The Integrated Stress Response Is Tumorigenic and Constitutes a Therapeutic Liability in Somatotroph Adenomas

Somatotroph adenomas are the leading cause of acromegaly, with the nearly sparsely granulated somatotroph subtype belonging to high-risk adenomas, and they are less responsive to medical treatment. The integrated stress response (ISR) is an essential stress-support pathway increasingly recognized as a determinant of tumorigenesis. In this study, we identified the characteristic profiling of the integrated stress response in translocation and translation initiation factor activity in somatotroph adenomas, normal pituitary, or other adenoma subtypes through proteomics. Immunohistochemistry exhibited the differential significance and the priority of eukaryotic translation initiation factor 2β (EIF2β) in somatotroph adenomas compared with gonadotroph and corticotroph adenomas. Differentially expressed genes based on the level of EIF2β in somatotroph adenomas were revealed. MetaSape pathways showed that EIF2β was involved in regulating growth and cell activation, immune system, and extracellular matrix organization processes. The correlation analysis showed Spearman correlation coefficients of r = 0.611 (p < 0.001) for EIF2β and eukaryotic translation initiation factor 2 alpha kinase 1 (HRI), r = 0.765 (p < 0.001) for eukaryotic translation initiation factor 2 alpha kinase 2 (PKR), r = 0.813 (p < 0.001) for eukaryotic translation initiation factor 2 alpha kinase 3 (PERK), r = 0.728 (p < 0.001) for GCN2, and r = 0.732 (p < 0.001) for signal transducer and activator of transcription 3 (STAT3). Furthermore, the invasive potential in patients with a high EIF2β was greater than that in patients with a low EIF2β (7/10 vs. 4/18, p = 0.038), with a lower immune-cell infiltration probability (p < 0.05). The ESTIMATE algorithm showed that the levels of activation of the EIF2 pathway were negatively correlated with the immune score in somatotroph adenomas (p < 0.001). In in vitro experiments, the knockdown of EIF2β changed the phenotype of somatotroph adenomas, including cell proliferation, migration, and the secretion ability of growth hormone/insulin-like growth factor-1. In this study, we demonstrate that the ISR is pivotal in somatotroph adenomas and provide a rationale for implementing ISR-based regimens in future treatment strategies.


Introduction
Somatotroph adenomas are a kind of functional pituitary adenomas that cause acromegaly due to the excessive secretion of growth hormone (GH) and insulin-like growth factor (IGF-1) [1]. Excess GH/IGF-1 in somatotroph adenoma patients causes the overgrowth of some tissues and multiple metabolic abnormalities by affecting body composition, protein dynamics, and molecular mechanisms in adults [2]. The mortality risk is related to the duration of the illness before diagnosis and the lesional degree of cardiovascular disease and diabetes at the time of diagnosis [3]. Improvements in diagnostic modalities have increased the prevalence of somatotroph adenomas through the identification of active 2 of 15 acromegaly (high IGF-1 and/or GH levels) in clinically asymptomatic patients, also called silent somatotroph adenomas [3]. Compared to other subtypes, up to 50% of somatotroph adenoma patients have clear phenotypic changes, whereas another~50% have either mild symptoms or are visibly asymptomatic for metabolic abnormality, including polyamine anabolism and sugar anabolism [4]. Initial therapy generally constitutes transsphenoidal surgery, although the remission rate is approximately 50%; somatostatin analog (SSA) therapy is reserved for cases not cured by surgery [5]. The WHO 2017 classification suggests that subtypes of pituitary adenoma, such as male lactotrophs, silent corticotroph and Crooke cells, sparsely granulated somatotrophs, and silent plurihormonal POU Class 1 Homeobox 1-positive tumors, should be considered as "high risk" tumors [6].
The integrated stress response (ISR) is an essential stress-support pathway increasingly recognized as a determinant of tumorigenesis via tuning protein synthesis rates [7]. The various stresses are sensed by four specialized kinases-eukaryotic translation initiation factor 2 alpha kinase 1 (HRI), eukaryotic translation initiation factor 2 alpha kinase 2 (PKR), eukaryotic translation initiation factor 2 alpha kinase 3 (PERK), and eukaryotic translation initiation factor 2 alpha kinase 4 (GCN2)-that converge on the phosphorylation of a single serine on the eukaryotic translation initiation factor eIF2 [8]. The salient feature of ISR signaling lies in the modulation of the cellular concentration of the ternary complex (TC), which is composed of the heterotrimeric eukaryotic translation initiation factor and eukaryotic translation initiation factor 2 (EIF2), consisting of an α, β, and γ subunit in equal molar amounts [9]. Phosphorylation of the eIF2 subunit regulates the heterotrimer activity [10]. Oncogenic roles of eIF2α are well-known in chronic myeloid leukemia [11], colorectal cancer [12], prostate cancer [13], and so on. eIF2β forms part of the nucleotide-binding pocket in eukaryotes and appears to bind GTP or GDP [14]. Integrated genomic data in breast cancer has shown that eIF2β is highly amplified and correlates to the aggressive behavior of luminal breast tumors [15]. eIF2γ has a catalytic site for GTPase activity and is involved in binding of the ternary complex to mRNA; the eIF2γ variant is associated with hypothalamo-pituitary development, hypopituitarism, and pancreatic dysfunction and glucose regulation [16]. The ISR is a double-edged sword with both pro-survival and pro-death activities, and regulation of ISR signaling components is a promising strategy in cancer therapy [17,18].
While the ISR is upregulated in many cancers, its precise function in tissue-specific cancer initiation and progression remains poorly understood [19]. In this study, we explored somatotroph adenomas using multi-omic methods and showed the unique characteristics of the eIF2 pathway and ribosome-related genes in somatotroph adenoma patients compared to patients with other subtypes of pituitary adenomas. We demonstrated the relationship of the ISR with disease phenotype and immune cell infiltration in somatotroph adenomas together with the strong therapeutic benefits of its pharmacological inhibition.

Characteristic Profiling in Somatotroph Adenomas Compared to Normal Pituitary, Corticotroph Adenomas, and Gonadotroph Adenomas
Differentially expressed proteins were identified by nanoLC-MS/MS among five normal pituitary tissues, five somatotroph adenomas, five gonadotroph adenomas, and five corticotroph samples. Proteomics analysis identified 46,383 peptides that mapped to 4751 proteins. A total of 274 proteins were differentially expressed (p < 0.05, with an iTRAQ ratio >1.5 or <0.7); 130 were upregulated; and 144 were downregulated in somatotroph adenomas, as shown in Figure 1A,B. Using the Metascape database, pathway and process enrichment analyses showed that the top three pathways were SRP-dependent cotranslational protein targeting to the membrane, peptide chain elongation, and eukaryotic translation elongation, as shown in Figure 1C. The top three GO biological processes were cytoplasmic translation, protein localization to the plasma membrane, and regulation of neurotransmitter levels, as shown in Figure 1D. The summary of enrichment analyses in TRRUST showed that the differentially expressed proteins were regulated by STAT3, REST, TFAP2A, SNAI1, etc., as depicted in Figure 1E. Somatotroph adenoma and lactotroph both belong to the PIT-1 lineages of PAs. In this study, we analyzed the difference between somatotroph and macro-lactotroph adenomas, as shown in Figure S1. The volcano map shows the top 20 DEGs in Figure S1A. Using the Metascape database, DEG pathway enrichment shows the top 10 GO biological processes in Figure S1B and the top 10 KEEG pathways in Figure S1C,D.
EIF2 is a key component of the ternary complex whose role is to deliver initiator tRNA into the ribosome. Mass-spectrometric data showed increased subunit α/β and reduction in γ subunit, mainly subunit β. Immunohistochemistry exhibited the differential significance and the priority of EIF2β in somatotroph adenomas compared with gonadotroph and corticotroph adenomas, as depicted in Figure 2. Somatotroph adenoma and lactotroph both belong to the PIT-1 lineages of PAs. In this study, we analyzed the difference between somatotroph and macro-lactotroph adenomas, as shown in Figure S1. The volcano map shows the top 20 DEGs in Figure S1A. Using the Metascape database, DEG pathway enrichment shows the top 10 GO biological processes in Figure S1B and the top 10 KEEG pathways in Figure S1C,D.
EIF2 is a key component of the ternary complex whose role is to deliver initiator tRNA into the ribosome. Mass-spectrometric data showed increased subunit α/β and reduction in γ subunit, mainly subunit β. Immunohistochemistry exhibited the differential significance and the priority of EIF2β in somatotroph adenomas compared with gonadotroph and corticotroph adenomas, as depicted in Figure 2.

Differentially Expressed Gene-Enrichment Analysis between EIF2β Groups
According to the H-score of EIF2β (H-score: 114), patients with somatotroph adenomas were divided into a high-EIF2β group (n = 10) and a low-EIF2β group (n = 18); six samples were excluded because of staining loss. Based on different levels of EIF2β, there were 244 differentially expressed genes (I fold changeI ≥ 1, p < 0.05). The top three pathways were cytokine signaling in the immune system, hemostasis, and extracellular matrix organization ( Figure 3A). The top three GO biological processes were complement activation, classical pathway complement activation, and humoral immune response mediated by circulating immunoglobulins ( Figure 3B). Correlation analysis showed the following Spearman correlation coefficients: EIF2β and HRI (r = 0.611, p < 0.001), PKR (r = 0.765, p < 0.001), PERK (r = 0.813, p < 0.001), GCN2 (r = 0.728, p < 0.001), and STAT3 (r = 0.732, p < 0.001) ( Figure 3C). Furthermore, we analyzed the clinical relevance of EIF2β in 28 patients with somatotroph adenoma ( Table 2). Patients with a high EIF2β had a greater invasive potential (p = 0.038). There was a tendency toward a more sparsely granulated adenoma phenotype in 28 somatotroph adenoma patients with a high EIF2β than in patients with a low EIF2β.  Furthermore, we analyzed the clinical relevance of EIF2β in 28 patients with somatotroph adenoma ( Table 2). Patients with a high EIF2β had a greater invasive potential (p = 0.038). There was a tendency toward a more sparsely granulated adenoma phenotype in 28 somatotroph adenoma patients with a high EIF2β than in patients with a low EIF2β.

Immunologic Signatures in Patients with Somatotroph Adenomas
First, we described the immune cell types and percentages with a heatmap for patients with somatotroph adenomas ( Figure 4A). The associations between the infiltrate levels of immune cell signatures and levels of EIF2β are illustrated in Figure 4B; there was a statistically significant difference in lymphocytes, macrophages, and dendritic cells between patients with a high EIF2β and patients with a low EIF2β (p < 0.05) ( Figure 4C). Moreover, according to the unpaired t test, there were significant differences in the immune score (p = 0.020) and ESTIMATE score (p = 0.024) between patients with a high EIF2β and patients with a low EIF2β ( Figure 4D). tients with somatotroph adenomas ( Figure 4A). The associations between the infiltrate levels of immune cell signatures and levels of EIF2β are illustrated in Figure 4B; there was a statistically significant difference in lymphocytes, macrophages, and dendritic cells between patients with a high EIF2β and patients with a low EIF2β (p < 0.05) ( Figure 4C). Moreover, according to the unpaired t test, there were significant differences in the immune score (p = 0.020) and ESTIMATE score (p = 0.024) between patients with a high EIF2β and patients with a low EIF2β ( Figure 4D).

Enumeration and Spatial Distribution of Immune Cells in Pituitary Adenomas Determined by Immunohistochemistry Staining
Innate immune cells and adaptive immune cells are very important for host defense, and the balance of immune responses is critical in controlling cancer progression. An immunohistochemistry experiment revealed that the mean number of immune cells was 7.41 ± 3.87% (range 0-32.9%) in somatotroph adenomas, 5.35 ± 3.65% (range 0-27.1%) in gonadotroph adenomas, and 4.97 ± 3.27% (0-42.9%) in corticotroph adenomas ( Figure 5).
correlated. (D): Immune scores of 28 patients with somatotroph adenomas according to ESTIMATE.

Enumeration and Spatial Distribution of Immune Cells in Pituitary Adenomas Determined by Immunohistochemistry Staining
Innate immune cells and adaptive immune cells are very important for host defense, and the balance of immune responses is critical in controlling cancer progression. An immunohistochemistry experiment revealed that the mean number of immune cells was 7.41 ± 3.87% (range 0-32.9%) in somatotroph adenomas, 5.35 ± 3.65% (range 0-27.1%) in gonadotroph adenomas, and 4.97 ± 3.27% (0-42.9%) in corticotroph adenomas ( Figure 5).

Effects of RNAi-EIF2β on Cell Proliferation and Growth Hormone Release by GH3 Cells
To test the role of EIF2β in somatotroph adenomas, we constructed the RNA interference fragments of EIF2β and packaged the viral vectors. The interfering efficiency of shRNA1-4 were 0.97 ± 0.15, 0.57 ± 0.02, 0.47 ± 0.03, and 0.91 ± 0.19-fold, respectively, of control GH3 cells after 72 h transient transfection ( Figure 6A) (p < 0.05). Western blot experiments also showed the same tendency ( Figure 6B). The sh2 and sh3 fragments were used in further experiments. The cell viabilities in RNAi groups were 78.4 ± 4.7% and 84.5 ± 5.1% of that in the control group after 24 h infection, 67.3 ± 3.9% and 71.5 ± 4.1% after 48 h infection, and 60.4 ± 3.8% and 65.2 ± 3.7% after 72 h infection, respectively ( Figure 6C, p < 0.05). ELISA showed that GH levels in the cell culture were 5.7 ± 1.1 ng/mL and 6.1 ± 0.9 ng/mL in the RNAi-EIF2β groups compared with 11.7 ± 0.7 ng/mL in the control GH3 group (p < 0.05) after 72 h treatment ( Figure 6D). IGF-1 levels in the cell culture were 7.7 ± 1.2 ng/mL and 9.2 ± 1.1 ng/mL in the RNAi-EIF2β groups compared with 16.3 ± 0.9 ng/mL in the control group (p < 0.05) after 72 h treatment ( Figure 6E). Transwell experiments showed that the trans-membrane-positive cells in the RNAi-sh2 group (226 ± 27/high field) and RNAi-sh3 group (254 ± 32/high field) were reduced compared with the control group (403 ± 67/high field) ( Figure 6F). We also found a decline in STAT3 and STAT5b in the sh2 and sh3 groups ( Figure 6G). group (403 ± 67/high field) ( Figure 6F). We also found a decline in STAT3 and STAT5b in the sh2 and sh3 groups ( Figure 6G).

Discussion
Somatotroph adenomas usually induce a decrease in body fat, osteoporosis, and increase water retention. Recent studies suggest that abnormal fat metabolism promotes

Discussion
Somatotroph adenomas usually induce a decrease in body fat, osteoporosis, and increase water retention. Recent studies suggest that abnormal fat metabolism promotes protein anabolism and relieves amino acid oxidation. Protein synthesis dysregulation is a burgeoning field in cancer biology. The ISR is an essential stress-support pathway, increasingly recognized as a determinant of tumorigenesis by fine-tuning protein-synthesis rates. Translation initiation lies downstream to deregulated signal pathways in cancer and is regulated by activated oncogenes or mutated tumor suppressors. Therefore, controlling protein synthesis has significant potential for developing innovative therapy in somatotroph adenomas.
In this study, we found significant differences in the eIF2 pathway in somatotroph adenomas compared to normal pituitary and other subtypes of pituitary adenomas through multi-omics analysis. Furthermore, we evaluated the effects of the eIF2 pathway on cell proliferation, invasion, and growth hormone release through ex vivo experiments.
Our understanding of the central mechanisms that govern the ISR has advanced vastly. The ISR's central regulatory hub lies in the EIF2-EIF2B complex, which controls the formation of the eEIF2/GTP/methionyl-initiator tRNA TC, a prerequisite for initiating new protein synthesis [7]. EIF2 and EIF2B are complex proteins of three (α-γ) and five (α-ε) nonidentical subunits, respectively [20]. Through integrated multi-omics analysis, we found an abnormally increased profile of ribosomal proteins in somatotroph adenomas compared with normal pituitary glands or other subtypes of pituitary adenomas. In the diversity of differentially expressed proteins, the top three pathways were: SRP-dependent cotranslational protein targeting the membrane, peptide chain elongation, and eukaryotic translation elongation by KEGG pathway enrichment, as shown in Figure 1C. Moreover, we found an imbalance in the EIF2 complex in somatotroph adenomas, i.e., increased α/β subunit and reduction in the γ subunit of eIF2.
Oncogenic roles of eIF2α have been extensively studied in several human cancers [21]; however, few studies demonstrate the roles of eIF2β in carcinogenesis. An integrated genomic search of novel therapeutic targets for breast cancer indicated that eIF2β is highly amplified in luminal breast tumors, which are an aggressive subtype of breast tumors [22]. In in vitro experiments, the knockdown of eIF2β manipulations induced G1 cell cycle arrest in both NCI-H358 and NCI-H460 cells [23]. Deleting the polylysine stretches in the amino-terminal region of eIF2β decreased protein synthesis, inhibited cell proliferation and viability, increased cell death, and induced G2 cell cycle arrest in Hek293TetR cells [24]. Interestingly, we noticed that the levels of EIF2β in somatotroph adenomas were related to the disease phenotype. Patients with a high EIF2β had a higher invasive potential and greater chances of sparsely granulated tumors than patients with low immune-cell infiltration with low EIF2β. The phosphorylation of eIF2α is carried out by a family of four kinases, PERK, PKR, GCN2, and HRI, which primarily respond to a distinct type of stress or stresses [25]. In this study, we observed a high correlation between EIF2α/β and PKR and PERK expression, suggesting that simultaneous upregulation of these molecules increases translation initiation, contributing to enhanced synthesis and secretion of GH/IFF-1 by somatotroph tumor cells. We found a decrease in cell viability and the level of GH and IGF-1 in the cell culture supernatant of GH3 cells in both sh2-EIF2β and sh3-EIF2β groups.
Enrichment analysis in TRRUST showed that the differentially expressed proteins were regulated by STAT3, REST, TFAP2A, SNAI1, etc. It was reported that STAT3 specifically binds the GH promoter and induces transcription in the GH3 cell line [26]. Additionally, C188-9, an inhibitor of STAT3, could downregulate the levels of eIF2 and STAT3 in GH3 cell lines. We confirmed a positive correlation between EIF2β and STAT3 based on transcriptome data and immunohistochemistry staining. In vitro experiments verified that RNAi-EIF2β relieved cell proliferation, invasion, and secretion of GH/IGF-1. Western blot experiments also verified the relationship between EIF2β and STAT3. We hypothesized that EIF2β was mainly involved in tumor-related pathways in somatotroph adenomas, including the cell adhesion molecule pathway and the JAK-STAT signaling pathway. We speculated that eIF2β might inhibit T cell-mediated immunity by promoting the Treg response. In addition, the expression of eIF2β was positively correlated with STAT3 and STAT5b. Related studies have shown that some signaling molecules are involved in the M2 polarization of macrophages, such as PI3K/AKT-ERK signaling, STAT3, HIF1α, and STAT6. These results suggested that eIF2β might regulate tumor macrophage infiltration, affecting the tumor microenvironment.
The tumor immune microenvironment is increasingly being recognized as a key contributor to tumorigenesis and tumor progression and recurrence. Cancer cells, stromal cells, and immune cells, along with their extracellular factors, have profound effects on either promoting or repressing anti-cancer immunity. These effects were mostly detected by immunohistochemistry; immune cell infiltration varied greatly among samples, which indicated that pituitary-hormone-secreting adenomas have substantial heterogeneity. For example, it has been reported that CD68+ macrophage infiltration was higher than that in corticotroph adenomas and densely granulated somatotroph adenomas [27]. In this study, we found a greater probability of immune cell infiltration in corticotroph adenomas and somatotroph adenomas compared with gonadotroph adenomas. Since corticotroph adenomas and somatotroph adenomas have greater invasive potential, we speculated that immune cell infiltration was related to invasive behavior. However, we found a negative relationship between the levels of EIF2β and immune score in 28 patients with somatotroph adenoma based on the "ESTIMATE" algorithm, which was further validated by immunohistochemistry data. The enriched GO terms by DEG analysis for biological processes were complement activation, classical pathway complement activation, and humoral immune response mediated by circulating immunoglobulins, whereas KEGG pathway terms enriched by DEG analysis were cytokine signaling in the immune system, hemostasis, and extracellular matrix organization; these may all have roles in the modulation of inflammatory infiltration in tumor tissue. The specific mechanism by which inflammatory infiltration leads to tumor invasion requires further investigation; this may also be relevant for further research into and development of targeted drugs for somatotroph adenomas.
However, our study has some limitations. First, the small sample size may have contributed to a bias error; the relationship between eIF2β and sparse granularity needs further confirmation in large samples. Moreover, the transcriptome data analyzed were generated through bulk RNA sequencing; bulk sequencing is likely to cause loss of some information because of the heterogeneous nature of tumor tissues, and this may be explored in more detail with single-cell RNA-seq techniques.
In conclusion, in the present study, we uncovered the unique characteristics of the eIF2 complex and abnormal integrated stress response through omics data in somatotroph adenoma patients. Furthermore, we characterized the landscape of the somatotroph adenoma microenvironment using different methodologies, including immunohistochemistry and transcriptomics. We speculated that eIF2β promoted the synthesis and secretion of GH/IGF-1 in somatotroph tumor cells and inhibited immune cell infiltration. GH3 cells were purchased from ATCC and cultured in F-12K medium (ATCC, Manassas, VA, USA) with 2.5% fetal bovine serum and 10% horse serum in a humidified incubator at 37 • C and 5% CO 2 .

Protein Preparation and Proteomics Analysis
Protein preparation and proteomics analysis were completed in accordance with the methods used in our previous study [28]. A mass of 100 µg of each pooled sample was then denatured, reduced, and alkylated as described in the iTRAQ protocol (Thermo Fisher, Waltham, MA, USA) and digested overnight with 0.1 µg/µL trypsin solution at 37 • C. Peptides were analyzed by nanoLC-MS/MS on a Q Exactive mass spectrometer (Thermo Fisher, USA). Chromatography was performed with solvent A (Milli-Q (Millipore, Billerica, MA, USA) water with 2% acetonitrile and 0.1% formic acid) and solvent B (90% acetonitrile with 0.1% formic acid). The Q Exactive instrument was operated in information-dependent data acquisition mode to switch automatically between MS and MS/MS acquisition. MS spectra were acquired across the mass range of 350-2000 m/z. All MS/MS data were analyzed using Mascot (Matrix Science, London, UK; version 2.3.0). Carbamidomethylation of cysteine residues was specified as a fixed modification. Oxidation of methionine residues was specified in Mascot as a variable modification. Only proteins with a p < 0.05 were accepted.

RNA Microarray and PCR Verification
Microarray hybridization and qRT-PCR were conducted in accordance with the methods used in our previous study [28]. Total RNA of 30 samples was extracted and purified using mirVana™ miRNA Isolation Kit (Ambion, Austin, TX, USA) following the manufacturer's instructions. Labeled cRNA was purified using an RNeasy mini kit (QIAGEN, Hilden, Germany). Each slide was hybridized with 1.65 µg Cy3-labeled cRNA using a Gene Expression Hybridization Kit (Agilent Technologies, Santa Clara, CA, USA) in a Hybridization Oven (Agilent Technologies), according to the manufacturer's instructions. After 17 h hybridization, slides were washed in staining dishes (Thermo Shandon, USA) with a Gene Expression Wash Buffer Kit (Agilent Technologies), following the manufacturer's instructions. Slides were scanned with an Agilent Microarray Scanner using default settings: Dye channel: green, scan resolution = 3 µm, PMT 100%, 20 bit. Data were extracted with Feature Extraction software 10.7 (Agilent Technologies). Raw data were normalized using the Quantile algorithm and limma package in R. The selected mRNAs were grouped into functional categories based on the Gene Ontology database (GO: http://www.geneontology.org/, accessed on 1 August 2022), and functional pathways (KEGG and BIOCARTA) were also analyzed using an online SAS analysis system.

Cell Proliferation and Migration Assays
GH3 cells were adjusted to a density of 1 × 10 5 cells/mL. An aliquot of 100 µL of cell suspension was plated into each well of 96-well plates and cultured overnight. At 24, 48, and 72 h post-transfection, 20 µL 3-(4,5-diethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-etrazolium inner salt (MTS) solution was added to each well, followed by further incubation for 4 h. Absorbance of each well at 490 nm was measured using an ELISA plate reader (Thermo, USA). Cell migration was measured using fibronectin and Matrigel-coated polycarbonate filters, respectively, and modified transwell chambers (Corning, MA, USA). GH3 cells (5 × 10 4 cells) were added into the upper chambers. After 24 h of incubation, migrating cells that adhered to the lower membrane were fixed in 4% paraformaldehyde and stained using hematoxylin (Zhongshan Company, Beijing, China). No invading cells were removed with cotton swabs. Images were photographed in three randomly selected fields of view at 200× magnification. Experiments were performed in triplicate.

ELISA
The levels of GH and IGF-1 were detected using an ELISA kit (APPLYGEN) according to the protocol. An aliquot of 10 µL of cell culture supernatant per well was used. The absorbance of each well was measured at 450 nm using an ELISA plate reader (Thermo Fisher). The amount of GH and IGF-1 was calculated from a standard curve prepared using the recombinant protein. Positive controls were supplied in the kit.

Statistical Analysis
Chi-square test and Fisher's exact test were used to determine the significance of categorical variables. One-way ANOVA was applied to the immunohistochemical results in patients or in vitro experiments. All p values are two-sided, and 0.05 was applied as the significance level.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All the data generated or analyzed in this study are included in this published article and its additional files.