The word mucin (derived from the Greek word for slimy) was originally used to refer to the glycoproteins in mucus lining the surfaces of glandular epithelial cells. In the 1970s, the best-characterized mucins were from ovine and bovine submaxillary glands. These mucins are composed of >50 % carbohydrates by weight and are O-linked to the protein core through serine and threonine residues [1]. The high O-linked carbohydrate content results in a dense, hydrophilic, negatively charged coat. This hydrophilic coat allows these proteins to form a viscoelastic gel that protects epithelial surfaces.

By the 1980s, mucins had been found to be the major components of many types of mucus, including epithelial surfaces of the respiratory, gastrointestinal, and genitourinary tracts; in sebaceous and salivary glands; and in the human milk fat globule (HMFG) membrane of breast milk [24]. Mucins were also found to be secreted in breast and pancreatic cancers, and thus monoclonal antibodies were developed to further characterize these tumor-associated mucins [57]. However, research into mucins using direct biochemical techniques such as sequencing was limited because of their large size and the high number of glycosidic side chains.

A major breakthrough in the understanding of mucin structure occurred in 1990, when 4 independent groups isolated cDNA clones from mammary and pancreatic mucins that encoded an identical protein core. The core of this protein, now called mucin 1 (MUC1), was sequenced, followed by the entire protein sequence [813]. The cloning of other mucins soon followed and, to date, 21 different mucins have been identified (MUC1-MUC21), which have similar protein cores, although the genes encoding these mucins are found on different chromosomes.

All mucins have regions encoding N-terminal protein sequences, followed by sequences containing a variable number of tandem repeats (VNTR), a transmembrane region, and a cytoplasmic tail. Variations in exon splicing result in different forms of these molecules, such as secreted and truncated forms [13]. There are 2 distinct groups of mucins, transmembrane mucins (MUC1, MUC3A, MUC3B, MUC4, MUC12-18, MUC20, and MUC21) and secreted gel-forming mucins (MUC2, MUC5AC, MUC5B, MUC6-8, MUC19).

The mucin VNTR region and the whole protein are rich in serine and threonine, which are the sites of O-glycosylation, and have few asparagines, which are the sites of N-linked glycosylation [14]. MUC1 contains 35 % serine and threonine residues at the N-terminal region, 25 % in the VNTR region, and 50 % at the C-terminal region [15]. The heavy glycosylation gives mucins a high density, hydrodynamic volume, and viscosity necessary for some of their biological functions, such as lubrication and protection of the gastrointestinal tract [16]. Although the VNTR regions of all mucins are rich in serine, threonine, and proline, there is little crossreactivity of the antibodies to different mucins.

The polypeptide backbone of MUC1 is synthesized in the rough endoplasmic reticulum as one large precursor containing only N-linked glycans, and carbohydrates are then added sequentially by a series of specific glycotransferases in the Golgi apparatus [15, 16]. Once fully glycosylated, MUC1 is proteolytically cleaved into alpha- and beta-subunits. The alpha-subunit consists of the extracellular domain and contains the VNTR region, and the beta-subunit consists of the cytoplasmic and transmembrane domains. The alpha and beta subunits remain associated in a non-covalent manner within the membrane [15]. This tight, non-covalent association between the subunits is not broken by treatment with low pH, boiling, high salt, peroxide, sulfhydryl reduction, or urea, but treatment with sodium dodecyl sulfate results in dissociation. The extracellular domain may be shed from the cell via an unknown mechanism, leading to secreted MUC1.

MUC1, found on chromosome 1q21, has features of an integral membrane protein and is a member of a family of structurally related, mucin-like glycoproteins, including the hemopoietic and transmembrane glycoprotein sialomucins CD34 (L-selectin ligand), CD43 (leukosialin), CD162 (PSGL-1; a ligand for P-selectin), CD164, glycosylation-dependent cell adhesion molecule 1 (GlyCAM1), and mucosal vascular addressin cell adhesion molecule 1 (MAdCAM–1) [17].

MUC1 nomenclature

Many different names have been used for MUC1 over the last 25 years, including epithelial membrane antigen (EMA), H23AG, H23 antigen, Krebs von den Lungen-6 (KL-6), episialin, polymorphic epithelial mucin (PEM), cell membrane-associated polymorphic mucin, tumor-associated epithelial membrane antigen, tumor-associated mucin (PEMT), DF3 antigen, breast carcinoma-associated antigen DF3, PAS-O, MAM6, mucin-1, Muc1, MUC-1, MUC1/REP, MUC1/TR, transmembrane mucin 1, MUC1 apomucin, mammary serum antigen (MSA), human milk fat globule antigen (HMFG), CAM 123-6, polymorphic urinary mucin (PUM), peanut-reactive urinary mucin, and the commonly used serum markers for breast cancer CA15.3 and CA27.29.

MUC1 (humans) and Muc1 (other species) are the most common name used, as assigned at the 1st International Workshop on Carcinoma-Associated Mucins in San Francisco (USA) in 1990, and in accord with the Human Genome Project mapping conventions [18]. At the 7th International Workshop on Human Leucocyte Differentiation Antigens in Harrogate (UK) in 2000, MUC1 was assigned as a cluster of differentiation (CD) molecule, CD227 [19].

MUC1 structure

MUC1 is a large cell surface glycoprotein rich in O-linked glycosylation sites in its extracellular domain. MUC1 has a molecular weight of 120–225 kDa, which is increased to 250–500 kDa with glycosylation (Fig. 1).

Fig. 1
figure 1

Schematic representation of MUC1. The extracellular region contains the large variable number of tandem repeats (VNTR) region, which contains O- and N- linked glycosylation sites. The VNTR consists of a nearly identical sequence of 20 amino acids (aa) (the most common sequence is PAHGVTSAPDTRPAPGSTAP) repeated 25–125 times (for a total of 500–2,500 aa). The MUC1 VNTR region includes an immunodominant hydrophilic knob (SAPDTR) within the hydrophilic beta turn PDTRPAPGST. An extended spacer region is shown, which spans aa GSTAPPAHG. Seven variants of the consensus VNTR sequence are shown. Because of its rigid structure, MUC1 extends 200–500 nm far above the 10 nm glycocalyx. In addition to the VNTR region, the extracellular region of MUC1 also contains the 104 aa N-terminus to the VNTR and 228 aa C-terminus to the VNTR. Proteolysis at the glycine and serine residues shown connects the extracellular (alpha) and intracellular (beta) subunits, creating the heterodimer structure. The intracellular subunit includes 28 aa spanning the transmembrane region and a 72 aa cytoplasmic tail. The cytoplasmic tail includes tyrosines (shown together with some of their binding partners, kinases and adapter proteins) that are phosphorylated, leading to signal transduction events. Other non-tyrosine residues are shown with their binding partners, including protein kinase C (PKC)-delta, GSK3beta, and beta-catenin. The inbox shows a schematic representation of the secondary structure of MUC1

MUC1 is a type I transmembrane heterodimer composed of 2 subunits. The alpha (larger) subunit is extracellular and consists of the N-terminus (104 amino acids), a 20 amino acid VNTR segment, which is repeated 25–125 times depending on the individual polymorphism, and a C-terminus of 170 amino acids. Because each VNTR segment contains 5 prolines and 5 potential O-linked glycans (via serine and threonine residues), MUC1 was predicted to exist as a rod-like structure protruding 200–500 nm above the plasma membrane, far above all other membrane-associated proteins within the 10 nm glycocalyx [15]. This was confirmed by electron microscopy of purified MUC1 and by visualizing the molecule at the cell surface [15] (Figs. 1, 2).

The alpha-subunit is non-covalently bound to the beta (smaller) transmembrane subunit, which consists of a small extracellular region of 58 amino acids, a transmembrane region of 28 amino acids, and a cytoplasmic tail of 72 amino acids (Figs. 1, 2). The transmembrane domain anchors MUC1 to the apical surface of epithelial cells. The cytoplasmic tail is homologous in widely divergent species and has tyrosine phosphorylation sites. Phosphorylation can be enhanced by okadaic acid treatment, indicating that there is a high potential for phosphorylation of MUC1 [20].

Fig. 2
figure 2

Schematic representations of MUC1 isoforms. The full-length transmembrane protein, also known as MUC1/REP and MUC1/TM, consists of two subunits. The alpha-subunit (MUC1–N) is the extracellular domain and is non-covalently bound to the beta-subunit (MUC1-C), which is mainly intracellular. The MUC1 alpha-subunit contains a cleaved-off signal peptide sequence, followed by an N-terminal 30 amino acid (aa) sequence (N30), the N74 sequence that flanks the VNTR region at the N-terminal, the large VNTR region, a 108 aa sequence that flanks the VNTR region at the C-terminal, and a final 62 aa region. The small extracellular domain of the MUC1 beta-subunit contains 58 aa, followed by the transmembrane region (28 aa) and the cytoplasmic tail (72 aa). The 62 aa at the C-terminal of the alpha-subunit and 58 aa of the extracellular region from the beta-subunit form the 120 aa sea urchin sperm protein–enterokinase–agrin (SEA) module. The MUC1 cytoplasmic domain (MUC1-CD) is shown. The MUC1/SEC (secreted) isoform includes MUC1-N and a novel 11 aa sequence (SVIGLSFPMLP). MUC1/Y does not undergo cleavage, and consists of the signal peptide, N30, part of the SEA domain (not checked region), a transmembrane region, and the cytoplasmic tail, and has a 9 aa sequence at the join region between MUC1-N and MUC1-C. The MUC1/X isoform is completely cleaved, forming Xalpha- (signal peptide, N30, and 62 aa C-terminus) and Xbeta- (58 aa extracellular region, transmembrane region, and cytoplasmic tail) subunits. MUC1/ZD consists of only 73 aa, N30 and 43 novel amino acids. MUC1* is membrane tethered, similar to MUC1-C, with only 45 aa in its extracellular domain. MUC1/SEC is a ligand for MUC1/Y and initiates signal transduction

The 120 amino acid sea urchin sperm protein-enterokinase-agrin (SEA) module is the extracellular region adjacent to the transmembrane domain and contains cleavage sites allowing release of the extracellular domain (Figs. 1, 2). The SEA module is also a component of other membrane-bound proteins such as MUC3, MUC12, MUC13, F4/80, and epidermal growth factor-like mucin-containing receptor, and is highly conserved among mucin-like glycoproteins, in particular the amino acids glycine (G) and serine (S) at the cleavage site (Fig. 1) [21].

There is little homology between human MUC1 and Muc1 from other species; only 30 % of human MUC1 and murine Muc1 molecules are identical. However, the cytoplasmic domains are highly conserved between species, and the cytoplasmic domains of human MUC1 and murine Muc1 are 90 % identical [22, 23]; thus, this region is thought to play a major functional role.

The VNTR segment of MUC1 was originally thought to consist of the exact same sequence of 20 amino acids, repeated 25–125 times; i.e., (PAPGSTAPPAHGVTSAPDTR)25-125. However, sequencing a number of proteolytic fragments corresponding to the entire VNTR region and sequencing the MUC1 VNTR region from different sources (from T47D, MCF-7, and HPAF breast cancer cell lines and from milk) demonstrated that several alternative repeats could occur. The most common variations observed were P9 to A or Q, H11G12 to VR, and D18T19 to ES (PAPGSTAPP9AH11G12VTSAPD18T19R), and the DT to ES variant was detected in 50 % of the repeats in the longer alleles [12, 24, 25] (Fig. 1).

A hydrophilic region extending through residues PDTRPAPGST [26] is the chief determinant of the secondary structure of the VNTR, and two beta-turns are predicted to occur within this region, one located at PDTR and the other within the sequence PGST (Fig. 1). Nuclear magnetic resonance (NMR) studies confirmed that the residues PDTR form a type 1 beta-turn [27, 28]. The remaining portion of the tandem repeat sequence is hydrophobic, with little to no secondary structure. This model suggests that the hydrophilic beta-turn region PDTRPAPGST is at the surface of the mucin molecule (therefore accessible to monoclonal antibodies [Mabs]), and that the rest of the VNTR region is buried within the molecule (Fig. 1). Both the MUC1 Mab C595 (anti–urinary mucin, which detects MUC1) and the Mab NCRC-11 (anti-breast carcinoma MUC1) bind to epitopes within the MUC1 VNTR region (RPAP and RPA, respectively) within the hydrophilic turn region, which indicates that this region is on the surface of the glycoprotein [29] and provides support for this overall structure.

Circular dichroism analysis of the peptide core of MUC1 VNTR indicates that it forms a beta helix, as the peptide is dominated by proline in the trans form [30], and there is no potential for beta-pleated sheet or alpha-helix formation within the 20 amino acid VNTR region [30]. In addition, intrinsic viscosity measurements of a 3-repeat (20 amino acids × 3) MUC1 peptide detected a rod shape [30] rather than a random coil conformation [31].

Mucins from cancer cells are underglycosylated relative to those from healthy cells [32], which alters their overall structure, resulting in more beta-turns. In cancer-associated MUC1 VNTR, the 7 amino acid epitope (PDTRPAP) is exposed due to premature termination of the carbohydrate side chains, and the core protein epitope (PDTRP) may be detected by the Mab SM-3, which binds to MUC1 from cancer cells but not from healthy cells. The 9 amino acid peptide (TSAPDTRPA) forms a type 1 beta-turn when unglycosylated but a rigid and extended structure in the glycosylated peptide [33]. Glycosylation had a similar effect in a 20 amino acid VNTR repeat peptide: the GVTSA region had an extended, rod-like secondary structure when the threonine was glycosylated within the GVTSAP region, whereas the APDTR region, where T was not glycosylated, formed a flexibly organized turn structure. In addition, in breast cancers, it would appear that every tandem repeat S and T are glycosylated, although with reduced length of carbohydrate chains, whereas in normal MUC1 from lactating breast, there is about 70 % glycosylation rate for the S and T [34].

In a number of models, a correlation has been shown between beta turns and increased immunogenicity; indeed, in the unglycosylated MUC1 VNTR, as seen in cancer cells, there is increased immunogenicity for both antibodies and T cells within the TSAPDTRPA region [3541]. A computer model was developed of a 60 amino acid VNTR peptide in the polyproline beta-turn helix conformation assuming that the MUC1 sequence exists in a poly-type I turn conformation [32]. The model showed a rod-like backbone with amino acids radiating outward that were completely exposed and explains why the A and DT residues (of the APDTRP epitope) allow substitutions: when the turn is formed, the P and R residues are in the same space and, therefore, accessible to antibodies [32]. Further, NMR analysis revealed a protruding knob with the immunodominant SAPDTR turn at the tip, and the knobs are connected by an extended spacer structure, which has a beta-strand and polyproline conformation [42] (Fig. 1). These studies also found that the number of repeats determined where glycosylation occurred; with a 3 tandem repeat, O-glycosylation with beta-D-N-acetylgalactosamine occurred at TS and ST rather than at the T of the APDTR [42]. Low-energy conformations of the 15 amino acid sequence PPAHGVTSAPDTRPA revealed that the peptide backbone in APDTRP and GVTSAP resembled an S-shaped bend, which is formed by consecutive inverse gamma-turn conformations partially stabilized by hydrogen bonds (Fig. 1) [43].

MUC1 isoforms

During gene expression, alternative splicing occurs, which enhances protein diversity and plays a major role in gene evolution and gene regulation. Alternative splicing determines whether some exons of a gene are included within or excluded from the final, processed mRNA, ultimately resulting in proteins that differ in their amino acid sequences and in their biological roles. In humans, approximately 95 % of pre-mRNAs which contain >1 exon are alternatively spliced, which allows for physiological adaptation of cells. Under normal conditions, splice variants that interfere with protein function are eliminated by post-transcriptional quality control mechanisms; however, alternative splicing has been shown in a large number of genetic disorders and in cancer [44]. A number of MUC1 isoforms have been identified by RNA studies and several forms were identified by antibodies which only a few are currently available.

The best known MUC1 isoform is the entire MUC1 molecule (MUC1/TR or MUC1/REP). MUC1 contains 7 exons; the alpha-subunit is encoded by exons 1-4 and the beta-subunit is encoded by exons 4–7. The VNTR region is encoded by exon 2. The best known isoforms lacking the VNTR are MUC1/X, MUC1/Y, MUC1/Z, MUC1/ZD, MUC1*, MUC1/A, and MUC1/B [45] (Fig. 2). More than 80 other alternatively spliced variants that encode different isoforms of MUC1 have been identified, although the full-length sequences have only been reported for some. Alternative splicing also results in a secreted MUC1 form (MUC1/SEC), where the extracellular domain (MUC1-N) is dissociated from the intracellular domains (MUC1-C). MUC1/SEC contains the VNTR region but is missing the transmembrane region and the cytoplasmic tail.

The MUC1 variants MUC1/A, MUC1/B, MUC1/X, MUC1/Y, and MUC1/Z are more frequently expressed in malignant tumors than in benign tumors. Seventy-eight MUC1 isoforms were isolated from human ovarian and breast cancer cell lines and from human activated T cells [46]. All but 2 isoforms were newly identified, and most were short isoform proteins, one of which (MUC1/Y-LSP) was shown to inhibit the growth of tumors in vivo [46]. The precise role of all the MUC1 isoforms in cancer and other disease and in normal cells is not clear.


The MUC1 protein is expressed and secreted by a variety of cells. The secreted MUC1 isoform, MUC1/SEC, terminates prematurely at a stop codon within intron 2 and thus is missing the transmembrane and cytoplasmic regions [47]. MUC1/SEC contains a short stretch of 11 hydrophobic amino acids (SVIGLSFPMLP) located 48 amino acids upstream of the C-terminus (Fig. 2). From its length and hydropathicity profile, it was concluded that this region did not act as a transmembrane domain, and that MUC1/SEC was secreted from the cell. Mabs (SEC-1, SEC-2, and SEC-3) produced to the 14 amino acid peptide sequence strongly bound to MUC1/SEC from breast and colon cancer cells, to sera of breast cancer patients, and to human milk [4749]. All primary ovarian cancers are positive for MUC1, whereas the variant MUC1/SEC is negative in ovarian cancers [50]. All variants tested were expressed in 8 cervical carcinoma cell lines. MUC1/SEC mRNA was present in endometrial carcinoma cell lines, endometrial tissue, primary cultured endometrial epithelial cells, and uterine flushings [51].


The 1.2-kb splice variant MUC1/Y is identical to MUC1 (traditional or transmembrane MUC1) in the transmembrane and cytoplasmic domains, but is missing the VNTR region. Mabs have been generated to the peptide TEKNAFNSS (“joining peptide”) that is derived from amino acids 58–62 of the N-terminus sequence (N30) (TEKNA) and from amino acids 343–346 of the C-terminus sequence (FNSS) (Fig. 2) [48, 49]. These antibodies (SP3.9, SP11, and IgG1) reacted with normal breast tissue (from ductal secretions) and with cancer tissue (gastric carcinomas), but not with normal colon, pancreas, salivary gland, stomach, kidney, liver, or lung [48, 49]. MUC1/Y is also overexpressed in ovarian, breast, pancreatic, and stomach carcinomas [52, 53], and in ductal carcinoma in situ of the breast [54]. The Mab 6E6/2 bound to the surface of MCF-7 breast cancer cells and to malignant epithelial cells from breast and ovarian cancer patients [52]. Although the full-length MUC1 is not expressed in HeLa cervical cancer cells, MUC1/Y is highly expressed in HeLa cells and in other cervical cancer cell lines [45].

MUC1/Y contributes to tumor initiation and progression in vivo, and it structurally resembles the type II cytokine receptors (with tyrosine phosphorylation sites in the cytoplasmic domain) that are known to be involved in ligand binding [55, 56], indicating that MUC1/Y may act in a similar manner to cytokine receptors. MUC1/Y undergoes tyrosine/serine phosphorylation and binds to Grb2, initiating cell signaling. In addition, MUC1/SEC binds to MUC1/Y (and full-length MUC1), which initiates phosphorylation of MUC1/Y [57]. Furthermore, MUC1/Y naturally forms a complex with MUC1 and MUC1/SEC in conditioned media. MUC1/SEC binds to the amino acid sequence INVHDVETQF [57] within the 44 amino acid stretch at the N-terminal to the transmembrane region. The MUC1/SEC binding site on MUC1/Y is homologous to the ligand binding sites of cytokine receptors, suggesting that MUC1/Y functions in a similar manner to cytokine receptors and is regulated by a secreted ligand.


Like MUC1/Y, MUC1/X is missing the VNTR region and results from alternative splicing at an alternative splice acceptor site located 18 amino acids upstream to the site that produces MUC1/Y [58]. The extracellular domain of MUC1/X consists solely of the SEA module in addition to 30 amino acids at the N-terminus (N30), and undergoes proteolytic cleavage at the same region as the full-length MUC1 (Fig. 2) [59] [60]. MUC1/X is highly expressed in cervical and ovarian cancer cells.


The MUC1/ZD mRNA is generated by alternative splicing at the same site that produces MUC1/Y. However, downstream to the splice acceptor site, the MUC1 reading frame has undergone a +1 reading frameshift, leading to a completely new C-terminal protein sequence of 43 amino acids, different from all other known MUC1 proteins [58]. The sequence TTTKSCRETFLK is unique to the MUC1/ZD 43 amino acid C-terminus and is part of the joining region. MUC1/ZD also contains 30 amino acids from the N-terminus (N30) and thus consists of 73 amino acids in total (Fig. 2). Using Mabs (ZUM7E7 and ZUM12D8) to the unique sequence, MUC1/ZD was found to be present in fluids aspirated from the sites of breast lumpectomies, breast cancer tissues, sebaceous glands, hair follicles, and epithelial cell layers of the skin, but not on fibroblasts [59]. Expression of MUC1/ZD does not correlate with expression of full-length MUC1, suggesting that a specific combination of MUC1 isoforms is expressed depending on the particular needs of the cell at any time.

MUC1/A and MUC1/B

MUC1/A expression is reduced in dry eye disease, and there is differential expression of MUC1/A and MUC1/B in prostate, ovarian, cervical, and breast cancer [6164]. The mechanism by which MUC1/A and MUC1/B may contribute to the pathogenesis of dry eye disease and cancer is unknown. The MUC1/A splice variant has 9 additional amino acids in the N-terminal domain that are not present in MUC1/B, and these variants differently regulate expression of the inflammatory cytokines tumor necrosis factor (TNF)-alpha, interleukin (IL)–1 beta, and IL-8 [63].


MUC1* is a membrane-bound MUC1 cleavage product that contains only 45 amino acids of the MUC1 extracellular domain (not including the VNTR region) and is highly expressed in cancer cells (Fig. 2). MUC1* also consists of a cytoplasmic tail and a transmembrane domain. The exact cleavage site is unknown [65]. MUC1* has growth factor-like activity where binding to its ligand NM23 stimulates cell growth, and is involved in the growth and differentiation of human embryonic stem cells.

MUC1-N, MUC1-C, and MUC1-CD

MUC1 undergoes autoproteolytic cleavage within the SEA module to form 2 non-covalently associated polypeptides, MUC1-N and MUC1-C [66]. MUC1-N (alpha-subunit) is the MUC1 N-terminal subunit, which contains the VNTR region. MUC1-N has anti-adhesive and lubricative properties (see below) and is shed from the cell surface following proteolytic cleavage (Fig. 2). MUC1-C (beta-subunit) is the transmembrane MUC1 C-terminal subunit that spans the cell membrane and consists of a 58 amino acid extracellular domain, a 28 amino acid transmembrane domain, and a 72 amino acid cytoplasmic domain (Fig. 2). MUC1-C is often referred to as MUC1 oncoprotein. MUC1–CD is the MUC1-C cytoplasmic domain. The cytoplasmic domain of MUC1-C consists of a CQC motif, which forms dimers in human breast and lung cancer cells that are disrupted by reducing agents [67].

The MUC1-C extracellular domain binds with galectin-3, forming a complex with epidermal growth factor receptor (EGFR) on the cell surface (Fig. 3). The MUC1-C cytoplasmic domain interacts with other key signaling molecules, including beta-catenin, p53, and NFkappaB, and as a result, the MUC1-C cytoplasmic domain is phosphorylated and activates signaling pathways [68]. MUC1-C is often transported to the mitochondria and to the nucleus, where it is involved in the regulation of transcription. Only MUC1-C and not MUC-N was expected to translocate to the nucleus, within the nuclear matrix, nucleoli, and the nuclear periphery [69]. However, recent studies indicate that MUC1-N is also found within the nucleus of normal epithelial tissues and in cancer cell lines [69]. In particular, MUC1-N is found in nuclear speckles (interchromatin granule clusters) and associates with spliceosomes. The different intranuclear distribution patterns of MUC1-N and MUC1-C correlate with their different functions.

Fig. 3
figure 3

Schematic representations of MUC1 function and interactions. The MUC1 extracellular domain contains numerous sialic acids, giving it a strong negative charge. As a result, the alpha-subunit has anti-adhesive properties, promoting escape of tumor cells and metastasis. The negative charge also prevents pathogens from binding to the mucosa. MUC1 inhibits binding to type-I and type-IV collagen, laminin, and fibronectin as well as integrin-mediated cell adhesion. Substances found to upregulate or downregulate MUC1 are shown. The function of the alpha-subunit is to hydrate, protect, and lubricate cell surfaces. Intercellular adhesion molecule 1 (ICAM-1) and E-selectin on endothelial cells interact with the alpha-subunit of MUC1, which increases metastasis and lymph node invasion of tumor cells. MUC1 also interacts with sialoadhesin (Siglec 1), DC-SIGN, C-type lectin receptor, the mannose receptor (MR), macrophage galectin lectin (MGL), anterior gradient 2 (AGR2), myelin-associated glycoprotein (MAG or Siglec-4a), MAL2 (a protein with 4 transmembrane regions), fibroblast growth factor receptor (FGFR), ErbB2 (Her-2/neu), galectin-3, and the epidermal growth factor receptor (EGFR), resulting in various functions. The main functions of the beta-subunit are to aid in tumor migration, survival, cell proliferation, apoptosis, mitogenesis, and angiogenesis. Galectin-3, ErbB2, estrogen receptor alpha (ERalpha), p53, GSKbeta, P120, heat shock proteins (Hsp 70 and 90), Src, and beta-catenin interact with the cytoplasmic tail, leading to cell signaling

MUC1-C is involved in lipid metabolism and in the upregulation of genes responsible for cholesterol and fatty acid synthesis [70]. MUC1-C is also involved in the regulation of glucose uptake and lactate production, indicating that MUC1-C promotes glycolysis in breast cancer cells [71]. Synthetic cell-penetrating peptide inhibitors (GO-201 and GO-203, containing a CQC motif), which target the MUC1-C oligomerization site, block the NF-kappaB p65 pathway, leading to cell death of prostate, breast, and lung cancers, multiple myeloma, and acute/chronic myelogenous leukemia cells in vitro and/or in nude mice in vivo [7178]. Similarly, the peptide inhibitor PMIP inhibits proliferation and estrogen receptor transcriptional activity in lung cancer cells [79]. Thus, disruption of the MUC1-C-terminal subunit (which is present in most isoforms) represents a novel immunotherapeutic approach for the treatment of MUC1-expressing cancers. Indeed, GO-203 has entered Phase I evaluation in patients with refractory solid tumors.

MUC1 expression

It was originally believed that MUC1 expression was restricted to virtually all mucosal epithelial tissues, and was overexpressed aberrantly in breast cancer cells [80, 81]. MUC1 is expressed in almost all epithelial tissues in the respiratory, gastrointestinal, urogenital, and hepatobiliary tracts, in sebaceous and salivary glands, and in hematopoietic cells (Table 1) [1, 82]. MUC1 has been found in most organs, including breast, esophagus, stomach, pancreas, liver, duodenum, lungs, kidney, bladder, prostate, endometrium, ovary, and testis (Table 1), but is not expressed in skin epithelial cells or mesenchymal tissues. In addition, extremely high serum MUC1 levels were found in pulmonary sclerosing hemangiomas [83] and in ductal adenomas of the breast [84]. MUC1 is also found in HMFG of breast milk [40, 85, 86], human tears [87], sweat glands, and sebaceous glands (Table 1) [88].

Table 1 Expression of MUC1 in normal cells, in cancer, and in other diseases

In human corneal and conjunctival epithelial cells, MUC1 is expressed at the apical surface and at the superficial layer of the cytoplasm (Table 1) [89]. MUC1 has a role in the stability of the tear film and has a protective role in preventing adhesion of foreign debris, cells, or pathogens to the ocular surface. MUC1 expression in the ocular surface epithelium is decreased in dry eye syndrome but is increased in Sjogren’s syndrome (an autoimmune disorder that destroys the exocrine glands that produce tears and saliva) and in squamous metaplasia, where different MUC1 variants are expressed, including MUC1/A, MUC1/SEC, MUC1/Y, MUC1/ZD, and MUC1/Y [61, 90]. MUC1 expression is upregulated during the early years of contact lens wear and is downregulated in androgen receptor-deficient patients. Mucins (including MUC1) are also expressed in the human Eustachian tube and middle ear epithelium and are upregulated in chronic otitis or mucoid otitis media [91]. In addition, MUC1 is expressed in salivary glands (parotid and submandibular glands, striated and excretory ducts, and in some serous acinar cells).

MUC1 is expressed by perineurial cells, myofibroblasts in particular (Table 1) [92]. MUC1 is also expressed in human hematopoietic stem/progenitor cells of cord blood (Table 1) [93] and mediates the growth of human pluripotent stem cells.

MUC1 also plays a role in gallstone formation (Table 1). Stone-containing gallbladders express higher levels of MUC1 than gallbladders without stones. Muc1—/− mice have decreased gallstone formation [94], and cholesterol cholelithogenesis was enhanced in transgenic mice containing the human MUC1 gene [95]. There is altered mucin gene expression and upregulation of MUC1 in black pigment stone-bearing gallbladders (with cholesterol stones and calcium bilirubinate stones). MUC1 is strongly expressed on the luminal surface of biliary epithelial cells of the small bile duct, in primary biliary cirrhosis, and in patients with chronic viral hepatitis, but MUC1 expression is infrequent in normal liver, extrahepatic biliary obstruction, and hepatolithiasis [96].

MUC1 in the reproductive system

MUC1 is present at the apical surface in uterine epithelial cells, and due to its rigid, rod-like structure, MUC1 acts as a barrier to microbial infections and enzymatic attack. Indeed, MUC1 knockout mice are more prone to bacterial infections [97].

In addition, MUC1 plays an important role during pregnancy, both in the uterus and in the embryo; MUC1 expression in the embryo downregulates MUC1 in the uterus, allowing for embryo attachment to the uterus (implantation). In the uterus, MUC1 is highly expressed on the apical surface of luminal epithelial cells throughout the menstrual cycle and during embyro implantation [98]. Muc1 expression is reduced during implantation in the luminal uterus epithelium of mice, baboons, pigs, and rats but is not altered in humans [99]. In the embryo, human blastocysts express MUC1 on the outer layer cells (trophoblasts) [100], and expression of MUC1 on trophoblast cells is involved in transendothelial blastocyst migration. When blastocysts in the non-adhesive phase were cultured with human endometrial epithelial cells, MUC1 expression levels were increased in the endometrial epithelial cells [101]. However, the opposite occurred when human blastocysts in the adhesion phase were cultured with endometrial epithelial cells: MUC1 expression was lost at the site of attachment. Furthermore, when human embryos were cultured with endometrial epithelial monolayers, MUC1 expression was lost from the epithelial cells at the point of and near embryo attachment, but MUC1 expression was normal on nearby cells [102]. MUC1 and mouse Muc1 expression were reduced at the site of embryo attachment in human MUC1 transgenic mice (20 and 98 % reduction, respectively) [103].

Injestion of arnebia, a herb used in traditional Chinese medicine to treat skin conditions, has been shown to inhibit the decrease in MUC1 expression at the point of blastocyst attachment, and thereby prevents embryo implantation [104]. In addition, in NOD mice, aberrant overexpression of IFN-gamma and MUC1 at the implantation point contributes to embryo loss in type 1 diabetes [105].

In women with unexplained recurrent pregnancy loss, it was noted that they had significantly reduced MUC1 expression in the peri-implantation endometrium compared with normal control women [107]. High levels of transmembrane and secreted MUC1 have been found to be localized to the apical surface of the tubal epithelium of the fallopian tube; thus, MUC1 may contribute to inhibiting ectopic implantation due to its anti-adhesive properties. In fact, a patient with 5 ectopic pregnancies had deficient MUC1 expression [108]. MUC1 levels are significantly reduced in patients with hydrosalpinx (fluid-filled, blocked fallopian tubes), which is associated with infertility due to reduced implantation and increased abortion rates.

MUC1 is overexpressed on epithelial cells isolated from patients with polycystic ovarian syndrome or endometriosis and on umbilical cord tissues [109, 110]. MUC1 is also expressed in the human placenta throughout pregnancy and is increased with gestational age. In severe preeclampsia placentas, MUC1 is significantly elevated (Table 1) [106].

In men, MUC1 is expressed in human testis, in the male germ cell lineage, and in ejaculated sperm. During normal spermatogenesis, MUC1 glycosylation variants are selectively distributed; pachytene spermatocytes and spermatids express high levels of MUC1 glycosylation, and MUC1 is aberrantly glycosylated during pathological conditions (Table 1) [111]. Vascular endothelial growth factor (VEGF) transgenic mice under the control of the Muc1 promoter show reduced male fertility due to impaired spermiogenesis, resulting in the rejection of male embryos. Thus, the presence of Muc1 in spermatocytes may be responsible for reduced fertility [112]. High expression of VEGF or MUC1 in human semen correlates with lack of pregnancy success following in vitro fertilization.

MUC1 on leukocytes

Nonepithelial expression of MUC1 has been described to occur on leukocytes, including T cells [113117], B cells [118], bone marrow mononuclear cells, monocytes, and dendritic cells (DCs) (Table 1) [117]. MUC1 is not expressed on resting or activated natural killer (NK) cells. Monocyte-derived DCs, murine bone marrow-derived DCs, murine splenic DCs, follicular DCs, and major histocompatibility complex (MHC) class II- skin Langerhans cells express MUC1 following in vitro maturation (Table 1) [125]. MUC1 is not expressed on blood DCs, thymic DCs, tonsil DCs, or MHC class II + skin Langerhan cells [125]. DCs isolated from synovial fluid from patients with arthritis express MUC1 [125].

Activated human T cells express MUC1 at the cell surface, but the level of expression is significantly lower than in normal epithelial cells or tumor cells [119]. The glycosylation of MUC1 in activated T cells is similar to normal epithelial cells, which is vastly different to the glycosylation in tumor cells. Anti-MUC1 peptide cytotoxic T lymphocytes (CTLs) are not able to lyse autologous MUC1-expressing activated T cells [120]. The molecular weight of MUC1 in resting human T cells is lower (250 kDa) than in activated T cells (>300 kDa). The MUC1 protein core and sialated epitopes are the same in cancer cells and in T cells, but there is a peptide region that is exposed on cancer cells due to aberrant glycosylation that is not exposed on activated T cells [116]. This suggests that the MUC1 expressed on activated T cells is different from that expressed by tumor cells.

In addition, MUC1 on activated T cells has an even distribution over the entire surface; however, when T cells are polarized in response to the migratory chemokine RANTES, MUC1 changes morphologically and is isolated to the leading edge of T cells. This differential distribution of MUC1 suggests that MUC1 may be involved in early interactions between T cells and endothelial cells.

MUC1 is a novel T cell costimulatory molecule and is associated with the regulation of T-cell responses on activated T cells [121, 122]. MUC1 is upregulated by IL-7 or IL-12 but not by IL–2, IL-4, IL-5, or IL-10 on CD4+, CD8+, CD25+, CD69+, CD45RA+, and CD45RO+ T cells [121, 123]. DCs also induce MUC1 expression in T cells at the DC-T cell synapse [123]. Furthermore, mitogen-activated human T cells secrete soluble MUC1, and MUC1 expression is downregulated after the mitogenic stimulus is removed.

The MUC1 cytoplasmic domain on activated T cells and DCs is tyrosine phosphorylated, indicating that it has a role in signaling [124]. When human monocyte-derived DCs were cultured in the presence of recombinant sialylated MUC1 glycoprotein (from cancer cells), expression of costimulatory molecules (CD40 and CD86), antigen-presenting molecules (CD1d and MHC), and the differentiation marker CD83 was decreased [126], while expression of the immature DC phenotype markers CD1a and the mannose receptor (CD206) was increased. These DC phenotypic changes correlated with defective T-cell stimulation, and hence may be a mechanism by which tumors escape immunosurveillance. Conversely, immature human myeloid DCs bound to the aberrantly glycosylated MUC1 found in cancer cells, which increased the cell surface expression of CD40, CD80, CD83, and CD86 and production of IL-6 and TNF-alpha, but did not induce T1 immune responses [127]. These data suggest that, in cancer patients, MUC1 binds to immature DCs and limits their ability to stimulate T1 T-cell responses. Muc1-knockout DCs were highly activated compared with Muc1-positive DCs, exhibiting higher expression of CD40, CD80, and CD86; increased secretion of TNF-alpha and VEGF; and stimulation of CD4+ T cells. Thus, Muc1 expression generates a suppressive DC phenotype [128]. In addition, Muc1 enhanced activation of DCs in response to toll-like receptor (TLR)4 and TLR5 receptor ligands.

Myeloid-derived suppressor cells (MDSCs) are characterized at different stages of differentiation as early myeloid progenitors, immature granulocytes, macrophages, and DCs, and these cells suppress cytotoxic activities of NK cells, CD4+, and CD8+ T cells and have tumor suppressive activities (Table 1). MUC1/SEC but not the MUC1 transmembrane form is associated with markedly lower levels of MDSC recruitment in vitro, and blocks the expression of arginase 1 and production of reactive oxygen species, which are involved in the suppression of T cells [129]. In a Muc1 knockout mouse model, Muc1 controlled the development of MDSCs, which directly affected tumorigenesis [130]. An understanding of how MUC1 regulates the development of myeloid progenitors into MDSCs would aid in the development of improved anti-MUC1 vaccines for immune suppression of tumors.

MUC1 is expressed in hematopoietic lineages. In bone marrow differentiating cells, MUC1 is expressed during erythropoiesis, on proerythroblasts and erythroblasts, and weakly during megakaryocytopoiesis and granulomonocytopoiesis, but is absent in circulating erythrocytes [131, 132]. During erythropoiesis, MUC1 is phosphorylated, indicating that there may be crosstalk between MUC1 on erythroblasts and other cells during erythropoiesis. Moreover, MUC1 is expressed on the surface of monocytes and monoblasts and in acute myeloid leukemia. The role of MUC1 on leukocytes (T cells, B cells, DCs, etc.) still requires study to understand its role in these cells.

Regulation of MUC1

MUC1 expression may be increased by a number of factors. Anterior gradient 2 (AGR2), a transforming growth factor (TGF)-beta responsive gene in human pancreatic carcinomas, is coexpressed with MUC1, and loss of SMAD4, which downregulates AGR2, converts TGF-beta from a tumor suppressor to a tumor promoter, resulting in an increase in AGR2 and MUC1 expression, leading to pancreatic adenocarcinoma (Fig. 3) [133].

On endometrial adenocarcinoma cells, MUC1 expression is stimulated by estradiol or raloxifene and is downregulated by tamoxifen or the combination of estradiol and tamoxifen or raloxifene [134]. MUC1 expression is upregulated by interferon (IFN)-gamma or TNF-alpha in several human cell lines, including epithelial, breast cancer, mammary epithelial, lung epithelial, nasal mucosa, and hematopoietic cell lines [135, 136]. IL-7 has been shown to stimulate MUC1 expression in multiple myeloma cells, and IL-24 activates MUC1 on colonic epithelial cells and in inflammatory bowel disease (Fig. 3). Neutralization of IL-10 secreted by monocytes/macrophages enhances IL-6 secretion and MUC1 expression on human colon cancer cells and is involved in cancer progression.

Fibroblast growth factor (FGF)-2 and -10 significantly stimulate MUC1 mRNA and protein levels on rat conjunctival epithelial cells, and retinoic acid, dexamethasone, and rebamipide increase MUC1 expression on human corneal epithelial cells [137139]. In addition, neutrophil elastase stimulates MUC1 gene expression in lung epithelial cells [140]. MUC1 is also regulated by progesterone, but not by estrogen or IL-1-beta (Fig. 3). Corticosteroids have been shown to increase MUC1 expression on nasal polyps as early as 2 weeks after exposure in humans.

In prostate cancer cells, MUC1 protein expression is regulated by the androgen receptor [141]. GATA3, a transcription factor that regulates luminal epithelial cell differentiation in the breast, has been shown to mediate the upregulation of MUC1 expression in breast cancers. Furthermore, human breast cancer cells express increased levels of MUC1 in the presence of ethanol (100 mM; the concentration reported to be present in alcoholics) [142]. The contraceptive microbicide nonoxynol-9, widely used for contraception, has been shown to decrease MUC1 protein and mRNA levels [143], indicating that nonoxynol-9 interrupts the protective barrier of MUC1 in the endometrium, which can lead to infections.

MUC1 function

Mucins function to hydrate, protect, and provide lubrication to mucosal and epithelial luminal surfaces of ducts (Fig. 3). Mucins are thought to protect against fluctuations in pH and osmolarity found in the epithelial environment, such as the bladder epithelium (urine), sweat, and milk gland epithelium. MUC1 expression is restricted to the apical surface facing the lumen of ducts, which lubricates epithelial surfaces and traps cellular debris. In milk, MUC1 stabilizes fat emulsions in the digestive tract until milk fat is hydrolyzed. MUC1 in milk also acts as a decoy for binding of infective agents such as bacterial and viral pathogens [144, 145].

During cancer, MUC1 expression is altered from the restricted apical surface to all over the cell surface, and this aberrant expression may play a role in metastasis [15]. Further, MUC1 inhibits efficient lysis by neutrophils, NK cells, and cytotoxic T cells, due to its extended and rigid rod-like structure extending high above the cell surface. As a consequence, MUC1 limits adhesion, allowing tumor cells to escape and thus leading to cancer metastasis [146]. MUC1 is also rich in sialic acid [147] residues, giving it a strong negative charge that may add to the anti-adhesion effect by charge repulsion (Fig. 3). Indeed, bovine Muc1 inhibits E. coli and S. typhimurium binding to human intestinal cells, and inhibition is due to sialic acids [148]. In addition, these sialic residues may make the glycans more bulky and thus may contribute to the rigidity of mucins.

MUC1 has a strong anti-adhesion effect on in vitro cell lines, as the aggregating capacity of MUC1 + transfected cells was strongly reduced compared with non-transfected cells [147]. This was confirmed by electron microscopy studies, which showed that MUC1-transfected cells from tumors in nude mice established only loose cell–cell contacts, which is the prerequisite for tumor metastasis. This suggests that MUC1 may have the same anti-adhesive properties in vivo. Furthermore, transfection of human pancreatic and gastric cells with MUC1 decreased adhesion to type I collagen, type IV collagen, fibronectin, and laminin, but increased motility and in vitro invasiveness [149]. In vivo, MUC1-transfected human gastric cancer cells grew considerably faster than non-transfected cells and invaded into the muscle layer, whereas the non-transfected cells did not [150]. Inhibition of MUC1 in a human pancreatic cell line resulted in slower proliferation and increased binding to collagen IV and fibronectin in vitro, and reduced lymph node metastasis of pancreatic cancer cells in vivo [151]. It is clear that overexpression of MUC1 promotes cancer progression and metastasis, and data suggest that MUC1 is involved in the oncogenic process.

The hydrophilic extracellular VNTR region of MUC1 and the high glycosylation are thought to prevent hydrophobic chemotherapeutic drugs from passing through the membrane and into the cell (Fig. 3). Indeed, MUC1 expression on circulating metastatic breast cancer cells correlated with a lack of treatment effectiveness, whereas the response rate was significantly higher and progression-free survival was longer in patients with MUC1- circulating cancer cells [152]. Furthermore, progression-free survival was also longer in patients with decreased levels of circulating MUC1 protein following chemotherapy. Thus, patients lacking MUC1 + tumor cells respond better to chemotherapy.

Gemcitabine downregulates MUC1 in a pancreatic cell line in vitro, suggesting that gemcitabine would be useful in MUC-expressing adenocarcinomas. Indeed, in mice the use of gemcitabine in combination with MUC1-based vaccines has been shown to be useful for the treatment of pancreatic cancer [153, 154]. In a human Phase IIb trial, response rates were increased in patients with advanced non-small cell lung carcinoma treated with TG4010 (vaccinia virus encoding MUC1 and IL-2) and gemcitabine [155].

A recent study showed that MUC1 modulates the hypoxic response in carcinoma cells, by regulating the expression of hypoxia inducible factor-1alpha, as shown in a pancreatic cancer cell model [156]. In addition, MUC1 binds with hypoxia inducible factor-1alpha and p300 and acts as a metabolic regulator, leading to hypoxic environments that help tumors survive and grow.


MUC1 binds to C-type lectin receptors, DC-SIGN (CD209), the mannose receptor, and macrophage galactose lectin (MGL), which are expressed on macrophages and DCs [157, 158]. Binding of MUC1 to the mannose receptor results in uptake into early endosomes, where MUC1 is retained long term without degradation, which prevents processing and presentation of the MUC1 antigen (Fig. 3) [157]. For this reason, to stimulate an immune response, MUC1 should be presented to DCs in a form that can allow MUC1 to be better processed, such as conjugation to mannan [38].

Sialoadhesin (CD169, Siglec-1) is a cell adhesion molecule expressed by macrophages that binds to neutrophils, monocytes, NK cells, B cells, and T cells. Sialoadhesin also binds to MUC1, as demonstrated using human breast cancer cells and erythroid cells (Fig. 3) [159]. Myelin-associated glycoprotein (MAG, Siglec-4a), a membrane-bound protein expressed on oligodendrocytes and Schwann cells that aids in the binding of myelin to neurons, also interacts with MUC1 [160]. The MAL2 protein, a component of lipid rafts, transports membrane-bound proteins to the apical surface. In breast and ovarian cancers, MAL2 is overexpressed and interacts with MUC1, suggesting that MAL2 may be involved in regulating MUC1 expression and its localization (Fig. 3) [161].

MUC1 induces galectin-3 expression by post-transcriptional mechanisms, and the glycosylated Asn-36 amino acid on MUC1-C is responsible for galectin-3 expression [162]. In turn, galectin-3 binds to MUC1-C at the glycosylated Asn-36, resulting in a regulatory loop. The interaction of MUC1 with galectin-3 has been shown to promote metastasis. Galectin-3 is markedly increased in the circulation of cancer patients, leading to increased adhesion of MUC1-positive cancer cells to blood vessel endothelial cells, promoting embolus formation and survival of disseminating tumor cells in the circulation and promoting metastasis [163, 164]. Moreover, serum levels of other galectins, particularly, galectin-2, -4, and -8, are increased up to 30 fold in colon cancer patients, particularly those with metastasis (Fig. 3) [165]. Further understanding of interactions between MUC1, galectin-3, and the other galectins will aid in the development of effective therapeutics for preventing metastasis.

The epidermal growth factor (EGF) family consists of 4 type I tyrosine kinase receptors, EGF receptor (EGFR or ErbB1), ErbB2/neu, ErbB3, and ErbB4, which are overexpressed in cancer and are involved in cancer initiation and progression. The MUC1 cytoplasmic tail interacts with all 4 ErbB family members, and promotes tumor proliferation through the activation of signaling pathways. Galectin-3 is essential for the MUC1/EGFR interaction [166, 167]. In human breast carcinoma cells, MUC1 associates with ErbB2 (also known as Her-2/neu) and treatment with heregulin increases MUC1-ErbB2 complexes.

Gamma-catenin (plakoglobin) is found in complexes with cell adhesion molecules, and its nuclear localization is dependent on MUC1 expression. The sequence SAGNGGSSL within the cytoplasmic tail of MUC1 is a binding site for beta- and gamma-catenins [168]. The level of nuclear gamma-catenin is directly correlated with activity of c-Myc and other oncogenes [169]. MUC1 binding to beta-catenin alters its ability to interact with E-cadherin in the formation of cellular adherens junctions, leading to diminished cell–cell adhesion and promoting cell migration and neoplastic progression (Fig. 3).

Intercellular adhesion molecule 1 (ICAM-1) is expressed on endothelial cells and cells of the immune system (macrophages and lymphocytes). MUC1 binds to domain 1 of ICAM-1 and mediates breast cancer cell migration (Fig. 3) [170]. ICAM-1 binds to MUC1 at the extracellular VNTR domain, which requires at least 6 repeats. In mammary epithelial cells, downregulation of ICAM-1 led to induction of MUC1 [171]. Enhanced metastasis of MUC1 + pancreatic cells was noted in nude mice, and MUC1 enhanced the invasiveness and motility in vitro of cancer cell line S2-013 and the MUC1-transfected cell line MKN74 [150]. In addition, it was shown that MUC1 is a ligand for E-selectin, a cell adhesion molecule expressed only in endothelial cells [172]. This finding demonstrates that the binding of MUC1 to E-selectin and to ICAM-1 plays a major role in the metastatic adhesion cascade. MUC1 overexpression in cancer cell lines also inhibits integrin-mediated cell adhesion to extracellular matrix components [53]. In addition, MUC1 regulates intracellular oxidant levels and the apoptotic response to oxidative stress [173].

Because of the aberrant overexpression of MUC1-C in human breast cancer cells, a complex forms between MUC1-C and the estrogen receptor on the Rab31 promoter, activating Rab31 gene expression, which is associated with a poor prognosis [174]. In in vitro studies, blocking the MUC1 C-terminal region in MUC1-expressing lung adenocarcinoma cells inhibited estrogen receptor transcriptional activity and cell proliferation [79], further demonstrating that MUC1 overexpression in cancer cells is involved in cancer progression and metastasis.

Cell signaling

MUC1 overexpression plays a role in tumorigenesis, cell adhesion, immunoregulation, and cell signaling. Overexpression of MUC1 lacking either the VNTR or the cytoplasmic tail correlates with enhanced invasiveness and a metastatic phenotype compared with overexpression of full-length MUC1 [175].

Both the MUC1 transmembrane form and MUC1/Y (which lacks the VNTR region) function in signal transduction pathways. MUC/SEC interacts with the extracellular domain of MUC1/Y, which induces phosphorylation. The 72 amino acid cytoplasmic tail of MUC1 contains 7 evolutionally conserved tyrosine (Y) residues that are phosphorylated. Phosphorylation on Y residues is the key step in signal transduction pathways mediated by membrane proteins, including MUC1. Kinases and adapter proteins have been shown to bind to 5 of the 7 Y residues in MUC1, including PI3 K (Y20), Shc (Y26), PLC-γ (Y35), c-Src (Y46), and Grb2 (Y60) (Fig. 3) [176, 177]. Following tyrosine phosphorylation, the MUC1 cytoplasmic domain interacts with SH2 domain-containing proteins and initiates a signal transduction cascade linked to the oncogenic process [176, 177].

Other proteins bind to nontyrosine sites on MUC1, including GSK3beta (S44), PKCdelta (T41), and beta-catenin (S50). Consensus sequences resembling an immunoreceptor tyrosine-based activation motif (ITAM) and an immunoreceptor tyrosine-based inhibitory motif (ITIM) are present in the cytoplasmic tail. Estrogen receptor alpha, p53, p120ctn, all ErbB members, heat shock protein (Hsp)70, and Hsp90 have been shown to bind to the MUC1 cytoplasmic tail (Fig. 3). The interaction between MUC1 and EGFR regulates Grb2/Sos/Ras-MEK-ERK2 and beta- and gamma-catenin signal pathways, which are associated with malignant transformation, prevention of apoptotic mechanisms, invasion, and metastasis (Fig. 3). Disruption of the signaling pathways of MUC1 may have therapeutic applications in the treatment of MUC1-expressing adenocarcinomas; indeed, the first MUC1-C inhibitor has entered Phase I clinical trials as a potential treatment for patients with breast cancer [68].

It is clear that MUC1 plays important biological roles in cell–cell interactions, cell–matrix interactions, modulating tumor progression, metastasis, and cell signaling [178182]. Further work is required to understand the molecular mechanisms and signaling pathways that control the aberrant expression of MUC1 underlying tumor progression in adenocarcinoma.

MUC1 and pathogens

MUC1 serves as an adhesion molecule for a variety of pathogens and is thought to provide protection from bacterial and chemical injury (Fig. 3).

The Muc1 ectodomain, when transfected into Chinese hamster ovary cells, was shown to serve as a binding site for Pseudomonas aeruginosa (P. aeruginosa) [183]. NEU1 sialidase is involved in the interaction between P. aeruginosa and MUC1 [167]. In addition, binding of P. aeruginosa to MUC1 was blocked by anti-flagellin antibody or purified flagellin. Thus, flagellin is an adhesin of P. aeruginosa responsible for its binding to MUC1 on the epithelial cell surface.

MUC1 also has an anti-inflammatory effect. Muc1 knockout mice (Muc1-/-) infected with P. aeruginosa showed reduced lung colonization and increased leukocytes, TNF-alpha, and IL-8 in bronchoalveolar lavage fluids compared with control mice [184]. When treated with flagellin, alveolar macrophages or tracheal epithelial cells showed higher levels of TNF-alpha and IL-8 in Muc1-/- mice compared with control mice, indicating that Muc1 suppresses P. aeruginosa-induced inflammation. Likewise, small interfering RNA (siRNA) of MUC1 from human bronchial epithelial cells enhanced IL-8 secretion in the presence of flagellin [184]. In addition, overexpression of MUC1 reduced secretion of TNF-alpha, not only in the presence of flagellin, a TLR-5 agonist, but also in the presence of agonists for TLR-2, TLR-3, TLR-4, TLR-7, and TLR-9 [185]. The mechanism by which MUC1 interacts with TLRs suppressing their signaling pathways is not known.

In respiratory syncytial virus (RSV) infection, there is an increase in TNF-alpha, which upregulates MUC1 expression via the TNF receptor on lung epithelial cells, suppressing stimulation of further TNF-alpha increase by RSV and forming a negative feedback loop that controls RSV-induced inflammation [186]. Likewise, nontypeable Haemophilus influenza activates TLR-2 in airway epithelial cells, leading to increased TNF-alpha and IL-8 cytokines, which activate MUC1 expression, leading to suppressed TLR signaling and IL-8 [187]. Hence, MUC1 controls nontypeable Haemophilus influenza infection via suppression of TLR-2 signaling. Thus, MUC1 plays a role in the control of excessive inflammation during airway bacterial or viral inflammation, and the cytoplasmic tail, not the extracellular domain, is responsible for the anti-inflammatory effect.

Conversely, intranasal inoculation of wild-type mice with murine adenovirus type 1 (MAV-1) greatly enhanced TNF-alpha cytokine levels, but did not alter Muc1 expression levels. A higher viral load was detected in Muc1-/- mice compared with non-Muc1-/- mice, suggesting that Muc1 may protect against MAV-1 respiratory infections. However, Muc1 does not have an anti-inflammatory effect in MAV-1 infections [188].

Mucins, including MUC1, protect the gastric epithelium. Helicobacter pylori (H. pylori) causes chronic gastritis and intestinal metaplasia that may evolve into gastric cancer. MUC1 expression is decreased in the gastric mucus gel layer where H. pylori resides in humans, and eradication of H. pylori from patients results in a return to normal MUC1 expression. In the presence of H. pylori, gastric epithelial cells (KATO III) stopped synthesizing MUC1 within 4 h, which could be partially reversed by 24 h [189]. The decrease in vivo in gastric mucins including MUC1 disrupts the protective barrier of the mucin layer, leading to gastric diseases.

MUC1 typically localizes at the apical surface of cells; however, during H. pylori infection, MUC1 is not restricted to the apical surface [190]. H. pylori infection leads to aberrant glycosylation of MUC1, exposing peptide core epitopes and enhancing the immunogenicity of MUC1. Indeed, anti-MUC1 IgG antibodies are present in patients with benign gastric disease, in chronic gastroduodenal diseases, and in gastric cancers [191]. The presence of anti-MUC1 antibodies in H. pylori infections suggests a role in modulating tumor immunity. H. pylori was shown to interact with the MUC1 VNTR region [192].

In the gastric mucosa of Muc1-/- H. pylori infected mice, there are higher levels of keratinocyte chemoattractant (mouse equivalent to human IL-8), TNF-alpha, and neutrophils and increased H. pylori colonization [193]. SiRNA knockdown of MUC1 in human gastric epithelial cells was associated with increased phosphorylation and enhanced IL-8 production, while overexpression of MUC1 reduced IL-8 production. Furthermore, it was noted that MUC1 interacted with beta-catenin and CagA, and that overexpression of MUC1 reduced this interaction and decreased nuclear localization of beta-catenin driven by H. pylori [194]. Furthermore, Muc1 binds to IKKgamma, blocking the stimulation of NF-kappaB by H. pylori and the resulting inflammatory responses [193]. Thus, promoting overexpression of MUC1 in the gastric epithelia may provide a therapeutic strategy to prevent H. pylori infection in the gastric mucosa.

Campylobacter jejeni (C. jejeni) is one of the most common causes of human gastroenteritis in the world. C. jejeni infection in mice results in a rapid increase in Muc1 expression in gastrointestinal cells, which demonstrates that cell surface mucins including Muc1 are critical components of the mucosal barrier to infection [195]. Likewise, infection of mice with Citrobacter rodentium (C. rodentium), a model for gastroenteritis in mice, increased Muc1 cell surface expression in the distal colonic epithelium, indicating that Muc1 has a protective role against bacteria reaching the epithelial surface [196]. Similarly, infection with Salmonella St. Paul, C. jejeni, or Clostridium difficile in humans induces MUC1 expression in the colon [196].

MUC1 is expressed in oral epithelial cells (KB cells) and is upregulated by IL-1beta, IL-6, TNF-alpha, or IFN-gamma. An increase in MUC1 transcript and protein levels is noted in KB cells in the presence of liposaccharide from Porphyromonas gingivalis [197]. Furthermore, mucins, including MUC1, are known to be important components of the mucociliary transport system in the middle ear and Eustachian tube, and MUC1 expression is increased during pneumococcal infection [198]. Thus, pneumococci contribute to the hyper-production of MUC1 during infection.

Human corneal and conjunctival epithelial cells synthesize and express MUC1. In mice, Muc1 protein and mRNA levels are present in the conjunctival and Harderian gland epithelia. Muc1−/− mice (C57BL/6 × SVJ129 background) are predisposed to developing blepharitis and conjunctivitis compared with wild-type mice with intact Muc1 [199]; however, Muc1−/− mice (C57BL/6 background) show a normal ocular surface phenotype with no ocular surface infections [200]. Thus, Muc1 has a protective role at the ocular surface, acting as a protective barrier to infection.

Over 45 years ago, it was reported that a history of mumps in women with ovarian cancer protected women against benign ovarian cysts [201]. This association suggested that mumps was involved in protection against ovarian cancer. Indeed, anti-MUC1 antibodies were present and significantly higher in mumps cases compared with other patients [202]. Although mumps can result in orchitis, sterility, meningitis, pancreatitis, and deafness, this study suggests that mumps infection may also have protected patients against ovarian cancer. Further studies are required to understand the role of MUC1 in mumps infections and the apparent subsequent protection against ovarian carcinoma. It is likely that the mumps virus expresses a sequence similar to MUC1 or a MUC1 mimic, which induces an anti-MUC1 immune response, leading to effective immune surveillance of MUC1 + ovarian cancer cells.

MUC1 in disease

MUC1 expression is altered in a number of diseases. Crohn’s disease, a chronic inflammatory disease, is genetically associated with the NOD2 gene and its protein. MUC1 expression is significantly reduced in the ileal mucosa from Crohn’s disease patients compared with normal ileal mucosa (Table 1) [203]. The decreased level of MUC1 (and other mucin genes) suggests that there is a mucosal defect in Crohn’s disease.

In inflammatory bowel disease (IBD), there is marked overexpression of MUC1, which is aberrantly glycosylated compared with luminal MUC1 from age-matched healthy controls [204]. As a result of the abberant glycosylation, anti-MUC1 antibodies are present in the sera of IBD patients. MUC1 may play a role in the pathogenesis of IBD and MUC1 may be used as a therapeutic target. In fact, immunization of mice (MUC1 transgenic mice crossed with IL–10–/- mice) against MUC1 delays IBD and prevents progression to colitis-associated colon cancer [205]. T helper (Th)1- and Th2-mediated colitis was shown to be more severe in Muc1-/- mice compared with control mice, with increased levels of Th17 cells and IL-17 production [206]. Muc1, which is upregulated by Th17 signaling, functions in a negative feedback pathway that prevents an excessive Th17 cell response in inflamed mouse colons. Disrupting this negative feedback pathway may play a role in the development of IBD.

Overexpression of MUC1 is seen in interstitial pneumonitis, but MUC1 expression is normal in alveolar pneumonia. MUC1 serum levels are increased in interstitial pneumonitis (pulmonary fibrosis, hypersensitivity pneumonitis, sarcoidosis, radiation pneumonia, or farmer’s lung disease), pulmonary alveolar proteinosis, and in bronchiolitis obliterans syndrome (Table 1). MUC1 levels in serum and bronchoalveolar lavage fluid are a sensitive marker for indicating the activity of fibrosing lung diseases. Purified MUC1 is chemotactic and chemokinetic for human fibroblasts in vitro, and chemotaxis is augmented in the presence of fibronectin [207]. These studies indicate that MUC1 is one of the chemotactic factors for fibroblasts, and increased MUC1 in the epithelial lining fluid in small airways may contribute to intra-alveolar fibrosis in fibrosing lung diseases [207]. In patients with idiopathic pulmonary fibrosis, high MUC1 levels are associated with a poor prognosis and low MUC1 levels with better outcomes.

Cystic fibrosis is an autosomal recessive, chronic, progressive genetic disease that affects mostly the respiratory system, causing accumulation of mucus and impaired lung function, and also affects the pancreas, liver, and intestines, leading to a shortened life expectancy. Cystic fibrosis cells overexpress and secrete high levels of MUC1 (Table 1) [208]. Elevated MUC1 levels are also present in serum from cystic fibrosis patients [209]. In these patients, the respiratory tract (bronchial epithelial cells), nasal epithelial cells, intestinal epithelial cells, and pancreatic cells all express high levels of MUC1. Cystic fibrosis mice expressing MUC1 have high levels of MUC1 (MUC1 transmembrane, MUC1/Y, and MUC1/SEC) in the intestine.

MUC1 secretion by goblet cells in the intestine has a protective effect; however, alterations in the quality or quantity of MUC1 may be a cause of ulcerative colitis. In addition, anti-MUC1 antibodies present in the serum of ulcerative colitis patients can destroy the colonic cells through antibody-dependent cell-mediated cytotoxicity [210].

Moreover, MUC1 is involved in the pathogenesis of neonatal acne (high sebum excretion rate) and post-adolescent acne, where longer VNTR regions are associated with severe acne [88]. It is clear that MUC1 plays a role in many diseases but the precise role of MUC1 in diseases is not entirely clear.

MUC1 in cancer

Overexpression of MUC1 is often associated with colon, breast, ovarian, lung, prostate, and pancreatic cancers, tumor progression, and poor prognosis. MUC1 is shed into the circulation and is elevated in the serum of carcinoma patients. MUC1 is highly expressed in virtually all adenocarcinomas (Table 1) and functions as an oncoprotein, making MUC1 a primary target for tumor defense strategies.

MUC1 is overexpressed in head and neck carcinomas (including tongue, larynx, pharynx, macillary sinus, and tonsillar ring) [211]. MUC1 is also expressed in esophageal squamous cell carcinomas, gallbladder adenocarcinomas, hepatocellular carcinomas, extrahepatic bile duct carcinomas, renal carcinomas, urinary tract carcinomas, mesotheliomas, and thyroid cancers (Table 1). MUC1 is overexpressed in endometrial adenocarcinomas; leiomyomas and leiomyosarcomas; prostate, uterine cervix, and rectal carcinomas, and extramammary Paget’s disease (Table 1) [212214].

MUC1 is also overexpressed in a number of hematological cancers, including in multiple myeloma cells, acute lymphoblastic leukemia, chronic lymphoblastic leukemia, blast cells, and CD34 + cells from acute myelogenous leukemia and hairy cell leukemia patients (Table 1). Aberrant overexpression of MUC1 has been noted in adult T-cell leukemia/lymphoma and plays a role in disease progression. In addition, MUC1 is highly expressed in gastric mucosa-associated lymphoid tissue (MALT) lymphomas, follicular lymphomas, plasmacytomas, anaplastic lymphoma kinase (ALK)-positive anaplastic large cell lymphomas, and diffuse large B-cell lymphomas (Table 1) [215, 216]. MUC1 was overexpressed in only a few cases of classic Hodgkin’s disease (1/20) or ALK-negative anaplastic large cell lymphomas (3/20) and in none of the primary cutaneous anaplastic large cell lymphomas tested. Clearly, MUC1 can be regarded as a universal antigen that is highly expressed in a number of different cancers.

Aberrant glycosylation

In normal tissues, MUC1 is localized at the secretory pole of the cell. Conversely, on cancer cells, although its expression is characteristically heterogeneous, MUC1 is found ubiquitously throughout the cell surface and is, therefore, accessible as a target for immunotherapy. In addition, MUC1 is underglycoslyated in cancer cells, which leaves the protein core exposed and exposes new tumor peptide and carbohydrate epitopes that can be detected by Mabs.

Several differences in glycoprotein profiles have been observed in tumor cells compared with normal cells. In tumor cells, glycoproteins such as MUC1 are incompletely glycosylated, generating novel oligosaccharides that are highly expressed in a variety of cancers, whereas they are weakly expressed in normal tissues, presumably because of further oligosaccharide chain elongation [217]. The linkage sugar is N-acetyl galactosamine (GalNac) in glycoproteins carrying O-linked sugars and is N-acetyl glucosamine (GlcNAc) in those carrying N-linked sugars. The incomplete glycosylation in cancer cells is due to an alteration in glycosyl transferase activity, leading to shortened carbohydrate side chains at normally O-linked core carbohydrate determinants, such as the monosaccharide GalNac linked to serine (Tn antigen) and the disaccharide Gal(beta1-3)GalNac or Thomsen-Friedenreich (TF or T) antigen, which are not exposed in normal cells. Activity of the normal colonic enzyme UDP-GlcNAc: GalNAc b3-GlcNAc-transferase, which is involved in O-linked glycans, is decreased in colon cancer and is not present in all cultured cells [218, 219]. The reduced activity of this enzyme favors the synthesis of cancer-associated T antigens.

Most cancer cells show reduced beta6-GlcNAc-transferase activity and express the L-type enzyme, which is normally expressed in leukocytes and is involved in synthesis of O-linked glycans, whereas L-type expression is decreased in breast cancer cells [218, 219]. This decrease in L-type activity favors the exposure of peptide (protein backbone) epitopes (primarily SAPDTRP amino acids within the VNTR) of epithelial MUC1, as well as T antigens. Interestingly, the majority of the monoclonal antibodies to MUC1 recognized SAPDTRPAP region of MUC1 VNTR, as this sequence is highly immunogenic in mice. However in humans, naturally occurring antibodies recognize the minimal epitope RPAPGS most frequently [220] and immunization of cancer patients with oxidized mannan-MUC1 results in antibodies against STAPPAHG and PAPGSTAP sequences [221].

Three different carbohydrate T antigens have been identified in cancer cells [218, 219]: the TF or T antigen, Tn antigen, and sialated Tn antigen. The TF/T antigen is an O-linked disaccharide (betaGal(1-3)aGalNAc-O-serine/threonine) found on all human lymphocytes, erythrocytes, and epithelial cells, which can be revealed by neuraminidase treatment. TF determinants and their monosaccharide precursors (alphaGalNAc-O-serine or Tn antigen) are expressed on approximately 85 % of human adenocarcinomas, both primary and metastatic, based on the reactivity of human carcinomas to synthetic TF and Tn Mabs [221]. This suggests that the TF antigen may be an important tumor marker and target for antitumor immune responses.

Sialated Tn (STn antigen) has a terminal structure of alpha2-6SialylGalNAc-O-serine/threonine in the core region of mucin glycoproteins, which is produced during one of the first steps in mucin glycosylation, and has also been associated with cancer. Expression of the STn epitope in colon cancer patients has been found to be associated with a poorer prognosis compared with colon cancers that are not STn + [222]. The presence of the STn epitope on circulating mucins is also a predictor of poor prognosis in ovarian and breast cancer patients [223], with a much higher 5-year survival rate of 85 % in ovarian cancer patients with STn- serum compared with 10 % in STn + patients [222224]. This suggests that STn may have some importance in the metastatic cascade and is thus an appropriate target for immunotherapy.

A Phase III clinical trial with the synthetic vaccine STn-KLH, which targets the STn antigen (Theratope, manufactured by Biomira Inc., now known as Oncothyreon Inc. together with Merck KGaA), was completed in 1030 breast cancer patients at more than 120 sites in 10 countries [225, 226]. The trial failed to meet the primary endpoints of time of disease progression and overall survival; however, a subset of patients on hormonal treatment in addition to Theratope showed trends towards higher survival rates. Studies are being conducted at other research laboratories to develop new improved STn cancer vaccines [227].

Serum tests

MUC1 is overexpressed on carcinomas both in the cytosol and on the cell surface membrane and, as a result, MUC1 is released from tumor cells into the serum, and can be detected with diagnostic serum antigen detection tests in patients with cancer. Well-known commercially available serum tests based on antibodies to MUC1 include cancer antigen (CA)15.3, CA19.9, CA27.29, CA549, cancer-associated serum antigen (CASA), mammary serum antigen (MSA), mucin-like carcinoma-associated antigen (MCA), MA6-MA9, and ovarian serum antigen (OSA).

CA15.3 is a commercially available test (originally by Centocor Diagnostics) that detects MUC1 using the anti-MUC1 VNTR Mab DF3 (epitope DTRPAP) as the catcher and radiolabeled anti-MUC1 carbohydrate Mab 115D8 (epitope STn) as the tracer. CA15.3 is an independent prognostic marker for breast cancer and is used routinely worldwide to monitor the response to breast cancer treatment and disease recurrence [228].

CA19.9 was probably the first mucin-based assay to be developed and has proved to be useful in monitoring disease recurrence and prognosis in patients with pancreatic cancer. CA19.9 (which detects the sialylated Lea epitope within the MUC1 glycoprotein) is also elevated in 67 % of patients with advanced adenocarcinomas of the upper gastrointestinal tract [229].

CA27.29 is a fully automated, commercially available method to measure MUC1 based on the anti-MUC1 VNTR Mab B27.29 (epitope PDTRPAP). It is commonly used to test for treatment effectiveness and to detect early recurrence in breast and ovarian cancer [230].

CA549 is a commercially available method to measure MUC1 using the anti-MUC1 VNTR Mab BC4E 549 (epitope TSAPDTRPAP) and is commonly used in the management of breast cancer [231].

CASA and OSA are quantitative serum tests for mucins produced by epithelial carcinomas that use Mabs to the MUC1 VNTR region [86]. The anti-MUC1 VNTR Mab BC2 (IgG1, epitope APDTR) is used in both assays as the capture antibody, and the anti-MUC1 VNTR Mabs BC3 (IgM, epitope APDTR) and OM-1 (IgM, epitope APDTRP) are used as the secondary Mabs for CASA and OSA, respectively. Both assays have been shown to be useful in monitoring ovarian cancer patients [232, 233]. Elevated CASA results were also found in 76.5 % of breast cancer patients but only 2 % of normal females [232, 233]. In addition, CASA and OSA were found to be more sensitive and far superior to CA-125 (MUC16) levels for the detection of small volume ovarian carcinomas [234].

The MSA test uses the Mab 3E1.2, which detects a carbohydrate epitope (O-linked to the mucin, containing N-glycolylneuraminic acid) of the MUC1 molecule [235]. This test was first developed as a research assay and was found to be superior to CA15.3 and carcinoembryonic antigen (CEA), particularly in the early stages of breast cancer [236]. MSA levels were elevated in 76 % of stage I/II and 86 % of stage III/IV breast cancer patients, while 1.9 % of healthy blood donors had raised MSA levels. Furthermore, MSA and CA15.3 in combination gave a specificity and predictive value of a positive result of 100 % [232].

The commercially available MCA assay (Roche Diagnostics) is a two-step solid phase enzyme assay using the anti-MUC1 VNTR Mab b12 (epitope APDTRPAP) as the capture and detection antibody. MCA serum levels have been shown to be a sensitive indicator of metastatic breast cancer [237].

MA6-MA9 is a sandwich radioimmunoassay using the anti-MUC1 VNTR Mabs MA6 and MA9. Serum MA6-MA9 levels were elevated by 63, 75, and 88 % in stage I, II, and III/IV breast cancer patients, respectively [238, 239]. The marker is not elevated in pregnant or lactating females. MA6-MA9 assay levels were also raised in 97 % of patients with metastatic breast cancer, while CEA levels were only raised in 46 % [238, 239].

With the rapid progression and understanding of the molecular structure of mucins in the last 25 years, significant progress has been made in producing anti-MUC1 Mabs and developing useful serum tests to monitor disease treatment and metastasis. However, currently available serum tests will detect antigens in serum from patients with advanced cancer patients but not from those in the early stage of the disease. Thus, better screening assays are needed for the early diagnosis of cancer.

A target for cancer immunotherapy studies

The idea that patients could be immunized with vaccines developed against MUC1 + adenocarcinomas was the logical extension of a number of independent observations [232, 240] as summarized below.

  1. (i)

    MUC1 is expressed in normal epithelial cells at low levels and is restricted to the apical surface; however, during cancer (breast, ovary, pancreas, colon, kidney, and lung), there is up to a 100-fold increase in MUC1 expression, and it is present over the entire cell surface.

  2. (ii)

    Cell surface MUC1 exists as flexible rods that protrude relatively great distances from the cell membrane surface and are the first points of contact by immune effector cells and antibodies.

  3. (iii)

    There is much information available on MUC1 and synthetic polypeptides from the sequence of MUC1; MUC1 was the first mucin to be cloned and the protein structure deduced.

  4. (iv)

    MUC1 is an attractive molecule for immunological studies. It is highly immunogenic in mice; almost all mice immunized with tumor cells or extracts produced antibodies that reacted with the highly immunogenic VNTR (PAPGSTAPPAHGVTSAPDTR)25-125, specifically with the APDTR amino acids within this region.

  5. (v)

    During malignant transformation, MUC1 is aberrantly glycosylated, altering the structure and exposing new carbohydrate epitopes, including the Tn antigen, STn antigen, and TF antigen. The role of Tn-, STn-, and TF-antigen glycan structures in tumor biology is not clear; however, their presence is correlated with poor prognosis in most adenocarcinomas.

  6. (vi)

    Further, as a result of the aberrant glycosylation, peptide epitopes (primarily SAPDTRPAP within the VNTR) are exposed that are rarely detected in normal tissues or secretions.

  7. (vii)

    Lymph nodes from patients with breast and pancreatic cancer and T cells from patients with ovarian cancer and multiple myeloma were shown to contain human leukocyte antigen (HLA)-unrestricted/anti-MUC1 reactive CTLs that reacted to the same MUC1 epitopes as detected by the murine Mabs. In a patient with breast cancer who later became pregnant (MUC1 is greatly increased in the lactating breast) and had a severe case of mastitis accompanied with heavy lymphocyte infiltration and a high frequency of CTLs to MUC1, there has been no recurrence of the cancer. It is possible that the non-MHC-restricted T cells were protective.

  8. (viii)

    MHC-restricted CD8 + T cells against the MUC1 protein core within and outside the VNTR region are present in breast cancer patients and in healthy multiparous women.

  9. (ix)

    In addition to anti-MUC1 T cells found in cancer patients, antibodies to MUC1 (within the VNTR regions) have been found in patients with ovarian, breast, pancreatic, or colon cancers and in patients with endometriosis.

  10. (x)

    Foxp3 + CD4 regulatory T cells accumulate in late-stage endometriosis in transgenic mouse models and in humans, suggesting a role for MUC1 in progression from ovarian endometriosis to ovarian cancer. In addition, MUC1 transgenic mice have higher levels of CD4 + CD25 high regulatory T cells compared with wild-type mice.

  11. (xi)

    Circulating soluble MUC1 is present in the serum of cancer patients, and a number of serum tests that detect MUC1 have been developed.

  12. (xii)

    Immune complexes of MUC1 and antibodies have been found in the serum from ovarian and breast cancer patients. Soluble MUC1 and MUC1/antibody immune complexes are also detected at higher levels in pregnant women than in non-pregnant women.

  13. (xiii)

    In mice and in humans, immunization with MUC1 VNTR, non-VNTR, Tn, STn, and TF peptides stimulates immune responses.


MUC1 is a high molecular weight, heavily O-glycosylated glycoprotein, which is expressed in all epithelial tissues, leukocytes (T cells, B cells, dendritic cells, and monocytes), the eye, ear, mouth, skin, milk, and gallbladder stones, and plays a key role in the reproductive system. MUC1 plays an essential role in forming protective mucus barriers against infections, has anti-adhesive properties due to its large rod-like structure, prevents cell death, promotes tumor invasion/metastasis, and is involved in intracellular signaling.

A number of alternatively spliced variants of MUC1 have been described that encode different isoforms of MUC1, some of which are more common in cancer cells than in normal cells. Overexpression, aberrant cellular localization (presence on the entire cell surface rather than localized to the apical surface as in normal cells), and changes in glycosylation of MUC1 have been associated with carcinomas. Research is continuing to determine the role of MUC1 in cancer and other diseases, its interactions with pathogens, and its role in the immune response. Here, we have presented only a small glimpse of this fascinating molecule.