The present study retrospectively reviewed 52 patients (mean age, 40.3 years; range 16–62 years) with somatotroph adenomas at Beijing Tiantan Hospital affiliated to Capital Medical University, between July 2012 and December 2016. The diagnosis of somatotroph adenoma was based on clinical symptoms and histopathological features, per the 2017 World Health Organization Classification (14). All patients were diagnosed with sporadic somatotroph adenomas. The inclusion criteria for recruitment of patients were as follows: i) Serum GH >1 ng/ml and IGF-1 above normal (115–307 ng/ml), or GH >2 ng/ml; ii) magnetic resonance imaging confirmed the existence of the saddle area; iii) acromegaly or facial changes; and iv) transmission electron microscopy revealed large secretory granules (400–600 nm). Patients presenting familial pituitary adenoma were excluded from the present study. In addition, three normal pituitary glands were obtained from a donation program at Tianjin First Central Hospital in China (two male and one female patients; age, 21–45). None of the donors had a history of pituitary disease. Progression-free survival (PFS) was defined as the interval between the date of surgery and the date of tumor recurrence or tumor regrowth. The protocols were approved by the Internal Review Board of Beijing Tiantan Hospital affiliated to Capital Medical University and the study was conducted according to the principles of the Declaration of Helsinki (no. KY-2013-02). Informed consent was obtained from all patients.

Tissue microarray construction and the IHC protocol were performed according to a previously described method (15). The following primary antibodies were used: Anti-SLC20A1 (cat. no. ab237527; 1:1,000; Abcam), anti-β-catenin (cat. no. ab32572; 1:2,000; Abcam), anti-Wif1 (cat. no. ab155101; 1:500; Abcam) and anti-Ki 67 (cat. no. ACK02; 1:1,000; Leica Microsystems, Inc.). A secondary antibody (cat. no. sc-2040; 1:5,000; Santa Cruz Biotechnology, Inc.) was used in combination with the Bond Polymer Refine Detection system (cat. no. DS9800; Leica Microsystems, Inc.). Staining intensity was stratified on a scale of 0–3 as follows: 0, no staining; 1, weak staining; 2, moderate staining; and 3, strong staining. The H-score was obtained by multiplying the staining intensity by a constant to adjust the mean to the strongest staining, to produce a score in the range of 0–300. Specifically, H-score=scale × percentage of strong staining (0–100%). H-score=1.0 indicated a weak percentage; 2.0 indicated a moderate percentage; and 3.0 indicated a strong percentage.

GH3 cells were purchased from the American Type Culture Collection (ATCC) and were cultured in F12K (ATCC) supplemented with 2.5% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) and 15% horse serum (Gibco; Thermo Fisher Scientific, Inc.) under 37°C and 5% CO₂.

The GH3 cells were plated onto 96-well dishes with 1×10⁴ cells and 100 µl medium in each well, and incubated overnight. SLC20A1 gene silencing in the GH3 cells was established by transfection of short hairpin RNA (1 µg for 1×10⁶ cells; OriGene Technologies, Inc.) or non-targeting control shRNA (scramble) for 72 h using 5 µl Opti-MEM (Gibco; Thermo Fisher Scientific, Inc.) and 0.3 µl Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's instructions. The sequences of the pGFP-C-shLenti plasmid were: Scramble, TTCTCCGAACGTGTCACGT; A, CACCGAAGTATGACAATCTGTGGATGCTC; B, GCAATGCTGTGTCTGACCTTCACTCAGAG; C, ATAGGAATCCTGTGTCTGAGGTAGTATGT; and D, GCTTCTGCTCCACCGAAGTATGACAATCT. Then, 20 µl MTS solution was added to each well and further incubated for 4 h. The absorbance at 490 nm of each well was measured using an ELISA plate reader (Thermo Fisher Scientific, Inc.).

Cell migration and invasion were measured using fibronectin- and Matrigel-coated polycarbonate filters, respectively, and modified transwell chambers (Corning Inc.). In total, 5×10⁴ GH3 cells in 200 µl F12K medium were added into the upper chambers and 600 µl medium [F12K + 2.5% FBS (cat. no. 10099-141; Gibco; Thermo Fisher Scientific, Inc.)+ 5% horse serum (cat. no. sh30074; HyClone; GE Healthcare Life Sciences] was added into lower chamber. After 24 h, migrating cells that adhered to the lower membrane were fixed in 4% paraformaldehyde (for 30 min at 4°C) and stained using hematoxylin (OriGene Technologies, Inc.) for 5 min at room temperature. The average number of migrated cells was quantified in five randomly selected fields of view under a fluorescence microscope (Zeiss GmbH; magnification, ×100) Experiments were performed in triplicate.

GH level was detected using an ELISA kit (cat. no. 710685; Shanghai Enzyme-linked Biotechnology Co., Ltd.) according to the manufacturer's protocol. The absorbance at 450 nm in each well was measured using an ELISA plate reader (M200 Pro; Tecan Group, Ltd.).

Total RNA was extracted from 24 frozen pituitary adenoma samples (~10 mg) using the RNeasy^® Mini Kit (Qiagen GmbH). Purified total RNA (1 µg) was reverse transcribed into complementary DNA (cDNA) using a Revert Aid First-Strand cDNA Synthesis kit (Thermo Fisher Scientific, Inc.), based on the manufacturer's protocol (25°C for 10 min, 37°C for 2 h, 95°C for 5 min). The primer details are presented in Table I. GAPDH was used as an endogenous control for normalizing the levels of target genes. RT-qPCR was performed in an ABI 7500 Fast Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.) using Platinum^® SYBR^® Green qPCR SuperMix-UDG (Invitrogen; Thermo Fisher Scientific, Inc.). The comparative quantification cycle (Cq) method was used to calculate the fold-change in differential expression of each gene (2^−ΔΔCq method) (16) the thermocycling conditions were the following: Initial denaturation at 95°C for 15 followed by 40 cycles of 94°C for 15 sec, 55°C for 30 sec and 74°C for 34 sec.

A total of 10 mg specimen was lysed to extract total protein in TNE buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl and 1 mM EDTA; Sigma-Aldrich; Merck KGaA) containing 1% Nonidet P-40 (Calbiochem; Merck KGaA) with protease and phosphatase inhibitor cocktails (Roche Applied Science). The total protein extracted from somatotroph adenomas or normal pituitary glands was centrifuged at 12,000 × g for 30 min at 4°C, and the protein concentration was determined using a bicinchoninic acid protein assay kit (Pierce; Thermo Fisher Scientific, Inc.). For western blot analysis, 40 µg protein per lane was separated electrophoretically using 4–12% Bis-Tris SDS-PAGE gels and blotted onto polyvinylidene fluoride membranes. Each membrane was blocked for 1 h with 5% milk in TBS at room temperature, incubated with antibodies against SLC20A1 (cat. no. ab237527, 1:2,000; Abcam), Wif1 (cat. no. ab155101, 1:1,000; Abcam), β-catenin (cat. no. ab32527, 1:5,000; Abcam), sFRP4 (cat. no. ab154167, 1:2,000; Abcam) and GAPDH (cat. no. G5262, 1:8,000; Sigma-Aldrich; Merck KGaA) at 4°C overnight, and then incubated 1 h with horseradish peroxidase-tagged secondary antibodies (cat. no. sc-2363, 1:5,000; Santa Cruz Biotechnology, Inc.) at room temperature. Finally, the membranes were visualized by enhanced chemiluminescence (cat. no. sc2048, Santa Cruz Biotechnology, Inc.) using a Versadoc XL imaging system (Amersham Imager 600; GE Healthcare Life Sciences), and densitometry was performed using the ImageQuant TL software (version 7.0; GE Healthcare Life Sciences). GAPDH levels were analyzed as a loading control.

The χ² exact test was used to determine the significance of clinicopathological characteristics and variables. Pearson's correlation coefficient was used to analyze the correlation between β-catenin/Wif1 expression and SLC20A1 score. Student's t-test was applied to examine the differential expression of SLC20A1, β-catenin, Ki67 and Wif1 in patient samples. Data are presented as the mean ± SD. One-way ANOVA was used in cell growth, GH release and migration of GH3 cells experiments. All P-values were two sided and P<0.05 was considered to indicate a statistically significant difference. All experiments were repeated three times. Analyses were performed using SPSS (vesion 19.0; IBM Corp.).

There were 28 males and 24 females in the present study (Table SI) with a mean age of 40.3 years (age range, 17–62 years). The median tumor volume was 6.25 cm³ (range, 2.24–18.7 cm³). The average level of GH was 22.83±7.31 ng/ml, and ranged from 7.9 to 76.2 ng/ml. The number of total resections was 31/52 (59.6%), while the number of partial resections was 21/52 (40.4%). The number of recurrent cases was 11/52 (21.2%). Tissues from 52 patients were stained with standard hematoxylin and eosin staining, and IHC was performed with antibodies against SLC20A1, β-catenin, sFRP4, Ki67 and Wif1. According to the Knosp classification, the 52 patients were classified into an invasive group (15 patients) and a non-invasive group (37 patients). The patients in the invasive group had a greater degree of headache and visual deficits than the patients in the non-invasive group. The preoperative serum GH level was 36.7±11.9 ng/ml in the invasive group and 17.2±6.3 ng/ml in the non-invasive group (P<0.05). Follow-up data revealed that total resection was carried out in 5/15 (33.3%) cases in the invasive group and 26/37 (70.3%) cases in the non-invasive group (χ²=4.611, P=0.032). Furthermore, recurrence in the invasive group occurred in 6/15 (40%) cases and in 5/37 (13.5%) cases in the non-invasive group (χ²=4.489, P=0.034) (Table II).

IHC was used to evaluate the levels of SLC20A1, Wif1, β-catenin and Ki67 in the 52 somatotroph adenoma samples (Fig. 1A). The H-score of SLC20A1 was 222.6±15.2 in the invasive group and 144.5±30.4 in the non-invasive group (P<0.01), while the H-scores of β-catenin were 210.1±21.4 and 134.9±32.7 in these two groups, respectively (P<0.05). The H-scores of Wif1 presented the opposite trend compared with β-catenin or SLC20A1, with values of 134.5±22.7 and 253.6±14.8 in the invasive and non-invasive groups, respectively (P<0.01). The Ki67 index was 6.3±2.2 in the invasive group and 1.9±1.3 in the non-invasive group. Western blotting also confirmed the same tendency, as depicted in Fig. 1B. In the present study, a high-level case was defined as having a H-score higher than the median value.

According to the H-score median of SLC20A1, Wif1 and β-catenin, 52 samples were divided into high and low H-score groups. The tumor volume in patients with high SLC20A1 H-scores (high SLC20A1 group) was 8.34±2.64 and 4.16±2.39 cm³ in patients with low SLC20A1 H-scores (low SLC20A1 group) (P<0.05), and the rate of tumor recurrence was 8/26 (30.8%) compared with 3/26 (11.5%) in these two groups, respectively. Patients with high Wif1 H-scores (high Wif1 group) had less tumor volume (3.14±2.35 vs. 9.36±3.64 cm³) than those with low Wif1 H-scores (low Wif1 group) (P<0.05) and lower recurrence rates compared with patients in the low Wif1 group, data not shown (2/26 vs. 9/26, χ²=5.65, P=0.017). The H-scores of SLC20A1 were negatively associated with the H-scores of β-catenin in somatotroph adenomas, with a correlation coefficient of 0.366 (Fig. 2A). The correlation coefficient was −0.424 between Wif1 and SLC20A1 (Fig. 2B). PFS associated with low SLC20A1 H-scores was longer than that associated with high SLC20A1 H-scores (P=0.033, Fig. 3A), while the PFS in the high Wif1 group was longer than that in the low Wif1 group (P=0.043), as shown in Fig. 3B.

The efficiency of four shRNA segments (Sh-A/B/C/D-SLC20A1) was measured in GH3 cells. RT-qPCR demonstrated that the relative percentages of SLC20A1 mRNAs were 32.3±10.4, 8.2±5.7, 17.7±8.4 and 64.9±22.4%, respectively, after 72 h of transfection (Fig. 4A). Western blotting also confirmed that the interference efficiency of segment Sh-B was 83.4% and that of Sh-C was 76.4% compared with the SLC20A1 protein levels in GH3 cells after 72 h of transfection (Fig. 4B). Cell proliferation experiments revealed that the cell viability of the Sh-B group was 76.3±4.5, 65.7±3.7 and 53.1±3.2% of control GH3 cells after 24, 48 and 72 h of transfection, respectively, and that of the Sh-C group was 86.4±5.7, 75.6±4.4 and 67.5±3.8% of the control group, respectively, after the same time (P<0.05; Fig. 5A). ELISA was used to measure the effect of SLC20A1 on GH secretion in GH3 cells. The GH levels in the Sh-B and Sh-C groups were 34.7±10.4 and 54.6±14.4% of that in control GH3 cells, respectively (P<0.05; Fig. 5B). The transmembrane invasion assay revealed that RNAi-SLC20A1 significantly suppressed cell invasion in the Sh-B and Sh-C groups (Fig. 5C).

RT-qPCR revealed that SLC20A1 silencing reduced the mRNA levels to 0.105±0.114 and 0.233±0.141-fold for N-cadherin (N-CAD), 0.216±0.185 and 0.346±0.172-fold for vimentin, 0.294±0.155 and 0.387±0.212-fold for vascular endothelial growth factor (VEGF), and 0.344±0.165 and 0.547±0.282-fold for matrix metalloproteinase-2 compared with the control group. There was no significant difference in the expression levels of E-cadherin, matrix metalloproteinase-9 and Snail. In additional, no significant difference was observed between the control group and scramble group (Fig. 5D).

Aberrant expression of the Wif1 and sFRP genes has been reported to be associated with the malignant and invasive characteristics of certain human tumors, including somatotroph adenomas (17,18). RT-qPCR revealed that the mRNA levels of Wif1 in the Sh-B and Sh-C groups were 4.3±1.2 and 3.6±0.7-fold, respectively, of those in the control GH3 cells, while the mRNA levels of sFRP4 were 5.8±1.3 and 4.6±0.9-fold of those in the control GH3 cells (Fig. 6). Western blotting detected that the Sh-B and Sh-C segments of SLC20A1 could evidently increase the levels of Wif1 and sFRP4, and significant decrease the β-catenin level in the Sh-B group (Fig. 6).