Introduction
Vestibular schwannomas are benign tumours that arise from Schwann cells of the vestibular portion of the VIIIth cranial nerve. They account for approximately 6 per cent of intracranial tumours, and occur in two distinct populations: unilateral vestibular schwannomas, which are predominately sporadic,Reference Kasantikul, Netsky, Glasscock and Hays1, Reference Erickson, Sorenson and McGavran2 and bilateral vestibular schwannomas that occur almost exclusively in patients with germline neurofibromatosis type 2 (NF2).Reference Martuza and Ojemann3, Reference Baldwin, King, Chevretton and Morrison4
A frequent mutation in vestibular schwannoma is a mutation in the NF2 coding region located on chromosome 22 band q11-13.1. It encodes a 595-amino-acid protein that has been named ‘merlin’ (moesin-ezrin-radixin-like protein).Reference Neff, Welling, Akhmametyeva and Chang5 Sporadic and NF2 vestibular schwannomas are associated with loss of functional merlin in both alleles, in 66 per cent and 33 per cent respectively.Reference Neff, Welling, Akhmametyeva and Chang5 Merlin is a known regulator of cell signalling pathways, encompassing cell–matrix adhesion, proliferation and survival. Merlin's tumour suppressor function is mainly anti-proliferative, and is mediated through the inhibition of cell cycle regulators such as downstream target cyclin D1.Reference Xiao, Gallagher, Shetler, Skele, Altomare and Pestell6 Mutations in merlin have been discovered in a spectrum of central nervous system tumours including schwannomas, meningiomas and ependymomas.
Cell cycle deregulation through the activation of oncogenes and inactivation of tumour suppressor genes is well described in both benign and malignant tumours.Reference Aarhus, Bruland, Saetran, Mork, Lund-Johansen and Knappskog7–Reference Sonoki, Harder, Horsman, Karran, Taniguchi and Willis14 The D-type cyclins (D1, D2 and D3) are induced by mitogens during the G1 cell cycle phase, and continued synthesis throughout the cell cycle depends on continuous growth factor stimulation. Cyclin D1 is a protein derived from the CCND1 gene on chromosome 11q13, and cyclin D3 is a protein closely related to cyclin D1, derived from the CCND3 gene on chromosome 6p21. The role of cyclin D1 in a number of human neoplasias such as mantle cell lymphoma is well established.Reference Donnellan and Chetty15 Cyclin D3 is also speculated to have a role in tumourigenesis.Reference Doglioni, Chiarelli, Macri, Dei Tos, Meggiolaro and Dalla Palma16
Ki-67 is a nuclear protein expressed by an individual cell when it is in the replication cycle.Reference Gerdes, Schwab, Lemke and Stein17 The percentage of tumour cells that express Ki-67 immunohistochemistry is termed the Ki-67 proliferation index. This is considered to be a surrogate marker for tumour proliferation, and in current medical practice it is used as an independent prognostic determinate in a variety of tumours.Reference Hunter and Pines18–Reference Xiong, Connolly, Futcher and Beach20
Previous studies of cell cycle regulatory proteins in patients with vestibular schwannoma have revealed inconsistent results.Reference Lassaletta, Del Rio, Torres-Martin, Rey, Patron and Madero21–Reference Neff, Oberstien, Lorenz, Chaudhury, Welling and Chang23 This study aimed to further evaluate the role of cyclin D1 and cyclin D3 proteins in vestibular schwannoma. A secondary aim was to determine whether proliferation index as measured by Ki-67 was associated with patient age at diagnosis.
Materials and methods
Patients
Following ethics committee approval, 180 consecutive patients with previously untreated vestibular schwannoma, who underwent surgical excision over a 15-year period (1998 to 2013), were identified for subsequent tissue sample retrieval from the case records of the Department of Head and Neck Surgery at St Vincent's Hospital, Sydney, Australia. There were 84 females and 96 males; patients were aged 14–81 years at the time of surgery (mean age of 54 years).
Tissue processing and microarray
The archived formalin-fixed paraffin-embedded tumour material was retrieved. The tissue was cut into 4-μm-thick sections for haematoxylin and eosin staining and Ki-67 immunohistochemistry. Representative areas were chosen, and biopsies of 2 mm diameter were taken from these donor blocks and included in recipient tissue microarray blocks using a precision tissue array instrument (Tissue-Tek Quick-Ray System, Sakura Finetek USA, Torrance, California, USA). From this tissue array block, 4-μm-thick sections were prepared for further immunohistochemistry analysis.
Immunohistochemistry
Tissue microarray sections were used for cyclin D1 and cyclin D3 immunohistochemistry, whereas sections of the original formalin-fixed paraffin-embedded tumour material were used for Ki-67 immunohistochemistry. Tissue sections of 4 μm thickness were cut from formalin-fixed, paraffin-embedded tissue. Immunohistochemistry staining was performed using the Ventana Benchmark Ultra automated slide stainer (Ventana Medical Systems, Oro Valley, Arizona, USA).
In brief, dewaxing and rehydration, followed by the blocking of endogenous peroxidase with a 3 per cent hydrogen peroxide and methanol mixture, was carried out. After undergoing heat-induced epitope retrieval using Ventana buffer (type CC1; Ventana Medical Systems), 36 minutes for cyclin D1, 92 minutes for cyclin D3 and 36 minutes for Ki-67, primary monoclonal antibodies with either cyclin D1 (1:100) (code M3642; Dako, Golstrup, Denmark), cyclin D3 (1:50) (CyclinD3; Leica Biosystems, Buffalo Grove, Illinois, USA) or Ki-67 (a ready-to-use antibody with no specified dilution) (Confirm Ki-67 (30-9); Ventana Medical Systems), were incubated, for 48 minutes for cyclin D1, 60 minutes for cyclin D3 and 28 minutes for Ki-67, at 36°C. Slides were then stained with a Ventana Ultraview Detection Kit. An additional amplification step was performed for cyclin D3 using a Ventana Amplification Kit. All washing procedures were performed in phosphate buffered saline. Slides were then counterstained with Ventana Hematoxylin I, followed by Ventana Bluing Solution (Ventana Medical Systems). Tissues expressing cyclin D1, cyclin D3 or Ki-67 protein were identified by deposition of brown chromogen, with the haematoxylin counterstain staining light blue.
Expression quantification
Normal tonsil tissue was used as a positive control for Ki-67, and breast carcinoma was used as a positive control for cyclin D1 and cyclin D3 (Figures 1a, 2a and 3a, respectively).Reference Vielh, Chevillard, Mosseri, Donatini and Magdelenat24–Reference Chen, Lee, Hsieh, Ho, Pan and Huang26
Fig. 1 Immunohistochemical localisation of Ki-67 in a normal tonsil tissue section magnified at ×100 (a) and a vestibular schwannoma tissue section magnified at ×200 (b). Intensity of nuclear labelling is strong in (a) and (b).
Fig. 2 Immunohistochemical localisation of cyclin D1 (×200) in a breast carcinoma tissue section (a) and a vestibular schwannoma tissue section (b). Nuclear and cytoplasmic labelling intensity is strong in (a) and mild in (b).
Fig. 3 Immunohistochemical localisation of cyclin D3 (×200) in a breast carcinoma tissue section (a), and a vestibular schwannoma tissue section (b). Intensity of nuclear labelling is strong in (a) and mild in (b)
Distribution of cyclin D1 or cyclin D3 nuclear expression was graded as the following: <10 per cent (absent), 11–25 per cent, 26–50 per cent or 51–100 per cent.Reference Liu, Minin, Huang, Seligson and Horvath27, Reference Pan, Tang, Xu, Lu, Lin and Qiu28 For intensity of nuclear labelling of the cyclin D expression, a four-tier grading scale consisting of the following was used at high power magnification (×200) (field of view area: 0.15 mm2): 0 (absent), + (mild), ++ (moderate) or +++ (strong).Reference Pan, Tang, Xu, Lu, Lin and Qiu28 For cyclins D1 and D3, the immunoreactivity was predominately nuclear; however, mild cytoplasmic labelling was also observed in cyclin D1. Within each patient tumour sample, the percentage of cell labelling with maximum nuclear intensity (‘Pos-Max’) was scored as follows: 0 (absent), <10 per cent, 10–50 per cent or >50 per cent.
Proliferative activity
The Ki-67 proliferation index was determined using an Olympus X51 microscope at an original magnification of ×200. A cell was classified as positive when any degree of specific chromogen deposition was localised to the nucleus. The mean Ki-67 proliferation index was determined by counting the number of positive tumour cell nuclei in three chosen fields divided by the total number of cells in the same three fields (approximately 1000 cells in total).Reference Aguiar, Tatagiba, Dankoweit-Timpe, Matthies, Samii and Ostertag29 The three chosen fields of preference were the three highest areas of proliferation or Ki-67 staining.
Statistical analysis
The relationships between Ki-67 proliferation index, and cyclin D1 and cyclin D3 protein immunohistochemistry distribution and intensity were compared for the two age groups, using the independent samples t-test for continuous variables and the chi-square test for categorical variables. Differences were considered significant at a level of p < 0.05.
Results
In total, 180 patients (aged 54 years ± standard deviation (SD) 13.9; range, 14–81 years) were studied. Patients were divided into 2 age groups: 35 patients (18 males and 17 females) were ≤40 years of age and 145 patients (78 males and 67 females) were >40 years of age.
Proliferation index
The proliferation was statistically significantly higher index (p = 0.01) in the ≤40 years age group (mean 4.52 ± SD 2.56) compared to the >40 years age group (mean 3.27 ± SD 1.79). There was no statistically significant difference (p = 0.62) between male-to-female ratios in relation to proliferation index. Respective immunohistochemistry immunoreactivity data for Ki-67 are shown in Figure 1.
Cyclin D1 expression
Cyclin D1 was overexpressed in 68 per cent of the patients. There was overexpression of cyclin D1 immunoreactivity in 63 per cent of patients (22 out of 35) in the ≤40 years age group and in 69 per cent (100 out of 145) in the >40 years age group. The cyclin D1 immunoreactivity distribution scoring system is demonstrated in Figure 4a. There was no statistically significant difference (p = 0.49) in the distribution of labelling between the two age groups when negative labelling (≤10 per cent) was compared with positive labelling (>10 per cent).
Fig. 4 Grading scale of cyclin D1 distribution of labelling in a normal saphenous nerve tissue section magnified at ×100 (a), and vestibular schwannoma tissue sections magnified at ×40 showing: 11–25 per cent (b), 26–50 per cent (c) and 51–100 per cent grading (d).
The cyclin D1 nuclear labelling was dichotomised into low intensity (0, +) or high intensity (++, +++). Figure 5 summarises the grading scale used for intensity of nuclear labelling. There was no statistically significant difference in intensity of cyclin D1 nuclear labelling between the two age groups (p = 0.49).
Fig. 5 Grading scale of cyclin D1 (×200) intensity of nuclear labelling in a normal saphenous nerve (score 0) tissue section (a), and vestibular schwannoma tissue sections showing mild (+) (b), moderate (++) (c) and strong (+++) grading (d).
Similarly, the cyclin D1 protein Pos-Max was dichotomised into low per cent distribution (0, <10) and high per cent distribution (10–50, >50), with no statistically significant difference between the two groups (p = 0.54).
Tables I–III summarise the cyclin D1 protein immunoreactivity between the two groups.
Table I Summary of cyclin d1 protein immunoreactivity: distribution data
*Data represent numbers of patients; †data represent numbers (percentages) of patients
Table II Summary of cyclin d1 protein immunoreactivity: intensity of nuclear labelling data
*Data represent numbers of patients; †data represent numbers (percentages) of patients
Table III Summary of cyclin d1 protein immunoreactivity: pos-max data
*Refers to percentage of cell labelling with maximum nuclear intensity. †Data represent numbers of patients; ‡data represent numbers (percentages) of patients
Cyclin D3 expression
Cyclin D3 was overexpressed in 44 per cent of the patients. Cyclin D3 protein expression exhibited strong nuclear immunoreactivity (Figure 3b). There was overexpression of cyclin D3 immunoreactivity in 57 per cent of patients (20 out of 35) in the ≤40 years age group and 41 per cent (60 out of 145) in the >40 years age group (p = 0.09). The distribution of cyclin D3 labelling of tumours in the ≤40 years age group ranged from 11 to 100 per cent, compared to <26 per cent in all tumours in the >40 years age group.
Cyclin D3 was dichotomised into low intensity (0, +) and high intensity (++, +++) immunoreactivity. Figure 6 summarises the grading scale used for intensity of nuclear labelling.
Fig. 6 Grading scale of cyclin D3 (×200) intensity of nuclear labelling in a normal saphenous nerve (score 0) tissue section (a), and vestibular schwannoma tissue sections showing mild (+) (b), moderate (++) (c) and strong (+++) grading (d).
The Pos-Max was dichotomised in a similar manner to the cyclin D1 protein. There was a significant increase in Pos-Max cyclin D3 nuclear protein staining in the younger age group compared to the older age group (p = 0.02).
Tables IV–VI summarise the cyclin D3 protein immunoreactivity between the two groups.
Table IV Summary of cyclin d3 protein immunoreactivity: distribution data
*Data represent numbers of patients; †data represent numbers (percentages) of patients
Table V Summary of cyclin d3 protein immunoreactivity: intensity of nuclear labelling data
*Data represent numbers of patients; †data represent numbers (percentages) of patients
Table VI Summary of cyclin d3 protein immunoreactivity: pos-max data
*Refers to percentage of cell labelling with maximum nuclear intensity. †Data represent numbers of patients; ‡data represent numbers (percentages) of patients
Discussion
The loss of functional merlin leads to deregulation of a pathway associated with the contact-dependent inhibition of normal Schwann cells, ultimately leading to increased cell growth, tumour formation and tumour cell invasion. Merlin functions as a negative cell cycle regulator of the Rac-dependent signalling pathway.Reference Lutchman and Rouleau30, Reference Manchanda, Jones, Lee, Pringle, Zhang and Yu31
The regulation of merlin is complex and poorly understood, although studies suggest it may be dependent on its molecular conformation.Reference McClatchey and Fehon32, Reference LaJeunesse, McCartney and Fehon33 Merlin inactivation occurs when it is phosphorylated at S518 by p21-activated kinase (PAK), which is activated by Rac1 (ras-related C3 botulinum toxin substrate 1), interrupting the c-terminal domain and maintaining the folded state.Reference Kissil, Johnson, Eckman and Jacks34, Reference Okada, Lopez-Lago and Giancotti35 The disrupted folded state occurs by a folding of the alpha-helical portion and c-terminal portion of merlin, such that it blocks the four-point-one, ezrin, radixin, moesin (‘FERM’)-binding domain, which allows it to mediate cell–cell attachment, cell motility, membrane receptor availability and signal transduction.Reference McClatchey and Fehon32, Reference Okada, Lopez-Lago and Giancotti35–Reference Shimizu, Seto, Maita, Hamada, Tsukita and Hakoshima39 The Rac pathway has been associated with tumourigenesis, and merlin tumour suppressor function is likely mediated by its dephosphorylated accumulation, inhibiting Rac signalling.Reference Okada, Lopez-Lago and Giancotti35
Rac activates a variety of intracellular signalling pathways involved in transcriptional activation, transformation and proliferation. Downstream signalling regulated by Rac includes PAK, mitogen-activated protein kinase (MAPK), MET signalling, Jun-N terminal kinase (JNK), p38 and nuclear factor kappa beta (NF-κB).Reference Shrestha, Schafer, Boehm, Thomas, He and Du40, Reference Xiao, Chernoff and Testa41 Rac activation is important for progression of the G1/S (synthesis) phase transition of the cell cycle, which occurs through Rac's ability to propagate transcription, and translation of cyclin D1, driving the retinoblastoma protein (pRb)-cyclin-dependent kinase (CDK) pathway. In contrast, overexpression of merlin induces G1 cell cycle arrest, highlighting an important association with merlin and cyclin D1 regulation.Reference Xiao, Gallagher, Shetler, Skele, Altomare and Pestell6 In a previous study, the loss of functional merlin in vestibular schwannoma was associated with decreased p21 protein and mRNA levels, when compared to normal myelinated nerve, and subsequently elevated cyclin D1 protein.Reference Wu, Chen, Wang, Li, Li and Zhang42
Previous studies have indicated important roles for cyclin D1 protein in the regulatory control of Schwann cell proliferation.Reference Lassaletta, Del Rio, Torres-Martin, Rey, Patron and Madero21–Reference Neff, Oberstien, Lorenz, Chaudhury, Welling and Chang23 The current study is the largest to date investigating the expression of cyclin D1 and cyclin D3 in vestibular schwannoma. Cyclin D1 was overexpressed in 68 per cent of patients. This is consistent with a study of 64 sporadic vestibular schwannomas, in which 67 per cent of the tumours were positive.Reference Lassaletta, Del Rio, Torres-Martin, Rey, Patron and Madero21 Cyclin D3 immunoreactivity in the current study was overexpressed in 44 per cent of patients. Again this is comparable to previous findings; in a study of 15 vestibular schwannomas, cyclin D3 immunoreactivity was overexpressed in 50 per cent of tumours.Reference Neff, Oberstien, Lorenz, Chaudhury, Welling and Chang23 Our study provides further evidence to support the potential association between the overexpression of cyclin D1 protein and merlin deregulation.
While nuclear labelling was required to confirm positive tumour sections for D cyclins, cytoplasmic labelling was present in most tumour sections for cyclin D1. Cyclin D1 is regulated by phosphorylation at its threonine 286 site via glycogen synthase kinase 3β (GSK-3β), which allows binding of cyclin D1 with nuclear exportin CRM1, which places cyclin D1 in the cytoplasm for proteolysis.Reference Diehl, Cheng, Roussel and Sherr43 As such, we did not consider cytoplasmic positivity as significant.
Cyclin D3 protein is expressed in variable abundance in proliferating cell populations in several types of human tumours, including breast, colorectal, melanoma, and head and neck cancers.Reference Bartkova, Zemanova and Bartek44 The overexpression in human cancers is unlikely to be as frequent as that of cyclin D1 immunoreactivity in head and neck cancers, and the reason for this is not well understood. The functional role of cyclin D3 in vestibular schwannoma pathogenesis remains to be determined. Schwann cell cultures exposed to polypeptide growth factor heregulin and adenylyl cyclase activator forskolin increased steady state levels of CCAAT/enhancer binding protein-b (C/EBPb) in Schwann cells which cyclin D3 binds to.Reference Fuentealba, Schworer, Schroering, Rahmatullah and Carey45 In cultured melanoma cell lines, cyclin D3 overexpression is regulated by fibronectin-mediated phosphatidylinositol 3-kinase/AKT signalling, but not by the mitogen-activated kinase kinase (MEK) pathway.Reference Spofford, Abel, Boisvert-Adamo and Aplin46 In the same study, RNA interference experiments demonstrated that cyclin D3 contributed to G1/S cell cycle progression and proliferation. In our study, the younger age group had a higher expression of cyclin D3 compared to the older age group, but only when assessed by the Pos-Max method.
Previous imaging and clinical studies have shown a statistically significant higher tumour growth rate in younger individuals (≤40 years) compared to those who are older (>40 years).Reference Charabi47, Reference Stangerup, Tos, Caye-Thomasen, Tos, Klokker and Thomsen48 This is the first study to demonstrate an association between patient age and tumour proliferation index as measured by Ki-67. Cyclin D1 and cyclin D3 expression in this study did not correlate with the Ki-67 proliferation index, which is consistent with previous studies.Reference Oyama, Kashiwabara, Yoshimoto, Arnold and Koerner12, Reference Shoker, Jarvis, Davies, Iqbal, Sibson and Sloane49
• Loss of functional merlin is common in both sporadic and neurofibromatosis type 2 related vestibular schwannomas
• This loss contributes to cell cycle pathway deregulation, with loss of Rac signalling inhibition
• This may lead to cyclin D protein overexpression, promoting vestibular schwannoma growth
• Younger patients have increased Ki-67 proliferation index compared to older patients, corresponding to increased tumour diameter and growth
• Overexpression of cyclin D1 and cyclin D3 was found in 68 and 44 per cent of tumours, respectively
• Elucidating molecular targets in vestibular schwannoma will aid treatment, enabling residual disease and radioresistant tumour morbidity to be addressed
A better understanding of the tumour biology of vestibular schwannoma will enhance the management paradigm of this condition. Identifying predictors of growth may in future enable more selective management strategies for patients with this condition. Furthermore, this understanding is critical if potential new targeted molecular therapies are to be considered as treatment modalities in the future. Studies are currently underway to further evaluate the role of multiple molecular proteins and determine their importance in vestibular schwannoma pathogenesis.
Conclusion
Vestibular schwannomas show a small but statistically significant increase in Ki-67 proliferation index in patients ≤40 years old compared to those aged >40 years. Overexpression of cyclin D1 and D3 proteins is a common feature in vestibular schwannomas, and may have a role in tumour biology. Further investigations of the Rac signalling pathway in vestibular schwannoma are indicated, and may lead to the identification of molecular targets to augment current treatment modalities.
Acknowledgements
Much appreciation is given to all scientists and technical assistants at SydPath (Department of Anatomical Pathology at St Vincent's Hospital, Darlinghurst, Australia) for their assistance. This work was supported by the St Vincent's Hospital Darlinghurst Microsurgical Skills Lab Ear Trust Fund and funding from SydPath.
