Identification of Dysregulated Expression of G Protein Coupled Receptors in Endocrine Tumors by Bioinformatics Analysis: Potential Drug Targets?

Background: Many studies link G protein-coupled receptors (GPCRs) to cancer. Some endocrine tumors are unresponsive to standard treatment and/or require long-term and poorly tolerated treatment. This study explored, by bioinformatics analysis, the tumoral profiling of the GPCR transcriptome to identify potential targets in these tumors aiming at drug repurposing. Methods: We explored the GPCR differentially expressed genes (DEGs) from public datasets (Gene Expression Omnibus (GEO) database and The Cancer Genome Atlas (TCGA)). The GEO datasets were available for two medullary thyroid cancers (MTCs), eighty-seven pheochromocytomas (PHEOs), sixty-one paragangliomas (PGLs), forty-seven pituitary adenomas and one-hundred-fifty adrenocortical cancers (ACCs). The TCGA dataset covered 92 ACCs. We identified GPCRs targeted by approved drugs from pharmacological databases (ChEMBL and DrugBank). Results: The profiling of dysregulated GPCRs was tumor specific. In MTC, we found 14 GPCR DEGs, including an upregulation of the dopamine receptor (DRD2) and adenosine receptor (ADORA2B), which were the target of many drugs. In PGL, seven GPCR genes were downregulated, including vasopressin receptor (AVPR1A) and PTH receptor (PTH1R), which were targeted by approved drugs. In ACC, PTH1R was also downregulated in both the GEO and TCGA datasets and was the target of osteoporosis drugs. Conclusions: We highlight specific GPCR signatures across the major endocrine tumors. These data could help to identify new opportunities for drug repurposing.


Introduction
G protein-coupled receptors (GPCRs) are the largest family of membrane receptors involved in many types of cellular responses. The GPCR family represents approximately 4% of the human genes, with more than 800 members [1]. GPCRs are involved in important functions, such as cardiac function, hormone regulation, immune responses and neurotransmission. Abnormal expression or activity is associated with several human diseases [2]. As a result, GPCRs are considered as therapeutic targets in many diseases. Drugs that target GPCRs account for about 34% of the current medicines, making it the

Collection of GEO Datasets and TCGA Data
To comprehensively evaluate GPCRs expression between endocrine tumors and nontumoral tissues, we incorporated the expression data of GPCRs from GEO. We selected four types of endocrine tumors: medullary thyroid cancer (MTC), pituitary neuroendocrine tumors (PitNETs), pheochromocytoma (PHEO) and paraganglioma (PGL) and adrenocortical carcinoma (ACC). An electronic search of the GEO databases was carried out with the following key terms: "medullary thyroid", "adrenocortical", "pituitary" "pheochromocytoma", "paraganglioma", AND "cancer OR carcinoma OR tumor". The GEO datasets published before 1 April 2020 were included. We filtered data by 'Expression profiling by array' in 'homo sapiens'. The reference lists were manually reviewed for further identification of relevant studies. We selected the studies with the following inclusion criteria: studies contained mRNA expression data in tumoral and non-tumoral tissues in previous cited endocrine tumor. The exclusion criteria were: cell lines, xenograft studies, miRNA analyses, no comparison with normal tissue, duplicate. The microarray datasets were downloaded from the GEO database for each endocrine tumor and are summarized in Table 1.

Identification of Differentially Expressed Genes
DEGs between tumoral and normal samples were identified using the GEO2R web application (http://www-ncbi-nlm-nih-gov.proxy.insermbiblio.inist.fr/geo/geo2r/, accessed on 7 January 2022). GEO2R is an interactive online tool, which compares two groups of samples (i.e., normal vs. tumoral) to obtain genes with different expressions under the same experimental conditions. It uses Bioconductor R packages to transform and analyze GEO data with GEO query and limma analyses [36]. We applied adjustment to the P-values with the Benjamini & Hochberg method (false discovery rate, FDR) and a log 2 transformation to the data. We saved the NCBI annotations where possible. Then, we filtered, from the set of genes studied, a list of 396 GPCRs (Table S1). The DEGs with an adjusted p < 0.05 were considered as the cut-off criteria. Then, to strengthen the significance of the data, we chose to select DEGs obtained from all independent datasets of tumor tissue samples rather than combine the GSE data, as in previous publications [11,37]. We applied for Venn software online (http://www.interactivenn.net/, accessed on 26 June 2020) to obtain the common DEGs in all independent cohorts for each endocrine tumor [38].

Identification of GPCRs as Targets for Approved Drugs
We evaluated drug targets among GPCRs differentially expressed between normal and tumoral samples and the list of approved drugs for each of these receptors. This was conducted using two pharmacological databases: ChEMBL and DrugBank. DrugBank is a clinically oriented drug database that provides information about drug and drug action for more than 500,000 drugs [39]. ChEMBL is an open database containing information extracted from the medicinal chemistry literature regarding the compounds tested and their primary target [40]. Approved drugs were verified with those listed by the Food and Drug Administration (FDA) (https://www.fda.gov, accessed on 3 January 2022) and by the European Medicines Agency (EMA) (https://www.ema.europa.eu, accessed on 3 January 2022).

Survival Analysis
The GEPIA database (http://gepia.cancer-pku.cn, accessed on 14 February 2022) was used for survival analysis of the dysregulated GPCR genes. Data from TCGA-ACC project (88 patients with ACC) were analyzed. To analyze the prognostic value of GPCR genes, the patient samples were split into 2 groups according to the median expression of the biomarker. The two patient cohorts were compared by a Kaplan-Meier survival plot, and the logrank P value was calculated.

Medullary Thyroid Cancer
For medullary thyroid cancer (MTC), only one study was available from the GEO database (GSE27155). The samples consisted of four normal thyroid tissues and two medullary carcinomas. The analysis revealed 14 DEGs for GPCRs (seven downregulated and seven upregulated) ( Table 2). The TSH receptor (TSHR) was the main receptor downregulated in MTC compared to normal thyroid tissue (log 2 FC: −2.22) as well as GPRC5A; an orphan receptor was also found to be downregulated (log 2 FC: −1.93). For the upregulated GPCRs, the majority were weakly upregulated, such as dopamine receptor (DRD2, log 2 FC: 0.62) and cholecystokinin receptor (CCKBR, log 2 FC: 0.59). The two most upregulated GCPRs compared to normal thyroid tissue were the adhesion receptor ADGRG2 (log 2 FC: 1.42) and the receptor LGR5 (Leucine Rich Repeat Containing G Protein-Coupled Receptor 5, log 2 FC: 1.43).

Pheochromocytoma and Paraganglioma
Four gene expression profiles (GSE50442, GSE39716, GSE19422, GSE60459) were acquired from the GEO database for pheochromocytoma and paraganglioma. The dataset contained from two to sixty-one pheochromocytoma samples, four to twenty-four paraganglioma and from three to eight normal samples (medullar adrenal tissue). All PHEOs were non-metastatic. For PGL, GSE39716 and GSE60459 included ten and three metastatic paraganglioma, respectively.
The data for PHEO were quite heterogeneous among the four studies. The number of GPCRs with a significant difference in expression between healthy and tumor tissue varied between 12 and 86 ( Figure 2, Table S3). Although there was no receptor significantly common to all four series, the angiotensin II type 2 receptor (AGTR2) was found to be downregulated in the tumor tissue in all the studies (GSE60459, log 2 FC = −1.1, p = 0.03; GSE50442, log 2 FC = −0.56, p = 0.493; GSE39716, log 2 FC = −0.75, p = 0.003; GSE19422, log 2 FC = −0.60, p = 0.003).
In PGL, the analysis revealed from 37 to 69 DEGs for GPCRs. Venn diagram analysis revealed that seven GPCRs were down regulated compared with normal adrenal gland in the four datasets (AVPR1A, MC2R, NPY5R, NPY6R, RXFP1, LGR4, PTH1R) ( Figure 3, Table S4). When we compared metastatic PGL and non-metastatic PGL in GSE39716 and GSE60459, no significant DEG were found.

Pheochromocytoma and Paraganglioma
Four gene expression profiles (GSE50442, GSE39716, GSE19422, GSE60459) were acquired from the GEO database for pheochromocytoma and paraganglioma. The dataset contained from two to sixty-one pheochromocytoma samples, four to twenty-four paraganglioma and from three to eight normal samples (medullar adrenal tissue). All PHEOs were non-metastatic. For PGL, GSE39716 and GSE60459 included ten and three metastatic paraganglioma, respectively.
The data for PHEO were quite heterogeneous among the four studies. The number of GPCRs with a significant difference in expression between healthy and tumor tissue varied between 12 and 86 ( Figure 2, Table S3). Although there was no receptor significantly common to all four series, the angiotensin II type 2 receptor (AGTR2) was found to be downregulated in the tumor tissue in all the studies (GSE60459, log2FC = −1.1, p = 0.03; GSE50442, log2FC = −0.56, p = 0.493; GSE39716, log2FC = −0.75, p = 0.003; GSE19422, log2FC = −0.60, p = 0.003).
In PGL, the analysis revealed from 37 to 69 DEGs for GPCRs. Venn diagram analysis revealed that seven GPCRs were down regulated compared with normal adrenal gland in the four datasets (AVPR1A, MC2R, NPY5R, NPY6R, RXFP1, LGR4, PTH1R) ( Figure 3, Table  S4). When we compared metastatic PGL and non-metastatic PGL in GSE39716 and GSE60459, no significant DEG were found. (A)

Adrenocortical Carcinomas
For ACC, five gene expression profiles (GSE14922, GSE90713, GSE12368, GSE19750, GSE33371) were acquired from the GEO database. The dataset contained from four to fifty-seven cancer samples and from four to ten normal samples. The analysis of these datasets revealed six, eight, fifteen, nine and fifty-two DEGs, respectively. Venn diagram analysis revealed that PTH1R was the only gene downregulated in the five datasets (Figure 4, Table S5). To validate this finding, TCGA data (TCGA-ACC Project) were used to quantify differential expression of PTH1R by comparing tumors against adrenal normal tissue from the Gene Tissue Expression Project (GTEx) database. In the TCGA cohort, PTH1R was also downregulated (log2FC: −3.37, p < 0.0001).

Adrenocortical Carcinomas
For ACC, five gene expression profiles (GSE14922, GSE90713, GSE12368, GSE19750, GSE33371) were acquired from the GEO database. The dataset contained from four to fiftyseven cancer samples and from four to ten normal samples. The analysis of these datasets revealed six, eight, fifteen, nine and fifty-two DEGs, respectively. Venn diagram analysis revealed that PTH1R was the only gene downregulated in the five datasets ( Figure 4, Table S5). To validate this finding, TCGA data (TCGA-ACC Project) were used to quantify differential expression of PTH1R by comparing tumors against adrenal normal tissue from the Gene Tissue Expression Project (GTEx) database. In the TCGA cohort, PTH1R was also downregulated (log 2 FC: −3.37, p < 0.0001).

Identification of Dysregulated G Protein Coupled Receptors as Targets for Approved Drugs
Among the dysregulated GPCR, we determined which receptors were targeted by approved drugs and how many drugs were available for these targets. The data were summarized in the Table 3. In MTC, among the 14 DEG, CCKBR, D2R and Adenosine A2B receptor (ADORA2B) were the target of drugs. DRD2 was the most frequently targeted receptor with over 40 anti-psychotic agents acting as DRD2 antagonists. For PitNET, any drug did not target FZD7. Interestingly, three up-regulated GPCR in invasive PitNET were targeted receptor: SSTR1, HTR4 and PTGER2. Among the seven down-regulated GPCRs in paraganglioma, Arginine Vasopressin Receptor 1A (AVPR1A), ACTH receptor, MC2R, and PTH1R were the target of 2 to 4 drugs each. PTH1R also found to be downregulated in adrenocortical cancer, was targeted by 3 agonists used in osteoporosis or hypoparathyroidism.

Identification of Dysregulated G Protein Coupled Receptors as Targets for Approved Drugs
Among the dysregulated GPCR, we determined which receptors were targeted by approved drugs and how many drugs were available for these targets. The data were

Survival Analysis
Survival data were only available for ACC from TCGA. We, therefore, chose to study the gene expression of PTH1R for ACC. Using the GEPIA database to explore the association between gene expression and survival, we found that PTH1R expression was neither significantly correlated with overall survival nor disease-free survival of ACC patients (log-rank p = 0.72 and log-rank p = 0.88, respectively) ( Figure 5).

Survival Analysis
Survival data were only available for ACC from TCGA. We, therefore, chose to study the gene expression of PTH1R for ACC. Using the GEPIA database to explore the association between gene expression and survival, we found that PTH1R expression was neither significantly correlated with overall survival nor disease-free survival of ACC patients (log-rank p = 0.72 and log-rank p = 0.88, respectively) ( Figure 5).

Discussion
This study constitutes the first atlas of GPCRs within major endocrine tumors after thyroid cancer [14]. Drug repurposing strategies provided a new approach for revealing new clinical applications of an approved drug and for developing antitumor drugs. The use of GPCRomic approaches leads to the discovery of GPCRs that could contribute to cancer pathophysiology and, thus, may be therapeutic targets.

Discussion
This study constitutes the first atlas of GPCRs within major endocrine tumors after thyroid cancer [14]. Drug repurposing strategies provided a new approach for revealing new clinical applications of an approved drug and for developing antitumor drugs. The use of GPCRomic approaches leads to the discovery of GPCRs that could contribute to cancer pathophysiology and, thus, may be therapeutic targets.

Medullary Thyroid Cancer
We confirmed the upregulation of cholecystokinin 2 receptor (CCKBR/CCK2R) expression in MTC, which is targeted by radiolabeled peptide analogs for molecular imaging and targeted radiotherapy of different human tumors, such as MTC [41,42]. The dopamine receptor DRD2 was also upregulated. It is known that L-dopa inhibits calcitonin secretion in MTC, and a new chimeric somatostatin-dopamine analog has recently shown an anti-tumor effect in vitro [43]. Interestingly, DRD2 is the target of many drugs already approved and is, therefore, accessible for drug repurposing. Like DRD2, the adenosine receptor ADORA2B was upregulated and the target of approved drugs. However, sparse data are available on the role of these treatments in thyroid cancer, although previous studies have shown a role for adenosine receptors in calcitonin secretion in C-cells [44]. ADGRG2/GPR64 was also over-expressed in MTC (log 2 FC = 1.42, p = 0.0049) as in a number of carcinomas derived from breast cancer, parathyroid tumors, Ewing sarcomas, prostate, kidney or lung, and the inhibition of its expression promotes invasiveness and metastatic spread [45,46]. GPRC5A was downregulated (log 2 FC = 1.93, p = 0.0002) as previously reported in other cancers, especially in lung cancer, where it displayed a tumor suppressive role [47]. FZD1 was also downregulated as previously described in non-medullary thyroid cancer, in which inhibition increased invasiveness, indicating a possible pathogenesis role [48]. However, to have only one study with a very small number of samples limits the significance of the findings for medullary thyroid cancer. In addition, the different histological origin between the normal tissue (mainly of follicular origin) and tumor tissue (medullary origin) probably explains part of the results. This is the case, for example, with the downregulation of the TSH receptor (TSH-R), which is strongly expressed in normal thyroid tissue but was negligible in MTC, as expected [49].

Pituitary Neuroendocrine Tumors (PitNETs)
Five studies were available in the GEO database comparing pituitary adenomas and the normal pituitary. Data regarding aggressiveness were not available. Independent of tumor characteristics (secreting or non-secreting, invasive or non-invasive), Frizzled receptor 7 (FZD7) was downregulated in PitNETs. Interestingly, in invasive adenomas compared to the normal pituitary gland (GSE51618) FZD7 was also found to be downregulated. To date, this data was not described before in the literature. However, SFRP2 and 4 (secreted-frizzled related protein), Wnt antagonists, considered as tumor suppressor genes, were downregulated in pituitary adenomas, in particular in invasive tumors, suggesting a role of the Wnt pathway in the progression of PitNETs [50][51][52]. Among the differentially expressed GPCR genes between non-invasive and invasive tumors, the somatostatin type 1 receptor (SSTR1) was known and already targeted by somatostatin analogs. HTR4 and PTGER2 were also targeted by agonists. The antitumoral and antiangiogenic effects of HTR4 agonists were thus shown [53,54].

Pheochromocytoma or Paraganglioma
For PHEO, no receptor was significantly common to all four series. However, even if, in one study, angiotensin II type 2 receptor (AGTR2) was not statistically downregulated, it was found to be downregulated in the tumoral tissue in all the studies. Indeed, the absence or a very low expression of AGTR2 has been described in PHEO [55]. Interestingly, this receptor has been also described for its antiproliferative role in cancer, including in the pheochromocytoma PC12 cell line in response to angiotensin II [56]. However, unlike the angiotensin II type 1 receptor, no approved drug targets the AGTR2. In addition, it would be interesting to have metastatic pheochromocytomas to compare the profile between metastatic and non-metastatic, as for paragangliomas. However, no data were available.
For PGL, seven GPCRs were found to be downregulated when compared to normal medulla adrenal tissue. However, no data are available in the literature regarding their expression level or their role in paragangliomas. Interestingly, the arginine vasopressin receptor, AVPR1A, was also found to be downregulated in thyroid cancer and associated with progression-free survival, while it was upregulated in castration-resistant prostate cancer with an antitumoral effect of a selective AVPR1A antagonist [14,57]. Moreover, AVPR1A agonist and antagonists are approved drugs and could be tested as proof of concept. When we compared malignant paraganglioma and benign paraganglioma, no significant DEGs were found. The patient numbers were small in the two datasets, probably limiting the statistical power of the analyses.

Adrenocortical Carcinoma
In ACC, PTH1R was found downregulated in 5/5 datasets. This result was confirmed in ACC samples from the TCGA cohort. PTH1R is a major endocrine regulator of skeletal development and mineral homeostasis. It was also found that PTH1R enhanced the secretion of steroid hormones by human adrenocortical cells via a signaling mechanism involving the activation of both the Gs and Gq pathways [58]. PTH1R was previously immunolocalized in the cytoplasm in a normal adrenal cortex [59]. The PTH1R mRNA expression level was significantly higher in aldosterone-producing adenomas than in cortisol-producing adenoma [59] but seemed to be similar between ACCs and adenomas [60]. To the best of our knowledge, this is the first comparison between ACCs and a normal adrenal cortex. We found a downregulation of PTH1R, while PTHrp seemed to have a positive effect on the proliferation of the H295R lineage [60]. This could be explained by a downregulation of PTH1R by PTHrP, as described in vascular smooth muscle, osteoblastic or kidney cell lines [61]. The role of PTH1R in tumor progression is still poorly understood. Indeed, PTH1R knockdown increased cell migration and invasion and decreased tumor differentiation in neuroblastoma but had an opposite effect in osteoblastoma [62]. Finally, the PTH1R expression was neither significantly correlated with overall survival nor disease-free survival in ACC patients. However, there were only 88 patients in the study, probably limiting the statistical power of the analyses. Moreover, the absence of an association with survival should not, in our opinion, compromise potential therapeutic targets, as in thyroid cancer, where TSH inhibition influences survival even though its receptor expression level in the tumor is not associated with overall survival [14,63].
RNA-seq and microarrays that assess GPCR expression have certain limitations. First, RNA expression does not necessarily reflect protein expression. However, methods to assess GPCR protein expression are limited. Immunological techniques are limited by the lack of well-validated and specific GPCR antibodies [64]. The low expression of GPCRs compared to other proteins makes detection of GPCRs complicated. Moreover, recent studies found a closer relationship between the mRNA and protein, especially for GPCRs [65,66]. Another alternative to validate GPCR expression would be cell signaling and functional assays. Second, the clinical data were not associated with the microarrays data in the GEO repository, which limits the evaluation of the prognostic impact of these GPCRs on recurrence or progression. The data from the GPCRomic analyses need to be validated by further prospective studies for specific GPCRs with clinical evaluation.

Conclusions
This study presented all the data available to date regarding GPCRs in endocrine tumors. The data could help to identify new potential pathways worth targeting, eventually with approved drugs after proving the tumoral impact in preclinical models. The highthroughput screening of clinically relevant GPCRs with approved drugs would then be an interesting tool for the evaluation of these molecules in endocrine tumors.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/cells11040703/s1, Table S1: List of GPCR genes included in bioinformatics analysis. Table S2: List of DEGs selected from analysis of pituitary adenoma datasets. Table S3: List of DEGs selected from analysis of pheochromocytoma datasets. Table S4. List of DEGs selected from analysis of paraganglioma datasets. Table S5: List of DEGs selected from analysis of adrenocortical cancer datasets.  Data Availability Statement: The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript or in the decision to publish the results.