If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Two recent genome-wide association studies reported that single nucleotide polymorphisms (SNPs) in (or near) TERT (5p15), CCDC26 (8q24), CDKN2A/B (9p21), PHLDB1 (11q23), and RTEL1 (20q13) are associated with infiltrating glioma. From these reports, it was not clear whether the single nucleotide polymorphism associations predispose to glioma in general or whether they are specific to certain glioma grades or morphologic subtypes. To identify hypothesized associations between susceptibility loci and tumor subtype, we genotyped two case-control groups composed of the spectrum of infiltrating glioma subtypes and stratified the analyses by type. We report that specific germ line polymorphisms are associated with different glioma subtypes. CCDC26 (8q24) region polymorphisms are strongly associated with oligodendroglial tumor risk (rs4295627, odds ratio [OR] = 2.05, P = 8.3 × 10−11) but not glioblastoma risk. The opposite is true of RTEL (20q13) region polymorphisms, which are significantly associated with glioblastoma (rs2297440, OR = 0.56, P = 4.6 × 10−10) but not oligodendroglial tumor. The SNPs in or near CCDC26 (8q24) are associated with oligodendroglial tumors regardless of combined 1p and 19q deletion status; however, the association is greatest for those with combined deletion (rs4295627, OR = 2.77, P = 2.6 × 10−9). These observations generate hypotheses concerning the possible mechanisms by which specific SNPs (or alterations in linkage disequilibrium with such SNPs) are associated with glioma development.
Gliomas cause significant morbidity and mortality. Approximately 18,500 people in the United States are diagnosed with glioma each year. Because most gliomas are biologically aggressive, approximately 12,800 people in the United States die as a result of these tumors annually (
). Like many diseases, glioma development is hypothesized to be associated with relatively common germ line alterations, each with limited penetrance. Two genome-wide association studies reported that single nucleotide polymorphisms (SNPs) in (or near) TERT (5p15), CCDC26 (8q24), CDKN2A/B (9p21), PHLDB1 (11q23), and RTEL1 (20q13) are associated with glioma development (
). Our group analyzed glioblastoma (GBM) and anaplastic astrocytoma (AA) and identified the CDKN2A/B and RTEL1 associations, and observed the TERT association during the discovery phase but not the replication phase (
). We hypothesized that the different results were primarily due to varied glioma case mixes in the two studies. Differential histological associations of risk loci have been reported in lung cancer (
). To test our hypothesis, we genotyped two case-control groups composed of a spectrum of glioma subtypes.
With 582 cases and 532 control subjects from the Mayo Clinic for discovery and 864 cases and 602 control subjects from UCSF for confirmation, we examined the top 13 SNPs in the five associated regions from the two published genome-wide association studies (
). To our knowledge, data regarding patients with oligodendroglial tumors (n = 390) and grade II astrocytoma (n = 103) have not been previously published.
Materials and methods
Study populations
Mayo Clinic case-control study
The Mayo Clinic case group included 582 individuals with glioma newly diagnosed between 2005 and 2009. Cases were identified within 24 hours of diagnosis, except for those who had their initial diagnosis elsewhere, followed by verification at the Mayo Clinic. Pathologic diagnosis was confirmed by review of the primary surgical material for all cases by two Mayo Clinic neuropathologists, C. Giannini and B. Scheithauer, on the basis of surgically resected material. The control group consisted of consented individuals who underwent a general medical examination at the Mayo Clinic and matched to cases by sex, date of birth (within 2.5 years), self-identified race (Hispanic white, non-Hispanic white, American Indian, African American, Asian, Pacific Islander, other), and residence. Geographic region of residence was matched in three zones according to the distance to Mayo Clinic Rochester: Olmsted County; the rest of Minnesota, Wisconsin, Iowa, North Dakota, and South Dakota; and the rest of the United States and Canada. Excluded were individuals under the age of 18 and those with a history of brain tumor. The 1p/19q deletion data was obtained from the medical record, or if these data were unavailable, by routine fluorescence in situ hybridization (FISH) analysis (
). The Mayo Clinic case and control enrollment research protocol was approved by the Mayo Institutional Review Board. Mayo Clinic subject characteristics are summarized in Supplemental Table 1.
UCSF case-control study
UCSF cases and control subjects were taken from the San Francisco Bay Area Adult Glioma Study (AGS). Details of subject recruitment for AGS have been provided previously (
). Briefly, patients aged 20 or older diagnosed with histologically confirmed incident gliomas (International Classification of Diseases for Oncology, morphology codes 9380–9481) were recruited from the local population-based registry, the Northern California Rapid Case Ascertainment program and the University of California, San Francisco Neuro-oncology Clinic, between 1991 and 2009. Regarding pathology review, all 191 UCSF cases with oligodendroglial features were either diagnosed or reviewed by a UCSF neuropathologist or, for one patient whose materials were unavailable for review, the pathology report, for indicated deletion of 1p/19q. The 1p/19q deletion data were obtained from the medical record, or if these data were unavailable, by FISH analysis (
). For the other histologies, 454 of 520 (87%) GBM, 89 of 95 (94%) AA, and 56 of 58 (97%) grade 2 astrocytoma were either diagnosed or reviewed by a UCSF neuropathologist. Neuropathologists Kenneth Aldape and Tarik Tihan reviewed most cases used in this study—49 and 23%, respectively. AGS control subjects aged 20 years or older from the same residential area as cases were identified by means of random-digit dialing and were frequency matched to cases on the basis of age, sex, and ancestry. Consenting participants provided blood and/or buccal specimens and information during an in-person or telephone interview. UCSF subject characteristics are summarized in Supplemental Table 1.
Genotyping
Single nucleotide polymorphism (SNP) genotyping was performed on all Mayo Clinic cases and control subjects, 191 UCSF AGS cases with an oligodendroglial component, and 192 AGS control subjects by the custom GoldenGate Illumina genotyping platform (Illumina, San Diego, CA). GoldenGate assays were developed for all 13 SNPs of most interest from the prior two genome-wide association studies (
) as well as 89 nonredundant SNPs across the associated 8q24 region. Each GoldenGate assay batch included appropriate within-run and between-run control subjects as well as a large number (n = 86) of additional SNPs. Excluding occasional SNP failures, all within-run and between-run comparisons were completely concordant. We previously described SNP genotyping and quality control measures for UCSF GBM, AA, and astrocytoma grade 2 cases and 602 control subjects with the Illumina 370duo array panel (
Concordance in interplate, intraplate, and overall subject replicates were summarized to investigate potential genotyping error. Subject level call rates were calculated, and those subjects (12 of 1,497, 0.8%) with call rates of <0.9 were excluded from further analysis. Individual SNP call rates were summarized and SNPs (4 of 192, 2%), with call rates of <0.9 were excluded from the analysis. The minor allele frequency was calculated for each SNP, and SNPs (3 of 192, 1.6%) with a minor allele frequency of <0.01 were excluded from further analysis.
The above analyses were performed on the complete set of data, and each analysis was repeated separately for each plate to investigate any potential plate effects. The genotype frequency distribution for each SNP in the control subjects was tested against the Hardy–Weinberg equilibrium by a chi-square test. SNPs with Hardy–Weinberg equilibrium P-values of <0.001 were excluded from the analysis because these might be SNPs with possible genotyping errors. Identity by descent estimates were calculated for all pairs of individuals to assess the relationships among subjects, as completed in PLINK (
) from the SNPs on the custom Illumina array, with no apparent population stratification observed. Therefore, association analysis was completed using all subjects.
However, it should be noted that the SNPs genotyped on the Illumina array had limited information regarding ancestry. Assessment of each SNP with disease status was computed by logistic regression under a log-additive genetic model, with genotypes coded as 0, 1, or 2 copies of the minor allele. Analysis of the association of glioma risk was performed for UCSF and Mayo separately as well as for UCSF and Mayo together. Separate histology- and grade-specific stratified analyses comparing gliomas to control subjects were performed for the following glioma groups: grade 2 astrocytoma, AA, GBM, oligodendroglioma, mixed oligoastrocytoma, oligodendroglioma/mixed oligoastrocytoma, grade 2 oligodendroglioma/mixed oligoastrocytoma, grade 3 oligodendroglioma/mixed oligoastrocytoma, 1p/19q deleted tumors, and 1p/19q intact tumors. The 1p/19q deleted versus the 1p/19q intact case–case comparisons were also completed. These analyses included both age and sex as covariates in the logistic model. Assessment of genetic heterogeneity between sites was also assessed by including a site-by-SNP interaction term in the pooled analysis. No significant genetic heterogeneity in SNP effects was detected between the two sites. To adjust for multiple comparisons, a Bonferroni approach was used, with P-values of <4.17 × 10−3 considered statistically significant.
Ethics
These studies were approved by the Mayo Clinic Office for Human Research Protection and the University of California San Francisco Committee on Human Research. Informed consent was obtained from all study participants.
Results
Stratification of germ line risk alleles by oligodendroglial components
Of the five chromosomal regions examined, only 8q24 (CCDC26) was significantly associated with oligodendroglioma or oligoastrocytoma (Figure 1 and Supplemental Table 2). All three of the previously reported 8q24 SNPs were highly significant (e.g., for rs4295627, odds ratio [OR] = 2.05, 95% confidence interval = 1.65–2.54, P = 8.3 × 10−11). These associations were found in both case-control groups and in tumors with either oligodendroglioma or mixed oligoastrocytoma morphology (Supplemental Table 2). 8q24 polymorphisms were associated with oligodendroglial tumors, irrespective of grade or 1p/19q deletion status, although the effect size was highest for codeleted tumors (for rs4295627 OR = 2.77; 95% confidence interval = 1.98–3.88). GBM was not associated with 8q24 polymorphisms, whereas non-GBM astrocytoma showed an association that did not reach statistical significance in our sample (Figure 1 and Supplemental Table 3).
Figure 1Glioma risk stratified by chromosomal region and morphology type, oligodendroglioma grade and 1p/19q codeletion status. The SNP within each region that was most strongly associated is illustrated. For comparative purposes, some SNP-associated risks (and their 95% confidence intervals) have been inverted (Supplemental Tables 2 and 3). P-values are provided for case-control comparisons significant after correction for multiple testing. MOA, mixed oligoastrocytoma; Oligo, oligodendroglioma; GBM, glioblastoma; AA, anaplastic astrocytoma; A2, astrocytoma grade 2; codel, codeleted.
Figure 2 shows fine mapping of loci within the 8q24 region by tumor histologic type. Two peaks of association were noted for oligodendroglial tumors, but not GBM or lower-grade astrocytoma. Potential explanations for the distinct peaks include independent risk loci and/or a recombination event within the at-risk haplotype.
Figure 2Fine mapping of loci within the 8q24 region by tumor histologic type log10(p) values for 89 SNPs across the 8q24 region stratified by glioma morphologic subtype. Combined Mayo Clinic and UCSF data are shown. Mayo Clinic glioma cases (n = 582) and control subjects (n = 532), and UCSF oligodendroglioma (oligo) and mixed oligoastrocytoma (MOA) cases (n = 191) and control subjects (n = 192) were genotyped by the Illumina VeraCode platform. UCSF astrocytic glioma cases (n = 673) and control subjects (602) were evaluated by the Illumina 370duo platform. For these latter gliomas, only 28 SNPs are illustrated. Arrows indicate the SNPs reported by Shete et al.
Stratification of germ line risk alleles by astrocytoma grade
The other four glioma risk regions demonstrated less specificity than that seen for 8q24, although RTEL1 region (20q13) risk estimates were strongest for GBM (e.g., for rs2297440, OR = 0.56, 95% confidence interval = 0.47–0.68, P = 4.6 × 10−10) (Figure 1 and Supplemental Tables 2 and 3). Risk estimates were similar across glioma subtypes for the 5p15 region, although only GBM and AA associations were significant after adjusting for multiple testing. Although the 9p21 region was strongly associated with GBM and null for oligodendroglioma, risk intervals for all histologic types overlap. Although the 11q23 glioma risk SNP rs498872 was significantly associated with oligodendroglioma, similar risk estimates were observed for other non-GBM gliomas, although this was not statistically significant (Figure 1).
Discussion
We show that 8q24 polymorphisms were associated with gliomas containing an oligodendroglial component but not with GBM. Conversely, the 5p15, 9p21, and 20q13 regions were associated with GBM risk but were not strongly associated with oligodendroglial tumors. This pattern supports the current model of glioma initiation and progression. For example, because 8q24 associations are observed for risk of oligodendroglioma regardless of 1p/19q deletion, as well as potentially with non-GBM astrocytoma, we predict 8q24 SNPs may be associated with gliomas containing IDH1/2 mutations (
). Indeed, it is plausible to hypothesize that 8q24 alterations may facilitate the acquisition of IDH1/2 mutations or interact with an IDH1/2 mutation to facilitate tumorigenesis (
). Unfortunately, we had a small number of cases, and therefore statistical power was insufficient to generate meaningful data for grade 2 astrocytoma and secondary GBM.
The associated 8q24 region in glioma is proximal to the regions near MYC associated with the development of other cancers (
Clinical significance of alterations of chromosome 8 detected by fluorescence in situ hybridization analysis in pathologic organ-confined prostate cancer.
). It is unknown whether the other tumors with the 8q24 association are of different morphologic subtype or grade. Our glioma results suggest that 8q24 associations with risk in the other tumors may also be linked to lower grade and/or less aggressive behavior. The 8q24 region—including MYC—is frequently gained or amplified in many cancers (including gliomas); the amplicon often includes CCDC26 (
Clinical significance of alterations of chromosome 8 detected by fluorescence in situ hybridization analysis in pathologic organ-confined prostate cancer.
). It is not known whether the SNPs within 8q24 are associated with somatic gain of 8q24.
SNPs within or near the TERT (5p15), CDKN2A/B (9p21), and RTEL1 (20q13) regions were most strongly associated with GBMs. TERT is a component of the telomerase system, which maintains telomere length. RTEL1 has been hypothesized to be involved in the resolution of D loops that occur during homologous recombination (
). Other genes in the 20q13 region near RTEL1 may also predispose to aneusomy. STMN3, which facilitates tubulin depolymerization and is regulated during mitosis (
). Germ line alterations in STMN3 might affect tubulin binding and thus affect mitotic segregation. The CDKN2A/B region contains several tumor suppressor genes; deletions in this region are present in hereditary glioma and melanoma syndromes (
In summary, our report is novel in that we show that specific germ line alterations predispose to gliomas with different morphologic subtypes, histologic grades, and somatic genetic alterations. Further studies will be necessary to determine whether the known or additional genomic risk regions are associated with other morphologic or molecular genetic subgroups of glioma.
Acknowledgments
The research at Mayo Clinic was funded by NIH grants P50 CA108961 (the Mayo Clinic Brain Tumor SPORE) and RC1 NS068222. Work at University of California, San Francisco was supported by NIH grants R01 CA52689 and UCSF Brain Tumor SPORE, P50CA097257, as well as by grants from the National Brain Tumor Foundation, the UCSF Lewis Chair in Brain Tumor Research and by donations from families and friends of John Berardi, Helen Glaser and Elvera Olsen. We thank the University of Georgia Genome Center and the Mayo Clinic Genotyping Core for performing the custom SNP analyses. In addition to C. Gianinni, we thank B. Scheithauer for his careful histological review of all the gliomas collected at the Mayo Clinic. The San Francisco Adult Glioma Study thanks the Northern California Cancer Center for glioma patient case finding; we also thank Kenneth Aldape for pathology review and the pathology departments of Alexian, Alta Bates, Brookside, California Pacific, Drs. Pinole, Eden, El Camino, Good Samaritan, Highland, John Muir, Kaiser Redwood City, Kaiser San Francisco, Kaiser Santa Teresa, Los Gatos, Los Medanos, Marin General, Merrithew, Mills Peninsula, Mt. Diablo Hospital, Mt. Zion, Naval Hospital, O’Connor, Ralph K. Davies, Saint Louise, San Francisco General, San Jose, San Leandro, San Mateo County, San Ramon Valley, Santa Clara Valley, Sequoia, Seton, St. Francis, St. Luke’s, St. Rose, Stanford, Summit, UC San Francisco, Valley Livermore, Veterans Palo Alto, Veterans SF, and Washington Hospitals and Medical Centers for providing tumor specimens for review. We used HapMap data to generate the recombination frequencies surrounding the associated 8q24 region. We thank the International HapMap Consortium (
Clinical significance of alterations of chromosome 8 detected by fluorescence in situ hybridization analysis in pathologic organ-confined prostate cancer.
00557-1&id=gr1.jpg){kind=link}
00557-1&id=gr2.jpg){kind=link}