TERT promoter mutations in primary and secondary WHO grade III meningioma.

PURPOSE: TERT promoter mutation (TERTpMut ) has a strong association to recurrence and has been suggested to act as a driver mutation for malignant transformation of WHO grade I and II meningiomas. TERTpMut has been investigated in selected high-grade meningioma samples. The existence of TERTpMut across recurrent tumors in a population-based cohort needs to be investigated in order to identify when TERTpMut emerges across recurrent samples and to validate prognostic impact among WHO grade III tumors.
METHODS: We gathered material from a consecutive single-center cohort of 40 patients with malignant meningioma (WHO grade III) treated between 2000 and 2018, including specimens from primary and secondary malignant meningiomas with the corresponding earlier benign specimens and later malignant recurrences. In total 107 tumor samples were studied by Sanger sequencing for TERT promoter mutational status.
RESULTS: Seven of 40 patients (17.5%) harbored TERTpMut thus validating the incidence of TERTpMut in previous non-population-based cohorts. In 6/7 patients, the TERTpMut was present at initial surgery (WHO grade I-III) while in one patient the TERTpMut was found de novo when the meningioma became malignant. The incidences were 2/1.000.000/year for TERTpMut WHO grade III meningioma and 8/1.000.000/year for TERTpwt WHO grade III meningioma in our catchment area. We found a 1.7 times higher recurrence rate (CI 95% 0.65-4.44) and a 2.5 higher mortality rate per 10 person-years (CI 95% 1.01-6.19) for TERTpMut compared to TERTpwt .
CONCLUSION: TERTpMut can occur independently of malignant progression in meningioma and was most often present from the first tumor sample across recurring tumors. TERTpMut in WHO grade III may represent a marker of an aggressive subset of tumors.

Introduction

Meningiomas account approximately for 37% of all intracranial tumors (17). A subset of meningiomas, WHO grade III, are malignant. These tumors comprise approximately 1%–3% of all primary meningiomas, with a dismal prognosis with significant morbidity and mortality (9, 17). Better knowledge of pathophysiology is desirable to improve management of these patients, as hitherto published retrospective data show limited efficacy of aggressive surgery and adjuvant therapy once malignant criteria are met (18).

WHO grade III meningiomas comprise of primary malignant meningiomas and secondary that develop through malignant transformation of grade I or II meningiomas. Previous reports have described CDKN2A/B loss (4), accumulation of chromosome gains and losses (6), and telomerase reverse transcriptase (TERT) gene alterations (5). TERT gene alterations (TERT‐alt) may enforce cell immortalization by counteracting telomere shortening, thus promoting growth (1). TERT‐alt comprise, but are not limited to, promoter mutations, gene translocations and DNA amplifications (1). The most common alterations are specific point mutations (C250T and C228T) of the TERT promoter region (TERTpMut) (24). TERT promoter mutations either occur already in initial low‐grade meningiomas or emerge during malignant transformation of meningiomas (5). A recent meta‐analysis of individual patient data including 677 patients with grade I–III meningiomas provided evidence that TERT‐alt is a negative prognostic biomarker independent of the tumors’ WHO‐grade (11). The previous studies (8, 18, 20, 21) included 21–58 patients possessing WHO grade III meningiomas with 14%–23% expressing TERTpMut. TERTpMut was reported to accumulate in grade III tumors and thought to contribute to malignant progression. TERTpMut was a negative prognostic factor among grade III tumors in one study (20). The previous studies were neither population based nor included consecutive patients with clinical follow‐up. Only one study (18) compared expression of TERTpMut in primary (5/28) and secondary malignant meningiomas (3/29), but did not include preceding grade I–II samples from the secondary grade III meningiomas. Observational data do not indicate TERTpMut as a common cause for WHO grade shifts although accumulation in higher grades has been observed and considered a possible driver mutation (20). Serial samples that include pre‐malignant meningiomas of patients with secondary grade III tumors allow the analysis of how TERTpMut is associated with the shift to a grade III tumor.

We have investigated the frequency of the two point mutations C228T and C250T (TERTpMut) in a consecutive cohort of 40 WHO grade III meningiomas and occurrence of TERTpMut in a series of tumor samples obtained during malignant transformation of secondary WHO grade III meningiomas.

The aims were to validate the previously detected 14%–23% expression of TERTpMut in WHO grade‐III meningiomas in a consecutive, population‐based clinical material and to analyze detection of TERTpMut in longitudinal series of tumor samples from patients with secondary and primary WHO grade III meningiomas.

Materials and methods

Patient cohort

We recruited patients diagnosed and treated for WHO grade III meningioma between October 2000 and September 2018 at the Departments of Neurosurgery and Pathology at the Copenhagen University Hospital Rigshospitalet in Denmark (catchment area 2.2 million). We identified 40 consecutive patients, a subset of whom have been clinically described in a previous report (12). The linkage between Danish social security system (CPR number) and WHO grade III specimen numbers in the Danish Pathology Biobank enabled us to localize serial tumor tissue from initial as well as all subsequent surgeries for all 40 patients, in total 119 potential paraffin‐embedded formalin‐fixed tumor specimens. Tissue was successfully retrieved from 108/119 (91%) specimens in the pathology archive. The remaining specimens had either been destroyed or lost. A senior consultant in neuropathology (HB) classified each specimen according to the WHO 2016 classification (11) before moving further with the TERT analysis. After TERT analysis, the slides were reviewed again in order to identify histopathological distinctions in TERTpMut tumors (histological subtypes, necrosis, brain invasion, mitoses).

Ethical approval for the current study was obtained by the Danish Ethics Committee (approval number H‐6‐2014‐010).

Statistics

We applied time since first WHO grade III meningioma to radiologically verified recurrence according to the RANO criteria (23) or death as underlying time scale, whichever appeared first. The term “malignant progression” is used for WHO grade I/II tumors switching to grade III. Tumor “recurrence” was used to denote radiological progression or reappearance of a previously operated tumor. For investigation of differences in age and sex in TERTpMut and TERTpwt groups, we applied a t‐test and Fisher’s test, respectively.

For TERTpMut versus TERTpwt, we (i) applied the Aalen–Johansen estimator and Gray’s test to investigate the cumulative incidence of recurrence when considering death without recurrence as a competing risk; and (ii) estimated the recurrence and mortality rate per 10 person‐years.

We applied Cox proportional hazards regression analysis to investigate the effect of TERTpMut on death. The Cox regression was adjusted to age at diagnosis. We tested assumption of proportionality by investigating Schoenfeld residuals and found that the assumption was valid. Age at diagnosis was included as a continuous covariate, in which we found a linear association to be adequate after applying a restricted cubic spline regression (χ 2, P > 0.05). The probability of overall survival was estimated with Kaplan‐Meier curves, and we tested whether the curves were significantly different with a log‐rank test. We considered P < 0.05 significant. For computing, we used the open‐source software R version 3.7.1 to all analyses.

Mutational analysis

DNA extraction

DNA was extracted from 4 × 10 micrometer paraffin‐embedded formalin‐fixed tissue following the standard protocol. In short, TRIS buffer (TRIS 10 mM + 0.1 mM EDTA, pH 7.4) was added to the tissue and incubated at 95°C for 10 min. The samples were subsequently cooled down to 56°C and 20 µL of Proteinase K (Qiagen, Denmark) was added. The samples were then incubated overnight at 56°C. The following day, Proteinase K was inactivated by incubation for 10 min at 95°C. We measured DNA concentration with NanoDrop (ThermoFisher Scientific, Denmark). One of the 108 samples was excluded because of inadequate DNA quality and the amount leaving the total amount of samples available for Sanger sequencing was found to be 107.

Sanger sequencing

Bidirectional Sanger sequencing of the TERT promoter region including the two mutations, C228T and C250T, and was performed using previously applied methodology (10). In short, the target region was amplified by conventional PCR using the previously reported primers 5′‐CACCCGTCCTGCCCCTTCACCTT‐3′ (sense) and 5′‐GGCTTCCCACGTGCGCAGCAGGA‐3′ (antisense) in 20 µL reactions using the Platinum™ II Hot‐Start PCR Master Mix (Thermo Fisher Catalog #:14000013) with GC‐enhancer. The amount of DNA was between 10 ng and 100 ng depending on the quality. PCR conditions consisted of a Touch‐Down protocol with annealing temperatures ranging from 60 to 50°C. ExoSAP‐IT PCR Product Cleanup Reagent (Thermo Fisher) was used for PCR purification. The sequences were generated by Sanger sequencing at the KiGene core facility, Karolinska Institutet. Samples were sequenced with sense primer initially and subsequently verified with antisense primer in selected cases. All mutations and aberrations at the hotspots were confirmed and re‐analyzed with both sense and antisense primers.

Results

The diagnoses of 40 consecutive grade III meningiomas during 18 years in a catchment area of 2.2 million inhabitants correspond to an incidence of 10.1/100.000/year. Among our 40 patients with WHO grade III meningiomas, we found TERTpMut in seven patients (17.5%): 2/20 (10%) primary grade III meningiomas and 5/20 (25%) secondary grade III meningiomas. The patient and treatment characteristics of the TERTp and the TERTp groups are shown in Table 1. There was no significant difference in age (TERTpMut: mean 62 years, TERTpwt: mean 59, t‐test P = 0.68). Only one female belonged to the TERTpMut group (1/7 14.3%) compared to 19 in the TERTpwt group (19/33, 57.6%). The difference was not statistically significant (Fisher test; P = 0.09). Twenty patients (50%) had a primary malignant meningioma and 20 a secondary. Of the 40 WHO grade III meningiomas, four tumors were papillary (two primary and two secondary), four were rhabdoid (three primary and one secondary) and 32 were anaplastic (15 primary and 17 secondary). TERTpMut tumors of all grades did not show any distinctive histopathological features compared to the TERTp group, apart from TERTpMut being found solely in patients whose tumors were classified as anaplastic meningiomas.

Table 1

Patient and treatment characteristics of the TERTpMut and the TERTpWt group.

	TERTpMut (n = 7)	TERTpWt (n = 33)
Age at first WHO grade III diagnosis (mean in years ± SD (range))	62 ± 15 (32–81)	59 ± 16 (10–87)
Female	1 (14%)	19 (58%)
Number of surgeries, all grades (mean, range)	4 (1–11)	3 (1–8)
Tumor size of the first WHO gr. III (widest diameter in mm ± SD (range))*	56 ± 20 (16–88)	49 ± 16 (40–96)
Time from first grade III to death or end of follow‐up July 2019 in months (median, (range))	20 (0,3–80)	38 (0,5–201)
Location (first WHO grade III tumor)
Convexity	4 (57%)	17 (52%)
Falcine/parasagittal (non‐convexity)	1 (14%)	9 (27%)
Skull base	2 (29%)	4 (12%)
Intraventricular	0 (0%)	3 (9%)
Simpsongrade (first WHO grade III)
Grade I	1 (14%)	8 (24%)
Grade II	1 (14%)	13 (39%)
Grade III	2 (29%)	6 (18%)
Grade IV	2 (29%)	4 (12%)
Grade V	0 (0%)	0 (0%)
Missing data (Simpson grade I–III)	1 (14%)	2 (6%)
Received adjuvant radiotherapy (any tumor)	86%	67%

Not all preoperative MR scans were available for the first WHO grade III tumor. In the TERTpMut group, the tumor size was derived from the first meningioma (WHO grade II) in one patient and from the seventh meningioma, WHO grade III, in another patient. In TERTpWt the first meningioma (WHO grade II) was used in three patients, and in four patients the last WHO grade III preoperative scan was used.

Primary malignant meningiomas (n = 20)

The two patients with papillary and three patients with rhabdoid histology all had meningiomas containing TERTp. Two patients with primary anaplastic meningiomas (patient #1 and #7) had TERTpMut. Both had a C228T TERTpMut and none had the C250T mutation (Figure 1). Patient #1 was only operated once for the WHO grade III meningioma while patient #7 was operated for a grade III recurrence that retained the C228T TERTpMut (Figure 1).

Figure 1

TERT promoter mutations (TERTpMut), C228T and C250T, during malignant degeneration in the seven patients with TERTpMut from our cohort of 40 WHO grade III meningioma.

Secondary malignant meningiomas (n = 20)

The two patients with papillary and one patient with rhabdoid histology had meningiomas containing TERTp. Five patients with anaplastic meningioma had TERTpMut. The first tumor sample from one of these patients (patient #7) with a C228T TERTpMut contained, however, papillary elements while subsequent tumor samples were purely anaplastic.

Four patients had a C228T (patient #2, 3, 5 and 6) and one had a C250T (patient #4) mutation (Figure 1). The TERTpMut was present already in the first available meningioma sample (WHO grade I meningioma for patient #3 and grade II meningiomas for patients #2, 4 and 5) in four patients. The same mutation was detectable in samples from 2, 4 and 1 subsequent surgeries in patients #2, 4 and 5, respectively. One patient (patient #6) underwent five surgeries for a WHO grade I tumor where the first tumor sample was unavailable and four subsequent tumor samples showed TERTpwt. Material was again unavailable for five subsequent operations for WHO grade II tumors, while the final operation, which revealed a grade III meningiomas, showed TERTpMut.

One patient changed from TERTpMut to TERTpwt after the first surgery: Patient #3 had a C228T mutation in the tumor sample from the first surgery of a WHO grade I meningioma, while the first subsequent recurrence, which had progressed to WHO grade III, showed TERTpwt. Later recurrences of the grade III tumor were again TERTpMut (Figure 1). We reviewed the tumor blocks again from the TERTpwt tumor from patient #3 (two large pieces) and concluded that the sequenced material was representable for the tumor sample.

Recurrence and mortality in TERTpMut and TERTpwt patients

The cumulative incidence of recurrence after first WHO grade III diagnosis was 33% (CI 95%: 20–52) for the TERTpwt group and 43% (CI 95%: 16–82) for TERTpMut group after 1 year. Gray’s test showed no significant difference between cumulative incidence of recurrence in the TERTpMut and TERTpwt group (P = 0.48) (Figure 2). In contrast, incidence of recurrence was 6.8 (CI 95% 5.0–8.7) per 10 person‐years for TERTpMut patients compared with 4.0 for TERTpwt patients (CI 95% 3.5–4.5), translating into a 1.7 higher recurrence rate in the TERTpMut group (Figure 3).

Figure 2

Cumulative incidence of recurrence in TERTpMut and TERTpwt in our cohort of 40 WHO grade III meningioma (with death without recurrence as a competing risk).

Figure 3

Incidence rates (events of recurrence and mortality) per 10 person‐years for TERTpMut and TERTpwt. Incidence rate ratio for recurrence 1.7 (6.8/4) and for mortality 2.5 (3.7/1.5).

Cox proportional hazard regression analyses did not show a statistically significant effect on overall survival by TERT promoter status. Adjusted to age at diagnosis, the hazard ratio for TERTpMut patients (n = 7) was 1.9 (CI 95% 0.7–5.4) in reference to their wild‐type counterparts (n = 33). TERTpwt patients had a median survival of 52.0 months (CI 95% 19.9—not reached) compared with TERT Mut patients with a median survival of 20.4 (CI 95% 6.67—not reached) (Figure 4). A log‐rank test indicated no significant difference (P = 0.2). In contrast, the mortality incidence rate was 3.7 for TERTpMut patients (CI 95% 2.8–4.7) per 10 person‐years, and 1.5 in TERTpwt patients (CI 95% 1.3–1.7), yielding a 2.5 higher mortality rate in the TERTpMut group (Figure 3).

Figure 4

Kaplan–Meier Survival curve. Overall survival for TERTpMut and TERTpwt WHO grade III meningioma. P value of log‐rank test = 0.2.

Discussion

Main findings

We found TERTpMut in seven patients, 2/20 with primary and 5/20 with secondary, WHO grade III meningiomas. Already the first, premalignant, sample from 4/5 patients with secondary malignant meningiomas harbored TERTpMut. Our findings corroborated previous reports (5, 21) of C228T being the most common TERTpMut in meningioma, as we found C228T in six of the seven patients.

The proportion of TERTpMut in this population‐based consecutive cohort was 17.5%, which is similar to the 14%–23% in previous series with malignant meningioma, and two‐ to threefold higher than in series with benign meningioma (14). The incidences were 2/1.000.000/year for TERTpMut WHO grade III meningioma and 8/1.000.000/year for TERTpwt WHO grade III meningioma in our catchment area.

We found a higher incidence rate of recurrence and death when comparing incidence rates per 10 person‐years for TERTpMut to TERTpwt WHO grade III meningioma. TERTpMut were confined to WHO grade III tumors of the anaplastic subtype with one small exception; an anaplastic WHO grade III meningioma that showed some level of papillary morphology in the first tumor and in later recurrence the papillary morphology was absent.

Prognostic impact of TERTpMut in WHO grade III meningioma

We found a 1.7 times higher recurrence rate for TERTpMut compared to TERTpwt and a 2.5 higher mortality rate for TERTpMut compared to TERTpwt, which is well compatible with an independent prognostic impact by TERTpMut. Surprisingly, however, Cox regression analysis showed a difference that was not statistically significant between overall survival in the TERTpMut and TERTpwt groups. Similarly, the cumulative incidences of recurrence from the first WHO grade III tumor failed to show a statistically significant difference between the groups. In contrast, larger cohorts and meta‐analyses of WHO grade III meningioma have demonstrated associations also between TERTpMut and survival. Sahm et al reported that TERTpMut WHO grade III meningioma patients recurred statistically significantly earlier than TERTp meningioma patients (20). Our cohort, although large for primary and secondary grade III meningiomas, included only 40 patients. The difference between published literature and our cohort may reflect lower statistical power of the latter. WHO grade III meningiomas comprise a population where all patients experience recurrence and death within a relatively short time span and larger cohorts may be needed to detect statistically significant differences between subgroups. Importantly, however, cumulative incidence of recurrences and incidence rates reflect different qualities. Malignant tumors such as WHO grade III meningiomas recur if follow‐up is long enough, while time to recurrence may differ with the growth rates. In analogy, the Ki‐67 proliferation index reflects time to recurrence rather than cumulative incidence of recurrence during long‐term follow‐up of meningiomas (15). In this context, rates of recurrence and mortality may be preferrable prognostic markers. These rates were higher for TERTpMut meningiomas in our cohort; findings that agree with Mirian et al (14) who reported that TERT gene alterations (including promoter mutations, gene translocations and DNA amplifications: TERT‐alt), were independent prognostic markers that defined a subset of aggressive malignant meningiomas.

TERT gene alterations in the natural history of malignant meningioma

TERTpMut are typically somatic mutations (8) which, in our cohort, were already present in the first meningiomas operated not only in the two patients with primary anaplastic meningiomas, but also in the first available non‐malignant samples in 4/5 secondary anaplastic meningiomas with TERTpMut. The findings together with prognostic observations agreed that TERTpMut may be a marker of aggressive disease (5, 8, 14). Such a linkage between TERT‐mutation and aggressive disease could occur in two different ways.

One possibility is that TERT would function as a driver mutation in which case the TERT mutation would occur simultaneously with change from a benign grade I or II meningioma to grade III. Such mechanisms have been described for hepatocellular cancer (16). Observations in a small number of meningioma patients were compatible with this mechanism in a study by Juratli et al (8). They observed switches from TERTpwt to TERTpMut that were simultaneous with switches from benign to anaplastic meningiomas while pairs of WHO grade I, II or III tumors that retained their original grades also retained wild‐type TERT (8); in this study TERTpMut appeared to be a feature of progression and anaplastic histology.

The other possibility is represented by findings of Goutigny et al who described that TERTpMut could be present already in grade I–II samples (5). In this study, TERTpMut was interpreted to represent a meningioma subgroup that was particularly prone to malignant progression and the authors suggested that TERTpMut could serve to drive phenotypes from lower to higher grade meningiomas. Our population‐based findings from a similarly sized cohort as those above (5, 8), support the latter possibility; TERTpMut can be present early in a subgroup of meningiomas with a particularly bad prognosis possibly because of some genetic instability which also conferred an increased risk of malignant progression.

The incidence of TERTpMut grade III meningiomas is low and research will depend on availability of mutational and clinical data from different sources. Population‐based data are particularly useful and publication of our findings will not only support the latter of two above hypotheses but also provide material for meta‐analyses of TERTpMut in high‐grade meningioma.

Temporal and spatial heterogeneity

Investigation of TERTpMut in serial samples obtained during malignant progression in meningioma is complicated by spatial heterogeneity of grade II–III meningiomas (2, 13). Previous studies showed that WHO grade II–III meningioma maintained very few mutations across serial recurrences (2, 3). Spatial heterogeneity regarding TERTpMut is evident in other cancers (22). For meningiomas, Juratli et al (8) analyzed spatially different tumor samples obtained during one operation from three different patients. The tumors displayed simultaneous presence of both TERTpMut and TERTpwt subclones. Findings of heterogeneity suggest that TERTpMut may be an epiphenomenon and a passenger mutation. The association between worsening WHO‐grade and negative prognostic impact of TERTpMut (5, 8, 14, 20) could reflect that higher grade tumors simultaneously have worse prognoses and genetic instability with faulty telomere maintenance mechanisms reflected as TERTpMut which are inherent to many types of cancer (1).

Patient #3 (Figure 5) in our cohort clearly showed that biopsies from serial recurrences could show different TERT‐mutational status. The temporarily “disappearing” TERTpMut could suggest undersampling or differential dominance of subclones with different TERT‐mutational status during the course of disease; a mechanism well described by Juratli et al (8). The TERTpMut WHO grade I meningioma might have had enough genomic disruption to progress to a WHO grade III, where the dominant clonal component happened to be TERTpwt, while the last recurrence was again dominated by tumors with clonal components containing TERTpMut. Small subclones with TERTpMut cells might not have been reproducibly discovered at the first operation of the grade III tumor as a result of the sensitivity of Sanger sequencing (7, 19). Similarly, patient #6 with secondary malignant meningioma and first detection of TERTpMut when the meningioma became malignant, might have harbored a smaller TERTpMut subclone already in the initial tumors. The apparent temporal heterogeneity probably reflected spatial heterogeneity with co‐existence, but differential dominance of TERTpwt and TERTpMut subclones. Given temporal and spatial heterogeneity, TERTpMut could either be independent of a basic malignant geno‐and phenotype or represent a subclone with qualities to maintain malignant behavior and produce recurrences. Our findings and previous publications are compatible with both possibilities: TERTpMut might be either a driver mutation or a passenger mutation—it could also have different roles in different tumors.

Figure 5

Chromatograms in patient #3. A. C228T mutation in the patient’s first WHO grade I meningioma. B. The mutation was not evident in following WHO grade III recurrence. C,D. C228T mutation in the later WHO grade III recurrences.

Strengths and weaknesses

Malignant meningiomas are rare. Our cohort is large and comprehensive for the pathology but is still too small for high statistical power. A unique strength of our study is the availability of surgical tissues from a consecutive, population‐based cohort of primary and secondary WHO grade III meningiomas, including serial samples from pre‐malignant precursor meningiomas and recurrent grade III meningiomas. The weakness is the retrospective design which precluded systematic collection of tumor samples to account for spatial heterogeneity. Subsequently, analyses only reflect the tumor subclones that were available for analyses and interpretation may have been obfuscated by intratumoral heterogeneity and Sanger sequencing sensitivity.

Finally, we studied only two point mutations by Sanger sequencing. TERT promoter mutations comprise the majority of TERT‐gene alterations in cancers (1) and a recent systematic review showed that most TERT‐alterations in meningiomas were point mutations at C228T and C250T, but 5% still comprised RETREG1‐TERT and LPCAT1‐TERT fusions (14). Unbiased gene sequencing could thus have revealed other phenotypical relevant TERT gene alterations in our cohort of malignant meningioma.

Conclusion

We investigated the TERT promoter mutation (TERTpMut) status in a population‐based consecutive cohort of 40 WHO grade III meningioma patients (20 primary and 20 secondary). Seven of the 40 patients (17.5%) harbored TERTpMut, thus validating earlier reported incidence of TERTpMut in WHO grade III meningioma. The incidences were 2/1.000.000/year for TERTpMut WHO grade III meningioma and 8/1.000.000/year for TERTpwt WHO grade III meningioma in our catchment area.

We found a higher recurrence rate and mortality per 10 person‐years in the TERTpMut group. TERT‐mutations were not typically acquired during malignant progression. TERTpMut could occur independent of malignant progression in meningioma and TERTpMut could represent either driver‐ or passenger mutations.

Conflict of interest

The authors declare no conflict of interests.

Ethics approval

The study was approved by the Danish Ethics Committee. Approval number H‐6‐2014‐010.