A total of 474 TC-DEGs and 64 BRGs are identified
Four hundred and seventy-four DEGs were identified (P < 0.05; |log FC| >1) in patients with thyroid carcinoma compared with the normal controls using the available numerical mRNA expression values, including 295 up-regulated genes and 179 down regulated genes (Fig. 1). The 64 BRGs ((including 12 core clock genes) were mined by literature reviewing, NCBI database retrieval and STRING, which are presented in Table 1.
Table 1
The 64 biorhythm related genes
Gene. symbol | Entrez ID | Gene. symbol | Entrez ID | Gene. symbol | Entrez ID |
ARNTL● | 406 | CHEK11 | 1111 | DEC1 | 50514 |
ARNTL2 | 56938 | CIPC | 85457 | DLAT | 1737 |
ATR | 545 | CLOCK● | 9575 | DLD | 1738 |
BATF2 | 116071 | CLSPN | 63967 | EP300 | 2033 |
BHLHE40 | 8553 | CRY1● | 1407 | EPAS1 | 2034 |
BHLHE41 | 79365 | CRY2● | 1408 | GRP | 2922 |
BLM | 641 | CSNK1D | 1453 | GSK3B | 2932 |
BTRC | 8945 | CSNK1E● | 1454 | HDAC3 | 8841 |
CDC45 | 8318 | DBP | 1628 | HIF1A | 3091 |
CEBPA | 1050 | DDIT3 | 1649 | KPTAP23-1 | 3123 |
SERPINE1 | 5054 | SIRT1 | 23411 | TIMELESS | 8914 |
NRIP1 | 8204 | PGC | 5225 | PDPR | 55066 |
PDHA1 | 5160 | PHLPP1 | 23239 | PDXK | 8566 |
PDHB | 5162 | POLA1 | 5422 | PER1● | 5187 |
PDHX | 8050 | PPM1A | 5494 | PER2● | 8864 |
PDP1 | 54704 | PPTC7 | 160760 | PER3● | 8863 |
MCM2 | 4171 | NFIL3 | 4783 | TIPIN | 54962 |
MCM3 | 4172 | NONO | 4841 | RIPK4 | 54101 |
MCM7 | 4176 | NPAS2● | 4862 | ROR | 100885779 |
MED1 | 5469 | NR1D1● | 9572 | RORA● | 6095 |
NCOR1 | 9611 | NR1D2● | 9975 | RORB | 6096 |
RORC | 6097 | | | | |
Entrez ID is the number of genes in Entrez database of NCBI. The core genes of biological clock are marked with black dot superscript. |
TC-DEGs And BRGs Chromosomal Mapping
The chromosomal locations of 474 TC-DEGs and 64 BRGs were identified using the UCSC Genome Browser Gateway resources, and MapGene2Chrom web v2 online analytical tool was used to establish the chromosomal location map (Fig. 2). The chromosomal assignments of these genes were inhomogeneous. These genes were mainly distributed on Chrom01, 02, 11, 12, 04, 03 and X, none of the associated genes were distributed on sex Y chromosomes.
GO and KEGG pathway analysis
Using DAVID, Eighteen most significantly enriched GO process terms (FDR ≤ 0.05 and Count ≥ 5 genes as criteria) were obtained for TC-DEGs and BRGs, including “circadian rhythm”, “extracellular space”, “extracellular region”, “circadian regulation of gene expression”, “extracellular exosome”, “regulation of circadian rhythm”, “oxidation-reduction process”, “proteinaceous extracellular matrix”, “retinol metabolic process”, “arachidonic acid secretion”, “regulation of acetyl-CoA biosynthetic process from pyruvate”, “nuclear hormone receptor binding”, “pyruvate metabolic process”, “pyruvate dehydrogenase complex”, “negative regulation of glucocorticoid receptor signaling pathway”, “positive regulation of gene expression”, “response to drug” and “phospholipase A2 activity”, which are presented in Fig. 3A. KEGG pathway analysis showed that 18 significantly enriched pathways (P-value ≥ 0.05 and Count ≥ 10)for TC-DEGs and BRGs, including “Purine metabolism”, “Retinol metabolism”, “Circadian rhythm”, “Metabolism of xenobiotics by cytochrome P450”, “Transcriptional misregulation in cancer”, “Glycerophospholipid metabolism”, “Arachidonic acid metabolism”, “Drug metabolism - cytochrome P450”, “Chemical carcinogenesis”, “Pancreatic secretion”, “Tyrosine metabolism”, “Ether lipid metabolism”, “Glycolysis/Gluconeogenesis”, “Vascular smooth muscle contraction”, “Thyroid hormone signaling pathway”, “Complement and coagulation cascades”, “Tryptophan metabolism” and “p53 signaling pathway”. ARNTL and BHLHE40 belong to both the TC-DEGs and BRGs, and ARNTL gene is one of the core clock genes. PER2 is one of the core clock genes, and it also exists in the signal pathway of “Transcriptional misregulation in cancer”. NCOR1 and RXRG are the key links between “Transcriptional misregulation in cancer” and “Thyroid hormone signaling pathway”. CCND1 links “Thyroid hormone signaling pathway” with “p53 signaling pathway”. SERPINE1 links “p53 signaling pathway” with “Complement and coagulation cascades”. And, PLAU is the key link between “Transcriptional misregulation in cancer” and “Complement and coagulation cascades” (Fig. 3B). Our study identified 8 key genes of thyroid carcinoma associated with biological clock.
Construction The Prognostic Metabolic Gene Signature In TCGA
The 141 metabolic genes (77 up-regulated and 64 down-regulated) were included in TCGA-THCA and GSE33630 to train the prognostic model (Table 2). Univariate Cox regression model found 11 risk genes (p < 0.01), which are ITPKA, FBP2, AOX1, PLA2G7, GPAT3, HAGHL, PFKFB1, GGCT, PGK1, PAPSS2 and SGPP2 (Fig. 4A). GGCT (Fig. 4B), HAGHL (Fig. 4C) and ITPKA (Fig. 4D) are highly expressed in tumor cells (p < 0.01). PLA2G7 (Fig. 4E), GPAT3 (Fig. 4F), PAPSS2 (Fig. 4G), AOX1 (Fig. 4H), FBP2 (Fig. 4I), SGPP2 (Fig. 4J), PGK1 (Fig. 4K) and PFKFB1 (Fig. 4L) are lower expressed in tumor cells compared with the normal controls (p < 0.01).
Table 2
The 141 metabolic genes (77 up-regulated and 64 down-regulated).
Differential expression | Gene. symbol |
77 up-regulated genes | ADCY8 | AK1 | ALOX5 | UPP1 | PC |
PLA2G2E | PDE6A | CYP26A1 | PNP | ISYNA1 |
PLA2G12B | MTHFD1L | RDH5 | PLA2G10 | HPGDS |
ACOT12 | GUCY1B1 | PNMT | DPYS | PLA2G5 |
ALDH3B2 | SPHK1 | GSTA1 | CDA | CYP26C1 |
NT5C1A | ACER2 | LPL | P4HA2 | ETNK2 |
AWAT2 | GPT | ALDH1A3 | PDE1C | NT5E |
CYP2S1 | CHST13 | CA7 | HDC | HAGHL |
PLCD3 | ME3 | CA12 | ALOX12B | TDO2 |
AOC1 | TH | ALDH3B1 | PLPP2 | CA2 |
ENTPD2 | CYP4A22 | ENTPD1 | ITPKA | PCYT1B |
GGT6 | AMY2A | ENTPD8 | DHRS3 | PLA2G2F |
GALE | MBOAT2 | PDE5A | SULT2B1 | B4GALT6 |
ALOX15B | RENBP | PLA2G2C | PDE9A | CHIT1 |
CYP1B1 | CBR3 | GGCT | CPT1C | GLS2 |
DHDH | CHPT1 | | | |
64 down-regulated genes | GPAT3 | TPH2 | SGPP2 | ACACB | TPO |
IDO1 | DDC | LIPC | FMO1 | IDO2 |
CP | FBP2 | CA4 | RDH12 | FMO2 |
GBA3 | GSTM5 | PFKFB2 | PFKFB1 | ADH1A |
ACSM5 | ACSM2A | AK7 | ENPP6 | HGD |
HAL | PAPSS2 | CYP1A1 | GYS2 | AGPAT4 |
GSTM3 | DGKI | UGT8 | PLA2G4E | PDE11A |
GATM | OTC | CD38 | ALDH1A2 | BCO1 |
ADH6 | ADH1C | NOS1 | GLDC | ACSL6 |
IMPA2 | AOX1 | PLA2G7 | PLA2G2D | OGDHL |
ENPP1 | ADH1B | LDHAL6A | LIPF | CA5A |
SORD | PIK3C2G | NMNAT3 | CTH | MIOX |
PDE10A | UGT2B11 | ALDH1A1 | ALAS2 | |
Protein-protein Interactions Of Associated Genes
The protein-protein interaction network of key genes (ARNTL, BHLHE40, PER2, NCOR1, RXRG, CCND1, SERPINE1 and PLAU) and risk metabolic genes (ITPKA, FBP2, AOX1, PLA2G7, GPAT3, HAGHL, PFKFB1, GGCT, PGK1, PAPSS2 and SGPP2) of thyroid carcinoma associated form a closely related network diagram.
There are indirect interaction among ARNTL, BHLHE40 and 10 risk genes, which including ITPKA, FBP2, AOX1, PLA2G7, GPAT3(AGPAT9), PFKFB1, GGCT, PGK1, PAPSS2 and SGPP2 (Fig. 5).
Conserved MicroRNAs Interactions Of Associated Genes
The thyroid carcinoma risk metabolic genes associated with biological clock were mined, which are ARNTL, BHLHE40, PER2, NCOR1, RXRG, CCND1, SERPINE1, PLAU, ITPKA, FBP2, AOX1, PLA2G7, GPAT3, HAGHL, PFKFB1, GGCT, PGK1, PAPSS2 and SGPP2. The conserved microRNAs of these genes were identified using Targetscan. Contrast research found that ARNTL, BHLHE40, PAPSS2, CCND1, PLAU, SERPINE1, PER2 and NCOR1 are regulated by many common conserved microRNAs, such as hsa-let-7a-5p, hsa-miR-106, hsa-miR-181, hsa-miR-449, hsa-miR-34, hsa-miR-30, hsa-miR-27, hsa-miR-23 and so on (Fig. 6).
Gene Set Enrichment Analyses
Five hundred and seven patients (with a follow-up period longer than 30 days) were divided into a high- risk and low‐risk group with an optimal cut‐off of 0.5 for risk score (Fig. 7A). Seventy-two significantly enriched KEGG pathways in TCGA-THCA were mined by GSEA. Vast majority of the enriched KEGG pathways were related to metabolism or metabolic diseases, such as the metabolism of tyrosine, sphingolipid, pyrimidine, glutathione, glycerolipid, fatty acid, insulin signaling pathway, autoimmune thyroid disease, the maturity‐onset diabetes of the young and type II diabetes mellitus. Besides, the other enriched KEGG pathways were not related to metabolism but frequently pathways in carcinoma, such as the systemic lupus erythematosus, colorectal cancer, bladder cancer, pancreatic cancer and thyroid cancer. Besides, most of the metabolism‐related pathways were enriched in the low risk group, whereas most of the non-metabolism related pathways were enriched in the high‐risk group (Fig. 7B).