Reviews in Oncology Reviews in Oncology CIC Edizioni Internationali 2013 January-March; 1(1): 15–20. ISSN: 2282-6378

Oligodendroglial tumors: from biology to a better patient care

Patrizia Farina,1, 2 Giuseppe Lombardi,1, 2 Marta Rossetto,1, 3 Agusti Alentorn,1, 4 and Marc Sanson1, 4

1Université Pierre et Marie Curie-Paris 6, Centre de Recherche de l’Institut du Cerveau et de la Moelle Epinière (CRICM), Paris, France
2Medical Oncology, Oncology Institute of Veneto-IRCCS, Padova, Italy
3Neurosurgical Clinic, Universy of Padova, Padova, Italy
4AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Service de Neurologie 2-Mazarin, Paris, France

WHO classification relies on similarities between tumor cells and normal glial cells to identify astrocytomas, oligodendrogliomas, and mixed oligoastrocytomas. It is highly subjective and poorly reproducible, leading a high inter-observer variability (1). The importance of correct identification of oligodendroglioma in clinical practice was raised twenty years ago with the observation that some oligodendrogliomas, unlike common astrocytic gliomas, were particularly sensitive to chemotherapy (2). The chemosensitivity has been then associated to the presence of the combined loss of chromosomes 1p and 19q (3). Based on the CB-TRUS studies, oligodendrogliomas account for approximately 2% of all primary brain tumors and 6% of gliomas (4). In other series, oligodendrogliomas represent up to 20% of all gliomas. Oligodendroglioma is characterized by monomorphic cells with uniform round nuclei and perinuclear halos with honeycomb appearance. In addition significant mitotic activity, prominent microvascular proliferation or necrosis defines anaplastic oligodendroglioma (5). In clinical practice the diagnosis of oligodendroglioma remains challenging. The clinical behavior of oligodendroglioma is heterogeneous within the same pathological WHO grade, and particularly in grade III, and this is in part explained by the molecular profile.

Indeed, over the last years, major efforts have been accomplished to discover the genes driving oligodendroglial oncogenesis and to identify relevant biomarkers for clinical use. The first milestone in genetics of oligodendrogliomas was the detection of recurrent co-deletion of chromosome regions 1p and 19q. This chromosome imbalance, observed in ∼2/3 of pure oligodendrogliomas and 10–15% of mixed oligoastrocytomas, is associated with better outcome and better response to chemotherapy and should now be part of the pre therapeutic screening. More recently, pivotal studies using high throughput molecular biology technics have identified novel mutated genes with clinical and biological perspectives.

In the nineties, microsatellites analysis identified recurrent losses of heterozygosity (LOH) in chromosome regions 1p36 and 19q13 (6). This signature was associated with frontal location and classic oligodendroglial morphology (7). Early retrospective studies suggested that this codeletion of chromosome 1p and 19q, was the response to chemotherapy and longer progression-free survival in anaplastic oligodendrogliomas (3). It was then shown in low grade gliomas that 1p19q codeletion has a different natural history with slower growth rate (8). In fact until recently, it was still unclear whether 1p19q loss was predictive for chemotherapy or merely indicates a different natural history: two trials from the Radiation Therapy Oncology Group (RTOG 9402) and European Organisation for Research and Treatment (EORTC 26951) shows clearly near-doubling of median survival times (14.7y vs 7.3y in the RTOG study) of patients with 1p19q codeleted grade III gliomas treated with chemotherapy and radiation therapy (RT) vs RT alone, whereas patient without codeletion have a poor survival (2.6–2.7 y) and no significant benefit of adjuvant chemotherapy (9, 10).

The identification of the 1p19q codeletion is therefore critical for the clinician, but may be challenging and requires an adequate technique. Indeed the “true” 1p19q codeletion [i.e. with centromeric breackpoint, as identified by comparative genomic hybridization (CGHa)] must be distinguished from partial deletions of 1p which can be associated with 19q loss and are associated with poorer outcome (11). The 1p19q codeleted gliomas over express proneural genes (12): one of them is INA which encodes alpha-internexin (INA), a class IV neuronal intermediate filament. Immunohistochemical INA analysis can be used routinely, is useful to predict 1p19q codeletion and is a simple and valuable prognostic and predictive factor for adjuvant chemotherapy (13,14). INA expression was found to be a strong pronostic factor in patients with anaplastic oligodendroglial tumors enrolled in the EORTC trial 26951 (15).

This 1p19q co-deletion with centromeric breakpoints (11) was highly suggestive of an imbalanced reciprocal translocation, t (1; 19) (q10; p10), that has been proven by fluorescent in situ hybridization (FISH) analysis (16). No chimeric gene created by this chromosome fusion has been identified to date (17). In addition these cells are extremely difficult to grow in culture (18), and therefore relevant experimental models are missing.

EGFR amplification and chromosome arm 21q deletion are mutually exclusive with 1p19q codeletion and predictive of poor prognosis. EGFR amplification is associated with high rate of necrosis, older age of patients (19). In grade II and III oligodendrogliomas chromosome arm 10q loss is predictive of poor prognosis in terms of both progression free and overall survival (20), and in the same line, the looses of chromosome arms 9p and 10q have been associated with poor outcome in WHO grade II oligodendrogliomas, irrespective of 1p/19q status (21).

Over the recent last years, next generation sequencing technologies identified novel mutated genes in gliomas, and particularly in oligodendrogliomas with 1p19q codeletion.

The IDH1 mutation has been identified by high throughput sequencing of glioblastoma (22). In fact IDH1 mutation, which affect 40% of gliomas, is found in only 5–8% of primary glioblastoma, vs 75% of grade II and 50% of grade III gliomas (23). IDH1 encodes the cytoplasmic isoform of isocitrate dehydrogenase. Mutation affects the residue 132 of the IDH1 gene, the majority (>90%) being a CGT➝CAT change, leading to an Arg132➝His substitution. Some patients without IDH1 mutations harbor a mutation in the analogous amino acid residue (Arg172) of the mitochondrial isoform IDH2(24). Strikingly IDH1/IDH2 mutations are tightly associated with genetic profile, and are a constant feature in 1p19q codeleted gliomas (25), suggesting a potential link between both alterations. IDH1 R132H detection is now considerably facilitated by the development of two monoclonal antibodies specifically targeted against the IDH1 R132H mutation (26, 27).

IDH1/IDH2 mutated tumors have a better outcome, whatever grade considered (23, 28, 29). Whether IDH1/IDH2 mutation can predict response to treatment in gliomas needs to be further investigated (30).

The mutation causes the loss of the isocitrate dehydrogenase function and the gain of an α-ketoglutarate reductase function leading to the cellular accumulation of D-2-hydroxyglutarate (D-2HG) (31). The rate of D-2HG in IDH mutated tumors is increased by a factor >100, thus representing a diagnostic marker (this change is almost specific for gliomas) and prognostic (mutated gliomas have longer survival) of interest. Interestingly, D-2HG can be detected using proton magnetic resonance spectroscopy (MRS) (32, 33).

The accumulation of D-2HG results in a profound modulation of epigenome (genomic DNA methylation resulting in CIMP= CpG Island Methylated Phenotype and histone methylation), gene expression and inhibition of terminal differentiation (34, 35). As a consequence of diffuse CpG methylation, IDH1/2 mutated tumors, including all the 1p19q codeleted oligodendrogliomas, are tightly associated with methylation of MGMT promoter (and consequently gene silencing) (35, 36). MGMT gene encodes the Methyl-Guanyl-Methyl-Transferase enzyme which removes the methyl from the O6 residue of guanine and is therefore involved in the chemoresistance to alkylant agents, such as temozolomide or nitrosoureas. In the same line, a genome-wide methylation profiling study of oligodendroglial WHO grade III tumors from the EORTC study 26951 clinical trial, showed that: (i) CpG island hypermethylation phenotype (CIMP+) is associated with MGMT promoter methylation, 1p/19q codeletion, IDH mutation and (ii) CIMP+ tumors have better prognosis than CIMP- and (iii) CIMP+ is an independent prognostic factor in multivariate analysis. CIMP+ and IDH mutation correlation is in agreement with the functional link mentioned above (36). In anaplastic oligodendrogliomas, MGMT promoter methylation seems to be a part of the broader CIMP+, that explain that MGMT methylated tumors, whatever their treatment, have a better outcome. Similarly, recent studies have demonstrated that IDH mutated glioma have singular proteomic and metabolomics patterns (37, 38).

Very recently mutations involving TERT promoter (C228T and C250T, corresponding to the positions 124 and 146 bp upstream of the TERT ATG start site) have been reported in gliomas, involving 80% of glioblastoma and 100% of oligodendrogliomas with 1p19q codeletion (42). TERT encodes the telomerase reverse transcriptase involved in the maintenance of telomere length. In the absence of telomerase activity, telomeres shorten with each cell division. Unlimited cancer cells division requires some telomere maintenance mechanism to avoid cell death or senescence. It is particularly interesting to note that TERT mutation were mutually exclusive with gliomas with ATRX mutation: ATRX inactivation is involved in another alternative mechanism of telomere lenghtening (ALT).

We can define three subgroups of oligodendrogliomas/oligoastrocytomas, based on the genetic profile: (i) 1p/19q codeleted (most of them are CIC mutated and all are IDH and TERT promoter mutated), (ii) non 1p19q co-deleted and IDH mutated (which have the ATRX mutation instead of TERT promoter mutation) and (iii) double negative. These subgroups have distinct molecular patterns (epigenetic, transcriptomic, proteomic and metabolomics), distinct natural history and response to alkylating chemotherapy.

Oligodendrogliomas with 1p/19q co-deletion and IDH mutation (group 1) are characterized by better prognosis and better response to chemotherapy. They are associated with cerebral frontal location, classic morphology, absence of gene high-level amplification. Recently, it has validated through phase III clinical trials that 1p/19q status is critical for medical management of anaplastic oligodendroglioma patients (see below). The interactions between the key genetic alterations of this tumor subgroup (particularly 1p/19q codeletion, IDH mutation, CIC mutation, and TERT promoter mutation) need to be elucidated particularly in oligodendrocyte progenitor that been recently suggested as the cell of origin of oligodendroglioma. Novel sequencing technologies such as RNA-Seq and development of tumor cell lines and animal models harboring this abnormality will probably be helpful to dissect molecular oncogenesis of 1p/19q co-deleted oligodendrogliomas.

IDH mutated gliomas without 1p/19q codeletion (group 2), have an intermediate prognosis (poorer than 1p19q codeleted, but better than the non IDH mutated counterparts) (25), and include mostly astrocytomas, and occasionaly oligodendrogliomas and oligoastrocytomas. Mutation of P53 and ATRX are frequent (43). ATRX is part of a chromatin remodeling complex involved in telomere biology. Mutations of ATRX cause alternative lengthening of telomeres (ALT) and are mutually exclusive with TERT promoter mutations (42).

IDH wild type gliomas (and consequently non 1p/19q co-deleted-group 3) have poor prognosis and are mostly represented by mixed oligoastrocytomas. In grade II, group 3 is nearly similar to the “triple negative” recently identified subgroup of low grade glioma characterized by absence of 1p/19q codeletion, TP53 expression and IDH mutation (44) (1p19q deletion is virtually absent in 1p19q, and TP53 expression, mostly found in IDH mutated-non codeleted tumors, is infrequent in IDH wild-type tumors). Most of these LGG have oligodendroglial or mixed phenotype. In addition, they are frequently located in the insula, larger in size and associated with poor prognosis. These triple negative WHO grade II tumors have probably distinct oncogenic mechanism compared to the majority of oligodendrogliomas exhibiting IDH mutation (45).

Several evidences suggest that human oligodendroglioma have a white matter origin, and derive from OPCs (Oligodendrocytes Progenitor Cells) rather than Neural stem Cells (46). Tumors with intact 1p/19q are more frequent in the temporoinsular regions whereas 1p/19q codeleted tumors are more frequently located in the frontal lobe (7). Low grade oligodendroglioma with 1p/19q codeletion have higher regional cerebral blood volume (rCBV) and blood flow (rCBF) but a lower apparent diffusion coefficient (ADC) than their non codeleted counterparts (47, 48). Quantitative MR texture may also help to differenciate 1p19q codeleted and non codeleted tumors (49). However, the most promising approach today is the possibility to detect the D-2HG produced by IDH mutated tumors by spectro-MRI (32, 33). In the near future this feature may be useful for non-invasive diagnosis of IDH mutated gliomas and oligodendrogliomas.

The optimal treatment strategy for anaplastic oligodendroglioma tumor is evolving.

Low grade oligodendroglial tumors are slow-growing tumors, and share with astrocytic and mixed LGG invasive and malignant potential. Gross total surgical resection, whenever possible, is recommended. Radiotherapy is considered as a postoperative standard treatment for LGG, but the optimal timing of this treatment (i.e., immediate vs at progression) is still discussed. There are now consistent evidences suggesting that up-front chemotherapy, with temozolomide or PCV regimen, is a valuable alternative to radiotherapy, in non resectable and symptomatic LGG (50, 51). Moreover, the 1p19q codeleted oligodendrogliomas have a slower growing rate, a higher response rate to chemotherapy, and more sustained response (8,51). The recently completed EORTC 22033–26033 trial that randomises at progression temozolomide vs radiotherapy in patients with low grade gliomas will help to determine the best option, according to the genomic profile (ie 1p19q codeletion, IDH mutation). The treatment of anaplastic gliomas is based on post-operative radiotherapy. The EORTC Brain Tumor Group initiated in 1995 a prospective randomized phase III trial (EORTC study 26951) to determine whether adjuvant PCV given after 59.4 Gy of radiotherapy (RT) in fractions of 1.8 Gy would improve survival.

After a median follow-up of 7 years, the results showed an increase in progression-free survival (PFS) in adjuvant PCV-treated patients, but no statistically significant increase in overall survival (OS), whatever the molecular profile (52). A similar North American study (RTOG 9402) in which PCV chemotherapy was given with an intensified PCV regimen before 59.4 Gy of RT reached similar conclusions (53). In both studies median OS was not reached for the 1p/19q codeletion.

However six years later updated analysis with a median follow-up of 12 years demonstrated that patients with codeleted 1p19q grade III oligodendrogliomas, had a significant improvement in overall survival when treated with early chemotherapy with radiation compared with early radiation, and salvage chemotherapy at tumor relapse, and thus establishes the 1p19q codeletion as a predictive marker (9, 10). Radiotherapy alone should no longer be considered an adequate treatment for this patient population. However, there are still unanswered questions regarding the quality of life of this long term survivor category: should upfront chemotherapy, omitting/deferring radiotherapy, be the initial therapy for oligodendroglial tumors with codeleted 1p19q, in order to avoid late neurocognitive toxicity? Can temozolomide, an oral agent with a better toxicity profile, be substituted for PCV? Indeed randomized German trial (NOA-04) suggested that up-front radiotherapy or chemotherapy (with either temozolomide or PCV) achieved comparable results in patients with anaplastic gliomas (29).

In cases without 1p19q deletion (group 2 and 3), most neuro-oncologists still associate chemotherapy to radiotherapy into the upfront strategy: 1) in group 2 the RTOG 9402 data suggest a benefit of upfront PCV in IDH mutated, non codeleted anaplastic oligodendrogliomas, and 2) several neuro-oncologists treat the non mutated IDH grade III tumors (group 3) with concomitant and adjuvant temozolomide, considering their survival range similar to true glioblastomas. This question will be in fact answered by the ongoing EORTC 26053–22054 (CATNON) that evaluates the benefit of concomitant and/or adjuvant chemotherapy to fractionated radiotherapy in grade III non 1p19q codeleted gliomas, stratified according to MGMT promoter methylation and IDH1 mutated status.

The molecular profiling has revolutionized histo-molecular classification of oligodendroglial tumors, and identified at least three distinct prognostic profiles. Secondly, 1p19q codeletion in anaplastic grade III oligodendrogliomas and mixed oligoastrocytomas should be now considered not only as a prognostic but, most importantly, predictive factor of response to adjuvant PCV chemotherapy. Based on these results, medical management of patients with anaplastic gliomas has changed and requires now assessment of 1p/19q status to deliver the right treatment to 1p/19q codeleted anaplastic oligodendrogliomas patients (i.e. radiotherapy plus PCV). The parternship between IDH, CIC, FUBP1 and TERT mutations and 1p/19q co-deletion in oligodendroglioma genesis remains to be elucidated. These molecular alterations might be candidates to new molecular targeted agents. Further studies are warranted and the increasing understanding of molecular pathways involved may lead to more selective therapeutic targets in the future.