|
||||||||||||||||||||
|
|
Loss of SMARCB1/INI1 Immunoexpression in Chordoid Meningiomas
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.273647
Background: Chordoid meningiomas have an aggressive clinical course characterized by frequent recurrences. Recent whole-genome sequencing studies demonstrated Chr22 loss in chordoid meningiomas not accounted for by NF2 mutations. SMARCB1/INI1 is a candidate gene on Chr22, which has not been analyzed extensively in meningiomas. AKT1 mutation has been recently identified to be a driver of meningiomagenesis. Keywords: AKT1 mutation, chordoid, immunohistochemistry, INI1, meningioma, SMARCB1
Meningiomas are common tumors of the central nervous system (CNS) originating from the meningeal coverings of the brain and spinal cord. They are often associated with heterozygous inactivating mutations of the NF2 tumor suppressor gene located on chromosome 22q.[1] However, recent whole genome sequencing studies have identified novel mutations in TRAF7, KLF4, AKT1, and SMO in non-NF2 mutated meningiomas, thus expanding the genetic landscape of these tumors.[1],[2] SMARCB1/INI1 is a tumor suppressor gene located on chromosome 22q, which plays an important pathogenetic role in several rhabdoid and non-rhabdoid tumors. Loss of expression of SMARCB1/INI1 is an important feature for the diagnosis of various neoplasms including atypical teratoid/rhabdoid tumor (AT/RT) of CNS, renal and extrarenal rhabdoid tumors, cribriform neuroepithelial tumor (CriNET), and epithelioid sarcoma.[3],[4],[5],[6] Occasional reports on the mutation of SMARCB1/INI1 in meningiomas are available.[7],[8] However, this alteration has not been studied in detail in these tumors, and there is a paucity of data on SMARCB1/INI1 mutation and loss of protein expression in specific subtypes of meningiomas. Chordoid meningioma is a rare subtype of meningioma comprising approximately 0.5% of all meningiomas, and it shares a histological similarity to chordoma.[9] Initially described as “vacuolated” and “myxoid” meningiomas, these tumors were found to exhibit a very high rate of recurrence even after gross total resection. Therefore, they were accorded the World Health Organization (WHO) grade II.[10],[11],[12],[13],[14] While loss of chromosome 22q has been reported in chordoid meningiomas, the frequency of loss exceeds that of NF2 alterations.[1] This suggests that some other candidate genes may be present on chromosome 22 that are involved in the pathogenesis of these tumors. The INI1 gene located on Chr 22q is one such potential candidate gene that remains unexplored in these tumors. In addition, chordoid meningiomas bear morphological similarity to the chordoma, a neoplasm which has recently been reported to demonstrate loss of INI1 in a subset of cases.[15],[16] We, therefore, conducted this retrospective study to analyze chordoid meningiomas for SMARCB1/INI1 mutations, as well as loss of INI1 immunoexpression, to determine if INI1 has a role to play in the pathogenesis of these tumors.
All cases of chordoid meningioma diagnosed over a period of 8 years (2007–2014) were retrieved from the archives of the Department of Pathology of our Institute. Surgical specimens had been fixed in neutral buffered formalin, embedded in paraffin, and routinely processed. Five-micron thick tissue sections were cut from the representative paraffin blocks of each case for hematoxylin and eosin (H and E) staining and immunohistochemistry. H and E-stained sections were reviewed independently by three neuropathologists (AK, MCS and CS), and only cases with exclusive chordoid morphology were included in the study. Other subtypes of meningiomas were included for comparison including all grades and various subtypes. All experiments using human samples were approved by the Institutional Ethics Committee (reference number IESC/T-349/12.09.2014). Immunohistochemistry Immunohistochemistry was performed using antibodies directed against epithelial membrane antigen (EMA) (Dako, Denmark; 1:100), vimentin (Novocastra, United Kingdom; 1:100), MIB-1 (Dako, Denmark; 1:200), progesterone receptor (PR) (Dako, Denmark; 1:50), D240 (Dako, Denmark; 1:100), pancytokeratin (Neomarker, USA; 1:200), S-100 (Dako, Denmark, 1:100), CD34 (Dako, Denmark, 1:100), brachyury (SantaCruz Biotechnology, USA, 1:100), and INI1 (Cell Marque, USA; 1:100). Labelled streptavidin biotin kit (Universal, Dako, Denmark) was used as a detection system. Antigen retrieval was performed in a microwave oven using citrate buffer at pH 6.0. Staining of endothelial cells was used as internal positive control for INI1. The labeling index (LI) for MIB-1 was calculated in the highest proliferating area as the percentage of labeled nuclei per 1000 cells. SMARCB1 and AKT1 mutation analysis Cases with adequate material in paraffin blocks were taken up for genetic analysis. DNA isolation and SMARCB1 gene sequencing for all 9 exons were performed, as described previously.[15],[16] AKT1 p.E17K mutation positioned in exon 4 was analyzed using polymerase chain reaction (PCR) and direct DNA sequencing. The primers for PCR amplification and sequencing of AKT1 exon 4 were forward 5'-CTGGCCCTAAGAAACAGCTCC-3'and reverse 5'-CGCCACAGAGAAGTTGTTGA-3'. Reaction conditions for PCR amplification were 45 cycles at 95°C for 30 seconds, 60°C for 1 minutes, and 72°C for 1 minutes. PCR products were purified and bidirectional sequencing was performed using ABI 3730 sequencer (Applied Biosystems, USA). Nucleotide sequence analysis was carried out using DNA Star (DNASTAR, Inc. Madison, Wisconsin USA). Genetic variations identified were analyzed for their pathogenic effect by three different online tools, viz. Sequence and evolutionary conservation-based method, SIFT (Sorting Intolerant from Tolerant) (http://sift.jcvi.org/); Protein sequence and structure-based method, PolyPhen-2 (Polymorphism Phenotyping version 2) (http://genetics.bwh.harvard.edu/pph 2/); and supervised learning method, MutationTaster (http://www.mutationtaster.org/).
Sixteen cases of chordoid meningioma were identified in this retrospective study [Table 1] (Cases 1–16), with a mean age of 37.3 (range: 17–67 years) and a slight female predilection (M:F = 0.8). They accounted for 1.1% (16/1408) of the meningiomas diagnosed in 8 years. Sixteen meningiomas of other subtypes [Table 1] (Cases 17–32) were analyzed for comparison (mean age: 37.7 years; range: 10–65 years). They included 8 WHO grade I tumors, 5 clear cell (WHO grade II), 2 rhabdoid (WHO grade III), and 1 papillary (WHO grade III) meningiomas.
Histopathology and immunohistochemistry Of the 16 cases of chordoid meningiomas analyzed, 12 showed typical morphology with round-to-oval cells arranged in cords in a myxoid background, mimicking a chordoma [Figure 1]. In four cases, tumor cells were oval-to-spindle shaped, arranged in cords in an extensively myxomatous background. Mitotic figures were scarce, and none of the tumors showed necrosis. Underlying brain invasion was observed in two cases. All cases showed immunopositivity for EMA, vimentin, D240, and PR; one case was focally immunopositive for pancytokeratin, and three were positive for S-100 protein [Figure 2]. None of the tumors were immunopositive for brachyury, CD34, or GFAP. MIB-1 LI ranged from 1% to 20% (mean LI: 3.7%). Six of the 16 cases of chordoid meningioma (37.5%) showed loss of INI1 immunoexpression in the tumor cells [Figure 2]. None of the other subtypes of meningioma included for comparison showed loss of INI1 expression. Patients with INI1 loss were older (mean age = 39.7 years) than those with retained INI1 expression (mean age = 35.8 years), however, this difference was not statistically significant (P = 0.998; independent t-test). Cases with INI1 loss showed a higher MIB-1 LI (mean MIB-1 LI: 6.2%) as compared to cases with retained INI1 expression (mean MIB-1 LI: 2.2%) (P = 0.025, independent t-test).
SMARCB1 and AKT1 mutation analysis Six cases of chordoid meningioma (four cases showing INI1 loss and two INI1 immunopositive cases) with adequate material in paraffin blocks were taken up for mutation analysis. Bidirectional Sanger sequencing did not reveal pathogenic mutations in any of the cases analyzed. However, non-pathogenic polymorphisms in SMARCB1 gene were identified in four cases (66.7%). Genetic variations were observed in intron 9 (c1119-41G>A) in four cases, exon 7 (c897 C>A) in three cases (with two cases in heterozygous and one in homozygous state), and exon 4 (c500 + 21 T>A) in one case [Table 2]. AKT1 E17K missense mutation was identified in a single case (Case no. 10), which also showed loss of INI1 expression, and had the highest MIB-1 LI of 20% [Figure 3].
The last few years have seen a virtual explosion in data regarding the molecular characterization of meningiomas. While the involvement of 22q loss and mutations of the NF2 gene have long been known, recent studies have identified mutations in TRAF7, KLF4, AKT1, and SMO genes in non-NF2 altered meningiomas.[1],[2] Chordoid meningioma is a distinct histopathological subtype of meningioma, which is characterized by aggressive clinical behavior. A recent whole genome sequencing study included seven chordoid meningiomas, five of which showed loss of Chr 22 (71.4%). In addition, other alterations that overlapped with Chr 22 loss, viz. AKT1, TRAF7 mutations and isolated KLF4 mutations, were also identified.[1] Notably, only one of the tumors showing Chr 22 loss also harbored NF2 mutation, highlighting the fact that frequency of loss of Chr 22 significantly exceeds that of NF2 alterations, and that there may be other genes on Chr 22 which may be implicated in the pathogenesis of these tumors. SMARCB1/INI1 gene located at Chr 22q11.2 is a core subunit of the ATP-dependent chromatin remodeling complex,[17],[18] and the INI1 protein is ubiquitously expressed in the nuclei of all normal cells. Loss of SMARCB1/INI1 expression in rhabdoid tumors has been found to drive the cell cycle by activating Cyclin D1.[19] Although it was previously believed that absence of SMARCB1/INI1 staining was the hallmark of rhabdoid tumors alone, subsequent studies have revealed that several tumors, including those without rhabdoid morphology, also show INI1 loss. Some of these are medullary carcinoma of kidney, epithelioid sarcoma, epithelioid malignant peripheral nerve sheath tumor, extraskeletal myxoid chondrosarcoma, CriNET, and some cases of schwannoma, myoepithelial carcinoma, and poorly differentiated chordoma.[16],[17],[20],[21],[22],[23],[24] It has also been observed that many tumors with rhabdoid morphology do not show INI1 loss, including rhabdoid meningiomas, despite similar allelic loss of 22q.[25],[26] Two studies have evaluated INI1 in meningiomas because its location on Chr 22 makes it a suitable candidate gene to be involved in meningioma pathogenesis. Schmitz et al. reported that four out of 126 meningiomas showed a point mutation in exon 9 of SMARCB1 gene; however, loss of INI1 immunoexpression was not noted in these cases.[27] They proposed that SMARCB1 mutation cooperates with NF2 mutation, rather than itself playing a role in the pathogenesis of meningiomas. Riske et al. identified SMARCB1/INI1 mutation in 1 of 80 meningiomas analyzed.[28] Bruder et al. studied 41 meningiomas and 23 schwannomas and demonstrated polymorphic changes in intron 8 of SMARCB1 gene, however, no inactivating mutations were observed.[29] In the present study, we performed immunohistochemistry for SMARCB1/INI1 in 16 chordoid meningiomas and in 16 cases of other subtypes of meningiomas. Chordoid meningiomas showed INI1 loss in a significant proportion of cases; however, none of the other subtypes showed INI1 loss. Mutation analysis of chordoid meningiomas revealed AKT1 (E17K) mutation in a single chordoid meningioma, where it coincided with loss of INI1 expression and had the highest MIB-1 LI in the series. However, none of the cases harbored SMARCB1 mutation; only non-pathogenic polymorphisms were identified in intron 9 and in exons 4 and 7. Previously, Schmitz et al. described mutations in exons 4 and 7 of SMARCB1 in meningioma.[27] Our results coincide with previous studies on rhabdoid tumors and epithelioid sarcomas, where protein loss was observed despite an intact SMARCB1 region.[30],[31],[32] Thus, identification of INI1 loss in chordoid meningiomas further expands the family of INI1-deficient neoplasms. In the absence of mutations in the SMARCB1 gene in chordoid meningiomas, it is possible that this loss could be secondary to post-transcriptional or epigenetic changes, as observed in other neoplasms.[33],[34] Studies have shown that SMARCB1 alterations in epithelioid sarcoma are not caused either by promoter and/or histone hypermethylation,[33] but have suggested the role of miR-206, miR-381, and miR-671-5p that target SMARCB1 at the mRNA and protein level.[34] Kohashi et al. reported overexpression of miR193a-5p in malignant rhabdoid tumors, and suggested its role in downregulation of SMARCB1 mRNA.[35] These findings indicate the likely interplay between epigenetic modifiers and post-translation modifications in the genesis and/or progression of tumors associated with INI1 loss. It has been suggested that loss of INI1/SMARCB1 function affects the cellular actin cytoskeleton, providing a potential explanation for the aggressive behavior, recurrence, and metastatic nature of cancers.[36],[37] In our study, chordoid meningiomas with loss of INI1 had higher proliferation indices compared to those with retained expression, which indicates a propensity for aggressive behavior. Thus, we postulate that loss of INI1 expression is an important pathogenetic mechanism for chordoid meningiomas, possibly a second hit on Chr 22, and may play a role in their aggressive behavior.
This is the first study to demonstrate the loss of SMARCB1/INI1 expression in a significant proportion of chordoid meningiomas while being absent in other meningioma subtypes. This immunohistochemical marker may, therefore, be of utility as a diagnostic marker for these tumors, as an adjunct to histomorphology. Loss of INI1 expression appears to show an association with higher MIB-1 LI, indicative of more aggressive biological behavior, setting the stage for prospective studies with long-term patient follow-up. Further studies are also required to unravel the role of this loss of expression in their pathogenesis, as well as the mechanism by which this loss occurs. Identification of AKT1 mutation, known to activate PI3K/AKT pathway, opens up new avenues for targeted therapy in patients with these aggressive tumors. Establishment of the varied pathogenetic mechanisms among different meningioma subtypes will ultimately aid in planning and developing targeted therapeutic strategies for these tumors in the future. Acknowledgment The authors are thankful to Ms. Kiran Rani and Mr. Pankaj Kumar for their assistance in immunohistochemistry studies. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2]
|
|
|||||