- Email: [email protected]
Mitochondrial chaperone, TRAP1 as a potential pharmacological target to combat cancer metabolism
Journal Pre-proofs Mitochondrial chaperone, TRAP1 as a potential pharmacological target to combat cancer metabolism Balaji Ramkumar, Shrikant P. Dharaskar, Mounika Guntipally, Khanderao Paithankar, Amere Subbarao Sreedhar PII: DOI: Reference:
S1567-7249(19)30019-4 https://doi.org/10.1016/j.mito.2019.09.011 MITOCH 1417
To appear in:
Mitochondrion
Received Date: Revised Date: Accepted Date:
31 January 2019 16 September 2019 18 September 2019
Please cite this article as: Ramkumar, B., Dharaskar, S.P., Guntipally, M., Paithankar, K., Sreedhar, A.S., Mitochondrial chaperone, TRAP1 as a potential pharmacological target to combat cancer metabolism, Mitochondrion (2019), doi: https://doi.org/10.1016/j.mito.2019.09.011
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Elsevier B.V. and Mitochondria Research Society. All rights reserved.
Mitochondrial chaperone, TRAP1 as a potential pharmacological target to combat cancer metabolism Balaji Ramkumar1, Shrikant P. Dharaskar1,2,#, Mounika Guntipally1,#, Khanderao Paithankar1, Amere Subbarao Sreedhar1*.
1CSIR-Centre
for Cellular and Molecular Biology, Uppal Road, Hyderabad, 500 007, Telangana,
India 2Academy
of Scientific & Innovation Research, Government of India.
# Authors have equally contributed to the manuscript * Correspondence: AS Sreedhar, [email protected] Running Title: TRAP1 is a potential pharmacological target to combat cancer metabolism Key words: Hsp, TRAP1, cancer, mitochondria, metabolism.
1
Abstract The stress response forms the most ancient defense system in living cells. Heat shock proteins (Hsps) are highly conserved across species and play major roles in mounting the stress response. The emerging information now suggests that Hsp90 family of chaperones display additional cellular roles contributing to diseases like cancer. For this reason, pharmacological targeting of Hsp90 has emerged as a novel antitumor strategy. However, its mitochondrial homologue TRAP1 has not been implicated in cancer with conclusive mechanistic insights. Since understanding the mutational spectrum of cancer cells indicates the outcome of the disease as well as treatment response, we examined mutational spectrum of TRAP1. Our in silico analyses of TRAP1 SNPs and CNVs correlated with the aggressive cancer phenotypes, and are found to be predominant over Hsp90 itself. The increased CNVs have been correlated with increased expression of TRAP1 in metastatic cancer cells, increased ATP production, and decreased oxygen consumption rate of mitochondria. Examining TRAP1 knockdown as well as over expression in metastatic cancer cells furthered our understanding that TRAP1 likely to facilitate the altered energy metabolism in the functional compromise of mitochondrial OXPHOS. Interestingly, the increased ATP levels in the TRAP1 background are found to be independent of glucose oxidation. Our results suggest TRAP1 role in triggering the alternate energy metabolism in cancer cells. Since targeting tumor metabolism is considered as an alternate strategy to combat cancer, we propose pharmacological targeting of TRAP1 to target alternate energy metabolism.
Introduction The stress response is highly conserved across the species from archaebacteria to higher eukaryotes, where heat shock proteins (Hsps) form the most ancient defense system (Feder and
2
Hofmann, 1999). Having chaperoning functions, Hsps contribute to the maintenance of cellular protein homeostasis under normal physiological conditions, in addition, exhibits cytoprotective functions in response to various extra- or intra-cellular stressors (Katschinski, 2004). Mounting evidence suggests that among Hsps, the Hsp90 chaperone contribute to various pathological disorders by facilitating mutated proteins to function normally (Neckers, 2007). This could be the reason why high Hsp expression correlates with cancer, a pathological disease known for the accumulation of wide spectrum of mutations. Further, Hsp90 gained much attention due to its role in the conformational maturation and functional stabilization of a large number of mutated gene products, hence is also called as the 'cancer chaperone'. While Hsp90 and the endoplasmic reticulum analogue, Grp94 were considered as molecular targets to interfere with cell proliferation and cellular protein homeostasis, studies pertinent to its mitochondrial analogue, TRAP1 are limited (Sreedhar et al., 2004).
Cancer is a polygenic disease emerging from the accumulation of multigenic mutations. The cancer atlas indicates that in addition to frame-shift mutations, deletions and additions, increased copy number variations (CNVs) correlating with increased single nucleotide polymorphisms (SNPs) in various genes may contribute to cancer progression (Liu et al., 2018). However, the CNV and SNP analyses with respect to Hsps are limited (Zhang et al., 2015). Against the earlier understanding that Hsps are intron-less, several reports indicated that Hsp90 family of proteins also exhibit SNP variations (Jolly et al., 1999; Sreedhar et al., 2004; Urban et al., 2012). Since Hsp expression correlates with aggressive cancer phenotypes, understanding the cross talk between CNVs and SNPs with respect to Hsps is important. (Liu et al., 2018; Mace et al., 2018).
3
Mitochondria are thought to have entered the eukaryotes by endoreduplication over millions of years ago. Subsequently, majority of mitochondrial genes are being integrated with the cellular genome (Andersson et al., 2003). Therefore, the cumulative impact of mutations in cancer cells can also have influence on the fate of mitochondrion. The Warburg effect in cancer cells facilitates the altered energy metabolism, however in the functional compromise of mitochondria, especially the oxidative phosphorylation (OXPHOS) (Warburg, 1924; Liberti and Locasale, 2016). Inhibiting Hsp90 functions predisposes cells to autophagy (Vishal et al., 2011), however inhibiting TRAP1 appears to influence both mitochondrial integrity and energy metabolism.
In this study, we performed CNV and SNP analyses of Hsp90 family of genes and presented the data representing their correlation with cancer incidences. We show that despite playing a key role in cancer progression, Hsp90 could not score CNV and SNP variations as much as it was scored for TRAP1. Subsequently, we correlated TRAP1 expression in established secondary tumor cells facilitating the altered energy metabolism. Our findings suggest TRAP1 playing a major role in metabolic reprogramming.
Materials and Methods Functional Classification of SNPs from Hsp90 isoforms The database of Single Nucleotide Polymorphisms (dbSNP) and The Cancer Genome Atlas (TCGA) hosts wide range of information on the genomic alterations that are reported in the patient specific cancer cell lines. The gene names are submitted as query to the dbSNP and the chromosomal report from dbSNP for the hsp90 genes were obtained using the Perl script through
4
their NCBI utilities tool. From the sequence database, the isoforms of Hsp90, HSP90AA1 (Hsp90α), HSP90AB1 (Hsp90β), HSP90B1 (Grp94) and TRAP1 (TRAP1) were scored for SNPs. The positional occurrence and the functional consequence of the SNPs are used to classify them under various functional classes.
The isoform-specific chromosomal mapping of functional SNPs The information on the functional class of SNPs and the alleles involved are obtained from the chromosomal report from the dbSNP. The chromosomal region corresponding to the exons were obtained from the genome data viewer of the NCBI. The chromosomal positions of each SNP were mapped to the corresponding exon, and the frequency of the type of allelic transition in each of the exon was calculated.
SNP prevalence by cancer phenotype The results obtained from the dbSNP were converted into ensembl Variant Effect Predictor (VEP) tool input format and were submitted to the VEP ensemble web interface for categorizing the SNPs into deleterious and non-deleterious based on the prediction algorithms, SIFT and PolyPhen. The SNPs that were marked deleterious were checked in TCGA for the cancer incidences.
Copy Number Variations assessment ClinVar database (http://www.ncbi.nlm.nih.gov/clinvar) was used for Hsp90 specific genomic variations for potential cancer incidences. Subsequently the copy number variations in different
5
Hsp90 isoforms were analyzed and the potential variations were scored by quantitative real time PCR.
Cell culture maintenance and treatments Human embryonic kidney cells HEK293 (ATCC® CRL-1573™), human lung carcinoma A549 (ATCC® CCL-185™), human neuroblastoma IMR32 (ATCC® CCL-127™) were obtained from American Type Culture Collection(ATCC). Cells were authenticated and maintained in Dulbecco's modified eagle medium (DMEM, Thermo Fisher, cat. no. 12491-023) with 10% fetal bovine serum(FBS, Thermo Fisher, cat no. 12483-020) at 37°C in CO2 incubator. IMR-32 cells were treated with rotenone to inhibit OXPHOS complex I (ROT, 2 µM, Sigma-Aldrich, cat. no.557368-1GM).To inhibit glycolysis, 2-Deoxy-D-glucose(2DG,1 mM, MP Biomedicals, LLC, cat. no.194642) was used. Cobalt chloride (CoCl2, 200 µM, Sigma-Aldrich, cat. no.D8375-1G) was used to induce chemical hypoxia. Cells grown in 6-well plates were treated with respective drugs before processing at specified time periods. All the drug treatments were done for 18-24 h intervals unless otherwise indicated.
TRAP1 knockdown system TRAP1(NM_016292.2) shRNA was designed using online software (BLOCK-it RNA designer, Invivogen
and
Biosettia).
The
shRNA
were
custom
synthesized(Sense:
5'-
TTTGGGGTTCCACTTCCAAACATGACGAATCATGTTTGGAAGTGGAACCCTTTTT-3', Antisense:5'CTAGAAAAAGGGTTCCACTTCCAAACATGATTCGTCATGTTTGGAAGTGGAACCC-3') and cloned in mu6Pro vector (kind gift from Dr. David L. Turner at University of Michigan, Ann
6
Arbor, USA). IMR32 cells were transfected with shRNA TRAP1, stable cells were selected with G418for 21 days.
TRAP1 over expression system The full length coding sequence of TRAP1 was retrieved (NM_016292.3), and PCR amplified from
MCF-7
cDNA
using
gene
specific
primers,
forward
5'-
ATTGAATTCATGGCGCGCGAGC-3'; reverse 5'-ATTGGATCCGTGTCGCTCCAGG-3' and cloned into ECoRI and BamHI restriction sites of pEGFPN2 expression system. Stable IMR-32 cells over expressing TRAP1 were used for subsequent experiments.
Chromosomal DNA isolation Genomic
DNA
was
isolated
with
minor
modification
of
the
protocol
(http://cancer.ucsf.edu/_docs/cores/array/protocols/dna_cell_culture.pdf). In brief, cells were trypsinized (0.05% Trypsin-EDTA) and washed with PBS. After centrifugation (465xg), pellet was resuspended in Tris.Cl buffer (pH 8.0) and subjected to sodium dodecyl sulphate (SDS, 0.66%) and Proteinase K (100 µg/mL) treatments, mixed gently by inversion and incubated overnight at 55°C. Saturated sodium chloride was added followed by ethanol at room temperature to precipitate chromosomal DNA. The DNA was spooled, rinsed twice in 70% ethanol, air dried, and resuspended in TE buffer, then subjected to Phenol:Choloroform:Isoamyl alcohol (25:24:1) extraction. The aqueous phase further subjected to Choloform:Isoamyl alcohol (25:24) extraction to remove phenol and ethanol precipitated.
7
Quantitative Polymerase Chain Reaction (qPCR) The genomic DNA sequence for TRAP1 was retrieved from UCSC genome browser. (URL: genome-asia.ucsc.edu; Chr:16 3658037-3717597, +ve strand). The qPCR primers were designed by online software (Prime3Plus). The primer sequences for TRAP1 (Chr16 3658037-3717597) and
ZNF80
(Chr3:114234636-114237578)
are,
for
TRAP1,
forward:
5'-
GTGCGGCAAAATAACGAACG-3' and reverse: 5'-GTTCGTTTGTAGCCACGTGT-3', and for
ZNF80,
forward:
5'-AAAACCTGTGAGTGCAGACC-3'
and
reverse:
5'-
CTGACGTCGGTTTGAGTTCA-3'.
Scanning confocal imaging microscopic analysis Cells (0.2x106 cells/mL) were grown on cover glasses (22x22 mm, Fisher Scientifics) were used for immunofluorescence. Cells first incubated with Mitotracker Red (CMX-Ros, 200 nM, Dojindo, cat no.R237) for 15 min in the incomplete medium, washed with PBS, fixed in 3.7% formaldehyde for 10 min, and permeabilized with 0.1% Triton x100 (Sigma-Aldrich, cat. no.T8787-100ML) for 7 min. For immunofluoresecnce, cells were incubated with 2%BSA (Sigma-Aldrich, cat. No.A2153 ) for 1 h, TRAP1 antibodies (1:200, abcam, cat no.ab182775) for overnight (4C), and subsequently incubated with Alexa Fluor 488 conjugated secondary antibodies (1:1000, abcam, cat no. ab150077) at room temperature for 1 h. The cover glasses were mounted using mounting medium containing DAPI (Vectashield,cat. no. H1200) and observed under microscope (SP8, Leica).
Immunoblot analysis
8
Cells were lysed in RIPA buffer (pH 7.4) and proteins (20 µg/well) were resolved on 10% SDSPAGE, then transferred onto nitrocellulose membrane. The membrane was blocked with 3% BSA for 1 h and incubated with TRAP1 antibodies (1:1000, 4C, 12 h, BD Biosciences, cat. no.612345) followed by HRP conjugated secondary antibodies (1:5000, RT, 1 h, Santacruz, cat. no.sc2005). The antigen-antibody interactions were scored using BM Chemiluminescence western blot kit as per the manufacturer's instructions (Roche, cat. no.11520709001).
Oxygen Consumption Rate (OCR) Cells (0.2x105/ 1.0x105/ 2.0x105) were grown on 6-well NUNC plates at standard culture conditions to achieve 60% confluence. Untreated cells were processed immediately, and cells after treatments were washed with PBS and resuspended in incomplete medium (100 µL). The basal level respiration was measured at different time intervals using Oxygraph 2K (Oroboros Instruments Corp, Austria). Medium without cells was used as a blank. The values obtained were plotted as line as well as a bar diagram. In case of 2DG and rotenone treatments, we measured the overall respiration unless otherwise these drugs interfere with stage-specific respiration.
Total ATP analysis The luminescence ATP detection kit (abcam, cat no. ab113849) was used for measuring total cellular ATP as per the manufacturer's instructions. In brief, cells (0.02x105/100 µL/1.0x105/ 100 µL/ 2.0x105/µL) in a 96-well plate were first subjected to lysis (50 µL), and then was added the substrate solution (50 µL) incubated in dark for 10 min. The change in luminescence with time was measured and represented in a bar diagram.
9
Preparation of cells for metabolite analysis Whole cells were harvested and prepared for loading. Briefly, a volume of culture containing 1x107 cells were harvested and cellular metabolism was quenched by the addition of a 5x volume of 60% methanol supplemented with 0.85% (W/V) ammonium bicarbonate at a minimum temperature of -40C. Quenched cells were pelleted by centrifugation and the pellet rinsed before being resuspended in 0.5 mL of 100% methanol, flash frozen on liquid nitrogen and when thawed, pelleted by centrifugation. The supernatant was retained and the extraction process repeated. The pooled 1 mL of extracted metabolites were dried and stored at -80°C until NMR sample preparation.
Metabolite analysis by 2D NMR Metabolites present in NMR samples were identified by comparison of related chemical shifts with published NMR data of known metabolites with reference to online resources (Madison Metabolomics Consortium Database (MMCD) and the Spectral Database for Organic Compounds
(SDBS)
http://riodb01.ibase.aist.go.jp/sdbs/(National
Institute
of
Advanced
Industrial Science and Technology, October-December 2010). Initial targets for metabolite identification were drawn from the known metabolite components of DMEM medium, and subsequent cross comparisons made to the CD-CHO medium. All spectra were recorded at 37°C on a Unity INOVA 600 MHz NMR spectrometer (Varian Inc, PaloAlto, CA, USA) fitted with a 5 mm triple resonance HCN probe with z-shield gradients. 1D 1H spectra with WATERGATE solvent suppression were recorded over 32768 points, 512 transients and a spectral width of 8004.8 Hz. An acquisition time of 2.047 s and a relaxation delay of 1.5 s were used. This provides an overall excitation recovery time of 3.547 s and produces a total experimental time of
10
31min that enables efficient and fast data acquisition with minimal effect due to relaxation; 1H T1 values up to 1.5 s will experience an effective magnetization recovery greater than 90%. Data processing was performed by ACD Labs SpecManager V9.0 involved zero filling to 65536 points and multiplication of 1 Hz exponential line broadening before Fourier transformation.
Results Functional classification of SNPs from Hsp90 isoforms The SNP variation has been linked to disease outcome thus; proper identification, functional assignment, and validation play a key role in understanding their functional relevance (Batnyam et al., 2013). Principally, SNPs are classified based on their positional occurrences as intronvariant, splice acceptor variant, UTR-3 prime, UTR-5 prime, downstream –500bp, splice-donor, upstream –2 Kb. Based on their functional consequences, they are classified into synonymous, missense, frameshift, stop-gain, stop-lost. Interestingly, most of the SNPs are reported as intronvariant followed by missense variants in the exons, compared to other functional classes (Figure 1a). Further, Hsp90AA1 and TRAP1 showed enhanced intron-variants compared to Hsp90AB1 and Hsp90B1. Between Hsp90AA1 and TRAP1 the latter exhibited more number of SNPs. While Hsp90AA1 exhibited the variance in the order of C/T < A/G < C/G < A/C followed by others, TRAP1 exhibited the variance in the order of A/G < C/T < C/G < A/C followed by others. The other SNPs are found to be in the upstream regions i.e. above 2 Kb region. These results suggested us that A/G and C/T variations are predominant (Figure 1b). In gross terms, intron variations contribute to disease susceptibility, the prevalence Hsp90 variations may have role in cancer progression (Cooper, 2010).
11
The isoform specific chromosome mapping of functional SNPs and their prevalence by cancer phenotype The mutations in exonic regions usually will not have direct influence on gene transcription or functions of the transcripts hence considered silent despite their non-synonymous (missense) nature. However, these mutations can disrupt the splice sites hence can influence the pathogenic outcome (Wu and Hurst, 2016). The missense mutations on each exon of Hsp90 isoforms were scored and represented in the bar diagram. Hsp90 isoforms, Hsp90AA1 has 10 exons, Hsp90AB1 has 18 exons, Hsp90AB1 has 12 exons and TRAP1 has 18 exons. Interestingly, all of them exhibited some amount of missense mutations and without any specific pattern or predominance (Figure 2a). These findings suggested that unless and otherwise exon variants are validated for functional relevance with respect to cancer phenotype(s), the significance of these mutations (allelic transitions) cannot be extrapolated.
We could retrieve 16549, 3695, 4677, 20693 SNPs from the database for HSP90AA1, HSP90AB1, HSP90B1, TRAP1 respectively for Homo sapiens. Among these, 4, 3, 13 and 12 SNPs respectively were reported to be deleterious and are associated with cancer incidences (as per SIFT and PolyPhen scores, Figure 2b; Table 1). Intriguingly, it was found that compared to SNPs from TRAP1and HSP90B1, the SNPs from Hsp90AA1 and Hsp90AB1 have correlated with lower number of cancer incidences. Till date, TRAP1 is less implicated in cancer incidences due to limited studies pertinent to its contribution in cancer progression. However, the present findings suggest that TRAP1 may play significant role in the pathogenic outcome of cancer.
Copy number variation (CNV) as a predictive marker for disease assessment
12
Cancer is a polygenic disease, where tumor heterogeneity plays a significant role in disease progression (Meacham and Morrison, 2013). Similar to exon versus intron variations that reveal population differences (Zhan et al., 2014), SNPs within the tumor may provide information related to tumor heterogeneity. While SNPs themselves are considered to be a small variation with respect to gene expression pools that decides the cell fate or disease outcome, we have examined for the CNV that provides information related to gain or loss of segments of DNA (de Koning et al., 2011). We have used transformed, malignant and metastatic cancer cells, HEK293T, A549, and IMR-32 cells respectively to understand TRAP1 specific correlation with cancer phenotypes. While transformed and malignant cells showed lower CNVs, there is a significant increase observed in metastatic cells (Figure 2c).
Correlation of TRAP-1 expression with cellular energy metabolism Correlating with CNV data, we also observed enhanced TRAP1 protein expression in phenotype specific manner. There was a gradual increase in TRAP1 expression HEK293T < A549 < IMR32 cells. Although TRAP1 mitochondrial localization is not clear in IMR-32 cells, in HEK-293 and A549 cells, TRAP1 showed strong mitochondrial co-localization (Figure 3a). Subsequently, the immunoblot analysis showed its increased expression in A549 and IMR-32 cells (Figure 3b).
Mitochondria are the major source of ATP production (Bratic and Trifunovic, 2010). Cancer cells are known to bypass conventional energy metabolism due to their selection pressure and limited oxygen supply (Warburg, 1924). Hence, we examined hypoxia (using cobalt chloride) and OXPHOS inhibition (using rotenone) effects on TRAP1 expression in IMR-32 cells. While a marginal increase in TRAP1 expression was observed in response to hypoxia, rotenone treatment
13
has significantly decreased its expression (Figure 3c). This suggest that drug induced uncoupling may work independent of TRAP1. Next we examined whether TRAP1 differential expression in HEK293T, A549 and IMR-32 cells differentially regulates mitochondrial functions. Compared to HEK293T cells, A549 and IMR-32 cells showed enhanced ATP levels (Figure 3d). Subsequently, we observed decreased oxygen consumption rate (OCR) in IMR-32 cells compared to A549 and HEK293T cells (Figure 3e and 3f).
Correlation of TRAP-1 expression with cellular energy metabolism To understand the involvement of TRAP1 in mitochondrial function and energy metabolism, custom synthesized shRNA TRAP1 was stably expressed in IMR-32 cells (Figure 4a and 4b). These cells were used to measure OCR and ATP levels in comparison with parental cells, and we observed TRAP1 KD cells displaying decreased OCR. Next, IMR-32 cells stably over expressing the full-length TRAP1 was compared with KD cells. The TRAP1 OE cells showed further decrease in OCR compared to KD cells (Figure 4c). To understand the effect of 2DG and rotenone on basal level of respiration of parental, TRAP1 KD and TRAP1 OE cells, they were treated with the drugs and the rate of respiration was measured. TRAP1 KD cells are found to be sensitive to 2DG treatment compared to parental and TRAP1 OE cells (Figure 4e versus 4d and 4f). However, rotenone treatment has decreased respiration in parental and TRAP1 OE cells (Figure 4d and 4f). These results suggested that mitochondrial OXPHOS is minimized in the TRAP1 expression background. Subsequently, we have measured total cellular ATP levels in parental, TRAP1 KD and TRAP1 OE cells either alone or in response to pharmacological interventions of glycolysis (using 2DG) and OXPHOS (using rotenone). The 2DG treatment resulted in significant decrease in ATP levels in TRAP1 KD cells, but not in parental cells.
14
Whereas, rotenone treatment showed no significant influence on total ATP levels in all the cell types studied (Figure 4g, 4h and 4i). The TRAP1 OE cells exhibited similar response like in parental cells with respect to 2DG treatment indicating that in the TRAP1 background, inhibiting glucose oxidation had no effect on ATP production (Figure 4i). To understand both source and secondary metabolite status of these cells, NMR based metabolite analysis was carried out in parental, TRAP1 KD and TRAP1 OE cells. We observed a moderate increase in intracellular glutamine and glutamate levels in KD cells in comparison with parental cells (Figure 4j and 4k). In contrary, a drastic increase (nearing 4 folds) in these secondary metabolites was observed in TRAP1 OE cells (Figure 4l). These results suggest that in TRAP1 OE background, these metabolites are probably accumulated from the alternate energy metabolism.
15
Discussion The Cellular proteotoxicity contributes to a large number of pathological disorders including cancer (McConkey, 2017). Cells evolve with mechanisms to overcome proteotoxicity by increasing the expression of molecular chaperones to regulate cellular protein homeostasis. Hsps having the chaperoning functions clears the proteotoxicity by not only refolding the damaged proteins, but also facilitating the mutated/altered proteins to function normally. Thus, Hsps are attributed to various adaptations (Feder and Hofmann, 1999). The Hsp90 chaperone is considered as a cancer chaperone consequently its inhibition has emerged as a novel antitumor strategy to combat cancer (Solit and Rosen, 2006). However, its mitochondrial homologue, TRAP1 functions are not fully explored. There are reports that TRAP1 expression correlates with tumor aggression, however lacks conclusive experimental evidences. Since TRAP1 is a mitochondrial resident, and targeting the mitochondrial metabolism is emerging as a strategy, we examined its role in energy metabolism correlating with disease progression.
The chaperone, Hsp90 is in the centre of protein network forming the foldosome complex, thus, is critical for cellular protein stability, in addition is also involved in the functional maturation of mutated gene products (Csermely et al., 1998; Sreedhar et al., 2004; Kishinevsky et al., 2018). To a large extent, its homologue in the ER, Grp94 was also explored for protein quality control and antigen presentation (Argon et al., 2016). Despite several individual studies related to SNPs in Hsp90 family of genes, their clinical relevance with respect to cancer progression is not known. We asked whether SNPs analysis followed by CNVs unravels the isoform specific functions of Hsps. From the analyses we could confirm that among Hsp90 genes, TRAP1 exhibits more number of SNPs and CNVs (on chromosome 16) compared to Hsp90 isoforms
16
HSP90AA, HSP90AB1 and HSP90B1. It has been shown that chromosome 16 duplications are already observed and suggested to have correlation with disease susceptibility (Martin et al., 2004). While SNPs provided the firsthand information on TRAP1 functional relevance with cancer, the CNVs data confirmed that unless until there is a need for biochemical/phenotypic reformation, the increase in copy number may not be required.
We hypothesized that similar to Hsp90, TRAP1 may be involved in tumor progression, thus, can also be considered as a cancer chaperone. Till date there is no such information available that mitochondrial proteins directly contribute to cancer. In support of this, it was also proposed that cancer cells prefer alternate energy metabolism by bypassing the OXPHOS (Bratic, 2010). There were limited studies that correlate TRAP1 expression with altered energy metabolism (Matassa et al., 2018) and neoplastic transformation (Sciacovelli et al., 2013). It is established that hypoxia decreases mitochondrial OCR and thus affects the cellular ATP production. In our model, the basal levels of TRAP1 are high enough hence subjecting these cells to OXPHOS inhibition did not influence mitochondrial functions indicating that cells with enhanced TRAP1 expression may have adapted to sustain bypassing effects of OXPHOS. Strengthening this, TRAP1 OE cells also showed similar non-responsiveness to OXPHOS inhibition. Interestingly, the metabolite analyses revealed that there is no difference in the net glucose levels in parental cells compared to TRAP1 KD cells. However, there was a significant increase in glutamate and glutamine levels in the KD cells. Further, TRAP1 KD has sensitized these cells to glycolysis inhibition and thus to ATP production indicating that TRAP1 is probably involved in the metabolic reprogramming of cancer cells. Intriguingly, TRAP1 OE cells showed nearly 4 folds increases in glucose, glutamine and glutatmate levels indicating cellular adaptations to alternate energy metabolism.
17
The metabolic reprogramming is now considered as one of the hallmarks of cancer (Hanahan and Weinberg, 2011; Li and Le, 2018). Although it was thought that targeting mitochondria is sufficient to target cancer (Skoda et al., 2018), extrapolating studies from Warburg (Warburg, 1924) unraveled several alternate pathways, thus metabolite sensing is suggested to be an important parameter for deciding the therapeutic outcome (Han and Patten, 2018; Papa et al., 2018). In the context that metabolite targeting acts as an efficient tumor suppressor mechanism, our studies by unraveling TRAP1 involvement in metabolic reprogramming suggests that this mitochondrial chaperone can be a potential pharmacological target to combat cancer metabolism.
Acknowledgements Authors thank Dr. Anant Patel for helping us with NMR experiments. Authors thank Science and Engineering Research Board (SERB), Department of Science and Technology, Ministry of Science, Government of India for the financial support. Mr. Shrikant is supported by fellowship from CSIR for his PhD.
Conflict of Interest: Authors declare the conflict of interest as NONE.
References Andersson, S.G., Karlberg, O., Canback, B., Kurland, C.G., 2003. On the origin of mitochondria: a genomics perspective, Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences. 358, 165–179. 18
Argon, Y., Marzec, M., Hawkes, C.P., Eletto, D., Boyle, S., Rosenfeld R., Hwa, V., Wit, J.M., van Duyvenvoorde, H.A., Oostdijk, W., Losekoot, M., Pedersen, O., Yeap, B.B., Flicker, L., Barzilai, N., Atzmon, G., Grimberg, A., 2016. A Human Variant of Glucose-Regulated Protein 94 That Inefficiently Supports IGF Production, Endocrinology, 157,1914–1928.
Batnyam, N., Gantulga, A., Oh, S., 2013. An Efficient Classification for Single Nucleotide Polymorphism (SNP) Dataset. Computer and Information Science. Studies in Computational Intelligence. 493, 171-185.
Bratic, I., Trifunovic, A., 2010, Mitochondrial energy metabolism and ageing, Biochim. Biophys. Acta. 1797, 961-967.
Cooper, D.N., 2010. Functional intronic polymorphisms: Buried treasure awaiting discovery within our genes. Human Genomics. 4, 284-288.
Csermely, P., Schnaider, T., Csaba, S., Prohaszka, Z., Nardai, G., 1998. The 90-kDa Molecular Chaperone Family: Structure, Function, and Clinical Applications. A Comprehensive Review. 79, 129-168.
de Koning, A.P.J., Gu, W., Castoe, T.A., Batzer, M.A., Pollock, D.D., 2011. Repetitive Elements May Comprise Over Two-Thirds of the Human Genome. PLOS Genetics. 7,1-12.
19
Feder, M.F., Hofmann, G.E., 1999. Heat Shock Proteins, Molecular Chaperones, and The Stress Response. Evolutionary and Ecological Physiology, Annu. Rev. Physiol. 61,243–282.
Han, C.Y., Patten, D.A., Richardson, R.B., Harper, M.E., Tsang, B.K. 2018. Tumor metabolism regulating chemosensitivity in ovarian cancer. Genes Cancer. 9,155-175.
Hanahan, D., Weinberg, R.A., 2011. Hallmarks of Cancer: The Next Generation, Cell. 144,646674.
Jolly, C., Vourc’h, C., Robert-Nicoud, M., Morimoto, R.I., 1999. Intron-independent association of splicing factors with active genes. J Cell Biol. 145, 1133-1143.
Katschinski, D.M., 2004. On heat and cells and proteins; News Physiol. Sci. 19,11–15.
Kishinevsky, S., Wang, T., Rodina, A., Chung, S.Y., Xu, C., Philip, J., Taldone, T., Joshi, S., Alpaugh, M.L., Bolaender, A., Gutbier, S., Sandhu, D., Fattahi, F., Zimmer, B., Shah, S.K., Chang, E., Inda, C., Koren, J., Saurat, N.G., Leist, M., Gross, S.S., Seshan, V.E., Klein, C., Tomishima, M.J., Erdjument-Bromage, H., Neubert, T.A., Henrickson, R.C., Chiosis, G., Studer, L. 2018. HSP90-incorporating chaperome networks as biosensor for disease-related pathways in patient-specific midbrain dopamine neurons. Nature communications. 9,1-15.
Li, T., Le, A., 2018. Glutamine Metabolism in Cancer. Adv Exp Med Biol. 1063, 13-32.
Liberti, M.V., Locasale, J.W., 2016. The Warburg Effect: How Does it Benefit Cancer Cells? Trends Biochem Sci. 41,211-218. 20
Liu, J., Zhou, Y., Liu, S., Song, X., Yang, X.Z., Fan, Y., Chen, W., Akdemir, Z.C., Yan, Z., Zuo, Y., Du, R., Liu, Z., Yuan, B., Zhao, S., Liu, G., Chen, Y., Zhao, Y., Lin, M., Zhu, Q., Niu, Y., Liu, P., Ikegawa, S., Song, Y.Q., Posey, J.E., Qiu, G., 2018. The coexistence of copy number variations (CNVs) and single nucleotide polymorphisms (SNPs) at a locus can result in distorted calculations of the significance in associating SNPs to disease. Hum Genet. 137,553-567.
Mace, A., Kutalik, Z., Valsesia, A., 2018. Copy Number Variation. Methods Mol Biol.1793,231258.
Martin, J., Han, C., Gordon, L.A, Terry, A., Prabhakar, S., She, X., Xie, G., Hellsten, U.,Chan, Y,M., Altherr, M., Couronne, O., Aerts, A., Bajorek, E., Black, S., Blumer, H., Branscomb, E., Brown, N.C., Bruno, W.J., Buckingham, J.M., Callen, D.F., Campbell, C.S., Campbell, M.L.,Campbell, E.W., Caoile, C., Challacombe, J.F., Chasteen, L.A., Chertkov, O., Chi, H.C.,Christensen, M., Clark, L.M., Cohn, J.D., Denys, M., Detter, J.C., Dickson, M.,Dimitrijevic-Bussod, M., Escobar, J., Fawcett, J.J., Flowers, D., Fotopulos, D., Glavina, T., Gomez, M., Gonzales, E., Goodstein, D., Goodwin, L.A., Grady, D.L., Grigoriev, I., Groza, M.,Hammon, N., Hawkins, T., Haydu, L., Hildebrand, C.E., Huang, W., Israni, S., Jett, J., Jewett, P.B., Kadner, K., Kimball, H., Kobayashi, A., Krawczyk, M.C., Leyba, T., Longmire, J.L., Lopez, F.,Lou, Y., Lowry, S., Ludeman, T., Manohar, C.F., Mark, G.A., McMurray, K.L., Meincke, L.J., Morgan, J., Moyzis, R.K., Mundt, M.O., Munk, A.C., Nandkeshwar, R.D., Pitluck, S., Pollard, M., Predki. P., Parson-Quintana, B., Ramirez, L., Rash, S., Retterer, J., Ricke, D.O., Robinson, D.L.,Rodriguez, A., Salamov, A., Saunders, E.H., Scott, D., Shough, T., Stallings, R.L., Stalvey M., Sutherland, R.D., Tapia, R., Tesmer, J.G., Thayer, N., Thompson,
21
L.S., Tice, H., Torney, D.C.,Tran-Gyamfi, M., Tsai, M., Ulanovsky, L.E.,Ustaszewska, A., Vo, N., White, P.S., Williams, A.L., Wills, P.L., Wu, J.R., Wu, K., Yang, J., Dejong, P., Bruce, D., Doggett, N.A., Deaven, L., Schmutz. J., Grimwood, J., Richardson. P., Rokhsar, D.S., Eichler, E.E., Gilna, P., Lucas, S.M., Myers, R.M.,Rubin. E.M., Pennacchio, L.A., 2004. The sequence and analysis of duplication-rich humanchromosome 16. Nature. 432,988-994.
Matassa, D.S., Agliarulo, I., Avolio, R., Landriscina, M., Esposito, F., 2018. TRAP1 Regulation of Cancer Metabolism: Dual Role as Oncogene or Tumor Suppressor. Genes.5, 9.
McConkey, D.J., 2017. The integrated stress response and proteotoxicity in cancer therapy. Biochem Biophys Res Commun. 482,450-453.
Meacham, C.E., Morrison, S.J., 2013. Tumour heterogeneity and cancer cell plasticity. Nature. 501,328-337.
Neckers, L., 2007. Heat shock protein 90: the cancer chaperone. J Biosci. 32,517-530.
Papa, S., Choy, P.M., Bubici, C. 2018. The ERK and JNK pathways in the regulation of metabolic reprogramming.Oncogene. ISSN 1476-5594 (online).
Sciacovelli, M., Guzzo, G., Morello, V., Frezza, C., Zheng, L., Nannini, N., Calabrese, F., Laudiero, G., Esposito, F., Landriscina, M., Defilippi, P., Bernardi, P., Rasola, A., 2013. The mitochondrial chaperone TRAP1 promotes neoplastic growth by inhibiting succinate dehydrogenase. Cell Metabolism. 17, 988-999. 22
Skoda, J., Borankova, K., Jansson, P.J., Huang, L.H.M., Veselska, R., Richardson, D.R., 2018. Pharmacological targeting of mitochondria in cancer stem cells: An ancient organelle at the crossroad of novel anti-cancer therapies.Pharmacol Res. 139, 298-313.
Solit, D.B., Rosen, N., 2006. Hsp90: A Novel Target for Cancer Therapy. Current Topics in Medicinal Chemistry. 6, 1205-1214.
Sreedhar,
A.S.,
Kalmar,
E.,
Csermely,
P.,
Shen,
Y.F.,
2004.
Hsp90
isoforms:
functions,expression and clinical importance. FEBS Lett. 562,11-15.
Urban, J.D., Budinsky, R.A., Rowlands, J.C., 2012. An evaluation of single nucleotide polymorphisms in the human heat shock protein 90 kDa α and β isoforms. Drug Metab Pharmacokinet. 27,268-278.
Vishal, C., Kumar, J.U., Brahmendra Swamy, C.V., Nandini, R., Srinivas, G., Kumaresan, R., Sreedhar, A.S., 2011. Repercussion of Mitochondria Deformity Induced by Anti-Hsp90 Drug 17AAG in Human Tumor Cells. Drug Target Insights. 5,11-32.
Warburg, O., Posener, K., Negelein, E., 1924. üeber den Stoffwechsel der Tumoren. Biochem Z 152,319–344. Reprinted in English in the book On metabolism of tumors by O. Warburg, Constable. London, 1930.
23
Wu, X., Hurst, L.D., 2016. Determinants of the Usage of Splice-Associated cis-Motifs Predict the Distribution of Human Pathogenic SNPs. Mol Biol Evol. 33,518-529.
Zhan, X., Dixon, A., Batbayar, N., Bragin, E., Ayas, Z., Deutschova, L., Chavko, J., Domashevsky, S., Dorosencu, A., Bagyura, J., Gombobaatar, S., Grlica, I. D., Levin, A., Milobog, Y., Ming, M., Prommer, M., Purev-Ochir, G., Ragyov, D., Tsurkanu, V., Vetrov, V., Zubkov, N., Bruford, M. W.2014. Exonic versus intronic SNPs: contrasting roles in revealing the population genetic differentiation of a widespread bird species. Heredity, 114,1-9. (2014).
Zhang, F., Gu, W., Hurles, M.E., Lupski, J.R. 2009. Copy number variation in human health, disease, and evolution. Annual Review of genomics and human genetics, 10, 451-481.
24
GENE
HSP90AA1
HSP90AB1
HSP90B1
TRAP1
CHR
14
SNP locus
dbSNP ID
Exon No.
Allele in reference assembly
Minor allele
Amino acid Transition
Cancer type
102085425
rs745469368
8
G
C
P301R
Adenomas and Adenocarcinomas; Cystic, Mucinous and Serous Neoplasms
102085341
rs945693652
8
A
G
V329A
Nevi and Melanomas
102086091
rs747407643
9
C
A
D188Y
Transitional Cell Papillomas And Carcinomas
102085341
rs945693652
8
A
G
V329A
Skin Cutaneous Melanoma
44248741
rs1331943025
2
A
G
I38V
Adenomas and Adenocarcinomas; Cystic, Mucinous and Serous Neoplasms
44251789
rs920937641
9
G
A/T
R456H, R456L
Breast Invasive Carcinoma
44250062
rs760864610
5
AAA
-
K186del
Uterine Corpus Endometrial Carcinoma
103943816
rs3209749
14
A
C
N657H
Squamous Cell Neoplasms
6
Adenomas and Adenocarcinomas, Epithelial Neoplasms, NOS, Squamous Cell Neoplasms,Transitional Cell Papillomas and Carcinomas Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous Neoplasms, Epithelial Neoplasms. NOS Colon Adenocarcinoma, Stomach Adenocarcinoma Head and Neck Squamous Cell Carcinoma Head and Neck Squamous Cell Carcinoma
103939608
rs1354079125
8
T
C
Y359H
103941439
rs1385298900
9
C
G
H374Q
103941638
rs368374645
10
C
T
R414C
103943204
rs762366467
13
C
G
A592G
103946665
rs372636341
15
C
T
P692L
103932846
rs1362809151
4
T
G
I105M
Glioblastoma Multiforme
103942642
rs774797218
12
T
C
I497T
Lung Adenocarcinoma
103943778
rs754939231
14
C
T
P644L
103932375
rs571691885
3
G
A
R84K
103946849
rs770752783
16
C
A/T
R724R, R724W
103938403
rs1445926074
7
G
A
D307N
103943757
rs1277290411
14
C
T
S637F
3674486
rs1307786417
10
C
A
W299C,W159C
3679729
rs1272474481
14
G
A/C
S178W, S125L, S125W,S38L,S38W
12
16
Liver Hepatocellular Carcinoma Uterine Corpus Endometrial Carcinoma Stomach Adenocarcinoma Bladder Urothelial Carcinoma Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous Neoplasms, Squamous Cell Neoplasms Cystic, Mucinous and Serous Neoplasms Nevi and Melanomas, Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous
25
3690838
rs1288142880
17
T
C
3664360
rs113510154
6
G
A/C
3674346
rs1422474678
10
G
C
3662060
rs192869535
3
G
A
3686127
rs192245771
15
G
A
R114W, R61W
Stomach Adenocarcinoma
T
E643K, E590K, E503K
Skin Cutaneous Melanoma
R658C
Colon Adenocarcinoma
3662000
rs145014945
3
C
E79G
Neoplasms, Epithelial Neoplasms, NOS Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous Neoplasms, Epithelial Neoplasms, NOS
R495C, R495G, R442C, R442G, R355C, R355G P346R, P206R, P293R R623C, R483C, R570C
Colon Adenocarcinoma Stomach Adenocarcinoma Uterine Corpus Endometrial Carcinoma
Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma Uterine Corpus Endometrial Carcinoma
3658834
rs139636268
2
G
A/T
3672727
rs1210505741
9
G
-
L380fs (L380C), L327C, L240C
3672762
rs375364304
9
C
T
R368H
Uterine Corpus Endometrial Carcinoma
3674464
rs13926
10
G
C,T
R254S, R254G
Multiple Myeloma
R658S
Figure Legends Figure 1. The functional classification of SNPs indifferent Hsp90 isoforms. a. SNPs reported in the isoforms of HSP90 are retrieved from dbSNP. The SNPs predominantly found in the intronic regions were scored and represented in the bar diagram. Note that HSP90AA1 (Hsp90α) and TRAP1 (TRAP1) showing more variations compared to HSP90AB1 (Hsp90β) and HSP90B1 (Grp94). b. The frequency of allelic transitions reported under each functional class for the HSP90 isoforms were scored and represented in the bar diagram. Note predominant allelic transitions in HSP90AA1 and TRAP1.
Figure 2. Allelic transitions based on exon number. a. Isoform specific chromosomal mapping of functional SNPs reported by the dbSNP. The number of transition mutations is more predominant in all the isoforms compared to transversion mutations. b. Cancer phenotype 26
specific SNP prevalence. The cancer incidences reported with missense mutations in the exonic regions for each isoform of HSP90 were scored and represented in the bar diagram. most of the cancer incidences are reported in the TRAP1 and HSP90B1 compared with the other isoforms. c. Copy number variation (CNV) versus cancer phenotype. Quantitative real time PCR analysis of CNV in TRAP1 in HEK293, A549, and IMR-32 cells. The TRAP1 PCR values are normalized using ZNF80. Note IMR-32 cells showing increase in CNV.
Figure 3. Correlation of TRAP-1 expression with altered cellular energy metabolism. a. Cytoimmunofluorescence analyses of TRAP1 localization in the mitochondrion. The mitochondria from HEK293T, A549 and IMR-32 cells were stained with CMX-Ros (red) and examined for TRAP1 (stained with Alexa fluor 488 (green)) colocalization. The nucleus was stained with DAPI (blue). Note low mitochondrial localization in IMR-32 cells despite increased levels of total TRAP1. Scale bar represents 25 µm. b. Immunoblot analysis of TRAP1 in HEK293T, A549 and IMR-32 cells. The numbers represents the densitometry values obtained for gene/actin ratio. c. Effect of CoCl2 and rotenone on TRAP1 expression. RT-PCR analysis of TRAP-1 expression. d. Measurement of total cellular ATP levels. The average luminescence values obtained were interpolated with the ATP standards and represented. Note that A549 and IMR-32 cells showed enhanced ATP levels. e. Analysis of oxygen consumption rate (OCR) in different cell types. Spectral analysis of OCR observed over time obtained from Oxygraph. f. statistical representation of OCR data. The data acquired were from three independent experiments. Values represented are Mean SD. The significance values are, * p<0.05, ** p<0.01, *** p<0.001.
27
Figure 4. Effect of TRAP1 knockdown on altered energy metabolism. a, b. custom synthesized shRNA TRAP1 (1 and 2 µg/ transfection) was used for making stable TRAP1 KD cells. c. The OCR analysis of parental, TRAP1 KD and TRAP1 OE cells. Note decrease in OCR in IMR-32 cells. Measurement of cellular respiration in parental (d), TRAP1 KD (e) and TRAP1 OE (f) cells. Total cellular ATP levels in parental (g), TRAP KD (h) and TRAP1 OE (i) cells in response to 2DG and rotenone. Metabolite analysis using NMR spectra of total cell lysates from parental (j), TRAP1 KD (k) and TRAP1 OE (l) cells. Note significant increase in glucose, glutamine and glutamate levels in TRAP1 OE cells.
28
29
30
31
32
Highlights
TRAP1 exhibits more number of SNPs in cancer cells compared to Hsp90
TRAP1 exhibits increased copy number variations (CNVs) in metastatic cancer cells
TRAP1 CNVs correlate with increased expression in aggressive cancer cells
TRAP1 appears to facilitate alternate energy metabolism
33
S1567-7249(19)30019-4 https://doi.org/10.1016/j.mito.2019.09.011 MITOCH 1417
To appear in:
Mitochondrion
Received Date: Revised Date: Accepted Date:
31 January 2019 16 September 2019 18 September 2019
Please cite this article as: Ramkumar, B., Dharaskar, S.P., Guntipally, M., Paithankar, K., Sreedhar, A.S., Mitochondrial chaperone, TRAP1 as a potential pharmacological target to combat cancer metabolism, Mitochondrion (2019), doi: https://doi.org/10.1016/j.mito.2019.09.011
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Elsevier B.V. and Mitochondria Research Society. All rights reserved.
Mitochondrial chaperone, TRAP1 as a potential pharmacological target to combat cancer metabolism Balaji Ramkumar1, Shrikant P. Dharaskar1,2,#, Mounika Guntipally1,#, Khanderao Paithankar1, Amere Subbarao Sreedhar1*.
1CSIR-Centre
for Cellular and Molecular Biology, Uppal Road, Hyderabad, 500 007, Telangana,
India 2Academy
of Scientific & Innovation Research, Government of India.
# Authors have equally contributed to the manuscript * Correspondence: AS Sreedhar, [email protected] Running Title: TRAP1 is a potential pharmacological target to combat cancer metabolism Key words: Hsp, TRAP1, cancer, mitochondria, metabolism.
1
Abstract The stress response forms the most ancient defense system in living cells. Heat shock proteins (Hsps) are highly conserved across species and play major roles in mounting the stress response. The emerging information now suggests that Hsp90 family of chaperones display additional cellular roles contributing to diseases like cancer. For this reason, pharmacological targeting of Hsp90 has emerged as a novel antitumor strategy. However, its mitochondrial homologue TRAP1 has not been implicated in cancer with conclusive mechanistic insights. Since understanding the mutational spectrum of cancer cells indicates the outcome of the disease as well as treatment response, we examined mutational spectrum of TRAP1. Our in silico analyses of TRAP1 SNPs and CNVs correlated with the aggressive cancer phenotypes, and are found to be predominant over Hsp90 itself. The increased CNVs have been correlated with increased expression of TRAP1 in metastatic cancer cells, increased ATP production, and decreased oxygen consumption rate of mitochondria. Examining TRAP1 knockdown as well as over expression in metastatic cancer cells furthered our understanding that TRAP1 likely to facilitate the altered energy metabolism in the functional compromise of mitochondrial OXPHOS. Interestingly, the increased ATP levels in the TRAP1 background are found to be independent of glucose oxidation. Our results suggest TRAP1 role in triggering the alternate energy metabolism in cancer cells. Since targeting tumor metabolism is considered as an alternate strategy to combat cancer, we propose pharmacological targeting of TRAP1 to target alternate energy metabolism.
Introduction The stress response is highly conserved across the species from archaebacteria to higher eukaryotes, where heat shock proteins (Hsps) form the most ancient defense system (Feder and
2
Hofmann, 1999). Having chaperoning functions, Hsps contribute to the maintenance of cellular protein homeostasis under normal physiological conditions, in addition, exhibits cytoprotective functions in response to various extra- or intra-cellular stressors (Katschinski, 2004). Mounting evidence suggests that among Hsps, the Hsp90 chaperone contribute to various pathological disorders by facilitating mutated proteins to function normally (Neckers, 2007). This could be the reason why high Hsp expression correlates with cancer, a pathological disease known for the accumulation of wide spectrum of mutations. Further, Hsp90 gained much attention due to its role in the conformational maturation and functional stabilization of a large number of mutated gene products, hence is also called as the 'cancer chaperone'. While Hsp90 and the endoplasmic reticulum analogue, Grp94 were considered as molecular targets to interfere with cell proliferation and cellular protein homeostasis, studies pertinent to its mitochondrial analogue, TRAP1 are limited (Sreedhar et al., 2004).
Cancer is a polygenic disease emerging from the accumulation of multigenic mutations. The cancer atlas indicates that in addition to frame-shift mutations, deletions and additions, increased copy number variations (CNVs) correlating with increased single nucleotide polymorphisms (SNPs) in various genes may contribute to cancer progression (Liu et al., 2018). However, the CNV and SNP analyses with respect to Hsps are limited (Zhang et al., 2015). Against the earlier understanding that Hsps are intron-less, several reports indicated that Hsp90 family of proteins also exhibit SNP variations (Jolly et al., 1999; Sreedhar et al., 2004; Urban et al., 2012). Since Hsp expression correlates with aggressive cancer phenotypes, understanding the cross talk between CNVs and SNPs with respect to Hsps is important. (Liu et al., 2018; Mace et al., 2018).
3
Mitochondria are thought to have entered the eukaryotes by endoreduplication over millions of years ago. Subsequently, majority of mitochondrial genes are being integrated with the cellular genome (Andersson et al., 2003). Therefore, the cumulative impact of mutations in cancer cells can also have influence on the fate of mitochondrion. The Warburg effect in cancer cells facilitates the altered energy metabolism, however in the functional compromise of mitochondria, especially the oxidative phosphorylation (OXPHOS) (Warburg, 1924; Liberti and Locasale, 2016). Inhibiting Hsp90 functions predisposes cells to autophagy (Vishal et al., 2011), however inhibiting TRAP1 appears to influence both mitochondrial integrity and energy metabolism.
In this study, we performed CNV and SNP analyses of Hsp90 family of genes and presented the data representing their correlation with cancer incidences. We show that despite playing a key role in cancer progression, Hsp90 could not score CNV and SNP variations as much as it was scored for TRAP1. Subsequently, we correlated TRAP1 expression in established secondary tumor cells facilitating the altered energy metabolism. Our findings suggest TRAP1 playing a major role in metabolic reprogramming.
Materials and Methods Functional Classification of SNPs from Hsp90 isoforms The database of Single Nucleotide Polymorphisms (dbSNP) and The Cancer Genome Atlas (TCGA) hosts wide range of information on the genomic alterations that are reported in the patient specific cancer cell lines. The gene names are submitted as query to the dbSNP and the chromosomal report from dbSNP for the hsp90 genes were obtained using the Perl script through
4
their NCBI utilities tool. From the sequence database, the isoforms of Hsp90, HSP90AA1 (Hsp90α), HSP90AB1 (Hsp90β), HSP90B1 (Grp94) and TRAP1 (TRAP1) were scored for SNPs. The positional occurrence and the functional consequence of the SNPs are used to classify them under various functional classes.
The isoform-specific chromosomal mapping of functional SNPs The information on the functional class of SNPs and the alleles involved are obtained from the chromosomal report from the dbSNP. The chromosomal region corresponding to the exons were obtained from the genome data viewer of the NCBI. The chromosomal positions of each SNP were mapped to the corresponding exon, and the frequency of the type of allelic transition in each of the exon was calculated.
SNP prevalence by cancer phenotype The results obtained from the dbSNP were converted into ensembl Variant Effect Predictor (VEP) tool input format and were submitted to the VEP ensemble web interface for categorizing the SNPs into deleterious and non-deleterious based on the prediction algorithms, SIFT and PolyPhen. The SNPs that were marked deleterious were checked in TCGA for the cancer incidences.
Copy Number Variations assessment ClinVar database (http://www.ncbi.nlm.nih.gov/clinvar) was used for Hsp90 specific genomic variations for potential cancer incidences. Subsequently the copy number variations in different
5
Hsp90 isoforms were analyzed and the potential variations were scored by quantitative real time PCR.
Cell culture maintenance and treatments Human embryonic kidney cells HEK293 (ATCC® CRL-1573™), human lung carcinoma A549 (ATCC® CCL-185™), human neuroblastoma IMR32 (ATCC® CCL-127™) were obtained from American Type Culture Collection(ATCC). Cells were authenticated and maintained in Dulbecco's modified eagle medium (DMEM, Thermo Fisher, cat. no. 12491-023) with 10% fetal bovine serum(FBS, Thermo Fisher, cat no. 12483-020) at 37°C in CO2 incubator. IMR-32 cells were treated with rotenone to inhibit OXPHOS complex I (ROT, 2 µM, Sigma-Aldrich, cat. no.557368-1GM).To inhibit glycolysis, 2-Deoxy-D-glucose(2DG,1 mM, MP Biomedicals, LLC, cat. no.194642) was used. Cobalt chloride (CoCl2, 200 µM, Sigma-Aldrich, cat. no.D8375-1G) was used to induce chemical hypoxia. Cells grown in 6-well plates were treated with respective drugs before processing at specified time periods. All the drug treatments were done for 18-24 h intervals unless otherwise indicated.
TRAP1 knockdown system TRAP1(NM_016292.2) shRNA was designed using online software (BLOCK-it RNA designer, Invivogen
and
Biosettia).
The
shRNA
were
custom
synthesized(Sense:
5'-
TTTGGGGTTCCACTTCCAAACATGACGAATCATGTTTGGAAGTGGAACCCTTTTT-3', Antisense:5'CTAGAAAAAGGGTTCCACTTCCAAACATGATTCGTCATGTTTGGAAGTGGAACCC-3') and cloned in mu6Pro vector (kind gift from Dr. David L. Turner at University of Michigan, Ann
6
Arbor, USA). IMR32 cells were transfected with shRNA TRAP1, stable cells were selected with G418for 21 days.
TRAP1 over expression system The full length coding sequence of TRAP1 was retrieved (NM_016292.3), and PCR amplified from
MCF-7
cDNA
using
gene
specific
primers,
forward
5'-
ATTGAATTCATGGCGCGCGAGC-3'; reverse 5'-ATTGGATCCGTGTCGCTCCAGG-3' and cloned into ECoRI and BamHI restriction sites of pEGFPN2 expression system. Stable IMR-32 cells over expressing TRAP1 were used for subsequent experiments.
Chromosomal DNA isolation Genomic
DNA
was
isolated
with
minor
modification
of
the
protocol
(http://cancer.ucsf.edu/_docs/cores/array/protocols/dna_cell_culture.pdf). In brief, cells were trypsinized (0.05% Trypsin-EDTA) and washed with PBS. After centrifugation (465xg), pellet was resuspended in Tris.Cl buffer (pH 8.0) and subjected to sodium dodecyl sulphate (SDS, 0.66%) and Proteinase K (100 µg/mL) treatments, mixed gently by inversion and incubated overnight at 55°C. Saturated sodium chloride was added followed by ethanol at room temperature to precipitate chromosomal DNA. The DNA was spooled, rinsed twice in 70% ethanol, air dried, and resuspended in TE buffer, then subjected to Phenol:Choloroform:Isoamyl alcohol (25:24:1) extraction. The aqueous phase further subjected to Choloform:Isoamyl alcohol (25:24) extraction to remove phenol and ethanol precipitated.
7
Quantitative Polymerase Chain Reaction (qPCR) The genomic DNA sequence for TRAP1 was retrieved from UCSC genome browser. (URL: genome-asia.ucsc.edu; Chr:16 3658037-3717597, +ve strand). The qPCR primers were designed by online software (Prime3Plus). The primer sequences for TRAP1 (Chr16 3658037-3717597) and
ZNF80
(Chr3:114234636-114237578)
are,
for
TRAP1,
forward:
5'-
GTGCGGCAAAATAACGAACG-3' and reverse: 5'-GTTCGTTTGTAGCCACGTGT-3', and for
ZNF80,
forward:
5'-AAAACCTGTGAGTGCAGACC-3'
and
reverse:
5'-
CTGACGTCGGTTTGAGTTCA-3'.
Scanning confocal imaging microscopic analysis Cells (0.2x106 cells/mL) were grown on cover glasses (22x22 mm, Fisher Scientifics) were used for immunofluorescence. Cells first incubated with Mitotracker Red (CMX-Ros, 200 nM, Dojindo, cat no.R237) for 15 min in the incomplete medium, washed with PBS, fixed in 3.7% formaldehyde for 10 min, and permeabilized with 0.1% Triton x100 (Sigma-Aldrich, cat. no.T8787-100ML) for 7 min. For immunofluoresecnce, cells were incubated with 2%BSA (Sigma-Aldrich, cat. No.A2153 ) for 1 h, TRAP1 antibodies (1:200, abcam, cat no.ab182775) for overnight (4C), and subsequently incubated with Alexa Fluor 488 conjugated secondary antibodies (1:1000, abcam, cat no. ab150077) at room temperature for 1 h. The cover glasses were mounted using mounting medium containing DAPI (Vectashield,cat. no. H1200) and observed under microscope (SP8, Leica).
Immunoblot analysis
8
Cells were lysed in RIPA buffer (pH 7.4) and proteins (20 µg/well) were resolved on 10% SDSPAGE, then transferred onto nitrocellulose membrane. The membrane was blocked with 3% BSA for 1 h and incubated with TRAP1 antibodies (1:1000, 4C, 12 h, BD Biosciences, cat. no.612345) followed by HRP conjugated secondary antibodies (1:5000, RT, 1 h, Santacruz, cat. no.sc2005). The antigen-antibody interactions were scored using BM Chemiluminescence western blot kit as per the manufacturer's instructions (Roche, cat. no.11520709001).
Oxygen Consumption Rate (OCR) Cells (0.2x105/ 1.0x105/ 2.0x105) were grown on 6-well NUNC plates at standard culture conditions to achieve 60% confluence. Untreated cells were processed immediately, and cells after treatments were washed with PBS and resuspended in incomplete medium (100 µL). The basal level respiration was measured at different time intervals using Oxygraph 2K (Oroboros Instruments Corp, Austria). Medium without cells was used as a blank. The values obtained were plotted as line as well as a bar diagram. In case of 2DG and rotenone treatments, we measured the overall respiration unless otherwise these drugs interfere with stage-specific respiration.
Total ATP analysis The luminescence ATP detection kit (abcam, cat no. ab113849) was used for measuring total cellular ATP as per the manufacturer's instructions. In brief, cells (0.02x105/100 µL/1.0x105/ 100 µL/ 2.0x105/µL) in a 96-well plate were first subjected to lysis (50 µL), and then was added the substrate solution (50 µL) incubated in dark for 10 min. The change in luminescence with time was measured and represented in a bar diagram.
9
Preparation of cells for metabolite analysis Whole cells were harvested and prepared for loading. Briefly, a volume of culture containing 1x107 cells were harvested and cellular metabolism was quenched by the addition of a 5x volume of 60% methanol supplemented with 0.85% (W/V) ammonium bicarbonate at a minimum temperature of -40C. Quenched cells were pelleted by centrifugation and the pellet rinsed before being resuspended in 0.5 mL of 100% methanol, flash frozen on liquid nitrogen and when thawed, pelleted by centrifugation. The supernatant was retained and the extraction process repeated. The pooled 1 mL of extracted metabolites were dried and stored at -80°C until NMR sample preparation.
Metabolite analysis by 2D NMR Metabolites present in NMR samples were identified by comparison of related chemical shifts with published NMR data of known metabolites with reference to online resources (Madison Metabolomics Consortium Database (MMCD) and the Spectral Database for Organic Compounds
(SDBS)
http://riodb01.ibase.aist.go.jp/sdbs/(National
Institute
of
Advanced
Industrial Science and Technology, October-December 2010). Initial targets for metabolite identification were drawn from the known metabolite components of DMEM medium, and subsequent cross comparisons made to the CD-CHO medium. All spectra were recorded at 37°C on a Unity INOVA 600 MHz NMR spectrometer (Varian Inc, PaloAlto, CA, USA) fitted with a 5 mm triple resonance HCN probe with z-shield gradients. 1D 1H spectra with WATERGATE solvent suppression were recorded over 32768 points, 512 transients and a spectral width of 8004.8 Hz. An acquisition time of 2.047 s and a relaxation delay of 1.5 s were used. This provides an overall excitation recovery time of 3.547 s and produces a total experimental time of
10
31min that enables efficient and fast data acquisition with minimal effect due to relaxation; 1H T1 values up to 1.5 s will experience an effective magnetization recovery greater than 90%. Data processing was performed by ACD Labs SpecManager V9.0 involved zero filling to 65536 points and multiplication of 1 Hz exponential line broadening before Fourier transformation.
Results Functional classification of SNPs from Hsp90 isoforms The SNP variation has been linked to disease outcome thus; proper identification, functional assignment, and validation play a key role in understanding their functional relevance (Batnyam et al., 2013). Principally, SNPs are classified based on their positional occurrences as intronvariant, splice acceptor variant, UTR-3 prime, UTR-5 prime, downstream –500bp, splice-donor, upstream –2 Kb. Based on their functional consequences, they are classified into synonymous, missense, frameshift, stop-gain, stop-lost. Interestingly, most of the SNPs are reported as intronvariant followed by missense variants in the exons, compared to other functional classes (Figure 1a). Further, Hsp90AA1 and TRAP1 showed enhanced intron-variants compared to Hsp90AB1 and Hsp90B1. Between Hsp90AA1 and TRAP1 the latter exhibited more number of SNPs. While Hsp90AA1 exhibited the variance in the order of C/T < A/G < C/G < A/C followed by others, TRAP1 exhibited the variance in the order of A/G < C/T < C/G < A/C followed by others. The other SNPs are found to be in the upstream regions i.e. above 2 Kb region. These results suggested us that A/G and C/T variations are predominant (Figure 1b). In gross terms, intron variations contribute to disease susceptibility, the prevalence Hsp90 variations may have role in cancer progression (Cooper, 2010).
11
The isoform specific chromosome mapping of functional SNPs and their prevalence by cancer phenotype The mutations in exonic regions usually will not have direct influence on gene transcription or functions of the transcripts hence considered silent despite their non-synonymous (missense) nature. However, these mutations can disrupt the splice sites hence can influence the pathogenic outcome (Wu and Hurst, 2016). The missense mutations on each exon of Hsp90 isoforms were scored and represented in the bar diagram. Hsp90 isoforms, Hsp90AA1 has 10 exons, Hsp90AB1 has 18 exons, Hsp90AB1 has 12 exons and TRAP1 has 18 exons. Interestingly, all of them exhibited some amount of missense mutations and without any specific pattern or predominance (Figure 2a). These findings suggested that unless and otherwise exon variants are validated for functional relevance with respect to cancer phenotype(s), the significance of these mutations (allelic transitions) cannot be extrapolated.
We could retrieve 16549, 3695, 4677, 20693 SNPs from the database for HSP90AA1, HSP90AB1, HSP90B1, TRAP1 respectively for Homo sapiens. Among these, 4, 3, 13 and 12 SNPs respectively were reported to be deleterious and are associated with cancer incidences (as per SIFT and PolyPhen scores, Figure 2b; Table 1). Intriguingly, it was found that compared to SNPs from TRAP1and HSP90B1, the SNPs from Hsp90AA1 and Hsp90AB1 have correlated with lower number of cancer incidences. Till date, TRAP1 is less implicated in cancer incidences due to limited studies pertinent to its contribution in cancer progression. However, the present findings suggest that TRAP1 may play significant role in the pathogenic outcome of cancer.
Copy number variation (CNV) as a predictive marker for disease assessment
12
Cancer is a polygenic disease, where tumor heterogeneity plays a significant role in disease progression (Meacham and Morrison, 2013). Similar to exon versus intron variations that reveal population differences (Zhan et al., 2014), SNPs within the tumor may provide information related to tumor heterogeneity. While SNPs themselves are considered to be a small variation with respect to gene expression pools that decides the cell fate or disease outcome, we have examined for the CNV that provides information related to gain or loss of segments of DNA (de Koning et al., 2011). We have used transformed, malignant and metastatic cancer cells, HEK293T, A549, and IMR-32 cells respectively to understand TRAP1 specific correlation with cancer phenotypes. While transformed and malignant cells showed lower CNVs, there is a significant increase observed in metastatic cells (Figure 2c).
Correlation of TRAP-1 expression with cellular energy metabolism Correlating with CNV data, we also observed enhanced TRAP1 protein expression in phenotype specific manner. There was a gradual increase in TRAP1 expression HEK293T < A549 < IMR32 cells. Although TRAP1 mitochondrial localization is not clear in IMR-32 cells, in HEK-293 and A549 cells, TRAP1 showed strong mitochondrial co-localization (Figure 3a). Subsequently, the immunoblot analysis showed its increased expression in A549 and IMR-32 cells (Figure 3b).
Mitochondria are the major source of ATP production (Bratic and Trifunovic, 2010). Cancer cells are known to bypass conventional energy metabolism due to their selection pressure and limited oxygen supply (Warburg, 1924). Hence, we examined hypoxia (using cobalt chloride) and OXPHOS inhibition (using rotenone) effects on TRAP1 expression in IMR-32 cells. While a marginal increase in TRAP1 expression was observed in response to hypoxia, rotenone treatment
13
has significantly decreased its expression (Figure 3c). This suggest that drug induced uncoupling may work independent of TRAP1. Next we examined whether TRAP1 differential expression in HEK293T, A549 and IMR-32 cells differentially regulates mitochondrial functions. Compared to HEK293T cells, A549 and IMR-32 cells showed enhanced ATP levels (Figure 3d). Subsequently, we observed decreased oxygen consumption rate (OCR) in IMR-32 cells compared to A549 and HEK293T cells (Figure 3e and 3f).
Correlation of TRAP-1 expression with cellular energy metabolism To understand the involvement of TRAP1 in mitochondrial function and energy metabolism, custom synthesized shRNA TRAP1 was stably expressed in IMR-32 cells (Figure 4a and 4b). These cells were used to measure OCR and ATP levels in comparison with parental cells, and we observed TRAP1 KD cells displaying decreased OCR. Next, IMR-32 cells stably over expressing the full-length TRAP1 was compared with KD cells. The TRAP1 OE cells showed further decrease in OCR compared to KD cells (Figure 4c). To understand the effect of 2DG and rotenone on basal level of respiration of parental, TRAP1 KD and TRAP1 OE cells, they were treated with the drugs and the rate of respiration was measured. TRAP1 KD cells are found to be sensitive to 2DG treatment compared to parental and TRAP1 OE cells (Figure 4e versus 4d and 4f). However, rotenone treatment has decreased respiration in parental and TRAP1 OE cells (Figure 4d and 4f). These results suggested that mitochondrial OXPHOS is minimized in the TRAP1 expression background. Subsequently, we have measured total cellular ATP levels in parental, TRAP1 KD and TRAP1 OE cells either alone or in response to pharmacological interventions of glycolysis (using 2DG) and OXPHOS (using rotenone). The 2DG treatment resulted in significant decrease in ATP levels in TRAP1 KD cells, but not in parental cells.
14
Whereas, rotenone treatment showed no significant influence on total ATP levels in all the cell types studied (Figure 4g, 4h and 4i). The TRAP1 OE cells exhibited similar response like in parental cells with respect to 2DG treatment indicating that in the TRAP1 background, inhibiting glucose oxidation had no effect on ATP production (Figure 4i). To understand both source and secondary metabolite status of these cells, NMR based metabolite analysis was carried out in parental, TRAP1 KD and TRAP1 OE cells. We observed a moderate increase in intracellular glutamine and glutamate levels in KD cells in comparison with parental cells (Figure 4j and 4k). In contrary, a drastic increase (nearing 4 folds) in these secondary metabolites was observed in TRAP1 OE cells (Figure 4l). These results suggest that in TRAP1 OE background, these metabolites are probably accumulated from the alternate energy metabolism.
15
Discussion The Cellular proteotoxicity contributes to a large number of pathological disorders including cancer (McConkey, 2017). Cells evolve with mechanisms to overcome proteotoxicity by increasing the expression of molecular chaperones to regulate cellular protein homeostasis. Hsps having the chaperoning functions clears the proteotoxicity by not only refolding the damaged proteins, but also facilitating the mutated/altered proteins to function normally. Thus, Hsps are attributed to various adaptations (Feder and Hofmann, 1999). The Hsp90 chaperone is considered as a cancer chaperone consequently its inhibition has emerged as a novel antitumor strategy to combat cancer (Solit and Rosen, 2006). However, its mitochondrial homologue, TRAP1 functions are not fully explored. There are reports that TRAP1 expression correlates with tumor aggression, however lacks conclusive experimental evidences. Since TRAP1 is a mitochondrial resident, and targeting the mitochondrial metabolism is emerging as a strategy, we examined its role in energy metabolism correlating with disease progression.
The chaperone, Hsp90 is in the centre of protein network forming the foldosome complex, thus, is critical for cellular protein stability, in addition is also involved in the functional maturation of mutated gene products (Csermely et al., 1998; Sreedhar et al., 2004; Kishinevsky et al., 2018). To a large extent, its homologue in the ER, Grp94 was also explored for protein quality control and antigen presentation (Argon et al., 2016). Despite several individual studies related to SNPs in Hsp90 family of genes, their clinical relevance with respect to cancer progression is not known. We asked whether SNPs analysis followed by CNVs unravels the isoform specific functions of Hsps. From the analyses we could confirm that among Hsp90 genes, TRAP1 exhibits more number of SNPs and CNVs (on chromosome 16) compared to Hsp90 isoforms
16
HSP90AA, HSP90AB1 and HSP90B1. It has been shown that chromosome 16 duplications are already observed and suggested to have correlation with disease susceptibility (Martin et al., 2004). While SNPs provided the firsthand information on TRAP1 functional relevance with cancer, the CNVs data confirmed that unless until there is a need for biochemical/phenotypic reformation, the increase in copy number may not be required.
We hypothesized that similar to Hsp90, TRAP1 may be involved in tumor progression, thus, can also be considered as a cancer chaperone. Till date there is no such information available that mitochondrial proteins directly contribute to cancer. In support of this, it was also proposed that cancer cells prefer alternate energy metabolism by bypassing the OXPHOS (Bratic, 2010). There were limited studies that correlate TRAP1 expression with altered energy metabolism (Matassa et al., 2018) and neoplastic transformation (Sciacovelli et al., 2013). It is established that hypoxia decreases mitochondrial OCR and thus affects the cellular ATP production. In our model, the basal levels of TRAP1 are high enough hence subjecting these cells to OXPHOS inhibition did not influence mitochondrial functions indicating that cells with enhanced TRAP1 expression may have adapted to sustain bypassing effects of OXPHOS. Strengthening this, TRAP1 OE cells also showed similar non-responsiveness to OXPHOS inhibition. Interestingly, the metabolite analyses revealed that there is no difference in the net glucose levels in parental cells compared to TRAP1 KD cells. However, there was a significant increase in glutamate and glutamine levels in the KD cells. Further, TRAP1 KD has sensitized these cells to glycolysis inhibition and thus to ATP production indicating that TRAP1 is probably involved in the metabolic reprogramming of cancer cells. Intriguingly, TRAP1 OE cells showed nearly 4 folds increases in glucose, glutamine and glutatmate levels indicating cellular adaptations to alternate energy metabolism.
17
The metabolic reprogramming is now considered as one of the hallmarks of cancer (Hanahan and Weinberg, 2011; Li and Le, 2018). Although it was thought that targeting mitochondria is sufficient to target cancer (Skoda et al., 2018), extrapolating studies from Warburg (Warburg, 1924) unraveled several alternate pathways, thus metabolite sensing is suggested to be an important parameter for deciding the therapeutic outcome (Han and Patten, 2018; Papa et al., 2018). In the context that metabolite targeting acts as an efficient tumor suppressor mechanism, our studies by unraveling TRAP1 involvement in metabolic reprogramming suggests that this mitochondrial chaperone can be a potential pharmacological target to combat cancer metabolism.
Acknowledgements Authors thank Dr. Anant Patel for helping us with NMR experiments. Authors thank Science and Engineering Research Board (SERB), Department of Science and Technology, Ministry of Science, Government of India for the financial support. Mr. Shrikant is supported by fellowship from CSIR for his PhD.
Conflict of Interest: Authors declare the conflict of interest as NONE.
References Andersson, S.G., Karlberg, O., Canback, B., Kurland, C.G., 2003. On the origin of mitochondria: a genomics perspective, Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences. 358, 165–179. 18
Argon, Y., Marzec, M., Hawkes, C.P., Eletto, D., Boyle, S., Rosenfeld R., Hwa, V., Wit, J.M., van Duyvenvoorde, H.A., Oostdijk, W., Losekoot, M., Pedersen, O., Yeap, B.B., Flicker, L., Barzilai, N., Atzmon, G., Grimberg, A., 2016. A Human Variant of Glucose-Regulated Protein 94 That Inefficiently Supports IGF Production, Endocrinology, 157,1914–1928.
Batnyam, N., Gantulga, A., Oh, S., 2013. An Efficient Classification for Single Nucleotide Polymorphism (SNP) Dataset. Computer and Information Science. Studies in Computational Intelligence. 493, 171-185.
Bratic, I., Trifunovic, A., 2010, Mitochondrial energy metabolism and ageing, Biochim. Biophys. Acta. 1797, 961-967.
Cooper, D.N., 2010. Functional intronic polymorphisms: Buried treasure awaiting discovery within our genes. Human Genomics. 4, 284-288.
Csermely, P., Schnaider, T., Csaba, S., Prohaszka, Z., Nardai, G., 1998. The 90-kDa Molecular Chaperone Family: Structure, Function, and Clinical Applications. A Comprehensive Review. 79, 129-168.
de Koning, A.P.J., Gu, W., Castoe, T.A., Batzer, M.A., Pollock, D.D., 2011. Repetitive Elements May Comprise Over Two-Thirds of the Human Genome. PLOS Genetics. 7,1-12.
19
Feder, M.F., Hofmann, G.E., 1999. Heat Shock Proteins, Molecular Chaperones, and The Stress Response. Evolutionary and Ecological Physiology, Annu. Rev. Physiol. 61,243–282.
Han, C.Y., Patten, D.A., Richardson, R.B., Harper, M.E., Tsang, B.K. 2018. Tumor metabolism regulating chemosensitivity in ovarian cancer. Genes Cancer. 9,155-175.
Hanahan, D., Weinberg, R.A., 2011. Hallmarks of Cancer: The Next Generation, Cell. 144,646674.
Jolly, C., Vourc’h, C., Robert-Nicoud, M., Morimoto, R.I., 1999. Intron-independent association of splicing factors with active genes. J Cell Biol. 145, 1133-1143.
Katschinski, D.M., 2004. On heat and cells and proteins; News Physiol. Sci. 19,11–15.
Kishinevsky, S., Wang, T., Rodina, A., Chung, S.Y., Xu, C., Philip, J., Taldone, T., Joshi, S., Alpaugh, M.L., Bolaender, A., Gutbier, S., Sandhu, D., Fattahi, F., Zimmer, B., Shah, S.K., Chang, E., Inda, C., Koren, J., Saurat, N.G., Leist, M., Gross, S.S., Seshan, V.E., Klein, C., Tomishima, M.J., Erdjument-Bromage, H., Neubert, T.A., Henrickson, R.C., Chiosis, G., Studer, L. 2018. HSP90-incorporating chaperome networks as biosensor for disease-related pathways in patient-specific midbrain dopamine neurons. Nature communications. 9,1-15.
Li, T., Le, A., 2018. Glutamine Metabolism in Cancer. Adv Exp Med Biol. 1063, 13-32.
Liberti, M.V., Locasale, J.W., 2016. The Warburg Effect: How Does it Benefit Cancer Cells? Trends Biochem Sci. 41,211-218. 20
Liu, J., Zhou, Y., Liu, S., Song, X., Yang, X.Z., Fan, Y., Chen, W., Akdemir, Z.C., Yan, Z., Zuo, Y., Du, R., Liu, Z., Yuan, B., Zhao, S., Liu, G., Chen, Y., Zhao, Y., Lin, M., Zhu, Q., Niu, Y., Liu, P., Ikegawa, S., Song, Y.Q., Posey, J.E., Qiu, G., 2018. The coexistence of copy number variations (CNVs) and single nucleotide polymorphisms (SNPs) at a locus can result in distorted calculations of the significance in associating SNPs to disease. Hum Genet. 137,553-567.
Mace, A., Kutalik, Z., Valsesia, A., 2018. Copy Number Variation. Methods Mol Biol.1793,231258.
Martin, J., Han, C., Gordon, L.A, Terry, A., Prabhakar, S., She, X., Xie, G., Hellsten, U.,Chan, Y,M., Altherr, M., Couronne, O., Aerts, A., Bajorek, E., Black, S., Blumer, H., Branscomb, E., Brown, N.C., Bruno, W.J., Buckingham, J.M., Callen, D.F., Campbell, C.S., Campbell, M.L.,Campbell, E.W., Caoile, C., Challacombe, J.F., Chasteen, L.A., Chertkov, O., Chi, H.C.,Christensen, M., Clark, L.M., Cohn, J.D., Denys, M., Detter, J.C., Dickson, M.,Dimitrijevic-Bussod, M., Escobar, J., Fawcett, J.J., Flowers, D., Fotopulos, D., Glavina, T., Gomez, M., Gonzales, E., Goodstein, D., Goodwin, L.A., Grady, D.L., Grigoriev, I., Groza, M.,Hammon, N., Hawkins, T., Haydu, L., Hildebrand, C.E., Huang, W., Israni, S., Jett, J., Jewett, P.B., Kadner, K., Kimball, H., Kobayashi, A., Krawczyk, M.C., Leyba, T., Longmire, J.L., Lopez, F.,Lou, Y., Lowry, S., Ludeman, T., Manohar, C.F., Mark, G.A., McMurray, K.L., Meincke, L.J., Morgan, J., Moyzis, R.K., Mundt, M.O., Munk, A.C., Nandkeshwar, R.D., Pitluck, S., Pollard, M., Predki. P., Parson-Quintana, B., Ramirez, L., Rash, S., Retterer, J., Ricke, D.O., Robinson, D.L.,Rodriguez, A., Salamov, A., Saunders, E.H., Scott, D., Shough, T., Stallings, R.L., Stalvey M., Sutherland, R.D., Tapia, R., Tesmer, J.G., Thayer, N., Thompson,
21
L.S., Tice, H., Torney, D.C.,Tran-Gyamfi, M., Tsai, M., Ulanovsky, L.E.,Ustaszewska, A., Vo, N., White, P.S., Williams, A.L., Wills, P.L., Wu, J.R., Wu, K., Yang, J., Dejong, P., Bruce, D., Doggett, N.A., Deaven, L., Schmutz. J., Grimwood, J., Richardson. P., Rokhsar, D.S., Eichler, E.E., Gilna, P., Lucas, S.M., Myers, R.M.,Rubin. E.M., Pennacchio, L.A., 2004. The sequence and analysis of duplication-rich humanchromosome 16. Nature. 432,988-994.
Matassa, D.S., Agliarulo, I., Avolio, R., Landriscina, M., Esposito, F., 2018. TRAP1 Regulation of Cancer Metabolism: Dual Role as Oncogene or Tumor Suppressor. Genes.5, 9.
McConkey, D.J., 2017. The integrated stress response and proteotoxicity in cancer therapy. Biochem Biophys Res Commun. 482,450-453.
Meacham, C.E., Morrison, S.J., 2013. Tumour heterogeneity and cancer cell plasticity. Nature. 501,328-337.
Neckers, L., 2007. Heat shock protein 90: the cancer chaperone. J Biosci. 32,517-530.
Papa, S., Choy, P.M., Bubici, C. 2018. The ERK and JNK pathways in the regulation of metabolic reprogramming.Oncogene. ISSN 1476-5594 (online).
Sciacovelli, M., Guzzo, G., Morello, V., Frezza, C., Zheng, L., Nannini, N., Calabrese, F., Laudiero, G., Esposito, F., Landriscina, M., Defilippi, P., Bernardi, P., Rasola, A., 2013. The mitochondrial chaperone TRAP1 promotes neoplastic growth by inhibiting succinate dehydrogenase. Cell Metabolism. 17, 988-999. 22
Skoda, J., Borankova, K., Jansson, P.J., Huang, L.H.M., Veselska, R., Richardson, D.R., 2018. Pharmacological targeting of mitochondria in cancer stem cells: An ancient organelle at the crossroad of novel anti-cancer therapies.Pharmacol Res. 139, 298-313.
Solit, D.B., Rosen, N., 2006. Hsp90: A Novel Target for Cancer Therapy. Current Topics in Medicinal Chemistry. 6, 1205-1214.
Sreedhar,
A.S.,
Kalmar,
E.,
Csermely,
P.,
Shen,
Y.F.,
2004.
Hsp90
isoforms:
functions,expression and clinical importance. FEBS Lett. 562,11-15.
Urban, J.D., Budinsky, R.A., Rowlands, J.C., 2012. An evaluation of single nucleotide polymorphisms in the human heat shock protein 90 kDa α and β isoforms. Drug Metab Pharmacokinet. 27,268-278.
Vishal, C., Kumar, J.U., Brahmendra Swamy, C.V., Nandini, R., Srinivas, G., Kumaresan, R., Sreedhar, A.S., 2011. Repercussion of Mitochondria Deformity Induced by Anti-Hsp90 Drug 17AAG in Human Tumor Cells. Drug Target Insights. 5,11-32.
Warburg, O., Posener, K., Negelein, E., 1924. üeber den Stoffwechsel der Tumoren. Biochem Z 152,319–344. Reprinted in English in the book On metabolism of tumors by O. Warburg, Constable. London, 1930.
23
Wu, X., Hurst, L.D., 2016. Determinants of the Usage of Splice-Associated cis-Motifs Predict the Distribution of Human Pathogenic SNPs. Mol Biol Evol. 33,518-529.
Zhan, X., Dixon, A., Batbayar, N., Bragin, E., Ayas, Z., Deutschova, L., Chavko, J., Domashevsky, S., Dorosencu, A., Bagyura, J., Gombobaatar, S., Grlica, I. D., Levin, A., Milobog, Y., Ming, M., Prommer, M., Purev-Ochir, G., Ragyov, D., Tsurkanu, V., Vetrov, V., Zubkov, N., Bruford, M. W.2014. Exonic versus intronic SNPs: contrasting roles in revealing the population genetic differentiation of a widespread bird species. Heredity, 114,1-9. (2014).
Zhang, F., Gu, W., Hurles, M.E., Lupski, J.R. 2009. Copy number variation in human health, disease, and evolution. Annual Review of genomics and human genetics, 10, 451-481.
24
GENE
HSP90AA1
HSP90AB1
HSP90B1
TRAP1
CHR
14
SNP locus
dbSNP ID
Exon No.
Allele in reference assembly
Minor allele
Amino acid Transition
Cancer type
102085425
rs745469368
8
G
C
P301R
Adenomas and Adenocarcinomas; Cystic, Mucinous and Serous Neoplasms
102085341
rs945693652
8
A
G
V329A
Nevi and Melanomas
102086091
rs747407643
9
C
A
D188Y
Transitional Cell Papillomas And Carcinomas
102085341
rs945693652
8
A
G
V329A
Skin Cutaneous Melanoma
44248741
rs1331943025
2
A
G
I38V
Adenomas and Adenocarcinomas; Cystic, Mucinous and Serous Neoplasms
44251789
rs920937641
9
G
A/T
R456H, R456L
Breast Invasive Carcinoma
44250062
rs760864610
5
AAA
-
K186del
Uterine Corpus Endometrial Carcinoma
103943816
rs3209749
14
A
C
N657H
Squamous Cell Neoplasms
6
Adenomas and Adenocarcinomas, Epithelial Neoplasms, NOS, Squamous Cell Neoplasms,Transitional Cell Papillomas and Carcinomas Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous Neoplasms, Epithelial Neoplasms. NOS Colon Adenocarcinoma, Stomach Adenocarcinoma Head and Neck Squamous Cell Carcinoma Head and Neck Squamous Cell Carcinoma
103939608
rs1354079125
8
T
C
Y359H
103941439
rs1385298900
9
C
G
H374Q
103941638
rs368374645
10
C
T
R414C
103943204
rs762366467
13
C
G
A592G
103946665
rs372636341
15
C
T
P692L
103932846
rs1362809151
4
T
G
I105M
Glioblastoma Multiforme
103942642
rs774797218
12
T
C
I497T
Lung Adenocarcinoma
103943778
rs754939231
14
C
T
P644L
103932375
rs571691885
3
G
A
R84K
103946849
rs770752783
16
C
A/T
R724R, R724W
103938403
rs1445926074
7
G
A
D307N
103943757
rs1277290411
14
C
T
S637F
3674486
rs1307786417
10
C
A
W299C,W159C
3679729
rs1272474481
14
G
A/C
S178W, S125L, S125W,S38L,S38W
12
16
Liver Hepatocellular Carcinoma Uterine Corpus Endometrial Carcinoma Stomach Adenocarcinoma Bladder Urothelial Carcinoma Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous Neoplasms, Squamous Cell Neoplasms Cystic, Mucinous and Serous Neoplasms Nevi and Melanomas, Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous
25
3690838
rs1288142880
17
T
C
3664360
rs113510154
6
G
A/C
3674346
rs1422474678
10
G
C
3662060
rs192869535
3
G
A
3686127
rs192245771
15
G
A
R114W, R61W
Stomach Adenocarcinoma
T
E643K, E590K, E503K
Skin Cutaneous Melanoma
R658C
Colon Adenocarcinoma
3662000
rs145014945
3
C
E79G
Neoplasms, Epithelial Neoplasms, NOS Adenomas and Adenocarcinomas, Cystic, Mucinous and Serous Neoplasms, Epithelial Neoplasms, NOS
R495C, R495G, R442C, R442G, R355C, R355G P346R, P206R, P293R R623C, R483C, R570C
Colon Adenocarcinoma Stomach Adenocarcinoma Uterine Corpus Endometrial Carcinoma
Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma Uterine Corpus Endometrial Carcinoma
3658834
rs139636268
2
G
A/T
3672727
rs1210505741
9
G
-
L380fs (L380C), L327C, L240C
3672762
rs375364304
9
C
T
R368H
Uterine Corpus Endometrial Carcinoma
3674464
rs13926
10
G
C,T
R254S, R254G
Multiple Myeloma
R658S
Figure Legends Figure 1. The functional classification of SNPs indifferent Hsp90 isoforms. a. SNPs reported in the isoforms of HSP90 are retrieved from dbSNP. The SNPs predominantly found in the intronic regions were scored and represented in the bar diagram. Note that HSP90AA1 (Hsp90α) and TRAP1 (TRAP1) showing more variations compared to HSP90AB1 (Hsp90β) and HSP90B1 (Grp94). b. The frequency of allelic transitions reported under each functional class for the HSP90 isoforms were scored and represented in the bar diagram. Note predominant allelic transitions in HSP90AA1 and TRAP1.
Figure 2. Allelic transitions based on exon number. a. Isoform specific chromosomal mapping of functional SNPs reported by the dbSNP. The number of transition mutations is more predominant in all the isoforms compared to transversion mutations. b. Cancer phenotype 26
specific SNP prevalence. The cancer incidences reported with missense mutations in the exonic regions for each isoform of HSP90 were scored and represented in the bar diagram. most of the cancer incidences are reported in the TRAP1 and HSP90B1 compared with the other isoforms. c. Copy number variation (CNV) versus cancer phenotype. Quantitative real time PCR analysis of CNV in TRAP1 in HEK293, A549, and IMR-32 cells. The TRAP1 PCR values are normalized using ZNF80. Note IMR-32 cells showing increase in CNV.
Figure 3. Correlation of TRAP-1 expression with altered cellular energy metabolism. a. Cytoimmunofluorescence analyses of TRAP1 localization in the mitochondrion. The mitochondria from HEK293T, A549 and IMR-32 cells were stained with CMX-Ros (red) and examined for TRAP1 (stained with Alexa fluor 488 (green)) colocalization. The nucleus was stained with DAPI (blue). Note low mitochondrial localization in IMR-32 cells despite increased levels of total TRAP1. Scale bar represents 25 µm. b. Immunoblot analysis of TRAP1 in HEK293T, A549 and IMR-32 cells. The numbers represents the densitometry values obtained for gene/actin ratio. c. Effect of CoCl2 and rotenone on TRAP1 expression. RT-PCR analysis of TRAP-1 expression. d. Measurement of total cellular ATP levels. The average luminescence values obtained were interpolated with the ATP standards and represented. Note that A549 and IMR-32 cells showed enhanced ATP levels. e. Analysis of oxygen consumption rate (OCR) in different cell types. Spectral analysis of OCR observed over time obtained from Oxygraph. f. statistical representation of OCR data. The data acquired were from three independent experiments. Values represented are Mean SD. The significance values are, * p<0.05, ** p<0.01, *** p<0.001.
27
Figure 4. Effect of TRAP1 knockdown on altered energy metabolism. a, b. custom synthesized shRNA TRAP1 (1 and 2 µg/ transfection) was used for making stable TRAP1 KD cells. c. The OCR analysis of parental, TRAP1 KD and TRAP1 OE cells. Note decrease in OCR in IMR-32 cells. Measurement of cellular respiration in parental (d), TRAP1 KD (e) and TRAP1 OE (f) cells. Total cellular ATP levels in parental (g), TRAP KD (h) and TRAP1 OE (i) cells in response to 2DG and rotenone. Metabolite analysis using NMR spectra of total cell lysates from parental (j), TRAP1 KD (k) and TRAP1 OE (l) cells. Note significant increase in glucose, glutamine and glutamate levels in TRAP1 OE cells.
28
29
30
31
32
Highlights
TRAP1 exhibits more number of SNPs in cancer cells compared to Hsp90
TRAP1 exhibits increased copy number variations (CNVs) in metastatic cancer cells
TRAP1 CNVs correlate with increased expression in aggressive cancer cells
TRAP1 appears to facilitate alternate energy metabolism
33