Aberrant O‐glycosylation contributes to tumorigenesis in human colorectal cancer

Abstract Aberrant O‐glycosylation is frequently observed in colorectal cancer (CRC) patients, but it is unclear if it contributes intrinsically to tumorigenesis. Here, we investigated the biological consequences of aberrant O‐glycosylation in CRC. We first detected the expression profile of Tn antigen in a serial of human CRC tissues and then explored the genetic and biosynthetic mechanisms. Moreover, we used a human CRC cell line (LS174T), which express Tn antigen, to assess whether aberrant O‐glycosylation can directly promote oncogenic properties. It showed that Tn antigen was detected in around 86% human primary and metastatic CRC tissues. Bio‐functional investigations showed that T‐synthase and Cosmc were both impaired in cancer tissues. A further analysis detected an occurrence of hypermethylation of Cosmc gene, which possibly caused its loss‐of‐function and a consequent inactive T‐synthase. Transfection of LS174T cells with WT Cosmc restored mature O‐glycosylation, which subsequently down‐regulated cancer cell proliferation, migration and apoptotic‐resistant ability. Significantly, the expression of MUC2, a heavily O‐glycosylated glycoprotein that plays an essential role in intestinal function, was uniformly reduced in human CRC tissues as well as in LS174T cells. These data suggest that aberrant O‐glycosylation contributes to the development of CRC through direct induction of oncogenic properties in cancer cells.

Mucin-type O-glycosylation, which is the most common posttranslational modification of many membrane bound and secreted glycoproteins, occurs in the Golgi apparatus and is regulated by a series of glycosyltransferases. 12 Normally, the biosynthesis of O-glycans is initiated by polypeptide α-N-acetylgalactosaminyltransferases (ppGal-NAc-Ts), a family of enzymes which add GalNAc to Ser/Thr residues to form the Tn antigen. 13 Tn antigen is the only precursor for all mucintype O-glycans. For further elongation, T-synthase (core 1 β1,3-galactosyltransferase) is the key enzyme during the O-glycosylation process. 14 Interestingly, the expression and activity of T-synthase require an endoplasmic reticulum(ER)-resident molecular chaperone named Cosmc. [15][16][17] Knockout of either T-synthase or Cosmc in mice causes Tn expression in vivo. [18][19][20] Besides, in gastrointestinal tract, Tn antigen can also be elongated into core 3 O-glycans by specific glycosyltransferase-C3GnT. 21

| Immunohistochemical staining of Tn antigen
The formalin-fixed, paraffin-embedded tissues (n = 186) were cut into 5 μm sections. The sections were firstly stained with haematoxylin and eosin (H&E) using a standard protocol. For immunohistochemistry of Tn antigen, deparaffinized sections were boiled for 20 minutes in 0.01 mol/L citrate buffer at pH 6.0 for epitope retrieval. The sections were blocked using 0.3% H 2 O 2 and 5% BSA. Then, sections were incubated overnight at 4°C with a specific anti-Tn IgM monoclonal antibody (5 μg/mL, CA3638, clone 12A8-C7-F5, kindly provided by Dr. Tongzhong Ju, Emory University, Atlanta, Georgia, USA) 13,20,26 followed by horseradish peroxidase-conjugated antimouse IgM antibody (Abcam, ab97230) for 1 hour at room temperature. Finally, the sections were developed with DAB reagent (ZSGB-BIO, China) and counterstained with haematoxylin.  Reverse 5′-TCCTAACCAAACTATTCTAACTACAAAC-3′, T-synthase-F   5′-GGTGATTTTTGTTTTTTTGGGTAGT-3′, T-synthase-R 5′-TCAAAA TCTTAAAACTAATACATAACCTT-3′. An additional T7 promoter tag was added to each reverse primer for in vivo transcription, and a 10mer tag was added to the forward primer to balance the PCR primer length. PCR products were treated with Shrimp alkaline phosphatase (SAP) to dephosphorylate the unincorporated dNTPs. At the same time, the RNaseA was added to cleave the transcripts (T-cleavage).

| Exome sequencing
The DNA methylation was quantitatively analysed by the MassAR-RAY platform (SEQUENOM).

| T-synthase activity assay
The assessment of the T-synthase activity was conducted using a fluorescence-based assay. 27

| Lentiviral-mediated Cosmc transfection
The GV367-EGFP-Cosmc vector encoding Cosmc and the control vector GV367-EGFP-control were purchased from Shanghai Genechem Co. LTD (China). The Cosmc expression vector and the control were transfected into LS174T cells with polybrene (Genechem, Shanghai, China). The transfected cells were cultured for 2 days and then selected with puromycin.

| Cell proliferation, migration and apoptosis
The cell proliferation was measured by the Cell Count Kit-8 (CCK-8).
The 1 × 10 4 cells were seeded into 96-well plates and incubated at 37°C. Before assay, medium was removed, 10 μL of the CCK8-solution was added to each well of the plate and incubated at 37°C for 2 hours. The optical densities (OD) were determined at 450 nm by spectrophotometer.
Cell migration was determined using wound healing assay and transwell assay. Briefly, 2 × 10 6 cells were seeded into a 6-well Finally, all sections were counterstained with DAPI and mounted with anti-fade mountants. The inForm software (PerkinElmer) was used for analysing the multispectral images.

| Statistical analysis
All data were analysed with the SPSS 19.0 statistical software (SPSS, Chicago, IL, USA) and GraphPad Prism 6.0 (GraphPad software, La Jolla, CA, USA). All real-time PCR reactions were repeated three times. The relative quantity of the gene expression was normalized to GAPDH using 2 −▵▵ct methods. The methylation level was analysed by EpiTYPE software(SEQUENOM). Continuous data were presented as the mean ± SEM and analysed with Student's t-test.
Categorical data were analysed by chi-square test. P < .05 were considered statistically significant. where all non-tumour tissues were Tn negative (Figure 1). Tn antigen was most often stained in the apical cell membranes, mucin droplet and the cytoplasm of the cancer tissues ( Figure S1), suggesting a fre-

| Exploration of the mechanisms leading to Tn antigen expression in human CRC
We further investigated the mechanisms responsible for Tn antigen expression detected in patients with CRC. It is known that mature O-glycosylation depends specifically on T-synthase activity and its single chaperone-Cosmc. 31 Besides, in colonic tissues, defective core 3 O-glycosylation could also lead to the exposure of Tn antigen. 21,32 Here we found that the expression of C3GnT that controls core 3 O-glycosylation in cancer tissues expressing Tn antigen (Tn-positive) was comparable to that in Tn-negative cancerous tissues, thereby suggesting that there were only possible defects in T-synthase and/ or Cosmc (Figure 2A). As expected, we showed that the expression and activity of T-synthase were much lower in Tn-positive CRC tissues relative to Tn-negative cancerous tissues (Figure 2A,B). Concomitantly, the expression of Cosmc, the key chaperone required for T-synthase activity and expression, was also found to be decreased in these Tn-positive cancer tissues (Figure 2A). In addition, ppGalNAc-Ts catalyze the initial O-glycosylation step, 13

Human colorectal carcinoma LS174T cells harbour mutations in
Cosmc gene and have expression of truncated O-glycans such as Tn antigen. 23    translational modification of many proteins and lipids, and subsequently alters their functions involved with oncogenic properties. 13 Of these O-glycosylated proteins, the secretory mucin MUC2 is stored in bulky apical granules of the goblet cells and is a most important factor maintaining intestinal homeostasis. 38 As the predominant O-glycoprotein, MUC2 is heavily modified by O-glycosylation and is generally considered to be essential for epithelial protection. MUC2lacking mice spontaneously developed colitis and colorectal cancer. 53 Therefore, we assume that defective O-glycosylation may affect the expression and/or function of MUC2 that is required for suppressing intestinal cancer. Our results showed that MUC2 expression was much reduced in Tn-positive human colonic cancer tissues relative to normal tissues, suggesting that MUC2 may require mature O-glycosylation for its proper expression and stability. However, the alterations in MUC2 may also be due to the loss of epithelia as a result of severe inflammation/ulceration. To specifically address this issue, we com-

CONF LICT OF I NTEREST
The authors declare that they have no conflict of interest.