Intestinal Goblet Cells and Mucins in Health and Disease: Recent Insights and Progress

Mucins are highly glycosylated large glycoproteins with protein backbone structures rich in serine and threonine, which are linked to a wide variety of O-linked oligosaccharide side chains that make up more than 70% of the weight of the molecule [3, 4]. Up to 20 different mucin genes have been identified, MUC1 to MUC20 according to order of their discovery. Mucin genes are expressed in tissue and cell type–specific manner and are broadly classified into two types, secretory and membrane-associated. Gel-forming secretory mucins such as MUC2, MUC5AC, MUC5B, and MUC6 are localized on chromosome 11.5.5 as a cluster. In small and large intestine, MUC2 is the major secretory mucin synthesized and secreted by goblet cells, whereas goblet and absorptive cells express membrane-bound mucins, MUC1, MUC3, MUC4, MUC13, and/or MUC17, in the apical membrane [3, 4, 15]. Secretory and membrane mucins have distinct structural features and biosynthetic pathways.

MUC2 is the first human secretory mucin to be identified and characterized [15, 16]. MUC2 mucin has structural and physicochemical properties similar to those of other gel-forming secretory mucin such as MUC5AC, MUC5B, and MUC6, expressed in gastric and respiratory glandular epithelium. MUC2 mucin monomer has more than 5000 amino acids and consists of central tandem repeat domains rich in proline, serine, and threonine, the latter of which are linked O-glycosidically to many oligosaccharide side chains of varying lengths and compositions (also called “PTS” domains). The highly glycosylated central tandem repeat domains of MUC2 mucin monomers are flanked on either side by the cysteine-rich domains, including C-terminal cysteine knot domain and four D domains of von Willebrand factor (vWF), which are involved in dimerization and oligomerization, respectively, resulting in highly viscous gel-forming mucin network (Fig. 1) [16‐18]. The rodent homologue of MUC2 is designated as Muc2, and has a similar structure motif and cell and tissue type–specific expression.

Intestinal mucosal epithelia cells also express epithelial membrane-bound mucins, MUC1, MUC3, MUC4, MUC12, MUC13, and MUC17, which have structural similarity. MUC3 mucin, the most abundantly expressed membrane mucin in the small intestine, consists of two subunits, an extracellular subunit containing heavily O-glycosylated tandem repeat domains and two epidermal growth factor (EGF)—like domains separated by sperm protein, enterokinase, and agrin (SEA) module, a proteolytic cleavage site during biosynthesis of MUC3 in the endoplasmic reticulum, and a membrane-associated subunit with trans-membrane domain and cytoplasmic tail with potential phosphorylation sites involved in signaling (Fig. 1) [19‐21]. MUC3 expression in the apical membrane of absorptive and goblet cells shows a maturational gradient with increasing expression from crypt to villus. These membrane mucins extend rod-like 200–1500 nm above the cell surface and form the glycocalyx (Fig. 1). The membrane mucins may be shed from the cell surface by the activation of membrane-associated metalloproteinases, by the separation of two subunits in the SEA domain, or by the alternative splicing contributing to mucus layer formation [3]. The cysteine-rich EGF-like domains of the mouse Muc3 and human MUC17 mucins have been shown to inhibit apoptosis and stimulate cell migration, implying a bioactive role in maintaining the integrity of the surface epithelial layer [22, 23].

Like other gel-forming secretory mucins, MUC2 mucin monomers form dimers in the endoplasmic reticulum via intermolecular disulfide bonding between the c-terminal cysteine knot domains. During transit through Golgi apparatus, MUC2 mucin proteins become heavily O-glycosylated and undergo trimerization via disulfide bonding in the amino terminal region, resulting in the formation of very large polymers [17, 18]. Proteolytic processing may also occur in the late secretory compartments. It is possible that other mucin proteins, such as MUC5B, may polymerize with a different pattern, forming linear chains rather than trimers at the amino terminal region as with MUC2 [24]. Numerous and diverse oligosaccharide side chains linked O-glycosidically to apomucin are synthesized by the sequential addition of five different sugars by the action of a series of Golgi resident glycosyltransferases. These oligosaccharides can generate diverse signals by binding to many different ligands such as bacteria, viruses, lectins, adhesion molecules, growth factors, and cytokines [25, 26•]. The glycosylated MUC2 monomer has a mass of about 2.5 MDa and the polymer may be more than 100 MDa. The fully glycosylated and processed MUC2 mucin is densely packed and stored in secretory granules/vesicles and released by two pathways, constitutive/basal and Ca2⁺-dependent stimulated/regulated pathways described in the previous section. Massive secretion of MUC2 mucin by exocytosis in stimulated secretion is triggered by a wide array of bioactive factors, including cholinergic agonist, hormones (neuropeptides), microbes and microbial products and toxins, inflammatory cytokines, and reactive oxygen and nitrogen species. These mucin secretagogues signal through secondary messengers such as intracellular Ca²⁺, cyclic adenosine monophosphate (cAMP), and diacylglycerol that activates protein kinase [14].

Mucin secretion is frequently coupled with increased synthesis of mucins, but chronic secretion may result in goblet cell depletion and decreased synthesis of mucins. The syntheses of mucins are regulated by many bioactive factors that function as mucin secretagogues. Increasing evidence indicate that mucin expression is controlled either by transcriptional or epigenetic regulation [4, 27, 28]. Transcriptional regulation of MUC2 is mediated by activation of signaling pathways targeting the transcription factors that bind to specific sites on MUC2 promoter. Bioactive factors including microbes, microbial products, toxins, cytokines, hormones/neuropeptides, and growth factors have been reported to be involved in positive or negative regulation of MUC2 transcription [4, 28]. Activation of a transcription factor, nuclear factor (NF)-κB, is a common event during inflammation in the gastrointestinal tract and MUC2 mucin has been shown to have NF-κB binding sites in the promoter. Lipopolysaccharide (LPS) from gram-negative Pseudomonas aeruginosa upregulates MUC2 transcription through activation of NF-κB mediated by the Ras–mitogen-activated protein kinase (MAPK) pathway in colon epithelial cells. P. aeruginosa also binds to a glycolipid receptor, asialo GM1, at the cell surface through flagellin, causing the release of adenosine triphosphate (ATP), which in turn increases the intracellular Ca²⁺ concentration, leading to the activation of NF-κB mediated by downstream signaling pathways [29]. An inflammatory cytokine, tumor necrosis factor (TNF)-α (Th1 type cytokine), upregulates MUC2 transcription through activation of NF-κB mediated by PI3K/Akt pathways. TNF-α also has an inhibitory effect on MUC2 transcription through activation of JNK pathway, but overall effect of TNF-α treatment of colon cancer cells is a net increase in MUC2 transcription [30]. Th2 type cytokines, interleukin (IL)-4 and IL-13, also upregulate MUC2 transcription through NF-κB activation mediated by MAPK [31]. By contrast, a neuropeptide hormone, vasoactive intestinal peptide (VIP), upregulates MUC2 transcription through activation of transcription factors CREB/ATF1 mediated by MAPK and p38 pathways [32]. A prostaglandin, PGE2, also induced MUC2 transcription through activation of CREB/ATF1. To summarize, both Th1 and Th2 cytokines and microbial products, LPS, lipoteichoic acid (LTA), lipopeptide (LP), and flagellin induce MUC2 transcription through activation of NF-κB, whereas activation of CREB/ATF1 is involved in neuropeptide/hormone-induced MUC2 transcription. Recently, the SPDEF (SAM pointed domain-containing Ets) transcription factor has been shown to work downstream of Math1 to promote differentiation of secretory progenitor cells into goblet and Paneth cells, and is a major regulator of secretory gene products such as MUC2 [33].

Epigenetics describes heritable changes in gene expression that are not associated with changes in genomic DNA sequences. Epigenetic regulation that leads to the turning on or off of genes involves close interaction of DNA methylation, histone modifications, and microRNA silencing. Studies of MUC2 in mucinous and nonmucinous colon cancers showed that methylation of CpG islands in the specific regions of MUC2 promoter downregulates MUC2 expression [34]. A recent study also showed that MUC2 gene expression is regulated by a tightly associated epigenetic mechanisms of DNA methylation and histone modifications in the 5′ flanking region of MUC2 promoter [35].

Intestinal mucus layers secreted by goblet cells consist mainly of compact mesh-like network of viscous, permeable, gel-forming MUC2 mucin, which provides the frontline host defense against endogenous and exogenous irritants and microbial attachment and invasion but allows the transport of nutrients. Mucus layers also contain other goblet cell products, TFF3, RELMβ, and Fcgbp; antimicrobials peptides such as β defensin and lysozymes secreted by Paneth cells; and secretory IgA secreted by enterocytes [1, 2, 5, 10]. There are two mucus gel layers in the gastrointestinal mucosa, an inner firmly adherent layer and an outer more loosely adherent layer, both consisting largely of MUC2 mucin in the intestine (Fig. 1) [36, 37]. Gastric mucus in the human stomach consists of MUC5AC and MUC6 mucins [38, 39]. A study of the thickness of two mucus layers in the rat gastrointestinal mucosa showed that the inner firmly adherent layer was much thicker and continuous in stomach, ileum, and colon compared to jejunum, which had a much thinner, patchy inner mucus layer (Table 1) [36]. Microfold (M) cells in the intestinal mucosal epithelium are devoid of overlying mucus layers and have the ability for transepithelial transport of foreign antigens and microbes to be captured by the underlying dendritic cells [40]. The outer loosely adherent mucus layer formed by proteolytic and glycosidic degradation of highly polymerized gel-like MUC2 mucin was similar in thickness in stomach and jejunum, but markedly increased in ileum, and thickest in colon (Table 1) [36]. Microbes are associated mostly in the outer loose mucus layer and absent from the inner firm mucus layer, indicating that the inner firm mucus layer functions as a critical protective barrier against bacterial adhesion and invasion of underlying epithelial cells [41••]. The outer loose mucus layer provides a good habitat for microbial colonization, because oligosaccharides of MUC2 mucin provide numerous microbial attachment sites and energy source (Fig. 1) [42••]. Muc2-deficient mice lack the intestinal mucus layers and demonstrate increased permeability and bacterial adherence to the epithelial cell surface [41, 43]. The thickness of mucus layers is maintained by a balance between synthesis, secretion, and degradation, modulated by the microbial glycosidases and proteases and the mechanical shear forces of peristalsis. The important role of membrane-bound mucins at the apical cell surface and at the interface of the cell surface and the inner mucus layer was demonstrated by Muc1-deficient mice showing increased susceptibility to invasion by Campylobacter jejuni [44].

The human adult intestinal microbiota colonizing the outer loosely adherent mucus layer is made up of more than 10¹⁴ bacteria with increasing gradient of concentration and complexity of bacterial population from jejunum to colon; stomach, duodenum, and jejunum with 10²–10³ aerobic bacteria per gram luminal content, and 10⁷–10⁸ and 10¹¹–10¹² predominantly anaerobic bacteria per gram in distal ileum and colon, respectively [45•]. Recent studies using culture-independent techniques of 16S ribosomal RNA gene sequencing and meta genomic sequencing methods estimate the presence of 200–300 species per individual from more than 1000 prevalent bacterial species in human adult intestine. It was further shown that 99% of intestinal resident/commensal microbiota belongs to only four phyla: Firmicutes (composed mostly of Clostridium XIX and IV groups) and Bacteroidetes (which together constitute more than 90% of the total intestinal microbes), with Proteobacteria and Actinobacteria making up the rest. However, the analysis of mucosally associated bacteria showed enrichment of streptococcal and Lactobacillus spp (Bacillus subgroup of Firmicutes) [46, 47].

The normal intestinal mucosal epithelium has tolerance to commensal microbiota because of its ability to distinguish commensal microbiota from pathogenic microorganisms by their molecular patterns, such as microbe-associated molecular patterns and pathogen-associated molecular patterns, through pattern recognition receptors (PRRs) such as cell surface Toll-like receptors (TLRs) and cytoplasmic nucleotide-binding oligomerization domain (NOD) proteins [48, 49]. These PRRs are thought to activate a common signaling pathway leading to the activation of NF-κB that stimulates the production of inflammatory cytokines and co-stimulatory molecules. However, the normal intestinal epithelial cells express low levels of TLR4 and TLR2 and are hyporesponsive to the stimuli caused by the respective ligands.

Intestinal commensal microbiota depends on mucus and undigested dietary carbohydrates for binding sites and energy source and affects intestinal epithelial functions, including those of goblet cells and mucus layers, by a “cross talk” feedback mechanism. Diverse oligosaccharides of mucin glycoproteins were evolved to promote symbiotic relationship with commensal bacteria and to evade pathogens [1, 2, 42••]. Microbiota and microbial products can modulate mucin synthesis and secretion, either by direct activation of diverse signaling cascades or through bioactive factors generated by epithelial and lamina propria cells. Enteric pathogens circumvent the protective function of mucus layer by developing motility, mucolytic activity, and other virulence factors, causing the degradation and penetration of mucus layers and subsequent attachment and invasion of epithelial cells. The bacteria that gain access to the intestinal mucosal cell surface adhere more frequently to the glycocalyx of the most luminal portion of the villi or upper crypt in the intestine. Most commensal and pathogenic bacteria attach to the intestinal mucosal cells through interaction of adhesins to the mucosal receptors, such as integrins, carcinoembryonic antigen-related cell adhesion molecules (CEACAM) or sialylated, galactosylated, or mannosylated glycoproteins or glycolipids. Probiotics such as Lactobacillus plantarium were reported to induce MUC2 and MUC3 mucins and inhibit the adherence of EPEC (enteropathogenic Escherichia coli), indicating that enhanced mucus layers and glycocalyx overlying the intestinal epithelium and the occupancy of the microbial binding sites by Lactobacillus spp provide protection against invasion by the pathogens [50, 51]. Probiotics also cause qualitative alterations in intestinal mucins, preventing the pathogen binding.

The trefoil factor family comprises a group of small peptides (6.5–12 kDa) with three intramolecular disulphide bonds, which is highly expressed in the mucus-producing cells of the gastrointestinal tract and plays a significant role in epithelial restitution (Table 2) [8, 52]. TFF3/intestinal trefoil factor (ITF) is expressed and secreted by goblet cells in the intestine, whereas TFF1 and TFF2 are expressed and secreted by gastric surface foveolar cells and mucous neck cells/pyloric glands, respectively. TFF3 is the second most abundant goblet cell product present in the theca of mature goblet cells, and monomeric (6.6 kDa) and dimeric (13 kDa) forms of TFF3 can be found in vivo. In vitro and in vivo studies indicate that TFF3 facilitates not only intestinal epithelial restitution but also mucosal protection [8, 52]. TFF3 and mucin together were more effective in protecting epithelial cells in vitro when compared with either one alone [53]. Mice overexpressing rat TTF3 in the intestine showed increased resistance to intestinal damage and ulceration, whereas TFF3-deficient mice were more susceptible to dextran sodium sulfate (DSS)-induced colitis. Furthermore, administration of recombinant TFF3 improved DSS-induced colitis by restoring the capacity for restitution [8, 52]. The recent studies on the molecular mechanisms through which TFF3 promotes epithelial restitution indicate the involvement of multiple interactive mechanisms. TFF3 facilitates cell migration but not cell proliferation, unlike other motogenic molecules, such as transforming growth factor β (TGFβ) or hepatocyte growth factor (HGF). In addition, TFF3 blocks apoptosis and contributes to the innate immune response mediated by the mucosal sensor systems for commensal microbiota, such as Toll-like receptor family members, TLR2 and TLR4 [54]. TLR2 stimulation induced TFF3 transcription via Ras/MEK/MAPK and PI3K/Akt pathways but not MUC2 transcription in vivo in mice. TLR2-deficient mice showed a selective defect in TFF3 and TFF3 administration improved DSS-induced colitis by suppressing mucosal apoptosis [54]. Furthermore, TFF3 has been suggested to increase the viscosity of mucin by binding to the vWF C domain of MUC2, enhancing the structural integrity of the intestinal mucus barrier [55]. Recently, TFF3 was reported to form disulfide-linked heteromer with Fcgbp protein, which could interact with MUC2 mucin in a covalent and non-covalent manner, contributing to the stability of mucin network in the mucus layer [56].

RELMβ belongs to a family of resistin-like cytokine molecules consisting of small, cysteine-rich secreted proteins. It contains 12.5-kDa subunits that form disulfide-dependent dimeric units. RELMβ is produced by goblet cells in the intestine, with an increasing anterior to caudad gradient, and secreted into the intestinal lumen at high concentration as a homodimer (Table 2) [57]. RELMβ is highly induced in the intestinal goblet cell by colonization with normal enteric bacteria and in mouse models of gastrointestinal helminth infection and inflammatory bowel disease [58, 59]. RELMβ may have an immunoregulatory function. RELMβ induced by Th2-associated nematode infection had antiparasitic activity through IL-4 and IL-13 dependent mechanism [60]. Induction of Th2 cytokines, IL-4 and IL-13, caused expulsion of parasites such as Nocardia brasiliensis and Heligmosomoides hypogyrus from the intestinal lumen, by inducing RELMβ, which functions not only as Th2 cytokine immune effector molecule but also as an inhibitor of chemotaxis of parasites, interfering with parasite nutrition by directly binding to the chemosensory components of the parasites. Induction of RELMβ was much more significant than Muc2 and TFF3 on Th2 immune stimulation [60]. Furthermore, RELMβ upregulates Muc2 transcription and secretion, contributing to the mucosal barrier integrity. It showed a markedly preventive effect on trinitrobenzene sulfonic acid (TNBS)—induced colitis in mice [61]. RELMβ-deficient mice showed reduced susceptibility to T-cell-independent DSS-induced colitis, but increased susceptibility to T-cell-dependent TNBS-induced colitis, indicating that RELMβ may trigger protective or proinflammatory effects, depending on the nature of the chemically induced colitis [59].

Fc-γ binding protein (Fcgbp) is expressed in the mucus granule of goblet cells [62]. Fcgbp contains 13 vWF D domains and was reported to bind IgG antibodies through Fc part of IgGs. Recent proteomic analyses of the two mucus layers from the colon of mice and humans showed that N terminal parts of Fcgbp was covalently attached to Muc2 mucin (Table 2) [37]. It was suggested that covalent attachment of Fcgbp protein to MUC2 contributes to cross linking and stabilization of mucin network in the inner mucus layer [37].

In most intestinal infections, induction of goblet cells and mucin synthesis and secretion occur frequently during acute phase. However, chronic infection results in the depletion of goblet cells and quantitative and qualitative alteration in mucus layers due both to altered synthesis and secretion of mucins and to microbial glycosidases and proteases [1, 2, 48]. Rapid and massive secretion of goblet cell mucus not only facilitates the expulsion of pathogens but also maintains the integrity of mucus protective layers. Intestinal parasitic infections cause profound changes in the goblet cells and mucins of the small intestine. Many experimental helminth infections such as Nippostrongylus brasiliensis and Trichinella spiralis have been shown to cause intestinal goblet cell hyperplasia and increased mucus secretion, mediated by Th2 immune responses (IL-13 and IL-4), contributing to increased intestinal mucosal protection and worm expulsion [48, 63]. Expulsion of the nematode, Trichuris muris, in resistant, susceptible, and Muc2-deficient mouse strains showed that the increased Muc2 production, observed exclusively in resistant mice, correlated with worm expulsion, whereas worm expulsion from the intestine was significantly delayed in Muc2-deficient mice [64••]. In addition, the mucus barrier in resistant mice was less permeable than that of susceptible mice. These findings underscore the important innate defense function of mucins in enteric infection.

However, with sustained hypersecretion of stored mucins, depletion of goblet cells occurs, resulting in defective mucin synthesis and mucus barrier. In N. brasiliensis infection in rodents, goblet cell hyperplasia is accompanied by alteration of terminal sugar residues of goblet cell mucins, such as increased sialylation, which is mainly responsible not only for the expulsion of damaged worms but also for the prevention and attachment/migration of normal worms [65]. The pathogens have evolved mechanisms to invade the intestinal mucosal cells by penetrating through the mucus layers. A protozoan parasite, Entamoeba histolytica colonizes mucus layer of the colon by adhering to mucin oligosaccharides. The parasite then invades through the mucus layers by secreting cysteine proteases, which cleave the MUC2 mucin, resulting in defective mucus barriers, through which it can invade and attach to the intestinal mucosal cell surface [66].

Despite the well-known clinical manifestation of mucus accumulation in the lung, intestine, and other organs in CF, the mechanisms underlying mucus-associated pathology remain unclear [67]. Within goblet cells, mature mucin polymers are thought to exist in a highly condensed form, due mainly to neutralization of the repulsive forces of the polyanionic charge of the oligosaccharide side chains by the high concentration of H⁺ and Ca⁺⁺ within the granules. During exocytosis, H⁺ and Ca⁺⁺ are removed from the anionic sites by extracellular HCO ₃ ⁻ causing rapid expansion of compact mucin polymers into the meshed network of viscous mucus gels due to anionic repulsive electrostatic forces [68, 69•]. The loss of CFTR in CF results in the loss of Cl⁻ and HCO ₃ ⁻ transport [69•]. HCO ₃ ^- is critical for normal mucus gel formation, and aggregated mucus observed in CF may be caused by defective HCO ₃ ⁻ transport [67, 69•]. Increased fucosylation of mucin glycans due to induction of fucosyl α1-2 glycosyltransferase may also contribute to the increased viscosity of mucin [70]. CF mice exhibited decreased intestinal motility with longer exposure to the bacteria entrapped in the mucus and small intestinal bacterial overgrowth, which responded to treatment with antibiotics and laxatives [71].

IBD is thought to be caused by continuous pathologic immune responses to altered and/or dysbiosis of commensal microbes and microbial products [45, 72]. Accumulating evidence indicates that complex interactions of the following events are involved in the pathogenesis of IBD: 1) altered and/or dysbiosis of commensal microbiota, such as decreased ratio of protective to aggressive commensal bacterial species or functional alteration of commensal bacteria; 2) defective bacterial killing and processing by the intestinal mucosal cells due to genetic defect; 3) defective mucosal barrier function resulting from abnormal synthesis and or processing of mucins and TFF3; and 4) defective host immune response, such as defective innate and or adaptive immunity. Alterations of mucin production and glycosylation occur in IBD, but whether they contribute to initiation of inflammation or are the result of inflammation is unknown.

Goblet cells are reduced in number and size in ulcerative colitis. Recent studies in mouse models of colitis highlight the importance of the role of mucin in maintaining the integrity of protective mucus barriers whose breakdown can result in colitis. Muc2-deficient mice with no morphologically identifiable goblet cells and absent Muc2 expression (TFF3 and RELMβ were expressed) in the intestine recently were reported to have markedly deficient mucus layers with increased permeability and enhanced bacterial adhesion to the mucosal cell surfaces. These mice developed spontaneous colitis and were susceptible to DSS-induced colitis [41, 43]. Two strains of mice with single missense mutations in oligomerization domains (vWF D3 and vWF C, respectively) showed aberrant Muc2 synthesis, reduced storage of mucin in the theca of goblet cells, diminished mucus layers with increased permeability and increased production of Th1 and Th2 cytokines, and developed spontaneous distal colitis and increased susceptibility to DSS-induced colitis [73••]. These mice showed aberrant assembly and processing of mucin complex and biochemical and histologic evidence of endoplasmic reticulum stress in goblet cells and Paneth cells, the phenomena observed in the patients with ulcerative colitis [73••].

Changes in mucin glycosylation have been reported to cause defective mucus barrier and increased permeability and increased susceptibility to DSS-induced colitis, which may be caused by early thinning and weakening of the mucous layer allowing increased bacterial translocation [74]. Mice lacking core 3 β1,3-N-acetylgulcosaminyl-transferase enzyme with defective O-glycans in mucins showed decreased Muc2 synthesis, increased intestinal permeability and increased susceptibility to DSS-induced colitis [75••]. Similarly, mice lacking core 2 β1,6-N-acetylglucosaminyl-transferase enzyme that initiate core 2 type protein O glycosylation in mucin, demonstrated increased intestinal permeability and increased susceptibility to DSS-induced colitis [76].

Increased numbers of mucosa-associated (adherent invasive) E. coli are observed in IBD. A recent study on tissue-associated microflora using 16S rRNA gene sequencing in patients with IBD showed that the IBD group had depletion of commensal bacteria, especially members of the phyla Firmicutes (largely Clostridium XIVa and IV groups) and Bacteroidetes with concomitant increases in Proteobacteria and Actinobacteria [46]. Although the causes of the altered pattern of microflora (dysbiosis) in IBD patients are not clear, a recent observation that coordinated regulation of glycan degradation and polysaccharide capsule synthesis can occur in the intestinal lumen, underscores the importance of the dynamic interplay between mucin glycan metabolism and microbial ecology [42••].

Mucinous carcinoma is characterized by pools of abundant amounts of extracellular mucin, and defined as the tumor with mucin representing more than 50% of the tumor mass. By contrast, signet-ring cell carcinoma is defined by the presence of more than 50% of tumor cells with prominent intracytoplasmic mucin, and has worse prognosis than mucinous carcinoma. Mucinous carcinoma occurs more frequently in the colon, where it accounts for 6–19% of all colorectal cancer cases, than in the small intestine, where it occurs mostly in the appendix. Mucinous carcinoma has clinicopathologic properties and molecular changes that are different from nonmucinous cancer, such as higher stage at diagnosis, more frequent invasion of adjacent organs, lymph node metastasis and peritoneal dissemination, and more frequent occurrence of microsatellite instability (MSI), CpG island methylator phenotype (CIMP), and BRAF mutations [77].

The changes in mucins that occur in colorectal cancer may be broadly classified into two main types: aberrant mucin gene regulation and glycosylation. Mucinous carcinoma expresses high levels of MUC2 goblet cell mucin whereas MUC2 expression is downregulated in nonmucinous cancer. Upregulation of MUC2 transcription in mucinous carcinoma is caused by altered epigenetic and genetic regulation of MUC2, such as MUC2 promoter hypomethylation and increased binding of goblet cell lineage associated transcription factor, HATH1, to MUC2 promoter [13, 34]. Mucinous carcinoma ectopically expresses gastric mucin, MUC5AC, which is absent from normal intestine, more frequently than nonmucinous cancer. In addition, MUC1 membrane mucin is expressed at lower level in mucinous carcinoma compared to nonmucinous cancer [78]. Changes in glycan structures of mucins in epithelial cancers include truncation of oligosaccharide side chains (resulting in the expression of T, Tn, sialylTn antigenic determinants) and neosynthesis of glycans (resulting in the expression of sialylLeA and sialylLeX antigens). Studies have shown that these glycan antigens are highly expressed in both mucinous and nonmucinous cancers [3, 4, 25, 79].

Accumulating evidence indicates that mucin plays an important role in several stages of metastatic processes of colorectal cancer. First, colon cancer cell lines selected for high capacity for metastasis express high levels of MUC2. Second, colon cancer cells selected for high mucin production showed increased liver colonization and metastasis. Third, inhibition of mucin O-glycosylation in colon cancer cells decreased cancer cell binding to E-selectin expressed on endothelial cells, and liver colonization. Fourth, downregulation of MUC2 expression by antisense techniques decreases the metastatic activity of human colon cancer cells in nude mouse model system. To summarize, altered level and pattern of mucin gene expression and altered glycosylation in cancer cells affect their biologic properties, such as cell proliferation, adhesion, motility, invasion, escape from host immune surveillance, tumorigenicity, and metastasis [79].