Molecular predictive and prognostic factors in ependymoma
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.177630
Source of Support: None, Conflict of Interest: None
An ependymoma is an uncommon glial tumor, which arises from different parts of the neuroaxis. Considerable variation in presentation and survival in tumors in different locations after an optimum treatment indicates inherent molecular and genetic differences in tumorigenesis between them. A number of genetic aberrations have been identified to distinctly characterize different subgroups of ependymomas that include a posterior fossa tumor, a supratentorial tumor, and a pediatric tumor. These different groups have substantial genetic alterations, and also distinct demography, clinical characteristics, and prognosis. This article is intended to review the diverse molecular and genetic aberrations that may be helpful in prognostication and prediction of survival in patients suffering from an ependymoma.
Keywords: Ependymoma; genetic; molecular; prognosis
Ependymomas are relatively rare glial tumors arising from the ependymal cells in the brain and constitute around 5% of the central nervous system (CNS) tumors. They are significantly more common in male than in female patients. Ependymomas are broadly divided into three histological grades, with myxopapillary ependymomas and subependymomas being classified as Grade I tumors; the tanycytic, cellular, papillary, and clear cell subtypes being considered as Grade II tumors; and, the anaplastic variant being considered as a Grade III tumor. Unlike other glial tumors, the histological grade of the tumor has a conflicting significance as a prognostic and predictive marker., In addition to the histological grade, the site of the tumor, and age at initial diagnosis have been the important prognostic factors for survival in patients with an ependymoma. In the recent years, the importance of tumor location has emerged as one of the most important determinants of outcome. Ependymomas are now classified as supratentorial ependymomas, posterior fossa ependymomas, and spinal ependymomas. This classification with a distinct molecular association has brought an in-depth insight into the prognosis of different types of ependymoma. A tailored treatment protocol and incorporation of futuristic molecular targeted therapy for optimizing outcome has also emerged. Surgery is considered as the definitive curative treatment in ependymomas, and an attempt towards gross total excision is desirable to optimize outcomes. Consensus on this treatment regimen, however, is lacking.,, At present, the prognosis and adjuvant treatments are based solely on the pathological grading and extent of surgery. With local recurrence being the predominant mode of failure, local adjuvant radiotherapy is often added with the expectation of an improved outcome.,, The role of adjuvant chemotherapy is not clear except in pediatric cases where it is aimed to defer radiotherapy as a modality of treatment., In the recent years, great enthusiasm has been noted in exploring the molecular predictive and prognostic factors which may be useful in further tailoring treatment. In this review, we intend to review the various new prognostic and predictive markers that are helpful in the management of ependymomas.
The use of MIB-1 labeling indices along with the grade of the tumor may help in better predicting the nature of disease and the chances of recurrence in ependymomas. Prayson reported a MIB-1 labeling index of 4 or more was associated with more aggressive tumors with more chances of recurrence. Bennetto et al., evaluated Ki-67 labeling index in 74 pediatric patients with an ependymoma and found it to be a strong predictor for survival. However, there are also reports that although Ki-67 correlates well with the grade of the tumor, it may not be of prognostic significance in ependymomas.
The most common genetic abnormality seen in ependymomas includes 6q loss seen in about 60%, and 5p gain seen in about 40% of the cases. Huang et al., reported a loss of heterozygosity (LOH) of chromosome 6 in about 30%, and LOH of chromosome 9 in 27% of these patients. The authors hypothesized that important tumor-suppressor genes in ependymomas may be seen on chromosomes 6 and 9.
Zheng et al., reported loss of genetic material in chromosome 22q may be seen in 71%, and that in chromosome 16 may be seen 57% of the patients with an ependymoma. Less common is the loss of material in chromosome 17 seen in 46%, chromosome 6 seen in 39%, chromosome 19q seen in 32%, and chromosome 1p seen in 29% patients. Allelic loss of chromosome 22q may be seen in about 36–70% of patients with an ependymoma and is one of the commonnest abnormalities seen in sporadic ependymomas.,, It was also seen that the allelic losses on 22q were seen more frequently in intracranial anaplastic ependymomas in children and in intraspinal ependymomas in adults. Ebert et al., also concluded from his study of 62 samples of an ependymoma that spinal ependymomas are more often associated with molecular events involving chromosome 22. This genetic abnormality may, therefore, play an important role in the genesis of spinal ependymomas. Similarly, a gain on chromosome 7 is an almost exclusive feature of spinal cord ependymomas. They also postulated that the better prognosis seen in spinal ependymomas may be due to the existing differences in the genetic abnormalities in spinal and intracerebral ependymomas. LOH of the chromosome 11q may also be seen in patients with an ependymoma and is associated with a significant inverse correlation with LOH of 22q.
Gain of 1q, one of the most common molecular abnormalities seen in human cancers, is also seen in ependymomas. It is associated with a poor prognosis in many malignancies such as a neuroblastoma and Wilms' tumor. There are also reports that gain of 1q may be seen in about 20% of the patients harboring an ependymoma and is associated with a poorer prognosis when associated with these tumors. A chromosomal gain of 1q may be seen more frequently in pediatric patients, patients with intracranial lesions, and in Grade III ependymomas. It was also seen that a gain of 1q was associated with a higher chance of tumor recurrence in intracranial ependymomas.
Amplification of epidermal growth factor receptor (EGFR) located on chromosome 7p11.2, may also be seen in patients with an ependymoma. It has been seen that it is a poor prognostic factor in intracranial ependymomas as observed in the case of other glial tumors. However, it may not serve as a prognostic marker in intraspinal tumors. The epidermal growth factor receptor (EGFR) may be used as a therapeutic target in future trials. Similarly, studies have shown that tenascin, vascular endothelial growth factor (VEGF), and EGFR expression may be seen more often in a high-grade tumor, and may also serve as significant factors predicting progression-free survival.,,,
p53 tumor-suppressor gene is the most commonly observed mutation in human malignancies. Its role in the pathogenesis of astrocytic glial tumors, including glioblastomas, is proven. But, various studies have shown its doubtful significance in the pathogenesis or progression of human ependymomas., Zamecnik et al., evaluated 31 patients of pediatric intracranial ependymomas to study the prognostic factors in ependymomas and found that p53 mutation and bcl-2 positivity were associated with a poorer prognosis in ependymoma patients. Rushing et al., also reported that the presence of a higher MIB-1 labeling index or p53 immunolabeling may be an indicator for the presence of a higher grade of tumor, in patients harboring an ependymoma, even when the routine histologic criteria does not reveal a high-grade ependymoma. It may, therefore, be useful detect these markers as they are associated with a poorer prognosis and signify the presence of a higher grade of ependymomas.
Similarly, unlike other gliomas, retinoblastoma (RB) gene alterations and p16 deletions are seen in <25 % cases of ependymoma and studies to correlate it with tumor grade, recurrence, or death have been shown to be negative. It may also be noted that RB gene alterations and p16 deletions may not be playing a prominent role in the malignant progression of ependymomas. Survivin is a gene that is important in the development of the normal mammalian embryo. Preusser et al., evaluated 63 patients with an intracranial ependymoma and reported that survivin expression in an ependymoma correlates with anaplasia and with a high Ki-67 index. High-expression of survivin was shown to be a significant prognostic factor in the univariate analysis, but lost its significance in the multivariate analysis.
The ErbB family of proteins also contains receptor tyrosine kinases and is structurally related to EGFR. ErbB-2 receptor signaling also leads to an aggressive disease behavior in ependymomas by promoting tumor cell proliferation, and inhibition of ErbB tyrosine kinase activity may be used as a useful target for the control of ependymoma progression. C-mos gene is an important gene from the mitogen-activated protein kinase transduction pathway and acts via its protein product mos. Athanasiou et al., in his study involving 34 patients diagnosed with an ependymoma showed that the expression of mos had a significant negative association with recurrence-free interval. But, this association was not significant for overall survival. Hypermethylated in Cancer 1 (HIC-1) is a tumor-suppressor gene, which is hypermethylated or associated with loss of expression in ependymomas and other brain tumors. The hypermethylation of HIC-1 gene is seen in nonspinal ependymomas and not in spinal tumors and thus, these may be a genetically distinct entity. This may account for the differences in survival outcomes between spinal and non-spinal ependymomas.
Rousseau et al., evaluated the methylation of 3 tumor-suppressor genes (CDKN2A, CDKN2B, and p14 ARF) in 152 ependymal tumors. They found that tumor-suppressor gene CDKN2B methylation was more frequent in extracranial tumors, and that CDKN2B and p14 ARF methylation were more frequent in low-grade tumors. Spinal ependymomas were characterized by high-expression levels of HOXB5, PLA2G, and CDKN2A, and pediatric ependymomas have higher levels of LDHB and STAM genes expressed. Based on the various genetic abnormalities in ependymoma, Modena et al., have developed an ependymoma-specific gene expression signature using various abnormal expressions of developmental and differentiation pathway genes. But, it needs to be validated before being routinely used in a clinical setting. Different molecular aberrations with a prognostic impact are tabulated in [Table 1].
A myxopapillary ependymoma usually occurs in the filum terminale of the spinal cord and is generally associated with a better prognosis. Barton et al., evaluated the molecular characteristics of pediatric ependymomas and found that expression of homeobox (HOX)-HOXB13, neurofilament light peptide (NEFL), and platelet-derived growth factor receptor (PDGFR) alpha was seen in myxopapillary ependymomas and these were not expressed in sub-ependymomas. They concluded that HOXB13 is important in the tumorigenesis of myxopapillary ependymomas, and PDGFR alpha (with the availability of inhibitors of PDGFR alpha) may be a potential therapeutic target in myxopapillary ependymomas.
Sub-ependymomas are also Grade I ependymomas and do well with surgery alone without adjuvant radiotherapy. A subependymoma is grouped in the same group as a myxopapillary ependymoma in the World Health Organization classification, and is genetically distinct from a myxopapillary ependymoma. The expression of ETV6, YWHAE, TOP2A, TLR2, IRAK1, TIA1, and UFD1L genes is seen in sub-ependymomas, but not in myxopapillary ependymomas.
Global acetylation of lysine position 9 of histone 3 (H3K9Ac) pattern may also be of prognostic significance in ependymomas. Ebrahimi et al., reported that tumors with <20 % of acetylated nuclei had a higher probability of recurrence, and that recurrent tumors had significantly lower acetylated nuclei compared to primary ependymomas. Sub-ependymomas had more H3K9Ac-positive nuclei than other histological subtypes of ependymomas. This may contribute to its better prognosis and fewer chances of recurrence than the other histological subtypes. They also found that intracranial parenchymal ependymomas had significantly fewer H3K9Ac-positive nuclei. It was also seen that ependymal tumors with more than or equal to 20% H3K9 acetylated cells had a lower MIB-1 expression than those with <20 % acetylated cells. Multivariate analysis confirmed the prognostic significance of H3K9Ac-positivity that was independent of the tumor location. Thus, the variable expression H3K9Ac may be responsible for a better prognosis in some tumor types and some tumor locations.
Clear cell ependymomas are considered as Grade II tumors, but they are not uniform. There may be a group of clear cell ependymomas (about 35%) that may show a high MIB-1 labeling index compared to other Grade II ependymomas and may require special attention when planning adjuvant therapy., Genetic material loss of chromosome 9 is regarded as the molecular hallmark of clear cell ependymomas by some authors. Aberrations involving chromosomes 1q and 3 are also seen in a high frequency in clear cell ependymomas.
An anaplastic ependymoma is considered as a Grade III tumor, and a gain of 1q and a loss on 9 are frequently seen associated with grade III tumors. Classic and anaplastic ependymomas, with a gain of 1q, usually occur in the posterior fossa and have an aggressive clinical course. Recently, Witt et al., reported two distinct cohorts of posterior fossa ependymomas. Transcriptional profiling revealed two demographically, molecularly, transcriptionally, and clinically distinct groups, with the Group A tumors frequently seen in younger patients, with laterally located tumor and having a high risk of recurrence compared to the Group B tumors. There was upregulation of laminin alpha 2 (LAMA2) and neural epidermal growth factor like-2 (NELL2) in Group A and Group B, respectively. Witt et al., in an extensive genetic and epigenetic analysis of posterior fossa ependymomas hardly found any recurrent somatic mutation. However, the global DNA methylation pattern revealed a higher extent of CpG island hypermethylation in Group A tumors, and an overlap with the polycomb repressive complex 2 in embryonic stem cells being responsible for the normal cellular differentiation. However, this phenomenon was absent in Group B tumors. Supra-tentorial ependymomas forms a genetically distinct group with 70% showing a fusion of v-rel avian reticuloendotheliosis viral oncogene homolog A (RELA) and C11orf95. This C11orf95-RELA fusion is exclusively seen in supra-tentorial ependymomas. The RELA fusion protein accumulates in the nucleus and activates transcription of NF-kB. In a landmark article, Pajtler et al., attempted to establish a uniform molecular classification using DNA methylation profiling for all ependymal tumors across all CNS compartments, histopathological grades, and age groups. They came up with nine subgroups [highlighted in [Figure 1]. The tumors are classified according to their location: Supra-tentorial, posterior fossa, and spine. Each subsite comprises of a subpendymoma that is designated as a Grade I tumor. In addition, rest of the supratentorial tumors have been classified into two distinct groups depending on the characteristic RELA fusion and YAP-1 fusion. Similarly, posterior fossa tumors have been classified based on their NELL-2 and LAMA-2 composition. Spinal tumors have two additional groups viz. myxopapillary and anaplastic, characterized by NF-2 mutation , [Figure 1].
Pediatric ependymomas are generally associated with a poorer prognosis than their adult counterparts. Along with other factors, this may also be because pediatric and adult ependymomas are different biologically. Abnormal karyotyping may be seen in up to 90% of the pediatric ependymoma cases with abnormalities involving chromosomes 22 and 1q being the most common ones. The major difference from the adult ependymoma is that genomic gain of 1q is seen more commonly in pediatric ependymomas, whereas gains of chromosomes 7, 9, and 12 seem more prevalent in adult ependymomas. Overexpression of genes LDHB and STAM and down-regulation and deletion of members of the protein 4.1 superfamily may be characteristically seen in pediatric ependymomas.
Ridley et al., evaluated the possible predictive markers for better outcomes in pediatric ependymomas. They evaluated 97 cases and reported that a low expression of nucleolin was a very important biological predictor of outcome in pediatric intracranial ependymomas. They also found that telomerase reactivation and maintenance of telomeric repeats are important in the pathways of ependymoma progression. Urioste et al., reported telomeric alterations and mutations in the MEN1 gene to be important steps in the development of ependymomas in humans. Human telomerase reverse transcriptase (hTERT) is the catalytic subunit of the enzyme telomerase. Tabori et al., reported that hTERT expression a strong predictor for overall survival and progression free survival. In addition there was a good correlation of fTERT expression and telomerase activity. On multivariate analysis, it was found that hTERT expression was the strongest predictor for survival and its prognostic significance was independent of other clinical and pathologic prognostic markers in pediatric ependymomas. Similarly, loss of DNA sequences from chromosome arm 17p is common in sporadic pediatric ependymomas whereas loss of DNA sequences from chromosome arm 22q is seen in adult ependymomas.
There has also been a considerable amount of radio-resistance in ependymomas with a relatively high incidence of local recurrence of these tumors. One of the proposed mechanisms for local control is the p53-mediated growth arrest in ependymomas. p53 mutations may be seen more frequently in cases with an anaplastic ependymoma than with a lower grade of tumor. Gaspar et al., studied this aspect in 24 pediatric patients with an ependymoma and found that irradiation yielded a significant tumor growth delay and tumor regression in the p53 functional tumors and postulated that p53-mediated growth arrest may be responsible for radio-resistance in some cases of ependymoma.
O (6)-methylguanine-DNA-methyltransferase (MGMT) is a DNA repair protein in glial cells. Hypermethylation of this MGMT gene is associated with the inactivation of MGMT and with increased sensitivity to chemotherapy as seen in patients with a glioblastoma. Buccoliero et al., evaluated 12 recurrent cases of anaplastic ependymoma and found the MGMT promoter hypermethylation to be absent in all these cases, and that 75% of the patients had a high MGMT protein expression. This may be one of the reasons for the high chemoresistance of these tumors. This knowledge will be important while choosing chemotherapeutic agents in future chemotherapy trials for patients with an ependymoma both for the primary and for recurrent tumors. Thus, MGMT testing may be done prior to adding alkylating agents to these tumors in future trials.
Korshunov et al., retrospectively evaluated tumor specimens of 76 patients with intracranial ependymomas who received a combined treatment and found that low-grade ependymomas had a significant expression of metallothioneins, glutathione S-transferase pi, and P-glycoprotein. Together, these three are called chemoresistance-related proteins and may be one of the main reasons for the development of chemoresistance in low-grade ependymomas. It was also found that these markers may be helpful in predicting local tumor progression in ependymomas.
In the recent years, significant improvement has taken place in the understanding of the biology of ependymomas. In a landmark article, Pajtler et al., have classified the ependymal tumors into nine subgroups depending on their location and behavior., This subgroup allows for a better understanding of the nature of the disease. The subependymal tumors are considered as Grade I tumors and surgical resection alone appears to be the optimum treatment. Majority of the grade III tumors have been found to have characteristic molecular aberrations [Figure 1]. However, there is ample opportunity to intensify treatment even after surgery and adjuvant radiation. These distinct molecular aberrations also open a new window of research. Newer molecules such as gamma-secretase inhibitor pertinent to the notch signaling pathway, hTERT inhibitor (telomerase inhibitor), dual inhibitor of EGFR (ERBB1) and ERBB2 (lapatinib), anti-VEGF (bevacizumab), anti-EGFR (erlotinib), and mTOR inhibitor (everolimus) merit evaluation in well-designed trials [Figure 2].
Traditionally, the histological grade, site of the tumor, and age at diagnosis have remained the main prognostic factors in patients harboring an ependymoma. With increasing knowledge of molecular biology, better predictive and prognostic factors have been detected. The prognostic significance of these factors is due to the different tumor biology of different varieties of these tumors at varying locations even though all of them have been clubbed under the common heading of 'ependymoma.' Surgery and radiotherapy have remained the main treatment options for an ependymoma. Better knowledge of the molecular biology of these tumors also gives us insight into their relative chemoresistance and radio-resistance. Thus, knowledge gained regarding these predictive factors help us in tailoring adjuvant treatment for these patients. It also helps us in selecting the proper chemotherapy regimens.
More than knowledge, it is the the clinical validation of these factors and their logical use in therapeutic practice that will help in improving the prognosis of patients harboring an ependymoma. Molecular classification of ependymomas considerably help us in this regard.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2]