Genetic Overview of Astrocytomas (WHO grades II and III)

In order to plan the most effective strategy possible to manage your tumour, it is imperative to discover its genetic basis through laboratory testing of tumour tissue sample. The following discussion applies mainly to astrocytomas of WHO grades II and III (diffuse and anaplastic astrocytomas).

In recent years, genetic sequencing has increased the understanding of these tumours most significantly. In 2006 and 2010, detailed gene-expression and mutation studies of glioblastomas classified these tumours into three or four different subgroups: Proneural, Mesenchymal and Proliferative in the first study [1]; or Proneural, Mesenchymal, Classical and Neural in the second study [2].

In 2012, a similar analysis was carried out for lower-grade astrocytomas (WHO grades II and III). In this study, astrocytomas were classified into neuroblastic (NB), early progenitor-like (EPL) and pre-glioblastoma (PG) subtypes based on molecular and clinical characteristics [3].

A follow-up study later in 2012 by the same group of researchers at Memorial Sloan-Kettering Cancer Center identified a newly discovered mutation, ATRX, as being the defining genetic marker distinguishing early progenitor-like tumours from neuroblastic tumours [4].

A year after this, in October 2013, a thorough genetic analysis of seven diffuse astrocytomas (grade II) and 16 anaplastic astrocytomas (grade III) along with several oligoastrocytomas and glioblastomas was published in Oncotarget (online journal) by researchers from the Preston Robert Tisch Brain Tumor Center at Duke University [5].

In the words of the authors, this study marks the “largest exome sequencing study of progressive astrocytomas to date”. Exons are the sections of DNA within each gene which are eventually translated into protein and the exome is the totality of exons within the genome. Progressive astrocytoma in this context is referring to lower grade astrocytomas of WHO grades II and III. The significance of this study is that the coding region of every gene in the genome of these tumours was sequenced for mutations, giving an excellent overview of the “genetic landscape” of these tumours.

The chief finding was that only 3 mutations were common in grade II and III astrocytomas: IDH1, TP53, and ATRX.

The seven grade II diffuse astrocytomas contained an average of 13 somatic mutations, anaplastic astrocytomas an average of 36, and primary glioblastomas an average of 42. However for the lower grade astrocytomas, only the three mutations mentioned above (IDH1, TP53, ATRX) were commonly found.

Mutational landscape and clonal architecture in grade II and III gliomas

On April 13, 2015, the most comprehensive genetic analysis of lower grade gliomas (WHO grades II and III) was published online in the journal Nature Genetics by a team of Japanese researchers [13]. This study compiles comprehensive genetic analysis from both a large cohort of Japanese samples as well as samples from The Cancer Genome Atlas. The landscape of somatic mutations and copy number alterations in the entire cohort of over 700 patients is thoroughly investigated in this landmark publication. Low grade glioma specimens can be divided into three different types.

Type 1 – IDH mutant oligodendroglioma

Lower grade glioma type 1 is characterized by IDH1 (or IDH2) mutation and codeletion of one copy of chromosomes 1p and 19q. These two alterations define this group, and were found in 100% of 288 specimens. Other alterations found in the majority of cases in this group are TERT promoter mutations (98%), and CIC mutations on chromosome 19q (58%). Other common mutations in this group occur in the genes FUBP1 on chromosome 1p (31%), NOTCH1 (21%), PIK3CA (11%) and PIK3R1 (8%), and in the chromatin remodelling genes ARID1A (7.3%) and ARID1B (5.2%). Potentially druggable pathways related to these genes are the PI3K pathway (PIK3CA) and the NOTCH pathway (NOTCH1). Drugs inhibiting these pathways are currently under development.

Type 2 – IDH mutant astrocytoma

Lower grade glioma type 2 is characterized by IDH1 mutation, TP53 mutation, and ATRX mutation, and corresponds to IDH1-mutant astrocytoma. In this group of 313 samples, all carried somatic mutations in IDH1 or IDH2, and 99% of these also carried mutations in the TP53 gene. 77% carried mutations in the ATRX gene. These three mutated genes were by far the most common alterations in this group. The next most commonly mutated gene was SMARCA4, involved in chromatin remodelling with a 6% mutation rate. No other gene in this group was mutated in over 5% of cases.

Type 3 – IDH non-mutant lower grade gliomas

This type carries genetic alterations that are characteristic of glioblastoma, and are not IDH mutated. 156 glioma samples were placed in this group, with the most common somatic mutations involving the TERT promoter (33% of cases) and EGFR (22%). Other common mutations in this group are also characteristic of glioblastoma, such as TP53 (18%), PTEN and NF1 (both mutated in about 15% of cases).

2015.4 Suzuki Supp image 5

Landscape of somatic mutations and copy number alterations in the three basic types of lower grade glioma. Combined data from Japanese cohort and the Cancer Genome Atlas, n=757.

The importance of copy number-neutral loss of heterozygosity of chromosome 17p in type 2 gliomas (IDH-mutant astrocytoma)

Multiple prior studies have clarified the importance of IDH mutation status to the prognosis of lower grade gliomas. A new study by a French team, published in the September 2016 edition of The Oncologist, demonstrates that a further genetic subdivision of IDH-mutant astrocytomas may help strengthen prognostic ability even further [14]. This new genetic marker is copy number-neutral loss of heterozygosity of chromosome 17p, which contains the TP53 gene.

Copy number-neutral loss of heterozygosity” simply means that one copy of the chromosomal section has been deleted while the remaining copy has been duplicated. The net result is that the cell still has a total of two copies of the gene or chromosomal segment, but instead of having two different copies (one from the mother and one from the father), a single copy has been duplicated.

In this study, copy number-neutral loss of heterozygosity of chromosome 17p (CNLOH 17p) was nearly exclusively associated with IDH1-mutant astrocytomas with TP53 mutations. The TP53 gene is located on chromosome 17p, and all cases of CNLOH 17p included the TP53 locus. The result of CNLOH 17p in these cases was the deletion of the normal copy of TP53, and the duplication of the mutated copy of TP53. CNLOH 17p was detected in 56% – 58% of IDH-mutant astrocytomas.

Surprisingly, CNLOH 17p was found to be a significant prognostic factor, with better survival outcomes for those with the CNLOH 17p alteration. When “type 2” gliomas, that is, IDH-mutant astrocytomas without 1p/19q co-deletion, were subdivided into those with or without CNLOH 17p, the group with CNLOH 17p was found to have improved survival, comparable to “type 1” gliomas (1p/19q codeleted oligodendrogliomas).


Group 1 refers to IDH-mutant, 1p/19q codeleted oligodendrogliomas. Group 2 refers to IDH-mutant, 1p/19q non-codeleted astrocytomas. Group 2 has been subdivided into those with or without copy number-neutral loss of heterozygosity of chromosome 17p. Group 3 refers to IDH non-mutant gliomas.

A separate analysis of 142 lower grade gliomas from The Cancer Genome Atlas confirmed the improved prognosis of IDH-mutant astrocytomas with CNLOH 17p.

This study does not attempt to answer the question why IDH-mutant astrocytomas with copy number-neutral LOH of chromosome 17p (two copies of mutant p53, and deletion of wild-type p53) have improved survival versus those without CNLOH 17p. It may be that the improved survival in this subgroup depends on conventional treatments such as radio- or chemotherapy.

Farewell to “oligoastrocytoma”

In August 2014, a German study [12] was published online in Acta Neuropathologica which calls into question the use of the term “oligoastrocytoma” as a distinct entity. This research was based on 43 IDH1-mutant oligoastrocytoma samples from Muenster, Tuebingen, and Heidelberg Germany, diagnosed between 1989 and 2013.

In 30 out of 43 cases (70%), IDH1 R132H neoplastic cells made up the oligodendroglial component of the tumour, and also showed evidence of 1p/19q chromosomal co-deletion. The astrocytic component of these tumours lacked IDH1 mutations or any other alteration associated with neoplastic transformation such as nuclear p53 accumulation, loss of ATRX protein, or 1p/19q codeletion. These areas were considered to be areas of reactive astrogliosis, misidentified as areas of astrocytoma. In other words, these 30 cases were pure oligodendrogliomas with reactive astrocytes mistaken for astrocytoma cells.

In 12 of 13 of the remaining cases, or 28% of the 43 cases, the oligodendroglial and astrocytic components of each tumour shared identical molecular and genetic features. 11 of these cases exhibited ATRX loss and p53 accumulation in both the astrocytic and oligodendroglial areas of the tumour. These tumours are more properly classified as astrocytomas. The 12th case had co-deletion of 1p/19q in both the oligodendroglial and astrocytic areas, without loss of ATRX or p53 accumulation. This case was re-interpreted as a true oligodendroglioma. These 12 cases represent common clonal origin of the entire tumour and fall mostly into the astrocytoma category, with one oligodendroglioma, when judged on a molecular rather than morphological basis.

The final case had molecular features of both astrocytoma and oligodendroglioma, including 1p/19q loss in all, ATRX loss in all, and nuclear p53 in nearly all tumour cells. This tumour was a recurrence and the only tumour of the 43 samples known to have been previously irradiated. The initial tumour had no loss of ATRX and no p53 accumulation. This tumour sample was genetically sequenced, and no TP53 or ATRX mutation was found. This tumour also did not manifest complete deletion, but only partial deletion on chromosomes 1p and 19q, unlike pure oligodendrogliomas. It is possible that the later accumulation of p53 protein and loss of ATRX expression in this case was a result of radiation treatment. Of all 43 cases, this was the only case which could not be definitively ascribed to either the astrocytoma or oligodendroglioma categories on the basis of molecular features.

The authors conclude that “the molecular findings and the clinical study data strongly argue against OA.” In other words, “oligoastrocytoma” as a distinct entity either does not exist, or is very rare. These cases can for the most part be re-interpreted as either astrocytomas or oligodendrogliomas based on the presence of molecular markers such as 1p/19q co-deletion (in oligodendrogliomas); or absence of ATRX protein (or ATRX mutation) and presence of nuclear p53 accumulation (or TP53 mutation) in the case of astrocytoma. These authors propose to refrain from a diagnosis of “oligoastrocytoma” and instead use IDH1 mutation, 1p/19q status as determined by FISH or other methods, and ATRX protein expression to make a diagnosis of either astrocytoma or oligodendroglioma. p53 nuclear accumulation does not necessarily signify TP53 mutation and may instead be a sign of reactive gliosis. Therefore loss of ATRX expression may be the more reliable marker of astrocytoma. Especially, reactive gliosis must be considered and must not be misdiagnosed as an astrocytoma component of an oligodendroglial tumour. In the roughly 1% of cases which display molecular features of both astrocytoma and oligodendroglioma, the term oligoastrocytoma may provisionally be used.

Three recommended tests for lower grade astrocytomas

In view of the contrasting genetic, epigenetic and clinical characteristics which divide IDH-mutant from IDH-wild type tumours, it is highly advisable that any patient with a grade II or III astrocytoma have their tumour tissue tested for IDH1 mutation with immunohistochemical methods, or IDH1 and IDH2 mutations by DNA sequencing. IDH mutation is the most important diagnostic marker, which defines the tumour genetically.

TP53 mutations usually co-occur with IDH1 mutations in astrocytoma and may predict response to certain chemotherapeutic agents (such as nitrosoureas and irinotecan) [7, 8]. The most common method to test p53 status is through immunohistochemical staining for p53 protein accumulation in the cell nuclei, though this is only an indicator, and not concrete proof of mutation in the TP53 gene. A study appearing online in March 2014 in the Journal of Neuro-Oncology [10] found that the best correlation between positive nuclear staining for the p53 protein and a mutated TP53 gene in low grade gliomas occured when at least 10% of cells were positive for p53 (a labeling index of 10%). In other words, a potential sign of TP53 mutation is presence of p53 protein found in around 10% or more of cells in a tissue sample. In this study neither p53 positivity nor TP53 mutation were found to be prognostic for survival.

Mutation of the ATRX gene typically leads to a phenomenon called Alternative Lengthening of Telomeres (ALT) which allows the cells to divide indefinitely, unlike normal cells which may only undergo a finite number of cell divisions [9]. Immunohistochemical testing for the presence or absence of ATRX is perhaps a more reliable method than p53 to establish an IDH-mutant tumour as being an astrocytoma, rather than oligodendroglioma.

If there is uncertainty as to the precise diagnosis, or if given the diagnosis of “oligoastrocytoma”, co-deletion of one entire copy of chromosome 1p and 19q indicates oligodendroglioma, while absence of ATRX protein and extensive presence of p53 protein indicates astrocytoma.

The importance of MGMT testing

MGMT (O6-methylguanine DNA methyltransferase) is a DNA repair enzyme which specifically removes alkyl groups (such as methyl groups) from the O6 position of guanine within the DNA. Agents such as temozolomide and chloroethylating nitrosoureas create DNA damage most potently by adding methyl groups (in the case of TMZ) or chloroethyl groups (in the case of nitrosoureas) to the O6 position of guanine. Therefore, a tumour with a high degree of MGMT activity will be resistant to chemotherapies which target DNA at this location. If the promoter region of the MGMT gene is unmethylated, the gene will be potentially active, whereas if the promoter region of MGMT is hypermethylated, the gene will be silenced. Numerous clinical studies have shown that patients with a hypermethylated MGMT promoter have increased survival following alkylator chemotherapy, while those with an unmethylated MGMT promoter show limited benefit from these agents. The cost of the test is in the $500-$750 range. A simpler, more routine MGMT test is immunohistochemical staining for the presence of the MGMT enzyme, though this test is not considered to be as reliable as promoter methylation status in terms of clinical outcome.

A 2011 study using sensitive pyrosequencing of the MGMT gene revealed that virtually all (127 of 129 or 98%) of IDH1/2 mutant gliomas had significant methylation of a selection of 16 CpG sites most critical for determining MGMT protein expression. This study will be discussed further on the Closer Look at IDH Mutations page.

Comprehensive and Integrative Genomic Characterization of Diffuse Lower Grade (2 & 3) Gliomas – Daniel Brat

  1. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Phillips et al. 2006.
  2. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Verhaak et al. 2010.
  3. IDH mutation and neuroglial developmental features define clinically distinct subclasses of lower grade diffuse astrocytic glioma. Gorovets et al. 2012.
  4. Whole-exome sequencing identifies ATRX mutation as a key molecular determinant in lower-grade glioma. Kannan et al. 2012.
  5. The genetic landscape of anaplastic astrocytoma. Killela et al. 2013.
  6. Upregulation of ASCL1 and inhibition of Notch signaling pathway characterize progressive astrocytoma. Somasundaram et al. 2005.
  7. Differential sensitivity of malignant glioma cells to methylating and chloroethylating anticancer drugs: p53 determines the switch by regulating xpc, ddb2, and DNA double-strand breaks. Batista et al. 2007.
  8. p53 disruption profoundly alters the response of human glioblastoma cells to DNA topoisomerase I inhibition. Wang et al. 2004.
  9. Loss of ATRX, genome instability, and an altered DNA damage response are hallmarks of the alternative lengthening of telomeres pathway. Lovejoy et al. 2012.
  10. TP53 and p53 statuses and their clinical impact in diffuse low grade gliomas. Gillet et al. 2014.
    READ ABSTRACT (Email me for a PDF copy)
  11. Frequent ATRX, CIC, FUBP1 and IDH1 mutations refine the classification of malignant gliomas. Jiao et al. 2012.
  12. Farewell to oligoastrocytoma: in situ molecular genetics favor classification as either oligodendroglioma or astrocytoma. Sahm et al. 2014.
    READ ABSTRACT Email me for a PDF copy
  13. Mutational landscape and clonal architecture in grade II and III gliomas. Suzuki et al. 2015.
    READ ABSTRACT Email me for a PDF copy
  14. Chromosome 17p Homodisomy Is Associated With Better Outcome in 1p19q Non-Codeleted and IDH-Mutated Gliomas. Labussiere et al. 2016.
    READ ABSTRACT Email me for a PDF copy