International Journal of Molecular Sciences Article Glycosyltransferase B4GALNT2 as a Predictor of Good Prognosis in Colon Cancer: Lessons from Databases Michela Pucci, Nadia Malagolini and Fabio Dall’Olio * Department of Experimental, Diagnostic and Specialty Medicine (DIMES), General Pathology Building, University of Bologna, Via San Giacomo 14, 40126 Bologna, Italy; [email protected] (M.P.); [email protected] (N.M.) * Correspondence: [email protected]; Tel.: +39-051-2094704 Abstract: Background: glycosyltransferase B4GALNT2 and its cognate carbohydrate antigen Sda are highly expressed in normal colon but strongly downregulated in colorectal carcinoma (CRC). We previously showed that CRC patients expressing higher B4GALNT2 mRNA levels displayed longer survival. Forced B4GALNT2 expression reduced the malignancy and stemness of colon cancer cells. Methods: Kaplan–Meier survival curves were determined in “The Cancer Genome Atlas” (TCGA) COAD cohort for several glycosyltransferases, oncogenes, and tumor suppressor genes. Whole expression data of coding genes as well as miRNA and methylation data for B4GALNT2 were downloaded from TCGA. Results: the prognostic potential of B4GALNT2 was the best among the glycosyltransferases tested and better than that of many oncogenes and tumor suppressor genes; high B4GALNT2 expression was associated with a lower malignancy gene expression profile; differential methylation of an intronic B4GALNT2 gene position and miR-204-5p expression play major roles in B4GALNT2 regulation. Conclusions: high B4GALNT2 expression is a strong predictor of good prognosis in CRC as a part of a wider molecular signature that includes ZG16, ITLN1, BEST2, and Citation: Pucci, M.; Malagolini, N.; GUCA2B. Differential DNA methylation and miRNA expression contribute to regulating B4GALNT2 Dall’Olio, F. Glycosyltransferase expression during colorectal carcinogenesis. B4GALNT2 as a Predictor of Good Prognosis in Colon Cancer: Lessons Keywords: glycosylation; colon cancer; Sda antigen; glycosyltransferases; epigenetic regulation from Databases. Int. J. Mol. Sci. 2021, 22, 4331. https://doi.org/10.3390/ ijms22094331 1. Introduction Academic Editor: Takeshi Sato Glycosylation plays a crucial role in a variety of biological processes, including intracel- lular transport, cell adhesion, cell growth, and apoptotic death [1]. In cancer, glycosylation Received: 30 March 2021 changes consist of up- or downregulation of numerous carbohydrate structures affecting tu- Accepted: 19 April 2021 a Published: 21 April 2021 mor invasion and progression [2–4]. The Sd antigen is a carbohydrate structure expressed on erythrocytes and a few other tissues in the vast majority of individuals. The minimal a α Publisher’s Note: MDPI stays neutral structure of the Sd antigen consists of a 2,3-sialylated galactose to which a GalNAc β with regard to jurisdictional claims in residue is 1,4-linked (Figure1)[ 5]. This epitope can be carried out by various underlying published maps and institutional affil- sugar chains, which are indicated as R1–R5 in Figure1. The enzyme responsible for the a iations. last step of Sd biosynthesis is β1,4N-acetylgalactosaminyltransferase 2 (B4GALNT2), first identified in our laboratory [6] and successively cloned in our and other labs [7–9]. The few Sda-negative individuals display missense mutations in the C-terminal portion of the B4GALNT2 gene [10]. The expression of B4GALNT2 is very high in normal colonic mucosa but undergoes a dramatic downregulation in colorectal carcinoma (CRC) [11,12]. However, Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. the level of expression in cancer samples is highly variable. In many specimens, it is unde- This article is an open access article tectable, while it is relatively conserved in other cases, predicting a longer survival [13]. In distributed under the terms and mice, B4galnt2 expression influences microbiota composition [14]. The B4GALNT2 gene is conditions of the Creative Commons comprised of 11 coding exons. There are two different alternative first exons, both contain- Attribution (CC BY) license (https:// ing a translational start codon. The polypeptide encoded by the transcript including Exon creativecommons.org/licenses/by/ 1S (short form) contains a cytoplasmic tail of conventional length, whereas the one encoded 4.0/). by the transcript with Exon 1L (long form) possesses an exceptionally long cytoplasmic tail. Int. J. Mol. Sci. 2021, 22, 4331. https://doi.org/10.3390/ijms22094331 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 2 of 15 Int. J. Mol. Sci. 2021, 22, 4331 coded by the transcript including Exon 1S (short form) contains a cytoplasmic tail of2 con- of 14 ventional length, whereas the one encoded by the transcript with Exon 1L (long form) possesses an exceptionally long cytoplasmic tail. The short form exhibits conventional Golgi localization, while the long form is localized in post-Golgi compartments as well as The short form exhibits conventional Golgi localization, while the long form is localized in on the plasma membrane [15]. post-Golgi compartments as well as on the plasma membrane [15]. Type 1 Type 2 O-linked O-linked O-linked lactosaminic lactosaminic Core 1 Core 2 Core 3 R R Ser/Thr Ser/Thr Ser/Thr β6 β3 R1 β4 R2 β3 R3 β3 β4 R4 β3 R5 β4 α2,3 sialyltransferases R1-5 α2,3 N-acetylgalactosamine (GalNAc) B4GALNT2 N-acetylglucosamine (GlcNAc) R1-5 Galactose (Gal) β1,4 Sialic acid (Sia) α2,3 Sda antigen a Figure 1. Schematic representation of the structure and biosynthesis of the Sd antigen. Several α α Figurecarbohydrate 1. Schematic structures representation (R1–R5) terminating of the structure with and galactose biosynthesis are 2,3 of the sialylated Sda antigen. by different Several 2,3 carbohydratesialyltransferases. structures The resulting (R1–R5) sialylatedterminating structure with galactose is a substrate are α2,3 for sialylated B4GALNT2, by different which generates α2,3 a sialyltransferases.the Sd antigen. The resulting sialylated structure is a substrate for B4GALNT2, which generates the Sda antigen. The molecular bases of B4GALNT2 downregulation in colon cancer tissues and cell linesThe have molecular not been bases fully of elucidated, B4GALNT2 although downregulation methylation in colon appears cancer to playstissues a and role. cell In linesfact, have treatment not been of different fully elucidated, colon cancer although cell linesmethylation with the appears methylation to plays inhibitor a role. In 5-aza- fact, 0 treatment2 -deoxycytidine of different resulted colon in cancer a partial cell lines activation with the of B4GALNT2methylationexpression inhibitor 5- [aza16,-172′].-deox- The x x ycytidinecarbohydrate resulted antigen in a sialylpartial Lewis activation(sLe of), B4GALNT2 a well-known expression ligand for[16,17 the]. cellThe adhesioncarbohy- dratemolecules antigen of thesialyl selectin Lewis family,x (sLex is), a ectopically well-known expressed ligand for by manythe cell cancers adhesion and molecules is associated of x α thewith selectin malignancy family, [2 is]. ec Thetopically sLe antigen expressed is formed by many by acancers fucose and residue is associated1,3-linked with to ma- the α lignancyGlcNAc residue[2]. The of sLe a x 2,3antigen sialylated is formed type 2by chain, a fucose a structure residue on α1,3 which-linked B4GALNT2 to the GlcNAc acts to a B4GALNT2 residuesynthesize of a the α2,3 Sd sialylatedantigen [type5]. In 2 cell chain lines,, a structure the forced on expression which B4GALNT2 of acts resultedto synthe- in Sda expression and sLex inhibition, demonstrating that biosynthesis of the two antigens size the Sda antigen [5]. In cell lines, the forced expression of B4GALNT2 resulted in Sda is mutually exclusive [12,18]. These B4GALNT2-expressing cell lines displayed reduced expression and sLex inhibition, demonstrating that biosynthesis of the two antigens is mu- metastatic ability [18,19], which was attributed to sLex inhibition rather than to de novo tually exclusive [12,18]. These B4GALNT2-expressing cell lines displayed reduced meta- Sda expression. However, our recent work [13,20] has shown that B4GALNT2 expres- static ability [18,19], which was attributed to sLex inhibition rather than to de novo Sda ex- sion decreases malignancy and stemness in different models of colon cancer cell lines, pression. However, our recent work [13,20] has shown that B4GALNT2 expression de- independently of sLex inhibition. The COADREAD (colon and rectal adenocarcinoma) creases malignancy and stemness in different models of colon cancer cell lines, inde- cohort of “The Cancer Genome Atlas” (TCGA) contains data from 626 cases. Matched pendently of sLex inhibition. The COADREAD (colon and rectal adenocarcinoma) cohort tumor–normal tissue samples have been molecularly characterized to identify alterations providing insights into the biology of CRC. Molecular data are derived from multiple types of analysis including gene expression, whole genome sequences, DNA methylation, and miRNA expression. In addition, numerous metadata, including clinical information about Int. J. Mol. Sci. 2021, 22, 4331 3 of 14 participants are also available. The purpose of the present study is to gain insights into the mechanisms of regulation of B4GALNT2 expression and its association with malignancy through an in-depth analysis of data available in TCGA. 2. Results 2.1. The Level of Expression of Only a Few Oncogenes and Tumor Suppressor Genes Is Associated with Patient