A role for the p53 pathway in the pathology of meningiomas with NF2 loss

Abstract

The neurofibromatosis 2 locus (NF2) is inactivated through mutation and loss of heterozygosity (LOH) in 40–65% of all sporadic meningiomas, while the role of the p53 tumor suppression pathway in meningioma initiation and progression is still unclear. This study aims to determine if a p53 codon 72 arginine-to-proline polymorphism, found to be correlated with cancer development and cancer patient survival in other tumors, is associated with sporadic meningioma initiation or progression. We investigated Pro72 incidence in a cohort of 92 sporadic meningiomas and analyzed its association with histological grade (WHO classification) and with NF2 LOH (determined using polymorphic microsatellite markers on 22q). The Pro72 allele was not found to be selected for in the cohort. However, in the subgroup of meningiomas with NF2 LOH and carrying Pro72, 50.0% had high grade tumors (WHO grades II and III) compared to only 14.3% of those without NF2 LOH (OR = 6.0, CI = 1.56–23.11, P = 0.012). The significant association occurred only when considering subgroups of meningiomas with or without NF2 LOH, suggesting that not including NF2 status when analyzing study cohorts may explain the variability seen in the literature where all meningiomas were grouped together. Our data suggests a role for the p53 pathway in the progression of meningiomas in which NF2 is inactivated, and highlights the importance of accounting for NF2 LOH in future studies of meningiomas and the p53 pathway.

Introduction

Arising from the arachnoid coverings of the brain and spinal cord, meningiomas account for 15–20% of primary neoplasms of the central nervous system, and have an estimated annual incidence of about 6 per 100,000 individuals [1]. Loss of heterozygosity (LOH) of chromosome 22, containing the NF2 gene, has been found in 40–65% of all meningiomas [2-4]. NF2 mutations are found in up to 60% of meningiomas and are usually associated with NF2 LOH [3, 5], suggesting that LOH of NF2 is an early genetic event in meningioma pathogenesis. In addition to meningioma formation, NF2 inactivation has also been associated with later tumor progression to higher grades [6, 7], but this pattern of tumor development is not present in all tumors since the majority of meningiomas still occur as benign tumors [8]. These results suggest that although about half of meningiomas are associated with NF2 LOH, only a small group of tumors are at risk for tumor progression. Our hypothesis is that other underlying genetic factors exist that are responsible for meningioma progression.

TP53, the gene encoding p53, mediates a major tumor suppression pathway of the cell. Since the first report in 1989 [9], p53 mutations have been found in approximately half of all cancers [10, 11]. Despite this clear prominence of the p53 pathway in cancers, conflicting results are present in the literature over the role of the p53 pathway in meningiomas. Several lines of evidence suggest that the p53 pathway is not important in meningioma etiology such as the absence of TP53 mutations in meningiomas [12-14], the lack of abnormal expression of p53 or Mdm2 in meningiomas or meningioma cell lines [15], and the similarity of p53 expression in orbital meningiomas of different histological grades [16]. Meanwhile, other studies suggest the involvement of the p53 pathway in meningioma development: the correlation of p53 protein expression with histological tumor grade and meningioma recurrence [17], the loss of detectable Mdm2 protein in high grade meningiomas [18], and a defective p53 response to gamma ray stress in meningioma cells [19]. In addition, the NF2 protein product was reported to increase p53 stability through downregulating Mdm2 levels in mouse fibroblast [20]. It follows that loss of NF2 may increase the likelihood of p53 suppression, thus decreasing tumor suppression activity and providing a possible mechanism for the involvement of the p53 pathway in meningiomas. We hypothesize that a genetic factor that modulates the activity of the p53 pathway may help regulate meningioma pathogenesis.

The proto-oncogene product Mdm2, an E3 ubiquitin ligase, acts to both repress p53's transcriptional activity in the nucleus [21, 22] and induce p53 degradation via the ubiquitin-proteasome system [23-25]. Excessive p53 downregulation by Mdm2 has been associated with accelerated tumor growth and metastatic cancer phenotypes [26-29]. Contrastingly, Mdm2 has also been implicated to be necessary for a tumor suppressive function of p53. Mdm2-mediated monoubiquitylation of p53 has been found to be necessary for p53 translocation from the nucleus to the mitochondria, where the p53 induces apoptosis in a stressed cell [30-32].

The p53 codon 72 arginine-to-proline polymorphism alters a cell's apoptotic potential via differential interaction with Mdm2 [33]. The Arg72 allele has been found to have increased interaction with Mdm2, causing enhanced shuttling to the mitochondria and induction of transcriptionally independent apoptosis [33]. Meanwhile, the Pro72 allele, which is found to be correlated with cancer development [34] and decreased cancer patient survival [35], is associated with decreased translocation of p53 from the nucleus [33]. In this case, p53 interaction with Mdm2 would suppress p53 and cause tumor development. Since NF2 has been found to inhibit Mdm2 [20], we hypothesize that NF2 LOH in meningiomas increases Mdm2 function and allows for subsequent tumor progression to occur with additional presence of the Pro72 allele.

Materials and methods

Subjects and sample collection

This study was approved by the Institutional Review Board of Massachusetts General Hospital/Partners HealthCare and informed written consent was obtained from all individuals donating tissue. Between August 2000 and December 2004, 142 meningioma patients had tumors resected at Massachusetts General Hospital and consented to banking of excess tumor tissues for research. From the 142 patients, 50 were excluded from this study due to lack of adequate blood or tumor DNA for analysis or presence of multiple intracranial tumors suggestive of Neurofibromatosis 2 or the syndrome of multiple meningiomas. Medical records review was performed using the Research Patient Database Registry, an internal research tool providing access to demographic information, radiology reports and pathology reports. Blood and tumor DNA were extracted using a standard extraction kit as previously described [36].

NF2 loss of heterozygosity

Loss of heterozygosity (LOH) of the NF2 locus was determined as previously described [36]. Individuals heterozygous at NF2, as determined by analysis of blood DNA, were checked for loss of heterozygosity in tumor DNA at markers D22S193 (centromeric to NF2), NF2TET and D22S929 (intragenic), and D22S268 and D22S430 (telomeric to NF2).

p53 codon 72 polymorphism analysis

Determination of the polymorphism at codon 72 used a modification of the protocol described by Fan et al. [37] as follows. A 150 bp fragment containing the polymorphism was amplified in 100 ng of genomic DNA using the oligo-nucleotide primers 5′-AGATGAAGCTCCCAGAATGC-3′ (forward) and 5′-GTAGGTTTTCTGGGAAGGGA-3′ (reverse). A 10 μl total volume PCR contained 4 pM of each primer, 0.2 mM of each dNTP, 2 μl of Q solution, 1 μl of 10X buffer, and 0.25 U of Taq polymerase (Qiagen, Valencia, CA). Initial denaturation was carried out at 95°C for 5 min, followed by 35 cycles of 95°C for 30 s, 57°C for 30 s, 72°C for 1 min, and a final extension of 7 min at 72°C. PCR products were then digested using the endonuclease BstUI in 30 μl volumes according to the manufacturer's specifications (New England Biolabs, Beverly, MA). The restriction fragment length polymorphism was detected on an 8% polyacrylamide gel and visualized using ethidium bromide. No restriction site is present in the Arg72 allele, while the 150 bp PCR product is digested into two fragments of 114 and 36 bp in the Pro72 allele.

Statistical analysis

Histological grades of tumors were available from the pathology reports as assessed by the pathologist according to the criteria of WHO 2000 [38]. For the purpose of statistical analysis, atypical (WHO grade II) and malignant (WHO grade III) meningiomas were grouped together as high grade meningiomas and compared to low grade tumors (WHO grade I). Analysis of the p53 codon 72 polymorphism was performed by combining those patients heterozygous and homozygous for Pro72 into one single group whenever the group with homozygous Pro72 was too small to be analyzed independently. Associations between two dichotomous variables, for example, p53 codon 72 and NF2 LOH, were tested using Fisher's exact test (two tailed). The corresponding odds ratios are conditional maximum likelihood ratio estimates. Tests were conducted at the P = 0.05 level of significance and confidence intervals are reported at the 95% level.

Results

Cohort characteristics are delineated in Table 1. Similar to previous findings, a 2.4:1 female to male ratio was found [39]. Meanwhile, a higher typical incidence of high grade tumors is present (29.3%; 27/92) [8], possibly because Massachusetts General Hospital is a tertiary care center. In our cohort, high grade tumors were seen in 24.6% (16/65) of women and 40.7% (11/27) of men (odds ratio (OR) = 2.1, confidence interval (CI) = 0.81–5.46, P value = 0.138). LOH of the NF2 locus was seen with an incidence of 52.1% in the cohort, similar to previously reported values. NF2 LOH was also significantly associated with high tumor grade, with 41.7% (20/48) of tumors having NF2 loss being high grade tumors, compared to 15.9% (7/44) of those retained at the NF2 locus (OR = 3.8, CI = 1.40–10.17, P = 0.011). However, when segregating by gender, the significance of this association was restricted to tumors from men. More than half (55.6%; 10/18) of the meningiomas from men with NF2 LOH were high grade, while only 11.1% (1/9) of the tumors from men without NF2 LOH were high grade (OR = 10.0, CI = 1.03–97.50, P = 0.042). Such a difference was not found to be significant among women (P = 0.157). LOH of the NF2 locus was also associated with an early average age of tumor resection in men (52.3 years for tumors lost at NF2 compared to 65.8 years for tumors without NF2 LOH, unpaired Student's t-test, P = 0.02); this result was not found among tumors from women (59.2 years for tumors with NF2 LOH compared to 55.7 years for tumors without NF2 LOH). The reason for the increased prevalence of meningiomas among women has not yet been identified, but our results suggest that NF2 LOH in meningiomas is more common among men, and is also associated with high tumor grade and earlier age of onset in a gender specific manner.

Table 1.

Demographic and clinical characteristics of the sporadic meningioma patient cohort

Sex	Men: 27	Age of tumor resection (yrs)	57 ± 12.5
	Women: 65
Ethnicity	Caucasian: 83	Tumor grade	Benign: 65
	African: 2		Atypical: 23
	Asian: 2		Malignant: 4
	Unrecorded: 5	NF2 status	Loss: 48
			Retained: 44

Data involving p53 codon 72 are presented in Table 2. Distribution of the p53 codon 72 polymorphism among meningioma cases was found to be consistent with Hardy–Weinberg equilibrium, with allelic frequencies of 45.6% (Arg72/Arg72), 44.6% (Arg72/Pro72), and 9.8% (Pro72/Pro72), and did not significantly differ from previous reports using primarily Caucasian control groups [40, 41]. This supports previous studies that found no association between p53 polymorphisms and meningiomas in Nordic, British, and Slavic cohorts [40, 42, 43]. In our patient cohort, LOH of the NF2 locus in tumors was seen in 61.9% (26/42) of patients carrying the Arg72/Arg72 genotype compared to only 44.0% (22/50) of patients carrying at least one Pro72 allele, but without reaching statistical significance (OR = 2.1, CI = 0.90–4.77, P = 0.098). Age of tumor resection was not found to differ between codon 72 genotypes even after considering patients' gender and NF2 status in the tumor (data not shown).

Table 2.

Distribution of the p53 codon 72 polymorphism among sporadic meningioma patients

p53 codon 72 alleles	NF2 status	Women			Men			All
		High grade	Low grade	Total	High grade	Low grade	Total	High grade	Low grade	Total
Arg72/Arg72	NF2 LOH	5	11	16	4	6	10	9	17	26
	NF2 retained	2	9	11	1	4	5	3	13	16
	Total	7	20	27	5	10	15	12	30	42
With Pro72 [percentage of row total for (2 × 2) values]	NF2 LOH	5	9	14	6	2	8	11 (50)	11 (50)	22
	NF2 retained	4	20	24	0	4	4	4 (14.3)	24 (85.7)	28
	Total	9	29	38	6	6	12	15	35	50
All [percentage of row total for (2 × 2) values]	NF2 LOH	10	20	30	10 (55.6)	8 (44.4)	18	20 (41.7)	28 (58.3)	48
	NF2 retained	6	29	35	1 (11.1)	8 (88.9)	9	7 (15.9)	37 (84.1)	44
	Total	16	49	65	11	16	27	27	65	92

Italicized values (2 × 2) were found to have P < 0.05 with the Fisher's exact test

To determine the effect of the polymorphism on meningioma progression we compared the frequency of high-grade tumors among the different genotypes. There was no association between the overall polymorphism distribution and tumor grade, with high grade tumors accounting for 28.6%, 31.7%, and 22.3% of the patients carrying the Arg72/Arg72, Arg72/Pro72, and Pro72/Pro72 genotypes, respectively. When considering the effect of NF2 LOH on tumor grade in patients carrying at least one Pro72 allele (Pro72/Pro72 or Arg72/Pro72), we found that patients with NF2 LOH in the meningioma developed high grade tumors in 50.0% (11/22) of the cases, compared to only 14.3% (4/28) of patients without NF2 LOH (OR = 6.0, CI = 1.56–23.11, P = 0.012). In contrast to this association found among patients with the Pro72 allele, no significant association was found among patients with the Arg72/Arg72 genotype (OR = 2.3, CI = 0.52–10.20, P = 0.316), demonstrating a change correlated with the p53 codon 72 genotype.

Discussion

This study of a human cohort presents bipartite findings involving NF2 LOH in meningiomas. First, NF2 LOH has been found to be associated with both high tumor grade and an early average age of tumor resection among men, but not women. Second, aggressive meningiomas were found to be associated with NF2 LOH only among individuals with the Pro72 allele of the p53 codon 72, but not among those with the Arg72/Arg72 genotype.

Meningiomas have long been known to exhibit receptors to sexual steroids [44], suggesting that hormonal influences may explain the sexually dimorphic characteristics of this disease. However, discrepancies on the proliferative effects of sexual hormones on meningiomas and failure of variations in sexual receptor expression to explain the increased prevalence of meningiomas in women suggests that more complex factors are at play [45]. Though gender differences are clearly observed in meningiomas, their cause(s) still remain elusive. The fact that NF2 has not been independently linked to any sort of sexual dimorphism suggests that some additional mechanism of meningioma development may occur in women. If such a mechanism were distinct from NF2 LOH, it serves as one rationalization for why our first set of findings might occur with significance only among men.

The distribution of the p53 codon 72 polymorphism among our meningioma patients was not significantly different from that found in previously reported control groups [40, 41], suggesting that alteration of the p53 pathway by a change at codon 72 does not affect meningioma initiation. The findings in our study of a human cohort are consistent with a previous study of an NF2^flox2/flox2 meningioma mouse model, which received adCre injections into the cerebral spinal fluid to target NF2 at the level of the leptomeninges. When introducing the additional genetic background of p53 hemizygosity into the mice, no increase in the rate of meningioma formation was found [46].

When considering the additional aspect of meningioma progression, we found that the presence of the Pro72 allele predisposed to an increased relevance of NF2 LOH in affecting tumor progression. In accord with several previous studies [18, 19, 47], our findings support a role for the p53 pathway in the progression of meningiomas, particularly among those having NF2 LOH. The study of the NF2^flox2/flox2 meningioma mouse model had also considered the possible effect of p53 hemizygosity upon meningioma progression, but was unable to observe any differences [46]. The p53 hemizygous mice developed aggressive sarcomas that necessitated premature sacrifice, possibly masking a contribution of p53 inactivation on meningioma progression.

NF2 is one of several proteins known to downregulate the inhibitory action of Mdm2 on p53 [20, 48-50]. Thus, a potential model for the role of NF2 LOH in meningiomas is that depletion of NF2 increases Mdm2 levels, thereby decreasing the activity of the p53 tumor suppression pathway and predisposing to tumor progression. Hence, we find increased tumor progression susceptibility among meningiomas that exhibit NF2 LOH as an early genetic event. In accord with a previous study [18], our findings suggest that increased susceptibility of meningioma progression can arise through abnormally strong inhibition of the p53 pathway caused by loss of a factor that inhibits Mdm2-mediated negative regulation of p53. Since our findings depend on segregating the cohort according to NF2 status, perhaps the inadvertent merging (into one group) of meningiomas with NF2 LOH and those arising via other mechanisms has masked this relevance of the p53 pathway in previous studies that have suggested that alterations in the p53 pathway do not appear to contribute to the risk of sporadic meningioma formation or progression [15, 16, 40, 42, 43].

Using inducible cell lines, the Arg72 allele of p53 codon 72 was previously reported to have superior binding to Mdm2 as compared to the Pro72 allele. This was associated with enhanced tumor suppressive activity via Mdm2-dependent localization of p53 to the mitochondria for induction of transcriptionally independent apoptosis [33]. Meanwhile, the Pro72 allele remains more in the nucleus and must therefore mediate tumor suppression via transcriptional means. Interaction of this Pro72 allele with Mdm2 would therefore act to inhibit the p53 activities and increase the possibility of tumor development. Our study finds a selection for NF2 LOH in high grade tumors among meningiomas with the Pro72 allele, suggesting that the primary mechanism of p53-mediated tumor suppression in meningiomas with NF2 LOH might be the one associated with the Arg72 allele: mitochondria mediated apoptosis. A decreased activity of this mechanism due to the presence of the Pro72 allele makes the meningioma more likely to progress to higher grade. Further experiments are necessary to test these hypotheses.

In summary, this study supports the importance of NF2 LOH in affecting the nature of meningioma pathology and prognosis. With a mechanistic perspective of the p53 codon 72 polymorphism, this study finds evidence for a two-tiered mechanism in the formation of high grade meningiomas: after NF2 LOH occurs in tumorigenesis, the p53 pathway, when provoked, induces protective apoptosis through action at the mitochondria; polymorphism variants that disrupt the balance of components in this branch of the p53 pathway increase the risk of progression to high grade tumors. Further study and biochemical work is needed to fully understand the factors at work here, but our initial results suggest that future rational therapies for those meningiomas with NF2 inactivation might include options that impact the p53 pathway.

We are aware of the limited number of samples in our cohort, so conclusions in this study are made with caution. Nevertheless, the findings contribute to a better understanding of meningiomas pathogenesis, particularly on the role of NF2 LOH with regard to gender and the p53 pathway. Altogether, the data offers the suggestion that meningiomas arising with and without NF2 LOH may perhaps present two different molecular microenvironments having the same deleterious macroscopic manifestation.