Journal of Cancer Research and Therapeutics

ORIGINAL ARTICLE
Year
: 2010  |  Volume : 6  |  Issue : 4  |  Page : 521--529

Distinctive microRNA signature of medulloblastomas associated with the WNT signaling pathway


Amit Gokhale1, Ratika Kunder1, Atul Goel2, Rajiv Sarin1, Aliasgar Moiyadi1, Asha Shenoy3, Chandrasekhar Mamidipally4, Santosh Noronha4, Sadhana Kannan1, Neelam Vishwanath Shirsat1,  
1 Advanced Centre for Treatment, Research and Education in Cancer, Tata Memorial Centre, Kharghar, Navi Mumbai 410210, Maharashtra, India
2 Department of Neurosurgery, Seth G. S. Medical College and K.E.M. Hospital, Parel, Mumbai 400012, Maharashtra, India
3 Department of Pathology, Seth G. S. Medical College and K.E.M. Hospital, Parel, Mumbai 400012, Maharashtra, India
4 Department of Chemical Engineering, School of Bioscience and Bioengineering, Systems and Control, Indian Institute of Technology, Mumbai 400076, Maharashtra, India

Correspondence Address:
Neelam Vishwanath Shirsat
Advanced Centre for Treatment, Research and Education in Cancer, Tata Memorial Centre, Kharghar, Navi Mumbai 410210, Maharashtra
India

Abstract

Aim: Medulloblastoma is a malignant brain tumor that occurs predominantly in children. Current risk stratification based on clinical parameters is inadequate for accurate prognostication. MicroRNA expression is known to be deregulated in various cancers and has been found to be useful in predicting tumor behavior. In order to get a better understanding of medulloblastoma biology, miRNA profiling of medulloblastomas was carried out in parallel with expression profiling of protein-coding genes. Materials and Methods: miRNA profiling of medulloblastomas was carried out using Taqman Low Density Array v 1.0 having 365 human microRNAs. In parallel, genome-wide expression profiling of protein-coding genes was carried out using Affymetrix gene 1.0 ST arrays. Results: Both the profiling studies identified four molecular subtypes of medulloblastomas. Expression levels of select protein-coding genes and miRNAs could classify an independent set of medulloblastomas. Twelve of 31 medulloblastomas were found to overexpress genes belonging to the canonical WNT signaling pathway and carry a mutation in CTNNB1 gene. A number of miRNAs like miR-193a, miR-224/miR-452 cluster, miR-182/miR-183/miR-96 cluster, and miR-148a having potential tumor/metastasis suppressive activity were found to be overexpressed in the WNT signaling associated medulloblastomas. Exogenous expression of miR-193a and miR-224, two miRNAs that have the highest WNT pathway specific upregulation, was found to inhibit proliferation, increase radiation sensitivity and reduce anchorage-independent growth of medulloblastoma cells. Conclusion: Expression level of tumor/metastasis suppressive miRNAs in the WNT signaling associated medulloblastomas is likely to determine their response to treatment, and thus, these miRNAs would be important biomarkers for risk stratification within the WNT signaling associated medulloblastomas.



How to cite this article:
Gokhale A, Kunder R, Goel A, Sarin R, Moiyadi A, Shenoy A, Mamidipally C, Noronha S, Kannan S, Shirsat NV. Distinctive microRNA signature of medulloblastomas associated with the WNT signaling pathway.J Can Res Ther 2010;6:521-529


How to cite this URL:
Gokhale A, Kunder R, Goel A, Sarin R, Moiyadi A, Shenoy A, Mamidipally C, Noronha S, Kannan S, Shirsat NV. Distinctive microRNA signature of medulloblastomas associated with the WNT signaling pathway. J Can Res Ther [serial online] 2010 [cited 2022 Jul 7 ];6:521-529
Available from: https://www.cancerjournal.net/text.asp?2010/6/4/521/77072


Full Text

 Introduction



Medulloblastoma is a highly malignant brain tumor that occurs predominantly in children. Medulloblastomas are located in the cerebellar region of the brain and have a tendency to spread through the cerebrospinal fluid. Therefore, standard post-operative treatment not only includes local radiotherapy but also craniospinal radiation and chemotherapy. [1],[2] One-third of the patients are incurable, while the long term survivors suffer from permanent neurological deficits resulting from the intensive therapies administered to the developing child brain. All medulloblastomas are classified pathologically as grade IV tumors. Molecular markers for risk stratification are required, so that standard risk patients can be spared from excessive treatment and survival of high risk patients can be improved. Understanding of the molecular mechanism/s underlying the pathogenesis of medulloblastomas is crucial for designing novel targeted therapies, which could be more effective and free of undesirable side effects.

Most microRNA expression analyses of human cancers have arrived at the common conclusions that miRNAs are deregulated in cancer, and miRNA expression profile represents tumor biology better than the expression profile of protein-coding genes. [3],[4] In order to get a better understanding of medulloblastoma biology, miRNA profiling of medulloblastomas was carried out using Taqman Low Density array v 1.0 having 365 human microRNAs. In parallel, genome-wide expression profiling of protein-coding genes was carried out using Affymetrix gene 1.0 ST arrays. Both the profiling studies segregate medulloblastoma tumor tissues into almost identical molecular subtypes.

 Materials and Methods



Tumor tissue specimens of sporadic medulloblastomas and normal cerebellar tissues were procured with the approval of the Institutional Review Board after getting informed consent from the patients. Immediately following surgery, tumor tissues were snap-frozen in liquid nitrogen and stored at −70°C. Normal cerebellum tissues were obtained from the Brain Tissue Repository, National Institute of Mental Health and Neurosciences, Bangalore, India.

Total RNA was extracted from tumor tissues using mirVANA kit Ambion, Austin, TX, USA, as per the manufacturer's protocol after ensuring at least 80% tumor cell content. Normal cerebellar tissues labeled NC01 and NC02 were from 6-month and 2-month old infants, while NC03 and NC04 were from 4-year and 35-year old males, respectively. Total RNA (100 ng) was reverse transcribed using stem-loop RT primer pools from Applied Biosystems (Foster City, CA, USA). Polymerase chain reactions (PCRs) were carried out using the Taqman Low Density Arrays v 1.0 on ABI 7900HT Fast real time RT-PCR system. Relative quantities (RQ) of each miRNA in each of the tissue samples as compared to the endogenous control small RNA RNU48 were computed by comparative Ct method.

Total RNA extracted, as described before, was further purified using RNeasy columns (Qiagen,Valencia, CA,USA), as per the manufacturer's instructions. RNAs having more than 7.0 RIN value and no detectable genomic DNA contamination were used for the analysis (Bioanalyzer, Agilent Technologies, Palo Alto, CA, USA). RNA (100 ng) was reverse transcribed, amplified, and labeled with biotin using the whole transcript sense target labeling kit and hybridized to gene 1.0 ST arrays (Affymetrix, Santa Clara,CA, USA), as per the manufacturer's instructions.

Data normalization was done using GCRMA algorithm in the Bioconductor package of the R statistical environment ( http://www.bioconductor.org ). Protein-coding genes and miRNAs significantly differentially expressed in each cluster were identified by Significance Analysis of Microarrays (SAM) analysis and t-test, respectively ( http://www.TM4.org ). Hierarchical clustering and bootstrap analysis steps were implemented using MeV module of TM4 package ( http://www.TM4.org ) . miRNA target prediction common to at least four different target prediction softwares was done using miRecords database ( http://www.mirecords.umn.edu ). Pathway identification and functional classification of differentially expressed genes and that of predicted targets of miRNAs was done using DAVID tool (http://david.abcc.ncifcrf.gov).

The differential expression of a select set of protein-coding genes and miRNAs was confirmed by real time RT-PCR analysis. Total RNA (500 ng) was treated with amplification grade RNase-free DNase (Invitrogen, Carlsbad, CA, USA) and reverse transcribed using MMLV-RT (MBI Fermentas, Burlington, Canada). Primers were designed such that they correspond to two adjacent exons, and wherever possible, were located at exon boundaries to avoid amplification of genomic DNA. Expression was analyzed by SYBR Green assay using GAPDH as a housekeeping gene control. Expression of each miRNA was analyzed using specific Taqman assay. Each assay was validated using RNA expressing specific miRNA as a positive control and RNA (no RT) as a negative control. RNU48 was used as an endogenous control RNA. Relative expression levels were quantified by comparative Ct method. Analysis was also done using RNU44 as an endogenous control for confirmation (data not shown). Supplementary-([SUPPORTING:1]) [Table 1] lists sequences of the primers used.{Table 1}

Human medulloblastoma cell line Daoy (ATCC, Manassas, VA, USA) was grown in Dulbecco's Modified Eagle Medium DMEM supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA, USA) in a humidified atmosphere of 5% CO 2 . Daoy cells were transfected with 100 nM of miR-193a mimic, miR-224 mimic, or miR-23b mimic using Dharmafect 2 reagent, as per the manufacturer's protocol (Dharmacon, Lafayette, CO, USA) for a period of 48 h. miRNA levels in transfected cells were estimated by real time RT-PCR analysis using RNU48 as an endogenous control RNA. As a negative control, Daoy cells were transfected with 100 nM of siGLO (a RISC-free control siRNA) or siRNA negative control (Dharmacon, Lafayette, CO, USA). microRNA mimic negative controls from Dharmacon were found to affect proliferation of Daoy cells. The transfected cells were allowed to recover for a period of 24 h before analyzing their growth characteristics. For MTT reduction assay, miRNA transfected Daoy cells were plated at a density of 500 cells/well of a 96-well microtiter plate. [5] Growth of these cells was followed over a period of 8 days with replenishment of the medium every 3rd day. 20 μl of MTT (5 mg/ml) was added to each well at the end of the incubation period and the cells were incubated further for a period of 4 h. 100 μl of 10% sodium dodecyl sulfate (SDS) in 0.01 N HCl was added per well to dissolve the dark blue formazan crystals. Optical density was read on an Enzyme-linked immunosorbent assay (ELISA) reader at a wavelength of 540 nm with a reference wavelength of 690 nm.

For thymidine incorporation assay, 2500 miRNA transfected cells were plated per well of a 96-well microtiter plate. The cells were incubated in the presence of 1 μCi of tritiated thymidine (specific activity 240 Gbq/mmole, Board of Radiation and Isotope Technology, Navi Mumbai, India) per well for a period of 20 h before harvesting by trypsinization. Tritiated thymidine incorporated was estimated by scintillation counting. Radiation sensitivity of the miRNA transfected cells was studied by clonogenic assay. 1000 miRNA transfected cells were plated per 55 mm plate and then irradiated at a dose of 6 Gy (Cobalt-60 gamma irradiator, developed by Bhabha Atomic Research Centre, India). The medium was changed 24 h later and the cells were allowed to grow for 6-8 days until microscopically visible colonies formed. The cells were fixed by incubation in chilled methanol/acetic acid and the colonies were visualized by staining with 0.5% crystal violet.

Anchorage-independent growth of the miRNA transfected cells was studied by their potential to form colonies in soft agar. 10,000 cells were seeded in DMEM/10% FBS medium containing 0.3% agarose. The cells were seeded onto a basal layer containing 1% agarose. The cells were incubated for about 3-4 weeks and the colonies formed were counted. All the experiments were performed in triplicates. Student's t-test was performed to evaluate statistical significance of the difference in the performance of miRNA-transfected cells as compared to siGLO or siRNA negative control transfected cells.

 Results



Gene expression profiling of 19 medulloblastoma tumor tissues was done using Affymetrix gene 1.0 ST array that contains probe sets for 28,869 genes. Unsupervised hierarchical clustering using 1000 most differentially expressed genes segregates tumor tissues into four clusters, viz. 'A', 'B', 'C', and 'D' with a bootstrap support of 100% for each cluster [[Figure 1]A]. Hierarchical clustering of medulloblastomas using the most significantly differentially expressed genes (FDR < 1%) is shown in Supplementary-[[SUPPORTING:5]] [Figure 1].{Figure 1}

Of the 365 miRNAs studied, 216 were found to be expressed in medulloblastomas. Hierarchical unsupervised clustering using these 216 miRNAs segregates tumors into clusters/subtypes similar to those identified by expression profiling of protein-coding genes [[Figure 1]B]. [Figure 2] is a heat map depicting the expression of selected miRNAs that are significantly differentially expressed in the four medulloblastoma subtypes and normal cerebellar tissues. The protein-coding genes and the miRNAs significantly differentially expressed in each subtype are listed in Supplementary [Table 2]-([SUPPORTING:2]) and [Table 3]-([SUPPORTING:3]), respectively. Real-time PCR analysis confirmed the expression of selected set of protein-coding genes and miRNAs in the four molecular subtypes. Molecular subtyping of an independent set of 12 medulloblastomas was done by analyzing the expression of these selected marker genes and miRNAs [Figure 3].{Figure 2}{Figure 3}

Subtype A (six tumors) is characterized by the overexpression of a number of genes involved in the canonical WNT signaling pathway like WIF1, DKK1, DKK2, DKK4, AXIN2, LEF1, NKD1, and MYC. Based on the overexpression of WIF1 and GABRE, six out of 12 tumors from an independent set of medulloblastomas were found to belong to subtype A [Figure 3]. WNT pathway activation in these tumors was confirmed by sequencing exon 3 of CTNNB1 gene that codes for the N-terminal region of β-catenin protein. A point mutation was found in all 12 subtype A tumors that modified either the serine residues S33 or S37 which get phosphorylated or the neighboring D32 or I35 residue in the N-terminal region of the β-catenin protein [Supplementary-[[SUPPORTING:6]] [Figure 2]]. [6]

Subtype A has the most robust miRNA signature with 16 miRNAs differentially expressed as compared to the normal cerebellar tissues as well as all other subtypes [Table 1]. A number of miRNAs like miR-193a, miR-224/miR-452 cluster, miR-182/miR-183/miR-96 cluster, miR-365, miR-135a, miR-148a, miR-23b/miR-24/miR-27b cluster, miR-204, miR-146b, miR-449/miR-449b cluster, miR-335, and miR-328 are overexpressed by 3-100 fold almost exclusively in tumors associated with the WNT signaling pathway [Figure 2] and Supplementary [Table 3]-[[SUPPORTING:3]]]. Real time RT-PCR analysis confirmed significant overexpression of miR-224, miR-193a, miR-365, miR-148a, miR-182, and miR-23b in the WNT signaling associated medulloblastomas [Figure 3]. miR-224 and miR-452 belong to a single cluster located in the intron of GABRE gene coding for GABA receptor. The gene GABRE and miR-224/miR-452 are specifically expressed in subtype A tumors. miR-224 cluster, therefore, appears to be co-expressed with GABRE gene. GABRE is overexpressed exclusively in all nine tumors having WNT pathway activation of Kool et al., data set as well. [7]

Three tumors show overexpression of SHH signaling components that include HHIP, ATOH1, MYCN, PTCH1, and GLI2. One tumor from the additional set belongs to subtype B as it shows overexpression of MYCN and ATOH1. All four subtype B tumors underexpress TRPM3 gene and miR-204 as well as miR-135b [Figure 3],[Table 1]. Two of the B subtype tumors, HMED13 and HMED22, show overexpression of MYCNOS. MYCNOS is a MYCN related gene that is expressed from the DNA strand complementary to the MYCN coding strand. MYCNOS and MYCN are known to be co-expressed in MYCN amplified tissues. A 30-60 fold increase in copy number of MYCN gene was confirmed by real-time PCR analysis of genomic DNA from these two tumors (data not shown). miR-23b, miR-27b and miR-24 belong to a single miRNA cluster that is located in an intron of C9orf3 gene. Two subtype B tumors lacking MYCN amplification overexpress C9orf3 and miR-23b, while the two having MYCN amplification underexpress miR-153. C9orf3 and miR-23b cluster miRNAs are overexpressed in all subtype A tumors of our data set as well as Kool et al., data set. C9orf3 is overexpressed in 8 out of 15 subtype B tumors of Kool et al., data set. [7]

Remaining 10 medulloblastomas segregate into cluster C (2 tumors) and cluster D (8 tumors). Non-WNT, non-SHH subtype medulloblastomas in Kool et al., study segregate into three subtypes, viz. C, D and E, with overlapping gene signature. [7] The genes specifically expressed in tumors belonging to C, D and E subtypes of Kool et al., data set as well as those belonging to C and D subtypes of our data set include transcription factors involved in brain development, viz. EOMES and FOXG1B, a testes specific gene LEMD1, and genes involved in neuronal migration like UNC5D and EPHA8. Expression of neuronal differentiation related genes in subtypes C and D and the expression of retinal differentiation related genes in subtypes D and E distinguish the three subtypes C, D and E of Kool et al,. data set. Subtype C tumors from our data set overexpress retina-specific genes like CRX, NRL, TULP1, PDE6H and underexpress most neuronal differentiation genes like GRM1, GRM8, MYRIP, thus resembling subtype E of Kool et al., data set. Equivalence of our subtype C to subtype E of Kool et al., data set is supported further by the expression of subtype E specific genes like GABRA5, SMARCD3 as well as the overexpression of cell cycle genes and a number of ribosomal protein coding genes in subtype C tumors of our data set.

Subtype D tumors of our data set express a number of neuronal differentiation genes including those encoding synaptic proteins like MYRIP, SYN2, SYT6, SYT13, those involved in transmission of nerve impulse like GRM1, GRM8, GABARAPL1, GABBR2, GABRG2, as well as genes involved in axon guidance like EPHA6, EPHB1, EFNB1, RND1, RND2, and SEMA3A. None of the subtype D medulloblastomas from our data set overexpress retinal differentiation genes like CRX, NRL, TULP1, and PDE6H. Therefore, subtype D of our data set is equivalent to subtype C of Kool et al., data set.

miR-135b is upregulated in tumors belonging to subtypes C and D. miR-135b is located in an intron of LEMD1 gene that is overexpressed in the C and D subtype medulloblastomas. miR-204 and miR-153 are underexpressed in subtype B and C medulloblastomas. miR-204 is located in an intron of TRPM3 gene that is downregulated in B and C subtype tumors. TRPM3 gene is underexpressed in all subtype B and 9 out of 11 subtype E medulloblastomas of Kool et al.,data set. [7]

HMED16 and HMED18 form a sub-cluster of cluster D on hierarchical cluster analysis of microarray profiling data of the protein-coding genes. These tumors segregate into a distinct cluster (D2) on hierarchical cluster analysis of miRNA data. Sub-cluster D2 (two tumors) differs from sub-cluster D1 (six tumors) based only on the overexpression of genes encoding various extracellular matrix proteins and TGF-β signaling components. Gene expression profile of cluster D2 is a characteristic wound healing signature. [8] This wound healing signature is also evident in all cluster B tumors and some cluster A tumors of our data set as well as those of Kool et al., data set. miR-214 and miR-199a are overexpressed in all these tumors having overexpression of wound healing pathway genes.

Five tumors from the additional set belong to subtype C or D as they overexpress LEMD1, KHDRBS2 and miR-135b. Three out of the five tumors most likely belong to subtype C as they underexpress TRPM3, miR-204 and GRM8 (a neuronal differentiation marker). miR-153 is underexpressed in two out of these three tumors. Thus, overexpression of LEMD1, KHDRBS2 and miR-135b is specific for subtypes C and D, while underexpression of TRPM3, GRM8 and miR-204 further distinguishes subtype C from subtype D [Figure 3] and [Table 1].

miR-17-92, a polycistronic miRNA cluster, has been reported to be overexpressed in a wide variety of human cancers. Overexpression of miR-17-92 cluster miRNAs in A, B and C subtype medulloblastomas is consistent with their reported upregulation by MYC, MYCN and E2F transcription factors. [9],[10] Normal adult cerebellums have the least expression of miR-17-92 cluster miRNAs. miR-106b and miR-25, which belong to miR-17-92 paralog cluster, are also overexpressed in all medulloblastomas as compared to the normal cerebellar tissues.

miR-379/miR-656 cluster miRNAs located within an imprinted region on chromosome 14 are underexpressed in subtype A, B and C tumors as compared to normal cerebellar tissues and subtype D tumors. [11] miR-379/miR-656 cluster miRNAs may play a role in neural differentiation as suggested by their predominant expression in the brain. miR-127/miR-432/miR-433 miRNA cluster on chromosome 14 is also similarly underexpressed in subtype A, B and C tumors. miR-127 has been reported to be underexpressed in various other cancers as a result of promoter hypermethylation. [12] miR-124a is also considerably downregulated in A, B and C subtype medulloblastomas and has been shown to promote neural differentiation by triggering brain-specific pre-mRNA alternate splicing. [13]

To understand the functional significance of miRNAs overexpressed in medulloblastomas associated with WNT signaling activation, three miRNAs, viz. miR-193a, miR-224 and miR-23b, were exogenously expressed in Daoy cell line established from human sporadic medulloblastoma. miR-193a and miR-224 are the most highly and specifically upregulated miRNAs in cluster A tumors, while miR-23b is overexpressed in both A and B subtype tumors. miR-193a and miR-224 expression in Daoy cells is comparable to normal developing cerebellar tissues. miR-23b expression in Daoy cells is higher than that of miR-193a or miR-224, while it is still about four-fold lower than that in normal developing cerebellar tissues. Transfection of 100 nM of miRNA mimics in Daoy cells resulted in 10-100 fold increase in miRNA expression. A 50-100 fold overexpression of miR-193a in Daoy cells resulted in 50-60% growth inhibition, while 10-15 fold overexpression of miR-23b resulted in 1.6-1.8 fold increased proliferation of Daoy cells as judged by thymidine incorporation assay [[Figure 4]A]. miR-193a induced growth inhibition and miR-23b mediated proliferation stimulation of Daoy cells was also evident on analysis by MTT assay [[Figure 4]B]. A 10-15 fold miR-224 overexpression, on the other hand, showed a marginal difference on proliferation of Daoy cells by the thymidine incorporation assay, while the MTT assay demonstrated growth inhibitory effect. Five hundred cells were plated per well for the MTT assay, while 2500 cells/well were plated for the thymidine incorporation assay. It, therefore, appears that the difference in the behavior of miR-224 transfected Daoy cells is likely to be due to the difference in plating density, indicating increased growth factor requirement of miR-224 transfected Daoy cells. This observation is further supported by the fact that plating efficiency of miR-224 transfected Daoy cells was found to be reduced by 50% in clonogenic assay [[Figure 5]A]. Thus, miR-224 appears to reduce proliferation of Daoy cells in a density-dependent manner. The plating efficiency of miR-193a transfected Daoy cells was found to be reduced by almost 80%, while the plating efficiency of miR-23b transfected Daoy cells did not change significantly from control cells. Irradiation at a dose of 6 Gy resulted in about 70% reduction in the number of colonies formed by control siGLO or siRNA transfected Daoy cells in clonogenic assay. miR-193a overexpressing Daoy cells on irradiation at a dose of 6 Gy failed to form any colonies, while irradiation of miR-224 overexpressing Daoy cells resulted in more than 90% reduction in colony formation. No significant change was observed in radiation sensitivity of miR-23b overexpressing Daoy cells [[Figure 5]A]. miR-224 and miR-193a overexpression in Daoy cells was found to bring about 60 to 90% reduction in soft agar colony formation [[Figure 5]B]. There was no significant difference in the number of soft agar colonies formed by miR-23b overexpressing cells as compared to siGLO or control siRNA transfected cells.{Figure 4}{Figure 5}

 Discussion



Genome-wide expression profiling of protein-coding and miRNA coding genes identified almost identical four molecular subtypes of medulloblastomas. 38% (12 out of 31) medulloblastomas in our study were found to carry a mutation in CTNNB1 gene and thereby WNT pathway activation. Median age at diagnosis for WNT signaling associated medulloblastomas is reported to be higher (10.4 years) than that reported for medulloblastomas (6 years). [7] Median age at diagnosis of subtype A tumors in our data set is also high (12 years). Four out of 12 medulloblastoma patients belonging to the WNT subtype in our study are adults. Prevalence of SHH signaling associated medulloblastomas is reported in children less than 3 years of age. [7] Lack of medulloblastomas from children less than 3 years of age can explain lower number of medulloblastomas with SHH signaling activation in our data set. Even if SHH signaling associated tumors are not taken into account, medulloblastomas associated with WNT signaling appear to be more common in the Indian subcontinent (38% incidence vs. reported incidence of 10-15%), which needs to be confirmed on a larger data set. [7],[14] Six out of 12 (50%) subtype A tumors belong to female patients, while 5 out of the rest 19 (20%) patients in our data set are females. In Kool et al., data set as well, 44% (4/9) of subtype A tumors as compared to 32% (12/37) of the rest of the medulloblastomas (excluding subtype B tumors) belong to female patients. [7] Prevalence of medulloblastomas resulting from deregulated WNT pathway activation in females over 3 years of age probably explains better survival of female patients in this age group as reported from the retrospective analysis of 1226 medulloblastoma cases. [15]

Medulloblastomas having WNT pathway activation have been reported to have lower metastatic potential and better survival rates. [7],[14] One out of 10 informative subtype A patients of our data set had metastasis at the time of diagnosis as compared to 4 out of 13 subtype C and D patients [Supplementary [Table 4]-[[SUPPORTING:4]]]. Higher incidence of metastasis at diagnosis in subtype C and D tumors has also been reported by Kool et al. [7] Expression profile of subtype A tumors seems paradoxical to this observation. Subtype A tumors have higher expression of genes encoding ribosomal proteins, cell cycle regulators and genes encoding components of RAS-MAPK, TGF-β and NOTCH signaling pathways as compared to the subtype D tumors. Robust overexpression of a number of miRNAs in subtype A tumors is similar to the robust expression of negative regulators of WNT signaling like WIF1, DKK family genes, AXIN2, NKD1, NKD2. Many of these miRNAs are likely to be direct/indirect transcriptional targets of mutant β-catenin protein and may target components of WNT signaling machinery. Predicted targets of subtype A specific miRNAs include the WNT signaling components. miR-135a has been shown to target APC and its levels correlate with APC levels in colorectal cancer. [16] APC gene is downregulated in subtype A medulloblastomas. WNT1-inducible signaling pathway protein 2 (WISP2) has been shown to be a direct target of miR-449, another miRNA overexpressed in subtype A tumors. [17]

Overexpression of miR-193a and miR-224, the two most upregulated miRNAs in subtype A tumors, was found to inhibit proliferation, increase radiation sensitivity and inhibit anchorage-independent growth of medulloblastoma cells. Overexpression of miR-224 has been shown to promote apoptosis of hepatocarcinoma cells and API5 has been shown to be a target of miR-224. [18] miR-193a expression has been found to be downregulated in oral squamous cell carcinoma cell lines as a result of tumor-specific hypermethylation of CpG islands and its ectopic expression has been found to be growth inhibitory to these cell lines. [19] miR-193a has been reported to be downregulated in various types of solid tumors in a study that was done on 2532 tumor tissues. [20] miR-23b cluster miRNAs have been shown to inhibit migration of hepatocellular carcinoma cells and inhibit TGF-β signaling by targeting SMAD proteins. [21],[22]

Among other miRNAs overexpressed in subtype A tumors, miR-148a has been shown to be downregulated as a result of promoter hypermethylation in cancer cell lines established from lymph node metastasis and further shown to inhibit motility, tumor growth and metastasis on overexpression. [23] Overexpression of miR-183 in lung cancer cells has recently been shown to inhibit migration and invasion of lung cancer cells and Ezrin has been identified as a bonafide target of miR-183.[24] Thus, while the expression of proliferation stimulating and apoptosis inhibitory genes like MYC, CCND1, BIRC5 drives tumorigenesis resulting from activated WNT signaling, expression of miRNAs like miR-193a, miR-224, miR-148a, miR-183 appears to contribute to lower metastatic potential and better response to radiation therapy and thereby better survival rate of these medulloblastomas.

Subtype C medulloblastomas have been reported to have the highest metastatic potential followed by subtype D tumors. [7] miR-135b overexpressed in subtype C and D medulloblastomas has been shown to be upregulated in relapsed prostate cancer patients, indicating the oncogenic nature of this miRNA. [25] miR-124a and miR-137 are known to induce differentiation of glioma stem cells. [26] Relatively higher expression of these miRNAs in subtype D medulloblastomas is consistent with the expression of various differentiation related genes in these tumors. Subtype B and C tumors have lower expression of miRNAs like miR-204 and miR-153 whose predicted targets include components of TGF-β signaling pathway. Ferreti et al., have reported underexpression of miR-153 in high risk medulloblastomas which are either metastatic or belong to children less than 3 years of age. [27] These high risk tumors are likely to belong to the subtypes B and C. Thus, potential oncogenic miRNA miR-135b is overexpressed while the potential tumor-suppressive miRNAs like miR-204 and miR-153 are underexpressed in C and D subtype tumors having higher metastatic potential.

In summary, genome-wide expression profiling of both protein-coding genes and miRNAs segregates medulloblastomas into four molecular subtypes. These four molecular subtypes closely match the four molecular variants reported in a recent study of genome wide expression profiling coupled with DNA copy number alterations in medulloblastomas. [28] Relative expression levels of a select set of protein-coding genes and miRNAs could successfully identify these molecular subtypes in an independent set of medulloblastomas and thus they can serve as markers for molecular subtyping. A number of miRNAs having potential tumor/metastasis suppressive role were found to be overexpressed in WNT signaling associated medulloblastomas. Exogenous expression of miR-193a and miR-224, two miRNAs that have the highest WNT pathway specific upregulation, was found to inhibit proliferation, increase radiation sensitivity and reduce anchorage-independent growth of medulloblastoma cells. Detailed functional studies on miRNAs differentially expressed in WNT signalling associated medulloblastomas and correlation of their expression with clinical outcome on a larger sample size would indicate if these miRNAs could serve as important biomarkers for risk stratification.

 Acknowledgements



We are grateful to Prof. David Bowtell for his valuable suggestions and help in microarray profiling. We thank Dr. S. K. Shankar for providing normal cerebellar tissues and Mr. Anant Sawant for technical assistance.

References

1Paulino AC. Current multimodality management of medulloblastoma. Curr Probl Cancer 2002;26:317-56.
2Packer RJ, Cogen P, Vezina G, Rorke LB. Medulloblastoma: Clinical and biologic aspects. Neuro Oncol 1999;1:232-50.
3Wu W, Sun M, Zou GM, Chen J. MicroRNA and cancer: Current status and prospective. Int J Cancer 2007;120:953-60.
4Zhang B, Pan X, Cobb GP, Anderson TA. microRNAs as oncogenes and tumor suppressors. Dev Biol 2007;302:1-12.
5Mosmann T. Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxic assays. J Immunol Methods 1983;65:55-63.
6Legoix P, Bluteau O, Bayer J, Perret C, Balabaud C, Belghiti J, et al. Beta-catenin mutations in hepatocellular carcinoma correlate with a low rate of loss of heterozygosity. Oncogene 1999;18:4044-6.
7Kool M, Koster J, Bunt J, Hasselt NE, Lakeman A, van Sluis P, et al. Integrated genomics identifies five medulloblastoma subtypes with distinct genetic profiles, pathway signatures and clinicopathological features. PLoS One 2008;3:e3088.
8Chang HY, Sneddon JB, Alizadeh AA, Sood R, West RB, Montgomery K, et al. Gene expression signature of fibroblast serum response predicts human cancer progression: Similarities between tumors and wounds. PLoS Biol 2004;2:E7.
9Woods K, Thomson JM, Hammond SM. Direct regulation of an oncogenic micro-RNA cluster by E2F transcription factors. J Biol Chem 2007;282:2130-4.
10Northcott PA, Fernandez LA, Hagan JP, Ellison DW, Grajkowska W, Gillespie Y, et al. The miR-17/92 polycistron is up-regulated in sonic hedgehog-driven medulloblastomas and induced by N-myc in sonic hedgehog-treated cerebellar neural precursors. Cancer Res 2009;69:3249-55.
11Glazov EA, McWilliam S, Barris WC, Dalrymple BP. Origin, evolution, and biological role of miRNA cluster in DLK-DIO3 genomic region in placental mammals. Mol Biol Evol 2008;25:939-48.
12Lujambio A, Esteller M. CpG island hypermethylation of tumor suppressor microRNAs in human cancer. Cell Cycle 2007;6:1455-9.
13Makeyev EV, Zhang J, Carrasco MA, Maniatis T. The MicroRNA miR-124 promotes neuronal differentiation by triggering brain-specific alternative pre-mRNA splicing. Mol Cell 2007;27:435-48.
14Ellison DW, Onilude OE, Lindsey JC, Lusher ME, Weston CL, Taylor RE, et al. beta-Catenin status predicts a favorable outcome in childhood medulloblastoma: The United Kingdom Children's Cancer Study Group Brain Tumour Committee. J Clin Oncol 2005;23:7951-7.
15Curran EK, Sainani KL, Le GM, Propp JM, Fisher PG. Gender affects survival for medulloblastoma only in older children and adults: A study from the Surveillance Epidemiology and End Results Registry. Pediatr Blood Cancer 2009;52:60-4.
16Nagel R, le Sage C, Diosdado B, van der Waal M, Oude Vrielink JA, Bolijn A, et al. Regulation of the adenomatous polyposis coli gene by the miR-135 family in colorectal cancer. Cancer Res 2008;68:5795-802.
17Iliopoulos D, Bimpaki EI, Nesterova M, Stratakis CA. MicroRNA signature of primary pigmented nodular adrenocortical disease: Clinical correlations and regulation of Wnt signaling. Cancer Res 2009;69:3278-82.
18Wang Y, Lee AT, Ma JZ, Wang J, Ren J, Yang Y, et al. Profiling microRNA expression in hepatocellular carcinoma reveals microRNA-224 up-regulation and apoptosis inhibitor-5 as a microRNA-224-specific target. J Biol Chem 2008;283:13205-15.
19Kozaki K, Imoto I, Mogi S, Omura K, Inazawa J. Exploration of tumor-suppressive microRNAs silenced by DNA hypermethylation in oral cancer. Cancer Res 2008;68:2094-105.
20Volinia S, Galasso M, Costinean S, Tagliavini L, Gamberoni G, Drusco A, et al. Reprogramming of miRNA networks in cancer and leukemia. Genome Res 2010;20:589-99.
21Salvi A, Sabelli C, Moncini S, Venturin M, Arici B, Riva P, et al. MicroRNA-23b mediates urokinase and c-met downmodulation and a decreased migration of human hepatocellular carcinoma cells. FEBS J 2009;276:2966-82.
22Rogler CE, Levoci L, Ader T, Massimi A, Tchaikovskaya T, Norel R, et al. MicroRNA-23b cluster microRNAs regulate transforming growth factor-beta/bone morphogenetic protein signaling and liver stem cell differentiation by targeting Smads. Hepatology 2009;50:575-84.
23Lujambio A, Calin GA, Villanueva A, Ropero S, Sanchez-Cespedes M, Blanco D, et al. A microRNA DNA methylation signature for human cancer metastasis. Proc Natl Acad Sci U S A 2008;105:13556-61.
24Wang G, Mao W, Zheng S. MicroRNA-183 regulates Ezrin expression in lung cancer cells. FEBS Lett 2008;582:3663-8.
25Tong AW, Fulgham P, Jay C, Chen P, Khalil I, Liu S, et al. MicroRNA profile analysis of human prostate cancers. Cancer Gene Ther 2009;16:206-16.
26Silber J, Lim DA, Petritsch C, Persson AI, Maunakea AK, Yu M, et al. miR-124 and miR-137 inhibit proliferation of glioblastoma multiforme cells and induce differentiation of brain tumor stem cells. BMC Med 2008;6:14.
27Ferretti E, De Smaele E, Po A, Di Marcotullio L, Tosi E, Espinola MS, et al. MicroRNA profiling in human medulloblastoma. Int J Cancer 2009;124:568-77.
28Northcott PA, Korshunov A, Witt H, Hielsher T, Eberhart CG, Mack S, et al. Medulloblastoma comprises four distinct molecular variants. J Clin Oncol 2010 [Epub ahead of print].