Abstract
Background:
Gastrointestinal stromal tumour (GIST) is mainly initialised by receptor tyrosine kinase gene mutations. Although the tyrosine kinase inhibitor imatinib mesylate considerably improved the outcome of patients, imatinib resistance still remains a major therapeutic challenge in GIST therapy. Herein we evaluated the clinical impact of microRNAs in imatinib-treated GISTs.
Methods:
The expression levels of microRNAs were quantified using microarray and RT–qPCR in GIST specimens from patients treated with neoadjuvant imatinib. The functional roles of miR-125a-5p and PTPN18 were evaluated in GIST cells. PTPN18 expression was quantified by western blotting in GIST samples.
Results:
We showed that overexpression levels of miR-125a-5p and miR-107 were associated with imatinib resistance in GIST specimens. Functionally, miR-125a-5p expression modulated imatinib sensitivity in GIST882 cells with a homozygous KIT mutation but not in GIST48 cells with double KIT mutations. Overexpression of miR-125a-5p suppressed PTPN18 expression, and silencing of PTPN18 expression increased cell viability in GIST882 cells upon imatinib treatment. PTPN18 protein levels were significantly lower in the imatinib-resistant GISTs and inversely correlated with miR-125a-5p. Furthermore, several microRNAs were significantly associated with metastasis, KIT mutational status and survival.
Conclusions:
Our findings highlight a novel functional role of miR-125a-5p on imatinib response through PTPN18 regulation in GIST.
Similar content being viewed by others
Main
Gastrointestinal stromal tumour (GIST) is the most common mesenchymal tumour of the gastrointestinal tract with an incidence of 1–2 cases/100 000 inhabitants per year (Nilsson et al, 2005). The main initial event in GIST tumorigenesis is gain-of-function mutations of the receptor tyrosine kinase genes KIT or PDGFRA. These mutations occur in approximately 85% of GISTs (Corless et al, 2004) and cause ligand-independent activation of the encoded receptor tyrosine kinases that results in aberrant and uncontrolled cell growth and tumour formation (Corless et al, 2004). The development of imatinib mesylate, a small-molecule inhibitor that targets several receptor tyrosine kinases, including KIT and PDGFRA, markedly improved patient outcome. However, the efficacy of imatinib treatment varies depending on the mutated domains of KIT and PDGFRA (Gramza et al, 2009). Still, 10% of GIST patients have primary resistance to imatinib, and 40–50% of initially responding patients develop secondary resistance and progress within 2 years (Gramza et al, 2009).
Previous studies have suggested that acquisition of second site mutations in the kinase or loop domain of KIT or PDGFRA is the main mechanism of resistance to imatinib therapy in GIST (Gramza et al, 2009). However, second site mutations are not detected in a substantial proportion of imatinib-resistant GISTs (Antonescu et al, 2005; Wardelmann et al, 2006), suggesting the involvement of additional mechanisms. Furthermore, the mechanism of primary resistance remains unclear. Therefore, a better understanding of the underlying mechanisms in imatinib resistance is needed for improved treatment planning and development of additional treatment strategies.
microRNAs (miRNA) are short RNAs of ∼22 nucleotides in length that regulate many physiological and pathological processes, by forming imperfect hybrids with the 3′ untranslated region sequences of their target mRNAs that leads to mRNA degradation and/or translational repression (Bartel, 2004). Distinct miRNA expression signatures have been found in a variety of cancers, and they are often associated with patient outcome, suggesting their potential use as diagnostic and prognostic markers. Recent studies have shown their importance in drug resistance of different cancer types (Zheng et al, 2010). In GIST, specific miRNA expression signatures are associated with chromosome 14q loss (Choi et al, 2010; Haller et al, 2010), anatomical site (Choi et al, 2010; Niinuma et al, 2012), KIT or PDGFRA mutation (Haller et al, 2010), tumour risk (Choi et al, 2010; Niinuma et al, 2012) and overall survival (Niinuma et al, 2012). However, no study has addressed the functional role of miRNAs in imatinib response in GISTs, and it is unclear whether miRNAs could be used as prognostic markers for GIST patients who received neoadjuvant imatinib therapy.
In this study, we evaluated the clinical impact of miRNA deregulation in relation to imatinib response, clinical features and KIT mutational status of GISTs. We further investigated the functional role of miR-125a-5p and its potential targets in imatinib resistance of GIST cells and assessed the target protein expression levels in GIST clinical specimens.
Materials and methods
Clinical samples
A total of 30 snap-frozen tumours from 24 GIST patients who had received neoadjuvant imatinib treatment were included in this study. All tissue samples were collected at the Karolinska University Hospital Biobank between 2003 and 2012. All samples were obtained with informed consent, and study of the tissue materials was approved by the Karolinska Institutet ethical committee at Karolinska Hospital, Stockholm, Sweden. The patients were followed up until September 2012 or the time of death. The clinical, histopathological and follow-up details of all cases are detailed in Table 1, and a subset of the cases were included in a previous study (Ahlen et al, 2012). The diagnosis of GIST was based on the routine histopathological examination and positive immunoreactivity for CD117 (KIT). Clinical imatinib response was evaluated by fluorodeoxyglucose positron emission tomography and/or contrast-enhanced computed tomography for (i) tumour size, (ii) tumour characteristics, such as tumour density, glucose uptake, and intra-tumour alterations (e.g., addition of denser subparts of the tumour) and (iii) peritoneal and/or liver metastasis. Tumours that continued to progress under imatinib treatment were classified as imatinib resistant, while the tumours that shrunk obviously or showed significant cystic changes or necrosis were grouped as imatinib sensitive.
GIST cell lines
Two GIST cell lines (GIST882 and GIST48) were kindly provided by Dr Jonathan Fletcher at Brigham and Women’s Hospital, Boston, MA, USA. The GIST882 cell line was established from an untreated GIST with a homozygous missense mutation in KIT exon 13 (K642E), and it is imatinib sensitive (Tuveson et al, 2001). GIST882 cells were grown in RPMI 1640 media supplemented with 15% fetal bovine serum, 2 mM L-glutamine and 1% penicillin–streptomycin. The GIST48 cell line was established from a progressed GIST with a primary homozygous KIT exon 11 missense mutation (V560D) and a heterozygous KIT secondary exon 17 (D820A) mutation, and it is imatinib resistant (Bauer et al, 2006). GIST48 cells were grown in F-10 media supplemented with 15% fetal bovine serum, 2.5 μg ml−1 MITO+ serum extender (BD Biosciences, Franklin Lakes, NJ, USA), 5 μg ml−1 bovine pituitary extract (BD Biosciences), 2 mM L-glutamine and 1% penicillin–streptomycin. Both cell lines were cultured at 37 °C with 5% CO2. The authentication of the cell lines was confirmed by KIT mutation analysis and CD117 immunoreactivity (Berglund et al, 2013). The possibility of cross-contamination by other cell lines was ruled out by short tandem repeat (STR) profiling performed by Bio-Synthesis Inc. (Lewisville, TX, USA). The STR profiles of both cell lines are shown in Supplementary Table S1.
DNA and RNA extraction
Genomic DNA was purified using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany), and total RNA isolation was performed using the mirVana miRNA Isolation Kit (Life Technologies, Carlsbad, CA, USA). DNA and RNA concentrations were measured using the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA).
KIT and PDGFRA mutation analysis
All GIST samples were analysed for KIT (exons 9, 11, 13 and 17) and PDGFRA (exons 12 and 18) mutations using a direct sequencing method. Mutation analyses of 12 samples had been performed in the routine pathology diagnostic laboratory at Karolinska University Hospital, while 18 samples were analysed in this study. Primer sequences for amplification of genomic DNA were previously described (Sihto et al, 2005) and are available in Supplementary Table S2. PCR amplifications were performed in 25 μl reactions containing 1 unit Platinum Pfx DNA Polymerase (Life Technologies), 1 × PCR buffer, 0.2 mM dNTPs, 2 mM MgSO4, 0.2 μ M forward and reverse primers and 250 ng genomic DNA. A touch-down protocol was used with the following conditions: 2 min at 94 °C; 10 cycles of (30 s at 94 °C, 45 s at 66 °C–56 °C (by decreasing the annealing temperature 1 °C per cycle) and 30 s at 68 °C); 35 cycles of (30 s at 94 °C, 1 min at 56 °C and 30 s at 68 °C); followed by extension for 7 min at 68 °C. The PCR products were cleaned up using the ExoSAP-IT (USB Corporation/Affymetrix, Santa Clara, CA, USA) and sequenced at the KIGene facility (Stockholm, Sweden).
Genome-wide miRNA expression profiling
Global miRNA expression profiling was performed in 17 GISTs using the Human Agilent’s miRNA microarray system (Agilent, Santa Clara, CA, USA) with probes matching 903 human miRNAs (miRBase release 14). Array hybridisations and data analyses were performed as previously described (Caramuta et al, 2010). Cluster 3.0 software (http://bonsai.hgc.jp/~Bmdehoon/software/cluster/software.htm) was used for normalisation and median centering. Normalised miRNAs with <20% missing values were included for Significance Analysis of Microarrays (SAM) (http://www-stat.stanford.edu/Btibs/SAM/) and clustering analysis. Visualisation of the clustering data was done using Java TreeView (http://jtreeview.sourceforge.net). The data from microarray analysis have been deposited at NCBI Gene Expression Omnibus (GEO accession number, GSE45901).
TaqMan reverse transcription quantitative PCR (RT–qPCR) analysis
The expression of 10 selected mature miRNAs, i.e., miR-301a-3p, miR-30c, miR-134, miR-1915, miR-125a-5p, miR-365, miR-107, miR-150-3p, miR-1207-5p and miR-638, as well as PTPN18 and STARD13 mRNAs, was quantified using commercially available TaqMan RT–qPCR assays (Life Technologies). cDNAs were synthesised from 40 ng total RNA using TaqMan MicroRNA Reverse Transcription Kit (Life Technologies), or from 100 ng total RNA using High Capacity cDNA Reverse Transcriptase Kit (Life Technologies). RT–qPCR was performed using the TaqMan Universal PCR Master Mix (Life Technologies) in a 7900HT Real-time PCR System (Life Technologies). RNU6B and 18S rRNA was used as endogenous controls for miRNAs and mRNAs, respectively. All reactions were performed in triplicate, and relative expression levels were determined with the ΔCT method and reported as 2−ΔCT.
miRNA transfection in GIST882 and GIST48 cells
GIST882 and GIST48 cells were transfected using Nucleofector technology (Lonza AG, Basel, Switzerland). In brief, 4 × 106 cells were resuspended in 100 μl Nucleofector solution V and mixed with 100 pmol of miRNA mimic (pre-miR-125a-5p, pre-miR-211 or pre-miR-944), pre-miR Negative control (pre-miR-CTR), miRNA inhibitor (anti-miR-125a-5p) or anti-miR Negative control (anti-miR-CTR) (Life Technologies). Cells were then electroporated using the program T-20, seeded in T-25 flasks with 4 ml media and cultured at 37 °C for 48 h before imatinib cytotoxicity assay. Transfection efficiency was determined by measuring the cellular miR-125a-5p levels using RT–qPCR (Supplementary Figure S1A).
Transfection of short hairpin RNAs (shRNAs)
shRNA against human PTPN18 (target sequence: 5′-GAGGGACACAGCGACTACA-3′; Gensler et al, 2004) was cloned into pcDNA3-U6M2 vector as described previously (Taft et al, 2011). Primers are included in Supplementary Table S2. A total of 3 × 106 GIST882 cells were transfected with 2 μg of plasmid DNA containing the insert (shPTPN18) or without the insert (shControl) as a negative control, using the Amaxa nucleofection system (Lonza AG) 48 h before imatinib treatment. Transfection efficiency was determined by measuring the cellular PTPN18 levels using RT–qPCR and western blotting analyses (Supplementary Figure S1B).
Imatinib cytotoxicity assay
The effect of miR-125a-5p overexpression or inhibition on imatinib response was evaluated using the WST-1 colorimetric assay (Roche Applied Science, Mannheim, Germany), as previously described (Caramuta et al, 2010). After 48 h of transfection, 30 000 cells per well were seeded in 96-well plates and were left either untreated (as mock control) or treated with three different concentrations of imatinib (0.05 μ M, 0.1 μ M and 0.5 μ M for GIST882; 1 μ M, 5 μ M and 10 μ M for GIST48). All experiments were conducted in five wells for each condition and replicated at least three times in independent experiments. The percentage of surviving cells was calculated by comparing the absorbance values of the samples after background subtraction and normalised to the respective cells without imatinib treatment. Imatinib mesylate was kindly provided by Novartis Pharma (Basel, Switzerland).
Target prediction for miR-125a-5p
The potential mRNA targets of miR-125a-5p were predicted by miRecords (http://mirecords.biolead.org/). The targets that are predicted by at least six programs were chosen for subsequent analyses. Functional annotation and gene ontology analyses were performed by GeneCodis3 (http://genecodis.cnb.csic.es/), using a combination of GO biological process and molecular function parameters.
Western blotting analysis
Cells and frozen tissues were homogenised in NP-40 Lysis buffer (Life Technologies) with addition of protease inhibitor (Complete protease inhibitor cocktail; Roche) and 1 mM of phenylmethanesulfonyl fluoride (Sigma-Aldrich, St Louis, MO, USA). Quantification was done using Pierce BCA Protein Assay Kit (Thermo Scientific, Waltham, MA, USA). Lysates (60 μg for cells and 100 μg for tissues) were separated by Novex 4%–12% NuPAGE Bis-Tris gels (Life Technologies) and transferred to nitrocellulose membranes (Life Technologies). After blocking with 5% non-fat milk diluted in TBS, membranes were incubated with anti-PTPN18 antibody (No. 8311, Cell Signaling Technology, Danvers, MA, USA) at 1 : 1000 dilution, followed by anti-rabbit IgG-HRP (1 : 3000; No. 170-6515; Bio-Rad Laboratories, Hercules, CA, USA), or with anti-StARD13 (1 : 200; sc-67843; Santa Cruz Biotechnology, Santa Cruz, CA, USA), followed by anti-goat IgG-HRP (1 : 5000; 31402; Thermo Scientific). Detection was carried out with Novex ECL HRP chemiluminescent substrate reagent (No. WP20005; Life Technologies). Precision Plus Protein All Blue Standards (No. 161-0373; Bio-Rad Laboratories) was used to determine the relative molecular weights. Protein levels were quantified on immunoblots using the Image Gauge v.4.0 software (FujiFilm Science Lab, Tokyo, Japan). Incubation of the membranes with a GAPDH antibody (sc-47724; Santa Cruz Biotechnology) at 1 : 5000 was performed for normalisation purposes.
Generation of imatinib-resistant clones (GIST882R)
GIST882 cells were exposed to 1 μ M imatinib for 7 days, leading to the cell death of 60–70% of cells. The remaining cells were harvested and continuously grown in growth media containing 1 μ M imatinib. Proteins and RNAs were extracted after 1 month of growth with 1 μ M imatinib.
Statistical analysis
All statistical tests were performed in Statistica 8.0 (StatSoft Inc., Tulsa, OK, USA) or MS Office Excel 2007 (Albuquerque, NM, USA), unless otherwise stated. Unpaired Student’s t-test was conducted to compare the miRNA expression levels in different groups, and paired Student’s t-test was performed to analyse transfection experiments. Association of clinico-pathological features with imatinib response and miRNA expression was evaluated using the Fisher’s exact test. The inter-relationship of miRNA expression with survival was studied using the Kaplan–Meier plots, and significant differences between curves were evaluated using the log-rank test. Seventeen cases for which the primary tumour was studied were included in the survival analysis. Recurrences were excluded from the survival analyses as they may carry alterations that are not representative for the primary tumour status and may have different impact on patient survival. Furthermore, two cases with multiple tumours (GIST9 and GIST10) were excluded because of high variation in miRNA expression between individual tumours. Tumours were classified into different groups with high or low expression of each miRNA according to median expression level. All P-values obtained in this study were two-tailed, and P-values ⩽0.05 were considered as significant.
Results
Relationship between imatinib response and clinical/genetic features of GISTs
Our sample cohort consisted of 30 GIST samples from 24 patients who received neoadjuvant imatinib treatment (Table 1). Fourteen tumours were imatinib resistant, whereas 16 tumours were imatinib sensitive. GIST9 and GIST10 had four tumours each: three of the tumours in each patient showed progressive growth (imatinib resistant) and one tumour in each patient (GIST9-3 and GIST10-3) partially responded to the treatment (imatinib sensitive). Of the 24 patients, 14 had developed tumour metastasis and the remaining patients had no recurrence or metastasis during the follow-up. Thirteen of the tumours harboured a single non-synonymous mutation in KIT and 11 tumours had double KIT mutations. The GIST15 tumour harbouring a silent mutation in KIT (P585P) was considered as wild type, three tumours had a single mutation in PDGFRA, whereas two were wild type for both KIT and PDGFRA. We evaluated the relationship between imatinib response and clinical/genetic characteristics, including age, gender, metastasis and KIT mutational status, in our cohort. As shown in Supplementary Table S3, imatinib resistance was significantly correlated with metastasis (P=0.007; Fisher’s exact test) but not with double KIT mutations, age or gender.
Distinct miRNA expression patterns in relation to imatinib response and metastasis
We characterised miRNA expression patterns of 17 tumours (10 imatinib sensitive and 7 imatinib resistant) using a microarray-based platform. After normalisation and filtering, 155 miRNAs were included for unsupervised clustering analysis, which is based on similar miRNA expression patterns without prior knowledge of sample identity. The analysis classified the samples into two main subgroups (Figure 1A and Supplementary Figure S2). All imatinib-resistant tumours, except 1 (GIST2), were grouped in cluster 1; while 7 out of the 10 imatinib-sensitive GISTs were found in cluster 2.
miRNA expression levels in GIST samples. (A) Dendrogram based on unsupervised clustering of miRNA expression profiles by microarray in 17 GISTs. (B) Box plots illustrate the relative expression levels of miR-125a-5p and miR-107 in 14 imatinib-resistant and 16 imatinib-sensitive GIST specimens regardless of KIT/PDGFRA mutational status by RT–qPCR after normalisation to RNU6B. (C) Box plots show the relative expression levels of miR-125a-5p, miR-107, miR-134, miR-301a-3p and miR-365 in four imatinib-resistant and nine imatinib-sensitive tumours among GISTs carrying a single KIT mutation. (D) Relative miR-125a-5p levels in GIST882 (imatinib-sensitive) and GIST48 (imatinib-resistant) cell lines. Differences between groups were calculated using the unpaired Student’s t-test, and P⩽0.05 were considered significant. Error bars refer to s.d. between replicates.
Eleven of the 17 GISTs were from patients with metastasis detected either at diagnosis or during follow-up. The other six cases did not have any metastasis during follow-up of at least 25 months (range 25–98); five of which were still alive at the end of follow-up and one died of unrelated cause. The unsupervised clustering analysis grouped 9 out of the 11 metastatic GISTs in cluster 1, and all 6 non-metastatic GISTs in cluster 2 (Figure 1 and Supplementary Figure S2). The three imatinib-sensitive samples (i.e., GIST11, GIST12 and GIST20) that clustered together with imatinib-resistant tumours had metastasis at diagnosis. Notably, GIST20 showed two distinct areas with tumour necrosis or viable tumour cells in pathological evaluation, indicating that some imatinib-resistant tumour cells were present in this tumour. The survival analysis did not show any significant differences between the two clusters (data not shown).
miRNAs associated with imatinib resistance
Using SAM analysis on the microarray data, we found 27 overexpressed and 17 underexpressed miRNAs with a false discovery rate (FDR) <15% in the imatinib-resistant GISTs compared with imatinib-sensitive GISTs (Supplementary Table S4).
To verify the array results and to evaluate the significance of the findings, we performed RT–qPCR analyses on 10 selected miRNAs from among the differentially expressed miRNAs in a larger cohort of 30 GIST samples. The selected miRNAs consisted of five overexpressed (miR-125a-5p, miR-301a-3p, miR-30c, miR-365 and miR-107) and five underexpressed miRNAs (miR-150-3p, miR-134, miR-1915, miR-638 and miR-1207-5p) in imatinib-resistant GISTs compared with imatinib-sensitive GISTs. These miRNAs were selected among those with the highest score in the SAM analysis or because of their involvement in drug resistance. We found that the expression of miR-125a-5p and miR-107 was significantly higher (fold changes=2.84 and 2.39, respectively) in the imatinib-resistant tumours (P=0.032 and P=0.041, respectively; t-test) (Figure 1B). However, in the two patients (GIST9 and GIST10) with multiple tumours harbouring double mutations in KIT, the expression levels of these two miRNAs were not obviously different between the resistant and sensitive tumour pairs (Supplementary Figure S3). The remaining eight miRNAs were not significantly different between the two groups (data not shown).
Given that GISTs harbouring double KIT mutations or wild-type KIT and PDGFRA are known mechanisms of imatinib resistance, we performed an independent evaluation of the samples harbouring only a single KIT mutation (all in exon 11). We showed that expression levels of miR-125a-5p (4.7-folds; P=0.029; t-test), miR-107 (4.1-folds; P=0.031), miR-134 (38.1-folds; P=0.015), miR-301a-3p (4.4-folds; P=0.044) and miR-365 (4.2-folds; P=0.05) were significantly higher in the imatinib-resistant tumours among the GISTs with a single KIT mutation (Figure 1C).
We also measured the expression of miRNAs in GIST cell lines. miR-125a-5p expression level was higher in the imatinib-resistant cell line (GIST48) and lower in the imatinib-sensitive cell line (GIST882) (P=0.05; t-test), which corresponded well with its higher level in the imatinib-resistant GISTs. However, the expression pattern of other miRNAs in the GIST cell lines did not corroborate with the expression pattern observed in the clinical samples (Supplementary Figure S4). Given that miR-125 family members are shown to confer drug resistance (Zhou et al, 2010) and regulate cell death pathways in cancer cells (Kim et al, 2012), we further investigated the functional role of miR-125a-5p in GIST cells.
Effects of miR-125a-5p expression on imatinib response in GIST cell lines
Based on our results in clinical samples, we hypothesised that miR-125a-5p could act as a modulator of imatinib response among GISTs with a single KIT mutations. We next evaluated the functional consequences of miR-125a-5p expression on imatinib response in GIST cell lines with a single (GIST882) and double (GIST48) KIT mutations. As shown in Figure 2A, overexpression of miR-125a-5p in GIST882 cells resulted in a significantly higher (up to 26%) cell viability compared with the cells treated with pre-miR-negative control after 24 h of imatinib treatment, while its suppression caused a significantly lower (up to 9%) cell viability compared with the cells treated with anti-miR-negative control. However, overexpression or suppression of miR-125a-5p expression in GIST48 cells did not result in any significant change in cell viability compared with the negative control (Figure 2A). Our results may indicate that miR-125a-5p can modulate imatinib response of GIST cells harbouring a single KIT mutation but not in GIST cells harbouring KIT double mutations. To further support the specificity of miR-125a-5p effect on imatinib response, we also evaluated the effects of two additional miRNAs (miR-944 and miR-211), which were not differentially expressed between imatinib-sensitive and -resistant GISTs, as negative controls. Modulation of miR-944 and miR-211 did not result in any significant change in imatinib response of GIST882 cells (Supplementary Figure S5).
Effects of miR-125a-5p modulation on imatinib response and candidate targets in GIST cells. (A) Proportion of surviving cells in cultures transfected with pre-miR-125a-5p, anti-miR-125a-5p or their respective negative controls after 24 h of treatment with different concentrations of imatinib, as evaluated by WST-1 assay. (B) Western blotting analysis of miR-125a-5p candidate targets in GIST882 cells treated with pre-miR-125a-5p or pre-miR-CTR. Western blots show the protein expression levels of PTPN18 and STARD13 in three independent transfection experiments. Bar graphs showing the fold change of protein levels relative to pre-miR-CTR-treated cells after normalisation to GAPDH. Data represent the mean of three independent experiments, and the error bars refer to s.d. from the mean. Statistical differences between the two groups were calculated using the paired Student’s t-test, and P⩽0.05 was considered significant. Abbreviations: FC=fold change; NS=not significant.
Analysis and evaluation of miR-125a-5p targets in GIST
To better understand the possible role of miR-125a-5p on imatinib response, we applied a computational tool (miRecords) to identify the candidate targets of miR-125a-5p. The analysis revealed 41 common targets predicted by at least six different target prediction programs (Supplementary Table S5). The predicted targets were then subjected to functional annotation using GeneCodis3. We found several functional groups that were significantly enriched (P<0.05), including anti-apoptosis, cell cycle, cell differentiation, transcription, development, protein tyrosine kinase activity, protein phosphorylation and signal transduction (Supplementary Table S6). Among the predicted targets, we selected PTPN18 and STARD13 for further validations because of their involvement in KIT or tyrosine kinase signalling and their regulations by miR-125 family members (Guo et al, 2012; Tang et al, 2012). We evaluated the effect of miR-125a-5p overexpression on PTPN18 and STARD13 protein levels using western blotting analysis. As shown in Figure 2B, we observed that both PTPN18 (P=0.006; t-test) and STARD13 (P<0.001; t-test) protein levels were decreased upon overexpression of miR-125a-5p in GIST882 cells, suggesting that they are potential targets of miR-125a-5p in GIST cells.
To strengthen the functional evidence of miR-125a-5p in imatinib resistance, we generated imatinib-resistant subclones of GIST882 cells and compared the expression level of miR-125a-5p and its targets in imatinib-resistant GIST882R subclone with its imatinib-sensitive parental cell line. We observed an increase of miR-125a-5p expression in the imatinib-resistant GIST882R cells (P=0.018; t-test). PTPN18 and STARD13 expression levels were not significantly different at mRNA level (Figure 3A), but the protein expression levels of PTPN18 (P=0.006; t-test) and STARD13 (P=0.004; t-test) were significantly lower in the GIST882R cells compared with its parental cell line (Figure 3B).
Quantification of miR-125a-5p , PTPN18 and STARD13 expression levels in GIST882 parental cells and its imatinib-resistant subclone (GIST882R). (A) Quantification of miR-125a-5p, PTPN18 and STARD13 expression levels by RT–qPCR after normalisation to RNU6B or 18S rRNA. (B) Quantification of PTPN18 and STARD13 protein expression levels by western blotting analysis. GAPDH was used as a normalisation control. Data indicate mean values of three independent replicates, and error bars refer to s.d. between replicates. Statistical differences between the two groups were calculated using the paired Student’s t-test, and P⩽0.05 was considered significant.
We also measured the expression levels of these candidate targets in 27 GIST samples using western blotting analysis. As shown in Figure 4, we observed that PTPN18 protein levels were significantly lower in imatinib-resistant as compared with imatinib-sensitive samples (P=0.003; t-test) and inversely correlated with miR-125a-5p levels (r=−0.40; P=0.036). STARD13 was not significantly different between imatinib-resistant and -sensitive tumours (P=0.18; t-test), and no correlation was found with miR-125a-5p expression (r=−0.11; P=0.55).
Quantification of PTPN18 and STARD13 protein expression levels in GIST clinical samples by western blotting analysis. (A) Western blotting images for PTPN18 and STARD13 (160 and 125 kDa) in clinical samples. Bar graphs showing the protein levels after normalisation to GAPDH. (B) Correlation of protein expression levels and miR-125a-5p expression levels in imatinib-resistant and -sensitive GISTs. r=correlation coefficient. Statistical differences between the two groups were calculated using the paired Student’s t-test, and the expression correlation was evaluated by Pearson’s correlation analysis. P⩽0.05 were considered significant.
To further evaluate the direct functional role of PTPN18 on imatinib response, we determined the effect of PTPN18 suppression on imatinib response in GIST882 cells. As shown in Figure 5, suppression of PTPN18 in GIST882 cells resulted in a significantly higher cell viability upon imatinib treatment, as compared with the cells treated with control plasmid.
Effect of PTPN18 silencing on imatinib response in GIST882 cells. Proportion of surviving cells in cultures transfected with short hairpin RNAs against PTPN18 (shPTPN18) or shRNA control plasmid (shControl) after 72 h of treatment with different concentrations of imatinib, as evaluated by WST-1 assay. Statistical differences between the two groups were calculated using the paired Student’s t-test, and P⩽0.05 was considered significant.
miRNA expression levels are associated with metastasis, KIT mutational status and survival
We performed SAM analysis on the microarray data set to identify the most significant miRNAs that could distinguish metastatic tumours from non-metastatic tumours. We found 19 overexpressed and 25 underexpressed miRNAs with a FDR<15% in the metastatic GISTs compared with the non-metastatic GISTs (Supplementary Table S7). Notably, 8 of the 10 selected miRNAs (miR-150-3p, miR-301a-3p, miR-30c, miR-107, miR-134, miR-1207-5p, miR-638 and miR-1915) chosen for RT–qPCR validation from the SAM analysis of imatinib-resistance vs imatinib-sensitive GISTs were also present in the SAM list generated for metastatic vs non-metastatic GISTs (Supplementary Tables S4 and S7). In concordance with the microarray results, the expression levels of miR-150-3p and miR-301a-3p (P=0.009 and P=0.041, respectively; t-test) were significantly different between the metastatic and non-metastatic GISTs by RT–qPCR (Figure 6A). We also evaluated whether the miRNA expression levels correlated with metastatic potential. We found significant correlations between the metastatic and non-metastatic tumours with the expression level of miR-301a-3p, miR-150-3p, miR-1915 and miR-1207-5p (Table 2).
RT–qPCR analysis of miRNA expression levels in relation to metastasis, KIT mutational status and survival in GISTs. Box plots show relative expression levels of (A) miR-150-3p and miR-301a-3p in 20 cases of metastatic and 10 non-metastatic GISTs, and (B) miR-150-3p in 13 single and 11 double KIT-mutated tumours. Comparisons between groups were calculated using the unpaired Student’s t-test, and P⩽0.05 were considered significant. (C) Kaplan–Meier curves show significant association of miR-1915 expression with disease-free survival and overall survival in GIST patients. Differences in survival were calculated using the log-rank test. High and low expression levels refer to above and below the median level of all tumours.
To determine whether KIT double mutated tumours differ from KIT single mutated tumours based on their miRNA expression levels and to identify the most significant deregulated miRNAs between these two groups, we performed SAM analysis on the samples harbouring only KIT mutations (four double KIT-mutated vs nine single KIT-mutated tumours) in the microarray data set. We found 54 overexpressed and 24 underexpressed miRNAs with a FDR<15% in the KIT double mutated GISTs compared with the single mutated GISTs (Supplementary Table S8). Eight of the differentially expressed miRNAs were also evaluated by RT–qPCR. The analysis revealed that tumours harbouring double KIT mutations had significantly lower expression levels of miR-150-3p (P=0.025; t-test) compared with the tumours carrying a single KIT mutation (Figure 6B). Using the Fisher’s exact test, we found that the expression levels of miR-301a-3p (P=0.012) and miR-1915 (P=0.012) were significantly correlated with KIT mutation status (Table 2).
We also evaluated the association of miRNA expression (based on the RT–qPCR data) with disease-free and overall survival using the Kaplan–Meier survival plots and log-rank analyses. We found low expression of miR-1915 is correlated with shorter disease-free (P=0.003; log-rank test) and overall (P=0.001; log-rank test) survival of GIST patients (Figure 6C).
Discussion
In this study, we evaluated the clinical impact of miRNA deregulation in imatinib-treated GISTs. We found specific miRNAs associated with imatinib resistance, metastatic disease, KIT mutational status and survival in GIST patients.
miR-125a-5p modulates imatinib response in GISTs
We identified a subset of miRNAs significantly associated with imatinib resistance in GIST. Among them, we validated miR-125a-5p in a larger cohort of samples using RT–qPCR and functionally characterised its role in GIST cell lines. We show that miR-125a-5p expression modulates imatinib responsiveness in the single KIT-mutated GIST882 cells but not in the double KIT-mutated GIST48 cell line. Given that complete genetic changes are not available for both cell lines, it is possible that other genetic factor(s) (besides KIT mutation) may contribute to the different imatinib response observed in the two cell lines. Furthermore, miR-125a-5p expression was not different between the sensitive and resistant tumours of the two patients with multiple tumours harbouring secondary KIT mutation. Together, our data suggest the role of miRNA(s) as an alternative resistance mechanism to second site KIT mutations in GIST.
Aberrant expression of miR-125a-5p has been reported in several cancer types (Guo et al, 2009; Wang et al, 2009). Low miR-125a-5p expression is associated with more aggressive disease and poorer survival in gastric cancer (Nishida et al, 2011). Decreased expression of miR125a-5p is also associated with docetaxel resistance in human breast cancer cell lines (Kastl et al, 2012). Our findings support the notion that miR-125a-5p may have tumour-promoting or -suppressing functions depending on the cellular contexts. Several studies have reported diverse functions of miR-125 family members in different cell types as reviewed by Sun et al (2013).
In addition to miR-125a-5p, other miR-125 family members were shown to induce drug resistance (Zhou et al, 2010) and regulate cell death pathways, such as NF-κB (Kim et al, 2012) and p53 (Leotta et al, 2014). Functionally, miR-125a-5p has a role in cell proliferation, migration, invasion and apoptosis in different cancer cell types (Wang et al, 2009; Nishida et al, 2011). Using a computational approach, we found that the predicted targets of miR-125a-5p are functionally associated with several biological processes, such as anti-apoptosis, cell cycle, cell differentiation, signal transduction and protein phosphorylation, which have been implicated in drug resistance of different cancer types. In line with the functional annotation of the miR-125a-5p predicted targets, deregulation of genes involved in cell cycle control, apoptosis and muscle differentiation has been associated with imatinib response in GIST patients (Agaram et al, 2007; Romeo et al, 2009).
Among the predicted targets, we chose PTPN18 and STARD13 as candidate targets of miR-125a-5p. We show that both candidates are regulated by miR-125a-5p in GIST cells, and we further validated the association of PTPN18 expression with miR-125a-5p expression and imatinib response in GIST clinical samples and cell lines. PTPN18 is an interesting candidate for several reasons. First, it is involved in the negative feedback mechanisms controlling the Bcr-Abl fusion tyrosine kinase-signalling network by inhibiting the phosphorylation of Src family kinase (Rubbi et al, 2011). Second, PTPN18 can induce actin cytoskeleton reorganisation (Shiota et al, 2003), suggesting a plausible explanation of the morphological changes observed in the imatinib-resistant GIST cells (Mahadevan et al, 2007). Third, altered phosphorylation of tyrosine kinases is an alternative mechanism of imatinib resistance in GIST (Takahashi et al, 2013), suggesting the possible involvements of protein phosphatases in imatinib resistance of GIST. Here, we propose the role of PTPN18 in imatinib response through miR-125a-5p regulation, and this regulation may involve in imatinib resistance as an alternative mechanism to KIT second site mutations in GIST.
Metastasis-associated miRNAs in GIST
We show associations of two miRNAs (miR-301a-3p and miR-150-3p) with metastasis. In concordance with our results in GIST, miR-301a-3p overexpression was found in metastatic breast cancer (Shi et al, 2011). Functionally, miR-301a-3p has been demonstrated to promote cell growth, migration and invasion in breast cancer cells (Shi et al, 2011) and hepatocellular carcinoma cells (Zhou et al, 2012). These findings support the role of miR-301a-3p as an oncogene in progression and development of metastasis in various cancer types. On the other hand, very little is known about miR-150-3p. Interestingly, its upregulation was shown in colorectal cancers with liver metastasis (Lin et al, 2011).
GISTs with a single or double KIT mutations
Despite numerous studies revealed that double mutations confer acquired imatinib resistance in GISTs, we (in this study) and others (Antonescu et al, 2005; Wardelmann et al, 2006) found a proportion (40–53%) of imatinib-resistant GISTs harbouring a single KIT mutation, suggesting the involvement of alternative resistance mechanisms. Furthermore, we showed that several miRNAs were significantly deregulated between the GISTs with a single mutation and double mutations. Those differentially expressed miRNAs may be involved in partly distinct pathways in GISTs with single and double KIT mutations. Although Allander et al (2001) reported homogeneous gene expression profiles in GISTs with different KIT mutations, distinct clusters were also observed among the GISTs with KIT mutations (Allander et al, 2001; Yamaguchi et al, 2008), suggesting subgroups of GIST patients harbouring KIT mutations. Because of the lack of information on secondary mutations in those studies, we could not determine whether the different clusters observed in the gene expression profiles reflect the single and double mutations in KIT. In line with the KIT mutations in GIST, acute myeloid leukemias with single or double CEBPA mutations have distinct gene expression profiles and different clinical outcomes (Wouters et al, 2009). Further investigations are warranted to determine the biological and clinical differences between GISTs with single and double KIT mutations.
Prognostic miRNAs in GIST patients
We identified miR-1915 as a candidate prognostic marker for survival in GIST under imatinib therapy. Interestingly, miR-1915 expression is also significantly correlated with metastasis and double KIT mutation (Table 2). Recently, Niinuma et al (2012) reported that miR-196a overexpression was significantly associated with high-risk grade, metastasis and poor survival in GIST, while Choi et al (2010) revealed a different subset of miRNAs associated with tumour risk. The discrepancy in the findings is likely due to sample selection. There is no information given whether the studied cohorts in both studies were treated with imatinib on a neoadjuvant basis.
In summary, we report the identification of miRNAs associated with imatinib resistance, metastasis, KIT mutational status and survival in GIST. Our findings suggest that differential expression of these miRNAs may contribute to the prognosis of GIST patients treated with imatinib preoperatively. Importantly, we provide functional evidence that miR-125a-5p can modulate imatinib response in single KIT-mutated GIST cells through PTPN18 regulation, suggesting the prognostic impact of miR-125a-5p and PTPN18 for GIST patients under neoadjuvant imatinib therapy.
Accession codes
References
Agaram NP, Besmer P, Wong GC, Guo T, Socci ND, Maki RG, DeSantis D, Brennan MF, Singer S, DeMatteo RP, Antonescu CR (2007) Pathologic and molecular heterogeneity in imatinib-stable or imatinib-responsive gastrointestinal stromal tumors. Clin Cancer Res 13: 170–181.
Ahlen J, Westerdahl J, Zedenius J, Branstrom R, Larsson C, Nilsson I-L (2012) Side-effects from imatinib treatment of advanced GIST–associated with a better outcome. J Cancer Ther Res 1: 1–9.
Allander SV, Nupponen NN, Ringner M, Hostetter G, Maher GW, Goldberger N, Chen Y, Carpten J, Elkahloun AG, Meltzer PS (2001) Gastrointestinal stromal tumors with KIT mutations exhibit a remarkably homogeneous gene expression profile. Cancer Res 61: 8624–8628.
Antonescu CR, Besmer P, Guo T, Arkun K, Hom G, Koryotowski B, Leversha MA, Jeffrey PD, Desantis D, Singer S, Brennan MF, Maki RG, DeMatteo RP (2005) Acquired resistance to imatinib in gastrointestinal stromal tumor occurs through secondary gene mutation. Clin Cancer Res 11: 4182–4190.
Bartel DP (2004) MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116: 281–297.
Bauer S, Yu LK, Demetri GD, Fletcher JA (2006) Heat shock protein 90 inhibition in imatinib-resistant gastrointestinal stromal tumor. Cancer Res 66: 9153–9161.
Berglund E, Berglund D, Akcakaya P, Ghaderi M, Dare E, Berggren PO, Kohler M, Aspinwall CA, Lui WO, Zedenius J, Larsson C, Branstrom R (2013) Evidence for Ca(2+)-regulated ATP release in gastrointestinal stromal tumors. Exp Cell Res 319: 1229–1238.
Caramuta S, Egyhazi S, Rodolfo M, Witten D, Hansson J, Larsson C, Lui WO (2010) MicroRNA expression profiles associated with mutational status and survival in malignant melanoma. J Invest Dermatol 130: 2062–2070.
Choi HJ, Lee H, Kim H, Kwon JE, Kang HJ, You KT, Rhee H, Noh SH, Paik YK, Hyung WJ (2010) MicroRNA expression profile of gastrointestinal stromal tumors is distinguished by 14q loss and anatomic site. Int J Cancer 126: 1640–1650.
Corless CL, Fletcher JA, Heinrich MC (2004) Biology of gastrointestinal stromal tumors. J Clin Oncol 22: 3813–3825.
Gensler M, Buschbeck M, Ullrich A (2004) Negative regulation of HER2 signaling by the PEST-type protein-tyrosine phosphatase BDP1. J Biol Chem 279: 12110–12116.
Gramza AW, Corless CL, Heinrich MC (2009) Resistance to tyrosine kinase inhibitors in gastrointestinal stromal tumors. Clin Cancer Res 15: 7510–7518.
Guo S, Bai H, Megyola CM, Halene S, Krause DS, Scadden DT, Lu J (2012) Complex oncogene dependence in microRNA-125a-induced myeloproliferative neoplasms. Proc Natl Acad Sci USA 109: 16636–16641.
Guo X, Wu Y, Hartley RS (2009) MicroRNA-125a represses cell growth by targeting HuR in breast cancer. RNA Biol 6: 575–583.
Haller F, von Heydebreck A, Zhang JD, Gunawan B, Langer C, Ramadori G, Wiemann S, Sahin O (2010) Localization- and mutation-dependent microRNA (miRNA) expression signatures in gastrointestinal stromal tumours (GISTs), with a cluster of co-expressed miRNAs located at 14q32.31. J Pathol 220: 71–86.
Kastl L, Brown I, Schofield AC (2012) miRNA-34a is associated with docetaxel resistance in human breast cancer cells. Breast Cancer Res Treat 131: 445–454.
Kim SW, Ramasamy K, Bouamar H, Lin AP, Jiang D, Aguiar RC (2012) MicroRNAs miR-125a and miR-125b constitutively activate the NF-kappaB pathway by targeting the tumor necrosis factor alpha-induced protein 3 (TNFAIP3, A20). Proc Natl Acad Sci USA 109: 7865–7870.
Leotta M, Biamonte L, Raimondi L, Ronchetti D, Martino MT, Botta C, Leone E, Pitari MR, Neri A, Giordano A, Tagliaferri P, Tassone P, Amodio N (2014) A p53-dependent tumor suppressor network is induced by selective miR-125a-5p inhibition in multiple myeloma cells in vitro. J Cell Physiol 229: 2106–2116.
Lin M, Chen W, Huang J, Gao H, Ye Y, Song Z, Shen X (2011) MicroRNA expression profiles in human colorectal cancers with liver metastases. Oncol Rep 25: 739–747.
Mahadevan D, Cooke L, Riley C, Swart R, Simons B, Della Croce K, Wisner L, Iorio M, Shakalya K, Garewal H, Nagle R, Bearss D (2007) A novel tyrosine kinase switch is a mechanism of imatinib resistance in gastrointestinal stromal tumors. Oncogene 26: 3909–3919.
Niinuma T, Suzuki H, Nojima M, Nosho K, Yamamoto H, Takamaru H, Yamamoto E, Maruyama R, Nobuoka T, Miyazaki Y, Nishida T, Bamba T, Kanda T, Ajioka Y, Taguchi T, Okahara S, Takahashi H, Nishida Y, Hosokawa M, Hasegawa T, Tokino T, Hirata K, Imai K, Toyota M, Shinomura Y (2012) Upregulation of miR-196a and HOTAIR drive malignant character in gastrointestinal stromal tumors. Cancer Res 72: 1126–1136.
Nilsson B, Bumming P, Meis-Kindblom JM, Oden A, Dortok A, Gustavsson B, Sablinska K, Kindblom LG (2005) Gastrointestinal stromal tumors: the incidence, prevalence, clinical course, and prognostication in the preimatinib mesylate era—a population-based study in western Sweden. Cancer 103: 821–829.
Nishida N, Mimori K, Fabbri M, Yokobori T, Sudo T, Tanaka F, Shibata K, Ishii H, Doki Y, Mori M (2011) MicroRNA-125a-5p is an independent prognostic factor in gastric cancer and inhibits the proliferation of human gastric cancer cells in combination with trastuzumab. Clin Cancer Res 17: 2725–2733.
Romeo S, Debiec-Rychter M, Van Glabbeke M, Van Paassen H, Comite P, Van Eijk R, Oosting J, Verweij J, Terrier P, Schneider U, Sciot R, Blay JY, Hogendoorn PC (2009) Cell cycle/apoptosis molecule expression correlates with imatinib response in patients with advanced gastrointestinal stromal tumors. Clin Cancer Res 15: 4191–4198.
Rubbi L, Titz B, Brown L, Galvan E, Komisopoulou E, Chen SS, Low T, Tahmasian M, Skaggs B, Muschen M, Pellegrini M, Graeber TG (2011) Global phosphoproteomics reveals crosstalk between Bcr-Abl and negative feedback mechanisms controlling Src signaling. Sci Signal 4: ra18.
Shi W, Gerster K, Alajez NM, Tsang J, Waldron L, Pintilie M, Hui AB, Sykes J, P'ng C, Miller N, McCready D, Fyles A, Liu FF (2011) MicroRNA-301 mediates proliferation and invasion in human breast cancer. Cancer Res 71: 2926–2937.
Shiota M, Tanihiro T, Nakagawa Y, Aoki N, Ishida N, Miyazaki K, Ullrich A, Miyazaki H (2003) Protein tyrosine phosphatase PTP20 induces actin cytoskeleton reorganization by dephosphorylating p190 RhoGAP in rat ovarian granulosa cells stimulated with follicle-stimulating hormone. Mol Endocrinol 17: 534–549.
Sihto H, Sarlomo-Rikala M, Tynninen O, Tanner M, Andersson LC, Franssila K, Nupponen NN, Joensuu H (2005) KIT and platelet-derived growth factor receptor alpha tyrosine kinase gene mutations and KIT amplifications in human solid tumors. J Clin Oncol 23: 49–57.
Sun YM, Lin KY, Chen YQ (2013) Diverse functions of miR-125 family in different cell contexts. J Hematol Oncol 6: 6.
Taft RJ, Hawkins PG, Mattick JS, Morris KV (2011) The relationship between transcription initiation RNAs and CCCTC-binding factor (CTCF) localization. Epigenetics Chromatin 4: 13.
Takahashi T, Serada S, Ako M, Fujimoto M, Miyazaki Y, Nakatsuka R, Ikezoe T, Yokoyama A, Taguchi T, Shimada K, Kurokawa Y, Yamasaki M, Miyata H, Nakajima K, Takiguchi S, Mori M, Doki Y, Naka T, Nishida T (2013) New findings of kinase switching in gastrointestinal stromal tumor under imatinib using phosphoproteomic analysis. Int J Cancer 133: 2737–2743.
Tang F, Zhang R, He Y, Zou M, Guo L, Xi T (2012) MicroRNA-125b induces metastasis by targeting STARD13 in MCF-7 and MDA-MB-231 breast cancer cells. PLoS One 7: e35435.
Tuveson DA, Willis NA, Jacks T, Griffin JD, Singer S, Fletcher CD, Fletcher JA, Demetri GD (2001) STI571 inactivation of the gastrointestinal stromal tumor c-KIT oncoprotein: biological and clinical implications. Oncogene 20: 5054–5058.
Wang G, Mao W, Zheng S, Ye J (2009) Epidermal growth factor receptor-regulated miR-125a-5p—a metastatic inhibitor of lung cancer. FEBS J 276: 5571–5578.
Wardelmann E, Merkelbach-Bruse S, Pauls K, Thomas N, Schildhaus HU, Heinicke T, Speidel N, Pietsch T, Buettner R, Pink D, Reichardt P, Hohenberger P (2006) Polyclonal evolution of multiple secondary KIT mutations in gastrointestinal stromal tumors under treatment with imatinib mesylate. Clin Cancer Res 12: 1743–1749.
Wouters BJ, Lowenberg B, Erpelinck-Verschueren CA, van Putten WL, Valk PJ, Delwel R (2009) Double CEBPA mutations, but not single CEBPA mutations, define a subgroup of acute myeloid leukemia with a distinctive gene expression profile that is uniquely associated with a favorable outcome. Blood 113: 3088–3091.
Yamaguchi U, Nakayama R, Honda K, Ichikawa H, Hasegawa T, Shitashige M, Ono M, Shoji A, Sakuma T, Kuwabara H, Shimada Y, Sasako M, Shimoda T, Kawai A, Hirohashi S, Yamada T (2008) Distinct gene expression-defined classes of gastrointestinal stromal tumor. J Clin Oncol 26: 4100–4108.
Zheng T, Wang J, Chen X, Liu L (2010) Role of microRNA in anticancer drug resistance. Int J Cancer 126: 2–10.
Zhou M, Liu Z, Zhao Y, Ding Y, Liu H, Xi Y, Xiong W, Li G, Lu J, Fodstad O, Riker AI, Tan M (2010) MicroRNA-125b confers the resistance of breast cancer cells to paclitaxel through suppression of pro-apoptotic Bcl-2 antagonist killer 1 (Bak1) expression. J Biol Chem 285: 21496–21507.
Zhou P, Jiang W, Wu L, Chang R, Wu K, Wang Z (2012) miR-301a is a candidate oncogene that targets the homeobox gene Gax in human hepatocellular carcinoma. Dig Dis Sci 57: 1171–1180.
Acknowledgements
We thank the members of the Medical Genetics group for their help and suggestions and Inga-Lena Nilsson for helping us with clinical information. This study was supported by the Swedish Research Council (523-2009-3517, 521-2010-3518), the Swedish Cancer Society, Åke Olsson’s Foundation for Haematological Research, Åke Wiberg’s Foundation, Axel and Signe Lagerman’s Donation Foundation, Cancer Research Funds of Radiumhemmet, Karolinska Institutet and Stockholm County Council.
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
Competing interests
The authors declare no conflict of interest.
Additional information
Supplementary Information accompanies this paper on British Journal of Cancer website
Supplementary information
Rights and permissions
This work is licensed under the Creative Commons Attribution-NonCommercial-Share Alike 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/
About this article
Cite this article
Akçakaya, P., Caramuta, S., Åhlen, J. et al. microRNA expression signatures of gastrointestinal stromal tumours: associations with imatinib resistance and patient outcome. Br J Cancer 111, 2091–2102 (2014). https://doi.org/10.1038/bjc.2014.548
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/bjc.2014.548
- Springer Nature Limited
Keywords
This article is cited by
-
Anesthetics may modulate cancer surgical outcome: a possible role of miRNAs regulation
BMC Anesthesiology (2021)
-
Molecular biomarkers for prognosis of gastrointestinal stromal tumor
Clinical and Translational Oncology (2019)
-
MicroRNA-125b reverses oxaliplatin resistance in hepatocellular carcinoma by negatively regulating EVA1A mediated autophagy
Cell Death & Disease (2018)
-
MicroRNAs as Mediators of Resistance Mechanisms to Small-Molecule Tyrosine Kinase Inhibitors in Solid Tumours
Targeted Oncology (2018)
-
Serum miR-518e-5p is a potential biomarker for secondary imatinib-resistant gastrointestinal stromal tumor
Journal of Biosciences (2018)