Role of mTOR signaling pathway in the pathogenesis of subependymal giant cell astrocytoma – A study of 28 cases
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.190274
Source of Support: None, Conflict of Interest: None
Keywords: Brain tumors; hamartin; mTOR; Phospho-p70S6 kinase; subependymal giant cell astrocytoma; tuberin; tuberous sclerosis complex
Subependymal giant cell astrocytomas (SEGA) are World Health Organisation (WHO) grade I intraventricular tumors occurring in 6–14% of individuals with tuberous sclerosis complex (TSC). TSC is an autosomal dominant disorder resulting from mutations of genes TSC1 and TSC2, diagnosed clinically, based on the standard criteria.,, Presence of SEGA is a major criterion for the diagnosis of TSC. Detectable as early as 23 weeks after birth, these slow-growing tumors typically present with features of obstructive hydrocephalus in the second decade of life. Occurrence of SEGA outside of TSC is debatable. TSC1 and TSC2 genes function as tumor suppressors, encoding for the proteins hamartin and tuberin. These proteins form the central component of the pathway that modulates insulin-like growth factor-mediated cell signaling, which regulates cell growth and proliferation by inhibiting the mammalian target of rapamycin (mTOR) pathway., Mutations in TSC genes and resultant inactivation of hamartin or tuberin cause excessive phosphorylation of mTOR signaling proteins leading to defective cell migration and cytomegaly, thus resulting in the characteristic lesions of TS, such as tubers, nodules, and hamartomas in multiple organs, including the brain, kidney, skin, and heart., Mutations in TSC1 and TSC2 have also been shown to affect neuronal differentiation, resulting in a stem cell phenotype. Till date, 135 mutations of TSC1 and 360 mutations of TSC2 gene have been reported in TSC, with causative mutations occurring in almost all exons. No mutational hot spots are defined, a significant drawback considering the large size of the genes. The pathogenesis of SEGA in TSC is uncertain. Several studies have attempted to define the phenotype of the different cells types found in SEGA, viz., giant cells, dysmorphic neurons, and astrocytes. However, the results have not been conclusive. With this background, we aimed to look for TSC1 and TSC2 mutations, analyze the immunoexpression of tuberin and hamartin, and elucidate the role of mTOR pathway activation in the pathogenesis of SEGA.
All cases of SEGA resected for treatment of medically intractable epilepsy, diagnosed between 1982 and 2013, were retrieved from the archives of two institutes, All India Institute of Medical Sciences, New Delhi (AIIMS), and G.B. Pant Hospital, New Delhi (GBPH). Hematoxylin-and-eosin–stained slides were reviewed for confirmation of the diagnosis by two neuropathologists (M.C.S., V.S.). Clinical data were also obtained.
Five-micron-thick sections cut from formalin-fixed, paraffin-embedded (FFPE) tissue blocks were immunolabeled with the following antibodies using standard protocol: Hamartin (dil 1:400; Novacastra, UK), tuberin (dil 1:25; Acris, Germany), phospho-p70S6 kinase (Thr389; dil 1:100; Cell Signalling, MA, USA), S6 ribosomal protein (dil 1:50; Cell Signalling, MA, USA), phospho-S6 ribosomal protein (dil 1:400; Cell Signalling, MA, USA), phospho-4E-BP1 (Thr37/46; dil 1:1600, Cell Signalling, MA, USA), Stat3 (dil 1:600; Cell Signalling, MA, USA), and phospho-Stat3 (dil 1:200; Cell Signalling, MA, USA). Sections from normal brain cortex were used as controls. Staining for phospho-p70S6 kinase, S6 ribosomal protein, phospho-S6 ribosomal protein, phospho-4E-BP1, Stat3, and phospho-Stat3 was estimated as follows: No staining, 0; faint positivity in <10% of cells, 1+; moderate positivity in 10–50% of cells, 2+; and, strong positivity in ≥50% of cells, 3+. Only 2+ and 3+ staining were considered as positive.
DNA extraction and sequencing
DNA was extracted from tumor samples with adequate material in the FFPE blocks. Eight serial sections of 10-µ thickness were collected from each block. Recover All™ Total Nucleic Acid Isolation Kit (Ambion, Texas, USA) was used to extract DNA as per the manufacturer's instructions. Exons 15 and 17 of TSC1 gene and exons 33, 39, and 40 of TSC2 gene were selected for mutation analysis, as they are the most frequently mutated warm spots. Polymerase chain reaction (PCR) amplification was carried out using forward and reverse primers [Table 1], in a 10-µl reaction mixture containing 50 ng of DNA, 1 µl of 10× PCR buffer, 0.8 µl of 10 mM dNTPs, 0.25 µl each of forward and reverse primers, and 0.2 µl of AmpliTaq Gold PCR Master mix (Applied Biosystems, California, USA). Initial denaturation was carried out at 95°C for 1 minute, annealing at 57°C for 45 seconds, and extension at 72°C for 2 minutes. Bidirectional sequencing was performed using ABI 3730 sequencer (Applied Biosystems, California, USA). Mutation analysis was done using “Mutation Taster” online software, for predicting disease-causing potential of sequence alterations.
A total of 28 cases of SEGA were identified [Table 2], 21 from AIIMS and 7 cases from GBPH. The median age of the patients was 15 years (range, 2–60 years), with a male:female ratio of 3:1. Of the 28 patients, 10 met the criteria for clinical diagnosis of TSC, with presence of other associated lesions such as adenoma sebaceum, tubers, angiomyolipoma, and polycystic kidney disease. Tumors were located predominantly in the left lateral ventricle (19 of 28 cases, 68%), whereas 32% (9 of 28 cases) were in the right lateral ventricle. All patients presented with seizures, with the duration of symptoms ranging from as less as 15 days to as long as 7 years (median, 0.8 years). The mean duration of seizures in patients with TSC was considerably longer (1.3 years; range: 0.08 to 7 years) as compared with those without TSC (0.62 years; range, 15 days to 2 years).
All cases showed a similar histology. Microscopic examination revealed moderate cellularity, composed of three different populations of cells, viz., large pleomorphic ganglion-like cells with abundant glassy eosinophilic cytoplasm, gemistocytic astrocytes, and spindled cells. Areas of dystrophic calcification, lymphocytic and mast cell infiltrates, and presence of hyalinized and thickened blood vessels were frequent, while mitoses, microvascular proliferation, and necrosis were absent [Figure 1]. One case also showed the presence of a tuber in the adjacent brain parenchyma.
Expression of hamartin and tuberin
Loss of hamartin immunoexpression alone was seen in
eight cases (28%), while loss of tuberin alone was seen in nine cases (32%). One tumor (4%) showed loss of both hamartin and tuberin, while 36% (10 of 28) did not show loss of either protein. Loss was observed in the giant cell, spindle cell, and astrocytic components of the tumors. Loss of immunoexpression was irrespective of a clinical diagnosis of TSC (P = 0.99). The single case of SEGA that showed loss of both hamartin and tuberin also showed this dual loss in the cortical tuber (case 11) [Figure 2]a and [Figure 2]b.
Activation of mTOR cascade
In SEGA, overexpression of p70S6 kinase was seen in all cases (100%). Ribosomal S6 protein was overexpressed in 23 cases (82%), whereas phospho-S6 ribosomal protein was seen in 18 cases (64%). Immunoreactivity for phospho-4E-BP1 was seen in 18 cases (64%). All cases showed strong immunoreactivity for Stat3 (100%), while only four cases (14%) showed strong but focal expression of phospho-Stat3. Normal adult neocortex included as control showed faint immunostaining of neurons for phospho-p70S6 kinase, ribosomal S6, phospho-S6 ribosomal protein, and Stat3. Immunostaining for phospho-4E-BP1 and phospho-Stat3 was negative in the normal brain cortex. There was no significant difference in the expression of mTOR pathway proteins between cases with loss of either hamartin or tuberin and cases with preserved protein expression [Table 3]; [Figure 2]c,[Figure 2]e,[Figure 2]f,[Figure 2]g,[Figure 2]h,[Figure 2]i.
Of the 20 cases, 5 (25%) showed pathogenic mutations in TSC genes [Table 4], of which 2 patients were known cases of TSC. Three cases showed point mutation in exon 15 of TSC1 gene, while two cases showed point mutation in exon 33 of TSC2 gene. All cases except one showing point mutation also showed loss of either tuberin or hamartin on immunohistochemistry (IHC). Polymorphisms were also identified solely in exon 33 of TSC2 gene.
The current WHO classification of central nervous system (CNS) tumors classifies SEGA as a slow-growing tumor (WHO grade I) of mixed glioneuronal lineage. Role of TSC1 and TSC2 in the molecular pathogenesis of TSC is well established. However, it was only recently shown in studies that inactivating mutations in TSC1 and TSC2 are linked to hyperactivation of the mTOR pathway., To analyze the hypothesis that dysregulated mTOR signaling due to TSC mutation leads to SEGA, we performed IHC for tuberin and hamartin, as well as mTOR pathway markers in the present study. We found loss of hamartin and tuberin in 32% and 36% of SEGA, respectively. Loss of protein expression was seen in all the tumor components of SEGA. Simultaneous loss of hamartin and tuberin was a rare event, seen only in one case. Surprisingly, in two cases having neuronal and extraneuronal manifestations, and fulfilling the definite clinical diagnostic criteria for TSC, SEGA did not show immunoloss of tuberin or hamartin. Jozwiak et al., found loss of hamartin in all cases (9 of 9) and loss of tuberin in 67% of cases (6 of 9) with SEGA examined by them. In another study, tuberin was absent in 55% of SEGA cases (6 of 11) in epithelioid and spindle cell components, while giant cells showed faint staining. We found pathogenic mutations in 20% of SEGA analyzed, both in TSC-associated and non–TSC-associated cases, although we could not perform complete screening of all exons owing to the large size of the two genes. Mutations have been reported in almost all the exons of the TSC genes, and no mutation accounts for more than 2% of all mutations seen. Comprehensive mutation studies in TSC patients have led to the detection of mutations in approximately 85% of patients studied. Most mutations are located on exon 15 of TSC1 gene and exon 32 of TSC2 gene. Mutations in exon 35–39 of TSC2 gene encoding GAP domain are associated with a higher frequency of neurological lesions. In concordance with the aforementioned observations, we found mutations in exon 15 of TSC1 gene and exon 33 of TSC2 gene. Most cases with mutation showed loss of protein immunoexpression, except for a single mutated case that did not exhibit loss of either hamartin or tuberin. Most TSC1 mutations are point mutations that lead to the truncation and loss of protein, while TSC2 mutations are largely missense mutations, in-frame deletions and large rearrangements, which lead to premature termination. It is possible that some mutations may produce an unstable or nonfunctional intact protein that may be detectable by IHC, as in our case. Although the “two-hit” mechanism of germline and somatic mutations has been proposed for tubers in TSC patients, studies have not shown convincing evidence of loss of heterozygosity in brain lesions, indicating preservation of one of two alleles., Hence, cases may show preserved protein expression although they may be having clinical manifestations of the disease, owing to synthesis of some protein product of the retained allele, which may not be functional. This should be kept in mind while using hamartin or tuberin as surrogate markers for mutation testing or for diagnosing SEGA on small biopsies.
mTOR is a 280-kDa serine/threonine protein kinase that regulates cell growth, ribosome biogenesis, and autophagy, and is composed of two complexes: mTORC1, which is rapamycin-sensitive, and mTORC2, which is rapamycin-insensitive. Disruption of the mTOR pathway has been implicated in various diseases, such as diabetes, obesity, neurodegenerative disorders, as well as cancers, including gliomas., Recent studies have shown mTOR to play a pivotal role in brain development, regulation of neuronal size and its morphology, dendrite formation, and axon elongation. Experiments in mouse models and cultured mammalian cells lacking either TSC1 or TSC2 have demonstrated constitutive high-level phosphorylation of p70S6 kinase, 4E-BP1, and Stat3.,, Ribosomal S6 protein is phosphorylated by these kinases, and recent studies have shown that phospho-S6 ribosomal protein immunoreactivity is a biomarker for hyperactive mTOR pathway., The present study showed overexpression of mTOR pathway proteins, viz., phoshpo-p70S6 kinase, S6 ribosomal protein, phospho-S6 ribosomal protein, and phospho-4E-BP1. Our study shows that SEGA are definitely associated with mTOR pathway activation, thus supporting the findings of two previous studies that demonstrated hyperactive mTOR pathway in SEGA., However, this activation could not be attributed to loss of TSC1 and TSC2. In normal cells, TSC1–TSC2 complex acts as an inhibitor of mTOR. Studies on Drosophila have led to placement of dTsc genes downstream of the Drosophila protein kinase, Akt., Morever, tuberin also acts as substrate for other upstream kinases, which have modulatory effects on downstream targets of tuberin. Hence, it is possible that alterations in these kinases upstream of the TSC complex may lead to hyperactivation of mTOR pathway, leading to excessive cell proliferation in SEGA. However, distinctive neurological involvement suggests additional roles of hamartin–tuberin complex, which remain to be elucidated. Recent studies have also shown phospho-S6 ribosomal protein immunoreactivity in TSC, FCD IIB, and related malformative cortical lesions such as hemimegalencephaly and gangliogliomas, without any demonstrable loss of TSC gene activity.,, Demonstration of robust expression of phosphorylated downstream mTOR signaling proteins in cases showing strong protein expression of tuberin and hamartin in other malformative lesions possibly suggests that all these lesions involve activation of the mTOR pathway. However, whether this hyperactivation is secondary to loss of TSC1 or TSC2 function or to some other intermediary molecular link needs to be investigated further.
An interesting feature of low-grade gliomas is their propensity to spontaneously stop growing or to even regress. This slow growth rate and inability to progress to a higher histological grade may be attributed to the phenomenon of oncogene-induced senescence (OIS). Recent studies have demonstrated the role of the JAK/STAT3 signaling pathway in the induction of OIS, mediated by various cytokines., However, whether OIS is induced in SEGA is still unexplored. A few previous studies have shown increased expression of Stat3 in tubers., We performed IHC for Stat3 and phospho-Stat3 to elucidate the link between TSC2–mTOR signaling and Stat3 in regulating cell proliferation. While cytoplasmic Stat3 expression in giant cells was increased in all cases, only a small proportion of cases showed nuclear overexpression of phospho-Stat3. This is a significantly lower frequency of phospho-Stat3 expression as compared with a previous study on SEGAs and tubers that showed 100% positivity., However, this finding suggests that, while hyperactivation of mTOR pathway leads to dysregulated proliferation of cells, increased cytoplasmic Stat3 expression in these cells may be responsible for their ability to escape uncontrolled proliferation and thus limit tumor growth.
Mutations in TSC1 and TSC2 genes, and loss of proteins hamartin and tuberin play a definite role in the pathogenesis of SEGA. However, they are of limited value as diagnostic markers because loss of protein expression is not seen in all cases and analyzing TSC1 and TSC2 genes for mutation is not always feasible. mTOR pathway activation is a feature of SEGA, but its association with TSC gene alterations remains unclear and warrants further investigation to completely understand its role in the pathogenesis of these tumors.
We are thankful to Mr Pankaj and Mrs Kiran Rani for performing the immunohistochemistry.
Financial support and sponsorship
Conflicts of interest
None of the authors has any conflicts of interest to disclose.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]