Selective estrogen receptor modulators decrease invasiveness in pituitary adenoma cell lines AtT‐20 and TtT/GF by affecting expression of MMP‐14 and ADAM12

Clinically, pituitary adenomas (PAs) become detrimental when invasive. We conducted an efficacy analysis of estrogen receptor modulators (SERMs) in PA cells. All SERMs blocked invasiveness of PA cells, correlating with expression levels of invasion‐related metalloproteases ADAM12 and MMP‐14. Our findings support the notion that SERMs are able to block invasiveness of PA cells by reducing their expression levels.

Pituitary adenomas (PAs) may exert a mass effect on neighboring tissues, and they may lead to endocrine dysfunction or cranial nerve deficiencies. Therapeutic options comprise surgery, radiation, and pharmacotherapy. The extent of surgical resection is often limited in invasive or recurrent PAs. Radiation therapy, which is usually applied with remnants or recurrence of PAs at an inoperable site, may affect neighboring tissues and induce tumors. Pharmacotherapy is useful only in certain PAs, depending on the hormones and receptors expressed in the tumor cells [1].
In 2011, McCormack et al. [2] published a careful analysis of reports on the potential usefulness of temozolomide (TMZ) in aggressive PAs. In 2018, the European Society of Endocrinology has recommended TMZ as first-line chemotherapy in cases of aggressive PAs that are refractory to standard treatment [3]. However, a positive treatment effect of TMZ was only observed in 47 percent of cases [3]. Other than that, there is no widely recommended or evidence-based chemotherapeutic option to treat aggressive PAs.
Several in vitro studies suggest antitumorigenic effects of selective estrogen receptor modulators (SERMs) like resveratrol [4][5][6] or clinically applicable drugs like bazedoxifene, clomiphene, and raloxifene [7] on rodent PA cells. In acromegalic men, the SERM raloxifene has been shown to reduce insulin-like growth factor 1 levels [8]. There are indications that SERMs have an impact on cell-extracellular matrix interaction which may subsequently affect invasiveness, for example, in breast cancer cells: Tamoxifen, a SERM, has been shown to reduce extracellular transforming growth factor-beta 1 (TGF-beta1) expression, while inhibition of matrix metalloproteinase (MMP) activity restored TGF-beta1 levels in estrogen receptor-positive breast cancer cells [9]. Reduced disease-free survival time has been reported in patients with estrogen receptor-negative and MMP-9-positive breast cancer [10]. Induction of MMP-9 expression by laminin in a human cervical cancer cell line [11] and upregulation of MMP-9 as well as coordinated expression of MMP-2 and MMP-14 by fibronectin in a human T lymphocyte cell line [12] have been demonstrated. MMP-14 and a disintegrin and metalloproteinase-12 (ADAM12) levels in human PA tissue samples are correlated with cavernous sinus invasion [13].
Here, we aim to investigate the impact of three clinically applicable SERMs, namely bazedoxifene, clomiphene, and raloxifene, on invasiveness in vitro and correlate their effects with gene and protein expression levels of MMP-14 and ADAM12 in two murine PA cell lines mainly expressing estrogen receptor alpha (ER-α, 7), AtT-20 and TtT/GF. Since to date, no human PA cells are available, these cell lines might serve as a model for human studies.

Viability assay
One hundred microlitre medium per well containing 5000 cells per well was placed in 96-well plates for 24 h. Thereafter, SERMs were added at the indicated concentrations. After another 48 h, 50 µL CellTiter-Glo ® (Promega, Mannheim, Germany) per well were added. The plate was then vigorously shaken for 10 min and incubated for 20 min at room temperature in the dark. A FLUOstar OPTIMA Microplate Reader (BMG Labtech, Offenburg, Germany) was employed to detect luminescence. All experiments on viability were conducted in triplicates.

Invasion assay
Eight-micrometre transwell inserts (Greiner Bio-One, Frickenhausen, Germany) were used with 24-well plates. The inserts were coated on the upper side with 50 μL Matrigel (Corning ® Matrigel ® Matrix, Basement Membrane Matrix Growth Factor Reduced, Corning Incorporated, Corning, NY, USA). After 1 h at 37°C under 5% CO 2 , the Matrigel was solid. The transwell insert was then turned upside down, and 25 000 TtT/GF cells per well resp. 50 000 AtT-20 cells per well in 50 μL medium containing 0.5% FBS were seeded on the other side. After 4 h allowed for adherence, the transwell insert was turned upside up again. To build an FBS gradient, 250 μL medium containing 10% FBS was applied on top of each Matrigel layer, while 750 μL medium containing 0.5% FBS was placed in each of the 24 wells. After 24 h allowed for invasion, cells were fixed with 4% paraformaldehyde (Sigma), treated with 0.3 % octoxynol-9 (Triton ™ X-100 buffer; Sigma), and stained with 4 0 ,6-diamidin-2-phenylindole (DAPI; Sigma). Cells were counted in five randomly chosen viewing fields [respectively 10 randomly chosen viewing fields with the small interfering ribonucleic acid (siRNA) experiments]. The percentage of cells that had invaded into the Matrigel was determined.
Quantitative real time-polymerase chain reaction (qPCR) tube and gently mixed with 500 µL isopropanol. After 10 min at room temperature, tubes were centrifuged again (15 600x g, 15 min, 4°C). The supernatant was removed, and the pellet was washed with 1 mL 75% ethanol followed by centrifugation and careful removal of the supernatant. The sediment was dissolved in RNAse free water. RNA quantification was performed using a NanoPhotometer NP80 (Implen, Munich, Germany). cDNA was obtained using the 'RNA to cDNA Ecodry ™ Premix Kit' (Takara Bio Europe, Saint-Germain-en-Laye, France) according to the manufacturer's protocol. Quantitative polymerase chain reaction (qPCR) analyses were performed in triplicates using the 'Precision FAST MasterMix with ROX' (Primer Design, Southampton, UK). QuantiTect primers for the detection of RNA encoding mouse MMP-1, MMP-9, MMP-14, ADAM8, ADAM12, and Basigin were obtained from Qiagen. We used XS13 as housekeeping gene with primers (fwd-TGGGCAAGAACACCATGATG, rev-AGTTTCTCCAGCTGGGTTG) purchased from Microsynth (Balgach, Switzerland). A StepOnePlus ™ qPCR system (Thermo Fisher Scientific, Waltham, MA, USA) was employed. Results were normalized to blank controls using the 2 ÀΔΔCt method to estimate relative quantities for each gene.
Proteins from equal amounts of lysate were separated by SDS/PAGE. Proteins were then transferred onto nitrocellulose membranes. To block any unspecific binding, membranes were immersed and gently shaken in 5% nonfat, dried milk in Tris-buffered saline Tween-20 [TBST; 50 mM Tris, pH = 7.5; 150 mM NaCl; 0.1% Tween-20 (CARL ROTH, Karlsruhe, Germany)] for 1 h, after which they were incubated with primary antibodies at 4°C overnight. After three washing steps with TBST, blots were incubated with the secondary antibody for 1 h. After another washing step, signals were detected with Western Bright Chemiluminescence Substrate Sirius (Biozym Scientific, Hessisch Oldendorf, Germany) and chemiluminescence was determined using a Chemostar Imager (Intas, Göttingen, Germany) and IMAGEJ (National Institutes of Health (NIH), Bethesda, MD, USA) for quantification. Several exposure times were compared to confirm linearity of the signals.

siRNA transfection
FlexiTube Gene Solution siRNA for mouse ADAM12, mouse MMP-14, and AllStars-negative control siRNA (Qiagen) was applied in AtT-20 and TtT/GF cells. Transfection was carried out using HiPerFect reagent (Qiagen) according to the manufacturer's protocol with 5 nM siRNA applied. The resulting silencing of proteins was confirmed by western blotting as described above. Invasion rates were established according to the invasion assay protocol given above. means, ANOVA and t testing with a correction for type I error were applied.

Results
At first, we determined dose-response curves for the three SERMs bazedoxifene, clomiphene, and raloxifene in AtT-20 and TtT/GF cells. According to the 50% inhibitory concentration (IC 50 ) values, we found that in AtT-20 cells, bazedoxifene exerted the most pronounced impact on viability, while in TtT/GF cells clomiphene had the highest impact on viability which was comparable to that of bazedoxifene in the same cells (Fig. 1A,B, Table 1).

SERMs decrease invasiveness in murine PA cells
Invasion assays using Matrigel were performed with AtT-20 and TtT/GF cells ( Fig. 2A). We observed a significant impact on invasiveness after treatment of AtT-20 cells with clomiphene (P = 0.027) and raloxifene (P = 0.009) and after treatment of TtT/GF cells with clomiphene (P = 0.022) and bazedoxifene (P = 0.002) at the respective IC 25 (Fig. 2B,C). The effects of the three SERMs at IC 25 were compared to control media and to the effect of batimastat, a broadrange MMP inhibitor, which also exerted a significant effect on invasiveness in both cell lines (P < 0.01, Fig. 2B,C).    (Fig. 3). These results prompted us to particularly investigate the potential impact of SERMs on MMP-14 and ADAM12 gene expression in AtT-20 and TtT/GF cells. We could demonstrate a significant impact of bazedoxifene (P < 0.01), clomiphene (P < 0.01), and raloxifene (P < 0.01) on mRNA gene expression of ADAM12 in AtT-20 cells. We further observed a significant impact of bazedoxifene (P < 0.01) and raloxifene (P = 0.01) on the mRNA gene expression of MMP-14 in TtT/GF cells, as well as a significant impact of bazedoxifene (P < 0.001) and raloxifene (P < 0.001) on the mRNA gene expression of ADAM12 in TtT/GF cells, each as compared to blank controls (Fig. 4A,C,D). With or without SERMs, a relevant gene expression of MMP-14 in AtT-20 cells was not detected (Fig. 3, data after SERM treatment not shown). In either cell line, the gene expression levels of ADAM-related Basigin were not significantly downregulated by SERM treatment (Fig. 4B,E).

Gene silencing of ADAM12 and MMP-14 decreases invasiveness of murine PA cells
To evaluate the role of ADAM12 and MMP-14 in PA cell invasion, we analyzed the impact of ADAM12 and MMP-14 gene silencing on invasiveness in AtT-20 and TtT/GF cells. Transfection of AtT-20 and TtT/GF cells with ADAM12 and MMP-14 siRNA resulted in a significantly reduced protein expression of ADAM12 in both cell lines (Fig. 5A,C) and of MMP-14 in TtT/ GF cells (Fig. 5C). These cells were subjected to invasion assays and showed significantly decreased invasion rates of AtT-20 cells after ADAM12 silencing (P = 0.013, Fig. 5B), of TtT/GF cells after ADAM12 silencing (P < 0.001, Fig. 5D) and of TtT/GF cells after MMP-14 silencing (P = 0.011, Fig. 5D).

SERM treatment decreases MMP-14 and ADAM12 protein levels in murine PA cells
In AtT-20 cells, treatment with bazedoxifene, clomiphene, and raloxifene at IC 50 each obviously decreased ADAM12 protein expression (Fig. 6A) as compared to DMSO and blank controls. In this cell line, an impact of the three drugs on MMP-14 protein expression was not found according to the western blot (Fig. 6C) which, in AtT-20 cells, hardly detected any MMP-14 protein at all-in accordance with the qPCR results (Fig. 3).
In TtT/GF cells, treatment with bazedoxifene, clomiphene, and raloxifene at IC 50 each obviously decreased MMP-14 protein expression as compared to the blank control (Fig. 6D). In the same cells, treatment with bazedoxifene and raloxifene at IC 50 also obviously decreased MMP-14 protein expression as compared to the DMSO control, whereas treatment with clomiphene at IC 50 exerted only a moderate effect as compared to the DMSO control. A significant impact on ADAM12 protein expression in TtT/GF cells was only found for raloxifene according to the western blot (Fig. 6B).

Discussion
The impact of SERMs on viability and invasion in rodent PA cells and the correlation of MMP-14 and ADAM12 expression with invasiveness in human PAs have been demonstrated previously [4][5][6][7]13]. However, this is to our knowledge the first study that connects SERMs with reduced expression levels of MMP-14  Table 1   A significant impact on invasion was observed after treatment of AtT-20 cells with clomiphene and raloxifene and after treatment of TtT/GF cells with bazedoxifene and clomiphene. However, all the tested drugs moderated the invasive potential in both cell lines at least to some extent at IC 25 . These observations are somewhat different from those reported in [7], which may be explained by differences in incubation times and cell passages between studies.
ADAM12 and MMP-14 have been found to be related to cavernous sinus invasion in human PA tissue samples [13] as well as to growth in other tumors, for example, breast cancer [14]. We found the gene expression levels of ADAM12 in AtT-20 and TtT/GF cells and of MMP-14 in TtT/GF cells to be significantly downregulated after treatment with bazedoxifene, clomiphene, and raloxifene. Furthermore, we found that bazedoxifene and raloxifene decreased protein expression of MMP-14 in TtT/GF cells, and all three drugs decreased protein expression of ADAM12 in AtT-20 cells whereas only raloxifene significantly dowregulated protein expression of ADAM12 in TtT/ GF cells.
There are several limitations associated with this study. The absence of MMP-14 gene expression in AtT-20 cells as opposed to TtT/GF cells and the difference in ADAM12 protein expression in AtT-20 cells as opposed to TtT/GF cells indicate that there are distinct mechanisms of invasion involved in each of the cell lines, deserving future elucidation, including  investigation of the potential impact of SERMs on other invasion-related proteases. Another limitation is that it remains difficult to extrapolate from findings in rodent PA cells in vitro to potential applications in humans, since several challenges have to be overcome: In different PA cell lines from the same species, the same treatment may lead to distinct responses, as we have observed here. One may expect such differences to be similar or even more pronounced in other species. Furthermore, cultivation of human PA cells is still a major obstacle. Last but not least, apart from intravenously administered clomiphene, the concentrations of the drugs tested in this study are much higher than plasma concentrations usually achievable in humans [15][16][17] (Table S1).
We find it nonetheless very encouraging to see that in murine PA cells, SERMs downregulate two proteases that apparently play key roles in cavernous sinus invasion in human PAs [13].

Conclusion
Selective estrogen receptor modulators decrease invasiveness of murine PA cell lines AtT-20 and TtT/GF and regulate expression levels of MMP-14 and ADAM12, two important invasion-related proteins in murine PA cells as well as in human PA tissue samples. Our findings support the view that SERMs might become a potential salvage therapy for aggressive PAs.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Molecular characteristics of the three selective estrogen receptor modulators (SERMs) that were used to treat AtT-20 and TtT/GF cells in our study a as retrieved from http://www.drugbank.ca on May 21, 2020; b see reference number 15; c see reference number 16; d see reference number 17; for the respective inhibitory concentration (IC) values, please refer to table 1.