Correspondence to: H. Devaraj, Unit of Biochemistry and Glycotechnology, Life Sciences Building, University of Madras, Guindy campus, Chennai – 600025, India. E-mail: firstname.lastname@example.org
Mucins are high molecular weight glycoproteins designed for cellular protection and sensing the external environment. Aberrant glycosylation and altered mucin expression seen in cancers are implicated in mucin-dependent refraction to immunosurveilance and immunosuppressive induction around the tumour. Although mucins provide molecular targets for immune system's tumour recognition, their characteristics dictate that the nature of immune response required for recognition and lyses of mucin-expressing tumours needs to follow predominantly a MHC-unrestricted αβ TCR-mediated effector cell response. Frequent loss of dendritic cells maturation and elimination of reactive lymphocytes altered adhesive and anti-adhesive properties of the mucins, promote tumour survival and escape from the immune response.
Mucins are expressed by epithelial cells lining gastrointestinal and urogenetal tracts and glandular organs . Expression of mucin is cell- and tissue specific, and any alteration is taken as an indication of loss of tissue homoeostasis . Several studies, including our own, have characterized the shift in the mucin expression and its glycosylation pattern during carcinogenic transformation and used it as a biomarker for transformation [3-5]. Besides, presence of immunodominant tandem repeats and unique and altered glycosylation patterns makes it an ideal candidate for development of cancer vaccines . Nevertheless, development of tolerance to mucin immunization due to functional pliotrophism exhibited by mucins called for fresh studies that evaluated the immune regulative role of mucins to augment the cancer vaccine designs . This review overviews the mucin-dependent immune modulations to appreciate the basis behind tumour immunoevasion and vaccine development.
Mucin glycosylation shift in cancer
Mucin forms the crucial link that translates injury-mediated reactionary environment into a sustained genetic/physiological response that is pivotal to the initiation and progression of cancers. Persistent injury or infection activates lymphocytes to secrete pro-inflammatory cytokines that results in constitutive mucin sensing and aberrant expression . These aberrations arise as a consequence of the deregulation of expression of mucin core proteins and the enzymes that modify them, during the transformation of tumour cell [9, 10]. Transformation-related changes in mucin glycosylation and constitutive expression are therefore an inherent property of epithelial cancers . The nature of cytokine profile, the degree and duration of inflammation have a profound effect on mucin expression and play a causative role in initiating mucin-dependent oncogenic cell signalling and immunomodulation. The cell-specific and cytokine-dependent expressions of mucins are indeed natural healing processes subverted to aid the tumour formation and progression in an aberrant environment .
Cancer-associated mucin glycosylation is characterized by a general reduction of glycosylation and truncation of O-linked glycans [12, 13] (Fig. 1). Glycosylation shift from long-chained core 3 to short core 1 type in cancers results in expression of Thomson–Friedenrich (TF) and other fucosylated Sialyl-Lex and Sialyl-Lea antigens . One possible mechanism leading to increased core 1 structure in cancers may be a shift of O-glycan biosynthesis following changes in the peptide structure of mucin core  or by the relocalization of glycosyltranferases within the golgi complex as a direct pathological response to increase in intragolgi pH [16, 17]. For example, detection of Sialyl Tn initially in trans-golgi and later in all of Golgi compartments and rough ER during the adenoma–carcinoma sequence of colorectal cancers suggests that enzymes involved in the synthesis of Sialyl Tn progressively altered in their subcellular localization . Regulations in the Sialyl transferases and sulfotransferase activities, especially its upregulation, during the course of malignancy also explain the variations seen in the expression of sulphated and sialylated epitopes in most of the cancers [9, 19]. Inflammatory cytokines such as TNF-α are directly implicated in the activation of glycosyltransferases and sulfotransferases resulting in biosynthesis of sialylated and sulphated Lewisx epitopes [8, 20]. Further, mucins secreted by cancer cells induce several cytokines such as IL6 and PEG2 from peripheral blood monocytes/macrophages through orphan receptor activations and subvert them for prognosis of the cancer . Indeed, cancer cells show distinct changes in the cellular repertoire of glycosyltransferases, unique to the tissue of its origin, and express glycan epitopes that distinguish a cancer from the other .
Capacity to synthesis diverse carbohydrate epitopes is a prerequisite for a possible neoplastic transformation and provides the means with which a tumour can interact with host system . Multivalency exhibited by mucins in sialylated and/or fucosylated Lewis x/a epitopes increases the avidity with which selectins and other ligands bind to mucins . Besides, distinct combination of different o-glycans presented on the apomucin backbone creates specific binding sites for each selectin and is responsible for the uniqueness shown by each selectin in binding with mucins . Indeed, variations in the enzymes that alter the position and number of GalNAc residues attached to the mucin core polypeptides influence the metastatic abilities of colon carcinoma cells . Whereas cell surface mucins facilitate carcinoma cell interaction with leucocytes, platelets and endothelial cells, secreted mucins inhibit such interactions. Poor response of cellular immune response against tumour antigens is partly attributed to the soluble mucins that could prevent trafficking of tissue homing T lymphocytes and its adhesion and extravasion into tissues [26, 27]. Demonstration of L-selectins’ preferential binding to mouse embryonic MUC-10 mucin and not to adult MUC-10 in sub-mandibular glands  suggests that such mechanisms could be utilized by the tumour to express isoforms that has morphogenetic and cell–cell interactive roles.
Altered glycosylation and antigen processing
The tumour-protective ability of mucins against the host immune response is embedded on its structural peculiarity. The interested readers are directed to refer excellent reviews on mucin structural biology [29, 30] for a comprehensive account on this subject.
Mucins can be both immunostimulatory and immunosuppressive in their effects. MUC-1, for example, is a highly immunogenic tumour-associated antigen (TAA) that provides a unique immune system access to the MUC-1 over expressing breast, pancreas and ovarian carcinomas . If poorly glycosylated on its VNTR , it elicits humoral  and cellular immune responses , and the major epitope recognized by the antibodies is the PDTRPAP sequence with its o-glycosylation on its threonine residues [35, 36]. Interestingly, antigen processing of MUC-1 by dendritic cells (DC) or in human immunoproteasomes in vitro retains its o-linked glycans on its repeat domains. Its 20 amino acid tandem repeat (TR) posses three specific cleavage sites, being processed by human cathepsin L in low-density endosomes in a manner that is sensitive to o-glycosylation positions. Proteolysis of Thr-3-Ser-4 peptide bond in the TR does not occur if either amino acid is o-glycosylated, and this masking of cleavage site is responsible for inertness of tumour-associated MUC-1 glycoforms to effective DC processing . Further, it has been found that the processed SAPDT(GalNAc)RPAPG decameric glycopeptide containing a single sugar (GalNAc) binds strongly to MHC class I allele HLA A*0201, whereas the same sequence glycosylated with the disaccharide Gal-GalNAc does not bind at all . Processed MUC-1 TRs can use GalNAc to anchor on to the c-pocket of HLA class I (H-2 kb) molecule, and the number of anchors subsequently influences the affinity with which MUC-1 is presented on to the MHC class I . Low-affinity binding of the 9-mer MUC-1 peptide sequences (APDTRPA and STAPPAHGV) on to the HLA-A2 is partly due to the lack of high-affinity consensus motif and to the under glycosylation , and only HLA-A11 binding is close to the immunogenic value . Nevertheless, cytotoxic T lymphocytes (CTLs) generated against it are highly active and could lyse the human breast cancer cells expressing MUC-1 . Breast cancer cells therefore escape from autologous CTLs by expressing MUC-1-related antigenic epitopes more weakly or by modulating its antigenicity . Complete loss of MUC-1 is also observed in some breast tumour cell lines that are unresponsive or resistant to CTL cytotoxicity and characterized with antitumor immunity . Conversely, downregulation or loss of HLA class I expression in MUC-1 or c – erbB2 overexpressing NSCLC cells confer poor prognosis of the disease  and the mice lacking MHC- Class I made weak CTL response .
Modulation of cytokine profile and tolerance
Dendritic cells (DCs) form a crucial link between innate and adaptive immunity leading to specific T cell activation. MUC-1 in its aberrant glycosylated form on tumour cells is chemotactic to immature human myeloid DCs (through its polypeptide core) and a maturation and activation signal (through its carbohydrate moieties) but subverts the DC function by negatively affecting their ability to stimulate type 1 helper T cell responses . Human monocyte-derived DCs exposed to MUC-1 with sialylated core 1 (sialyl-T, ST) oligosaccharides, similar to those found in epithelial tumours in vivo, display a modified phenotype with decreased expression of costimulatory molecules (CD86, CD40), Ag-presenting molecules (DR and CD1d) and differentiation markers (CD83). Besides, markers associated with immature DC phenotype, such as, CD1a and CD206 (mannose receptor), are increased in its expression . Further, by altering the cytokine repertoire of monocyte – derived DCs and switch them into IL-10high IL-12low expressing antigen presenting cells (APCs), the tumour derived mucin cripple DCs immunostimulatory (Th 1 dependent) capacity and represses their functional differentiation and maturation . Mucin-dependent regulation of DC functions results in inadequate/impaired presentation of tumour antigens to T cells resulting in tolerance to TAAs and converts them into suppressor/regulatory T cells . Increased secretion of IL-10 interns causes T cell tolerance and anergy (Fig 2).
Although direct implication of MUC-1/DF 3 antigen in the apoptosis of activated T cells  is partially retracted, fresh studies on T cell suppression and induction of tolerance by MUC-1 suggest that upon MUC-1 challenge, expression of αβTCR and CD28 gets downregulated on CD8+ T cells resulting in the absence of detectable CTL activity and induction of peripheral tolerance . Active CTLs that infiltrate the pancreatic tumour microenvironment become cytolytically anergic and are tolerized to MUC-1 antigen, partly due to tumour microenvironment and to the presence of CD4+ CD25+ T regulatory cells that secrete IL-10 . MUC-1 also suppresses the T cell proliferation, which can be reversed by IL2 . However, the inhibition of cytolytic activity of human natural killer (NK) cells by ovarian cancer CA125 antigen could not be reversed by IL2 and did not involve alterations in proliferation or apoptotic induction, but related to major downregulation of CD16, suggesting that different mucins or its carbohydrate epitopes have different immune suppressive effects . Thus, while expression of Sialyl Tn antigen on colorectal cancer mucins inhibits natural killer T (NKT) cell cytotoxicity , aberrant glycosylated forms of Lea/Leb glycans on colorectal cancers interact with DC-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN) – C-type lectins – and impair its differentiation and functions , thereby influencing the prognosis of the cancer.
Ovarian cancer antigen, MUC-16/CA 125, is a known counter receptor for galectin-1  that can induce apoptosis of activated T cells in a CD45-dependent manner . Dose–response analysis demonstrates that a higher concentration (10 μm) of galectin-1 is required to induce T cell apoptosis, indicating that dimerization of galectin-1 is necessary for induction of apoptosis [55, 56]. However, increased deposition of galectin-1 on MUC-16/CA-125 in ovarian carcinoma results in facilitated dimerization and efficient presentation of galectin-1 leading to the death of tumour-infiltrating T cells even at a very low concentration . In fact, sequestration of galectin-1 by MUC-16 is so efficient that despite overexpression of galectin-1 by ovarian cancer cells, the serum concentration is lower than that of normal individuals . Association of galectin-3, a relative of galectin-1, with MUC-2 in colon carcinoma cells prevents tumour apoptosis and promotes proliferation and growth of the tumour [57, 58]. Further, galectin-3, by interacting with cancer-associated MUC-1 via TF, promotes cancer cell adhesion to endothelium by revealing epithelial adhesion molecules that are otherwise concealed by MUC-1 .
Tumour recognition and lysis
Inefficient tumour lysis characterizes most of the mucin overexpressing cancers. For instance, overexpression of MUC-4/SMC or MUC-16 inhibits lymphokine-activated killer (LAK) cells-mediated tumour lysis by masking the surface antigens on the tumour target cells [60, 61]. Efficient lyses of tumour cells by immune effectors require a clear distinction in their approach when it comes to mucin-expressing tumours. For one, mucins are superbly designed to protect the cells from both internal and external insults, and the penetration of mucin barrier and access to the tumour cells require both physical and physiological overturns. For example, expression of mucin antigens and membrane-spanning glycoprotein, Cancer Antigen (CA)-125, in ovarian cancer exerts immunosuppressive effects by entrapping/shedding effectors of the complement cascade and attenuates complement lysis of antibody-sensitized cells [62, 63]. Furthermore, lysis of episialin− melanoma cells by CLTs and LAKs involves a broad spectrum of adhesion molecules, whereas only LFA-1/ICAM-1 and CD2/LFA-3 pathways are exclusively utilized for the lysis of episialin + melanomas, blocking of which results in complete inhibition of cytolytic ability . Similarly, innate immune response is capable of recognizing chemotactic signals of secreted MUC-1 from DA3 mammary tumours expressing MUC-1/Sec phenotype. Secretary MUC-1 is capable of recruiting 3–4 times as many macrophages/APC as transmembrane phenotype (MUC-1/TM) and is mainly due to upregulation of MCP-1 (CCL-2) by MUC-1/Sec expression . Naturally, DA3/sec tumours are more susceptible to CTL-mediated rejection than DA3/TM tumours and therefore fail to develop in Balb-c mice . Recruitment of macrophages and monocytes involves interaction of GluNAc residues of mucin with calcium-type human macrophage lectin. Along with the neutrophils and eosinophils in the inflammatory infiltrate, they intern influence the mucin secretion by producing cytokines that trans activate EGF receptors [66, 67].
Mucin characteristics dictate that the nature of immune response required to address the recognition and subsequent lyses of mucin-expressing tumours should follow a MHC-unrestricted αβ TCR-mediated effector cell response [34, 68]. Frequent loss of DC maturation and ineffective MUC-1 processing qualitatively restricts MHC-dependent recognition and lysis of tumour cells. MUC-1, by far the most ubiquously expressed TAA, plays an important role in providing molecular targets for immune system tumour recognition [31, 35]. Prostate metastatic cancers that lack HLA class I expression are recognized and lysed by CD8+ CD56− T cells and CD8+ CD56+ natural killer T (NKT) cells in a manner that needs synergistic action of tumour-specific MUC-1, IL2 and IL12 and needs no MHC class I and CD1 expression . HLA-unrestricted CTL recognition of tumour-associated epitopes of MUC-1 involves TCR αβ, CD3 and CD8 and not the HLA type [70, 71], suggesting that expression of underglycosylated MUC-1 exposes highly antigenic repetitive multiple epitopes on the peptide core that crosslinks and aggregates TCR on the mucin-specific T cells [70, 71]. Both CD4+ and CD8+ T cells recognize MUC-1 epitopes in an HLA-unrestricted manner and produce appropriate responses . Presence of low level of MUC-1 antibodies in the normal individuals suggests that precursors of HLA-unrestricted anti-MUC-1 CD4+ T cells already exist in the peripheral blood and get activated once MUC-1 is overexpressed in cancers .
Conclusion and perspectives
Despite numerous investigations, the exact role of mucins in immune regulation is not fully elucidated, partly due to diversity of mucin molecules and heterogeneity in functions. Attempts using MUC-1-dependent vaccines evolved over the years with the advent of knowledge on tumour immunomodulation by mucins and by the partial clinical failures associated with the development of tolerance. From simple MUC-1-immunodominant peptide or variable repeat (VNTR) vaccines it has graduated to employ recombinant mucin peptides engineered with glycomoieties, mannan-MUC-1 fusion protein (MFP)-pulsed dendritic cell-based vaccines (to activate T cells), and MUC-1 tripatriate vaccines, having multiple components such as immunoadjuvant Pam3CysSK4, a peptide Thelper epitope and an aberrantly glycosylated MUC-1 peptide, MUC-1+/CEA+ tumour cell – DC fusion vaccines (for CTL induction), synthetic multimeric Tn/STn MUC-1 glycopeptides (to override tolerance) or MUC-6-Tn glycoconjugates, and adaptive and passive immunization protocols employing ex-vivo expanded tumour-specific T cells exposed to MUC-1 peptides/MUC-1-expressing cell lines. Of late, tumour-targeting approach using chimeric Ag receptors (CAR)-grafted T cells (that can anchor to tumour-associated MUC-1-expressing cells) and its variants such as nanobody-engineered CAR, and flexible multi-valent constructs have been used for better efficiency [7, 73, 74]. These developments in vaccinations mirror the tumour immunoprotective challenges and opportunities that the mucin-expressing cancers provide. Further, induction of MUC-1 and MUC-1-dependent oscillations of calcium signalling in immune cells and its association with phenotypic alterations of T cells, especially to a T-reg type, requires a complete investigation. Besides the interface between mucin and immune cells goes well beyond the immediate cellular milieu of the cancer and the net of interactions both within and away from the cancer decides the outcome of the immune response.
One of the authors (AAK) is grateful to CSIR for NET-JRF/SRF Fellowship.