Abstract

   ^1H NMR is used to detect alterations in metabolites and their linkage
   to metabolic processes in a number of pathological conditions including
   breast cancer. Inositol 1, 4, 5 trisphosphate (IP[3]R) receptor is an
   intracellular calcium channel known to regulate metabolism and cellular
   bioenergetics. Its expression is up regulated in a number of cancers.
   However, its linkage to metabolism in disease conditions has not been
   evaluated. This study was designed to determine the association if any,
   of these metabolites with altered expression of IP3R in breast cancer.
   We used ^1H NMR to identify metabolites in the serum of breast cancer
   patients (n = 27) and performed Real-time Polymerase Chain Reaction
   analysis for quantifying the expression of IP[3]R type 3 and type 2 in
   tissues from breast cancer patients (n = 40). Principal Component
   Analysis (PCA) and Partial Least Square-Discriminant Analysis (PLS-DA)
   clearly distinguished patients with high/low IP[3]R expression from
   healthy subjects. The present study revealed high expression of IP[3]R
   type 2 and type 3 in human breast tumor tissue compared to adjacent
   non-tumorous tissue. Moreover, patients with ≥ 2-fold increase in
   IP[3]R (high IP[3]R group) had significantly higher concentration of
   metabolic intermediates compared to those with < 2-fold increase in
   IP[3]R (low IP[3]R group). We observed an increase in lipoprotein
   content and the levels of metabolites like lactate, lysine and alanine
   and a decrease in the levels of pyruvate and glucose in serum of high
   IP[3]R group patients when compared to those in healthy subjects.
   Receiver operating characteristic (ROC) curve analysis was performed to
   show the clinical utility of metabolites. In addition to the human
   studies, functional relevance of IP[3]Rs in causing metabolic
   disruption was observed in MCF-7 and MDA MB-231 cells. Results from our
   studies bring forth the importance of metabolic (or metabolomics)
   profiling of serum by ^1H NMR in conjunction with tissue expression
   studies for characterizing breast cancer patients. The results from
   this study provide new insights into relationship of breast cancer
   metabolites with IP[3]R.

Introduction

   Inositol 1,4,5-trisphosphate receptors (IP[3]Rs) are calcium (Ca^2+)
   channels that regulate autophagy and metabolism [[36]1].Three different
   tissue-specific isoforms of IP[3]Rs, namely IP[3]R type 1 (IP[3]R1),
   IP[3]R type 2 (IP[3]R2) and IP[3]R type 3 (IP[3]R3) have been reported
   so far [[37]2].These receptors regulate the transfer of Ca^2+ from
   endoplasmic reticulum (ER) to mitochondria via a mitochondrial membrane
   transport protein, which in turn regulates cellular bioenergetics
   [[38]1]. Altered IP[3]R activity and/or the remodeling of IP[3]R
   expression profiles may be exploited by cancer cells to promote growth
   and drug resistance. This becomes important since altered expression of
   IP[3]Rs have been reported in various cancer types [[39]3, [40]4]. One
   of the key areas that can be targeted for potential treatment of cancer
   is the regulation of metabolism through IP[3]Rs.

   IP[3]Rs regulate cell fate by interacting with a number of proteins
   involved in apoptotic as well as anti-apoptotic pathways [[41]2]. Under
   normal physiological conditions, IP[3]Rs lead to basal autophagy
   together with balanced mitochondrial bioenergetics. In contrast,
   IP[3]Rs in cancer cells cause Ca^2+ overloading leading to aggressive
   phenotype depicted as cell survival during therapy [[42]5].
   Deregulation of IP[3]Rs play an important role in tumor growth,
   aggressiveness and drug resistance via modulation of different
   signaling pathways such as autophagy and energy metabolism [[43]5,
   [44]6]. A number of studies reveal altered expression of Ca^2+ channels
   and pumps in many human cancers, including breast, ovarian, glioma,
   liver, pancreatic, prostate, melanoma, colon, lung, bladder, thyroid,
   and oral cancer [[45]7–[46]11].

   NMR Spectroscopy is one of the most widely used techniques in
   metabolomics studies. NMR-based metabolomics approach using various
   body fluids and tissue specimens has been applied for diagnosis and
   pathophysiological studies of various diseases such as coronary heart
   diseases [[47]12], liver grafts [[48]13], inborn metabolic disorders
   [[49]14], lung injury [[50]15, [51]16], hemorrhagic shock [[52]17],
   liver abscess [[53]18], and many cancers including breast, lung,
   prostrate, bladder and gall bladder cancers [[54]19–[55]23]. Being
   non-invasive, blood serum related studies have more potential than
   using other body fluids, and several studies have so far been reported
   using serum or plasma [[56]17, [57]19, [58]20, [59]22]. There are many
   studies which have reported proteomic analyses being used in prognosis
   of breast cancer. Regular analysis tools like histochemistry and other
   techniques like MALDI-TOF-MS and SELDI–TOF–MS have been used for
   studying various breast cancer biomarkers [[60]24–[61]26]. Few studies
   have been carried out for the metabolic profiling of serum in breast
   cancer patients using clinical test approach [[62]22, [63]27].

   In the present study, we used metabolomics to differentially identify
   various metabolites in serum of breast cancer patients with high and
   low expression of IP[3]R as compared to the ones in serum of healthy
   controls. Principle Component analysis (PCA) and Partial Least Square-
   Discriminant Analysis (PLS-DA) were applied to analyze the ^1H NMR
   spectra of metabolites in serum samples. The differential metabolites
   and their pathways along with their ROC analyses are presented.

Material and Methods

Collection of Tissues

   Blood and tissue specimens were obtained from patients undergoing
   surgery at Department of Endocrine Surgery, Sanjay Gandhi Post Graduate
   Institute of Medical Sciences, Lucknow, India. Blood was also obtained
   from healthy volunteers for use in metabolomics studies. Approval from
   the Institute's Ethics Committee was taken prior to collection of
   samples from participants in the study. Written informed consent was
   obtained from the participants on the format approved by the ethics
   committee. The study recruited 40 breast cancer patients from whom
   tissue samples were obtained post-operatively. Out of these subjects
   only 27 blood samples were available for analysis. Tissues were
   collected from the tumor zone (tissue within the tumor boundary) and
   the normal zone (distal normal tissue at least 10 mm from the outer
   tumor boundary). A fraction of each tissue was fixed in formalin and
   embedded in paraffin for routine histopathological analysis. The
   remaining tissue samples were frozen in liquid nitrogen and then stored
   at -80°C for RNA analysis. The study also recruited 15 age matched
   healthy controls who volunteered to provide blood sample for the study.
   Serum samples obtained after centrifugation (at 3000 rpm, 4°C for 10
   min) of the blood samples were stored at -80°C for subsequent NMR
   experiments. Details of patients are given in ([64]S1 Table).

Real-Time PCR Analysis

   Total RNA was isolated from stored tissues using Tri Reagent ^TM (Life
   Technologies, USA) and reverse transcribed to synthesize cDNA using
   SuperScript^TM VILO cDNA Synthesis Kit (Life Technologies, USA)
   according to the manufacturer's instructions. Real-time PCR was
   performed using a SYBR mix (Applied Biosystems, USA) as per the
   manufacturer's instructions. Fold changes in gene expression were
   calculated using the 2−^ΔΔCT method. Specific primers of the genes used
   for PCR are listed below:

   (IP[3]R1): F5'-AACCGCTACTCTGCCCAAAA-3',

   R5'-AGTTTGTTGAGTAGCACTGCGTCT-3'

   (IP[3]R2): F5'-GCGATCTGCACATCTATGCTG-3',

   R5'-AAGTATTAATGTAGGCCCAAGACCTATT-3'

   (IP[3]R3): F5'-GGGCTCTCGGTGCCTGA-3',

   R5'-GGAGGGCTTGCGGAGAA-3'

   (GAPDH): F5′-AGGGCTGCTTTTAACTCTGGT-3',

   R5′-CCCCACTTGATTTTGGAGGGA-3′

Immunohistochemistry

   Tissue sections of 3 micron were cut on poly-L-Lysine coated slides.
   After de paraffinization in xylene and hydration by gradient alcohol
   series, antigen retrieval was done by heat treatment in citrate buffer
   (10 mM, pH 6.0). The sections were incubated in 10% NSS (normal sheep
   serum) for 45 minute to block non-specific binding and further
   incubated with antibodies against IP[3]R2 or IP[3]R3 (1:200) in 0.1%
   PBST (Phosphate buffer saline Triton X100) overnight at 4°C. Sections
   were stained using Quick Universal ABC kit (Vector, USA) followed by
   peroxidase staining reaction with DAB/H[2]O[2] as chromogen. The
   stained sections were observed under bright field light microscope
   (Nikon Eclipse 80i; Nikon Instech Co. Ltd., Kanagawa, Japan).

NMR Studies

   Serum samples were thawed just before acquiring NMR spectra and 400 μl
   of each sample was used for the NMR experiment. A co-axial insert
   containing TSP (sodium salt of trimethylsilyl-2, 2, 3,
   3-tetradeuteropropionic acid) was used for deuterium lock as well as
   for external standard reference. NMR spectra were recorded at Bruker
   Biospin Avance III 800 MHz NMR (Bruker GmBH, Germany) spectrometer
   operating at proton frequency of 800.21 MHz and equipped with
   CryoProbe. Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence was used for
   acquiring 1D ^1H NMR spectra. It eliminates the broad resonances
   arising from the macromolecules (proteins and lipids). CPMG experiment
   provides spectra with smooth baseline which facilitates multivariate
   analyses. NMR spectra of patient and healthy control samples were
   recorded with constant parameters to ensure accuracy of the results.
   All spectra were recorded with 64 k time domain data points, 20.02 ppm
   spectral width, 128 scans, relaxation delay of 5 s, and constant
   receiver gain value of 80.6 and 400 ms echo time. TOPSPIN software
   (version 2.1) was used for phase and base line correction of spectra.
   Two-dimensional homo-nuclear (^1H-^1H COSY and^1H-^1HTOCSY) and
   hetero-nuclear (^1H-^13C HSQC) data were also acquired to confirm the
   resonance assignment. Assignment of metabolites was also confirmed from
   a standard database [[65]21].

Cell Culture

   The human cancer cell line, MCF-7 cells (Invasive ductal carcinoma,
   Estrogen/Progesterone receptor positive cell line), MDA MB-231
   (Adenocarcinoma, Estrogen/Progesterone/Her2+ receptor negative or
   triple cell line) and MCF 10A (non tumorous mammary cell line) were
   procured from National Centre for Cell Sciences (NCCS), Pune, India and
   American Tissue & Cell Culture (ATCC) were used in the present study.
   MCF-7 cells were cultured in DMEM-F12 (Life Technologies, USA) and MDA
   MB-231 in DMEM supplemented with 10% fetal bovine serum. MCF 10 A cells
   were cultured in DMEM/F12 media supplemented with supplemented with 5%
   horse serum (Life Technologies, USA) along with cholera toxin
   (100ng/ml), Epidermal growth factor (EGF 20ng/ml), hydrocortisone
   (500ng/ml), insulin (10μg/ml) (Sigma,USA) and 1% penicillin,
   streptomycin antibiotic mixture (Life Technologies, USA). Cells were
   incubated at 37°C and 5% CO[2] in specified media.

Cell Treatment

   MCF-7, MDA MB-231 and MCF 10A cells (2×10^6) were seeded in 6-well
   plates and treated with either vehicle (Control) or 25 μM XeC
   (inhibitor of IP3R mediated calcium release) (Cayman) in calcium free
   medium (EGTA pretreatment,1mM,was given to chelate extracellular
   calcium) for 48 hours. MCF-7, MDA MB-231 and MCF 10A cells were
   transfected using Lipofectamine 2000 transfect reagent (Invitrogen,
   USA) as per the manufacturer's protocol. Briefly, 2×10^6cells were
   transfected with 4μg siRNA directed against the human IP3R2 or IP3R3
   mRNA sequence (ON-TARGET plus, Dharmacon, USA) or control siRNA
   (siGENOME non-targeting siRNA; Dharmacon, USA) for 72 hours and further
   analyses were performed.

Measurement of Glucose Uptake

   Percent glucose uptake was determined in MCF-7, MDA MB-231 and MCF 10A
   cells plated in 96-well plates and treated with vehicle alone or 25 μM
   XeC for 24 hours, or with siC (non-targeted siRNA) or siIP[3]R2 or
   siIP3R3 for 72 hours following respective protocols provided by the
   cell based assay kit manufacturer (Cayman, USA). Glucose uptake was
   analyzed using {2-[N-(7-Nitrobenz-2-oxa-1, 3-diazol-4-yl]
   Amino}-2-Deoxyglucose (NBDG), a fluorescently labeled deoxy glucose
   analogue as a probe for detection of glucose taken up by cultured
   cells.

RT2 Profiler PCR Array

   MCF-7 cells were plated in 6-well plate, treated with vehicle alone or
   XeC for 24 hours. Total RNA was isolated from the treated- and
   untreated- MCF-7 cells by a single step method using TRIZOL^® reagent
   (MRC, USA). RNA was converted into first-strand cDNA using the RT2
   First Strand Kit (Qiagen, Germany). Further, the cDNA was mixed with an
   appropriate RT2 SYBR Green Mastermix (Qiagen, Germany). This mixture
   was aliquoted into the wells of the RT2Profiler PCR Array. Real time
   PCR was performed and relative expression was determined using 2-ΔΔCT
   method. The detailed protocol as per the manufacturer’s instructions
   (RT2 Profiler assay, Qiagen, Germany) was followed.

Statistical Analysis

   Multivariate statistical analysis was performed on 1D CPMG data
   collected from serum samples of patients and healthy control subjects.
   Further data reductions were applied for statistical analysis. 0.03 ppm
   bin file were generated using Amix software and exported to the
   ‘Unscrambler X’ Software package (Version10.0.1, Camo USA, Norway).
   Spectral alignment was checked prior to statistical analysis.
   Supervised Principal Component Analysis (PCA) and unsupervised Partial
   Least Square Discriminant Analysis (PLS-DA) was done using bin file.
   Model validation was checked by employing a 7-fold internal cross
   validation procedure which gives the model validity in the form of
   explained variance results (R^2) and predictive capability results
   (Q^2).

   Quantitative analysis was performed by taking relative integral areas.
   Integral values of these metabolites were compared for serum samples
   from both the groups. Student t-test was performed to test the
   significance of metabolite variations within groups. ROC analysis was
   performed for area under curve (AUC) of metabolites to predict the
   diagnostic significance of metabolites using Graph Pad Prism software.
   Metaboanalyst 3.0 was used for pathway analysis.

   All cell line experiments were repeated at least three times. The data
   were expressed as mean ±S.E. (Standard Error) from three independent
   experiments. Statistical analysis was performed by using ANOVA followed
   by post hoc Newman-Keuls test or Student’s t test (two-tailed) while
   comparing multiple or two groups. The criterion for statistical
   significance was taken as p< 0.05.

Results

Dysregulated Inositol 1, 4, 5-Trisphosphate Receptors (IP[3]Rs) in Breast
Cancer Tissue

   The complete scheme for the analysis of test results of patients
   employed in this study is shown in ([66]Fig 1). Expression levels of
   IP[3]R2 and IP[3]R3 mRNA transcripts were investigated in tissues
   obtained from breast cancer patients using quantitative real-time PCR.
   IP[3]R1 was not significantly different in breast tumor tissue from the
   adjacent non-tumorous tissue with average increase in tumor volume
   ~0.77-fold relative to non-tumorous tissue (S.E.±0.28, [67]Fig 2A).
   Average fold change of IP[3]R2 and IP[3]R3 transcript levels were
   significantly higher (~4- fold) in tumor tissue compared to the
   adjacent non-tumorous region from the same patient (p<0.001, n = 30,
   [68]Fig 2B and 2C). Further, immunohistochemistry revealed intense
   staining of IP[3]R2 and IP[3]R3 in the tumor region as compared to the
   adjacent non-tumorous region ([69]Fig 2D and 2E).

Fig 1. An overview of the workflow performed for the serum metabolic
profiling of breast cancer patients using ^1H NMR spectroscopy.

   [70]Fig 1
   [71]Open in a new tab

Fig 2. Analysis of expression of Inositol 1, 4, 5-trisphosphate receptors
type 2 (IP[3]R2) and type 3 (IP[3]R3) in tumor tissue of breast cancer
patients.

   [72]Fig 2
   [73]Open in a new tab

   Relative mRNA expression of (A) IP[3]R1, (B) IP[3]R2 (p<0.001) and (C)
   IP[3]R3 (p<0.001) from tumoral and extra-tumoral tissues of breast
   cancer patients was analyzed using Real -time PCR. Analysis of tumoral
   and extra-tumoral tissues was done by incubating the tissue sections
   with antibodies against (D) IP[3]R2 and (E) IP[3]R3, and detected using
   DAB staining.

Metabolic Profile of Serum in Breast Cancer by ^1H NMR

   ^1H CPMG spectra depicting comparison between healthy subjects and
   patients (High IP[3]R and Low IP[3]R) is reported in ([74]Fig 3A and
   3B). Spectrum of serum from patients and healthy subjects shows the
   presence of metabolites including amino acids like alanine, lysine,
   glutamine, glutamate, other branched chain amino acids, metabolites
   like lactate, N-acetyl glycoproteins (NAG), glucose, lipids,
   creatinine, acetate, pyruvate, citrate and lipoproteins (HDL, LDL and
   VLDL). There were no marked qualitative and quantitative differences
   between spectra from healthy subjects and low IP[3]R patients. In
   contrast, quantitatively spectra from healthy control and high IP[3]R
   patient groups differed significantly with spectral quality remaining
   same. The metabolites making significant contribution for group
   separation were glucose, lactate, lysine, alanine, glutamate, NAG,
   pyruvate and lipid content. These differences were further confirmed by
   applying multivariate analyses like PCA and PLS-DA.

Fig 3.

   [75]Fig 3
   [76]Open in a new tab

   Stack plot of representative ^1H NMR CPMG spectra of serum from A)
   healthy control and patient with high tissue expression of IP[3]R, B)
   healthy control and patient with low tissue expression of IP[3]R. All
   the spectra were plotted at same vertical scale for quantitative
   comparison.

Multivariate Analysis of ^1H NMR Spectra of Serum: Discrimination Between
Healthy Control and Patient (high IP[3]R and low IP[3]R) Groups

   The NMR spectra recorded for serum samples were subjected to
   multivariate data analysis to unravel changes in the serum metabolic
   profiles and identify the potentially altered metabolic pathways in
   patients with differentially expressing IP[3]R gene with respect to
   healthy controls. The bias in this study is negated with the help of
   multivariate analysis which itself normalizes the data and separates
   the groups after normalizing the bias in the data. PCA score plots were
   generated initially for an overview of the dataset. PCA scores for all
   groups were obtained as shown in ([77]Fig 4A). PCA analysis was
   performed with integral data for healthy control versus patients. The
   scattered score plot of PCA revealed that healthy controls were well
   clustered and distinguishable from patient groups (healthy control and
   high/low IP[3]R) ([78]Fig 4B and 4C). Moreover, the outcome of PCA
   showed clustering of samples according to the disease milieu.

Fig 4. Score Plot Generated From PCA And PLS-DA Analysis Of NMR Spectra Of
Serum From Healthy Control And Patient Groups.

   [79]Fig 4
   [80]Open in a new tab

   (A) PCA and (D) PLS-DA score plot generated for healthy control vs.
   patient group, (B) PCA and (E) PLS-DA score plot generated for healthy
   control and high IP[3]R group, (C) PCA and (F) PLS-DA score plot
   generated for healthy control and low IP[3]R group.

   To obtain a more specific statistical analysis, PLS-DA was employed on
   all the serum samples, and the results showed a clear separation and a
   clustered pattern among the various groups in the PLS-DA score plot as
   shown in ([81]Fig 4D). The goodness of fit and predictability for the
   PLS-DA models for healthy control versus patient group was reflected in
   the values of R^2X = 0.85, R^2Y = 0.77 and Q^2 = 0.85. A pair wise
   PLS-DA analysis (R^2X = 0.97, R^2Y = 0.85 and Q^2 = 0.92) between
   groups was also performed for healthy control versus high IP[3]R group
   ([82]Fig 4E) and healthy control versus low IP[3]R group ([83]Fig 4F).
   The analysis demonstrated well-segregated and clustered form of PLS-DA
   score plots. We further analyzed the metabolites which made the groups
   different as depicted in the loading plot ([84]Fig 5A–5C). The
   discriminating metabolites between the groups were identified as
   lipids, lactate, glutamate, glucose, lysine, alanine, pyruvate, and
   NAG.

Fig 5.

   [85]Fig 5
   [86]Open in a new tab

   Score plot and corresponding loading plot generated from PLS-DA
   analysis between (A) healthy control and patient group, (B) healthy
   control and high IP[3]R group, (C) healthy control and low IP[3]R
   group.

Quantitative Analysis of Metabolites

   Pair wise PLS-DA comparison of healthy control with patients with high
   tissue expression of IP[3]R revealed several metabolites as
   discriminatory metabolites. The list of these discriminatory
   metabolites is given in ([87]Table 1) wherein relative integral areas
   of these metabolites were calculated and compared between healthy
   controls and patient groups by use of statistical analysis (p<0.05) as
   reported in ([88]Fig 6).

Table 1. Metabolites contributing to the difference between healthy control
and high IP[3]R patient group.

   Metabolites Chemical shift (multiplicity) Trend AUC[89]^a of ROC p
   value
   High IP3R Low IP3R
   Glucose 3–4 ppm ↓ - 0.89 <0.001
   Lactate 4.10 (q) ↑ - 1.00 <0.001
   Glutamate 2.42 (m) ↑ - 0.85 <0.001
   Lysine 1.69 (m) ↑ - 0.71 <0.05
   Alanine 1.46 (d) ↑ - 0.70 <0.001
   Pyruvate 2.2 (s) ↓ - 0.73 <0.01
   NAG 2.1 (m) ↓ - 0.93 <0.001
   Lipids 1.3 (m) ↑ - 0.70 <0.05
   [90]Open in a new tab

   ^aAUC (Area under curve) of ROC. Integral areas of normalized spectral
   regions were calculated and significance was obtained using student t
   test. Significant differences shown with different symbols, represent
   significant increase in patient with high expression of IP[3]R as
   compared to healthy control; (↑) represents significant decrease
   compared to control; (↓) represents no significant difference; m =
   multiplet; q = quartet; d = doublet and s = singlet.

Fig 6. Representative box-cum-whisker plots for significant metabolites of
healthy control vs high IP[3]R patient group (with p>0.05).

   [91]Fig 6
   [92]Open in a new tab

   Relative integral areas are calculated from CPMG spectra and plotted as
   a.u.

   Resonances arising from lipids were significantly higher in patients
   with high tissue expression of IP[3]R as compared to healthy control.
   Patients with low tissue expression of IP[3]R showed no significant
   difference compared to healthy control. Among the branched chain amino
   acids (valine/leucine/isoleucine) there were no significant differences
   between healthy control and patient groups. There was no change in
   aromatic amino acids in any of the patient groups as compared to
   healthy control. A high energy molecule like glucose was found to be
   significantly decreased in patients with high expression of IP[3]R when
   compared to healthy control. Lactate was found to be significantly
   higher in patients with high tissue expression of IP[3]R when compared
   to patients with low tissue expression of IP[3]R and healthy control.
   Lactate concentration was higher in patients with low tissue expression
   of IP[3]R as compared to healthy controls. The high IP[3]R group had
   low levels of pyruvate as compared to healthy control, while there was
   no significant difference observed in case of low IP[3]R group for the
   resonance of pyruvate when compared to healthy control. Amino acids
   like alanine and lysine were significantly increased in all patient
   groups compared to healthy control, but only patients with high IP[3]R
   showed significant increase. Glutamate was found to be elevated in
   patients with high IP[3]R while there was no significant difference
   between low IP[3]R group and healthy control. NAG levels were found to
   be reduced in both the patient groups. The decrease in NAG levels was
   greater in case of patients with high IP[3]R than with low IP[3]R or
   healthy control.

Inhibiting or Silencing Inositol 1,4,5-trisphosphate Receptors (IP[3]Rs)
Compromises Cellular Bioenergetics in Breast Cancer Cells

   In order to demonstrate the functional relationship between IP[3]R and
   metabolites, we investigated the effect of pharmacological inhibition
   and siRNA silencing of IP[3]R in cancerous cell lines in 2 separate
   experiments. A significant decrease was observed in glucose uptake by
   blocking/silencing the IP[3]R receptor in MCF-7, MDA MB-231 and MCF 10
   cells ([93]Fig 7A–7F). However the reduction of glucose uptake was more
   pronounced in MCF 7 and MDMBA-231 cells as compared to MCF 1OA cells.
   Further, we performed RT^2profiler assay which revealed a significant
   fold reduction in the expression levels of key glycolytic and
   mitochondrial pathway genes in MCF-7 cells treated with IP[3]R
   inhibitor compared to vehicle treated cells ([94]Fig 7G p<0.001. These
   results showed that in vitro inhibition of IP[3]R result in compromised
   bioenergetics, both in terms of glucose and mitochondrial metabolism.

Fig 7. Inhibition of IP[3]R in MCF -7 breast cancer cells effects metabolism.

   [95]Fig 7
   [96]Open in a new tab

   Glucose uptake in MCF-7 cells was analyzed using NBDG a fluorescently
   labeled deoxy glucose analogue as a probe for detection of glucose
   taken up by cultured cells. Quantitative estimation of glucose uptake,
   using a cell based assay kit, was performed as per instructions
   provided by the manufacturer (Cayman, USA). Cells were plated in
   96-well plates and treated with 25 μM XeC for 24 hours or with siC
   (non-targeted siRNA) or siIP[3]R2 or siIP[3]R3 (72 hours,Fig7A and 7D)
   Representative graph in MCF-7 cells. (Fig 7B and 7E) Representative
   graph showing percentage of glucose uptake as estimated using cell
   based assay in MDA MB-231 cells. (Fig7C and 7F) Representative graph
   showing percentage of glucose uptake as estimated using cell based
   assay in MCF 10A cells. RNA was extracted from treated and untreated
   cells and cDNA was prepared (Fig 7G) RT profiler PCR array for glucose
   as well as mitochondrial metabolism genes was performed using cDNA
   prepared from mRNA of MCF-7 cells. Data represent mean ±SEM. *p< 0.05,
   ***p< 0.001 compared to vehicle.

Discussion

   Tumorigenesis is the process of cancer formation where by normal cells
   are transformed into cancerous cells. This process is associated with a
   series of metabolic events that fuel cells to multiply or become
   aggressive [[97]28]. IP[3]R channels are calcium channels which play a
   crucial role in regulating metabolism in cells [[98]1]. In the present
   study, we targeted patients with altered expression of IP[3]R.
   Elucidation of metabolic changes involved in the pathogenicity with
   respect to IP[3]R can be helpful for exploring novel biomarkers for the
   diagnosis and surveillance of the disease. This study brings forth
   several important findings of NMR-based metabolomics of human serum
   samples from breast cancer patients. We report altered expression of
   IP[3]R2 and IP[3]R3 while IP[3]R1 remain unaltered in tumor tissue
   compared to adjacent non-tumorous tissues from breast cancer patients.
   As IP[3]R is also known to regulate autophagy and metabolism [[99]1],
   we also compared serum metabolic profiles among healthy subjects and
   breast cancer patients with high IP[3]R and low IP[3]R. We determined
   that the NMR-derived fingerprint of the serum metabolic profile is able
   to help discriminate among breast cancer patients with high and low
   IP[3]R expression and healthy controls. Interestingly, PCA and PLS-DA
   multivariate analysis not only differentiated the groups, but also
   revealed potential biomarkers from the complex NMR spectra of serum.
   Absolute quantitation and measurement of important metabolites using
   integral values established accuracy of these results. Lastly, the ROC
   curve analysis of vital biomarkers provided the clinical relevance of
   the essential metabolites. In addition, the in vitro results showed
   that inhibition of IP[3]R resulted in reduced glucose uptake in breast
   cancer cells -MCF-7 and MDA MB-231.

   Dysregulated metabolic pathways in breast cancer patients with high
   IP[3]R were identified on the basis of the metabolites that exhibited
   significant change in the NMR spectra of serum samples. A simplified
   metabolic pathway is demonstrated in [100]Fig 8 based on Metaboanalyst
   software. The analysis showed alteration in a number of biochemical
   pathways like glycolysis, pyruvate metabolism,
   alanine/aspartate/glutamate metabolism and lysine degradation. The
   analysis was calculated based on the significance value (p<0.05) of the
   pathway enrichment analysis. Here “Impact” is the pathway impact value
   calculated from pathway topology analysis and a value equal to or
   greater than 0.1 was considered significant.

Fig 8. An integrated analysis based on Metaboanalyst software (pathway tool)
for a simplified view of contributing pathways.

   [101]Fig 8
   [102]Open in a new tab

   The panel shows a view of metabolism in cancers depicting glycolysis
   and Krebs cycle which are modified to different processes like lipid
   and amino acid synthesis to meet the requirement of proliferating
   cells. (Metabolites depicted with green/red are decreased/increased in
   the present study).

   Deregulated expression of IP[3]Rs has been reported in various cancer
   types [[103]3, [104]4]. Deregulation of IP[3]Rs aids in growth and
   aggressiveness of the tumor, and also drug resistance through different
   signaling pathways such as autophagy and energy metabolism [[105]5,
   [106]6]. Through this study, we are the first group to report
   dysregulated expression of IP[3]R2 and IP[3]R3 while IP[3]R1 remains
   unaltered in breast cancer tissue compared to their expression in
   adjacent non-tumorous tissues from the patients. Increased expression
   of IP[3]R3 has been associated with tumor aggressiveness in other
   cancer types such as colorectal carcinoma and gastric cancer [[107]3,
   [108]4]. Similarly, IP[3]R2 was found to be increased in lymphomas
   which makes it susceptible to be targeted for its association with
   Bcl-2 [[109]29]. We also observed that patients with more than 2-fold
   of IP[3]R (both IP[3]R3 and IP[3]R2) tissue expression also had higher
   concentration of metabolic intermediates in serum analyzed by ^1H NMR
   spectroscopy. It has been demonstrated that blocking/knocking out
   IP[3]R in MCF-7 breast cancer cells compromises bioenergetics and
   induces autophagy [[110]1].

   A number of studies have shown modified expression of Ca^2+ channels
   and pumps in a number of human cancers like breast, ovarian, glioma,
   liver, pancreatic, prostate, melanoma, colon, lung, bladder, thyroid,
   and oral cancers [[111]7, [112]8, [113]10, [114]11, [115]30, [116]31].
   Calcium transfer through these channels, especially IP[3]R is known to
   regulate metabolism. There are also various online datasets available
   showing the expression of IP3Rs as well as its link to metabolism. A
   search of GEO profiles database revealed high expression of IP[3]R3 in
   breast cancer patients (GEO accession: GDS1250) [[117]32]. Another
   study in GEO profiles database showed analysis of A549 lung carcinoma
   and M059K glioblastoma cells treated with dichloroacetate (DCA), an
   inhibitor of the mitochondrial pyruvate dehydrogenase kinase. The
   expression of IP[3]R3 was down regulated in this study showing its link
   to metabolism (GEO accession: GDS2444) [[118]33]. Furthermore, a genome
   wide expression analysis in HPV16 cervical cancer revealed altered
   metabolic pathways and increased expression of IP[3]R[[119]34].

   It has been demonstrated that ^1H NMR spectroscopy is a potential tool
   for identifying metabolic perturbations in a number of pathological
   conditions using various body fluids (e.g. cerebrospinal fluid, serum,
   urine, plasma etc.) [[120]35, [121]36]. Many studies to date have used
   NMR or MS methods to detect altered metabolic profiles in different
   types of malignancies due to their capability to analyze a large number
   of metabolites in a single experiment [[122]22, [123]37]. In
   particular, several investigators, by using a metabolomics approach,
   have focused on establishing breast cancer biomarkers
   [[124]38–[125]40]. Numerous metabolites, including glucose, lactate,
   lipids, choline, and amino acids were shown to correlate with breast
   cancer [[126]40–[127]42]. A majority of these investigations were
   focused on either breast cancer tumors or cell lines. All these studies
   utilized NMR methods alone except for a study that utilized a
   combination of NMR and MS methods [[128]43].

   In the present study, role of ^1H NMR spectroscopy was explored in
   serum of breast cancer patients with high tissue expression of IP[3]R
   in comparison to healthy controls after characterizing the patients
   under different groups based on their IP[3]R expression status. A total
   of 8 metabolites have been identified in studies associated with breast
   cancer [[129]27, [130]41]. These metabolites represent changes in one
   or the other metabolic pathways involved in cancer cell bioenergetics
   including amino acid metabolism (glutamic acid, alanine, and lysine),
   glycolysis or gluconeogenesis (glucose, lactate, and pyruvate) and
   glutamate metabolism (glutamate). Increased lactate, reduced glucose
   and pyruvate, as reported in our study, are among the early findings of
   metabolic changes reported for breast tumors. These changes are
   associated with high rate of glycolysis in cancer cells. Also,
   association of a number of amino acids, fatty acids, and organic acids
   with breast cancer has previously been shown [[131]22, [132]27]. High
   levels of glutamate in the present study can be associated with an
   activation of glutaminolysis through an increased activity of the
   mitochondrial enzyme glutaminase. Glutamate has been recognized in a
   few studies as a potential biomarker related to various types of
   cancers and more recently in breast tissues as a tumor biomarker
   [[133]44–[134]47]. As shown by our data, the mean concentrations for a
   number of these metabolites, including alanine and lysine, were found
   to be increased while that of N-acetyl-glycoprotein was decreased
   significantly in patients with high expression of IP[3]R (in comparison
   to healthy control). Alanine is a well-known tumor biomarker [[135]48,
   [136]49]. It is required for tumor proliferation and is a product of
   glycolysis/glutaminolysis. On the other hand, there was no significant
   change in the concentration of these metabolites in patients with low
   expression of IP[3]R when compared to healthy control. Most of the
   intermediates found in our study are of glycolytic pathway in patients
   with high IP[3]R, thereby revealing greater dependence on glucose
   metabolism. Our findings highlights the correlation of these
   metabolites with the expression of IP[3]R in breast cancer patients.

   Providing evidence regarding the functional relationship between
   IP[3]Rs and metabolic disruption albeit of importance cannot be done in
   patient material. As a rational alternative, we investigated the effect
   of inhibition of IP[3]R in MCF-7, MDA MB-231, MCF 10A cells using
   pharmacological inhibitor XeC and siRNA approach. IP[3]Rs are known to
   regulate cellular metabolism and bioenergetics. Blocking of IP[3]R by
   XeC as well as siRNA silencing of IP[3]R2 and IP[3]R3 in MCF-7 cells
   and MDA MB- 231 resulted in significant reduction in a key metabolic
   process such as % glucose uptake. MCF 10A which is normal cell line was
   not effected to the extent observed in cancer cell lines (MCF-7 and MDA
   MB-231) showing the differential effect of IP[3]R blocking / silencing.
   Further, RT^2 profiler assay revealed a significant fold reduction in
   the expression levels of key glycolytic and mitochondrial pathway genes
   in MCF-7 cells treated with IP[3]Rs blocker compared to vehicle treated
   cells. Thus, inhibiting IP[3]R affected the glucose and mitochondrial
   metabolism whose metabolites were utilized as a source of energy with
   high IP[3]R expression in breast cancer patients.

   Through this preliminary study, we have shown an association between
   IP3R and dysregulated metabolism in breast cancer patients. The
   relevance of our in vitro findings can be further explored by carrying
   out similar studies in primary cultures from human tissues as well as
   animal models.

   While metabolic reprogramming is one of the major hallmarks that
   differentiate cancer from normal cells, unchecked cell proliferation
   not only includes deregulation of proliferation but also alterations in
   energy metabolism to supply metabolites and cofactors required for cell
   growth and division [[137]50]. The modified metabolic schemes depend on
   glycolytic activity, the Warburg effect, and on elevated glutamine
   metabolism [[138]51]. While assessment of circulating metabolites may
   not reflect the tumor metabolism alone, it provides a global picture of
   the balance between tumor metabolism and the physiological condition of
   cancer patients. In current study we have presented NMR based
   metabolomics application to show the correlation between IP3R
   expression and serum metabolites. The differential metabolites can
   potentially be targeted for providing treatment options for future.

Conclusions

   In conclusion, our study provides serum metabolic profile of breast
   cancer tissues with high IP[3]R. Despite the variations among patients,
   similarities in metabolic alterations observed in the tumor samples
   resulted in clear group separation from the non-tumor samples. The
   findings from our study are consistent with the understanding of cancer
   metabolism. Further mechanistic insights need to be elucidated using in
   vitro studies for better understanding of the targets which are
   regulated through IP[3]R in cancer metabolism.

Supporting Information

   S1 Table. Summary of clinicopathological characteristics of breast
   cancer patients.

   Hormone receptors are receptors for estrogen and progesterone; HR-: at
   least one of the two receptors (estrogen and progesterone) is negative;
   HR+: both receptors are positive.

   (DOCX)
   [139]Click here for additional data file.^ (59.5KB, docx)

Data Availability

   We have submitted all the NMR data on a database metabolight
   Metabolight Accession number: MTBLS326

Funding Statement

   Mrs. Aru Singh, "the first author", received her fellowship from
   Council for Scientific and Industrial Research-University Grant
   Commission, New Delhi, Government of India. Besides this, funders had
   no role in study design, data collection and analysis, decision to
   publish, or preparation of the manuscript.

References