPKM value was < 1 in both pre- and post-
surgery tumour samples. We then tested whether the mean
fold change of expression of each such gene was statistically
significantly different between patients belonging to the HP-
exposed and unexposed groups. Multiple-testing correction
was done by the Benjamini–Hochberg [9] procedure. Genes
that turned out to be significantly different were those that
were significantly modulated by HP exposure alone, and
not by the combined effect of HP exposure and the act of
surgery.
Functional annotation ofgenes
Functional annotation enrichment was performed at p
value < 0.05 for significantly altered genes using Gene
Ontology terms for biological processes in Enrichr [10, 11].
To assess how the genes encompassing significantly changed
biological processes may be represented in protein–protein
interaction networks, we performed network analysis using
NetworkAnalyst [12]. In addition, to identify progesterone-
modulated genes, the significantly altered genes were com-
pared with gene sets curated in Molecular Signatures Data-
base (MSigDb) using the search keyword “progesterone”.
Validation ofexpression using qPCR
A subset of more significantly altered and biologically more
relevant genes identified by functional annotation analysis
was chosen for validation of their expression using custom
qRT-PCR arrays based on SYBR Green chemistry (SABio-
sciences, Qiagen). Six samples from each of exposed and
unexposed groups were randomly chosen for validation.
[Many of the genes were non-coding RNA genes; hence, val-
idation using qRT-PCR was not possible. Further, budgetary
constraints did not permit validation of all the remaining
significantly altered genes or using all samples.] Reactions
were performed in duplicate with TBP as the endogenous
control [13, 14]. For each pre- and post-surgery sample pair,
fold change was computed using 2−∆∆CT [15]. We identified
genes with highly concordant properties: same direction of
change (over- or under-expression) in at least 50% of sam-
ples, and altered expression levels in both exposed and unex-
posed groups assessed by both qPCR and RNA sequencing.
Data availability
The datasets generated during and/or analysed during the
current study are not publicly available due to an ongoing
analysis related to this study, but are available from the
corresponding author on reasonable request with adequate
justification.
Results
Tumour transcriptome prole obtained fromRNA
sequencing
Our study included samples from the three major breast can-
cer subtypes (Table1). The average number of raw reads
obtained on RNA sequencing per sample was 114 ± 3.5
million (Supplementary Fig. S1, Supplementary TableS1).
The mean percent of reads that aligned to the human ref-
erence transcriptome (33.04 ± 1.1%) and genome (hg19;
65.35 ± 1.4%) was 86 ± 2% (Supplementary Fig. S1a). Ribo-
somal RNA depletion method adopted in our study enabled
capture of non-coding RNA along with mRNA (Supplemen-
tary Fig. S1b). The vast majority (> 70%) of assembled tran-
scripts were highly (FPKM ≥ 1) abundant (Supplementary
Fig. S1d).
Signicantly modulated genes due
tohydroxyprogesterone exposure beforesurgery
After applying the requirements of FPKM, mean fold change
and concordant deregulation in at least 50% patients, we
Table 1 Number of breast
cancer patients of various
known subtypes included
from two subgroups, with
and without exposure to
hydroxyprogesterone before
surgery
*On each patient, a sample was collected from the tumour tissue before surgery and during surgery
Exposure to Hydroxypro-
gesterone Before Surgery
No. of
patients*
Age in years
(mean ± s.e.)
Sample sizes for Clinical subtypes
ER + PR ± HER2− ER ± PR ± HER2+ ER−
PR−
HER2−
Exposed 18 58.6 ± 2.8 6 6 6
Unexposed 13 58.5 ± 9.7 4 6 (ER+/HER2+)
2 (ER−/HER2+)
1

Breast Cancer Research and Treatment
1 3
found 375 ({GE}) genes to be altered in the exposed group
that were also differentially expressed between post-surgery
and pre-surgery samples in the Unexposed (surgery only)
group (Supplementary TableS2).
Following the methodology described earlier, we identi-
fied 207 genes whose expression levels were significantly
altered as a result of exposure to hydroxyprogesterone alone
(Supplementary Fig. S2, Supplementary TableS3). Of these
207 genes, 32 were highly overexpressed (fold change > 2)
in a large percentage (≥ 66%) of exposed patients (Supple-
mentary TableS3).
Identifying biological implications due
tohydroxyprogesterone exposure
Significantly (p < 0.05) enriched Gene Ontology Biological
Processes for these 207 genes were identified. Additionally,
the interactions among those genes that were found altered
in major biological processes were identified using network
analysis (Fig.1; Supplementary Fig. S3). UBC (ubiquitin
C) was identified as the central “hub” gene in the network
to which the other major genes that are significantly altered
are connected (Fig.1). Five major biological processes were
significantly enriched (Fig.1 and Table2) of which the most
relevant process was the one involved in response to: proges-
terone, steroid hormone, drug, regulation of hormone levels.
Of the 207 genes that were significantly altered, 18 genes
belonged to this interaction network (Fig.1, Supplementary
Fig. S3, Supplementary TableS3). In addition, 27 genes
were reported as being perturbed by progesterone (Supple-
mentary TableS4) in gene sets of MSigDb. Thus, 45 (22%)
of 207 genes that showed significant alterations in expres-
sion levels upon hydroxyprogesterone exposure are known to
be associated with response to progesterone. The remaining
four biological processes that were significantly enriched
were (a) negative regulation of tumour necrosis factor
Fig. 1 Summary of major significantly enriched biological processes.
Select modules are expanded to show internal connections of mem-
bers that are likely to perform together as a biological process; a
response to progesterone hormonal stimulus; b cellular response to
stress; c miscellaneous processes including gene expression regula-
tion; d protein targeting to endoplasmic reticulum; e negative regu-
lation of inflammatory processes; genes shown in the network were
found to be up- (shown in red) or down- (shown in green) regulated
post-surgery in the hydroxyprogesterone exposed group

Breast Cancer Research and Treatment
1 3
Table 2 Significantly (p < 0.05) enriched GO BP terms obtained from Enrichr analysis for 207 genes
Category GO BP terms Genes
Response to hormone Response to testosterone, steroid hormone, progesterone PLN, TSPO, CLDN4, ANXA1, CAV1, AKR1C3, CRYAB, AQP1
Negative regulation Negative regulation of tumour necrosis factor superfamily cytokine production,
tumour necrosis factor production, nitric oxide biosynthetic process, gliogen-
esis, endothelial cell proliferation, cytokine-mediated signalling pathway, acute
inflammatory response
GSTP1, TSPO, CAV1, ID4, CAV2, ITGA5, ANXA1
Cellular response Cellular response to starvation, reactive oxygen species, radiation, oxygen levels,
oxidative stress, osmotic stress, lipid, hypoxia, hydrogen peroxide, glucocorti-
coid stimulus, fatty acid, decreased oxygen levels, corticosteroid stimulus, acid
chemical
CAV1, AKR1C3, HIGD1A, ANXA1, AQP1, RHOB, HIST3H2A, CRYAB, NDRG1,
RBX1, MGST1, TSPO, DGAT2, GSTP1, PLSCR4, LAMTOR2, S100A10
Protein targeting Protein targeting to membrane, organelle, membrane, endoplasmic reticulum,
protein complex disassembly
RPL30, RPL10, RPL32, RPS29, RPL31, RPS5, RPL34, RPL24, RPS21, TIMM17B,
TSPO, IPO7, HDGF, HIST3H2A
Metabolic process Acylglycerol metabolic process, acylglycerol biosynthetic process, acyl-CoA meta-
bolic process, bicarbonate transport, diterpenoid metabolic process, cyclic purine
nucleotide metabolic process, cyclic nucleotide biosynthetic process, glycer-
ophospholipid biosynthetic process, glycerolipid metabolic process, glycerolipid
biosynthetic process, glutathione derivative metabolic process, glutathione
derivative biosynthetic process, mitochondrial electron transport, NADH to
ubiquinone, long-chain fatty-acyl-CoA metabolic process, lipid storage, regula-
tion of lipid metabolic process, regulation of lipid biosynthetic process
DGAT2, GPAM, ACSL1, CAV1, PNPLA2, SLC4A2, AQP1, CRABP2, RDH10,
AKR1C3, CALCRL, DPM3, CHPT1, GSTP1, MGST1, NDUFA7, NDUFB7,
NDUFB10, VAV3, ANXA1, GPR116, TSPO, SORBS1
Other Nuclear-transcribed mRNA catabolic process, nonsense-mediated decay, gene
expression, nucleosome assembly, protein-DNA complex assembly
RPL30, RPL10, RPL32, RPS29, RPL31, RPS5, RPL34, RPL24, RPS21, TAF1D,
EIF2S3, PTRF, HIST1H2BM, HIST1H3A, HIST1H2BL, HIST1H2BC

Breast Cancer Research and Treatment
1 3
production, nitric oxide biosynthetic process, inflammatory
response, (b) cellular response to osmotic stress; oxidative
stress; external stimulus, radiation; zinc ion; reactive oxygen
species, hypoxia, (c) protein targeting to endoplasmic retic-
ulum; mitochondria, protein oligomerization and (d) gene
expression, nonsense-mediated decay, nucleosome assembly,
protein-DNA complex assembly (Fig.1 and Table2).
Genes deregulated viacanonical & non‑canonical
progesterone pathways
We wanted to investigate whether (and how many of) the
207 genes identified as significantly differentially expressed
as a result of HP exposure were deregulated via the canoni-
cal progesterone signalling pathway. We compared these 207
genes with the results of published ChIP-seq datasets using
ChIP-Atlas (http://chip-atlas .org). Using a cut-off of 10kb
peak-call intervals for PgR targets [16], 106 of the 207 genes
were identified as direct target genes of progesterone recep-
tor (Supplementary TableS3). Some of these 106 genes
were also validated using qPCR. We next investigated the
differential genes between progesterone exposed and unex-
posed patients stratified by the expression of progesterone
receptor (PgR) on the primary tumour. Using the same meth-
odology and cut-offs described under ‘Methods’, 450 genes
were found to be significantly altered between PgR-positive
HP-exposed (n = 7) and PgR-positive HP-unexposed (n = 9)
patients (Supplementary TableS5). 67 of the 106 (63.2%)
ChIP-seq overlap genes were among these 450 genes. In the
comparison between PgR-negative HP-exposed (n = 11) and
PgR-negative HP-unexposed groups (n = 4), 24 genes were
identified as differentially altered.
The large difference in the number of significantly dif-
ferentially altered genes in the 2 subsets of patients—450
genes in PgR-positive and 24 genes in PgR-negative samples
could possibly be attributed to vagaries of small sample size.
We also speculate that non-canonical pathways may mediate
the effect of HP exposure in PgR-negative tumours using
fewer numbers of genes. Further, 49 pathways enriched by
the 24 genes altered among PgR-negative samples are com-
mon with those enriched by the 450 genes altered among
PgR-positive samples (Supplementary TableS5; detailed
results not presented).
Validation oftheresults obtained fromanalysis
ofRNA sequencing data
We have validated, using quantitative real-time PCR (qRT-
PCR) experiments, some of the major results obtained from
analyses of RNA sequencing data. Many of the significantly
modulated genes were non-coding RNA genes and hence
were not appropriate for qRT-PCR validation. We selected
for validation a subset of 14 genes (CD69, CLDN4, DPM3,
HIST1H2AG, HIST1H2BC, HIST1H4E, ID1, MLF1, MPP6,
PLN, RPL10, RPS5, TMCC1, TSPO) from among those that
were found significantly enriched in multiple biological
processes (such as response to progesterone, abiotic stimu-
lus and negative regulation of processes implicated in cell
survival, Fig.1). Concordant trend of levels of expression
between RNA-seq and qRT-PCR assays was obtained for
all except four genes (Fig.2, Supplementary TableS6), of
which seven (CLDN4, DPM3, TMCC1, HIST1H2BC, PLN,
RPS5, TSPO) are reported targets of PgR (as obtained from
ChIP-seq overlap genes).
Discussion
We had reported [5] that a single dose of depot hydroxypro-
gesterone administered before surgery to women with opera-
ble breast cancer led to a significantly improved disease-free
and overall survival in node-positive breast cancer patients.
However, the biological mechanisms underlying this obser-
vation have remained unclear. We posited that the beneficial
effect of hydroxyprogesterone observed in that trial, might
result from altered expression of some genes and biological
pathways governed by these genes. To test this, we have per-
formed a whole transcriptome study using RNA Sequencing.
We identified, using stringent criteria to reduce the chance
of false-positive inferences, that alterations in the expression
of 207 genes are associated with exposure to HP. Further,
we identified a number of pathways that could mediate the
downstream biological effect of these deregulated genes.
Expectedly, genes involved in response to hormone
were identified as significantly differentially expressed in
response to hydroxyprogesterone. This response was possi-
bly effected by up-regulation of TSPO, PLN and CLDN4; a
result that was also validated by qPCR analysis (Figs.1a, 2).
An overwhelming perturbation of lipid metabolic processes
(Table2) post-surgery in tumours of HP-exposed group, also
suggest a possible progesterone action [17]. In the group
Fig. 2 qPCR validation of significantly differentially modulated genes
in exposed and unexposed groups obtained from RNA-Seq. Ten
genes showed concordance to RNA-Seq results

Breast Cancer Research and Treatment
1 3
exposed to HP, genes related to response to oxidative stress,
osmotic stress, radiation and starvation (e.g. TSPO, Fig.1b,
Table2) were found to be up-regulated after surgery. This
result suggests that progesterone may be regulating response
to stress caused by surgery. Nonsense-mediated decay
(NMD) response, known to degrade unwanted proteins and
reduce cellular stress responses such as those to hypoxia
and nutrient starvation [17], comprised genes that were up-
regulated in the exposed but not unexposed group (Fig.1c).
Genes up-regulated by HP—but not by surgery alone—e.g.
TSPO and RPS5, are also known to participate in protein tar-
geting to Endoplasmic Reticulum (ER) (Figs.1d, 2). Thus,
downregulation of surgical stress response could potentially
underlie the biological effect of progesterone in the peri-
operative setting.
Negative regulation of inflammatory response and
tumour necrosis factor (TNF) production (Table2; Fig.1e)
was observed to be amongst the top GO BP terms. Chronic
inflammation and severity of inflammation have been identi-
fied as a risk factor for many cancers including breast cancer
[18–21] and as an aid to invasion and metastasis in cancer
[22, 23]. Peri-operative inflammation has been suggested as
a trigger for metastasis [24]. Further, TNF (an inflammatory
cytokine) via NF-kB axis is known to induce proliferative,
invasive and malignant behaviour in various cancers includ-
ing breast cancer [25–29]. Thus, negative regulation of
inflammation resulting from progesterone exposure may play
an important role in mediating the pro-survival effects of this
agent. Up-regulated genes in the exposed group were also
identified to participate in electron transport in our study
(Table2). Interestingly, progesterone treatment has been
shown to increase the activity of the electron transportand
decrease oxidative stress and damage in tissues [30–34].
UBC, predicted as the central node (Fig.1, Supplemen-
tary Fig. S3), also suggests a stress-responsive behaviour as
the latter has been reported to be induced in stressed con-
ditions in mammals [35]. Significantly altered genes that
form connected nodes to UBC (Fig.1, Supplementary Fig.
S3) suggest downstream stress-regulating activities. The lat-
ter could be a potential mechanism of progesterone action
independent of its receptor. Thus, altogether, these results
suggest that pre-operative hydroxyprogesterone action might
possibly be to curb cellular stress response.
Hydroxyprogesterone also modulated some genes
involved in processes other than stress. We observed up-
regulation of MLF1 (Supplementary TableS3, Fig.2) after
hydroxyprogesterone administration, but not surgery, sug-
gesting cell cycle arrest due to possible regulation of mitosis
at G1 to S phase [36].
Our study, motivated by a clinical trial result, has pro-
vided molecular genetic insights that may have important
clinical implications. Despite the restricted sample size
and possible tumour heterogeneity, it is unlikely that the
molecular genetic observations are false, since we have
used stringent criteria in our discovery effort. We have
provided evidence that pre-operative exposure to proges-
terone favourably modulates the effect of surgical stress.
This, and other interventions with similar effects, are wor-
thy of continued investigation in the peri-operative period.
Acknowledgements Authors would like to acknowledge Ms. Sriparna
Biswas, Dr. Subrata Patra and Mr. Sumanta Sarkar for their help in
performing the RNA-Seq laboratory experiments. We would like to
thank Prof. Bidyut Roy, Mr. Badal De and Ms. Anindita Ray for help-
ing in qPCR experiments. We would also like to thank Drs Nita Nair,
Shalaka Joshi, Vani Parmar, Rohini Hawaldar and Vaibhav Vanmali
for sample acquisition.
Author contributions PP, PPM, RAB and SG conceived, designed and
executed the study. RAB, RC, NG, AD and SG provided clinical sam-
ples for the study. SC and AM generated RNA-Seq data. SC, PP and
PPM performed RNA-Seq data analysis. SC, RC, NG, AM, AD, PP,
SG and PPM interpreted the results. SC, RC, NG, AD, SG, RAB, PPM
and PP participated in writing of manuscript.
Funding Authors acknowledge funding support from Department of
Atomic Energy and Tata Memorial Centre for conducting the study.
SC would like to acknowledge Indian Council for Medical Research
for her doctoral research fellowship. RC is supported by a research fel-
lowship from the Homi Bhabha National Institute (HBNI), Tata Memo-
rial Centre. NG is supported by Prime Minister’s Fellowship Scheme
for Doctoral Research, a public–private partnership between Science
and Engineering Research Board (SERB), Department of Science and
Technology, Government of India and Confederation of Indian Industry
(CII). PPM was supported by the J. C. Bose Fellowship of the Govern-
ment of India, Department of Science and Technology.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
Ethical standards This study was approved by the Institutional Ethics
Committees of the Tata Memorial Centre, Mumbai, and the National
Institute of Biomedical Genomics, Kalyani. Biospecimens were col-
lected from breast cancer patients with written informed consent.
References
1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo
M, Parkin DM, Forman D, Bray F (2015) Cancer incidence and
mortality worldwide: sources, methods and major patterns in
GLOBOCAN 2012. Int J Cancer 136(5):E359–E386. https ://
doi.org/10.1002/ijc.29210
2. Barry PA, Schiavon G (2015) Primary systemic treatment in the
management of operable breast cancer: best surgical approach
for diagnosis, biological evaluation, and research. J Natl Cancer
Inst Monogr 51:4–8. https ://doi.org/10.1093/jncim onogr aphs/
lgv00 8
3. Lemon HM, Rodriguez-Sierra JF (1996) Timing of breast cancer
surgery during the luteal menstrual phase may improve progno-
sis. Nebraska Med J 81(3):73–78

Breast Cancer Research and Treatment
1 3
4. Graham JD, Clarke CL (1997) Physiological action of proges-
terone in target tissues. Endocr Rev 18(4):502–519. https ://doi.
org/10.1210/edrv.18.4.0308
5. Badwe R, Hawaldar R, Parmar V, Nadkarni M, Shet T, Desai
S, Gupta S, Jalali R, Vanmali V, Dikshit R, Mittra I (2011) Sin-
gle-injection depot progesterone before surgery and survival in
women with operable breast cancer: a randomized controlled
trial. J Clin Oncol 29(21):2845–2851. https ://doi.org/10.1200/
jco.2010.33.0738
6. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR,
Pimentel H, Salzberg SL, Rinn JL, Pachter L (2012) Differential
gene and transcript expression analysis of RNA-seq experiments
with TopHat and Cufflinks. Nat Protoc 7(3):562–578. https ://doi.
org/10.1038/nprot .2012.016
7. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg
SL (2013) TopHat2: accurate alignment of transcriptomes in the
presence of insertions, deletions and gene fusions. Genome Biol
14(4):R36. https ://doi.org/10.1186/gb-2013-14-4-r36
8. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van
Baren MJ, Salzberg SL, Wold BJ, Pachter L (2010) Transcript
assembly and quantification by RNA-Seq reveals unannotated
transcripts and isoform switching during cell differentiation. Nat
Biotechnol 28(5):511–515. https ://doi.org/10.1038/nbt.1621
9. Benjamini Y (1995) Controlling the false discovery rate: a prac-
tical and powerful approach to multiple testing. J Roy Stat Soc
57(1):289–300
10. Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan
Q, Wang Z, Koplev S, Jenkins SL, Jagodnik KM, Lachmann
A, McDermott MG, Monteiro CD, Gundersen GW, Ma’ayan A
(2016) Enrichr: a comprehensive gene set enrichment analysis
web server 2016 update. Nucleic Acids Res 44(W1):W90–W97.
https ://doi.org/10.1093/nar/gkw37 7
11. Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV, Clark
NR, Ma’ayan A (2013) Enrichr: interactive and collaborative
HTML5 gene list enrichment analysis tool. BMC Bioinfo 14:128.
https ://doi.org/10.1186/1471-2105-14-128
12. Xia J, Gill EE, Hancock RE (2015) Network analyst for statistical,
visual and network-based meta-analysis of gene expression data.
Nat Protoc 10(6):823–844. https ://doi.org/10.1038/nprot .2015.052
13. Caradec J, Sirab N, Keumeugni C, Moutereau S, Chimingqi M,
Matar C, Revaud D, Bah M, Manivet P, Conti M, Loric S (2010)
‘Desperate house genes’: the dramatic example of hypoxia. Br J
Cancer 102(6):1037–1043. https ://doi.org/10.1038/sj.bjc.66055 73
14. Glenisson M, Vacher S, Callens C, Susini A, Cizeron-Clairac G,
Le Scodan R, Meseure D, Lerebours F, Spyratos F, Lidereau R,
Bieche I (2012) Identification of new candidate therapeutic target
genes in triple-negative breast cancer. Genes Cancer 3(1):63–70.
https ://doi.org/10.1177/19476 01912 44983 2
15. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expres-
sion data using real-time quantitative PCR and the 2(-delta delta
c(t)) method. Methods 25(4):402–408. https ://doi.org/10.1006/
meth.2001.1262
16. Rubel CA, Lanz RB, Kommagani R, Franco HL, Lydon JP,
DeMayo FJ (2012) Research resource: genome-wide profiling of
progesterone receptor binding in the mouse uterus. Mol Endo-
crinol 26(8):1428–1442. https ://doi.org/10.1210/me.2011-1355
17. Kalkhoff RK (1982) Metabolic effects of progesterone. Am J
Obstet Gynecol 142(6 Pt 2):735–738
18. Yang YC, Johnson MP, Schorpp KM, Boen CE, Harris KM (2017)
Young adult risk factors for cancer: obesity, inflammation, and
sociobehavioral mechanisms. Am J Prev Med 53(3S1):S21–S29.
https ://doi.org/10.1016/j.amepr e.2017.04.025
19. Vona-Davis L, Rose DP (2013) The obesity-inflammation-eicos-
anoid axis in breast cancer. J Mammary Gland Biol Neoplasia
18(3–4):291–307. https ://doi.org/10.1007/s1091 1-013-9299-z
20. Sun X, Glynn DJ, Hodson LJ, Huo C, Britt K, Thompson EW,
Woolford L, Evdokiou A, Pollard JW, Robertson SA, Ingman
WV (2017) CCL2-driven inflammation increases mammary gland
stromal density and cancer susceptibility in a transgenic mouse
model. Breast Cancer Research 19(1):4. https ://doi.org/10.1186/
s1305 8-016-0796-z
21. Rakoff-Nahoum S (2006) Why cancer and inflammation? Yale J
Biol Med 79(3–4):123–130
22. Afrasiabi K, Zhou YH, Fleischman A (2015) Chronic inflamma-
tion: is it the driver or is it paving the road for malignant transfor-
mation? Genes cancer 6(5-6):214–219. https ://doi.org/10.18632 /
genes andca ncer.64
23. Grivennikov SI, Greten FR, Karin M (2010) Immunity, inflamma-
tion, and cancer. Cell 140(6):883–899. https ://doi.org/10.1016/j.
cell.2010.01.025
24. Retsky MW, Demicheli R (2017) Perioperative Inflammation as
Triggering Origin of Metastasis Development, 1st edn. Springer
International Publishing, Berlin
25. Balkwill F (2009) Tumour necrosis factor and cancer. Nat Rev
Cancer 9(5):361–371. https ://doi.org/10.1038/nrc26 28
26. Leek RD, Landers R, Fox SB, Ng F, Harris AL, Lewis CE (1998)
Association of tumour necrosis factor alpha and its receptors with
thymidine phosphorylase expression in invasive breast carcinoma.
Br J Cancer 77(12):2246–2251
27. Baumgarten SC, Frasor J (2012) Minireview: inflammation:
an instigator of more aggressive estrogen receptor (ER) posi-
tive breast cancers. Mol Endocrinol 26(3):360–371. https ://doi.
org/10.1210/me.2011-1302
28. Mantovani A, Allavena P, Sica A, Balkwill F (2008) Cancer-
related inflammation. Nature 454(7203):436–444. https ://doi.
org/10.1038/natur e0720 5
29. Staller P, Sulitkova J, Lisztwan J, Moch H, Oakeley EJ, Krek W
(2003) Chemokine receptor CXCR4 downregulated by von Hip-
pel-Lindau tumour suppressor pVHL. Nature 425(6955):307–311.
https ://doi.org/10.1038/natur e0187 4
30. Subramanian M, Pusphendran CK, Tarachand U, Devasagayam
TP (1993) Gestation confers temporary resistance to peroxidation
in the maternal rat brain. Neurosci Lett 155(2):151–154
31. Irwin RW, Yao J, Hamilton RT, Cadenas E, Brinton RD, Nilsen J
(2008) Progesterone and estrogen regulate oxidative metabolism
in brain mitochondria. Endocrinology 149(6):3167–3175. https ://
doi.org/10.1210/en.2007-1227
32. Karatepe O, Altiok M, Battal M, Kamali G, Kemik A, Aydin T,
Karahan S (2012) The effect of progesterone in the prevention of
the chemically induced experimental colitis in rats. Acta Cir Bras
27(1):23–29
33. Dhote VV, Balaraman R (2007) Gender specific effect of proges-
terone on myocardial ischemia/reperfusion injury in rats. Life Sci
81(3):188–197. https ://doi.org/10.1016/j.lfs.2007.05.010
34. Jiang C, Wang J, Li X, Liu C, Chen N, Hao Y (2009) Progesterone
exerts neuroprotective effects by inhibiting inflammatory response
after stroke. Inflamm Res 58(9):619–624. https ://doi.org/10.1007/
s0001 1-009-0032-8
35. Ryu KY, Maehr R, Gilchrist CA, Long MA, Bouley DM, Mueller
B, Ploegh HL, Kopito RR (2007) The mouse polyubiquitin gene
UbC is essential for fetal liver development, cell-cycle progres-
sion and stress tolerance. EMBO J 26(11):2693–2706. https ://doi.
org/10.1038/sj.emboj .76017 22
36. Yoneda-Kato N, Tomoda K, Umehara M, Arata Y, Kato JY (2005)
Myeloid leukemia factor 1 regulates p53 by suppressing COP1 via
COP9 signalosome subunit 3. EMBO J 24(9):1739–1749. https ://
doi.org/10.1038/sj.emboj .76006 56