Identification of potential core genes in colorectal carcinoma and key genes in colorectal cancer liver metastasis using bioinformatics analysis

In this study, we identified genes and mechanisms underlying liver metastases in CRC patients. We identified 35 DEGs between metastatic tumor and primary tumor tissues, in which 30 were upregulated whereas 5 were downregulated. Another 142 DEGs between primary tumor and normal colon tissues were also identified. Of these, 35 were upregulated whereas 107 were downregulated.

Set one differently expressed SERPINC1, FGA, F2, HPX and PLG genes regulated the expression of coagulation factor. SERPINC1, also known as ATIII and THPH7, is a member of the serpin C family. SerpinC1 gene encodes antithrombin III (ATIII), a serine protease inhibitor⁹. SERPINC1 inhibits thrombin-induced tumor growth and angiogenesis, impairing proliferation and migration of cancer cells¹⁰. Fibrinogen alpha chain (FGA) also known as Fib2 participates in hemostasis and wound healing by inducing conversion of fibrinogen to fibrin. FGA protein consists of 2 sub-units, each composed of Aα, Bβ and γ 3 polypeptides encoded by FGA, FGB and FGG genes, respectively¹¹. Research shows that CRC cells express FGA. Intriguingly, high levels of phosphorylated FGA have been observed in CRC tissues¹².

Coagulation factor II, thrombin factor 2(F2) is also known as RPRGL2, THPH1 or PT. Proteolytic cleaving of coagulation factor II generates activated serine protease thrombin. Over-production of thrombin not only increases blood coagulation, but also promotes growth and metastasis of tumors. Accordingly, thrombin and factors contributing to thrombin production are treatment targets for cancer and cancer-associated thrombosis¹³. Hemopexin (HPX), also known as HX encodes a plasma glycoprotein that binds heme with high affinity. HPX also scavenges labile heme. Studies have demonstrated an inverse association between plasma labile heme and Hx in prostate cancer (PCa) patients. Furthermore, under-expression of Hx in PCa tissues correlates with poorly differentiation of the tumor tissues and early relapse¹⁴. Also, hemopexin promotes invasion of the pancreatic cancer cells. Our findings suggest that hemopexin promotes lymph node metastasis of CRC cells, thus it is a potential therapeutic target for CRC¹⁵.

PLG encodes a serine protease plasminogen that circulates in blood plasma as an inactive zymogen. It is however converted to the active protease plasmin by several plasminogen activators. Herein, PLG and FGG were over-expressed in NSCLC tissues, relative to paired adjacent normal tissues (P = 0.000). A similar trend was observed in urine of NSCLC patients vs healthy individuals (P = 0.000). PLG and FGG proteins are therefore potential markers for NSCLC diagnosis¹⁶. Particular, PLG expression is associated with favorable prognostic in patients with advanced grade III/IV FIGO) (advanced ovarian cancer)¹⁷. Increasing evidence shows that thrombin regulates every step of cancer metastasis: (1) invasion, detachment from primary tumor, migration; (2) entry in to circulation; (3) survival in the vasculature; (4) extravasation and (5) implantation in host organs. Recent molecular data shows that it participates in transendothelial migration, platelet/tumor cell interactions, angiogenesis and other processes. While, thrombin-antithrombin complex (TAT) and vascular endothelial growth factor (VEGF) are over-expressed in CRC patients. Surprisingly, these proteins tend to increase in the course of chemotherapy. SERPINC1, ITIH2, F2 and PLG genes are associated with serine protease^18,19. Inter-alpha-trypsin inhibitors (ITI) are a family of structurally related plasma serine protease inhibitors that participate in extracellular matrix stabilization and modulation of tumor metastasis. ITIH2 is usually expressed in normal brain tissue and low-grade CNS tumors, but its expression is lost in high-grade CNS tumors including gliobastoma multiforme, further underlining its potential role as an anti-invasive protein²⁰.

AHSG, α2-Heremans-Schmid glycol, also referred to as protein Fetuin-A, is almost entirely secreted and expressed in the liver. Expression pattern of AHSG accurately predicts the prognosis of liver cirrhosis and hepatocellular cancer²¹. The positive association between AHSG expression and the risk of developing CRC stems from the effect of AHSG on obesity and insulin resistance²². Several studies have demonstrated the strong relationship between insulin resistance as well as hyperinsulinemia and risk of developing CRC²³. Fetuin‐has been implicated in adhesion of tumor cells, which promotes metastases²⁴. Ceruloplasmin (CP) also known as CP-2 is a metalloprotein that binds most of the copper in plasma and participates in peroxidation of Fe (II) transferrin to Fe (III). Given that the expression of Ceruloplasmin is associated with advanced T stage and perineural invasion, it is a potential prognostic marker for bile duct cancer. Under-expression of CP is associated with for poor prognosis of Adrenal cortical carcinoma (ACC). Meanwhile, Ceruloplasmin is a promising (prognostic) marker for pancreatic ductal adenocarcinoma (PDAC) in patients negative for CA19-9^25,26. Apolipoprotein A (APOA2) also known as apoAII, Apo-AII or ApoA-II, encodes apolipoprotein (apo-) A-II, the second most abundant high density lipoprotein particles. Compared to CA19-9 alone, a combination of CA19-9 and ApoA2-ATQ/AT detects pancreatic cancer up to 18 months earlier than the traditional methods. As such, it can be used for initial pancreatic cancer diagnosis prior imaging²⁷. Expression of ApoA1, ApoA2 and ApoA4 is under-expressed, whereas those of tumor antigens (e.g. carcinoembriogenic antigen) and inflammatory markers (e.g., C-reactive protein) are up-regulated in CRC patients, relative to healthy individuals²⁸. Histidine rich glycoprotein (HRG) also known as HPRG, HRGP and THPH11, is found in plasma and platelets and contains two cystatin-like domains. Under-expression of HRG is an independent poor prognostic for pancreatic ductal adenocarcinoma (PDAC)²⁹. Modulated HRG expression in patients with advanced LC is associated with advanced disease stage and hypofibrinolysis³⁰.

In set two DEGs, members of TIMPs family including TIMP1 are natural inhibitors of matrix metalloproteinases (MMPs) including MMP1. TIMP metallopeptidase inhibitor 1(TIMP1) also known as EPA, HCl and CLGI gene, belongs to the TIMP family. Recent evidence demonstrates that even though senescent cells inhibit tumorigenesis in the initial stages of cancer development, they can promote tumor progression in the latter stages. TIMP1 deletion allows senescence cells to promote metastasis, whereas inhibition of senescent cells using a senolytic BCL-2 inhibitor impairs metastasis. Given that TIMP1 promotes tumorigenesis and metastasis of human colon cancer, it is a potential prognostic biomarker for the cancer. Presence of TIMP1 mRNA in platelet independently predicts the presence of colorectal cancer. Transportation of the TIMP1 RNAs to colorectal cancer cells by platelets promote development of colorectal cancer^31,32. Matrix metallopeptidase 1(MMP1)/CLG/CLGN gene encodes a member of the peptidase M10 family of matrix metalloproteinases (MMPs). Expression of MMP1 in ovarian cancer tissues correlates with poor prognosis. Moreover, EVs with MMP1 mRNA in cancer ascites induces apoptosis of mesothelial cells³³. MMP-8 and TIMP-1 in serum, but not MMP-9 are associated with poor prognosis of CRC. In CRC patients without systemic inflammation, expression of MMP-8 and TIMP-1 is associated with poor prognosis³⁴. TIMP-1 and MMP-7 are highly sensitive and accurate diagnostic biomarkers for metastatic colorectal cancer. The levels of TIMP-1 and MMP-7 levels strongly and positively correlate with the severity and prognosis of liver disease³⁵.

CXCL1 and CXCL12 are members of the CXC subfamily of chemokines. C-X-C motif chemokine ligand 1(CXCL1) also known as FSP, GROa or MGSA gene, encodes a member of the CXC subfamily of chemokines. Colorectal carcinoma cells secrete VEGFA, which stimulates tumor-associated macrophages to produce CXCL1 in the primary tumor. High levels of CXCL1 in premetastatic liver tissue recruits CXCR2-positive myeloid-derived suppressor cells (MDSC) to form a premetastatic niche that ultimately promotes liver metastases³⁶. Disrupting the CXCL1/8-CXCR2 axis can suppress SMAD4-negative colorectal cancer³⁷. C-X-C motif chemokine ligand 12 (CXCL12)/IRH/PBSF/SDF1 is an antimicrobial gene encodes a stromal cell-derived alpha chemokine member of the intercrine family. CXCL12/CXCR4 promotes invasion of ovarian cancer cells by suppressing ARHGAP10 expression via the VEGF/VEGFR2 signaling pathway³⁸.

Collagen type I alpha 2 chain (COL1A2)/OI4/EDSCV/EDSARTH2 gene encodes the pro-alpha 2 chain of triple helix type I collagen. COL1A 2 suppresses CRC, thus it is a potential therapeutic option for CRC³⁹. PI3K, Akt and p-Akt proteins are over-expressed in gastric cancer tissues, relative to adjacent normal tissues. Comparable trend has been observed for COL1A2, COL6A3 and THBS2 mRNA expression in gastric cancer tissues⁴⁰. Aurora kinase A (AURKA)/BTAK/STK6/PPP1R47 gene product is a cell cycle-regulated kinase that participates in microtubule formation and/or stabilization of the spindle pole during chromosome segregation. In gastrointestinal cancer cell lines with activated KRAS, AURKA phosphorylates RPS6KB1, promoting proliferation, survival and growth of xenograft tumors in mice. Inhibiting AURKA slows down the growth of gastrointestinal tumors by activating KRAS⁴¹. Moreover, overexpression of AURKA enhances Oxaliplatin-mediated killing of colon cancer cells. Conversely, AURKA knockdown significantly weakened the chemosensitivity of colon cancer cells to Oxaliplatin⁴². Ubiquitin conjugating enzyme E2 C (UBE2C)/UBCH10/dJ447F3.2 modifies cellular abnormal or short-lived proteins destined for degradation. UBE2C not only suppresses gastric cancer colony formation, but also inhibits biosynthesis of gastric cancer DNA⁴³. Over-expressed of UBE2C in rectal carcinoma modulates miR-381 expression, promoting proliferation invasion of rectal carcinoma cells but inhibits apoptosis of cells⁴⁴. DNA topoisomerase II alpha (TOP2A)/TOP2/TP2A gene encodes DNA topoisomerase enzyme that disentangle the topological problems of dsDNA during replication and mRNA transcription. DNA topoisomerases, particularly type IIA topoisomerases, are potential therapeutic targets for numerous anticancer therapies⁴⁵. TOP2A is an oncogene for colon cancer, and even after development, TOP2A is over-expressed in the cancer cells⁴⁶. Aldehyde dehydrogenase 1 is a member of aldehyde dehydrogenase family encoded by HGNC (ALDH1A1)/HEL-9/PUMB1/RALDH1 gene. Herein, we found ALDH1A isoforms in multidrug resistance colorectal cancer tissues. Besides, ALDH1A is a potential marker for cancer stem cell. This revelation has opened a new frontier in to treatment of colorectal adenocarcinoma and other tumors⁴⁷. ALDH1A1 expression is associated with poor differentiation and prognosis of primary tumors, and shorter overall survival of respective patients. Over-expression of ALDH1A1 in tumors is also associated with therapy resistance and liver metastases⁴⁸. The beta subunit of activated cAMP protein kinase encoded by HGNC (PRKACB)/CAFD2/PKACB/PKA C-beta gene is a member of the serine/threonine protein kinase family. Downregulated expression of PRKACB is associated with shorter OS of CRC patients⁴⁹.

Glucagon is encoded by HGNC (GCG)/GLP1/GLP2/GRPP/GLP-1 gene. The protein is composed of four distinct peptides. Glucagon is produced by the pancreas, and antagonizes the glucose-lowering action of insulin by stimulating glycogenolysis and gluconeogenesis. It is a ligand for a specific G-protein linked receptor whose signaling pathway regulates cell proliferation. Two of the four glycogen peptides are secreted in the gut of endocrine cells, and promotes nutrient absorption through various mechanisms. The fourth glycogen sub-unit is an active enteroglucagon that compares to glicentin⁵⁰.

Overall, under-expression of ITIH2, COL1A2, TIMP1 and AURKA is associated with longer overall survival of CRC patients; whereas COL1A2 expression correlates with longer CRC free survival. In set 1, over-expression of ITIH2 was associated with poor overall survival. According to the Human Protein Atlas, ITIH2 is enriched in normal liver tissues and liver cancer tissues. However, pathological section (Human Protein Atlas) shows that the gene(ITIH2) not detected in tumor cells in colon cancer. The reason may be a insufficient number of pathological data available. Therefore, these observations are worth further in-depth study and more clinical pathological data are required to observe. In set 2, over-expression of the 3 hub genes (TIMP1, COL1A2 and AURKA) was associated with poor overall survival. In order to more accurately analyze the genes involved in colon cancer, we performed disease-free survival analysis on 10 genes. over-expression of COL1A2 was associated with poor disease-free survival of CRC patients. It further illustrates the close relationship between the gene (COL1A2) and the incidence of colon cancer. At the same time, the pathological section (Human Protein Atlas) shows that the gene is significantly expressed in colon cancer.