Development of patient-derived xenograft models from a spontaneously immortal low-grade meningioma cell line, KCI-MENG1

Meningiomas are the most common primary tumors of the central nervous system accounting for approximately 35.5% of all primary brain tumors [1]. The age-adjusted annual incidence rate is 7.22 per 100,000 individuals with a nearly 2.3-fold higher incidence in women. Over 100,000 cases were reported in United States between 2005 and 2009 [1]. Meningiomas are composed of neoplastic meningothelial cells derived from arachnoid cap cells [2]. The World Health Organization (WHO) classifies them into three main histologic subtypes: benign (grade I), atypical (grade II); and malignant (anaplastic) meningiomas (grade III) [3]. Current therapies involve surgery, fractionated radiation therapy, and stereotactic radiosurgery. There is an important group of patients with inoperable or incompletely resected low-grade meningiomas, in addition to the high-grade tumors, who develop recurrent disease following surgery and radiation therapy. Effective treatment options for these patients are exceedingly limited at present [4, 5].

Progress in the development of new treatments for meningioma is limited by a paucity of in vitro cell line models effectively limiting the availability of suitable in vivo models. Most of the well-characterized cell lines were isolated from malignant meningiomas [6‐9]. Those from benign [10‐12] or atypical [13] meningiomas have been genetically modified to generate stable, immortal cell lines. Of these artificially-immortalized benign meningioma cell lines, the most common method employed was viral transduction of cells to generate expression of the telomerase catalytic subunit (hTERT). The endogenous expression of hTERT is found in 30–50% of all benign and nearly 100% of high-grade meningiomas [14, 15]. Expression of hTERT in recurrent meningioma has also been observed [14]. Therefore, hTERT expression is a logical choice to manipulate the tumor cell biology to allow for continued in vitro cell growth. However, despite the careful characterizations described by the authors of those studies [10‐13], it is difficult to know what aspects of the tumor cell biology may also have been altered that would confound the use of these cells as meningioma models. Moreover, two cell lines, MENII-1 [12] and Me3TSC [10], in addition to the hTERT, co-expression with human papillomavirus E6/E7 oncogenes and SV40 large T antigen, respectively, was required in order to attain immortalization, although these viral genes have not been associated with meningioma in vivo. Also of note, syngeneic mouse models of meningioma have been genetically engineered by conditional knockout of tumor suppressors such as neurofibromatosis 2 (NF2) [16, 17], however, as gene expression and regulation are considerably divergent between mouse and human [18], human tumor models may be more suitable for translational research purposes [19].

Meningiomas were among the first CNS solid tumors found to have consistent cytogenetic aberrations [20‐22], The most well explored observation is the loss of heterozygosity due to a loss of one copy of the long arm of chromosome 22 [23, 24], and this is usually the only chromosomal loss associated with benign meningioma [25]. Atypical and malignant meningiomas also have losses of the short arm of 1 and the long arm of 14q [26], and gains of the long arms of chromosomes 1, 9, 12, 15, 17, and 20 [25, 27]. These striking chromosomal abnormalities may be related to the hTERT expression and telomerase activity [28] found in some benign and almost all high-grade meningiomas [14, 15].

While there have been numerous studies examining the genetic alterations characteristic of meningiomas, these have yielded little in the way of efficacious treatment alternatives. As such, there is a critical need for development of pre-clinical tumor models to improve the understanding of the underlying pathobiology of meningiomas and for the development and testing of novel therapeutic approaches. Human cell culture systems represent an essential experimental tool. However, most studies use primary, early passage human meningioma cell lines that typically senesce after a few passages. Here, we report the isolation and characterization of a novel, spontaneously-immortalized cell line, which we have designated as KCI-MENG1, derived from a WHO grade I benign meningioma and used to develop mouse xenograft models.

Improved survival and reduced recurrence are expected following complete excision of the intracranial meningiomas [36, 37]. However, up to 5% of benign meningiomas [38] and 17–40% of atypical meningiomas recur at 5 years following complete resection [38, 39]. Not surprisingly, partial resection is associated with a significantly higher risk of tumor recurrence (87% for atypical meningioma) [37, 39]. Generally, 5-year survival rate is 95 and 61% after total and partial removal of the tumor, respectively [36, 37, 39]. Furthermore, up to 29% of recurrent benign meningiomas [26, 40] were reported to progress into more aggressive higher grades. The currently available treatment options following partial resection or recurrence of the tumor are surgery and radiotherapy [41‐43]. To date, there are limited effective chemotherapeutic options for the treatment of refractory or recurrent benign or high-grade meningiomas [4, 5].

One obstacle in the development of novel therapeutic agents for meningioma treatments is a relative lack of suitable in vitro and in vivo model systems. Most cell lines originate from malignant meningiomas [6‐9] or from benign [10‐12] or atypical [13] meningiomas that have been genetically modified for immortalization (see Table 1). In this manuscript, we have described KCI-MENG1, a native, apparently immortal cell line derived from a WHO grade I meningioma, and is the only such immortal cell line we have identified out of 58 primary cultures of benign meningiomas collected. KCI-MENG1-LP is a heterogeneous cell population comprised of two cellular morphologies, while at high passage (KCI-MENG1-HP), only one of the two cell types remains (likely due to selection during the culturing process). All cell lines derived from KCI-MENG1 retained the expression of the meningioma diagnostic markers EMA and vimentin, which were weakly and strongly stained in the patient tumor specimen, respectively. EMA expression varied between the two cell types (very strong in the round cells, weak in the spindle cells) but both cell types were strongly stained for vimentin. The smaller, round cell phenotype has stronger Ki-67 staining, and maintains the expression of EMA and vimentin through a high number of passages. The difference in the growth kinetics of the cells at low vs. high passages and the marked increase of telomerase activity seem to be congruent with the shift in population density reflecting the loss of the spindle-shaped cells. Furthermore, the higher proportion of cells with the aberrant karyotype and higher magnitude of amplification of cancer-related genes identified by aCGH, particularly in chromosomes 1, 9, 12, 15, 17, and 20 which are known chromosomal gains in atypical and malignant meningioma [25], suggest that, although initially a minor subclone in the original meningioma tumor, the round cells with the high proliferative activity (KCI-MENG1-HP) are likely to be the tumorigenic cells responsible for the tumor development and growth in the patient. This is also supported by our development of subcutaneous tumors from both KCI-MENG1-LP and KCI-MENG1-HP cells in immunocompromised mice, and the high proliferative activity of the tumors generated from subdural implantation. Previous studies of xenografts generated from the malignant meningioma cell lines IOMM-Lee and CH-157, and patient-derived cells demonstrated that meningioma cells with a complex karyotype, such as our KCI-MENG1-HP, more consistently generated tumors after subcutaneous implantation than those with a simple karyotype [44]. This likely manifested as an in vivo clonal selection and accounts for our finding that the KCI-MENG1-LPSX-CL cells isolated from the low passage subcutaneous tumor were a homogeneous population that resembled the KCI-MENG1-HP cells. Interestingly, absence of PR immunostaining in the more aggressive high passages cells is in keeping with the known association of loss of PR expression, cumulative karyotype abnormalities, and aggressive clinical behavior of progressive or recurrent meningiomas [45].

In addition to vimentin, we found expression of N-cadherin in both round and spindle-shaped cell types of the low passage cells, both of which are considered markers of mesenchymal phenotype associated with the invasive properties of some cancer types [46]. N-cadherin expression was previously described in a subset of WHO grade I meningiomas [47] and co-expression of vimentin and N-cadherin was also found in drug-resistant lung cancer [48] and pancreatic carcinoma [49]. Diminished E-cadherin expression is commonly found in all grades of meningioma [50], and in our study, the original patient tumor, mouse xenograft tumors, and all cell lines were negative for E-cadherin immunostaining (see Figure 4) despite the genomic amplification of the CDH1 gene (which encodes E-cadherin) identified in the cells (see Additional file 2: Table S2), implying that the mesenchymal phenotype predominates.

The property that cells from all cancer types have in common is the ability to propagate indefinitely. Cancer cells typically achieve this by expressing telomerase, which is absent in senescent, differentiated cells [51]. The telomerase activity and resulting telomere dysfunction contributes to genomic instability [52] and can lead to the generation of polyploid cells and enhance the tumorigenicity of those cells [28], which corresponds to our findings of the robust telomerase activity and near-triploid karyotype of the round phenotype KCI-MENG1-HP cells. In meningiomas, telomerase activity tends to correlate with WHO grade and is observed in up to 95% in anaplastic meningiomas [14, 53‐55], though rarely found in benign meningiomas [15]. Telomerase inhibitors are currently under development, but not for meningiomas or other brain tumors [51].

Collectively, despite KCI-MENG1 cells originating from a WHO grade I meningioma, our data suggest that these cells have a genomic complexity and a biological profile that is consistent with recurrent and/or high-grade meningiomas. In a cytogenetic study of recurrent, progressive meningioma, Al-Mefty et al. described similar findings [26]. When they assayed meningioma specimens from the initial low-grade tumors from patients that later developed high-grade tumors, they found cells with the aberrant high-grade karyotype in the benign tumors. In a similar line of research, cytogenetic heterogeneity was identified in 33.4% of meningiomas, and it was found that tumor progression and recurrence was predicted by the most advanced clone even if present in lesser abundance [56]. Although the patient from which KCI-MENG1 cells are derived currently has no clinical or radiographic evidence of tumor recurrence 4 years following surgery, the isolation of the near-triploid cells, including a gain of chromosome 1q (known to correlate with shorter progression-free survival in atypical meningioma [27] ), and with high telomerase activity (usually associated with high-grade meningioma [53‐55] and demonstrated in meningiomas undergoing malignant progression [57]) is worrisome and suggestive that this patient could be at high risk to have a progressive recurrence. This patient population is one we hope to serve with translational studies of meningioma tumor biology and disease progression utilizing our KCI-MENG1 in vitro and in vivo models, and as such will facilitate further development of novel therapies to improve treatment options for all grades of meningiomas.

Although derived from a typically good prognosis benign meningioma specimen, the newly-established spontaneously immortal KCI-MENG1 meningioma cell line can be utilized to generate xenograft tumor models with low- or high-grade features, dependent on the cell passage number (likely due to the relative abundance of the near-triploid, likely poor prognosis, cells that were present as a small proportion in the original tumor). These human meningioma mouse xenograft models will provide biologically relevant platforms from which to investigate differences in low- vs. high-grade meningioma tumor biology and disease progression as well as to develop novel therapies to improve treatment options for poor prognosis or recurrent meningiomas.