Differential Expression of MicroRNAs in Silent and Functioning Corticotroph Tumors

The potential role of miRNAs in the silencing mechanisms of pituitary neuroendocrine tumors (PitNETs) has not been addressed. The aim of the present study was to evaluate the expression levels and the potential associated role of some miRNAs, pathways, and transcription factors in the silencing mechanisms of corticotroph tumors (CTs). Accordingly, the expression of miR-375, miR-383, miR-488, miR-200a and miR-103; of PKA, MAP3K8, MEK, MAPK3, NGFIB, NURR1, PITX1, and STAT3 were analyzed via qRT-PCR in 23 silent and 24 functioning CTs. miR-200a and miR-103 showed significantly higher expression in silent than in functioning CTs, even after eliminating the bias of tumor size, therefore enabling the differentiation between the two variants. Additionally, miR-383 correlated negatively with TBX19 in silent CTs, a transcription factor related with the processing of POMC that can participate in the silencing mechanisms of CTs. Finally, the gene expression levels of miR-488, miR-200a, and miR-103 were significantly higher in macroadenomas (functioning and silent) than in microadenomas. The evidence from this study indicates that miRNAs could be involved in the pathophysiology of CTs. The translational implications of these findings suggest that pharmacological treatments specifically targeting these miRNAs could become a promising therapeutic option for these patients.


Introduction
Corticotroph tumors (CTs) are a pituitary neuroendocrine tumor (PitNET) subtype characterized by the autonomous hypersecretion of adrenocorticotropic hormone (ACTH), which gives rise to an associated endocrine clinical syndrome called Cushing disease [1]. CTs associated with this disease are called functioning CTs. On the other hand, there are other ACTH-secreting pituitary tumors named silent CTs, which are not associated with any of the most important clinical and biochemical features of Cushing disease (i.e., hypercortisolemia) [2]. Together, CTs represent up to 17% of all PitNETs [3,4] of which approximately 45% are silent CTs [5,6]. However, the silencing mechanisms and pathogenesis of CTs remain unclear. The most feasible hypothesis lies in post-transcriptional dysregulation of the POMC gene, a precursor of ACTH [7][8][9][10][11]. However, other genetic or epigenetic alterations such as DNA methylation or microRNA (miRNA) regulation may influence the clinical behavior of this PitNET subtype. Indeed, we have recently published the effect of the DNA methylation of some tumor suppressor genes in the functionality of the CTs [12], but to the best of our knowledge, no studies have been conducted to date in order to determine the presence and potential role of miRNAs in CTs.
As small non-coding RNA molecules, miRNAs [13] are involved in the post-transcriptional regulation of gene expression through their inhibitory action on messenger RNA. In the mouse corticotroph AtT-20 cell line, Zhang et al. [14] demonstrated that some miRNAs participate in the regulation of POMC expression, which is activated by the hypothalamic release of the corticotropin-releasing hormone (CRH) in response to environmental stimuli. CRH activates the PKA-MAP3K8-MEK-ERK1/2-NGFIB pathway together with POMC transcription, while miR-375 negatively regulates this pathway inhibiting the MAP3K8 expression [14]. In addition, other miRNAs such as miR-488, miR-383, and miR-200a have been reported as capable of inhibiting hypothalamic POMC expression [15]. Finally, it has been described in mice a transcriptional regulation of the POMC expression through the activation of the POMC promoter by transcription factors such as the bicoid-type homeodomain transcription factor pituitary homeobox 1 (PITX1) and the signal transducer and activator of transcription 3 (STAT3), which is activated by cytokines and presents two binding sites on the POMC promoter [16].
Although some authors have also been reporting changes (up-and down-regulation) in the expression of certain miRNAs in human pituitary tumors [17][18][19][20][21][22], few have addressed the presence and potential association of miRNAs in the silencing mechanisms of silent PitNETs, specifically the CT [23]. Therefore, the aim of the present study was to evaluate, for the first time, the expression levels and potential association of miR-375, miR-383, miR-488, miR-200a, and miR-103 and the PKA-MAP3K8-MEK-ERK1/2-NGFIB pathway plus the transcription factors PITX1 and STAT3 in the silencing mechanisms of CTs.

Patients and Samples
We selected 47 tumors of the corticotroph line, of which 23 were silent CTs, from our collection of PitNETs according to the quantity and quality of the genetic material. We collected anonymized clinical, radiological, immunohistochemical, and molecular data from the Spanish Molecular Registry of PitNETs (REMAH) database, a multicenter clinical-basic collaborative project of the Spanish Society of Endocrinology [24]. Identification of the PitNET subtypes was performed in light of the clinical and dominant relative gene expression of the pituitary-specific hormone genes and the specific transcription factor, Tpit, as previously described [5,6,25]. We defined micro-and macroadenomas according to maximum tumor diameter (MTD) and invasiveness according to Knosp classification [26]. Patients' demographic and clinical characteristics are shown in Table 1 and Supplementary Table S1.
The study was performed in compliance with the Declaration of Helsinki and other applicable legislation. The local ethics committee (CEIC General University Hospital of Alicante) approved the protocol. None of the patients were from a vulnerable population, and all donors or their relatives provided written informed consent.

DNA and RNA Extraction and cDNA Synthesis
All molecular studies took place in the laboratory of the research institute of the General University Hospital of Alicante (ISABIAL).
All samples were preserved immediately after surgery in RNAlater solution at 4 • C for 24 h and then stored at −20 • C. Biological samples were disintegrated in the TissueLyser (Qiagen, Hilden, Germany). The AllPrep DNA/RNA/Protein Kit (Qiagen) was employed for manual RNA and miRNA extraction and their concentration and purity were measured in the Nanodrop spectrophotometer (Thermo Scientific, Waltham, MA, USA).
For the RNA retro-transcription reaction, 2 µg of RNA were used in a total volume of 20 µL, employing the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, California, CA, USA). For the miRNA retro-transcription reaction, 10 ng of miRNA were used in a total volume of 15 µL, by means of the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems) with RT primers (5X) of each miRNA of interest.
For the quantification of miRNA expression, we used TaqMan Fast Advanced PCR Master Mix and assays based on hydrolysis probes (TaqMan Gene Expression Assays, Life Technologies). We selected the following assays (20X): has-miR-375 (000564), has-miR-200a (000502), has-miR-383 (000573), has-miR-103 (000439), and has-miR-488 (002357). As endogenous controls, we used has-miR-361 (000554) and has-miR-186 (002285). The qRT-PCR protocol was as follows: 95 • C for 20 s, 40 cycles of 95 • C for 1 s, and 60 • C for 20 s. The qRT-PCR protocol for IDT's Fast Cycling was as follows: 95 • C for 3 min, 35 to 45 cycles of 95 • C for 5 s, and 60 • C for 30 s. For the quantification of NEUROD1 and TBX19 mRNA expression, which our group carried out in a previous study [11], we used TaqMan Fast Advanced PCR Master Mix and an assay based on hydrolysis probes: NEUROD1 (Hs01922995_s1) and TBX19 (Hs00193027_m1) (TaqMan Gene Expression Assay, Life Technologies, California, CA, USA) and the same fast cycling protocol described above.
A pool of RNA/miRNA from nine normal pituitary samples obtained from autopsies served as a calibrator. All samples were analyzed in duplicate. We included negative controls to check that the probes did not amplify genomic DNA contaminant. The relative differences in gene expression were expressed as fold change and were obtained with the 2 −∆∆Ct method (SDS software, Applied Biosystems).

Statistical Analysis
We expressed qualitative variables, including PitNET subtypes, as absolute and relative frequencies.
Participant age and MTD were expressed as mean ± standard deviation (SD). We used the Shapiro-Wilk test to assess the distribution of the molecular variables (fold change (FC)), the Kruskal-Wallis and Mann-Whitney tests to compare qualitative variables and non-parametric quantitative molecular variables, and the student's t and ANOVA tests when the quantitative variables showed a normal distribution. We used Spearman and Pearson correlation tests to compare quantitative variables. The area under the receiver operating characteristics (ROC) curve (AUC) was used to assess the sensitivity and specificity of the miRNA biomarkers. p Values of less than 0.05 were considered statistically significant. Statistical analysis and ROC curves were performed with SPSS 24.0 software (IBM Software, Ohio, OH, USA). All other figures were performed using GraphPad Prism 6 (GraphPad Software Inc., California, CA, USA).

miR200a and miR-103 as Potential Diagnostic Biomarkers of Functionality in CTs
The miRNAs studied appeared up-or down-regulated in both functioning and silent CTs compared with normal pituitary glands, but only miR-200a and miR-103 showed significant differences between functioning CTs and their silent counterparts ( Figure 1). J. Clin. Med. 2020, 9, x FOR PEER REVIEW 4 of 12 expressed as fold change and were obtained with the 2 −ΔΔCt method (SDS software, Applied Biosystems).

Statistical Analysis
We expressed qualitative variables, including PitNET subtypes, as absolute and relative frequencies. Participant age and MTD were expressed as mean ± standard deviation (SD). We used the Shapiro-Wilk test to assess the distribution of the molecular variables (fold change (FC)), the Kruskal-Wallis and Mann-Whitney tests to compare qualitative variables and non-parametric quantitative molecular variables, and the student's t and ANOVA tests when the quantitative variables showed a normal distribution. We used Spearman and Pearson correlation tests to compare quantitative variables. The area under the receiver operating characteristics (ROC) curve (AUC) was used to assess the sensitivity and specificity of the miRNA biomarkers. p Values of less than 0.05 were considered statistically significant. Statistical analysis and ROC curves were performed with SPSS 24.0 software (IBM Software, Ohio, OH, USA). All other figures were performed using GraphPad Prism 6 (GraphPad Software Inc., California, CA, USA).

miR200a and miR-103 as Potential Diagnostic Biomarkers of Functionality in CTs
The miRNAs studied appeared up-or down-regulated in both functioning and silent CTs compared with normal pituitary glands, but only miR-200a and miR-103 showed significant differences between functioning CTs and their silent counterparts ( Figure 1). Eliminating the bias of tumor size, significant differences held between macro silent CTs and macro functioning CTs (p = 0.049 and p = 0.05, respectively). Therefore, we performed a ROC curve Eliminating the bias of tumor size, significant differences held between macro silent CTs and macro functioning CTs (p = 0.049 and p = 0.05, respectively). Therefore, we performed a ROC curve analysis in order to determine the potential usefulness of the expression of these two miRNAs in discriminating silent CTs from functioning CTs. Expression of both miR-200a and miR-103 enabled the differentiation between the two variants (AUC = 0.739, 95% CI 0.592 to 0.887, p = 0.007; AUC = 0.727, 95% CI 0.574 to 0.880, p = 0.011, respectively) (Supplementary Figure S1).
Additionally, in silent CTs, there were strong negative correlations between: (1)

TBX19 Could Be a Putative Target Gene of miR-383
Fortunately, the expression levels of TBX19 and NEUROD1, two transcription factors related with the processing of POMC, were available in the CT samples included in this study [11]. Therefore, we could compare the relationship between miR-383 and the expression of these two key genes. We found a negative correlation between miR-383 and TBX19 and a positive correlation with NEUROD1 in both functioning and silent CTs: TBX19, silent CTs rho = −0.583, p = 0.007 and functioning CTs rho = −0.470, p = 0.049; NEUROD1, silent CTs rho = 0.647, p = 0.005; functioning CTs rho = 0.445, p = 0.05. We found that NEUROD1 is a potential target for miR-383 using the three databases screened (Supplementary Table S2) according to the different computational algorithms used (Figure 3).

TBX19 Could Be a Putative Target Gene of miR-383
Fortunately, the expression levels of TBX19 and NEUROD1, two transcription factors related with the processing of POMC, were available in the CT samples included in this study [11]. Therefore, we could compare the relationship between miR-383 and the expression of these two key genes. We found a negative correlation between miR-383 and TBX19 and a positive correlation with NEUROD1 in both functioning and silent CTs: TBX19, silent CTs rho = −0.583, p = 0.007 and functioning CTs rho = −0.470, p = 0.049; NEUROD1, silent CTs rho = 0.647, p = 0.005; functioning CTs rho = 0.445, p = 0.05. We found that NEUROD1 is a potential target for miR-383 using the three databases screened (Supplementary Table S2) according to the different computational algorithms used (Figure 3).
Indeed, there was a positive correlation between MTD and both miR-488 and miR-200a expression in the global series (rho = 0.472, p = 0.003 and rho = 0.420, p = 0.010, respectively) and in silent CTs (rho = 0.435, p = 0.035 and rho = 0.594, p = 0.005, respectively). However, there were not significant differences in the expression of both miRNAs between invasive and non-invasive corticotroph tumors (Figure 4b).

Discussion
The present study analyzes the expression of a subset of transcription factors, kinases, and miRNAs related with the melanocortinergic pathway and with proliferation, differentiation, and transcription regulation in a large series of functioning and silent CTs. Approximately 2000 miRNAs have been identified in humans, and these seem to regulate a third of the genes in the human genome [27]. Furthermore, miRNAs have been shown to play a relevant role in many biological processes. Indeed, the altered miRNA expression has been involved in the pathogenesis of numerous cancers. As shown in Supplementary Table S3, in PitNETs different miRNA expression patterns and potential targets of these miRNAs have been described, depending on the tumor subtypes [28][29][30]. In this way, our research provides new information on the potential role of miRNA on the silencing mechanisms in some corticotroph tumors that do not produce Cushing syndrome. Among the most relevant

Discussion
The present study analyzes the expression of a subset of transcription factors, kinases, and miRNAs related with the melanocortinergic pathway and with proliferation, differentiation, and transcription regulation in a large series of functioning and silent CTs. Approximately 2000 miRNAs have been identified in humans, and these seem to regulate a third of the genes in the human genome [27]. Furthermore, miRNAs have been shown to play a relevant role in many biological processes. Indeed, the altered miRNA expression has been involved in the pathogenesis of numerous cancers. As shown in Supplementary Table S3, in PitNETs different miRNA expression patterns and potential targets of these miRNAs have been described, depending on the tumor subtypes [28][29][30]. In this way, our research provides new information on the potential role of miRNA on the silencing mechanisms in some corticotroph tumors that do not produce Cushing syndrome. Among the most relevant results obtained is the differential expression of miR-200a and miR-103 in silent CTs compared with functioning CTs, with a potential usefulness in the discrimination of both variants. Additionally, our results indicate that miR-383 is overexpressed in CTs, especially in silent CTs, and correlated negatively with TBX19, a transcription factor related with the proccesing of POMC.

miRNAs as Potential Diagnostic Biomarker of Silent CTs
As shown, miR-200a and miR-103 exhibited significantly higher expression in silent than in functioning CTs, even after eliminating the bias of tumor size, enabling the differentiation between the two variants. Given their stability in extracellular fluids, there is increasing interest in the use of miRNAs as circulating biomarkers [31], and correlations have been observed between the expression of miRNAs in serum or plasma and their expression in tumors [31,32]. However, no circulating miRNAs are being employed in the case of PitNETs. As Di Leva et al. reported in 2014 [33], since the pituitary is highly vascularized and releases hormones into the circulation, extracellular miRNAs would be useful as diagnostic biomarkers of PitNETs. Due to the distinct aggressive behavior of the PitNETs subtypes, the possibility of identifying its silent counterparts in a non-invasive way could be of great help to plan the surgical intervention. So, the tumor overexpression of miR-200a and miR-103 and the accuracy we observed in discriminating silent from functioning CTs, encourages the study of the diagnostic accuracy of its determination in peripheral blood before surgery.

Silencing Role of miRNAs in CTs
There is little data describing the role of miRNAs in the silencing mechanisms of PitNETs. Recently, Neou et al. [23] analyzed an important series of 134 PitNETs, 35 of whom were corticotroph tumors, finding that CT tumors were arranged in two clusters with different miRNA signatures, with clear differences between functioning tumors and their silent counterparts. The higher expression levels of miR-200a and miR-103 demonstrated in this study suggest that miRNAs could be involved in the silencing of this clinically relevant PitNET subtype. Since all microadenomas were functioning CTs and all silent CTs were macroadenomas, there is a bias regarding to their size. Therefore, we only compared functioning and silent macroadenomas. Excluding functioning microadenomas, miR-200a and miR-103 continued to be differentially expressed in both variants. Indeed, in silent but not in functioning CTs, we observed a strong negative correlation between miR-200a and miR-103 and the expression of a key functional gene in this PitNET subtypes such as MAP3K8. Moreover, miR-103 also correlated negatively with the expression of MEK. As previously described, the effect of CRH on POMC transcription is mediated by the MAPK pathway, which implicates different genes, such as MAP3K8 and MEK [14,34]. For that reason, miR-200a and miR-103 could be functionally connected with the POMC gene by inhibiting its expression. Indeed, in our study, both miRNAs appeared down-regulated in functioning but normo-regulated in silent CTs.
In addition, our results revealed that other miRNAs, such as miR-375, miR-488 and miR-383, could also be directly involved in the silencing of CTs. Indeed, we found a higher expression of miR-375 in silent CTs than in functioning ones, and it has been described that MAP3K8 is a direct target gene of miR-375 in AtT-20 mouse pituitary tumor cells [14]. Moreover, we observed that miR-488 was down-regulated and presented a negative correlation with NURR1, PITX1, and STAT3 (all POMC expression enhancers) in functioning CTs. Therefore, these results might suggest that miR-488 could exert a permissive effect on the expression of POMC in functioning CTs, probably through the STAT3 pathway [15,35]. In the same way, miR-383 was especially up-regulated in silent CTs and, as it showed a negative correlation with STAT3, this miRNA might also participate in the silencing mechanisms of CTs [16]. Indeed, we found that miR-383 correlated negatively with TBX19, a transcription factor necessary for the differentiation and maintenance of adult corticotroph cells [36,37] (Figure 3). Finally, as somatostatin receptor genes (SSTR2, 3 and 5) are putative targets for miR-383, and STAT3 has been proposed as a promising target candidate for pituitary tumor immunotherapy [38], our results might also suggest that miR-383 could be a therapeutic target in corticotroph PitNETs.

miRNAs: Tumor Size and Proliferation
Several miRNAs have been associated with PitNET size and growth [29]. In 2005, Bottoni et al. observed a negative correlation between tumor diameter and expression of miR-15a and miR-16-1 in somatotroph and lactotroph tumors [17]. Later, the same group described six miRNAs able to differentiate microadenomas from macroadenomas in non-secreting tumors [18]. In the same way, Butz et al. suggested that miR-629 and miR-214 may contribute to tumor growth in non-secreting pituitary tumors through an impaired regulation of apoptosis [39]. Finally, Renjie at al. demonstrated a tumor suppressor role of miR-15a, miR-16, and miR-132 in pituitary tumor cells (GH3 and MMQ cells) and in most PitNETs [40]. However, to the best of our knowledge, none of these studies has focused on the analysis of miRNAs and key regulatory genes in silent and functioning CTs. Specifically, our study revealed that miR-488, miR-200a, and miR-103 showed higher expression levels in macroadenomas than in microadenomas. As all microadenomas were functioning CTs and all silent CTs were macroadenomas, we focused on silent CTs. Thus, we found that miR-488 and miR-200a expression levels were positively correlated with tumor size, suggesting that these miRNAs could participate in the growth of these tumors. Our results are concordant with other studies reporting that miR-488 and miR200 play a role as tumor suppressor agents in several types of cancers [41][42][43][44][45][46][47]. In addition, we observed a strong negative correlation between the expression of MAP3K8 and miR-488, miR-200a and miR-103, and also between the expression of MEK and miR-488 and miR-103 in silent CTs. As MAP3K8 acts as an oncogene in numerous cancer types [48][49][50] and the MEK gene is also involved in a wide variety of cellular processes such as proliferation [51,52], the down-regulation in the expression of these miRNAs could facilitate the oncogenic effects of MAP3K8/MEK in the growth of CT PitNETs. These observations unveiled new conceptual and functional avenues in CTs, with potential therapeutic implications, which warrant further investigation.

Conclusions
In conclusion, the evidence from this study suggests that miRNAs could participate in the silencing and growth of CTs. Specifically, we demonstrated that miR-200a, miR-103, miR-375, miR-488 and miR-383 could participate in the silencing mechanisms of these clinically relevant tumors, which are not associated with Cushing syndrome. In addition, miR-488, miR-200a and miR-103 could also be involved in the growth of CTs. Altogether, the translational research implications of these findings indicate that these miRNAs might be involved in the pathophysiology of CTs, suggesting that pharmacological treatments specifically targeting these miRNAs could become a promising therapeutic option for patients in the future, and provide a relevant clinical conclusion, which should soon be tested in humans.