Altered glycosyltransferases in colorectal cancer

KEYWORDS:

• Glycome
• sialic acid
• polylactosamine
• O-glycosylation
• mucin
• colon
• cancer
• gastrointestinal
• sequencing
• mutation

The cellular glycome, comprised of the entire spectrum of glycan structures, is immensely complex and has gained increasing interest due to its ability to influence and regulate key cellular processes like cell adhesion, migration, cell–cell recognition, and immune surveillance [1]. Glycosylation, the addition of defined sugars to proteins and lipids, occurs in the endoplasmic reticulum or the Golgi apparatus [2] and is catalyzed by the collective action of a family of enzymes (glycosyltransferases) that transfer sugar residues from nucleotide-sugar donors to lipid’ protein or sugar acceptors, and enzymes (glycosidases) that catalyze the hydrolysis of glycosidic linkages [3,4]. The process is highly substrate-, site-, and context-specific, and leads to the formation of highly complex glycan structures that can be distinguished as N-linked or O-linked based on addition of the sugars onto side chains of Asn or Ser/Thr residues, respectively [4,5]. A multitude of glycosyltransferase enzymes catalyze the different glycan biosynthetic pathways, resulting in the formation of a glycocode that bears key information for critical cellular functions involved in biologic homeostasis [6]. Of note, a single glycosyltransferase can often be involved in a variety of glycosylation sub-network pathways [1].

Aberrations in N- and/or O-linked glycosylation can perturb the cellular glycan structures and severely hinder the normal functioning of proteins and lipids involved in critical biological processes [4,5]. Indeed, altered N- and/or O-linked glycosylation is a hallmark of several cancers, including colorectal cancer (CRC), that promote tumor progression and metastasis [4,5]. While aberrant glycosylation patterns have led to the identification of several cancer-associated antigens many of which are glycoconjugates [4], the molecular basis for aberrant glycosylation in cancer and the underlying mechanisms by which they contribute to tumor progression, however, remain poorly understood.

Some of the causative factors underlying altered protein glycosylation in cancers include aberrant expression levels of glycosyltransferases, altered localization of glycosyltransferases within subcellular compartments, bioavailability of protein substrates, and/or the absence of necessary cofactors that aid in normal glyco-biogenesis [1,4,5]. In particular, our previous studies in familial and sporadic CRCs identified functionally inactivating germline and somatic mutations in the gene encoding for GALNT12 (Polypeptide N-acetylgalactosaminyltransferase 12) [7], suggesting that aberrant glycosylation in cancers could stem from genetic defects in the glycosyltransferase genes. These findings prompted us to undertake an in-depth characterization of CRC-associated mutational alterations in genes encoding glycosylation enzymes using targeted next-generation sequencing [8]. Of the 430 glycosylation-associated genes assessed, 36 genes exhibited mutational alterations (missense, nonsense, splice site, and frameshift mutations) in CRCs, with 12 genes displaying significantly higher mutation rates than the expected background rate in CRCs [8]. Interestingly, further enrichment analysis of the 12 genes showed that three genes, B3GNT2 (β-1,3-N Acetylglucosaminyltransferase 2), B4GALT2 (β-1,4-Galactosyltransferase 2), and ST6GALNAC2 (α-N-Acetylgalactosaminide α-2,6-Sialyltransferase 2), belonged to polylactosamine and N- and O-glycosylation pathway [8]. While B3GNT2 and B4GALT2 are involved in polyLacNAc biosynthesis on N-linked tetraantennary structures and on Core 1, 2, and 3 O-glycan core structures [9–11], ST6GALNAC2 terminates glycan chain elongation by the addition of sialic acid to the peptide GalNAc of O-glycan Core 1 or Core 3 structures [10,12]. Similar to GALNT12, mutations in these three genes predominantly mapped to the catalytic domain of respective enzymes [8]. Biochemical analyses of wild-type and mutant B3GNT2, ST6GALNAC2, and B4GALT2 demonstrated significant alterations in respective glycosyltransferase function [8]. For instance, mutations in B3GNT2 led to either loss of enzymatic function or mislocalization of the protein. Genomic loss of the wild-type allele in B4GALT2 was accompanied by a missense mutation that disrupted the post-translation modification pattern of the B4GALT2 protein [8], and finally, mutant ST6GALNAC2 consistently showed hyper enzyme activity [8] as compared to its wild-type counterpart.

GALNT12, B3GNT2, B4GALT2, and ST6GALNAC2 are enzymes involved in mucin type O-glycosylation and catalyze the formation of complex O-glycan structures [10]. While the endogenous substrates of these enzymes remain to be determined, given that an individual glycosyltransferase enzyme may be involved in the post-translational modification and maturation of multiple protein substrates, it is possible that genetic alterations in glycosylation pathway genes may accordingly affect a multitude of proteins, through loss-of-function, dominant negative, or gain-of-function mechanisms. This can be considered comparable to the paradigm where a single microRNA can regulate multiple transcripts; defects in these microRNAs can therefore alter the expression of multiple target genes. Glycosylation defects, either in the form of impaired glycan synthesis or abnormal glycan structures, could likewise affect the regulation and function of several different proteins and signaling molecules. For example, GALNT12 catalyzes the initiation step of mucin type O-linked glycosylation [13]. Mucins are a group of glycoproteins with extensive O-glycan side chains that are highly expressed in normal intestinal tissue, alterations of which can lead to disease conditions such as cancer [14]. Indeed, increased levels of unglycosylated MUC1 (Mucin-1) protein were observed in CRCs bearing the GALNT12 mutations [7]. In addition, hypoglycosylation of MUC1 is known to affect protein stability and subcellular localization and may initiate oncogenic signaling [15], while MUC2 (Mucin-2)-deficient mice are observed to spontaneously develop colorectal tumors likely due to increased inflammation in the colonic mucosa [14]. Similarly, B3GNT2, B4GALT2, and ST6GALNAC2 are essential in formation of O-glycan core 1, 2, and 3, structures. O-Glycans are classified into different Core groups and expression levels of Core 1, 2, and 3 structures are deregulated in different cancers [16]. For example, the tumor-associated Tn antigen is a modified Core 1 O-glycan that is expressed in pancreatic ductal adenocarcinomas; formation of these incomplete Core 1 glycans are attributed to mutations in C1GALT1-specific chaperone 1 (COSMC), a chaperone protein that regulates the activity of Core 1 beta 1,3-galactosyltransferase 1 involved in Tn antigen elongation process [17]. On the other hand, the Core 3 glycans are of particular interest since they are primarily expressed in the gastrointestinal tissues, and are the primary structures of mucin-type glycoproteins [16]. Core 3 O-glycans help regulate intestinal homeostasis, and alterations in Core 3 O-glycan structure have been implicated in CRC development [14,16]. For example, decreased expression of Core 3 structures is frequently observed in CRCs and loss of Core 3 synthase expression is associated with the grade of colon neoplasia in familial adenomatous polyposis patients [16]. Furthermore, loss of Core 3 synthase activity enhances metastasis of colon carcinoma cells [16] and mice deficient in Core 3 O-glycans are highly susceptible to colitis and colon adenocarcinoma [18]. Consistent with these reports, our phenotypic functional studies demonstrated mutant B3GNT2 and ST6GALNAC2 glycosyltransferases as having a significant impact on the migratory potential of colon carcinoma cells [8]. Specifically, mutations in B3GNT2 result in a gain of oncogenic function, likely owing to altered surface polyLacNAc residues that have been suggested to play a role in migration and metastasis [19]. In contrast, mutant ST6GALNAC2 proteins were unable to suppress the migration of colon cancer cells, an observation that is analogous to previous reports of ST6GALNAC2’s function as a metastasis suppressor gene in breast cancer [20]. Collectively, these findings strongly suggest that genetic defects in the mucin-type O-glycosylation pathway contribute to the progression of a subset of CRCs.

Interestingly, given previous findings from our group and others [7,21] demonstrating a strong association between germline defects in GALNT12 and CRCs, it is possible that inherent defects in glycosyltransferase genes also potentially play a role in susceptibility to familial forms of CRC. In fact, estimates suggest that up to 20% of CRC cases have an inherited susceptibility to CRC; albeit the underlying genetic defects remain to be identified [22]. Our recent findings of additional mutated genes in the O-glycosylation pathway [8] will now allow us to further explore the contribution of glycosylation pathway defects to inherited CRC susceptibility.

Taken together, while genetic defects in the glycosylation pathway could be one of the underlying mechanisms contributing to aberrant glycosylation, deciphering the endogenous substrates of glycosyltransferases and the functional complexities associated with protein glycosylation will significantly aid in understanding the precise role of aberrant glycosylation in tumor pathogenesis [23]. In addition, recent advances in integrative peptide fragmentation, glycopeptide extraction, and high-throughput mass spectrophotometry for quantitatively analyzing global tumor glycan structures [24,25] should further enable the development of evidence-based biomarkers and new therapeutic strategies for cancers.

Declaration of interest

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

The authors were supported by the U.S. Department of Health and Human Services, the National Institutes of Health and the National Cancer Institute (grant numbers K08 CA148980, P50 CA150964, R01 CA204549, U01 CA152756, U54 CA163060).