Mutations in the GNAS Gene Prevent the Cell Invasion by Activating the MEG3/Wnt/β-catenin Axia in Growth Hormone-secreting Pituitary Adenoma

Approximately 30–40% of growth hormone-secreting pituitary adenoma (GHPA) harbor somatic mutations in the GNAS (α subunit of the stimulatory G protein) gene. However, the latent functional role of the mutations and relative molecular mechanism in GHPA remain unknown. The GNAS gene mutations were detected in GHPAs using a standard PCR-sequencing procedure. The mutation-associated MEG3 expression was measured by RT-qPCR. MEG3 was manipulated in GH3 cells using a lentiviral expression system. Alterations in mRNA pro�les in the MEG3-overexpressed cells were analyzed by RNA-seq. The cell invasion ability was measured using a Transwell assay, and the EMT-associated proteins were quanti�ed by immuno�uorescence and western blot. Finally, a tumor cell xenograft mouse model was applied to verify the effect of MEG3 on tumor growth and invasiveness. The percentage of invasive tumors was signi�cantly declined in GNAS-mutated GHPA tumors with the GNAS mutations compared to those tumors with the wild-type of GNAS. Consistently, the GH3 cell invasion capacity was decreased by expressing the mutant GNAS. MEG3 is uniquely expressed at high levels in GHPA harboring the mutated GNAS gene. Accordingly, the upregulation of MEG3 resulted in inhibiting cell invasion; and vice versa, the downregulation of MEG3 led to enhancing cell invasion. Mechanistically, the high level of MEG3 in mutated GNAS cells prevented the cell invasion via inactivation of the Wnt/β-catenin signaling pathway, which was further validated in vivo. The GNAS mutations inhibit the invasiveness of GHPA cells via inactivation of the MEG3/Wnt/β-catenin signaling pathway.


Introduction
Growth hormone-secreting pituitary adenoma (GHPA)accounts for 12.5% of pituitary neuroendocrine tumors, and excessive growth hormone resulted in acromegaly and systemic complications (Orme et al., 1998,Colao et al., 2004,Melmed, 2006).Acromegaly has been associated with a two-fold increase in mortality, mainly due to cardiovascular disease, which can be reversed by treatments for controlling the hormone over-production (Melmed, 2006).In addition, approximately half of patients with GHPA experienced high-risk relapse after the surgical reduction due to tumor cells in ltrating the surrounding tissues.Thus, supplementary chemotherapy still needs to control the tumor recurrence (Wilson et al., 2013).
It has been well-demonstrated that mutations found in the GNASgene, encoding α subunit of the stimulatory G protein, were detected in around 40% of GHPA (Goto et al., 2014,Hayward et al., 2001).The functional studies suggested that the GNAS mutations lead to the constitutive activation of adenylyl cyclase (AC), thereby inducing the cyclic AMP (cAMP) signaling pathway in pituitary tumors (Lania et al., 2012,Mantovani et al., 2010).The function of the GNAS mutations is thought to be involved in cell proliferation and hormone secretion (Stork and Schmitt, 2002).
Maternally expressed gene 3 (MEG3), a large non-coding RNA (lncRNA), was rst identi ed as a tumor suppressor in the pituitary (Zhang et al., 2003).The function of MEG3is associated with inhibition of cell invasion (Ma et al., 2019).In addition, the cAMP response element (CRE), located at the MEG3proximal promoter region, is critical for MEG3expression (Zhao et al., 2006).Furthermore, cyclic adenosine monophosphate-responsive element-binding protein (CREB), as a downstream target of mutated GNAS, is also implicated in upregulation of MEG3by binding to the CRE site (Zhao et al., 2006,Yamamoto et al., 1990).Therefore, we speculate that the GNAS mutations may participate in the upregulation of MEG3 expression.The present study aimed to ascertain whether the GNAS mutation inhibits theinvasiveness of GHPA cells mainly through MEG3-mediated inactivation ofthe Wnt/β-catenin pathway, which may provide a new therapeutic approach for treating GHPA.

Materials And Methods
Patients and clinical characteristics.
Tumor samples were collected fromforty-fourpatients with acromegaly who underwent endoscopic endonasal transsphenoidal surgery at the Department of Neurosurgery of Nanjing Jinling Hospital (Nanjing, China) between Nov. 2018 and Nov. 2019, including 21 males and 23 females.Additionally, 10 patients with clinically non-functioning pituitary tumors (NFPA) were included as negative controls.
Approval for the study was obtained from the Ethical Committee of Nanjing Jinling Hospital (2018NZKY-008-02) and informed consents were obtained from all the patients who participated in this study.
Pituitary adenomas were classi ed into invasive and non-invasive tumors, according to the degree of lateral extension to the cavernous sinus (CS) space by MRI scanning (Cottier et al., 2000).Knosp grade 3 and 4 were de ned as invasive pituitary adenomas, and Knosp grade 0 to 2 were de ned as non-invasive tumors, respectively (Knosp et al., 1993).Tumor volume was determined by (length x width x height x Π)/6.The clinical characteristics of patients were described in Table 1.

Detection of mutationsin theGNAS gene
Genomic DNA was extracted from 44 GHPA and 10 NFPA tissues using a DNA miniprep kit according to the manufacturer's protocol (Qiagen GmbH, Hilden, Germany).The underlying point mutations in the GNAS gene have been reported in tumor specimens were CGT-to-TGT mutation at codon 201 (Arg201Cys) and CAG-to-CTG mutation at codon 227 (Gln227Leu) (Goto et al., 2014).PCR ampli cation of codon 201 and 227 was performed using a Taq DNA-Polymerase (TTH Biotools Madrid, Spain) as previously described (Goto et al., 2014).The PCR products were puri ed by a PCR puri cation kit (Qiagen GmbH), and then directly sequenced by an ABI3730XL analyzer (Applied Biosystems, Thermo Fisher Scienti c, Inc., Carlsbad, USA).The primer sequences from PCR and DNA sequencing were listed in Table 2.
Cell culture and transduction GH3, a rat GH-secreting pituitary tumor cell line, which produces both growth hormone and prolactin, was purchased from the Cell Culture Centre, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China).GH3 cells were cultured in Ham's F12 mediumsupplemented with 10% FBS and 1% streptomycin and penicillin in a humidi ed 5% CO 2 incubator at 37°C.A pWPT lentiviral expression vector was used to clone the wild-type or mutant-type GNAS gene and the generated construct was termed as follows: pWPT-GNAS (expressing GNAS wild-type), pWPT-GNAS-Q227L (expressing the mutated GNAS at Q227L), and pWPT-GNAS-R201C(expressing the mutated GNAS at R201C), respectively.GH3 cells were plated in 6-well plates at 70% confluence and then injected with lentivirus to express the wild-type or mutant-type GNAS.
RNA extraction and quantitative reverse transcription PCR (RT-qPCR) Total RNAs were isolated from tissues or cells using Trizol reagent and reverse transcribed into complementary DNA (cDNA) using the TaqMan MicroRNA Reverse Transcription Kit (TaKaRa, Dalian, China).RT-qPCR was performed using an SYBR Green PCR Master Mix (Takara, Japan) according to the manufacturer's instructions.The sequences of qPCR primers were listed in Table 2.The lncRNA MEG3 was normalized by β-actin and the level of MEG3 in GHPA was further normalized by its level in NFPA.

RNA-seq
Total RNA was quali ed by Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), and the next-generation sequencing library was prepared according to the protocol provided by the manufacturer (NEBNext® Ultra™ RNA Library Prep Kit for Illumina® HiSeq system).The sequences and data analysis were performed by Genewiz, China.

Western blots
Proteins were extracted from tumor cells and tissues within RIPA buffer (Beyotime Biotech., Shanghai, China), and separated in SDS-Page gels and transferred onto polyvinylidene fluoride membranes.After blocking with 5% fat-free milk in a Tris-buffered saline with 0.1% Tween 20, the membranes were incubated at 4°C overnight with the primary antibodies against MMP-2, β-catenin, MMP-9, and β-actin, which were purchased from Cell Signaling Tech., Danvers, MA, USA).Subsequently, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody (Cell Signaling Tech.).The images were visualized using enhanced chemiluminescence (Beyotime).Pierce ECL Western Blotting Substrate (Thermo Scienti c) was used to detect the chemiluminescence signals.Densitometric analyses of the western blot bands were performed using the Bio-Rad Imaging system (Bio-Rad, Hercules, California, USA).

Transwell assay
Matrigel matrix diluent (300 μg/mL) was used to coat the bottom of the upper chamber of the Transwell chamber and in a pre-cooled environment at 4℃.The volume ratio of serum-free F12 medium to matrigel in the upper chamber was 4:1, the total volume was 100μl, and the matrigel was frozen overnight at 4℃.The next day, 500 μl of F12 medium containing 10% FBS was added into the lower chamber.Cells (1×10 5 ) were seeded in the upper chamber.After incubation for 48h, non-invading cells in the upper chamber were removed.Cells that invaded the bottom chamber were xed with 4% paraformaldehyde and stained using 0.1% crystal violet (Beyotime).Five random elds were selected for counting purposes under a microscope (magni cation, ×200).

Immuno uorescence
Tumor tissue sections from surgical resection of GHPA.Tissue slides were xed with 4% formaldehyde, blocked using 10% normal goat serum.Primary antibodies against β-catenin, E-cadherin, N-cadherin, Vimentin were diluted at 1:100 in PBS and added to the slides.All the antibodies were obtained from Cell Signaling Technology.After incubation overnight at 4˚C.After removal from the incubation chamber, slides were washed thrice with PBST.Then sections were incubated for 40 minutes in the dark in a humidi ed chamber at room temperature with goat polyclonal secondary antibody to rabbit IgG (1:300, Abcam Biotechnology, USA) reconstituted in PBS.Sections were washed three times with PBST.After the counterstaining procedure, sections were treated with glycerol/PBS (2:1) for 10 minutes in the dark at room temperature.counted in ve randomly chosen elds using an Axiovert 200 uorescent microscope.

Subcutaneous xenografts in nude mice
The animal experiments were approved by the Animal Experimentation Ethics Committee of the Jinling Hospital of Nanjing University.
Four-week-old female athymic BALB/c nude mice were purchased from the Shanghai SLAC Laboratory Animal Co. Ltd (Shanghai, China) and were housed and maintained in laminar air ow chambers under speci c pathogen-free conditions.GH3 cells with different levels of MEG3 (10 7 cells/0.1ml)were subcutaneously injected into the right back side of mice.One week after the injection.After the tumor formed, lithium chloride (60 mg/kg/d in 100 μl saline), a β-catenin activator, was administered daily by intraperitoneal injecting in the group with the high levels of MEG3.The other groups were injected with saline alone as the controls.Tumor volumes were measured with a vernier caliper twice a week and calculated as (length × width 2 )/2.Four weeks after injection, the mice were sacri ced by cervical dislocation, and dissected tumors were weighed and processed to determine the expression levels of the relative proteins, including β-catenin, E-cadherin, N-cadherin, MMP-2, MMP-9.

Statistical analysis
Statistical data analysis was done using SPSS 19.0.The Student's unpaired T-test and Fisher's exact test were used for intergroup analysis.The results were presented as mean ± standard deviation and the correlation was analyzed using a Spearman's correlation.P<0.05 was considered as statistical signi cance.

Results
The GNAS mutations are collected toGHPA Total 44 patients with GHPA and 10 patients with NFPA were enrolled in this study.The GNAS mutations were scanned by directly sequencing the genomic PCR products ampli ed from the patient's tumor sampleswith the speci c primer sets.The results showed 16 out of the 44 cases carrythe GNAS mutations, including 8 cases in codon 227 but other 8 cases in codon 201.In contrast, noGNAS mutation was detected in all 10 NFPA patients.Subsequently, the patients were divided into two groups (GNAS-WT, n=28, GNAS-MUT, n=16).Compared to the GNAS-WT group, the tumor volumes were signi cantly reduced in the GNAS-MUT group.Additionally, the percentage of invasive tumors (Knosp grade 3-4) in the GNAS-MUT tumors was also remarkably decreased (Table 1).There were no statistical differences in age or gender between the two groups.

The mutant GNAS upregulatesMEG3expression
It was speculated that the GNAS mutations may participate in the upregulation of MEG3 expression.To verify the prediction, the expression levels ofMEG3 in NFPA and GHPA tumor tissues were quanti ed by RT-qPCR.As expected,the levels of MEG3 strikingly increase in 44 GHPA tumor tissues compared to 10 NFPA tumor tissues (Fig. 2A).Intriguingly,MEG3 expression further highly increases in the 16 GHPA tumor tissues with mutant GNAS compared to 28 wild-type tumors (Fig. 2B).Consistently, the high levels of MEG3 were veri ed in the GHPA tumor tissues with a single mutant site atQ227L and R201C, respectively(Fig.2C).Importantly,tumor invasiveness was signi cantly declined in the high level of MEG3group compared to the low level of the MEG3 group (Fig. 2D), suggesting that MEG3 may negatively correlation with tumor cell invasion in theGHPAs.

MEG3inhibits the invasiveness of GH3 cells
To verify if MEG3 is able to inhibit cell invasion, MEG3 was manipulated using lentiviral expression systems, either ectopically expressed or knocked down in GH3 cells.After lentivirus infection the cells, MEG3 was quanti ed byimmuno uorescenceand RT-qPCR (Fig. 3A).Expectedly,compared tolentiviral vector control, theoverexpression of MEG3 resulted in decreasing cell invasion.In contrast, the silence of MEG3 led to increased cell invasion (Fig. 3B).Consistently, the protein expression levels ofMMP-2 and MMP-9 were decreased inMEG3-overexpressed cells, while their levels increased in MEG3-silenced cells(Fig 3C).These results suggest that the GNAS mutations inhibit the invasiveness of GHPA tumors partially through activating MEG3.

MEG3inhibits cell invasion via inactivating the Wnt/β-catenin signaling pathway
To further investigate the mechanism by which MEG3 inhibits GHPA cell invasion, MEG3 was overexpressed in GH3 cells and the RNA expression pro le was analyzed by RNA-seq.The results revealed that the Wnt/β-catenin signaling pathway is potential in the involvement of cell invasion regulation (Fig. 4A).Consistently, The mRNA level of β-catenin signi cantly decreased in GH3 cells with the GNAS mutations.Ectopic expression ofMEG3 inGH3 cellsresulted in reducingβ-catenin expression, conversely, the silence of MEG3 inGH3 cellsled to increasing the level ofβ-catenin (Fig. 4B and C).Likewise,the level of β-catenin inGHPAtumors carrying theGNAS mutations was apparently lower than that in wild-type tumors, which was associated with the high level ofMEG3 (Fig. 4D and E).The results suggested that MEG3 negatively regulatesβ-cateninand promotes cell invasion, particularly inGHPA cells carrying theGNAS mutations.
The GNAS mutations inhibitthe epithelial-to-mesenchymal transition (EMT) process EMT is widely recognized to play a fundamental role in the promotion of cell mobility and tumor metastasis and the Wnt/β-catenin signaling pathway is a key mechanism underlyingEMT (Li et al., 2019,Ghahhari andBabashah, 2015).To reveal the effect of the GNASmutations on the EMT process, the expression of β-catenin regulated EMT-associated proteins was quanti ed by immuno uorescence, such E-cadherin as a cell adherent marker, N-cadherin and Vimentin as mesenchymal markers.The results showed that E-cadherin increased but N-cadherin and Vimentin decreased inGHPA tumors carrying theGNAS mutations and the high constitutive levels ofMEG3, as compared to the wild-type of GHPA tumors(Fig.6A and B).Altogether, our nding suggests that the GNAS mutations inhibit the invasiveness of GHPA via downregulation of EMT, as illustrated in Fig. 6C.

MEG3inhibiting the cell invasion was validated in vivo
To further verify that MEG3 suppresses the GHPA cell invasion by inhibiting β-catenin-regulated EMT, βcatenin was manipulated by overexpressing MEG3 in GH3 cells and then treating withLicl.The cells were subcutaneously injected into nude mice for tumor formation.Compared to the control group injected with GH3 cells, the tumor volume and weight were reduced inMEG3-overexpressed cells, but dramatically increased byLicl, indicating thatLiclsu ciently abrogated the negative effect of MEG3 in the regulation of β-catenin (Fig. 7A-C).Consistent with the regulated levels of β-catenin in the formed tumors, E-cadherin increased by overexpressing MEG3, but further decreased by treating withLicl.In contrast, the expression levels of N-cadherin,β-catenin, MMP-2 and MMP-9 decreased by elevating MEG3 and increased byLiclmediated induction (Fig. 7D and E).The results con rmed that MEG3 negatively regulates EMT via downregulating β-catenin.

Discussion
GHPA is a typically benign tumor with a high incidence and large economic burden and often manifests with invasive growth (Katznelson et al., 2014).Numerous studies indicated that the presence of the paradoxical GNAS point mutations strongly re ects the biological characteristics of GHPAs, such as a tendency for densely granulated tumors and smaller tumor size (Landis et al., 1990, Spada et al., 1990, Spada et al., 1991).A previous study has suggested that Gsα protein encoded by the GNAS gene is a key for activating the cAMP-dependent pathway in pituitary cells for differentiation and proliferation (Billestrup et al., 1986).However, the potential role and explicit mechanism of GNAS gene mutation in the invasiveness of GHPAs remains to be fully elucidated.In this regard, the nding from this study indicated that the incidence of invasiveness was markedly reduced in GHPA tumors carrying the GNAS mutations, compared to the wild-type of GNAS.Furthermore, GH3 cells possessing Q227L or R201C mutation appeared a lower percentage of invasive cells in comparison with the corresponding wild-type control cells.Thus, it is speculated that GNAS gene mutations inhibit the invasiveness of GHPA cells.
The GNAS mutations have been proposed to involve the constitutive activation of cAMP formation, which plays a causal role in pituitary adenomas (Mantovani et al., 2010).Interestingly, MEG3 was identi ed as a tumor suppressor that is a downstream target of cAMP (Zhao et al., 2006, Zhang et al., 2010, Ma et al., 2019).The previous evidence indicated that the level of MEG3 is uniquely high in GHPA, but not in NFPAs (Gejman et al., 2008).Thus, we have speculated that the GNAS mutations suppress the invasiveness of GHPA mainly through activating MEG3.As expected, the high levels of MEG3 were only detected in GHPA tumors carrying the GNAS mutations.The MEG3 levels were also signi cantly increased in GH3 cells expressing GNAS gene mutations, compared to the cell expressing the wild-type of GNAS gene, suggesting that the GNAS mutations inhibit the GHPA cell invasion through MEG3 activation.
Furthermore, to ascertain that the effect of MEG3 in cell invasion, we manipulated MEG3 in GH3 cells.Ectopic expression of MEG3 resulted in reducing the cell invasion, and vice versa, the silence of MEG3 led to enhancing the invasiveness.Altogether, our results suggest that MEG3 plays an important role in the promotion of GHPA invasiveness.
The canonical Wnt/β-catenin signaling pathway is thought to be a key regulator in the EMT process and tumor progression (Krishnamurthy and Kurzrock, 2018).Upon activation of the Wnt pathway, β-catenin accumulates in nuclei and functions as a factor (Nusse and Clevers, 2017).In the present study, we showed that the GNAS mutations lead to an increase of MEG3 but a decrease of β-catenin.Furthermore, the activation of β-catenin by Licl enhanced the cell invasion and the inactivation of β-catenin by Dkk1 inhibited the cell invasion.Subsequently, Elevated MEG3 leads to downregulation of β-catenin in GHPA cells.In parallel, the silence of MEG3 upregulated the β-catenin expression, suggesting that MEG3 inhibits the invasiveness of GHPA cells by inactivating the Wnt/β-catenin pathway.
Anterior pituitary with an epithelial phenotype expresses multiple cadherin proteins like E-cadherin that functions for cell attachment (Fougner et al., 2010).EMT is a vital mechanism underlying tumor cell invasion and metastasis.It has been well documented that the loss of E-cadherin and/or the increase of N-cadherin are hallmarks of EMT (27,28).In a variety of types of tumors, the Wnt/β-catenin signaling pathway is constitutively active to promote EMT (Liang et al., 2017).The present study uncovered that βcatenin and EMT-related functional proteins are altered by MEG3 in GHPA tumors carrying the GNAS mutations.Accordingly, enforcedly elevated MEG3 in GHPA cells led to the upregulation of E-cadherin but downregulation of N-cadherin, and Vimentin via altering β-catenin transcriptional regulation.
In addition, MMPs, the important proteolytic enzymes in the degradation of extracellular matrix and basement membrane, are crucial for tumor cell invasion (Di Nezza et al., 2002).Among of MMPs, MMP-2 and MMP-9 have been demonstrated to play vital roles in tumor invasion due to their potent ability to degrade collagen types IV (Scheau et al., 2019).In particular, their functions have been involved in the invasiveness of GHPA tumors (Yang and Li, 2019).Similar to other EMT-related proteins, our study further showed the levels of MMP-2 and MMP-9 decrease in GHPA tumors carrying the GNAS mutations, as well as the increase of MEG3 expression in GHPA cells resulted in decreasing the expression of MMP-2 and MMP-9, con rming the previous nding that MMPs participate in the progression of GHPA.Furthermore, how MEG3 regulates β-catenin-mediated transcriptional activation is being currently investigated.

Conclusions
In summary, this study revealed that GNAS mutations inhibit the invasiveness of GHPA tumors by increasing the level of MEG3.The upregulation of MEG3 in GHPA cells supresses the cell invasion capacity through inhibiting the Wnt/β-catenin signaling pathway.The silence of MEG3 upregulates βcatenin and enhances EMT.The nding suggests that MEG3 may serve as a biomarker for the detection of GHPA phenotype and inhibition of the Wnt/β-catenin signaling pathway may provide a useful therapeutic approach in the intervention of GHPA.The    show the signi cances between the two groups as indicated.
The GNAS mutations inhibit EMT in GHPA tumors.A and B The expression of EMT-related proteins in GHPAs tumor tissues were quanti ed by immuno uorescence (magni cation, x200).C: Depiction of the suggested mechanism underlying the GNAS mutations inhibiting the invasiveness of GHPA.

Figure 2 The
Figure 2