Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Integrated Analysis of Mismatch Repair System in Malignant Astrocytomas

  • Irene Rodríguez-Hernández,

    Affiliations Molecular Medicine Unit, Department of Medicine, University of Salamanca, Salamanca, Spain, IBMCC and IBSAL, (USAL/CSIC/University Hospital), Salamanca, Spain

  • Juan Luis Garcia,

    Affiliations IBMCC and IBSAL, (USAL/CSIC/University Hospital), Salamanca, Spain, Institute for Health Science Studies of Castilla y León, Salamanca, Spain

  • Angel Santos-Briz,

    Affiliations IBMCC and IBSAL, (USAL/CSIC/University Hospital), Salamanca, Spain, Department of Pathology, University Hospital of Salamanca, Salamanca, Spain

  • Aurelio Hernández-Laín,

    Affiliation Department of Pathology, University Hospital 12 de Octubre, Madrid, Spain

  • Jose María González-Valero,

    Affiliations IBMCC and IBSAL, (USAL/CSIC/University Hospital), Salamanca, Spain, Institute for Health Science Studies of Castilla y León, Salamanca, Spain

  • Juan Antonio Gómez-Moreta,

    Affiliation Department of Neurosurgery, University Hospital of Salamanca, Salamanca, Spain

  • Oscar Toldos-González,

    Affiliation Department of Pathology, University Hospital 12 de Octubre, Madrid, Spain

  • Juan Jesús Cruz,

    Affiliations IBMCC and IBSAL, (USAL/CSIC/University Hospital), Salamanca, Spain, Department of Oncology, University Hospital of Salamanca, Salamanca, Spain

  • Javier Martin-Vallejo,

    Affiliation Department of Statistics, University of Salamanca, Salamanca, Spain

  • Rogelio González-Sarmiento

    Affiliations Molecular Medicine Unit, Department of Medicine, University of Salamanca, Salamanca, Spain, IBMCC and IBSAL, (USAL/CSIC/University Hospital), Salamanca, Spain

Integrated Analysis of Mismatch Repair System in Malignant Astrocytomas

  • Irene Rodríguez-Hernández, 
  • Juan Luis Garcia, 
  • Angel Santos-Briz, 
  • Aurelio Hernández-Laín, 
  • Jose María González-Valero, 
  • Juan Antonio Gómez-Moreta, 
  • Oscar Toldos-González, 
  • Juan Jesús Cruz, 
  • Javier Martin-Vallejo, 
  • Rogelio González-Sarmiento


Malignant astrocytomas are the most aggressive primary brain tumors with a poor prognosis despite optimal treatment. Dysfunction of mismatch repair (MMR) system accelerates the accumulation of mutations throughout the genome causing uncontrolled cell growth. The aim of this study was to characterize the MMR system defects that could be involved in malignant astrocytoma pathogenesis. We analyzed protein expression and promoter methylation of MLH1, MSH2 and MSH6 as well as microsatellite instability (MSI) and MMR gene mutations in a set of 96 low- and high-grade astrocytomas. Forty-one astrocytomas failed to express at least one MMR protein. Loss of MSH2 expression was more frequent in low-grade astrocytomas. Loss of MLH1 expression was associated with MLH1 promoter hypermethylation and MLH1 -93G>A promoter polymorphism. However, MSI was not related with MMR protein expression and only 5% of tumors were MSI-High. Furthermore, the incidence of tumors carrying germline mutations in MMR genes was low and only one glioblastoma was associated with Lynch syndrome. Interestingly, survival analysis identified that tumors lacking MSH6 expression presented longer overall survival in high-grade astrocytoma patients treated only with radiotherapy while MSH6 expression did not modify the prognosis of those patients treated with both radiotherapy and chemotherapy. Our findings suggest that MMR system alterations are a frequent event in malignant astrocytomas and might help to define a subgroup of patients with different outcome.


Malignant gliomas account for 70% of all primary brain tumors with an incidence rate adjusted to the European Standard Population of 5.27 per 100 000 persons per year [1]. Unfortunately, the majority of these patients display progressive disease and subsequent death. The most common and devastating brain tumor in adults is glioblastoma (grade IV) with a median survival of approximately 12-14 months despite optimal treatment [2,3]. Patients with anaplastic astrocytoma (grade III) survive for nearly 1.5 years, and those with low-grade astrocytomas (grade II) can survive for as long as 5-10 years [4,5]. Initiation and progression of malignant astrocytomas are related to their genetic and chromosomal alterations. In this context, recent molecular and genetic studies have identified different markers that help to determine prognosis and likelihood of therapeutic response [6-10].

Mismatch repair (MMR) system maintains DNA stability by repairing DNA mismatches and insertion/deletion loops acquired during DNA replication. Therefore, MMR system maintains genomic integrity and provides tumor suppressor functions. Defective MMR function is found both in sporadic tumors and in cancers related to Lynch syndrome [11] that is characterized by a predisposition to early onset tumors in the proximal colon as well as extracolonic malignancies such as astrocytomas [12-14]. This syndrome is due to germline mutations in one of the MMR genes, mostly MSH2 or MLH1, and less frequently MSH6 or PMS2 [15,16]. Mutations in these genes result in microsatellite instability (MSI) and/or loss of expression of the associated protein. However, MMR deficiency in sporadic cancers is mostly due to loss of MLH1 expression as a result of somatic hypermethylation of its promoter [16]. MLH1 promoter hypermethylation has been associated in colorectal cancer (CRC) with the MLH1 -93G>A promoter polymorphism [17,18].

We have performed a molecular characterization of MMR system defects in malignant astrocytomas and we have evaluated the influence of these alterations in patient outcome. Specifically, we have investigated the expression profile and the promoter hypermethylation status of MLH1, MSH2 and MSH6 genes, as well as the MSI levels in pretreated low- and high-grade primary astrocytomas. We have also conducted a mutational analysis of MMR genes in tumors with MMR defective function.

Materials and Methods

Ethics Statement

The study was approved by the local Ethics Committees of the University Hospital of Salamanca (Salamanca, Spain) and University Hospital 12 de Octubre (Madrid, Spain), and written consent was obtained from the patients. The study was conducted according to the principles expressed in the Declaration of Helsinki.

Patients and samples

A total of 96 newly diagnosed patients with primary astrocytoma grades II to IV (study cohort) were recruited from June 2000 until March 2006 at the University Hospital of Salamanca (Spain). Patients were followed up from diagnosis to the present at the Neurosurgery and Oncology Departments. Tumors were classified as 20 low-grade astrocytomas (grade II), 19 anaplastic astrocytomas (grade III) and 57 glioblastomas (grade IV) according to the 2007 WHO classification [19]. The clinicopathological features of the patients are summarized in Table 1 and Table S1.

Patients, No. (%)LGA (n=20)AA (n=19)GBM (n=57)
Male11 (55)12 (63)36 (63)
Female9 (45)7 (37)21 (37)
Median age, years [quartiles]35 [30.3-46.0]57 [47.0-66.0]63 [54.5-69.0]
Tumor Region
Temporal7 (35)7 (37)20 (35)
Frontal6 (30)5 (26)22 (39)
Parietal2 (10)1 (5)6 (10)
Occipital0 (0)2 (11)5 (9)
Other5 (25)4 (21)4 (7)
Tumor Side
Right8 (40)10 (53)34 (60)
Left7 (35)5 (26)19 (33)
Other5 (25)4 (21)4 (7)
Total resection12 (60)13 (69)41 (72)
Subtotal resection7 (35)5 (26)12 (21)
Partial resection1 (5)1 (5)4 (7)
No treatment10 (50)4 (21)5 (9)
Radiotherapy5 (25)8 (42)39 (68)
Radiotherapy and Chemotherapy5 (25)7 (37)13 (23)

Table 1. Summary of astrocytoma patient characteristics.

LGA: low-grade astrocytoma, AA: anaplastic astrocytoma, GBM: glioblastoma.
Download CSV

An independent cohort of 71 newly diagnosed patients with primary astrocytomas WHO grades III and IV (12 anaplastic astrocytomas and 59 glioblastomas) [19] was explored to validate the prognostic results. Patients were admitted from April 2004 until December 2010 to the University Hospital 12 de Octubre (Madrid, Spain) and followed up from diagnosis to the present at the Neurosurgery and Oncology Departments. The clinical characteristics of this validation cohort are listed in Table S1.

All blood and tissue samples were obtained at diagnosis before initiation of treatment. Matched DNA from peripheral blood and frozen tumor specimens were extracted by standard phenol/chloroform procedure. Tumor tissues were also fixed in formalin and embedded in paraffin for immunohistochemical analysis.

Tissue microarray design and Immunohistochemistry

Formalin-fixed paraffin-embedded tissue samples of all patients of the study cohort were used to prepare a tissue microarray (TMA) made with a tissue arrayer device (Beecher Instrument, MD). All astrocytomas were histologically reviewed and three 1-mm-diameter cylinders from different areas of morphologically representative non-necrotic sites of each tumor were included to ensure the quality, reproducibility and homogenous staining of the slides. Thus, 3 different TMA blocks were constructed, each containing the 96 astrocytomas and 8 different tissue controls (lymph node, skeletal muscle, cerebral cortex, cerebellum, spleen, adrenal gland, lung and kidney). Immunohistochemical staining was performed on these sections using MLH1 clone G168-15 (BD Pharmingen, San Diego, CA, USA), MSH2 clone FE11 (Biocare Medical, Concord, CA, USA) and MSH6 clone BC/44 (Biocare Medical, Concord, CA, USA) antibodies. MSH6 immunohistochemical staining was also performed in formalin-fixed, paraffin-embedded tumor tissue sections from the 72 patients in the validation cohort. MMR protein expression was evaluated semiquantitatively by a pathologist (A.S-B) and an author (I.R-H) blinded to clinical and molecular information and disagreements between reviewers were resolved by the analysis of the slides by a third author (R.G-S). Tumor area was considered positive when there was obvious nuclear staining in more than 50% of tumor cells. Staining of nuclei of adjacent normal cells and tumor-infiltrating lymphocytes was used as internal positive controls. Figure 1 shows representative examples of low- and high-grade astrocytomas staining.

Figure 1. Representative images of hematoxylin and eosin (H&E), MLH1, MSH2 and MSH6 staining on paraffin-embedded sections of samples from representatives low-grade astrocytomas (grade II) (A-D) and glioblastomas (grade IV) (E-H) according to the 2007 WHO classification [19].

Low-grade astrocytomas are well differentiated and slow-growing tumors with absence of necrosis and microvascular proliferation, whereas high-grade astrocytomas are characterized by high cellularity and mitotic activity, necrosis and microvascular proliferation (arrows). MLH1, MSH2 and MSH6 expression was visualized by staining with specific antibodies and their expression was considered positive when nuclear staining was detected in more than 50% of tumor cells (Magnification x400).

Methylation-specific multiplex ligation-dependent probe amplification (MS-MLPA)

The SALSA MS-MLPA Kit ME011 (MRC-Holland) was used to detect aberrant methylation in MLH1, MSH2 and MSH6 promoter regions using probes that recognized sequences containing a methylation-sensitive HhaI restriction site. The kit included 5 specific probes for MLH1 (located at -659, -383, -246, -13 and +206 relative to initiating ATG), 3 probes for MSH2 (located at -269, -193 and +124 to initiating ATG) and 3 probes for MSH6 promoter region (located at -317, -126 and -32 to initiating ATG). All reactions were carried out as described by the manufacturer with minor modifications using in each reaction 150 ng of tumor DNA. PCR reaction fragments were separated by capillary gel electrophoresis (ABI Prism 3100 Genetic Analyzer, Applied Biosystems) and quantified using Genemapper software (Applied Biosystems). MS-MLPA processing was performed using Coffalyser analysis tool developed at MRC-Holland and tumor samples with a cut off value >0.75 was considered to be extensively hypermethylated as described by Jeuken et al [20]. For MLH1 analysis, methylation status was calculated considering only the two specific probes related to gene silencing (-246 and -13 positions corresponding to C and D promoter regions) [21].

Determination of MLH1 -93G>A genotype

Genotyping was performed using DNA extracted from peripheral blood of 96 patients and 200 sex-matched healthy subjects over 60 years old without history of cancer. MLH1 -93G>A promoter polymorphism (rs1800734) status was determined using TaqMan SNP Genotyping Assay ID C_7535141 (Applied Biosystems) containing sequence-specific forward and reverse primers to amplify the polymorphic sequence and two probes labeled with VIC and FAM dyes to detect both alleles [22]. PCR reactions were carried out using TaqMan universal PCR Master Mix (Applied Biosystems) following manufacturer’s instructions in a Step-One Plus Real-time PCR system (Applied Biosystems). Genotype distribution in the control group was within the Hardy-Weinberg equilibrium (P > 0.1).

Microsatellite Instability (MSI)

MSI was assessed by PCR in paired peripheral blood and tumor DNA obtained from 88 patients using a panel of 8 markers: 3 mononucleotide markers (BAT25, BAT26 and BAT40), 3 dinucleotide markers (D2S123, D5S346 and D17S250) and 2 tetranucleotide markers (MYCL and PAX6). This MSI marker set included the National Cancer Institute recommended markers for MSI detection in Lynch syndrome [11,23]. The 5’ antisense primers were labeled with FAM for BAT26, D5S346, D17S250 and D2S123, HEX for BAT40, MYCL and PAX6, and TET for BAT25. PCR reactions were performed using the Go Taq Hot Start Polymerase (Promega). PCR products were separated using an ABI Prism 3100 Genetic Analyzer (Applied Biosystems) and results were analyzed with Genemapper software (Applied Biosystems). Tumors were classified as MSI-High (MSI-H) if ≥ 30% markers demonstrated instability, MSI-Low (MSI-L) if < 30% demonstrated MSI, and microsatellite stable (MSS) if no marker exhibited MSI according to the established criteria for MSI determination [11,23].

Mutational analysis of MLH1, MSH2 and MSH6 genes

Mutational analysis of MMR genes was performed by PCR-CSGE (Conformation-Sensitive Gel Electrophoresis) [24] in tumor DNA of patients with negative immunohistochemical staining for MLH1, MSH2 and/or MSH6 and in patients with MSI-H tumors. Those samples that showed a mobility shift in the CSGE analysis were additionally analyzed by direct sequencing to identify the nature of mutations. In addition, DNA extracted from peripheral blood was analyzed to determine the somatic or germline origin of tumor mutations. Primers sequences are available upon request.

Statistical analysis

In most analysis, WHO grade III and IV astrocytomas are analyzed together as high-grade astrocytomas. The results were expressed as percentages for categorical variables and as medians [quartiles] for continuous variables. Associations between molecular and clinicopathological features as well as Hardy-Weinberg equilibrium were analyzed using the χ2 contingency test and the Fisher’s exact test when necessary (expected values below 5). Survival models were used to find clinical and/or molecular parameters related to overall survival (OS). OS was defined as the time between diagnosis and death or last follow-up. Those patients lost during follow-up were censored at the last known follow-up date and patients with a survival time lesser than 30 days were eliminated since these patients might have died for reasons other than the disease itself. Survival curves were estimated by the Kaplan Meier method and compared among patient subsets using the log-rank test. Multivariate Cox model was used to identify independent prognostic factors correlated with survival in astrocytoma patients. All categories were first ordered from good to bad with respect to prognosis. Then the corresponding Hazard Ratios (HR) will be greater than one helping the interpretation. Differences with a P-value <0.05 were considered as statistically significant and all tests were two-sided. All statistical analyses were performed using SPSS software.


MMR protein expression

MLH1, MSH2 and MSH6 protein expression was determined in the 96 tumor tissues included in the tissue microarray. Analysis of MMR protein expression showed that 41 tumors (43%) presented loss of expression of at least one MMR protein. Loss of MLH1 expression was observed in 17 cases (18%), MSH2 in 21 cases (22%) and MSH6 in 31 cases (32%). In addition, 10 tumors failed to express two MMR proteins simultaneously (MLH1 and MSH2 expression were lost in 1 case, MLH1 and MSH6 expression in 4 cases, and MSH2 and MSH6 expression in 5 cases). Nine tumors did not express the three MMR proteins. Furthermore, MSH2 negative staining was significantly more frequent in low-grade astrocytomas (45%) than in high-grade astrocytomas (16%) (P=0.012), whereas MLH1 and MSH6 expression were not associated with tumor histopathology (Table 2).

Patients, No. (%)MLH1 ExpressionMSH2 ExpressionMSH6 Expression
Tumor grade0.7210.0120.805
Low-grade astrocytomas17 (85)3 (15)11 (55)9 (45)14 (70)6 (30)
High-grade astrocytomas62 (82)14 (18)64 (84)12 (16)51 (67)25 (33)
Promoter Methylation*0.0210.6990.713
Hypermethylated6 (54)5 (46)8 (73)3 (27)5 (62)3 (38)
No hypermethylated70 (86)11 (14)64 (79)17 (21)57 (68)27 (32)

Table 2. Relation between tumor grade, promoter methylation status and MLH1, MSH2 and MSH6 protein expression.

(* MLH1, MSH2 and MSH6 expression was associated with MLH1, MSH2 and MSH6 promoter methylation status respectively)
Download CSV

MMR promoter methylation

MMR promoter methylation was studied in 92 patients. Promoter hypermethylation of MLH1 was present in 11 cases (12%), MSH2 in 11 cases (12%) and MSH6 in 8 cases (9%). No differences in MMR methylation levels were identified between different tumor grades. Methylation status of MLH1 promoter was significantly associated with loss of MLH1 expression (P=0.021). Specifically, 46% of tumors with MLH1 promoter hypermethylation showed lack of MLH1 expression (Table 2). We did not find any association between MSH2 and MSH6 promoter methylation and their protein expression (Table 2).

Analysis of MLH1 -93G>A polymorphism

MLH1 -93G>A polymorphism was determined in 96 astrocytoma patients and 200 control subjects. The distribution of genotypes in control samples did not significantly differ from that expected from Hardy-Weinberg equilibrium. MLH1 -93GG genotype was found in 57 (60%), GA genotype in 31 (32%) and AA genotype in 8 (8%) patients. We did not find significant differences in distribution of MLH1 -93G>A genotypes between low-grade astrocytomas, anaplastic astrocytomas and control subjects. However, carriers of the MLH1 -93AA genotype were more represented in the group of glioblastoma patients (P=0.017) (Table 3). Furthermore, the AA genotype was more frequent in tumors with MLH1 promoter hypermethylation (P<0.001) and in tumors that showed lack of MLH1 expression (P=0.030) (Table 3).

Patients, No. (%)Genotype
GG + GAAAP-value
Diagnostic group
LGA19 (95)1 (5)0.540
AA19 (100)0 (0)1.000
GBM50 (88)7 (12)0.017
Controls193 (96)7 (4)
MLH1 Hypermethylation< 0.001
Hypermethylated6 (55)5 (45)
No hypermethylated78 (96)3 (4)
MLH1 Expression0.030
Positive75 (95)4 (5)
Negative13 (76)4 (24)

Table 3. Distribution of MLH1 -93G>A genotypes according to diagnostic group, MLH1 promoter hypermethylation and MLH1 protein expression.

Download CSV

Microsatellite Instability

MSI analysis was performed in 88 cases and revealed that the most frequent unstable markers in our series were BAT25 (15% of cases) and BAT40 (13% of cases). On the contrary, BAT26 was unstable in 1% of cases.

Forty-seven patients (53%) were classified as MSS, 37 patients (42%) as MSI-L and 4 patients (5%) as MSI-H, with no differences between tumor grades. The four MSI-H tumors were glioblastomas. Three of these cases showed expression of all MMR proteins and one case presented loss of MSH6 expression. Neither MMR protein expression nor MMR promoter methylation nor MLH1 -93AA genotype were associated with MSI status (Table S2).

Germline and somatic mutations of MMR genes

Mutation analysis of MMR genes was performed in the forty-four tumors that showed abnormalities in MMR protein expression or MSI-H. Six of these tumors carried mutation in MLH1, MSH2 or MSH6 genes (Table 4). Four tumors presented MSS with MMR germline mutations and lack of expression of the associated protein; one glioblastoma with MSI-L showed two mutations in the MSH2 gene, one mutation in the MLH1 gene and loss of MLH1, MSH2 and MSH6 protein expression. The remained case was a MSI-H glioblastoma with loss of MSH6 expression and a family history of colorectal cancer. This patient showed a pathogenic germline frameshift mutation in MSH2 gene and wild-type allele loss in tumor tissue.

MSICaseGeneMMR Germline MutationMMR Somatic MutationPathogenicityLoss of MMR expression
MSSGBMMLH1c.2146GA (p.Val716Met)[25]UncertainMLH1, MSH2
MSSLGAMSH2c.1159CG (p.Leu387Val)UncertainMSH2
MSSLGAMSH6c.4004AC (p.Glu1335Ala)UncertainMLH1, MSH2, MSH6
MSSGBMMSH6c. *(24_28)delGTTGAUncertainMSH6
MSI-LGBMMLH1c.1937AG (p.Tyr646Cys)[26]UncertainMLH1, MSH2, MSH6
MSH2c.1983delA (p.Lys661AsnfsX24)Pathogenic
MSH2c.1064GA (p.Arg355Lys)Uncertain
MSI-HGBMMSH2c.2239_2240delAT (p.Ile747ArgfsX2)[27]Wt allele loss in tumorPathogenicMSH6

Table 4. MMR gene mutations identified in tumors with loss of at least one MMR protein expression and/or MSI-H.

Download CSV

We described for the first time in this study three novel mutations in MSH2 gene c.1159CG (p.Leu387Val), c.1983delA (p.Lys661AsnfsX24) and c.1064GA (p.Arg355Lys), and two novel mutations in MSH6 gene c.4004AC (p.Glu1335Ala) and c. *(24_28) delGTTGA (Table 4).

Impact of MMR alterations on patient survival

Next, we investigated the prognostic impact of both MMR alterations and clinical parameters on 92 astrocytoma patients (20 low-grade astrocytomas and 72 high-grade astrocytomas). Univariate survival analysis revealed that loss of MSH6 expression was significantly associated with a better median OS in high-grade astrocytomas (13.8 months vs. 10.1 months) (Table 5A). Age of patients (<60 years vs. ≥60 years) and treatment (radiotherapy and chemotherapy vs. radiotherapy alone vs. no treatment) were also significantly associated with OS in high-grade astrocytomas (Table 5A). Furthermore, multivariate Cox model indicated that MSH6 expression, age and treatment were statistically significant independent prognostic factors for OS of high grade-astrocytomas in our series (Table 5A).

ParameterUnivariate AnalysisMultivariate Analysis
HR (95% C.I.)P-valueHR (95% C.I.)P-value
MSH6 expression1.76 (1.01-3.07)0.0451.84 (1.05-3.23)0.033
Age2.03 (1.18-3.50)0.0091.75 (1.02-3.02)0.042
Treatment2.53 (1.56-4.11)0.0012.42 (1.52-3.87)< 0.001
MSH6 expression1.30 (0.66-2.56)0.4431.56 (0.77-3.19)0.219
Age1.76 (0.99-3.12)0.0511.74 (0.98-3.11)0.060
Treatment1.55 (1.01-2.47)< 0.0011.70 (1.02-2.84)0.041

Table 5. Univariate and multivariate survival analysis in high-grade astrocytomas.

Download CSV

In order to validate the prognostic value of MSH6 expression on high-grade astrocytomas survival, we evaluated MSH6 protein expression in an independent cohort of 71 high-grade astrocytomas. Fifteen (21%) of these tumors showed loss of MSH6 expression (Figure S1). However, survival analysis revealed that MSH6 expression was not associated with prognosis, whereas treatment was an independent prognostic factor for OS and there was marginally significant correlation between age and OS time in this validation group of patients (Table 5B).

We further analyzed the differences between the study cohort and the validation cohort of patients in order to unravel the role of MSH6 protein expression in patients diagnosed of astrocytoma. We found that the number of patients who received radiotherapy and chemotherapy was significantly higher in the validation cohort compared to the study cohort (Table 6). Consequently, those patients belonged to the validation cohort had a better outcome compared to the study cohort (Table 6). Since treatment itself constitute a strong prognostic marker (Figure 2A), we stratified the entire patient set by treatment and we analyzed the association between MSH6 expression and clinical outcome in each treatment subgroup separately. Loss of MSH6 expression was associated with a better median OS time in the group of high-grade astrocytoma patients only treated with radiotherapy (HR=2.17, 95% C.I. 1.14-4.11, P=0.015) (Figure 2B); meanwhile no significant differences were found in the group of patients treated with radiotherapy plus chemotherapy (HR=1.21, 95% C.I. 0.64-2.26, P=0.558) (Figure 2C). Survival analysis could not be estimated in the group of patients that did not receive any treatment due to small number of events.

Patients, No. (%)Study cohort (n=72)Validation cohort (n=71)P-value
MSH6 expression0.198
Positive50 (69)56 (79)
Negative22 (31)15 (21)
<6031 (43)36 (51)
≥6041 (57)35 (49)
Treatment< 0.001
No treatment5 (7)2 (3)
Radiotherapy47 (65)12 (17)
Radiotherapy and Chemotherapy20 (28)57 (80)
Median survival, months [quartiles]12.7 [10.1-15.3]15.1 [10.8-19.4]0.012

Table 6. Differences in prognostic characteristics between the study cohort and the validation cohort of high-grade astrocytoma patients.

Download CSV
Figure 2. Kaplan-Meier estimates of overall survival in high-grade astrocytomas in the entire patient set according to treatment received.

Treatment with both radiotherapy and chemotherapy confers a significant increase in overall survival time (A). Survival analysis in each treatment group separately showed that loss of MSH6 expression correlated with a better overall survival in patients receiving radiation therapy alone (B), whereas MSH6 expression did not modify prognosis of patients receiving both radiotherapy and chemotherapy (C).


Malignant astrocytomas are one of the most devastating cancers with a dismal prognosis. Virtually all high-grade astrocytomas progress and locally relapse regardless of improved diagnosis and multi-modality treatment approach [2,3]. Therefore, identification of new markers may contribute to a better prediction of prognosis and response to therapy. Astrocytomas are characterized by an infiltrating and aggressive behavior directly related to their genetic alterations in core signaling pathways [28]. In this regard, MMR activity could be implicated in astrocytoma pathogenesis due to the fact that loss of MMR function accelerates the accumulation of mutations that are no longer repaired.

We have identified a large number of astrocytomas with defective MMR system expression. Forty-three percent of tumors included in our series failed to express at least one MMR protein at diagnosis, suggesting that this abnormality could be an intrinsic property of a subgroup of tumors. We also showed that MLH1 and MSH6 expression profiles were similar in both low- and high-grade astrocytomas whereas the lack of MSH2 expression was significantly more frequent in low-grade astrocytomas. In this sense, it has been previously reported an increased expression of MSH2 in high-grade astrocytomas compared with low-grade astrocytomas. Therefore, up-regulation of MSH2 levels may be related to an increased cell proliferation rate in astrocytomas [29,30]. Although MSH2 expression is typically lost in colorectal or endometrial tumors [13,14,16], high levels of MSH2 expression have also been described in more malignant and proliferative melanoma and salivary gland grade tumors [31,32].

Loss of protein expression could be due to aberrant DNA methylation of cytosine residues in CpG promoter islands that leads to transcriptional silencing of the associated genes [33]. We have found a specific association between MLH1 methylation of the proximal promoter region and the absence of MLH1 expression in astrocytomas. Thus, we further confirm that MLH1 proximal promoter methylation is important in inhibiting MLH1 transcription as it has been previously reported in colorectal cancer (CRC) [21]. Moreover, our results suggest that MSH2 and MSH6 expression are not regulated by their promoter methylation status in astrocytomas.

Analysis of the genotypic distribution of MLH1 -93G>A polymorphism, located in a promoter region required for maximal transcriptional activity [34], showed that the -93AA genotype was associated with MLH1 promoter methylation and deficient MLH1 expression never reported before in astrocytomas. The -93AA genotype was also associated with a higher risk of developing glioblastomas. This variant has been previously associated with loss of functional MMR system in colorectal and endometrial cancers [17,18,35] and with an increased risk of developing different tumors [17,36,37]. Nevertheless, further studies in larger series of patients are necessary to confirm our observation.

MSI analysis revealed a low incidence of MSI-H tumors in our series (5%) similar to that previously reported [38-41]. All MSI-H cases were classified as glioblastomas, suggesting a possible relation with higher proliferation cell levels. MSI is a molecular feature resulting mainly from inactivating alterations of the MMR system; however, loss of MMR protein expression was not related to MSI status in our study. MMR deficiencies are well associated with MSI status in several tumors, such as ovarian, endometrial or colorectal cancers [11,14,42], but in other tumors such as medulloblastoma or Ewing sarcoma this association has not been reported [43,44]. These findings suggest that MMR protein deficiencies are related to MSI depending on the tumor type [44]. In addition, we observed that BAT25 and BAT40 were the most frequent unstable markers in astrocytomas instead of BAT26 that is the most unstable marker in CRC. This difference might indicate that MSI in astrocytomas is promoted by different mechanisms than in CRC.

Search for mutations in MMR genes showed only two pathogenic mutations. A germline pathogenic mutation in MSH2 gene [27] was found in a MSI-H glioblastoma belonged to a Lynch syndrome family. The other pathogenic mutation was a novel somatic MSH2 mutation in a MSI-L glioblastoma that also carried two additional variants of unknown significance in MSH2 and MLH1 genes. One MLH1, one MSH2 and two MSH6 mutations of unknown significance that could be causing the lack of expression of the associated proteins were also detected. The absence of MMR gene mutations in most of the sporadic MSI-H astrocytomas was in accordance with data reported from sporadic CRC with high level of MSI that do not harbor mutations in the repair genes [15,16].

Loss of MSH6 expression was more common than the absence of MLH1 or MSH2 expression in our series, suggesting an important role of this protein in the pathogenesis of astrocytomas. It has been recently reported that MSH6 alterations arise in gliomas as a consequence of temozolomide treatment [45,46]. Nevertheless, MSH6 alterations have also been documented in pretreated astrocytoma tumors [47,48]. We must note that all cases included in our study were analyzed before therapy, confirming that loss of MSH6 expression in astrocytomas is not always secondary to therapy-induced mutagenesis.

Furthermore, survival analysis showed that loss of MSH6 expression was significantly associated with longer overall survival in high-grade astrocytoma patients in our series. Controversially, we did not find this association when we analyzed MSH6 expression in an independent validation cohort of high-grade astrocytoma patients. However, the treatment approach was different between the two cohorts of patients due to the different recruitment period of patients in each series. For decades, postoperative radiotherapy has been the standard treatment for newly diagnosed high-grade astrocytomas [49]. However, since the 2005 publication of Stupp regimen [50], the established standard therapy for newly diagnosed high-grade astrocytomas has been surgical resection followed by radiotherapy with concomitant and adjuvant chemotherapy with temozolomide with the consequent significant increase in overall survival time [3,50,51]. Therefore, survival analysis was performed in each treatment group separately and revealed that loss of MSH6 expression was significantly associated with longer overall survival in patients with high-grade astrocytomas treated with radiotherapy alone, whereas no differences were found in those patients that received radiotherapy plus chemotherapy. The chemotherapy agent used in the majority of these patients was the alkylating agent temozolomide.

Several studies have suggested that loss of MSH6 activity confers resistance to temozolomide treatment in glioma and may therefore contribute to progressive tumor growth and tumor recurrence [45,46]. However, several analyses concluded that MMR deficiency does not play a role in clinical resistance to alkylator therapy in malignant gliomas [47]. This finding was in accordance to our results that MSH6 expression does not modify the prognosis of those patients treated with radiotherapy and chemotherapy. On the other hand, the roles of MMR deficiency in radiotherapy response are less clear. In our series, loss of MSH6 expression confers a better prognosis in high-grade astrocytomas treated only with radiotherapy suggesting that MSH6 protein could modulate response to radiation therapy in these tumors. It has been reported that loss MSH6 expression in initial lesions was indicator of prolonged survival in a group of patients treated mostly with postoperative radiotherapy [48]. Ionizing radiation (IR) exposure induces a wide variety of lethal DNA damage, especially double-strand breaks (DSB). Although MSH6 plays a fundamental role in the repair of mismatched DNA bases, recent studies demonstrated that MSH6 contributes to DSB repair though the non-homologous end-joining (NHEJ) pathway following IR exposure by the interaction of MSH6 with Ku70 [52]. Thus, MSH6-deficient cells have a major DSB repair defect and are more sensitive to IR-induced cell death [52]. Therefore, our results indicate that MSH6 expression might constitute a prognostic marker for astrocytoma survival in patients treated only with radiotherapy.

In summary, our results demonstrate that MMR system alterations are a frequent event in malignant astrocytomas. We suggest that analysis of MMR genes allows to define a subset of astrocytomas with different outcome and could help to search for new therapeutic strategies.

Supporting Information

Table S1.

Detailed clinical characteristics of the study and validation cohort of astrocytoma patients.



Table S2.

MLH1, MSH2 and MSH6 expression and methylation status according to MSI levels.



Figure S1.

Representative examples of negative (A, B) and positive (C, D) MSH6 staining in four glioblastoma tumors (WHO grade IV). MSH6 expression was analyzed in formalin fixed, paraffin embedded tumor sections from all patients in the study cohort and the validation cohort. MSH6 staining on these sections was performed using MSH6 clone BC/44 (Biocare Medical, Concord, CA, USA) antibody and counterstained with hematoxylin and eosin. MSH6 staining was considered positive when there was obvious nuclear staining in more than 50% of tumor cells (Magnification, x400).




The authors thank Nieves Mateos and Mª Angeles Hernández for their technical assistance. We are especially grateful to Dr Luis Pérez del Villar for his encourage and critical review of the manuscript.

Author Contributions

Conceived and designed the experiments: IRH JLG RGS. Performed the experiments: IRH JLG ASB JMGV RGS. Analyzed the data: IRH JLG JMV RGS. Contributed reagents/materials/analysis tools: JLG ASB RGS. Wrote the manuscript: IRH JLG RGS. Sample collection and updating patient information: AHL JAGM OTG JJC.


  1. 1. Ohgaki H, Kleihues P (2005) Population-based studies on incidence, survival rates, and genetic alterations in astrocytic and oligodendroglial gliomas. J Neuropathol Exp Neurol 64: 479-489. PubMed: 15977639.
  2. 2. Wehming FM, Wiese B, Nakamura M, Bremer M, Karstens JH et al. (2012) Malignant glioma grade 3 and 4: How relevant is timing of radiotherapy? Clin Neurol Neurosurg 114: 617-621. doi:10.1016/j.clineuro.2011.12.024. PubMed: 22244251.
  3. 3. Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ et al. (2009) Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol 10: 459-466. doi:10.1016/S1470-2045(09)70025-7. PubMed: 19269895.
  4. 4. Ohgaki H (2009) Epidemiology of brain tumors. Methods Mol Biol 472: 323-342. doi:10.1007/978-1-60327-492-0_14. PubMed: 19107440.
  5. 5. Wen PY, Kesari S (2008) Malignant gliomas in adults. N Engl J Med 359: 492-507. doi:10.1056/NEJMra0708126. PubMed: 18669428.
  6. 6. Mellinghoff IK, Wang MY, Vivanco I, Haas-Kogan DA, Zhu S et al. (2005) Molecular determinants of the response of glioblastomas to EGFR kinase inhibitors. N Engl J Med 353: 2012-2024. doi:10.1056/NEJMoa051918. PubMed: 16282176.
  7. 7. Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA et al. (2009) IDH1 and IDH2 mutations in gliomas. N Engl J Med 360: 765-773. doi:10.1056/NEJMoa0808710. PubMed: 19228619.
  8. 8. The Cancer Genome Atlas (TCGA) Research Network (2008) Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 455: 1061-1068.
  9. 9. Weller M, Stupp R, Reifenberger G, Brandes AA, van den Bent MJ et al. (2010) MGMT promoter methylation in malignant gliomas: ready for personalized medicine? Nat. Rev Neurol 6: 39-51. doi:10.1038/nrneurol.2009.197.
  10. 10. Huse JT, Holland EC (2010) Targeting brain cancer: advances in the molecular pathology of malignant glioma and medulloblastoma. Nat Rev Cancer 10: 319-331. doi:10.1038/nrc2818. PubMed: 20414201.
  11. 11. Umar A, Boland CR, Terdiman JP, Syngal S, de la Chapelle A et al.. (2004) 2004) Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst 96: 261-268. doi:10.1093/jnci/djh034. PubMed: 14970275.
  12. 12. Watson P, Vasen HF, Mecklin JP, Bernstein I, Aarnio M et al. (2008) The risk of extra-colonic, extra-endometrial cancer in the Lynch syndrome. Int J Cancer 123: 444-449. doi:10.1002/ijc.23508. PubMed: 18398828.
  13. 13. Koornstra JJ, Mourits MJ, Sijmons RH, Leliveld AM, Hollema H et al. (2009) Management of extracolonic tumours in patients with Lynch syndrome. Lancet Oncol 10: 400-408. doi:10.1016/S1470-2045(09)70041-5. PubMed: 19341971.
  14. 14. Lynch HT, Lynch PM, Lanspa SJ, Snyder CL, Lynch JF et al. (2009) Review of the Lynch syndrome: history, molecular genetics, screening, differential diagnosis, and medicolegal ramifications. Clin Genet 76: 1-18. doi:10.1111/j.1399-0004.2009.01230.x. PubMed: 19659756.
  15. 15. Peltomäki P (2005) Lynch syndrome genes. Fam Cancer 4: 227-232. doi:10.1007/s10689-004-7993-0. PubMed: 16136382.
  16. 16. Kastrinos F, Syngal S (2012) Screening Patients With Colorectal Cancer for Lynch Syndrome: What Are We Waiting For? J Clin Oncol 30: 1024-1027. doi:10.1200/JCO.2011.40.7171. PubMed: 22355054.
  17. 17. Allan JM, Shorto J, Adlard J, Bury J, Coggins R et al. (2008) MLH1 -93G>A promoter polymorphism and risk of mismatch repair deficient colorectal cancer. Int J Cancer 123: 2456-2459. doi:10.1002/ijc.23770. PubMed: 18712731.
  18. 18. Mrkonjic M, Roslin NM, Greenwood CM, Raptis S, Pollett A et al. (2010) Specific variants in the MLH1 gene region may drive DNA methylation, loss of protein expression, and MSI-H colorectal cancer. PLOS ONE 5: e13314. doi:10.1371/journal.pone.0013314. PubMed: 20967208.
  19. 19. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC et al. (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114: 97-109. doi:10.1007/s00401-007-0243-4. PubMed: 17618441.
  20. 20. Jeuken JW, Cornelissen SJ, Vriezen M, Dekkers MM, Errami A et al. (2007) MS-MLPA: an attractive alternative laboratory assay for robust, reliable, and semiquantitative detection of MGMT promoter hypermethylation in gliomas. Lab Invest 87: 1055-1065. doi:10.1038/labinvest.3700664. PubMed: 17700563.
  21. 21. Capel E, Fléjou JF, Hamelin R (2007) Assessment of MLH1 promoter methylation in relation to gene expression requires specific analysis. Oncogene 26: 7596-7600. doi:10.1038/sj.onc.1210581. PubMed: 17546041.
  22. 22. Schleinitz D, Distefano JK, Kovacs P (2011) Targeted SNP genotyping using the TaqMan(R) assay. Methods Mol Biol 700: 77-87. doi:10.1007/978-1-61737-954-3_6. PubMed: 21204028.
  23. 23. Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR et al. (1998) A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58: 5248-5257. PubMed: 9823339.
  24. 24. Ganguly A, Rock MJ, Prockop DJ (1993) Conformation-sensitive gel electrophoresis for rapid detection of single-base differences in double-stranded PCR products and DNA fragments: evidence for solvent-induced bends in DNA heteroduplexes. Proc Natl Acad Sci U S A 90: 10325-10329. doi:10.1073/pnas.90.21.10325. PubMed: 8234293.
  25. 25. Barnetson RA, Cartwright N, van Vliet A, Haq N, Drew K et al. (2008) Classification of ambiguous mutations in DNA mismatch repair genes identified in a population-based study of colorectal cancer. Hum Mutat 29: 367-374. doi:10.1002/humu.20635. PubMed: 18033691.
  26. 26. Belvederesi L, Bianchi F, Loretelli C, Gagliardini D, Galizia E et al. (2006) Assessing the pathogenicity of MLH1 missense mutations in patients with suspected hereditary nonpolyposis colorectal cancer: correlation with clinical, genetic and functional features. Eur J Hum Genet 14: 853-859. doi:10.1038/sj.ejhg.5201628. PubMed: 16724012.
  27. 27. Caldes T, Godino J, de la Hoya M, Garcia Carbonero I, Perez Segura P et al. (2002) Prevalence of germline mutations of MLH1 and MSH2 in hereditary nonpolyposis colorectal cancer families from Spain. Int J Cancer 98: 774-779. doi:10.1002/ijc.10240. PubMed: 11920650.
  28. 28. Ricard D, Idbaih A, Ducray F, Lahutte M, Hoang-Xuan K et al. (2012) Primary brain tumours in adults. Lancet 379: 1984-1996. doi:10.1016/S0140-6736(11)61346-9. PubMed: 22510398.
  29. 29. Hussein MR, El-Ghorori RM, El-Rahman YG (2006) Alterations of p53, BCL-2, and hMSH2 protein expression in the normal brain tissues, gliosis, and gliomas. Int J Exp Pathol 87: 297-306. doi:10.1111/j.1365-2613.2006.00482.x. PubMed: 16875495.
  30. 30. Srivastava T, Chattopadhyay P, Mahapatra AK, Sarkar C, Sinha S (2004) Increased hMSH2 protein expression in glioblastoma multiforme. J Neuro Oncol 66: 51-57. doi:10.1023/B:NEON.0000013482.99032.b0. PubMed: 15015769.
  31. 31. Castrilli G, Fabiano A, La Torre G, Marigo L, Piantelli C et al. (2002) Expression of hMSH2 and hMLH1 proteins of the human DNA mismatch repair system in salivary gland tumors. J Oral Pathol Med 31: 234-238. doi:10.1034/j.1600-0714.2002.310407.x. PubMed: 12076327.
  32. 32. Rass K, Gutwein P, Welter C, Meineke V, Tilgen W et al. (2001) DNA mismatch repair enzyme hMSH2 in malignant melanoma: increased immunoreactivity as compared to acquired melanocytic nevi and strong mRNA expression in melanoma cell lines. Histochem J 33: 459-467. doi:10.1023/A:1014472314354. PubMed: 11931386.
  33. 33. Esteller M (2007) Epigenetic gene silencing in cancer: the DNA hypermethylome. Hum Mol Genet 16: Spec No 1: R50-. 59. PubMed: 17613547.
  34. 34. Arita M, Zhong X, Min Z, Hemmi H, Shimatake H (2003) Multiple sites required for expression in 5'-flanking region of the hMLH1 gene. Gene 306: 57-65. doi:10.1016/S0378-1119(03)00385-8. PubMed: 12657467.
  35. 35. Chen H, Taylor NP, Sotamaa KM, Mutch DG, Powell MA et al. (2007) Evidence for heritable predisposition to epigenetic silencing of MLH1. Int J Cancer 120: 1684-1688. doi:10.1002/ijc.22406. PubMed: 17230510.
  36. 36. Lee KM, Choi JY, Kang C, Kang CP, Park SK et al. (2005) Genetic polymorphisms of selected DNA repair genes, estrogen and progesterone receptor status, and breast cancer risk. Clin Cancer Res 11: 4620-4626. doi:10.1158/1078-0432.CCR-04-2534. PubMed: 15958648.
  37. 37. Park SH, Lee GY, Jeon HS, Lee SJ, Kim KM et al. (2004) 93G-- A polymorphism of hMLH1 and risk of primary lung cancer. Int J Cancer 112: 678-682. doi:10.1002/ijc.20359. PubMed: 15382050.
  38. 38. Eckert A, Kloor M, Giersch A, Ahmadi R, Herold-Mende C et al. (2007) Microsatellite instability in pediatric and adult high-grade gliomas. Brain Pathol 17: 146-150. doi:10.1111/j.1750-3639.2007.00049.x. PubMed: 17388945.
  39. 39. Felsberg J, Thon N, Eigenbrod S, Hentschel B, Sabel MC et al. (2011) Promoter methylation and expression of MGMT and the DNA mismatch repair genes MLH1, MSH2, MSH6 and PMS2 in paired primary and recurrent glioblastomas. Int J Cancer 129: 659-670. doi:10.1002/ijc.26083. PubMed: 21425258.
  40. 40. Viana-Pereira M, Lee A, Popov S, Bax DA, Al-Sarraj S et al. (2011) Microsatellite instability in pediatric high grade glioma is associated with genomic profile and differential target gene inactivation. PLOS ONE 6: e20588. doi:10.1371/journal.pone.0020588. PubMed: 21637783.
  41. 41. Martinez R, Schackert HK, Appelt H, Plaschke J, Baretton G et al. (2005) Low-level microsatellite instability phenotype in sporadic glioblastoma multiforme. J Cancer Res Clin Oncol 131: 87-93. doi:10.1007/s00432-004-0592-5. PubMed: 15672285.
  42. 42. Shah SN, Hile SE, Eckert KA (2010) Defective mismatch repair, microsatellite mutation bias, and variability in clinical cancer phenotypes. Cancer Res 70: 431-435. doi:10.1158/1538-7445.AM10-431. PubMed: 20068152.
  43. 43. Alldinger I, Schaefer KL, Goedde D, Ottaviano L, Dirksen U et al. (2007) Microsatellite instability in Ewing tumor is not associated with loss of mismatch repair protein expression. J Cancer Res Clin Oncol 133: 749-759. doi:10.1007/s00432-007-0220-2. PubMed: 17530287.
  44. 44. Viana-Pereira M, Almeida I, Sousa S, Mahler-Araújo B, Seruca R et al. (2009) Analysis of microsatellite instability in medulloblastoma. Neuro Oncol 11: 458-467. doi:10.1215/15228517-2008-115. PubMed: 19179424.
  45. 45. Yip S, Miao J, Cahill DP, Iafrate AJ, Aldape K et al. (2009) MSH6 mutations arise in glioblastomas during temozolomide therapy and mediate temozolomide resistance. Clin Cancer Res 15: 4622-4629. doi:10.1158/1078-0432.CCR-08-3012. PubMed: 19584161.
  46. 46. Cahill DP, Levine KK, Betensky RA, Codd PJ, Romany CA et al. (2007) Loss of the mismatch repair protein MSH6 in human glioblastomas is associated with tumor progression during temozolomide treatment. Clin Cancer Res 13: 2038-2045. doi:10.1158/1078-0432.CCR-06-2149. PubMed: 17404084.
  47. 47. Maxwell JA, Johnson SP, McLendon RE, Lister DW, Horne KS et al. (2008) Mismatch repair deficiency does not mediate clinical resistance to temozolomide in malignant glioma. Clin Cancer Res 14: 4859-4868. doi:10.1158/1078-0432.CCR-07-4807. PubMed: 18676759.
  48. 48. Stark AM, Doukas A, Hugo HH, Mehdorn HM (2010) The expression of mismatch repair proteins MLH1, MSH2 and MSH6 correlates with the Ki67 proliferation index and survival in patients with recurrent glioblastoma. Neurol Res 32: 816-820. doi:10.1179/016164110X12645013515052. PubMed: 20223108.
  49. 49. Laperriere N, Zuraw L, Cairncross G, Cancer Care Ontario Practice Guidelines Initiative Neuro-Oncology Disease Site G (2002) Radiotherapy for newly diagnosed malignant glioma in adults: a systematic review. Radiother Oncol 64: 259-273. doi: 10.1016/s0167-8140(02)00078-6
  50. 50. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B et al. (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352: 987-996. doi:10.1056/NEJMoa043330. PubMed: 15758009.
  51. 51. DeAngelis LM (2009) Anaplastic glioma: how to prognosticate outcome and choose a treatment strategy. J Clin Oncol 27: 5861-5862. doi:10.1200/JCO.2009.24.5985. PubMed: 19901101.
  52. 52. Shahi A, Lee JH, Kang Y, Lee SH, Hyun JW et al. (2011) Mismatch-repair protein MSH6 is associated with Ku70 and regulates DNA double-strand break repair. Nucleic Acids Res 39: 2130-2143. doi:10.1093/nar/gkq1095. PubMed: 21075794.