This article is mentioned in:

Introduction

Thymoma, which is derived from the epithelial cells of the thymus, is a rare malignant tumour type. At the time of diagnosis, patients are usually between 40 and 60 years of age (1); certain patients present with myasthenia graves or other symptoms, including superior vena cava syndrome, dysphagia, cough or chest pain (2). In general, upon diagnosis with thymoma, patients undergo surgical treatment. However, for patients with stage III and IV disease, the 5-year overall survival rates are 74% and <25%, respectively (3), indicating that those patients require further treatment. Furthermore, even after post-operative radiotherapy, the overall survival rates of patients were not significantly improved (4). Therefore, patients with advanced disease are candidates for chemotherapy. According to certain international guidelines, the first-line combination chemotherapy regimens for thymoma include the following: i) Cisplatin, doxorubicin and cyclophosphamide (CAP), ii) CAP with prednisone, iii) cisplatin, doxorubicin, cyclophosphamide and vincristine, iv) cisplatin and etoposide, v) etoposide, ifosfamide and cisplatin and vi) carboplatin/paclitaxel, while the second-line chemotherapy regimens include etoposide, ifosfamide, pemetrexed, 5-fluorouracil or its analogues, gemcitabine and paclitaxel, separately (5). However, even after chemotherapy, the 10-year disease-free survival rate is only 56% for patients with stage III thymoma and 33% for patients with stage IV disease (6).

Therefore, exploration of effective chemotherapeutic agents to treat thymoma is required. However, most drug development strategies primarily involve determining a novel therapeutic target and then searching for compounds that fit the target. Therefore, agent discovery is an expensive endeavor that frequently involves issues with bioavailability first and toxicity later (7). However, drug repositioning, with the aim of identifying novel applications for existing drugs, has been established as a cost-effective strategy. Over the last decade, burgeoning computer technologies have made structure-based compound screening a prevalent tool in early drug identification (8). Among them, Connectivity Map (CMap), which is based on RNA chip technology, is a useful data resource for studying drug mechanisms and drug reallocation (9). Furthermore, all drugs in the database are ranked according to a score, which is derived from conventional measures, including the Pearson correlation coefficient. For instance, those with a score of 1 are the most strongly positively correlated with the query signature, and those with a score of −1 are most strongly negatively correlated. Consequently, by using the differentially expressed genes (DEGs) between thymoma and normal tissues, CMap is able to identify the affected pathways and small-molecule drugs that may be considered potential therapeutic agents for treating thymoma. Accordingly, the use of RNA chip technology may provide novel ideas for the treatment of thymoma.

In the present study, connectivity mapping was performed based on DEGs identified in a large population of patients with thymoma in The Cancer Genome Atlas (TCGA) and GTEx databases, accompanied by the procurement of prospective candidate drugs for the future treatment of thymoma.

Materials and methods

Screening of DEGs in thymoma

Gene expression profiling interactive analysis (access: August 9, 2018; GEPIA; http://gepia.cancer-pku.cn/) was used to identify the DEGs in thymoma. GEPIA uses data from TCGA and GTEx projects and is a freely available tool to consign customizable functionalities (10). Four-way analysis of variance (ANOVA) was used to calculate differential expression based on variables, including sex, age, ethnicity and disease stage. First, on the website, all the expression profiles were transformed to the log2 [transcripts per million (TPM)+1] scale for further calculations. Simultaneously, the median (tumour-normal) was defined by the log2 [fold change (FC)]. The adjusted q-value was then obtained after multiple adjustments using the Benjamini and Hochberg false discovery rate method for each variable. The cut-off values for abnormally expressed mRNAs were a log2FC >3 and a q-value of <0.05.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses, and protein-protein interaction (PPI) network of the DEGs

GO and KEGG analyses were used to further understand the roles of DEGs in thymoma (11). The Database for Annotation, Visualization and Integrated Discovery (DAVID) 3.8 (access time: April 10, 2019, http://david.ncifcrf.gov) was applied to identify the GO terms and KEGG pathways in which the DEGs were enriched (12–14). A P-value of <0.05 was considered to indicate significance. Simultaneously, the Search Tool for the Retrieval of Interacting Genes/proteins (STRING) database (access time: April 15, 2019, http://string-db.org), with the cut-off criterion of a combined score of >0.9, was utilized to provide an appraisal and integration of the PPI network (15,16).

Discovery of potential therapeutic drugs depended on the CMap database

CMap was utilized to identify potential therapeutic drugs for thymoma treatment (17). First, Affymetrix gene chips, one of the most frequently used platforms for expression profiling, were used to convert a gene symbol to a signature in HG-U133A format (18). However, with the limitations of the website, a log2FC >4.0 and a q-value of <0.05 were selected as cut-off values for overexpressed genes. Cutoff values for the downregulated expression of mRNAs were a log2FC of >3.0 and a q-value of <0.05. After the data were inputted into CMap (access time: April 6, 2018, http://portals.broadinstitute.org/cmap/), the corresponding compounds with scores of <-0.75 were considered potential drugs for the treatment of thymoma. Finally, the top 10 compounds were subjected to further analysis, and their molecular structures were obtained from PubChem (access time: April 6, 2019, http://www.ncbi.nlm.nih.gov/pccompound) (19).

Construction of the drug-target network

A drug-target network was constructed to further examine the potential mechanisms of action of the compounds. The freely available STITCH software (access time: April 11, 2019, http://prion.bchs.uh.edu/stitch/) was used to construct the drug-target network and to provide genetic identification (20). A higher score indicated a greater likelihood that the drug targeted the tested gene product (21).

Molecular docking analysis of the interactions between proteins and compounds

The Surflex-Dock program in Sybyl version X-2.0 was used to verify the interactions between drugs and targets. The program obtained particular information on how these novel drugs exert their anti-tumour activity. Simultaneously, the docking scores and crash were calculated to represent binding affinities and the degree of inappropriate penetration into the protein by the ligand (22–24).

Functional annotation of the targets

To further explore the functions of the compounds in combination with the target, KEGG pathway and GO term analyses of the targets were performed using the tools mentioned above.

Screening compounds based on absorption, distribution, metabolism and excretion (ADME) parameters in the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP) database

The ADME parameters were used for drug screening and evaluation to further determine the potential clinical value of the compounds. The TCMSP database (access time: April 6, 2018, http://sm.nwsuaf.edu.cn/lsp/tcmsp.php) was used to query the ADME parameters of the top 10 drugs identified above (25). The compounds were then screened with the following criteria: Oral bioavailability (OB) ≥30%, drug-likeness (DL) ≥0.18 and drug half-life (HL) ≥4 h.

Results

GO and pathway analyses of the DEGs associated with thymoma

In total, 2,447 DEGs, including 2,204 upregulated and 243 downregulated genes, were identified in 118 thymoma patients and 339 normal samples (2 normal tissues and 337 normal blood samples) (Fig. 1; Table I). A GO analysis was further employed to determine the possible molecular mechanisms of the DEGs. In the category biological process (BP), these genes were most enriched in ‘translation’, ‘SRP-dependent co-translational protein targeting to membrane’, ‘viral transcription’, ‘nuclear-transcribed mRNA catabolic process, nonsense-mediated decay’ and ‘rRNA processing’ (Fig. 2A). In the category cellular component (CC), ‘ribosome’, ‘extracellular exosome’, ‘mitochondrial inner membrane’, ‘mitochondrion’ and ‘cytosolic large ribosomal subunit’ were the most prominent terms (Fig. 2B). In the category molecular function (MF), these genes were significantly involved in ‘poly(A) RNA binding’, ‘structural constituent of ribosome’, ‘protein binding’, ‘RNA binding’ and ‘NADH dehydrogenase (ubiquinone) activity’ (Fig. 2C). In the KEGG pathway analysis, ‘ribosome’, ‘oxidative phosphorylation’, ‘Huntington's disease’, ‘Parkinson's disease’ and ‘Alzheimer's disease’ were the most prominent pathways (Fig. 2D). Among these pathways, ‘ribosome’, with 93 genes included, was the most significant pathway. The PPI networks comprising the DEGs revealed that certain genes, including complement C3 (C3), Serum amyloid A1 (SAA1), C-X-C motif chemokine receptor 1 (CXCR1), C-X-C motif chemokine receptor 2 (CXCR2), C-X-C motif chemokine ligand 12 (CXCL12), C-C motif chemokine ligand 19 (CCL19), C-C motif chemokine ligand 25 (CCL25) and Formyl peptide receptor were closely linked with a high degree (Fig. 3).

Figure 1. DEGs between thymoma and normal tissues. (A) Locations of the DEGs based on GRCh38.p2 (NCBI). (B) Volcano plot displaying the DEGs. DEG, differentially expressed gene; NCBI, National Center for Biotechnology Information.

Figure 2. GO terms and KEGG pathway analysis for the differentially expressed genes. (A-C) GO terms in the categories (A) biological process, (B) molecular function and (C) cellular component; (D) KEGG pathway analysis. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; hsa, Homo sapiens.

Figure 3. The PPI network of the significant DEGs. The PPI network of DEGs with the cut-off criterion of a combined score >0.9 had a total of 47 nodes and 96 edges. PPI, protein-protein interaction; DEG, differentially expressed gene.

Table I. Clinicopathological features of the thymoma patients (n=118).

Table I.

Clinicopathological features of the thymoma patients (n=118).

Clinicopathological feature	Number
Age (years)
<40	10
40–60	51
>60	57
Sex
Male	62
Female	56
Ethnicity
White	98
Black	6
Asian	12
Not specified	2
Stage
I	35
II	60
III	15
IV	6
Not available	2

Potential therapeutic drugs identified from the CMap database

CMap analysis identified 5,000 compounds correlated with the DEGs. According to the score rankings, 769 drugs scored <-0.75 and were considered potential thymoma therapeutics. The top 10 compounds displaying the strongest negative correlation were fulvestrant, hesperetin, zidovudine, hydrocortisone, rolitetracycline, ellipticine, sirolimus, quinisocaine, oestradiol and harmine (Fig. 4; Table II).

Figure 4. Molecular structures of the top ten drugs. (A) Fulvestrant; (B) hesperetin; (C) zidovudine; (D) hydrocortisone; (E) rolitetracycline; (F) ellipticine; (G) sirolimus; (H) quinisocaine; (I) estradiol; (J) harmine.

Table II. Connectivity map results for the 10 strongest negative correlation compounds according to CMap.

Table II.

Connectivity map results for the 10 strongest negative correlation compounds according to CMap.

Drug name	Dose	Connectivity score	Up score	Down score
Fulvestrant	1 µM	−1.00	−0.25	0.38
Hesperetin	13 µM	−0.99	−0.25	0.37
Zidovudine	15 µM	−0.99	−0.25	0.37
Hydrocortisone	11 µM	−0.99	−0.24	0.38
Rolitetracycline	8 µM	−0.98	−0.23	0.38
Ellipticine	16 µM	−0.98	−0.22	0.40
Sirolimus	100 nM	−0.98	−0.23	0.39
Quinisocaine	13 µM	−0.98	−0.23	0.38
Estradiol	100 nM	−0.97	−0.24	0.37
Harmine	16 µM	−0.97	−0.21	0.40

[i] All data are for the MCF7 cell line. Up score and down score represent induced and repressed of the expression in the list respectively. The connectivity score combines the up score and the down score.

Construction of the drug-target network

The targets of the test drugs were identified from the STITCH database (20) (Fig. 5). The predicted targets of the top 10 drugs predicted for thymoma were confirmed, and the predicted scores for the interactions of DNA topoisomerase IIα (TOP2A), TOP2B and tumour protein 53 with ellipticine were 0.93, 0.86 and 0.84, respectively. Similarly, the scores for the interactions of estrogen receptor 1 (ESR1), ESR2 and cytochrome P450 family 19 subfamily A member 1 (CYP19A1) with fulvestrant were 0.99, 0.99 and 0.99, respectively. The score for the interaction of CYP11B1 and s-cortisol (hydrocortisone) was 0.90. Furthermore, the scores for the interactions between C-C motif chemokine ligand 18 and regulatory factor X-associated ankyrin-containing protein with azidothymidine (zidovudine) were 0.44 and 0.40, respectively. The score for the interaction of monoamine oxidase A with CHEMBL14355 (harmine) was 0.52 (Table III).

Figure 5. Drug-target networks of five compounds curated from the STITCH database. Pill-shaped and spheres nodes represent the compounds and proteins respectively.

Table III. Predicted targets of the drugs identified from the STITCH database.

Table III.

Predicted targets of the drugs identified from the STITCH database.

Drug/targets	Combined score
Ellipticine
TOP2A	0.93
TOP2B	0.86
TP53	0.84
CYP1A1	0.78
MDM2	0.75
CYP1B1	0.74
CYP3A4	0.73
CYP1A2	0.73
CASP7	0.70
CYP2C19	0.70
Fulvestrant
ESR1	1.00
ESR2	0.99
CYP19A1	0.99
AR	0.97
PGR	0.97
TFF1	0.95
PRL	0.94
CCND1	0.92
IGF1R	0.92
RB1	0.91
S-cortisol
CYP11B1	0.90
Azidothymidine (zidovudine)
CCL18	0.44
RFXANK	0.40
CHEMBL14355
MAOA	0.52

[i] TOP2A, DNA topoisomerase II alpha; TP53, Tumor protein P53; CYP1A1, Cytochrome P450 family 1 subfamily a member 1; MDM2, MDM2 proto-oncogene; CASP7, Caspase 7; ESR1, Estrogen receptor 1; AR, Androgen receptor; PGR, Progesterone receptor; TFF1, Trefoil factor 1; PRL, Prolactin; CCND1, Cyclin D1; IGF1R, Insulin like growth factor 1 receptor; RB1, RB transcriptional corepressor 1; CCL18, C-C motif chemokine ligand 18; RFXANK, Regulatory factor X associated ankyrin containing protein; MAOA, Monoamine oxidase A.

Molecular docking analysis of the interactions between proteins and compounds

Molecular docking analysis was performed to confirm the interactions between drugs and protein targets. The total scores, crash and polar for the fulvestrant and ESR1 interaction were 10.26, −3.72 and 2.08 respectively (Fig. 6A). The total scores, crash and polar for the tamoxifen and ESR1 interaction were 6.60, −4.02 and 0, respectively (Fig. 6B).

Figure 6. Molecular docking test for compounds and targets. (A) Interaction between fulvestrant and ESR1. (B) Interaction between tamoxifen and ESR1. ESR1, estrogen receptor 1.

Functional annotation of the drug targets

To further explore the functions of the compounds in combination with the target, functional annotation, including GO and KEGG pathway analyses, were performed. The results revealed that the targets for fulvestrant in the category BP were significantly involved in ‘mammary gland alveolus development’, ‘transcription initiation from RNA polymerase II promoter’, ‘transcription, DNA-templated’, ‘positive regulation of transcription, DNA-templated’ and ‘signal transduction’ (Fig. 7A). However, in the CC category, the targets were only significantly involved in ‘nucleoplasm’ (Fig. 7B). In the category MF, the targets were mainly enriched in ‘steroid binding’, ‘steroid hormone receptor activity’, ‘enzyme binding’, ‘RNA polymerase II transcription factor activity, ligand’ and ‘transcription factor binding’ (Fig. 7C). Finally, in the KEGG pathway analysis of fulvestrant, ‘prolactin signaling pathway’, ‘prostate cancer’, ‘glioma’, ‘melanoma’ and ‘oocyte meiosis’ were the most prominent pathways (Fig. 7D; Table IV). Target genes of ellipticine in the category BP were mainly enriched in ‘omega-hydroxylase P450 pathway’, ‘epoxygenase P450 pathway’, ‘drug metabolic process’, ‘steroid metabolic process’ and ‘monoterpenoid metabolic process’ (Fig. 8A). In the CC category, the targets were mainly enriched in ‘organelle membrane’, ‘intracellular membrane-bounded organelle’, ‘endoplasmic reticulum membrane’, ‘nucleoid’ and ‘DNA topoisomerase complex (ATP-hydrolyzing)’ (Fig. 8B). Regarding the enrichment of targets of ellipticine in the category MF, they were significantly involved in ‘enzyme binding’, ‘oxygen binding’, ‘monooxygenase activity’, ‘oxidoreductase activity’ (acting on paired donors with incorporation or reduction of molecular oxygen, reduced flavin or flavoprotein as one donor and the incorporation of one atom of oxygen) and ‘aromatase activity’ (Fig. 8C). In the KEGG pathway analysis of ellipticine, ‘chemical carcinogenesis’, ‘steroid hormone biosynthesis’, ‘metabolism of xenobiotics by cytochrome P450’, ‘linoleic acid metabolism’ and ‘tryptophan metabolism’ were the most prominent pathways (Fig. 8D; Table V).

Figure 7. Functional annotation of the targets of fulvestrant. (A-C) GO terms in the categories (A) biological process, (B) molecular function and (C) cellular component; (D) KEGG pathway analysis. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; hsa, Homo sapiens.

Figure 8. Functional annotation of the targets of ellipticine. (A-C) GO terms in the categories (A) biological process, (B) molecular function and (C) cellular component; (D) KEGG pathway analysis. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; hsa, Homo sapiens.

Table IV. Functional annotation of the targets of fulvestrant.

Table IV.

Functional annotation of the targets of fulvestrant.

ID	Description	Category/type	Gene symbol
GO:0060749	Mammary gland alveolus development	Biological_process	AR, CCND1, ESR1
GO:0006367	Transcription initiation from RNA polymerase II promoter	Biological_process	PGR, AR, ESR1, ESR2
GO:0006351	Transcription, DNA-templated	Biological_process	PGR, AR, CCND1, ESR1, RB1, ESR2
GO:0045893	Positive regulation of transcription, DNA-template	Biological_process	AR, ESR1, RB1, ESR2
GO:0007165	Signal transduction	Biological_process	PGR, IGF1R, AR, ESR1, ESR2
GO:0010629	Negative regulation of gene expression	Biological_process	PGR, ESR1, RB1
GO:0060736	Prostate gland growth	Biological_process	AR, CYP1A1
GO:0033327	Leydig cell differentiation	Biological_process	AR, CCND1
GO:0000122	Negative regulation of transcription from RNA polymerase II promoter	Biological_process	CCND1, ESR1, RB1, ESR2
GO:0007267	Cell-cell signaling	Biological_process	PGR, AR, ESR2
GO:0005654	Nucleoplasm	Cellular_component	PGR, AR, CCND1, ESR1, RB1, ESR2
GO:0016020	Membrane	Cellular_component	IGF1R, CCND1, ESR1, CYP19A1
GO:0000790	Nuclear chromatin	Cellular_component	AR, ESR1
GO:0005634	Nucleus	Cellular_component	PGR, AR, CCND1, ESR1, RB1, ESR2
GO:0005496	Steroid binding	Molecular_function	PGR, AR, ESR1, ESR2
GO:0003707	Steroid hormone receptor activity	Molecular_function	PGR, AR, ESR1, ESR2
GO:0019899	Enzyme binding	Molecular_function	PGR, AR, CCND1, ESR1, ESR2
GO:0004879	RNA polymerase II transcription factor activity, ligand	Molecular_function	AR, ESR1, ESR2
GO:0008134	Transcription factor binding	Molecular_function	AR, CCND1, ESR1, RB1
GO:0051117	ATPase binding	Molecular_function	PGR, AR, ESR1
GO:0003700	Transcription factor activity, sequence	Molecular_function	PGR, AR, ESR1, RB1, ESR2
GO:0038052	RNA polymerase II transcription factor activity, estrogen	Molecular_function	ESR1, ESR2
GO:0043565	Sequence-specific DNA binding	Molecular_function	PGR, AR, ESR1, ESR2
GO:0034056	Estrogen response element binding	Molecular_function	ESR1, ESR2
hsa04917	Prolactin signaling pathway	KEGG	CCND1, ESR1, ESR2, PRL
hsa05215	Prostate cancer	KEGG	IGF1R, AR, CCND1, RB1
hsa05214	Glioma	KEGG	IGF1R, CCND1, RB1
hsa05218	Melanoma	KEGG	IGF1R, CCND1, RB1
hsa04114	Oocyte meiosis	KEGG	PGR, IGF1R, AR
hsa05200	Pathways in cancer	KEGG	IGF1R, AR, CCND1, RB1
hsa05205	Proteoglycans in cancer	KEGG	IGF1R, CCND1, ESR1
hsa05219	Bladder cancer	KEGG	CCND1, RB1
hsa04913	Ovarian steroidogenesis	KEGG	IGF1R, CYP19A1
hsa04151	PI3K/Akt signaling pathway	KEGG	IGF1R, CCND1, PRL

[i] GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; hsa, Homo sapiens; AR, Androgen receptor; CCND1, Cyclin D1; ESR1, Estrogen receptor 1; PGR, Progesterone receptor; RB1, RB transcriptional corepressor 1; IGF1R, Insulin like growth factor 1 receptor; CYP1A1, Cytochrome P450 family 1 subfamily a member 1; PRL, Prolactin.

Table V. Functional annotation of the targets of ellipticine.

Table V.

Functional annotation of the targets of ellipticine.

ID	Description	Type	Genes
GO:0097267	Omega-hydroxylase P450 pathway	Biological_process	CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0019373	Epoxygenase P450 pathway	Biological_process	CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0017144	Drug metabolic process	Biological_process	CYP3A4, CYP1A1, CYP2C19, CYP1A2
GO:0008202	Steroid metabolic process	Biological_process	CYP3A4, CYP1B1, CYP1A1, CYP2C19
GO:0016098	Monoterpenoid metabolic process	Biological_process	CYP3A4, CYP2C19, CYP1A2
GO:0046483	Heterocycle metabolic process	Biological_process	CYP3A4, CYP2C19, CYP1A2
GO:0006805	Xenobiotic metabolic process	Biological_process	CYP3A4, CYP1B1, CYP2C19, CYP1A2
GO:0042738	Exogenous drug catabolic process	Biological_process	CYP3A4, CYP2C19, CYP1A2
GO:0016925	Protein sumoylation	Biological_process	TP53, MDM2, TOP2B, TOP2A
GO:0046677	Response to antibiotic	Biological_process	CYP1A1, TP53, MDM2
GO:0031090	Organelle membrane	Cellular_component	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0043231	Intracellular membrane-bound organelle	Cellular_component	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CASP7, CYP1A2
GO:0005789	Endoplasmic reticulum membrane	Cellular_component	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0009295	Nucleoid	Cellular_component	TOP2B, TOP2A
GO:0009330	DNA topoisomerase complex (ATP-hydrolyzing)	Cellular_component	TOP2B, TOP2A
GO:0005730	Nucleolus	Cellular_component	TP53, MDM2, TOP2B, TOP2A
GO:0016604	Nuclear body	Cellular_component	TP53, MDM2
GO:0043234	Protein complex	Cellular_component	TP53, MDM2, TOP2A
GO:0005654	Nucleoplasm	Cellular_component	CASP7, TP53, MDM2, TOP2B, TOP2A
GO:0005737	Cytoplasm	Cellular_component	CYP3A4, CASP7, TP53, MDM2, TOP2B, TOP2A
GO:0019899	Enzyme binding	Molecular_function	CYP3A4, CYP1A1, CYP2C19, TP53, MDM2, CYP1A2, TOP2B, TOP2A
GO:0019825	Oxygen binding	Molecular_function	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0004497	Monooxygenase activity	Molecular_function	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0016712	Reduced flavin or flavoprotein as one donor	Molecular_function	CYP3A4, CYP1B1, CYP1A1, CYP1A2
GO:0070330	Aromatase activity	Molecular_function	CYP3A4, CYP1B1, CYP1A1, CYP1A2
GO:0020037	Heme binding	Molecular_function	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0005506	Iron ion binding	Molecular_function	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
GO:0008395	Steroid hydroxylase activity	Molecular_function	CYP3A4, CYP1A1, CYP2C19
GO:0016491	Oxidoreductase activity	Molecular_function	CYP3A4, CYP1A1, CYP2C19, CYP1A2
GO:0016705	Incorporation or reduction of molecular oxygen	Molecular_function	CYP3A4, CYP1B1, CYP2C19
hsa05204	Chemical carcinogenesis	KEGG	CYP3A4, CYP1B1, CYP1A1, CYP2C19, CYP1A2
hsa00140	Steroid hormone biosynthesis	KEGG	CYP3A4, CYP1B1, CYP1A1, CYP1A2
hsa00980	Metabolism of xenobiotics by cytochrome P450	KEGG	CYP3A4, CYP1B1, CYP1A1, CYP1A2
hsa00591	Linoleic acid metabolism	KEGG	CYP3A4, CYP2C19, CYP1A2
hsa00380	Tryptophan metabolism	KEGG	CYP1B1, CYP1A1, CYP1A2
hsa00830	Retinol metabolism	KEGG	CYP3A4, CYP1A1, CYP1A2
hsa00982	Drug metabolism-cytochrome P450	KEGG	CYP3A4, CYP2C19, CYP1A2
hsa05206	MicroRNAs in cancer	KEGG	CYP1B1, TP53, MDM2
hsa05219	Bladder cancer	KEGG	TP53, MDM2
hsa04913	Ovarian steroidogenesis	KEGG	CYP1B1, CYP1A1

[i] GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; hsa, Homo sapiens; CYP1A1, Cytochrome P450 family 1 subfamily a member 1; TP53, Tumor protein P53; MDM2, MDM2 proto-oncogene; TOP2B, DNA topoisomerase II beta; CASP7, Caspase 7.

Screening of compounds based on ADME parameters in the TCMSP database

Hesperetin, oestradiol and harmine were searched in the TCMSP database. The OB (%), DL and HL for hesperetin were 70.31, 0.27 and 15.78, respectively; those for oestradiol were 53.56, 0.32 and 3.50, respectively; and those for harmine were 56.80, 0.13 and 5.04, respectively (Table VI).

Table VI. Absorption, distribution, metabolism and excretion parameters for three drugs from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform database.

Table VI.

Absorption, distribution, metabolism and excretion parameters for three drugs from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform database.

Drug name	Oral bioavailability (%)	Drug-likeness (0–1)	Drug half-life (h)
Hesperetin	70.31	0.27	15.78
Estradiol	53.56	0.32	3.5
Harmine	56.8	0.13	5.04

Discussion

In the present study, it was speculated that the potential therapeutic drugs for thymoma are compounds that are matched with DEGs known to be associated with the occurrence and development of tumours. First, by using the DEG data from the GEPIA database, correlations between the genes and previously known pharmaceutical compounds were revealed in CMap. Subsequently, the top 10 molecules with the lowest negative correlations were obtained and they were considered as potential therapeutic drugs for further analysis. In addition, a drug-target network was constructed to examine the potential mechanisms of action of the compounds. Molecular docking analysis was then performed to confirm the interactions between the drugs and protein targets. Furthermore, the ADME parameters were inquired to determine the potential clinical value of the compounds.

According to the GEPIA tool, 2,447 DEGs were identified in 118 thymoma patients and 339 normal samples. After functional annotation analysis, it was determined that the DEGs were enriched in ‘ribosome’, ‘oxidative phosphorylation’, ‘spliceosome’, ‘DNA replication’ and ‘cell cycle’, which indicated that the DEGs may affect cell growth and have an important role in the occurrence and development of thymoma. Then, based on the DEGs, potential therapeutic drugs were identified using the CMap database, including fulvestrant, hesperetin, zidovudine, hydrocortisone, rolitetracycline, ellipticine, sirolimus, quinisocaine, oestradiol and harmine. Then, to examine the effect of the potential drugs, a drug-target network was constructed and a molecular docking analysis was performed. Compared with the docking score of tamoxifen and ESR1, the score of fulvestrant and ESR1 was high, which indicates a high interaction between fulvestrant and ESR1. Among these compounds, tamoxifen, as an ESR1 inhibitor, remains the first-line of medication for the treatment of ESR1+ breast cancer (26). In addition, the ESR1 protein localizes to the nucleus and has been proven important in the pathological processes of several cancer types, including breast and endometrial cancers (27,28). Furthermore, functional enrichment analysis indicated that the targets of fulvestrant significantly accumulated in ‘prolactin signaling pathway’, ‘prostate cancer’, ‘glioma’ and ‘melanoma’. The results indicated that fulvestrant may combine with ESR1 to affect the treatment of thymoma. Finally, using the TCMSP database, it was determined that the compounds oestrogen, oestradiol and harmine may potentially be of high clinical value. In summary, the compounds may have an important role in the treatment of thymoma. Among these compounds, fulvestrant has been used to treat breast and prostate cancers. Regardless of endocrine tolerance or the levels of hormone receptor expression, fulvestrant significantly improves the survival of patients with non-progressive breast cancer (29–31). In addition, fulvestrant also represents a treatment option for patients with recurrent hormone receptor-positive or HER2-negative metastatic breast cancer (32).

Hesperetin is a bioflavonoid from citrus fruit. Based on experimental evidence, hesperetin possesses anti-oxidant and free radical-quenching activities. This compound also induces apoptotic cell death. In addition, hesperetin reportedly exerts an anti-cancer effect on various cancer cell lines, including breast cancer, prostate cancer, human colon adenocarcinoma and hepatocellular carcinoma cells (33–37). Based on ADME values of the compound molecules from the TCMSP database, it was revealed that hesperetin, estradiol and harmine may have good prospects. Among them, hesperetin caught our attention. The consequence of it for oral bioavailability was 70.31. A high OB (%) is frequently considered a key factor to judge the drug-like properties of compounds as therapeutic agents. Furthermore, as hesperetin is derived from citrus fruit, it may be easy to produce on a large scale. However, due to its poor water solubility, its clinical use is restricted. Numerous studies have been undertaken to improve the bioavailability of flavonoids (33). In summary, hesperetin may not only be used as a promising compound to treat thymoma in the future, but may also be commonly used in the treatment of other tumour types.

Zidovudine is an inhibitor of HIV replication that may reverse neurological dysfunction induced by HIV and ameliorate certain clinical abnormalities (38,39). Hydrocortisone is the major glucocorticoid and its synthetic counterpart is used to treat inflammation, allergy, shock and certain neoplasms (40). Rolitetracycline is a broad-spectrum antibiotic (41). Sirolimus is a potent immunosuppressant (42). Quinisocaine blocks nerve conduction when applied to nervous tissues at appropriate concentrations.

Estrogen has been consistently reported to affect the advancement of thymoma (43–45). However, the therapeutic value of fulvestrant and hesperetin for thymoma has not been previously reported. Simultaneously, the other compounds among the top 10 have also not been reported to be suitable for the treatment of cancer.

The limitations to the present study include the following: First, the DEGs should be further validated in vitro to determine their specific expression in thymoma. In addition, the exact DEGs between the different groups of patients and at different stages of the disease should be determined in vitro to further identify potential compounds, particularly in patients aged 40–60 years and in stage III/IV. Finally, further research should focus on in vitro and in vivo tests prior to the clinical application of these compounds.

In summary, the development of compounds or combinations of drugs remains a requirement in order to improve chemotherapy. The present study identified compounds that may represent novel treatments for thymoma and may reduce the range of potential drugs for treating thymoma.

Acknowledgements

Not applicable.

Funding

No funding was received.

Availability of data and materials

The datasets used and/or analysed during this study are available from the corresponding author on reasonable request.

Authors' contributions

The study was designed by GC, HY, QL and RL. XW, PL, YL, GC, HY, YH and QL were involved in the statistical analysis. HY, YH, QL and RL were involved in drafting the manuscript and critically revising it for important intellctual content. YH, QL and RL gave final approval for the version of the manuscript to be published.

Each author sufficiently participated in the work to take public responsibility for appropriate portions of the content and agreed to be accountable for all aspects of the work to ensure that questions regarding the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References