Globally, colorectal carcinoma (CRC) remains the third most common cancer and the second leading cause of cancer-related death. CRC is a highly heterogeneous and malignant tumor and its progression is influenced by a constellation of factors including genetics and environment. It is known that chronic inflammation precedes colitis-associated tumor development. This localized inflammatory microenvironment can then recruit specific myeloid subsets—called myeloid-derived suppressor cells (MDSCs). These myeloid cells are phenotypically heterogeneous and pathologically activated in the tumor condition, showing a remarkable ability to both suppress T-cell responses and exhibit pro-angiogenic activity. In tumor-bearing mice, MDSCs co-express Gr-1 and CD11b markers and are divided into the following two subtypes based on Ly6c and Ly6g expression: CD11b+Ly6c^intLy6g+ polymorphonuclear MDSCs (PMN-MDSCs) and CD11b+Ly6c^hiLy6g− monocytic MDSCs (M-MDSCs). Past work has shown that blocking the infiltration of MDSCs into colonic mucosa dramatically suppresses colonic chronic inflammation and subsequent colitis-associated tumorigenesis. Meanwhile, highly augmented vascularity is often correlated with aggressive histopathological features and poor patient survival in CRC, and CD133+hematopoietic progenitor cell (HPCs) initiate growth and metastasis of colorectal cancer cells. Moreover, anti-angiogenesis therapy was approved to improve the survival time for CRC patients by FDA in 2004. Thus, identifying novel regulators for this tumor-promoting microenvironment—especially for its immunosuppressive and angiogenic microenvironment—may uncover novel therapeutic targets for CRC.

Human intelectin 1 (ITLN1, also known as Omentin 1) is a 34-kDa adipocytokine and was first isolated from a small intestine cDNA library in 2001. ITLN1 is both secreted and lipid-anchored to the cell membrane; it is highly expressed in the small intestine by intestinal goblet and adipose tissue. Recent evidence has implicated the role of ITLN1 in cancer progression; for instance, ITLN1 is over-expressed in human malignant pleural mesothelioma (MPM) and serves as a biomarker for distinguishing MPM from lung cancer. Conversely, ITLN1 acts as a tumor suppressor in neuroblastoma, ovarian cancer, hepatocellular cell carcinoma, and gastric cancer. Additionally, insufficient ITLN1 has been correlated with obesity-related colorectal carcinogenesis. However, the exact functions and clinical significance of ITLN1 in CRC remain relatively unknown. Thus, we sought to explore the inhibitory role of ITLN1 in the tumor-permissive microenvironment that exists during the first occurrence and subsequent development of colorectal carcinoma (CRC).

First, we found that when compared with their normal counterparts, ITLN1 was significantly downregulated in human CRC tissues (p<0.0001), and CRC patients with higher ITLN1 expression (n=622) had better overall survival rates than those with lower expression (n=623, p=0.01, HR=0.51), indicating that ITLN1 was frequently lost in CRC tissues and ITLN1 to be an independent prognostic predictor of CRC. Moreover, ITLN1 was highly expressed in normal colon tissue, but was gradually lost in the trajectory of inflamed colon tissue to villous adenoma until it was completely absent in CRC tissue. Orthotopic and subcutaneous tumor xenograft approaches were then used to further confirm the protective role of ITLN1 during tumor progression.

Ectopic overexpression of ITLN1 dramatically increased ITLN1 in both CRC cell lysate and CM. However, neither increased ITLN1 in CRC cells nor secretory ITLN1 had an effect on tumor cell proliferation and migration. Given these results, we speculated that ITLN1 may exert its anti-tumor effect by modulating the formation of the tumor microenvironment. As expectedly, increased ITLN1 expression in CRC cells significantly inhibited local pre-existing vessels sprouting, EPC recruitment and the infiltration of immunosuppressive myeloid-derived suppressor cells (MDSCs) into tumor tissues. Comparatively, ITLN1-derived MDSCs had a lower suppressive effect on T cell proliferation, NOS2 expression, and ROS production. Additionally, ITLN1 overexpression markedly suppressed bone marrow (BM)-derived hematopoietic progenitor cells (HPC) differentiation into MDSCs as well as NOS2 activity on MDSCs. Using H-2b+YFP+ chimerism through bone marrow transplantation, increased ITLN1 in HCT116 significantly reduced the BM-derived EPCs and MDSCs in vivo mobilization.

 Does secretory ITLN1 affect HUVEC, MDSCs or EPCs directly? Unexpectedly, rhITLN1 had no effect on HUVEC angiogenic function or the chemotaxis of EPCs, excluding the possibility that HUVEC, EPCs or MDSCs was the direct effector of secretory ITLN1. Thus, we speculated that the inhibitory effect was mediated by membrane-anchored ITLN1 downstream targets. Since HCT116-ITLN1-CM possessed potent suppression of MDSCs infiltration and vascularization, we next performed a cytokine microarray assay to explore its possible targets. And IL-17D and CXCL2 (MIP2) were identified as potential target molecules. Mechanistically, results indicated ITLN1 inhibited tumor-derived IL-17D and CXCL2 (MIP2) through the KEAP1/Nrf2/ROS/IL-17D and p65 NF-ĸB/CXCL2 signaling cascades dependent on PI3K/AKT/GSK3ß. This effect was reversed by the PI3K selective inhibitor LY294002. Collectively, ITLN1 synergistically suppressed IL-17D and CXCL2-mediated tumor vascularization, bone marrow derived EPC recruitment, as well as MDSCs generation and trafficking.

 Taken together, these findings implicate that 1) the BM-derived EPCs, MDSCs and local preexisting ECs network are essential cellular targets of IL-17D and CXCL2 during tumor vascularization and immunosuppression in CRC; 2) ITLN1-IL-17D inhibits the generation and activation of BM-derived MDSCs by suppressing endothelial-derived IL-6, IL-8, and GM-CSF production; 3) tumor-derived CXCL2 along with host-derived CXCL2 modulated BM-derived MDSCs trafficking and recruitment; and 4) the inhibitory effect of ITLN1 on CRC progression relies on a mechanism involving blocking the formation of the tumor-permissive microenvironment. Thus, either the ITLN1-PI3K/AKT-IL-17D or CXCL2 axes might serve as an exploitable therapeutic strategy for CRC patients.