PCSK1N as a Tumor Size Marker and an ER Stress Response Protein in Corticotroph Pituitary Adenomas

Abstract

Context

Silent corticotroph adenoma (SCA) exhibits more tumor aggressiveness features than functioning adenomas (FCAs).

Objective

We aimed to investigate proprotein convertase subtilisin/kexin type 1 inhibitor (PCSK1N) expression in CA and examine if endoplasmic reticulum (ER) stress-induced responses affect cell survival in a corticotroph tumor cell model.

Methods

Clinical and imaging characteristics were recorded in 33 patients with FCA (20 women, 11 macroadenomas) and 18 SCAs (8 women, all macroadenomas). Gene expression of pro-opiomelanocortin (POMC), T-box transcription factor 19(TBX19)/TPIT, proprotein convertase subtilisin/kexin type 1 (PCSK1)/PC1/3, and its inhibitor PCSK1N, was measured by reverse transcription–quantitative polymerase chain reaction in adenoma tissue. Mouse pituitary corticotroph tumor (AtT-20) cells were treated with tanespimycin (17-AAG), an HSP90 chaperone inhibitor, to induce ER stress, followed by gene and protein analyses.

Results

POMC, TPIT, and PCSK1 expression were higher, whereas PCSK1N was lower in FCA compared to SCA. PCSK1N correlated with POMC (r_s = −0.514; P < .001), TPIT (r_s = −0.386; P = .005), PCSK1 (r_s = −0.3691; P = .008), and tumor largest diameter (r_s = 0.645; P < .001), in all CA. Induction of ER stress by 17-AAG in AtT-20 cells led to a decrease of Pomc and an increase of Pcsk1n gene expression at 24 hours. Moreover, a downregulation of cell cycle, apoptosis, and senescence pathways, and alterations in cell adhesion and cytoskeleton, were observed at the protein level.

Conclusion

PCSK1N is higher in SCA compared with FCA, and associated with corticotroph cell markers and tumor size. PCSK1N is likely to be part of the adaptive response to ER stress, potentially conferring a survival advantage to the corticotroph tumor cell in conjunction with other proteins.

Corticotroph pituitary adenomas (CAs) develop from corticotroph cells and are classified by their immunohistochemistry (IHC) staining for adrenocorticotropic hormone (ACTH) and/or the corresponding transcription factor T-box transcription factor 19 (TBX19)/TPIT (1). In a clinical setting, CA present with different degrees of functionality, based on the level of ACTH secretion, ranging from silent CA (SCA) to whispering, and ultimately to functioning CA (FCA) (2, 3). Due to the clinical symptoms caused by ACTH-stimulated cortisol secretion leading to Cushing disease, FCAs are generally detected at an earlier stage as compared to SCAs (4). SCAs are usually detected by the symptoms caused by the tumor volume, as they are mostly macroadenomas with extrasellar expansion, whereas FCAs most often are intrasellar microadenomas at presentation (3, 5). SCAs represent up to 15% of all nonfunctioning pituitary adenomas (6), and about 3% of all pituitary neuroendocrine tumors, while FCAs comprise up to 2% to 6% (7).

Dysfunctional processing of pro-opiomelanocortin (POMC) into ACTH has been hypothesized to contribute to the silence of SCA. Indeed, a previous study showed that SCA secrete more inactive ACTH molecules (8). The main prohormone-converting enzyme in the anterior lobe of the pituitary gland is proprotein convertase subtilisin/kexin type 1 (PCSK1) or PC1/3, sequentially cleaving POMC to ACTH (9). Previous studies have shown that SCAs have lower expression of PCSK1 as compared to their functional counterpart, potentially causing the diminished ACTH production (9, 10). In addition, it has been shown that PCSK1 expression is increased in certain cases of SCA that later develop into FCA (3, 11). In the early 2000, studies of the AtT-20 mouse pituitary corticotroph tumor cell line revealed that overproduction of a small protein called proprotein convertase subtilisin/kexin type 1 inhibitor (PCSK1N) or proSAAS, encoded by the PCSK1N gene, markedly reduced the processing of POMC, and this protein was shown to specifically inhibit PCSK1 (12, 13). We have previously shown that SCAs express higher protein levels of PCSK1N (14).

SCAs are recognized to have an aggressive clinical behavior due to rapid and invasive growth, tumor heterogeneity, and higher recurrence rates (5, 15). In recent years, several differences in the molecular profiles between FCA and SCA have been described. For example, SCA exhibited several features of epithelial-mesenchymal transition (EMT), with increased expression of mesenchymal genes (16), dysregulation of several lipid metabolic pathways (17), a distinct expression of extracellular matrix genes, and a reduced endoplasmic reticulum (ER) protein-folding ability (14). Impaired protein processing in the ER potentially leads to ER stress and the activation of several compensatory pathways, among them, the unfolded protein response (UPR) signaling pathway, to reestablish protein homeostasis (18). Recently, PCSK1N has been described as a stress-responsive protein, with an antiaggregation function, and produced as part of a physiological defense mounted by cells to reduce cell damage and maintain cell viability (19).

Several molecules are known to induce ER stress in experimental settings, some with the potential to be further developed into anticancer drugs (20). One of these molecules is 17-N-allylamino-17-demethoxygeldanamycin (17-AAG) or tanespimycin, an inhibitor of the 90-kDa heat shock protein (Hsp90), which has been used in cancer treatment clinical trials (21). Hsp90 plays important roles in protein-folding processes in the ER, and 17-AAG has been known to activate one branch of UPR pathways called IRE1 in myeloma cancer cells (22). The inhibition of Hsp90 by 17-AAG also induces compensatory adaptation for cell survival (23). However, how the Hsp90 inhibitor interacts with PCSK1 and PCSK1N is not clearly defined.

We hypothesized that PCSK1N plays important roles involving cell growth and survival processes, facilitating CA growth. PCSK1N is known to function as a PCSK1 inhibitor (12), and is an ER stress-responsive protein (19). Based on our previous data showing that PCSK1N protein is higher expressed in SCA (14), we here aimed to assess the expression of PCSK1N, in addition to corticotroph cell markers (POMC, TPIT, and PCSK1), in a large cohort of FCA and SCA, and correlate its expression to clinical features such as tumor diameter. Furthermore, we investigated the protein repertoire initiated by ER stress response in AtT-20 cells to decipher the mechanisms behind the growth potential/cell survival in SCA.

Material and Methods

Study Design and Sample Selection

Patients operated on for pituitary adenomas at our single center were included in a retrospective study between 1998 and 2009 (n = 240) and a prospective study between 2014 and 2022 (n = 216). For the present study, we included 51 patients (29 from the retrospective and 22 from the prospective study) diagnosed and operated on for Cushing disease (n = 33) and clinically nonfunctioning pituitary adenomas showing positive staining for TPIT and/or ACTH (SCAs) (n = 18) with available tumor tissue (38 tumors were snap-frozen samples and 13 were optimum cutting temperature embedded). All SCAs were clinically classified as nonfunctioning at the time of surgery, whereas the patients with FCAs presented with signs and symptoms of hypercortisolism. Twenty-four FCA and 5 SCA patients were previously included in published cohorts (14, 24, 25). The IHC subclassification of SCA was performed in a research setting as described (26, 27). Tumor diameter as described in the preoperative pituitary magnetic resonance imaging investigation was recorded as the largest diameter. Tumors with a diameter of less than 10 mm were defined as microadenomas, and macroadenomas were 10 mm or larger. Two SCA tumors had a diameter greater than 40 mm (57 and 62 mm). Morning preoperative plasma ACTH (pmol/L) and cortisol (nmol/L) concentrations were measured in the routine clinical setting. ACTH was available in 28 FCAs and 7 SCAs, whereas cortisol was available in 29 FCAs and 10 SCAs.

Written informed consent was obtained from all patients. This study was approved by the regional ethics committee (REK No. 2014/1680, and REK No. 2020/22301) and the hospital authority.

Sample Preparation

Tumor tissue was homogenized in Trizol (Life Technologies [RRID:SCR_008817], catalog No. 15596018). RNA was extracted by QIAGEN miRNeasy Mini Kit (QIAGEN [RRID:SCR_008539], catalog No. 74106), including RNase-Free DNase (QIAGEN [RRID:SCR_008539], catalog No. 79256) treatment, according to the manufacturer's protocol. RNA integrity was determined using the Agilent 2100 Bioanalyzer (Agilent 2100 Bioanalyzer Instrument [RRID:SCR_018043]), and concentrations were measured by optical density (OD) readings on a Nanodrop ND-1000 Spectrophotometer (Nanodrop Technologies [RRID:SCR_016517]) as described previously (28). RNA integrity numbers were above 7 in 41 patients, and between 5.6 and 6.9 in 10 patients.

Reverse Transcription–Quantitative Polymerase Chain Reaction

Reverse transcription (RT) was performed using a high-capacity complementary DNA (cDNA) reverse transcription kit (Applied Biosystems [RRID:SCR_005039], catalog No. 4368814) by a Labnet MultiGene Gradient Thermal Cycler (Labnet MultiGene Gradient PCR Thermal Cycler [RRID:SCR_025065], Labnet International Inc) according to the manufacturer's protocol, using a total of 1 µg RNA. After the reaction, the cDNA was diluted to a ratio of 1:10. RT-qPCR was performed in an ABI 7900 (Applied Biosystems [RRID:SCR_005039]) using Power SYBR Green Master Mix (Applied Biosystems [RRID:SCR_005039], catalog No. 4367659). Samples were dispensed in the corresponding wells by an automated pipetting system (Eppendorf epMotion 5070 CB [RRID:SCR_025064]). Samples were run in duplicates with 10 μL per reaction, including 2 μL of the cDNA at 1:10 dilution. Primer concentration per reaction was 375 nM. Primer pairs were found either in the literature or in the primer bank (29). All primers were tested for specificity using BLAST analysis (NCBI BLAST [RRID:SCR_004870]). All the polymerase chain reaction (PCR) reactions had an efficiency between 90% and 100% (−3.6 ≥ slope ≥ −3.3). The amplification efficiency and correlation coefficients for these primers (Sigma-Aldrich [RRID:SCR_008988]) were obtained from the slope of the standard curves. To avoid false-positive amplification due to DNA contamination, the genomic DNA sequence (downloaded from Ensemble Data Base, Ensembl [RRID:SCR_002344]) was used to check exon-intron borders by matching the primers to their location. Target gene cycle threshold (Ct) value was adjusted by using the ΔCt method for the geometrical Ct mean of 2 reference genes (GAPDH and ALAS1), as previously validated for pituitary tumor samples (28), and to Actb for AtT-20 samples. Human primers are presented in Supplementary Table S1 (30). Gene expression data for ER protein-processing genes (CALR, GRP78, GRP94, and UGGT2) were previously presented (14). In our CA tumor cohort the gene expression Ct values range were as follows: POMC (12.5-27.8), TPIT (18.2-29.7), PCSK1 (17.5-26.8), PCSK1N (18.1-29.9), GAPDH (17.2-20.5), and ALAS1 (21.7-26.3) cycles.

Cell Culture and 90-kDa Heat Shock Protein Inhibition

AtT-20 mouse pituitary corticotroph tumor cells (ATCC catalog No. CRL-1795, RRID:CVCL_4109) were cultured in Dulbecco's Modified Eagle's medium-high glucose (Sigma-Aldrich [RRID:SCR_008988], catalog No. D6429) supplemented with 10% fetal bovine serum (FBS, South America, Premium, Biowest [RRID:SCR_025063], catalog No. S181B) and 1% penicillin/streptomycin (Gibco Life Technologies [RRID:SCR_008817], catalog No. 15140122) at 37 °C in a 5% CO₂ incubator.

The cells were seeded in triplicate in 12-well plates containing 1.5 × 10⁵ cells/well and cultured for 72 hours before stimulation for 6 and 24 hours by 17-AAG (50 μM, MedChemExpress [RRID:SCR_025062], catalog No. HY-10211). Dimethyl sulfoxide (DMSO) was used as control. Mouse primers are presented in Supplementary Table S2 (30). For proteomic analysis, the media was removed, and cells were washed with phosphate-buffered saline and lysed using radioimmunoprecipitation assay lysis buffer (Merck Millipore, Merck [RRID:SCR_001287], catalog No. 20-188) and halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific [RRID:SCR_008452], catalog No. 78441), according to the manufacturer's protocol.

Quantitative Label-Free Mass Spectrometry–Based Proteomic Analysis

The protein lysates from cells treated for 24 hours with 17-AAG (4 biological replicates) and controls (DMSO, 7 biological replicates) were subjected to quantitative label-free mass spectrometry (MS)–based proteomic analysis. Protein concentration was first estimated by bicinchoninic acid assay (Pierce BCA Protein Assay Kits, Thermo Fisher Scientific [RRID:SCR_008452], catalog No. 23225), and for each replicate an equal amount (10 μg) of protein was precipitated on amine beads (31). The precipitated proteins were dissolved in 50-mM ammonium bicarbonate, reduced, alkylated, and digested with trypsin (1:50 enzyme:protein ratio; Promega) at 37 °C overnight. Digested peptides were acidified, and the peptides were loaded into Evosep C18 tips.

Liquid chromatography (LC)-MS/MS analysis was carried out using an Evosep LC system (Evosep [RRID:SCR_024590]) coupled to a timsTOF fleX mass spectrometer (Bruker Daltonics [RRID:SCR_023608]), using a CaptiveSpray nanoelectrospray ion source (Bruker Daltonics [RRID:SCR_023608]). A total of 200 ng of peptide digest was loaded on a capillary C18 Evosep column (15 cm length, 150 μm inner diameter, 1.5 μm particle size, 120; Evosep [RRID:SCR_024590]). Peptides were separated at 50 °C using a 44-minute gradient. The timsTOF fleX was operated in PASEF mode. Mass spectra for MS and MS/MS scans were recorded between m/z 100 and 1700. Ion mobility resolution was set to 0.60 to 1.60 V s/cm over a ramp time of 100 ms. Data-dependent acquisition was performed using 10 PASEF MS/MS scans per cycle with a near 100% duty cycle. A polygon filter was applied in the m/z and ion mobility space to exclude low m/z, singly charged ions from PASEF precursor selection. An active exclusion time of 0.4 minutes was applied to precursors that reached 20 000 intensity units. Collisional energy was ramped stepwise as a function of ion mobility.

Raw files from LC-MS/MS analyses were submitted to MaxQuant 2.0.3.0 software for protein identification and label-free quantification (MaxQuant [RRID:SCR_014485]). Parameters were set as follows: Carbamidomethyl (C) was set as a fixed modification and protein N-acetylation and methionine oxidation as variable modifications. First search error window of 20 ppm and mains search error of 6 ppm. Trypsin without proline restriction enzyme option was used, with 2 allowed miscleavages. Minimal unique peptides were set to one, and the false discovery rate (FDR) allowed was 0.01 (1%) for peptide and protein identification. The Uniprot mouse database was used (UniProt [RRID:SCR_002380]). Generation of reversed sequences was selected to assign FDR rate. Known contaminants as provided by MaxQuant and identified in the samples were excluded from further analysis.

Statistics

Statistical analyses were performed using IBM SPSS version 27.0.0.0 (IBM SPSS Statistics [RRID:SCR_016479]). Data are presented as mean ± 1 SD or as median (minimum-maximum) or interquartile range depending on distribution for continuous measures. Correlation analyses were performed using 2-tailed Spearman correlation tests, and correlation coefficient (r_s) and statistical significance level (P) presented. The Mann-Whitney U test was used to test the differences between the groups. Perseus version 1.6.15.0 (Perseus [RRID:SCR_015753]) was used for further analysis of MaxQuant data: label-free quantitation intensity values were lg10 transformed, a minimum of 50% valid values in at least one group was required, and missing values were imputed from the low end of normal distribution. The t test between the groups (permutation-based FDR < 0.05) was performed to find the significant differentially expressed proteins (DEPs), and q (P value adjusted) less than .05 was considered statistically significant. The heat map was generated in R version 4.2.2 using the gplots2::heatmap.2 function (parameter: scale = “row”).

Results

Patient Clinical Characteristics

Clinical characteristics of the patients are presented in Table 1. The groups were similar regarding sex, whereas patients with FCAs were significantly younger and had smaller tumors. All SCAs and half of the FCAs were macroadenomas. Morning plasma ACTH and cortisol levels, measured preoperatively, were significantly higher in FCA compared with SCA patients.

Gene Expression of Corticotroph Cell Markers

Fig. 1 shows that gene expression of POMC (Fig. 1A), TPIT (Fig. 1B), and PCSK1 (Fig. 1C) were significantly higher, whereas PCSK1N (Fig. 1D) was lower, in FCAs as compared to SCAs. As a median, the POMC, TPIT, and PCSK1 gene expression fold change between FCA and SCA was 16.6, 2.4, and 3.2 times higher. Conversely, PCSK1N was a median of 9.9 times higher in SCA as compared to FCA (see Table 1). All the corticotroph cell markers and the housekeeping genes were highly expressed in our CA tumor cohort (Ct values < 30 cycles).

Correlation Analyses in the Entire Cohort of Corticotroph Tumors

POMC gene expression correlated positively with TPIT (r_s = 0.720; P < .001; n = 51) and PCSK1 (r_s = 0.569; P < .001; n = 51), and negatively with tumor size (r_s = −0.491; P < .001; n = 46).

PCSK1N correlated negatively with POMC (r_s = −0.514; P < .001; n = 51) (Fig. 2A), TPIT (r_s = −0.386; P = .005; n = 51), and cortisol (r_s = −0.356; P = .026; n = 39). However, a positive correlation was shown with tumor size (r_s = 0.645; P < .001; n = 46) (Fig. 2B). The correlation between PCSK1N and tumor size was also statistically significant in the FCA group alone (r_s = 0.655; P < .001) but not in the SCA group alone (r_s = 0.264; P > .05). PCSK1N correlated negatively with genes involved in ER processing (CALR, GRP78, GRP94, and UGGT2) (−0.57 < r_s < −0.35; P < .05; n = 44). No correlation of PCSK1N with plasma ACTH levels was found (r_s = −0.208; P = not significant; n = 35).

Induction of Endoplasmic Reticulum Stress Response in AtT-20 Cells

Treatment of the AtT-20 mouse corticotroph tumor cell line with 17-AAG (50 µM), an HSP90 chaperone inhibitor, to induce ER stress, led to a decrease of Pomc gene expression at 24 hours (Fig. 3A) and an increase of Pcsk1n expression at 6 and 24 hours (Fig. 3B). Gene expression of several ER chaperones, including Grp94, Calr, and Grp78, increased at 24 hours, as expected, as a consequence of induced ER stress (Fig. 3C-3E). Furthermore, Uggt1, a molecule that provides quality control by selectively reglycosylating unfolded glycoproteins, showed increased expression at both time points (Fig. 3F).

As the induction of ER stress response leads to the activation of the UPR pathway, genes involved in UPR pathway were also measured. Ire1, showed a trend toward increased expression at 6 hours, becoming significant just at 24 hours, after 17-AAG treatment (Fig. 3G). Furthermore, a significant decrease of the s/uXbp1 ratio, one of the main readouts of the IRE1 pathway, was observed (Fig. 3L), suggesting a compensatory response of the cells already at 6 hours, becoming even more prominent at 24 hours. Atf4 and Atf6 showed similar patterns of regulation as for Ire1, but reached statistical significance only for Atf4 at 6 hours (Fig. 3H and 3I).

The ratio of Bax/Bcl-XL, molecules involving cell death and survival as an end readout of the entire UPR response pathway, showed increased expression both at 6 and 24 hours (Fig. 3J). In contrast, Bax/Bcl-2 ratio was not significantly regulated (Fig. 3K).

Proteomic Analyses of AtT-20 Cell Lysates

To assess the ER stress response at the protein level, we performed proteomic analysis and identified 1118 DEPs regulated by 17-AAG treatment at 24 hours (Supplementary Excel file) (30). Six Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathways were selected based on the expected response to ER protein-folding inhibition and their relevance to the study. All the DEPs within these pathways are presented as a heat map (Fig. 4). Furthermore, based on literature review and each protein's fold change, we present the most relevant proteins within these 6 pathways in Table 2.

As expected due to the compensatory mechanisms of the ER inhibition, there were many upregulated proteins involved in protein processing in ER (see Fig. 4). Among these, several ER chaperones (GRP78, GRP94, PDIA6, CANX) and ER-associated degradation (ERAD) proteins (HRD1/SYVN1, OS9, ERLEC1, SEL1L, VCP, EDEM3, and DNAJC10) were observed (see Table 2). Furthermore, several proteins that are involved in the IRE1 pathway were regulated, such as increased PDIA6 and PRKCSH and decreased PTPN1 (see Table 2). No clear evidence for the activation at the protein level of the UPR-activated signaling pathways ATF4 and ATF6 was found.

Overall, the 17-AAG treatment resulted in a downregulation of apoptosis, cell cycle, and senescence pathways in addition to changes in cell adhesion and cytoskeleton protein composition indicating less proliferative and less-cohesive cells (see Fig. 4 and Table 2).

Discussion

CAs represent a continuum of tumors from SCA, to whispering and FCA, having in common the expression of ACTH and/or TPIT. Here, we have shown that PCSK1N expression is closely associated with tumor size in both subtypes. Furthermore, it may be a part of the ER stress-activated response facilitating corticotroph tumor cell survival, and potentially contributing to tumor growth.

PCSK1N has been described as an endogenous inhibitor of PCSK1 in AtT-20 cells. Higher expression in SCA might reduce proteolytic cleavage of POMC, and consecutively lower ACTH production and secretion. Interestingly, our data showed a negative association between PCSK1N gene expression and plasma cortisol, but not ACTH levels, suggesting an effect on POMC cleavage and functional ACTH formation in human CAs. Furthermore, in concordance with our data, PCSK1 has been shown to be higher expressed in FCA than in SCA in several studies (10, 11, 32), suggesting an impairment of the POMC processing ability in SCA. In addition, the present results show an overlap in the expression of corticotroph cell markers between FCA and SCA, supporting that they are not entirely separate entities, but represent a continuum of the same type of tumors.

Previous studies have indicated that PCSK1N might have other functions besides inhibiting PCSK1, as it has been found to be increased during disease progression in neurodegenerative diseases (33). Moreover, expression of PCSK1N and PCSK1 were not coregulated in the adult brain of PCSK1N-knockout mice and in rat islet cells (34). Recently, PCSK1N has been described as a potent antiaggregation chaperone, preventing β-amyloid aggregation into fibrils in a dose-dependent manner (35, 36). Following ER stress, Pcsk1n levels increase, rescuing the cells from toxicity and providing a survival advantage (19). Since PCSK1N has an antiaggregation function, the increase of PCSK1N cellular levels may represent an adaptation to prevent more protein aggregation that will increase the severity of ER stress. In addition to the increase of Pcsk1n, our data show that ER stress activation leads to a downregulation of apoptosis, cell cycle, and senescence signaling pathways, which could confer a survival advantage to the corticotroph tumor cell, thereby contributing to tumor growth in the longer perspective. As such, this mechanism could explain the strong association between PCSK1N expression and tumor size.

The UPR pathway is activated by 17-AAG treatment, as shown by the upregulation of the ER stress-response molecules Grp78, Grp94, and Uggt1 in our data, to maintain homeostasis. GRP78 has been found to be overexpressed in several tumor types, including lung cancer (37) and showed to be induced by oxidative stress to protect the cell from damage and death (38). Therefore, GRP78 overexpression is an important factor for tumor resistance to various stressors and tumor development. One of the mechanisms by which GRP78 inhibits apoptosis is stabilizing mitochondrial permeability and reducing cytochrome C release through the Raf-1 proto-oncogene, serine/threonine kinase (Raf-1) expression (37). Raf-1 protein expression was shown to be decreased by 17-AAG treatment in our data set. However, Raf-1 itself has been known to involve different pathways including cell cycle and nuclear factor κB activation, so the downregulation in our data might reflect other biological roles (39). Besides antiapoptotic activity, GRP78 has also been shown to be involved in tumor invasion. Upregulation of GRP78 can induce EMT and contribute to metastasis, whereas downregulation strongly inhibits in vitro tumor cell invasion and metastasis (40, 41). GRP94 is another protein involved in tumor proliferation, invasion, and EMT (42, 43). The knockdown of GRP94 in a lung cancer cell line was shown to inhibit tumor cell proliferation and increase apoptosis through caspase-7 (44). Based on our data, we suggest that the increase of both GRP78 and GRP94 by ER stress promotes cell survival by counteracting the cell death signal initiated by the proapoptotic BAX.

XBP1, as a member of the IRE1 pathway, is essential for cell survival during tumor growth and hypoxic conditions (45). During the ER stress response, the targets of the spliced form (sXBP1) include ER chaperones (GRP78, GRP94), autophagic response, and ERAD components. The unspliced form (uXBP1) is a negative feedback regulator of sXBP1, and the overexpression of uXBP1 is involved in sXBP1 degradation during the recovery phase of ER stress (46). Thus, the decreased s/uXbp ratio in our data suggests that the cells were in the recovery phase after ER stress induction, and this seems to occur very early within the first 6 hours after 17-AAG treatment. However, despite the early decrease in sXbp1 expression, the expression of Grp78 and Grp94 still continued to increase at 24 hours, potentially suggesting a compensatory mechanism.

Among the proteins involved in ERAD, SYVN1 plays a role to antagonize ER stress-induced apoptotic cell death by promoting IRE1 ubiquitination and degradation (47). Besides involving the fate of cell death or survival, IRE1 has also been known to potentiate tumor growth. Several proteins involved in IRE1 pathways were regulated in our study. Among these, PDIA6 and PRKCSH enhance IRE1 activation, so the upregulation, shown in our proteomics data, contributes most probably to the resistance to ER stress-induced cell death (48, 49). On the contrary, the absence of PTPN1 or PTP-1B can impair ER stress-induced apoptosis (50). Being downregulated by 17-AAG in our proteomics results, this may antagonize apoptosis induced by ER stress, again contributing to cell survival.

The observed decrease in NCAM1, a neuron-neuron adhesion and neural development protein, might lead to decreased cell adhesion, and influence cell attachment and migration properties. Downregulation of NCAM1 has been shown to be correlated with poor prognosis in colorectal cancer, increased migration of pancreatic cancer cells, and tumor metastasis in thyroid tumor (51, 52).

Lastly, we found several molecules involved in autophagy, among them CALR, GRP78, and SQSTM1/p62, to be upregulated as a response to 17-AAG–induced ER stress. During ER stress that is not severe enough to cause cell death, autophagy is always activated (53) and contributes to the protection against apoptotic cell death. Autophagy is a conserved process, which involves recycling of molecules within the lysosomes (54). This process is important for cell survival during stress conditions since UPR needs 2 degradation systems, ERAD and selective autophagy, to reduce the load of unfolded proteins (55).

To summarize, ER stress induction by 17-AAG treatment in AtT-20 cells resulted in the increase of genes and proteins involved in ER stress response, including ER chaperones, activation of UPR response (mainly IRE1/XBP1 pathway), and induction of ERAD. The final outcome is cell protection against apoptotic cell death. Decreasing the apoptosis may in the long term result in increased growth potential that contributes to the increase of tumor size observed in SCA.

Study Limitations

Assessment of ACTH or cortisol overproduction was not performed systematically in the SCA, as they were diagnosed by tumor IHC analysis after surgery. Second, the lack of standardized magnetic resonance imaging–based measurement of tumor volume may weaken the correlations using size as an aggressiveness marker. However, almost all the SCAs were diagnosed at a later stage of tumor development and most of them showed extrasellar growth, making tumor volume of little use as a predictor for aggressiveness. Lastly, ER stress induction in vitro shows data limited to 24 hours’ exposure, which serves only as a model for the mechanisms that may occur in vivo in tumors with longstanding ER stress. The prolonged ER stress occurring in the corticotroph tumor probably involves several other mechanisms that play a role in increased cell proliferation and escaping cell death.

Conclusion

PCSK1N is negatively associated with corticotroph pituitary cell markers, and positively with tumor size, being likely a part of the adaptive response initiated by ER stress, and together with other proteins, may contribute to a survival advantage of the corticotroph tumor cell.

Funding

M.A. received scholarships provided by the South-Eastern Norway Regional Health Authority (Open Project No. 2020081/2020). The study did not receive any external funding. Mass spectrometry–based proteomic analyses were performed by the Proteomics Core Facility, Department of Immunology, University of Oslo and Oslo University Hospital, which is supported by the Core Facilities program of the South-Eastern Norway Regional Health Authority. This core facility is also a member of the National Network of Advanced Proteomics Infrastructure (NAPI), funded by the Research Council of Norway INFRASTRUKTUR-program (project No. 295910).

Disclosures

There is no conflict of interest that could be perceived as prejudicing the impartiality of this study.

Data Availability

Some or all data sets generated during and/or analyzed during the present study are not publicly available but are available from the corresponding author on reasonable request.

References

Abbreviations

 
• 17-AAG

  17-N-allylamino-17-demethoxygeldanamycin

 
• ACTH

  adrenocorticotropic hormone

 
• CAs

  corticotroph pituitary adenomas

 
• cDNA

  complementary DNA

 
• Ct

  cycle threshold

 
• DEPs

  differentially expressed proteins

 
• DMSO

  dimethyl sulfoxide

 
• EMT

  epithelial-mesenchymal transition

 
• ER

  endoplasmic reticulum

 
• FCA

  functioning corticotroph adenoma

 
• FDR

  false discovery rate

 
• Hsp90

  90-kDa heat shock protein

 
• IHC

  immunohistochemistry

 
• MS

  mass spectrometry

 
• PCSK1

  proprotein convertase subtilisin/kexin type 1

 
• PCSK1N

  proprotein convertase subtilisin/kexin type 1 inhibitor

 
• POMC

  pro-opiomelanocortin

 
• RT-qPCR

  reverse transcription–quantitative polymerase chain reaction

 
• SCA

  silent corticotroph adenoma

 
• TBX19

  T-box transcription factor 19

 
• UPR

  unfolded protein response