YAP Mediates Tumorigenesis in Neurofibromatosis Type 2 by Promoting Cell Survival and Proliferation through a COX-2–EGFR Signaling Axis Free

The Hippo–YAP signaling pathway has emerged as a major driver of tumorigenesis and metastasis in a wide spectrum of human cancers (1–3). The core of the pathway is composed of a well-defined kinase cascade composed of the MST1/2 kinases that form a complex with the scaffold protein WW45 and phosphorylate the LATS1/2 kinases. Phosphorylated LATS1/2, in complex with Mob1, bind and phosphorylate YAP, a transcriptional co-activator. The phosphorylation of YAP creates a binding site for 14-3-3 and this prevents p-YAP from entering into the nucleus where it can form transcriptionally active complexes with TEADs and other transcription factors to drive the expression of proproliferative or antiapoptotic genes such as CTGF, Cyr61, Axl, Myc, and BIRC5 (4).

As a regulator of cell fate, proliferation, and death, YAP can function as an oncogene. Several examples exist showing that YAP overexpression drives tumorigenesis, including mouse models in which liver-specific expression of an activated allele of YAP or knockout of the MST1 and MST2 alleles in the liver lead to development of hepatocellular carcinoma (HCC; refs. 5, 6). Additional evidence for an oncogenic role of YAP in human tumors stems from findings demonstrating amplification of genomic region 11q22, to which YAP localizes, in breast cancer and significant upregulation of YAP expression in breast, ovarian, lung, pancreatic, colorectal, and liver cancers (7, 8). More recent studies have shown that YAP can function as an oncogene in tumors that are addicted to KRAS. Specifically, in models of KRAS-addicted tumors (pancreatic and lung adenocarcinoma), the inhibition of KRAS leads to cell death, which can be rescued by YAP activation (9, 10). Finally, genetic evidence for an oncogenic role for YAP in human cancer comes from two diseases, uveal melanoma and neurofibromatosis type 2 (NF2). In uveal melanoma, 80% of patients harbor mutations in the GNAQ (Gq) and GNA11 (G11) genes, which code for alpha subunits of heterotrimeric G proteins. Previous work had indicated YAP can be activated by mutated Gq/11 (11), and subsequently, it was found that mutated Gq/11 oncogenic function is mediated via YAP, thus implicating YAP as a potential therapeutic target in uveal melanoma (12, 13).

NF2 is an inherited disorder with an incidence of approximately 1 in 30,000 births, caused by germline mutations of the NF2 gene. The disease is characterized mainly by development of schwannomas of the eighth cranial nerve (14). The NF2 tumor suppressor gene encodes a 69-kDa protein called Merlin that has been shown to function as a regulator of multiple signaling pathways at the cell membrane and to possess nuclear functions. Merlin was originally shown to function upstream of Hippo in flies and subsequently in mammalian cells. A number of studies demonstrated that Merlin and YAP function antagonistically, including in vivo studies in which liver-specific knockout of Yap was sufficient to rescue HCC driven by inactivation of the Nf2 gene (15).

Mechanistic details of the function of Merlin have emerged from studies that demonstrated Merlin acts synergistically with a newly identified Hippo pathway component, Kibra, to promote LATS1/2 phosphorylation (16) and regulate the spatial organization of Hippo pathway components at the cell membrane by directly binding to LATS1/2 and recruiting it to the plasma membrane, where it is phosphorylated and activated by a MST-WW45 complex (17). Merlin has also been shown to have a nuclear function as an inhibitor of the E3 ubiquitin ligase CRL4^DCAF1 (18). Recent studies suggest that CRL4^DCAF1 promotes YAP- and TEAD-dependent transcription by inhibition of LATS1/2 in the nucleus, and analysis of patient samples indicates that this pathway operates in NF2-mutant tumors (19).

Thus, while the evidence cited above strongly suggests that YAP function is required downstream of NF2 loss of function in tumorigenesis, the mechanisms underlying the requirement for YAP and for which downstream targets are critical to the oncogenic functions of YAP remain unknown. To identify these mechanisms and identify disease-relevant targets, we employed a combination of cell-based and in vivo approaches. Our findings indicate that YAP function is required in NF2-null Schwann cells to promote cell survival through regulation of an EGFR signaling axis via transcriptional regulation of prostaglandin endoperoxide synthase 2 (COX-2) and prostaglandin E2 (PGE₂) production. Importantly, our findings suggest that treatment with COX-2 inhibitors could prove beneficial in slowing the growth rates of NF2-associated schwannomas.

All animal experiments complied with NIH guidelines and were approved by The Scripps Research Institutional Animal Care and Use Committee. A total of 5 × 10⁴ SC4-Luc pLKO or SC4-Luc pLKO-YAP cells were injected intraneurally into the sciatic nerves of NOD/SCID mice (6–8 weeks old). Tumor progression was monitored by bioluminescence imaging on an IVIS-200 system (Xenogen). Treatment was commenced after detection of signal correlating to tumor sizes of 1 to 2 mm³ (total flux ≥ 10⁶ photons/s). For drug treatment, celecoxib (Cayman Chemical) was diluted in vehicle (22.2:66.6:11.2, ethanol:PEG300:water) to a final dose of 100 mg/kg and administered by oral gavage, daily. Control mice received vehicle/DMSO mixture.

SC4 Nf2-null mouse Schwann cells and HEI-193 human NF2-mutant Schwann cells were obtained in 2010 and previously described (20, 21). HSC2λ cells were obtained from the laboratory of Dr. Margret Wallace (University of Florida) in 2015. These cells are derived from normal human Schwann cells that were immortalized by expression of TERT and CDK4^R24C (Wallace, manuscript in preparation). SC4-Luc cells were previously described (22). SC4, HEI-193, and HSC2λ cells were authenticated by short tandem repeat (STR) DNA profiling (DDC Medical; March 2015).

Cell viability was determined by luminescent ATP-dependent assay (CellTiter-Glo, Promega), according to manufacturer's instructions. For measurement of proliferation, the BrdU Proliferation Assay (Millipore) was used according to the manufacturer's instructions. Statistical significance was determined by a two-tailed Student t test. Each condition at each time point represents the mean of three experiments in triplicate for a total of 9 wells.

Measurement of caspase-dependent cell death was achieved through the use of the Caspase-Glo 3/7 assay following the manufacturer's instructions (Promega). Briefly, cells were seeded into white, opaque 96-well culture plates at 1,500 cells per well and transfected with control or YAP siRNAs. Caspase-Glo reagent was added at 24 or 48 hours and incubated at room temperature for 30 minutes, after which the luminescence was measured.

SC4 cells were transfected with control or YAP SMARTpool siRNA for 48 hours, and total RNA was extracted using TRIzol reagent. For analysis, the sequencing reads in color space were mapped to the mm9 genome using Tophat (23). The number of reads falling into each gene defined in the RefSeq gene annotations was quantified using HTSeq-count (24). The DESeq software (24) was used to detect differentially expressed genes between samples. Samples from three independent experiments were sequenced, combined, and analyzed to produce the final DESeq data. The RNA-Seq data are publicly available through the NCBI GEO database with accession number GSE61528.

RNAs were extracted using the Qiagen RNeasy Kit and reverse-transcribed into cDNA with the SuperScript III Kit (Life Technologies). qPCR was performed with SYBR Green (Applied Biosystems). Relative gene expression between control and YAP-KD was calculated with the 2^−ΔΔCT method (25). For complete primer sequences, please see Supplementary Table S2.

Cell media from SC4 and HEI-193 cells treated with the indicated siRNA's was collected and diluted 2-fold with provided assay buffer (Mouse AREG ELISA and Human AREG ELISA, Raybiotech). Secreted AREG concentrations were assayed for three independent experiments in triplicate, according to manufacturer's instructions. AREG concentrations were determined by comparing recorded absorbance readings to a standard curve of diluted AREG.

To examine the role of YAP in NF2-null Schwann cells, three cell lines were used: SC4—Nf2-null mouse Schwann cells, HEI-193—NF2-null schwannoma cells isolated from an NF2 patient, and HSC2λ cells—immortalized normal human Schwann cells in which two shRNAs were used to knockdown the expression of Merlin (Supplementary Fig. S1A). As expected, the loss of Merlin expression resulted in increased cell numbers over time (Supplementary Fig. S1B).

Using two independent siRNA sequences and siRNA SMARTpool, the expression of YAP was knocked down and effects on cell numbers were examined. A significant decrease in cell numbers was observed, with numbers reduced 1.5-fold in SC4 cells, 2-fold for HEI-193 cells, and 2- to 2.5-fold for HSC2λ cells, over the course of 72 hours, respectively (Fig. 1A and B, Supplementary Fig. S1C–S1E). Next, verteporfin, a known inhibitor of the YAP–TEAD interaction was used to pharmacologically test the requirement of YAP in NF2-null Schwann cells (26). Treatment of SC4 and HEI-193 cells over 72 hours with 1 μmol/L verteporfin significantly reduced cell numbers by 3- to 8-fold over 72 hours, respectively (Supplementary Fig. S1F and S1G).

To determine whether the effects of YAP knockdown are due to reduced cell proliferation or increased cell death rates, we assessed BrdU incorporation in YAP-knockdown SC4 and HEI-193 cells. BrdU incorporation was slightly reduced over 72 hours, indicating that YAP knockdown has a relatively small, but reproducible, impact on cell proliferation (Fig. 1C and D). To assess the effects of YAP inhibition on cell death, we assessed the viability of SC4 and HEI-193 cells treated with verteporfin in a dose-dependent manner. VP concentrations as low as 1 μmol/L caused a 50% reduction in viability of both SC4 and HEI-193 cells, as measured by an ATP-dependent luminescent assay (Fig. 1E and F). That inhibition of YAP leads to increased cell death rates is further corroborated by increased levels of apoptotic markers, cleaved caspase-3, caspase-7, and PARP in SC4 cells infected with a vector expressing YAP shRNA (SC4 pLKO-YAP) compared with SC4 cells infected with a control shRNA vector (SC4 pLKO; Fig. 1G). Similarly, increased caspase-3/7 activity was detected in SC4 and HEI-193 cells in which YAP expression was knocked down (Supplementary Fig. S1H and S1I).

To investigate the requirement of YAP for Nf2-null Schwann cell tumorigenesis, we first assessed the ability of cells with YAP knockdown to form colonies in soft agar. SC4 pLKO and SC4 pLKO-YAP cells were plated at equal numbers, and colony formation was assessed after 14 days. The knockdown of YAP significantly impaired the ability of SC4 pLKO-YAP cells to form colonies and reduced colony numbers by 2-fold (Fig. 2A and Supplementary Fig. S2A). Moreover, assessing colony size revealed a typical size of approximately 50 μm for SC4 pLKO-YAP cells compared with approximately 200 μm for the SC4 pLKO cells (Supplementary Fig. S2B). To assess the requirement of YAP for Nf2-null Schwann cell tumorigenesis in vivo, luciferase-expressing SC4 cells (SC4-Luc) were infected with either a control vector (SC4-Luc pLKO) or a vector expressing YAP shRNA (SC4-Luc pLKO-YAP) to generate SC4-Luc cells with stable YAP knockdown (Fig. 2B), which lead to inhibition of SC4-Luc cell growth in culture (Fig. 2C). SC4-Luc pLKO and SC4-Luc pLKO-YAP cells were orthotopically implanted into the sciatic nerves of NOD/SCID mice, and schwannoma tumor formation and growth were monitored via bioluminescent imaging. Mice were imaged every other day, starting on day 6, and were sacrificed on day 20. Both control and YAP knockdown cohorts developed measurable tumor signals within 8 to 10 days postimplantation. However, compared with SC4-Luc pLKO controls, YAP knockdown significantly inhibited the growth of the Nf2-null schwannomas in vivo (Fig. 2D–F and Supplementary Fig. S2C). Subsequent ex vivo analysis of excised tumors indicated that tumors arising from SC4-Luc pLKO-YAP cells were significantly smaller than tumors in the control cohort (Fig. 2G). Taken together, our data indicate that YAP activity is required for the proliferation and survival of Nf2-null Schwann cells in vitro and tumor growth in vivo.

To understand the specific contribution of YAP to the survival of Nf2-null Schwann cells, an RNA sequencing approach was used to identify specific YAP-regulated genes. SC4 cells were treated with either nontargeting or pooled YAP siRNAs for 48 hours, followed by RNA isolation and sequencing. A number of genes were identified as significantly regulated by YAP and classified on the basis of functional annotation including genes involved in regulation of matrix metalloproteases, prostanoid biosynthesis, fatty acid oxidation, and G-coupled receptor signaling (Fig. 3A–C). We focused on PTGS2, which codes for the COX-2 protein, an enzyme that functions as a catalytic intermediary in the synthesis of prostaglandins from arachidonic acid. PTGS2 was validated as a YAP target by RT-PCR in SC4 and HEI-193 cells, as knockdown of YAP resulted in a 2- to 3-fold reduction in PTGS2 transcription, confirming the RNA-Seq results (Fig. 3D). The top 12 genes regulated by YAP also included AREG, which has been previously reported as a YAP target and codes for amphiregulin (AREG), an EGFR ligand (Fig. 3B) (27). YAP knockdown in both SC4 and HEI-193 cells resulted in 2- to 5-fold reduction in AREG transcription (Fig. 3E). Examination of intracellular levels of COX-2 protein by Western blotting in the presence or absence of YAP indicates that in response to YAP knockdown, COX-2 levels are drastically reduced. (Fig. 4A). To assess levels of AREG, we analyzed the cell culture media for secreted protein employing an ELISA approach. A highly reproducible decrease in secreted AREG was observed in the culture media of SC4 and HEI-193 cells in which YAP or AREG are knocked down (Fig. 4B and Supplementary Fig. S3A).

COX-2 functions as a mediator of prostanoid synthesis, including prostaglandin E₂ (PGE₂), previously shown to promote proliferation and survival of various types of tumor cells (28). Thus, the identification and confirmation of AREG and PTGS2 as targets of YAP-dependent transcription links two paracrine signaling mediators of cell survival to the Hippo–YAP pathway in NF2-null Schwann cells. To assess the specific role of COX-2 downstream of YAP in NF2-null schwannoma cells, SC4 and HEI-193 cells were treated with control or COX-2 siRNA, and the effect on cell numbers was assessed (Fig. 4A, C, and D and Supplementary Fig. S3B and S3C). In both cell lines, COX-2 knockdown reduced cell numbers by approximately 50% at 72 hours, although this was not as strong as YAP knockdown (compare to Fig. 1A and B), implicating additional downstream effectors in mediating YAP activity. To assess the requirement for COX-2 activity in driving YAP-dependent increase in cell numbers, we also employed the COX-2 inhibitor celecoxib, which significantly inhibited the growth of SC4 cells in a dose-dependent manner (Fig. 4E).

To determine whether the effects of YAP–COX-2 are mediated through PGE₂, SC4 pLKO-YAP cells were treated with exogenous PGE₂ in an attempt to rescue the phenotype conferred by YAP knockdown. PGE₂ treatment was able to partially rescue the effects of YAP knockdown by approximately 30% at 72 hours (Fig. 4F). These results indicate COX-2 and PGE₂ are essential effectors of YAP activity that are required, at least in part, for Nf2-null Schwann cell growth.

Given the partial rescue of the effects of YAP knockdown by PGE₂, we next assessed whether the combination of PGE₂ and AREG would result in a greater rescue of the YAP knockdown phenotype. While treatment with PGE₂ or AREG alone led to only a partial rescue of the YAP knockdown phenotype, the combined actions of both targets rescued the phenotype, increasing cell numbers to levels comparable to control SC4-pLKO cells (Fig. 4F–H). Importantly, treatment of control SC4-pLKO cells with the PGE₂-AREG combination did not have a significant impact on cell numbers (Supplementary Fig. S3D). Taken together, these findings strongly indicate that YAP regulates Nf2-null Schwann cell growth by coordinated regulation of AREG and COX-2.

PGE₂ is known to promote cell proliferation and inhibit apoptosis through Src-dependent transactivation of EGF receptor (EGFR) and PI3K/AKT activation downstream (28). We therefore examined the levels and status of EGFR phosphorylation in SC4, HEI-193, and HSC2λ-shNF2 cells in the absence of YAP. While total EGFR levels remain unchanged upon YAP knockdown, the activation of EGFR was greatly reduced as indicated by reduced phosphorylation on tyrosine 845 (Fig. 5A and B and Supplementary Fig. S4A and S4B). Similar results were found with knockdown of COX-2 in HEI-193 cells and by treatment of SC4 cells with the COX-2 inhibitor celecoxib (Fig. 5B and C). Examination of the activation status of AKT revealed similar findings. While the levels of total AKT were not significantly affected by the presence or absence of YAP or COX-2, the phosphorylation at serine 473 decreased in a YAP- and COX-2–dependent manner (Fig. 5D and Supplementary Fig. S4C). As we previously determined that PGE₂ can partially rescue the effects of YAP knockdown at the cellular levels (Fig. 4F), we next sought to determine whether the activation of EGFR and AKT by YAP–COX-2 is mediated by PGE₂. Indeed, short-term exposure of SC4 cells to PGE₂ caused a significant increase in EGFR-Y845 and AKT-S473 phosphorylation (Fig. 5E). Interestingly, PGE₂ treatment also resulted in elevated YAP levels, suggesting that a feedback loop may exist between YAP and COX-2.

Since we observed phosphorylation of EGFR on tyrosine 845 in response to treatment with PGE₂, we next examined the possible contribution of Src, which is known to phosphorylate EGFR at tyrosine 845 and previously shown to transactivate EGFR in response to PGE₂ in colorectal cancer cells (29, 30). Treatment with the Src inhibitor SU6656 alone decreased phosphorylation of EGFR-Y845, which is perhaps expected, given a basal level of EGFR activity supporting cell survival (Supplementary Fig. S4D). Significantly, pretreatment of cells with SU6656 inhibited PGE₂-induced EGFR phosphorylation, implicating Src as a mediator of EGFR activation by PGE₂ (Fig. 5F and G, Supplementary Fig. S4E and S4F). Taken together, these findings strongly indicate that YAP regulates Nf2-null Schwann cell proliferation and survival through an EGFR/AKT signaling pathway.

To determine the extent to which the effects of YAP on NF2-null Schwann cells are mediated by COX-2 and AREG, we employed a gain-of-function approach and assessed consequences of expressing a constitutively activated allele of YAP (YAP^S127A) in SC4 and HSC2λ-shNF2 cells. The expression of YAP^S127A in both cell lines led to increased expression of COX-2 and AREG (Supplementary Fig. S5A–S5H and Fig. 6A and B). Moreover, the expression of YAP^S127A also had a significant effect at the cellular level, as evidenced by increases in SC4 (2.5-fold at 72 hours) and HSC2λ (2−fold at 72 hours) cell numbers over time, compared with control transfected cells (Fig. 6A and B). The effects of YAP^S127A expression were dependent on COX-2, as knocking down COX-2 expression partially rescued the growth advantage conferred by expression of YAP^S127A (Fig. 6A and B, Supplementary Fig. S5I and S5J). Given the partial rescue by COX-2 knockdown, we next assessed whether the combination of COX-2 and AREG knockdown would result in a greater inhibition. Indeed, combined inhibition led to an almost complete rescue of the growth advantage (Fig. 6C and D), strongly suggesting that growth advantage conferred by expression of YAP^S127A is mediated through coordinated regulation of AREG and COX-2.

Our findings suggest that treatment with a COX-2 inhibitor, such as celecoxib, could inhibit the growth of Nf2-null schwannomas. To test this hypothesis in vivo, we employed the orthotopic model of NF2 described above. Tumor progression was monitored every other day by bioluminescence imaging, and total flux counts were recorded for each animal. Between 10 and 12 days postinjection, animals reached similar flux readings, indicating measurable tumors (flux = 10⁶ counts) and were enrolled randomly into control (vehicle only, oral, once daily) or celecoxib-treated cohorts (100 mg/kg, oral, once daily) for a period of 10 days. Analysis of the flux readings for the animals in the two cohorts demonstrated a significantly slower tumor growth rate in celecoxib-treated mice than in control mice (Fig. 7A–C). Taken together, these data demonstrate that inhibition of COX-2 significantly impaired Nf2-null Schwann cells growth and tumor formation in vivo.

Our findings suggest that AREG and PTGS2 should be co-regulated in NF2-null Schwannoma. To assess this, we examined the expression of these genes in an existing expression microarray dataset prepared from 31 vestibular schwannomas (31). AREG and PTGS2 displayed a Pearson correlation coefficient of 0.5784 (P = 0.0007), suggesting that a correlation exists between expression of both alleles at the transcriptional level (Fig. 7D). Interestingly, the expression of PTGS2 also shows a statistically significant correlation with CYR61 (r = 0.7101, P < 0.0001) and CTGF (r = 0.4623, P = 0.0088; Fig. 7E and F), which have been previously characterized as YAP target genes (32). To further confirm the correlation between AREG and COX-2 at the protein level, we assessed the levels of AREG and COX-2 in protein extracts prepared from NF2-null schwannomas. AREG and COX-2 are expressed in the majority of tumors examined (10 of 16), and in agreement with the expression array data, the levels of AREG and COX-2 protein appear to correlate in the majority of these tumors (Fig. 7G). In a limited number of tumors, we were also able to assess the levels of AKT and p-AKT S473, which appear to be consistent between the samples (Fig. 7G).

While mechanistic details of YAP regulation and downstream targets are emerging, relatively little is known about relevance and function of these targets in cancer. YAP has been shown to drive expression of a number of mRNAs and miRNAs implicated in various aspects of cell proliferation and transformation including connective tissue growth factor (CTGF), the receptor tyrosine kinase Axl, EGF receptor ligand amphiregulin (AREG), the cell-cycle regulator cyclin D1, and the FoxM1 transcription factor (32–35). Previous studies indicate that YAP can suppress PTEN levels through expression of miR-29 and regulate global miRNA biogenesis through microprocessor activity (36, 37). It is likely that YAP regulates many of these targets in a cell- and tissue-specific manner, and further work is required to establish the relevance of these targets in YAP-dependent tumors.

Merlin, the product of the NF2 tumor suppressor gene, is well-established as an upstream regulator of the Hippo signaling pathway. However, the role of YAP in NF2-null schwannomas has yet to be established. Our findings demonstrate YAP is required for the survival and proliferation of Nf2-null Schwann cells and tumor formation in vivo. From the broad range of transcripts regulated by YAP, we focused on PTGS2, which codes for COX-2, as it has recently been identified as a direct target of YAP in a model of pancreatic ductal adenocarcinoma (38). COX-2 expression is typically low in normal tissue but reported as elevated in multiple types of cancers including pancreatic, lung, head and neck, prostate, and breast tumors (28). COX-2 is a critical enzyme in biosynthesis of prostaglandins, of which PGE₂ has been identified as playing a major role in promoting tumor growth (39–41).

Prostaglandins impact tumorigenesis through multiple mechanisms including modulation of tumor cell proliferation and apoptosis, as well as modulation of the tumor microenvironment and immune responses. Most relevant to our studies are cell-autonomous functions of PGE₂. Specifically, PGE₂ promotes cell proliferation in colon and lung cancer cells through the Ras-MAPK and GSK3β/β-catenin signaling pathways (42–44), has been shown to promote colon tumor cell survival through activation of a PI3K/AKT/PPARγ axis (39, 45), and promotes colorectal cancer cell migration and invasion through activation of a EGFR/AKT signaling axis, through a β-arrestin/Src-mediated intracellular mechanism (46).

In our studies, we find that PGE₂-mediated activation of EGFR/AKT signaling promotes cell survival of Merlin-deficient schwannoma cells (Supplementary Fig. S6). This likely reflects a cell-type–specific difference in the role of this signaling axis between colorectal cancer and Schwann tumor cells. Indeed, one example of different roles undertaken by the same PGE₂-activated pathway includes recent studies showing that PGE₂, through a PI3K/AKT signaling cascade, can protect mouse embryonic stem cells from undergoing apoptosis (47). As far as Src involvement, our studies with SU6656 suggest that Src, or another family member, functions as a mediator of the effects of PGE₂, and further studies are required to clearly identify the responsible mediator. Interestingly, the effects of YAP knockdown on Nf2-null Schwann cells were only partially rescued by PGE₂ treatment and required supplementation of amphiregulin. It has been previously reported that PGE₂-mediated activation of EGFR in colorectal cancer cells is dependent on TGF-α, an EGFR ligand (30). Thus, it is possible that in Schwann cells, the YAP-coordinated expression of AREG and PTGS2 functions in a similar fashion to what is observed in colorectal cancer cells.

Although the Hippo–YAP pathway probably carries out additional functions in NF2-null Schwann cells, EGFR-mediated suppression of apoptosis appears to represent a major function in promotion of schwannoma tumorigenesis, as suggested by rescue of the YAP knockdown phenotype by AREG and PGE₂-mediated activation of EGFR/AKT. Indeed, EGFR has been previously implicated in the pathogenesis of sporadic and NF2-associated schwannomas with consistent overexpression and activation of EGFR family receptors reported (48). It has been suggested that Merlin regulates EGFR internalization and signaling in response to cell:cell contact inhibition (49). Our studies suggest the activation of EGFR is mediated through YAP-coordinated expression of AREG and COX-2/PGE₂/EGFR activation (Supplementary Fig. S6). Further studies are required to determine whether both mechanisms function in NF2-null Schwann cells.

To establish the relevance of YAP coordinated expression of AREG and COX-2 in patients with NF2, we analyzed the expression of COX-2/AREG in NF2-null schwannomas. COX-2 and AREG expression appeared to be correlated both at the mRNA and protein levels, suggesting they are indeed regulated coordinately. As the number of normal nerve tissue samples in our study is low, it is difficult to determine whether the fact that COX-2/AREG do not appear to be co-regulated in these samples is significant. Interestingly, PTGS2 expression also correlated with the expression of CYR61 and CTGF in patient samples, suggestive of a role for a YAP-regulated transcriptional program in these tumors. Clearly more work is needed to establish the extent of YAP-regulated pathways involvement and to determine the significance of the relative levels of COX-2/AREG expression in tumors versus normal tissue. The finding that AKT appears to be activated in all tumor samples examined, compared with normal, is not surprising given the central role of PI3K/AKT signaling in tumorigenesis and the fact that multiple upstream effectors feed into this signaling hub.

Previous studies describe elevated COX-2 expression in vestibular schwannomas, which appear to positively correlate with a higher proliferation index (50). Moreover, a retrospective analysis of patients with vestibular schwannomas demonstrated that aspirin use shows an inverse correlation with tumor growth (51). Our findings identify a tumor cell-intrinsic mechanism, through which the effects of daily administration of celecoxib elicited a pronounced antitumor effect, and suggest that long-term treatment with NSAIDs could provide a benefit to patients with NF2. A number of reports have indicated that celecoxib has anticancer activity that is COX-2–independent (52). However, the inhibitory effects we observed on Nf2-null Schwann cell proliferation with direct knockdown of COX-2, along with pharmacologic inhibition, underscore a significant role for COX-2 in NF2-null schwannomas. Future studies employing immunocompetent animal models of NF2 will be required to elucidate the effects of COX-2 inhibition on other well-documented functions of eicosanoids, including the promotion of inflammation and effects on the tumor microenvironment, the role of PGE₂ in immunosuppression and, perhaps highly relevant to vestibular schwannomas, proangiogenic activities.

No potential conflicts of interest were disclosed.

Conception and design: W. Guerrant, S. Kota, J.L. Kissil

Development of methodology: W. Guerrant, S. Kota, J.L. Kissil

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): W. Guerrant, S. Kota, S. Troutman, V. Mandati, A. Stemmer-Rachamimov, J.L. Kissil

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): W. Guerrant, S. Kota, S. Troutman, M. Fallahi, A. Stemmer-Rachamimov, J.L. Kissil

Writing, review, and/or revision of the manuscript: W. Guerrant, S. Kota, A. Stemmer-Rachamimov, J.L. Kissil

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.L. Kissil

Study supervision: J.L. Kissil

The authors thank Dr. David Lim (House Ear Institute) for sharing HEI-193 cells, Drs. Margaret Wallace and Hua Li (University of Florida) for immortalized human Schwann cells, and Dr. Kirill Martemyanov (Scripps Research Institute) for critical reading of the article. They also thank the Scripps Genomics Core and Gautam Shankar from the Scripps Informatics Core. W. Guerrant is a recipient of the YIA from the Children's Tumor Foundation.

This work was supported by the NIH (NS077952 and CA124495 to J.L. Kissil).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.