Background
Multiple myeloma (MM) is characterised by the clonal proliferation of malignant plasma cells within the bone marrow (BM) and is the second most common haematological malignancy. The key clinical manifestations of MM include the development of painful osteolytic bone lesions, renal insufficiency, suppressed haematopoietic function and increased BM angiogenesis [
1]. There are an array of clinical variants of the disease, ranging from the asymptomatic monoclonal gammopathy of undetermined significance (MGUS) and smouldering MM, to the more aggressive active MM and plasma cell leukaemia. Numerous genes, pathways and miRNAs have been identified in MM that function as predictive biomarkers of highly proliferative disease and likelihood of response to treatment [
2‐
6]. Although the introduction of novel therapies has seen a significant improvement in the median survival of some groups of MM patients, the survival for some subgroups of patients, particularly those with highly proliferative disease, remains poor [
7]. This highlights the need to identify new genes and pathways that may be involved in the pathophysiology of MM to aid in both prognosis and the development of novel therapeutics.
The C57BL/KaLwRij (KaLwRij) mouse strain, a closely related derivative of the C57BL/6 strain, is one of the best-studied pre-clinical animal models of MM. The KaLwRij strain is susceptible to developing benign monoclonal gammopathy and, in a small proportion of mice, MM at >2 years of age [
8,
9]. Additionally, KaLwRij mice exhibit an inherent ability to support the growth of exogenous malignant plasma cells. The intravenous injection of murine myeloma cell lines, such as the KaLwRij-derived lines 5T33MM, 5T2MM and 5TGM1, into KaLwRij mice results in a myeloma-like disease that closely resembles human MM [
10‐
14]. The mechanisms responsible for this susceptibility to the development of myeloma in this strain of mice remain largely unknown. Importantly, we [
15,
16] and others [
17] have previously utilised this model to identify key genes whose expression may play a role in the development of MM disease in these mice.
We have previously identified pituitary tumour transforming gene 1 (
Pttg1) (also known as securin, EAP1 and TUTR1) as a gene that displays significantly increased expression in KaLwRij mice compared with C57BL/6 controls [
15]. Notably,
PTTG1 is over-expressed in a vast array of malignancies including pituitary [
18,
19], colorectal [
20], thyroid [
21] and lung [
22] cancer, and high levels of PTTG1 are commonly associated with an enhanced proliferative capacity, increased tumour grade and high invasive potential [
23]. PTTG1 is a key regulator of sister chromatid segregation during mitosis and, additionally, is involved in DNA damage repair [
23]. An increase in
PTTG1 expression has previously been described in up to 63 % of MM patients [
24,
25]; however, the role played by PTTG1 in MM disease development has not been determined. In the present study, we confirm up-regulation of
PTTG1 in MM plasma cells from a subset of MM patients compared with both MGUS and healthy controls and show that elevated
PTTG1 expression is associated with an increase in cell cycle-related gene expression and is associated with poor survival. Furthermore, knockdown of
Pttg1 decreases cellular proliferation in vitro and reduces myeloma tumour burden in vivo in the KaLwRij model of MM. Collectively, these data support a role for PTTG1 in promoting MM disease pathogenesis, likely through cell cycle- and proliferation-related pathways.
Discussion
Over-expression of PTTG1 has previously been linked to neoplastic transformation in a wide range of cell types [
23,
30‐
32]. Increased expression of
PTTG1 has been extensively studied in pituitary adenomas, as well as a range of other endocrine cancers (reviewed by [
33]). By comparison, little is known about the role PTTG1 may play in haematological malignancies. Early studies demonstrated that
PTTG1 was highly expressed in approximately 70 % of patients with leukaemia, lymphoma or other myelodysplastic diseases but not in healthy donors [
26,
34]. More recently, high PTTG1 protein expression has been observed in 63 % of MM patients [
24,
25]; however, the biological and prognostic significance of PTTG1 over-expression was not investigated in these studies. In the current study, we show a significant increase in
PTTG1 expression in human MM patients, with approximately 36–70 % of patients showing increased expression of
PTTG1 above that seen in healthy controls, findings which were subsequently confirmed by analysis of our own MM patient specimens. Importantly, increased
PTTG1 expression was associated with poor overall survival, suggesting that high
PTTG1 expression may be implicated in MM disease severity and poor patient outcome. Additionally, our analyses show that PTTG1 is upregulated in MM patients, but not in the asymptomatic precursor MGUS, suggesting an association with disease progression.
Chromosomal rearrangements and duplications are early events in the development of myeloma with the presence of genomic instability being associated with poor prognosis [
35]. In this study, elevated
PTTG1 was specifically observed in a subgroup of MM patients displaying a hyperdiploid phenotype (HY group). This is consistent with the role of PTTG1 in regulating sister chromatid separation during mitosis [
36,
37] as dysregulation of this key function commonly leads to genomic instability [
38‐
42]. Over-expression of PTTG1 promotes dysregulated chromosome segregation resulting in aneuploidy in human cancer cell lines [
40,
41,
43]. The chromosomal instability induced by PTTG1 over-expression has been proposed as a mechanism whereby PTTG1 drives malignant transformation [
44]. Interestingly, gene expression profiling studies have shown that the presence of chromosomal instability in newly diagnosed myeloma patients is associated with a gene expression profile that includes upregulation of
PTTG1 [
35]. These data suggest that expression of
PTTG1 in MM patients may contribute to the development of chromosome duplications characteristic of the HY group.
In addition to its association with hyperdiploidy, we found a significant increase in
PTTG1 expression in subset of patients with a proliferation-related gene expression profile (PR group), which is associated with high-risk disease and poor prognosis [
4]. Elevated
PTTG1 has previously been identified as part of a gene signature that is associated with increased proliferative index and is an independent predictor of poor prognosis in newly diagnosed MM patients [
45]. The increase in
PTTG1 in MM patients with highly proliferative disease is consistent with data from other systems which show that increased expression of
PTTG1 correlates with high levels of cellular proliferation [
46‐
48]. While, in the short term, over-expression of PTTG1 prevents exit from mitosis, leading to cell cycle arrest and increased cell death [
49,
50], sustained; stable over-expression of PTTG1 generally leads to enhanced cellular proliferation [
32,
51,
52]. Notably, our
Pttg1 knockdown studies in the KaLwRij-derived 5TGM1 myeloma plasma cell line resulted in a reduction in cellular proliferation in vitro, as well as decreased tumour development in vivo
. This is consistent with a number of animal knockout and knockdown studies in a range of different cell types, which show that reduction of
PTTG1 expression inhibits cell proliferation [
52‐
59]. Taken together, these data are consistent with increased expression of
Pttg1 being a core requirement for the growth of malignant plasma cells in some patients.
In support of the pro-proliferative role of PTTG1 in myeloma, analysis of four independent MM patient gene expression datasets revealed that the majority (68.9 %) of genes significantly up-regulated in
PTTG1 high patients had proliferation-related functions, specifically cell cycle regulation, DNA replication, mitotic spindle formation, chromosome segregation and DNA damage pathways. These included 20 of the top 50 genes originally identified as being upregulated in the PR subgroup [
4]. Notable exclusions include the cancer/testis antigens
MAGEA1,
MAGEA3,
MAGEA6,
GAGE1,
GAGE2,
GAGE4 and
GAGE5, which were originally identified as being strongly upregulated in the PR subgroup [
4] but are not upregulated in our
PTTG1 high patients. This, combined with the strength of the association between
PTTG1 expression and the expression of cell cycle-associated genes, suggests that the upregulation of proliferation-related genes with high
PTTG1 expression is not simply due to an increased representation of PR patients in this group.
Indeed, we found that knockdown of
Pttg1 was associated with downregulation of cell cycle regulatory genes in the 5TGM1 cell line, suggesting a potential mechanism for the decreased proliferation observed in these cells. Cell cycle genes
CcnB1 (cyclin B1) and
Birc5 (survivin), which were among the most highly up-regulated genes identified in our patient analysis, were down-regulated by approximately 50 % by
Pttg1 knockdown in the 5TGM1 cells. Both
CCNB1 and
BIRC5 have been identified as part of gene expression signatures predictive of high-risk disease and poor prognosis in MM patients in several studies [
4,
5,
45,
60‐
63].
PTTG1,
CCNB1 and
BIRC5 expression are under tight transcriptional control during cell cycle progression, being switched on during G
2/M phase [
46,
49,
64,
65]. While this suggests that downregulation of
Ccnb1 and
Birc5 expression following
Pttg1 could be a consequence of a decrease in G
2/M phase cells, we saw no change in cell cycle distribution in the PTTG1-kd cells to support this. Cyclin B1, an essential regulator of cell cycle transition during mitosis, has previously been identified as a
PTTG1-regulated gene whose up-regulation is associated with enhanced proliferation in keratinocytes [
52]. BIRC5 is an inhibitor of apoptosis proteins (IAP) family member which regulates microtubule dynamics and chromosomal segregation during mitosis. Over-expression and knockdown studies in human myeloma cell lines have implicated BIRC5 in proliferation and protection of MM cells from apoptosis [
62,
64,
66]. Taken together, these data suggest a potential role for cyclin B1 and BIRC5 in PTTG1-mediated tumour growth.
In addition to its role as a securin during chromatid segregation, PTTG1 can directly regulate gene expression. PTTG1 has been shown to physically interact with p53, repressing its transcriptional activity [
36]. Additionally, p53 expression can be transcriptionally and translationally repressed by transient over-expression of PTTG1 [
67]. Notably, some of the genes found to strongly correlate with
PTTG1 expression in this study, including
CDK1,
CCNB1,
CCNB2 and
BIRC5 are known to be inhibited downstream of p53 [
68‐
70]; p53-mediated regulation of gene expression down stream of PTTG1 could be a focus of future studies. However, we found no correlation between PTTG1 and p53 expression in MM patients in any of the datasets examined (data not shown), suggesting that PTTG1 is unlikely to transcriptionally downregulate p53 expression in plasma cells from MM patients. Additionally, our data suggests that
PTTG1 upregulation in MM patients is not a consequence of deletion of the p53 locus 17p13, or associated
TP53 mutations [
71], which is observed in approximately 10 % of patients and is itself associated with poor prognosis [
72‐
74].
In addition to cell cycle genes, we found that expression of
DEPDC1 (Dishevelled, EGL-10, Pleckstrin domain containing 1) was significantly up-regulated in
PTTG1 high patients. Importantly,
Pttg1 knockdown in myeloma cells leads to a 40 % reduction in
DEPDC1 expression. While the function of DEPDC1 is unknown, its expression has been associated with poor prognosis in lung cancer [
75] and advanced disease in breast cancer [
76]. Knockdown of DEPDC1 or inhibition using a specific peptide results in decreased cell proliferation and apoptosis in bladder cancer cell lines [
77,
78]. In myeloma,
DEPDC1 expression has been associated with poor prognosis [
29]. In addition, shRNA-mediated
DEPCD1 knockdown in human myeloma cell lines was shown to significantly inhibit cell proliferation and induce accumulation in G2/M in
TP53 wild-type cells and marked apoptosis in
TP53 mutant cells [
29]. These data suggest that regulation of
DEPDC1 expression may be a mechanism whereby PTTG1 regulates cell proliferation in myeloma.
Although studies have identified a role for PTTG1 in regulating epithelial-mesenchymal transition (EMT) [
79‐
81] and recent studies by Azab and colleagues [
82] have identified a role for the EMT processes in the dissemination and homing of MM plasma cells to the BM, we showed no association between the expression of EMT-related genes and high PTTG1 expression in MM patients. In addition, PTTG1 has previously been shown to directly mediate pro-angiogenic pathways through regulation of secreted factors VEGF and FGF-2 [
19,
83,
84], which in turn are commonly expressed in MM and associated with increased disease severity [
85‐
87]. However, our analyses did not identify an increase in these, or other, secreted pro-angiogenic factors in the presence of increased PTTG1 expression in MM patients. These data suggest that PTTG1 is unlikely to function through EMT or angiogenic pathways in modulating MM disease.
Materials and methods
Mouse tissue and plasma cell isolation
C57BL/6 mice were obtained from the Animal Resources Centre (Perth, Australia). C57BL/KaLwRij mice, originally kindly provided by Prof. Andrew Spencer (Monash University, Melbourne, Australia), were bred and housed at the SA Pathology Animal Care Facility (Adelaide, Australia). Tissues from C57BL/6 and C57BL/KaLwRij mice were snap frozen in liquid nitrogen and homogenised in TRIzol (Life Technologies, Carlsbad, CA). Blood was obtained from mice by cardiac puncture, collected in microfuge tubes containing 50 μL 0.5 M EDTA and centrifuged for 10 min at 500g and the cell pellet resuspended in TRIzol by vigorous vortexing. Femora and tibiae from age- and sex-matched C57BL/6 and C57BL/KaLwRij mice were extracted and cleaned thoroughly. A 21-gauge needle was inserted into the BM cavity, and the BM was flushed with ice-cold PFE (PBS/2 %FCS/2 mM EDTA). The resulting cell suspension was subjected to Ficoll density gradient separation and collected in 10 mL PFE, followed by centrifugation at 300g for 5 min at 4 °C. BM cells were immediately lysed in TRIzol, or CD138+ plasma cells were isolated by FACS. Briefly, cells were resuspended at 1 × 107 cells/mL in PFE and blocked with 110 μg/mL murine gamma globulin (Jackson Laboratories, Bar Harbor, ME) for 30 min at 4 °C. Cells were stained for 30 min at 4 °C protected from light with rat anti-mouse CD138 (R & D Systems, Minneapolis, MN) or an isotype control, washed twice with PFE and stained with secondary goat anti-rat IgG PE (Southern Biotech, Birmingham, AL) for 30 min at 4 °C protected from light. Cells were washed three times with PFE, followed by sorting for CD138+ cells on a FACSAria II (BD Biosciences, San Jose, CA). Total RNA was isolated from sorted cells using an RNAqueous Micro kit (Life Technologies).
Microarray analysis
For comparison of
PTTG1 expression in CD138+ BM plasma cells, isolated by CD138-MACS, from newly diagnosed MM or MGUS patients or normal controls, three independent datasets were used: E-GEOD-6477 (normal,
n = 5; MGUS,
n = 11; MM,
n = 133 [
61]), E-GEOD-16122 (normal,
n = 15; MGUS,
n = 22; MM,
n = 73 [
88]) and E-MTAB-363 (normal,
n = 5; MGUS,
n = 5; MM,
n = 156 [
89]). Analysis of
PTTG1 expression in different gene expression-defined (UAMS) patient subsets was conducted in GSE4581 (
n = 414) [
4]. Analysis of patient survival in
PTTG1 high and
PTTG1 low newly diagnosed MM patients was carried out using GSE4581 analysing patients included in the total therapy 2 (TT2) trial (
n = 256). Four independent datasets were used for analysis of gene expression in PTTG1 high (quartile 4) and PTTG1 low (quartiles 1–3) patient subsets in CD138-selected BM plasma cells from newly diagnosed MM patients: E-GEOD-19784 (
n = 328) [
90], E-GEOD-26863 (
n = 304) [
91], E-MTAB-317 (
n = 226) [
45] and E-MTAB-363 (
n = 156). E-MTAB-363, E-GEOD-26863, E-MTAB-317 and GSE4581 were conducted on Affymetrix GeneChip Human Genome U133 plus 2.0 arrays; E-GEOD-6477 and E-GEOD-16122 were conducted on Affymetrix GeneChip Human Genome U133A arrays. For all datasets except GSE4581, raw microarray data (CEL files) were downloaded from ArrayExpress (EMBL-EBI) and were normalised by RMA using the bioconductor package (affy) [
92] and R (version 3.03) and log
2 transformed. One patient in E-MTAB-363 (V0681) failed quality control (normalised unscaled standard error [NUSE] >1.05) and was excluded, and the remaining 165 files were re-normalised. For GSE4581, MAS5-normalised data were downloaded from the Gene Expression Omnibus (GEO) and were log
2 normalised prior to analysis. GO annotation (
http://www.geneontology.org) and DAVID [
93,
94] were used to classify genes by related function.
MM patient samples
BM trephines and aspirates were collected, with informed consent, from patients with MM at the time of diagnosis and prior to initiation of therapy. This study was approved by the Royal Adelaide Hospital Human Research Ethics Committee (application # RAH 030206 and 131133).
CD138+ plasma cell isolation from MM patients
CD138
+ plasma cells were isolated from diagnostic MM patient BM samples using CD138 microbeads (Miltenyi Biotech, Auburn, CA) as previously described [
15]. Briefly, cryopreserved human BM samples were thawed into 10 mL DMEM (high glucose) with 15 % FCS and DNase I (80 U/mL). Samples were centrifuged at 300
g for 10 min and supernatant removed. The cell pellet was resuspended in MACS buffer (2 mM EDTA, 0.5 % deionised BSA in PBS) and CD138 microbeads added, followed by incubation on ice for 15 min. Cell-bead conjugates were washed in 1 mL MACS buffer and centrifuged at 300
g for 10 min. Cells were resuspended in MACS buffer, applied to a pre-rinsed MS column and washed three times with MACS buffer and eluted in 1 mL. Subsequent FACS analysis confirmed >85 % CD138
+ following MACS. Total RNA was isolated using an All Prep DNA/RNA Micro Kit (Qiagen, Valencia, CA).
Dual colour staining BM trephines
Paraffin-embedded BM trephine sections were mounted on silicane-coated slides and dried. Endogenous peroxidase was blocked with 0.5 % H2O2 in methanol at room temperature for 30 min, followed by blocking with 3 % normal horse serum (NHS) for 30 min. Slides were incubated with anti-PTTG1 antibody (diluted 1:50; DCS-280; Abcam) at room temperature overnight. Slides were washed twice in PBS and incubated with biotinylated anti-mouse IgG (Vector Laboratories, Burlingame, CA) diluted 1:250 for 30 min at room temperature, washed in PBS and incubated with streptavidin Alexa Fluor 488 (diluted 1:500; Life Technologies) diluted 1:500 for 1 h at room temperature. Slides were re-blocked with 3 % NHS for 30 min and incubated with mouse anti-human CD138 (diluted 1:40; MI15, Dako, Denmark) overnight. Slides were washed twice with PBS followed by incubation for 1 h at room temperature with anti-mouse Alexa Fluor 594 (diluted 1:500, Life Technologies) and mounted in aqueous mounting solution. Images were taken on a Zeiss LSM 700 confocal system (Zeiss, Oberkocken, Germany) at ×40 magnification.
Cell lines
The murine myeloma plasma cell line 5TGM1 was kindly provided by Ass. Prof. Claire Edwards (University of Oxford, Oxford, UK), and was maintained in Iscove’s modified Dulbecco’s medium (Sigma) with 20 % FCS. Unless otherwise specified, all culture medium was supplemented with 2 mM l-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 10 mM HEPES buffer (Life Technologies) and 1 mM sodium pyruvate.
Generation of stable PTTG1 knockdown lines
To generate stable knockdown cell lines, an RNA duplex targeting murine
Pttg1 (GGGAAATTGCAGGTTTCAACG) was cloned into the pFIV-H1-mCherry vector. A scrambled sequence was used as a control. pFIV-H1-mCherry was created by excising the
GFP cassette from pFIV-H1-GFP (System Biosciences, Mountain View, CA) using XbaI and SalI and replacing it with the
mCherry cassette from pMSCV-mCherry. Following lentiviral infection of 5TGM1-luc cells (expressing a dual GFP and luciferase reporter construct [
16,
96,
97]), single-cell clones were generated from the top 10 % GFP- and mCherry-expressing cells using preparative cell sorting and the automatic cell deposition unit on a Beckman Coulter Epics Altra HyperSort, using Expo MultiComp Software version 1.2B (Beckman Coulter, Miami, FL). Clonal 5TGM1-PTTG-kd and 5TGM1-SCRAM lines were used for subsequent in vitro and in vivo assays.
Real-time PCR
Total RNA was isolated using TRIzol (Life Technologies) as per the standard protocol (unless otherwise specified). For mouse and human CD138
+ plasma cells, RNA was reverse transcribed using Sensiscript (Qiagen). For all other tissues and cell lines, RNA (1 μg) was reverse transcribed with Superscript III (Life Technologies) as per the manufacturer’s protocol. Real-time PCR was conducted on the Corbett Rotorgene using the following primers: human β-actin (F: 5'-TTGCTGACAGGATGCAGAAG-3' and R: 5'-AAGGGTGTAAAACGCAGCTC-3'), human
PTTG1 (F: 5'-CGGCCTCAGATGAATGCGGCT-3' and R: 5'-TTGATTGAAGGTCCAGACCCCAGC-3'), mouse
Gapdh (F: 5'- ACCCAGAAGACTGTGGATGG-3' and R: 5'-CAGTGAGCTTCCCGTTCAG-3'), mouse β-actin (F: 5'-TTG CTGACAGGATGCAGAAG-3' and R: 5'-CAGTGAGCTTCCCGTTCA-3'), mouse
Pttg1 (F: 5'-GCTCCTGATGATGCCTACCC-3' and R: 5'-CGCCATTCAAGGGGAGAAGT-3'), mouse
Ccnb1 (F: 5'-GATGATGGGGCTGACCCAAA-3' and R: 5'-ACATGGTCTCCTGAAGCAGC-3'), mouse
Cdk1 (F: 5'-GTCCGTCGTAACCTGTTGAG-3' and R: 5'-TGACTATATTTGGATGTCGAAG-3') [
98], mouse
Rrm2 (F: 5'-GATTTAGCCAAGAAGTTCAAGTTACAG-3' and R: 5'-TCACACAAGGCATAGTTTCAATAGC-3') [
99], mouse
Birc5 (F: 5'-GAACCCGATGACAACCCGAT–3' and R: 5'-TGGTCTCCTTTGCAATTTTGTTCT-3') and mouse
Depcd1 (F: 5'-AGCTGCAGTGGAGAAACATCT-3' and R: 5'-TGGTCTCCTTTGCAATTTTGTTCT-3'). Gene expression was represented relative to β-actin or
Gapdh expression, calculated using the 2
-ΔCT method.
Western blot analysis
Cells were lysed in lysis buffer (1 % NP-40, 20 mM HEPES, 150 mM NaCl, 10 % glycerol, 2 mM Na3VO4, 10 mM Na4P2O7, 2 mM NaF and Complete EDTA-free Protease Inhibitor Cocktail (Roche, Mannheim, Germany)). Total lysate (10 μg) was loaded on an 11 % acrylamide gel and subjected to SDS-PAGE. Proteins were transferred to PVDF membrane overnight. Membrane was incubated for 1 h in blocking buffer (Tris-buffered saline containing 0.1 % Tween20 and 2.5 % ECL Blocking Agent (GE Healthcare, Little Chalfont, UK)) and for 2 h at RT with mouse monoclonal anti-PTTG1 antibody (DCS-280; Abcam; Cambridge, MA) diluted 1:1000 in blocking buffer, followed by alkaline phosphate-conjugated anti-mouse IgG (Millipore, Billerica, MA) diluted 1:4000 in blocking buffer for 1 h at RT. Proteins were visualised using ECL detection reagent (GE Healthcare) on a Typhoon FLA 7000 IP2 (GE Healthcare).
Proliferation assays
For WST-1 assays, 5TGM1 cells (PTTG-kd or SCRAM controls) were seeded at 1 × 105 cells/well in triplicate in 96-well plates and were incubated at 37 °C with 5 % CO2. At 24-h intervals, WST-1 reagent (Roche) was added to the cells and incubated for 2 h prior to reading absorbance at 450 nm.
For BrdU assays, 5TGM1 cells were seeded at 4 × 105 cells/well in triplicate in a 96-well plate. BrdU (Roche) was immediately added to the cells and incubated for 2 h at 37 °C with 5 % CO2. BrdU incorporation was measured using a BrdU Cell Proliferation ELISA kit (Roche) as per manufacturer’s protocol and absorbance measured at 370 nm.
For cell cycle analysis, cells were seeded at 4 × 105 cells/well in a six-well plate and incubated for 24 h at 37 °C with 5 % CO2. Cells were fixed in ice-cold 70 % (v/v) ethanol, washed twice in PBS and stained with propidium iodide (PI; 40 μg/mL; Sigma) containing 20 μg/mL RNase A (Qiagen) for 30 min prior to analysis on a Gallios flow cytometer (Beckman Coulter). Cell cycle distribution was analysed using FCS Express version 4.
Animals
Ethical approval for this study was obtained from the SA Pathology/Adelaide Health Service Animal Ethics Committee (application # 136/10). C57BL/KaLwRij mice at 6–8 weeks of age were injected with 5 × 10
5 luciferase-expressing 5TGM1-luc cells (5TGM1-SCRAM or 5TGM1-PTTG-kd) in 100-μl sterile PBS via the tail vein. At weekly intervals, mice were administered luciferin (150 mg/kg) i.p. and imaged using the Xenogen IVIS 100 bioluminescence imaging system (Caliper Life Sciences, Hopkinton, MA) until termination of the experiment at day 28, prior to the development of lethal disease (hind limb paralysis) [
16,
96]. Total tumour burden was measured as total flux (photons/second) for each animal using Living Image software (PerkinElmer, Waltham, MA), as described previously [
15,
16,
97,
100‐
103].
Statistical analyses
Statistical analysis was performed using GraphPad Prism version 6.03 for Windows (GraphPad Software, San Diego, CA). Variance between patient groups was assessed using Kruskal-Wallis tests with Dunn’s multiple comparison tests. In each of the E-GEOD-19784, E-GEOD-26863, E-MTAB-317 and E-MTAB-363 datasets, gene expression was compared in PTTG1 high and PTTG1 low patients, using t tests with Bonferroni’s correction for multiple testing and p values from the four datasets were combined using Fisher’s method. Survival curves were compared using the log-rank (Mantel-Cox) test with hazard ratios calculated using the Mantel-Haenszel calculation. In vivo data, WST-1 assays and cell cycle distribution were analysed by two-way ANOVA with Sidak’s multiple comparison tests. qRT-PCR data and BrdU incorporation were compared between groups using unpaired two-tailed t tests. A p value of 0.05 was considered statistically significant. Unless otherwise described, all plots depict mean + SEM of three independent experiments.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JN performed and designed experimental work and wrote the paper; KV performed in silico analyses and experimental work and wrote the paper; DH performed experimental work and wrote the paper; RB, SW and KM performed experimental work; CK performed in silico analyses; AZ facilitated research and wrote the paper. All authors read and approved the final manuscript.