1 Introduction

Ovarian cancer is the most lethal gynaecological malignancy and the fifth most common cause of cancer-related deaths in women in the USA, resulting in an estimated 22,280 new cases and 15,500 deaths in 2012 [1]. The high mortality rate is ascribed to the fact that more than 60 % of ovarian cancer patients already have advanced disease at diagnosis. Therefore, ovarian cancer is also referred to as the ‘silent killer’. Ovarian cancer treatment typically combines surgery and platinum–taxane-based chemotherapy. While the majority of patients initially respond to platinum-based therapy, in approximately 25 % of cases no response can be achieved. Although current approaches including surgery and combination chemotherapy and hormonal therapy yield responses in 60–80 % of patients with advanced disease, the majority of ovarian cancer patients eventually relapse and become refractory to additional treatment [2], with a median progression-free interval of 18 months [3]. Conventional cancer chemotherapy often results in severe side effects related to non-specific modes of action. Studies evaluating various cytotoxic agents in recurrent ovarian cancer have found response rates of 10–28 % with an accompanying progressive increase in the number of drug-resistant tumours [4]. Currently, there is no satisfactory adjuvant treatment following surgery and chemotherapy. Thus, novel therapeutic strategies are urgently needed to improve the outcome for this deadly disease.

Epithelial ovarian cancers (EOCs) account for 90 % of all ovarian cancers and can spread directly to adjacent organs; ‘seeding’ of the peritoneal cavity is frequently associated with ascites formation, the most common feature of ovarian carcinoma, particularly serous carcinoma [5]. There are six histological tumour types associated with the ovarian surface epithelial layer. The different histologies include serous, mucinous, endometrioid, clear cell, Brenner and undifferentiated tumours. Morphologically, the primary tumour initially presents as a multilocular ovarian cyst with tumour tissue growing on the inside. Tumour cells penetrate the capsule resulting in growth on the outside. Tumour cells detached from the outside of the primary tumours disperse in the abdominal cavity and implant on peritoneal surface preferably the omentum. Advanced EOCs are characterized by rapid growth of solid intraperitoneal (i.p.) tumours and production of large volumes of ascites. Intra-abdominal dissemination is the primary cause of death for patients with EOC, although the exact mechanisms involved in EOC progression remain unclear. The poor prognosis in patients with EOC is related to the degree of perito-dissemination of cancer cells. EOC is frequently diagnosed in the late stage, and nearly all late-stage patients will suffer disease recurrence and death.

The progression of EOC from primary to metastatic disease is associated with a number of molecular and genetic changes. A variety of changes in genomic structure, growth factor receptors, proto-oncogenes and tumour suppression genes have been identified in EOC [6]. These changes can affect the expression of specific tumour-associated antigens (TAAs). Identification of molecular aspects of ovarian cancer growth and the enhancement of cancer cell motility, detachment from the primary tumour, attachment to the peritoneum and invasion of subperitoneal tissue have become the central focus in the development of molecular-targeted therapy for ovarian cancer [5, 7]. Targeting cancer surface TAAs with an antibody or antibody conjugate using a targeted therapy is a new develo** area and may have a promising future for control of the late stage and recurrent EOC. For the targeting therapies, molecular therapeutics requires credentialed and validated targets. Until now there have been few candidate TAAs identified for EOC, and clinical effects are limited [8].

An interesting TAA is mucin 1 (MUC1) because it is overexpressed in most adenocarcinomas [9–12]. MUC1 is a highly glycosylated type I transmembrane glycoprotein overexpressed on the cell surface in more than 90 % of EOCs, including platinum-resistant tumours [13–17]. Tumour-associated MUC1 is structurally different from normal MUC1 in that the former has shorter and less dense O-glycan chains, exposing novel regions of the protein core [18]. Thus, tumour-associated MUC1 is a promising therapeutic target for develo** novel approaches for the treatment of EOC patients.

Here we review the important findings of MUC1 in EOC metastasis and progression, discuss the circulating MUC1 in blood as an additional biomarker for EOC diagnosis and provide a comprehensive summary of MUC1 expression on human EOCs. Moreover, we focus on some emerging innovative approaches to the management of the disease and explore potential therapeutic applications of MUC1-targeted strategies in EOC treatment.

2 Structure and roles of MUC1 in cancer metastasis

2.1 Structure of MUC1

In common with all mucins, MUC1 consists of a protein backbone containing highly glycosylated and unglycosylated regions [19]. It is synthesized as a single polypeptide chain but expressed as a cell surface heterodimer that consists of N-terminal and C-terminal subunits, which form a stable complex following cleavage of a single MUC1 polypeptide in the endoplasmic reticulum [20]. Structurally, MUC1 consists of a large extracellular subunit comprised of a mucin-type identical 20-amino acid tandem repeats (TRs) and a smaller subunit that includes a small extracellular domain and a transmembrane (TM) domain plus a cytoplasm tail (CT) (Fig. 1).

Fig 1
figure 1

Schematic diagram of MUC1 protein located at the cell surface, its multiple functions and potential signalling pathways in cancer metastasis

Full size image

The MUC1 glycoprotein is a complex molecule with a protein core containing a large domain of variable number of tandem repeats (VNTR). The VNTR domain represents major portions of the extracellular protein. Within each TR, two serines and three threonines represent five potential O-glycosylation sites. The large extracellular TR domain of MUC1 is heavily O-glycosylated [21, 22]. O-Glycosylation with complex oligosaccharides is crucial to MUC1 structure and function. The extent of glycosylation depends not only on the tissue expressing MUC1 but also on the profile of glycosyltransferases expressed in those tissues [23, 24]. Tumour-associated MUC1 has been reported to be both less glycosylated and more glycosylated than the forms expressed by normal tissues [25, 26]. Glycosylation is altered in cancer tissue, revealing immunodominant peptide sequences in every TR, which on normal tissues is masked by glycosylation. The density of glycosylation of TRs by different tumours and normal cells is also highly variable and is believed to contribute significantly to the normal and aberrant functions that are associated with MUC1 during the pathogenesis of cancer and other diseases [27]. Differential glycosylation patterns on the TR may affect adhesion properties that result in an increased ability of tumour cells to metastasise. Many murine antibodies reactive with the VNTR domain of MUC1 have now been produced by immunisation with diverse materials including milk fat globule membranes, tumour cells and isolated mucin preparations. These MUC1 monoclonal antibodies (MAbs) can be used for cancer diagnosis and targeting therapy, which will be discussed in the following sections.

The studies on MUC1 TM in cancer are rarely reported until now. In addition to the full-length TM isoform, additional variants of MUC1 exist. These isoforms are generated by alterative mRNA splicing, and a number of them lack the TR region (MUC1/X, MUC1/Y and MUC1/Z) or CT (MUC1/SEC) [28–32]. These variants could be useful for cancer diagnosis and immunotherapy. However, their functions need to be further investigated.

Another important domain of MUC1 is the CT, which is highly conserved across mammalian species and hypothesized to play a role in its post-translational processing, subcellular localization, signal transduction and intracellular localization [33]. Alterations to the CT may affect trafficking through the Golgi and thereby influence glycosylation of the TR domain. The oncogenic effects of MUC1 are believed to occur through the interaction of its CT with various signalling molecules. The CT of MUC1 is 69 amino acids long and has several tyrosine, serine and threonine phosphorylation sites that can bind to several proteins implicated in the cancer regulation by affecting the proliferation, apoptosis, metastasis and transcription of various genes [34, 35]. Both TRs and CT domains are believed to be of functional significance. In addition, TR and CT have a close link in regulating cancer metastasis and progression. One animal study using S2-013 pancreatic cancer cell line has demonstrated that alternations in MUC1, specifically deletion of either the TR or CT, resulted in an increased propensity of tumour cells to invade vessels and metastasize to lymph nodes compared with a cell line overexpressing full length MUC1 and suggested a cooperative relationship between the TR and CT domains of MUC1 [36]. The details of TRs (extracellular domain) and CT associated with cancer metastasis and progression are discussed in the following sections.

2.2 MUC1 in cancer invasion and metastasis

The expression of MUC1 in cancer progression and metastasis is characterized by increased levels, altered glycosylation and aberrant surface distribution patterns [9, 27]. It is generally accepted that the over-expression of MUC1 by the tumour cells facilitates the invasive growth and the metastasis [37, 38]. In particular, the extracellular domain of MUC1 facilitates cancer progression. The expression of MUC1 in tumours may disrupt cell–cell and cell–matrix adhesions and function as an anti-adhesion molecule, which inhibits cell–cell adhesion, including a release of cells from tumour nests and causing micrometastasis [39, 40]. It can also promote adhesion and presumably metastasis by presenting carbohydrate ligands to adhesion molecules on endothelial cells such as intercellular adhesion molecule-1 [41], E-selectin [42] and sialic acid-binding immunoglobulin superfamily lectins [43]. Using a triple transgenic (Tg) MUC1KrasPten mouse model, Budiu et al. recently showed that MUC1+ ovarian tumours have augmented capacity for loco-regional spread and increased accumulation of CD4+Foxp3+ immune-suppressive regulatory T cells in vivo and are accompanied by high serum MUC1 [44]. These data further support the role of MUC1 in EOC metastasis.

2.3 MUC1 and immune suppression

Despite its immunogenicity, MUC1 also contributes to esca** from tumour immune surveillance and metastasis formation. In tumour cells, MUC1 has an anti-adhesive function, which is not only of importance in cancer development [45] but also might be responsible for diminished adherence of immune effector cells to the malignant cells [46]. The biological function of MUC1 may be in part due to its large size and the extended rigid structure. It has been proposed that enhanced levels of MUC1 expression by cancer cells may mask extra-cellular domains from immune surveillance, confer a survival advantage on malignant cells and play an important role in the ability of tumours to invade and metastasise [47], as MUC1-expressing melanoma cancer cells are found to be less susceptible to T-cell- and natural killer (NK) cell-mediated lysis in vitro [46]. In addition, the tissues from late-stage EOC patients were found to be accompanied by considerable immune suppression [48–50], supporting that the MUC1-associated immune suppression is involved in EOC metastasis.

2.4 MUC1 and anti-apoptosis

Over-expression of MUC1 plays an important role in anti-apoptosis, activation of survival pathways and is involved in drug resistance in cancer metastasis. In one study, Ren et al. have demonstrated that MUC1 CT attenuates the apoptotic response to DNA damage by increasing anti-apoptotic Bcl-XL and PI3K/Akt pathways as well as in part through suppressing the release of apoptogenic factors from mitochondria, and this oncoprotein confers resistance to genotoxic anticancer agents [51]. In another study, Wei et al. from the same group further demonstrated that MUC1 CT binds directly to the p53 tumour suppressor domain and selectively promotes transcription of growth arrest genes, decreases transcription of apoptotic genes as a survival response to stress and thereby decreases cell death [52]. Furthermore, MUC1 also activates the survival-related FOXO3a transcription factor in response to oxidative stress [53]. The relationship between MUC1 expression and chemoresistance in EOC is rarely reported. It was found that the over-expression of MUC1 on the surface of human pancreatic cancer cells impedes the cytotoxic activity of chemodrug 5-FU [54]. The following study further confirmed that this over-expression reduces intracellular drug uptake, anti-neoplastic and anti-tumour drug effects, which may have important clinical implications in treatment [55]. These observations suggest that over-expression of MUC1 in cancers could potently contribute to tumour resistance by a variety of mechanisms. Investigating the relationship between MUC1 expression and chemoresistance in EOC is an interesting area and worthwhile being conducted in the future.

2.5 MUC1 and autophagy

Autophagy is an interesting research area and is involved in cancer metastasis. Autophagy is a cellular response to stress or nutrient deprivation, which is a way to supply amino acids as an alternative energy source by degradation of damaged cytoplasmic organelles or protein [56] and is involved in EOC pathogenesis and progression [57]. It was reported that autophagy can induce EOC dormancy which is highly related with cancer recurrence [58]. Up-regulated autophagy may confer chemo- and radioresistance to cancer cells and also a pro-survival advantage in cancer cells experiencing oxygen and nutrient shortage [57]. In a recent study, MUC1 has been shown to inhibit the induction of necrosis in response to the deprivation of glucose with the induction of autophagy in colon cancer cells in vitro [59]. These results indicate that the over-expression of MUC1 as found in human cancers could provide a survival advantage in microenvironments with low glucose levels. Therefore, it would be interesting to determine whether an association exists between MUC1 and autophagy as a survival advantage in EOC under nutrient-deprivation conditions to help develop novel therapeutic approaches for EOC treatment in the future.

2.6 MUC1 and epithelial–mesenchymal transition

Emerging evidence is suggesting that epithelial–mesenchymal transition (EMT) plays a crucial role in the aggressiveness in EOC including increasing migration and invasion ability, contributing to chemoresistance and cancer stem cell (CSC) populations. The investigation of association between MUC1 and EMT is another attractive research area in cancer metastasis. EOC is a highly metastatic disease during which cells undergoing EMT lose their epithelial morphology, reorganize their cytoskeleton and acquire a motile phenotype through the downregulation of adherent junctions proteins (like cadherins) and upregulation of mesenchymal markers (Snail, Slug and Vimentin) [60, 61]. In one study, Roy et al. demonstrated that over-expression of MUC1 in pancreatic cancer cells triggers the molecular process of EMT, which translates to increased invasiveness and metastasis [62]. In another study, Rajabi et al. further confirmed that the interaction between MUC1 C-terminal subunit (MUC1-C) and androgen receptor (AR) is associated with induction of the EMT and increased invasion; MUC1-C also confers a more aggressive androgen-independent phenotype that is sensitive to MUC1-C inhibition in prostate cancer [63]. The roles of EMT in EOC metastasis and chemoresistance have been recently reviewed [64, 65]. However, no evidence on the roles of MUC1 in ovarian cancer EMT currently exists. Understanding the roles of MUC1 and EMT associated with EOC metastasis could lead to the identification of targets for novel therapeutic interventions.

Taken together, these studies support the conclusion that the role of the MUC1-related cancer metastasis is involved in many different mechanisms. Although the current data are mostly from breast, colon and pancreatic cancer research, it will be very interesting to investigate these mechanisms in EOC.

3 MUC1-mediated signalling pathways in cancer progression

In addition to being a TAA, MUC1 is also an oncoprotein, playing an active role in various cellular pathways, which impacts cell growth, proliferation and migration. Over-expression of MUC1 mediates signal transduction events that stimulate the motility, invasion and metastasis of cancer cells [66]. MUC1 has a distinct role in tumour progression because it can stimulate cell proliferation through its interaction with growth factor receptor, β-catenin and oestrogen receptor α (ORα). Tumour-associated MUC1 mainly plays its important role in cancer metastasis through its CT part. By virtue of its CT, MUC1 can interact with a variety of proteins involved in neoplasia and cell adhesion such as β-catenin, GSK3β, c-Src, protein kinase C delta, Grb2, ErbB1, ErbB2, ErbB3, ErbB4 and p120ctn [67–74] and play an active role in the oncogenic process (Fig. 1). The regulation of MUC1-mediated signalling pathways in cancer metastasis is very complex. In this section, we only focus on two main signalling pathways, i.e. MUC1/ErbB and MUC1/β-catenin/Wnt pathways.

3.1 MUC1 and ErbB family pathways

The ErbB family is comprised of the ErbB1/EGF receptor (EGFR), ErbB2/HER2/Neu, ErbB3 and ErbB4. Through its CT, MUC1 binds with the ErbB family of growth factor receptor tyrosine kinases and potentiates ErbB-dependent signal transduction in the MUC1 transgenic breast cancer mouse model [70, 73, 75]. Following ligand binding and receptor activation, these receptors are endocytosed and transported to lysosomes where the receptor is degraded [76]. This downregulation of growth factor receptors is a complex and tightly regulated process. Through stabilizing and enhancing the ErbB signalling by MUC1–ErbB kinase interaction, MUC1 activates extracellular signal-regulated kinases (ERKs) 1 and 2 and thereby increases cell proliferation [73]. MUC1 also regulates ErbB-independent ERK signalling through modulating the transcription of the genes encoding MEK1, Raf-1 and c-jun [77]. It was reported that the loss of MUC1 significantly delays tumour onset in an EGFR-dependent (WAP-TGFα) transgenic breast cancer model [78]. Further study indicated that the loss of MUC1 expression results in a decrease in the interaction between EGFR and the CCND1 promoter, which translates to a significant decrease in cyclin D1 protein expression [79]. These data suggest that MUC1 can affect breast cancer tumour growth via regulating nuclear localization and function of the EGFR.

3.2 MUC1 and β-catenin/Wnt pathway

MUC1 is also able to influence cellular signalling through its interaction with β-catenin (Fig. 1). β-catenin is a key effector of the Wnt signalling pathway [80, 81]. Binding of the Wnt mitogen to its cell surface receptor leads to the cytosolic stabilization of β-catenin and its accumulation in the nucleus. The MUC1 CT can bind and signal through β-catenin and the mitogen-activated protein kinase (MAPK). In vitro studies showed that MUC1 interacts directly with β-catenin via a 50-SAGNGGSSL-59 amino acid sequence in its CT [67]. β-catenin plays important functions in the formation of the cell–cell junction via E-cadherin interaction [82]. A study showed that a fragment of the CT of MUC1 can be transported to the nucleus in association with β-catenin [83]. β-Catenin acts as a transcriptional co-activator to increase the expression of cell-cycle progression genes cyclin-D1 and c-myc. Dysregulation of β-catenin is of great importance to the development of diverse human malignancies [84]. Disruption of the β-catenin binding site in MUC1 suppresses its ability to induce anchorage-dependent and anchorage-independent growth, indicating β-catenin binding to MUC1 is a critical component of its tumourigenic activity [84, 85]. MUC1 binding to β-catenin also suppresses its ability to interact with E-cadherin at adherent junctions, leading to a breakdown in cell–cell interactions, and GSK3β-mediated disruption of the complex restores the E-cadherin/β-catenin interaction [86]. Hence, MUC1-induced abrogation of cell–cell interactions may be mediated both by steric hindrance by its mucinous extracellular domain as well as by the interaction of its intracellular domain with β-catenin. It will be very interesting to investigate MUC1 interactions with β-catenin/Wnt target genes and their mechanistic implications for EOC metastasis.

In addition to binding β-catenin, MUC1 CT can also bind the serine/threonine kinase GSK3β and binding to this kinase competes away binding between MUC1 and β-catenin in breast cancer cell lines [86]. Both c-Src and PKCδ increase the interaction between β-catenin and MUC1 by binding and phosphorylating different tyrosine sequences that are located near the β-catenin binding domains [71, 72]. Schroeder et al. have demonstrated that MUC1 and β-catenin biochemically interact in a tumour-specific manner in MMTV-Wnt-1 transgenic mice, and lack of MUC1 expression causes a significant delay in tumour onset in this model of breast cancer [68].

All these observations indicate the CT of MUC1 may provide a proto-oncogenic signal, and targeting or interfering the CT of MUC1 may block this signal transduction and prevent tumour differentiation, proliferation and metastasis. Targeting or interfering MUC1-CT signalling pathways for cancer treatment will be discussed in the following section.

4 MUC1 as a biomarker in blood for EOC diagnosis and monitoring

4.1 The limitation of CA125 in EOC diagnosis and monitoring

Serum CA125 (also known as MUC16) continues to be the best validated marker for ovarian cancer and is currently widely employed for monitoring response to therapy and for detecting disease recurrence [87, 88]. However, CA125 is insufficient as a single biomarker for EOC diagnosis [89]. Despite the benefits accompanying the use of CA125, many challenges exist that render it not as effective in early screening and monitoring EOC progression. One of the primary challenges is its decline in sensitivity in early-stage EOC. It was reported that serum levels of CA125 were found to be elevated in only 50 % of symptomatic stage I EOC patients compared with the levels in 80 % of advanced-stage cases [90]. Clearly, CA125 alone is inadequate as a biomarker for EOC diagnosis. Another challenge is that approximately 15 % of EOC patients do not express CA125 [91]. In addition, CA125 is also expressed by a variety of normal epithelial cell types and is relatively nonspecific [92]. A number of false-positive results could also occur with ovulation, endometriosis, fibroids and many other benign conditions. In the past decade, cancer markers that may complement CA125 have been intensively sought for EOC diagnosis and prognosis [93, 94]. Given the limited ability of current approaches to effectively manage treatment-resistant EOC, there is an urgent need to identify additional biomarkers to be used for early diagnosis, to stratify patients and reduce unnecessary treatment-associated morbidities from ineffective chemotherapy and to assess new treatment effects. As MUC1 is overexpressed in over 90 % of EOC and can be shed into blood circulation, detection of serum tumour-associated MUC1 is an appropriate option as a complementary biomarker for EOC early diagnosis and monitoring.

4.2 Serum MUC1 as a biomarker in EOC

Several assays to detect TAAs in serum samples are based on antibodies directed against circulating MUC1 antigens [95] and are used in the postoperative monitoring of breast cancer and EOC patients. The detection of serum tumour-associated MUC1 as a biomarker for EOC diagnosis in clinical studies is summarized in Table 1. Sekine et al. reported that serum MUC1 (DF3 antigen) can be detected by DF3 MAb using a sequential double-determinant enzyme-linked immunoassay (EIA) to monitor EOC progression. MUC1 levels were found to be obviously elevated in 21 of 45 EOC patients (47 %) compared with the normal control subjects and correlated with progression of disease, suggesting that MUC1 is distinct from CA125 and determining levels of both markers may enhance the sensitivity of monitoring the course of EOC [96]. Human milk fat globule (HMFG) is reactive with MUC1 variable number of TRs. Bast et al. demonstrated that the elevations of CA125, HMFG1 (anti-MUC1) and HMFG2 (anti-MUC1) in sera were observed among 47 EOC patients in 91, 77 and 62 %, respectively, and concluded that combination of these markers can improve the specificity of a cost-effective screening strategy for EOC [97]. Using an ‘in-house’ single determinant enzyme-liked immunosorbent assay (ELISA), Fisken et al. evaluated serum HMFG2 in serial samples from 215 primary EOC patients and found that 45 % of patients with stage I, 54 % with stage II, 61 % with stage III and 75 % with stage IV disease had elevated serum HMFG2 and that post-operative levels were significantly related with residual tumour volume (P < 0.005) [98]. These results indicated that serum MUC1 (HMFG2) can be used as an additional biomarker to monitor EOC progression and recurrence after operation. Addition of HMFG2 could be shown to increase sensitivity while retaining the specificity of CA125. Richards et al. detected the core protein of MUC1 epithelial mucin (a synthetic peptide corresponding to a 105-amino acid segment of the MUC1 TR region) using ELISA assay from the sera including EOC patients (n = 10), patients with benign epithelial ovarian tumours (n = 27) and normal health control subjects (n = 100)) and found that MUC1 is present in the serum of healthy women and women with cancer and there is no evidence for an augmentation of the humoral immune response to MUC1 in EOC [99]. The reason for this could be that different anti-MUC1 MAbs bind different epitopes of MUC1 antigen.

Table 1 The detection of serum tumour-associated MUC1 as a biomarker for EOC diagnosis in clinical studies
Full size table

Combined with other markers (HE4, Glycodelin, MMP7, SLPI, Plau-R, Inhibin A, PAI-1 and CA125), Havrilesky et al. evaluated the sera MUC1 levels from EOC patients (n = 200) and healthy age-matched control subjects (n = 396) by sandwich ELISAs including an anti-MUC1 MAb (M2C5, Abcam, Cambridge, MA, USA; Ab #8323) and commercially available tests [87]. The results indicated that the both sandwich ELISAs with new MAbs and commercially available tests can be used to distinguish the serum of MUC1 of patients with EOC from that of healthy controls and demonstrated the utility of several one- and two-step multi-marker algorithms with acceptable test characteristics for possible use in an EOC screening population.

Wandall et al. developed an O-glycopeptide microarray that detected IgG antibodies to aberrant O-glycopeptide epitopes in patients vaccinated with a keyhole limpet hemocyanin-conjugated truncated MUC1 peptide and detected tumour-associated MUC1 auto-antibody in sera from EOC patients (n = 20) [100]. These auto-antibodies (MUC1) represent a novel source of sensitive biomarkers for early detection of EOC. Budiu et al. recently found that increased serum MUC1 and high anti-MUC1 antibody levels are potential prognostic biomarkers for poor clinical response and reduced overall survival (OS) in platinum-resistant or platinum-refractory EOC [101]. In contrast to MUC1, neither soluble MUC16 (CA125) nor MUC16-specifc antibodies were significantly associated with clinical response or overall survival (OS) in this study. These results indicated that measuring serum MUC1 may be used as a prognostic biomarker in platinum-resistant EOC. PankoMab [102] is a novel MUC1 antibody explicitly tailored to recognise the tumour-associated MUC1 epitope [103]. Jeschke et al. recently developed a novel anti-MUC1PankoMab to test patient sera MUC1 with an ELISA, which is a useful tool for obtaining strong correlation in sera from patients with benign ovarian diseases versus normal sera, and conclusive information on the transformation process from benign to malignant state in ovarian tissues [104]. All above data support the idea that tumour-associated MUC1 can be detected in blood by different approaches, and combination of MUC1 and CA125 in sera can improve the detection rate for EOC early diagnosis and monitoring disease progression, especially for chemoresistance EOC patients.

5 MUC1 in human EOC tissues

Many research groups have studied the expression of MUC1 on human EOC tissues. MUC1 expression in normal ovaries, benign and EOC tissues tested by different MAbs is summarized in Table 2. Croghan et al. demonstrated that F36/22 MAb reacts with MUC1 in serous, endometrioid, mucinous, clear cell and undifferentiated EOC [105]. Similar findings were obtained for the expression of MUC1 defined by DF3 MAb, which is detectable in 95 % of primary tumours and metastatic lesions, regardless of histology [106]. Ward et al. compared the pattern of expression of MUC1 by HMFG2 MAb with the patterns of MUC1 by HMFG1, F36/22 and AUA1 MAbs in EOC and found that the expression of MUC1 defined by HMF2 MAb was similar to that defined by HMFG1 and F36/22 MAbs, whereas the percentage of expression of MUC1 defined by AUA1 MAb was lower and well-differentiated tumours expressed more MUC1 than poorly differentiated tumours [107]. They concluded that HMFG2 assay was a useful addition to CA125 assay in monitoring EOC patients. Ichige et al. who used DF3-P MAb found that MUC1 is detectable on the surface of EOC cells and benign tumours, but not in normal tissues, and suggested that the DF3-P epitope might be useful as a target for radioimaging or immunotherapeutic approaches to EOC [108]. Dong et al. found that most benign, low malignant potential and invasive EOCs showed high MUC1 reactivity on cell membranes and that increased MUC1 expression was not correlated with the stage and grade of EOC using BC2 MAb [14]. Feng et al. who used an anti-MUC1 and MUC1-core MAb (raised against the human breast cancer cell line ZR75-1) found that a high expression of MUC1 was associated with the disease stage and histological grade of EOC and concluded that MUC1 influences the metastatic ability of EOC [40]. Using E29 MAb (Dako), Drapkin et al. found that all primary ovarian carcinomas (papillary serous and endometrioid) were positive to MUC1 [109]. Using anti-MUC1 glycoprotein MAb (clone Ma-695, Novocastra), the MUC1 has been found to be positive in nine of nine (100 %) and 40 of 42 (95 %) of primary ovarian carcinomas in two different studies, respectively [110, 111]. However, the positivity was always cytoplasmic and often granular. Only occasional cases had also membranous staining. Lu et al. [112] and Rosen et al. [113], who used anti-MUC1 core glycoprotein MAb (clone Ma-552), found that MUC1 is detectable on 148 of 158 (94 %) and 200 of 322 (62 %) of EOC tissues, respectively, and the binding sites in most cases are in the cytoplasm. Lu et al. concluded that the combination with other markers might identify >99 % of EOCs despite the heterogeneity of the disease.

Table 2 Summary and comparison of MUC1 expression on human normal ovaries, benign ovarian tissues and EOC tissues using different MAbs
Full size table

Using anti-human CA15-3 MAb, Takano et al. demonstrated that chemoresistance to platinum-containing regimen was closely associated with the cytoplasmic over-expression of MUC1 in EOCs and suggested that the MUC1 might be a prognostic marker in EOC [114]. Using MY.1E12T MAb, Tamada et al. demonstrated that the increased expression of MUC1with sialoglycans is associated with advanced stage of clear cell EOC and also involved in chemoresistance and apoptosis [115]. Our recent findings with C595 MAb are in accordance with the previous reports from Ichige’s report and confirm the report from Feng’s study. No obvious difference was found in immunoreactivity with C595 MAb between frozen and paraffin-embedded sections [16]. However, the structure of stained sections is more easily identified in paraffin-embedded sections. Our results support that the epitope of MUC1 detected by C595 MAb is a useful therapeutic target for EOC therapy, especially for late-stage metastatic disease.

The 214D4 MAb recognises human MUC1 irrespective of its glycosylation pattern, SM3 MAb recognises the differentiation-dependent glycoforms and 5E5 MAb exclusively recognises tumour-associated glycoforms of MUC1 (MUC1-associated Tn and STn). With 214D4, SM3 and 5E5 MAbs, Van Elssen et al. demonstrated that aberrantly glycosylated, tumour-associated MUC1 is expressed in situ in most EOCs, but not in ovarian surface epithelium and serous cystadenomas, and the cancer-associated MUC1 epitopes are not only present in primary tumours but also in ovarian endometriosis, precursor lesions and in metastatic lesions [17]. In addition to binding EOC tissue, both 214D4 and SM3 MAbs also bind normal ovarian tissues and benign ovarian tissues, whereas 5E5 MAb only binds EOC tissues. These results suggest that 5E5 MAb is promising for EOC diagnosis and therapy.

The considerable variability in MUC1 expression reported in normal ovaries, benign and EOC tissues is dependent on differences in methodology such as antibody specificity, processing of tissue samples and the complexity of the disease itself or the small number of cases included in the earlier studies. Using a glycosylation-independent antibody may avoid the staining difference. These results support that certain epitopes of MUC1 with a high positive rate in EOC tissues and negative staining in normal ovaries such as C595 and 5E5 MAbs are useful therapeutic targets for EOC therapy, especially for metastatic recurrent disease.

6 MUC1 as a therapeutic target for EOC therapy

New strategies that target selected TAAs and result in an effective therapeutic index are needed for metastatic, recurrent EOC. MUC1 received the second highest priority score (after WT1) in a recent ranking of 75 tumour antigens based on pre-defined and pre-weighted criteria (including, among others, therapeutic function, immunogenicity, oncogenicity and specificity), emphasizing its potential for future translational studies and vaccine development [116]. The use of MAbs in the treatment of EOC in preclinical studies and clinical trials has been recently reviewed [8, 117, 118]. In this section, we summarize the recent progress in MUC1-associated treatments in preclinical and clinical studies in EOC and explore some emerging innovative approaches to use MUC1-targeted strategies for the management of the EOC.

6.1 C595 MAb in EOC targeting therapy

C595 is a murine anti-MUC1 core protein IgG3 MAb. C595 MAb has been labelled with γ-emitting radioisotope (111In) to test its capacity for cancer localization and identification in 19 patients with a clinical suspicion of ovarian malignancy and achieved final accuracies of 79 and 64 % compared with magnetic resonance imaging and ultrasound in relation to the final tumour histology [119]. These conjugates can be safely administered intravenously despite the fact that it is a murine MAb and capable of inducing a human anti-mouse antibody (HAMA) response. The targeting MUC1 using C595 MAb in preclinical studies in EOC is summarized in Table 3. After labelling with α-emitter (213Bi), our group demonstrated that 213Bi-C595 α-conjugate (AC) could be used to specifically target single MUC1+ EOC cells in vitro [120] and regress an OVCAR-3 i.p. ascites animal model in vivo [121]. This AC may have a good future for a clinical trial for the control of metastatic EOC.

Table 3 MUC1 targeting approaches for EOC treatment in preclinical studies and clinical trials
Full size table

MUC1 has been often studied as a target for antibody-based immunotherapy. In our recent studies, we also demonstrated that C595 MAb (targeting MUC1) alone could kill EOC cells in a dose-dependent manner and that low-dose C595 MAb combined with docetaxel (DTX) increased the sensitivity of several EOC cell lines and induced apoptosis in EOC cell lines in vitro [122]. Using an OVCAR3 i.p. ascite animal model, we further confirmed that combining C595 MAb and DTX can effectively inhibit i.p. tumour growth and ascites production and prolong survival of animals in the mouse xenograft model; this combination treatment can target tumour-associated MUC1, reduce angiogenesis and induce apoptosis [123]. This combination approach may be a potent therapeutic agent against advanced, recurrent and metastatic EOC disease.

6.2 HMFG1 MAb in EOC targeting therapy

The MAb HMFG1 is a murine MAb that recognises an extracellular portion of MUC1. The targeting MUC1 using MAb HMFG1 in preclinical studies and clinical trials in EOC is summarized in Table 3. A phase I trial using a murine anti-MUC1antibody (HMFG1) was conducted in 26 patients with persistent/recurrent EOC following platinum-based chemotherapy [124]. While no clinical responses were achieved, anti-HMFG1 and anti-MUC1 antibody responses were significantly elevated in those individuals completing the vaccination regimen. Radiolabelled HMFG1 has been previously used to image MUC1+ tumours in patients with primary and metastatic ovarian cancer with minor adverse effects [125]. The therapeutic efficacy of radioimmunotherapy with the murine HMFG1 MAb labelled with three different radionuclides (90Y, 186Re and 131I) was assessed in athymic BALB/c mice with intraperitoneally growing: OVCAR-3 EOC xenografts [126]. Each of the three radiolabelled immunoconjugate preparations, when given i.p., caused significant delay in ascites formation and death. The therapeutic efficacy of radiolabelled HMFG1 in the treatment of EOC has been evaluated in a phase I/II trial. The phase I/II trials with 90Yttrium-labeled HMFG1 (90Y-HMFG1) (up to 25 mCi per patient) showed that the agent was generally well-tolerated when injected i.p. [13, 127]. The disease-free survival of EOC patients treated with i.p. 90Y-HMFG1 was prolonged compared with that of matched historical controls [128].

Based on these promising results, a phase III trial has been undertaken. The Study of MAb Radioimmunotherapy (SMART) was a multicenter, randomized prospective trial of i.p. 90Y-HMFG1 (18–30 mCi, n = 224 patients) vs. standard treatment (n = 223 control patients) in ovarian cancer patients in complete clinical remission (CCR) [129]. Patients were followed for a median time of 3.5 years. This study did not show an improvement in time to relapse or overall survival (OS) [129], and there was no significant difference in terms of time to relapse and OS between the two treatment groups [129, 130]. Reported side effects of i.p. HMFG1 were nausea, fatigue, arthralgia, myalgia, thrombocytopenia and neutropenia [129]. Although there was no significant difference in time to relapse and OS in the SMART study, interestingly, there was a significant difference in pattern of disease recurrence [130]. Time to relapse was significantly longer in patients that were treated with i.p. 90Y-HMFG1, whereas significantly more extraperitoneal relapses were seen in the treatment arm compared to the standard arm (49 vs. 14 %). This observation suggests that i.p. 90Y-HMFG1 leads to i.p. disease control in ovarian cancer patients in CCR [130]. Further analysis of the data gathered in the SMART study considering the immune response of participating patients is still ongoing.

6.3 MUC1-related vaccine immunotherapy in EOC treatment

In addition to TAA-targeted therapies, vaccine approaches in EOC have also utilized whole tumour cell lysates and dendritic cells (DCs) in an attempt to boost host anti-tumour immune responses. In a pilot study, autologous DCs were pulsed with HER-2/neu or MUC1-derived peptides and administered to ten patients with advanced breast or EOC [131]. Half of the patients experienced peptide-specific cytotoxic T lymphocytes responses, but unfortunately these responses were not correlated with long-term outcomes. Sialyl-Tn (STn) is a carbohydrate associated with the MUC1 mucin on breast and ovarian cancer and is an ideal candidate for vaccine immunotherapy. With the Sialyl-Tn-keyhole limpet hemocyanin STn-KLH vaccine (Theratope), Holmberg et al. have tested 17 patients with stage III/IV EOC in a phase I trial and found humoral and cellular responses with minimal side effects. They concluded that Theratope may be a useful addition to high-dose chemotherapy regimens [132]. These results suggested that patients generate immunity against MUC1 produced by their tumours and defined MUC1 as a TAA and candidate for cancer therapy (vaccines) [133]. In a pilot study, a heptavalent vaccine containing MUC1 peptide conjugated with keyhole limpet haemocyanin and various mucin-derived carbohydrate epitopes, in combination with QS21 (a Quillaja saponaria saponin used as an immunological adjuvant), safely induced antibody response in patients with EOC [134].

Budiu et al. recently generated triple Tg MUC1KrasPten mice that express physiological levels of human MUC1 as self-antigen and progress to MUC1-overexpressing ovarian epithelial tumours upon intrabursal administration of Cre recombinase-encoding adenovirus (AdCre) [44]. This EOC model expresses strongly elevated MUC1 levels, and the primary tumours metastasize loco-regionally and are accompanied by high serum MUC1, closely mimicking the human disease. Vaccination of MUC1KrasPten mice with type 1 polarized dendritic cells (DC1) loaded with a MUC1 peptide (DC1–MUC1) can circumvent tumour-mediated immune suppression in the host, activate multiple immune effector genes and effectively prolong survival [44]. MUC1-related vaccine immunotherapy in EOC treatment is summarized in Table 3.

Recently, a humanized variant of the murine HMFG1, AS1402, has been developed and is now being studied in a phase II trial evaluating the efficacy of the addition of AS1402 to hormonal therapy in postmenopausal women with advanced breast cancer (www.antisoma.com). This humanized antibody is a potential treatment agent for patients with EOC.

6.4 Targeting MUC1 CT in cancer treatment

MUC1 is an important marker of malignancy and is a target for several immunotherapies currently under investigation [135]. MUC1 CT as a potential therapeutic target for ovarian cancer has been recently reviewed [15]. Targeting MUC1 CT and reducing its expression may have clinical significance for control of late stage and refractory EOC. However, MUC1-C has no kinase or enzymatic function that would allow for targeting a catalytic site. Therefore, one potential strategy is to disrupt MUC1 CT interactions with specific effectors that are linked to transformation. A potential approach for targeting MUC1 CT is through disruption of the MUC1 CT interaction with β-catenin. In this regard, a decoy GGSSLSY peptide was shown to block binding of MUC1 CT and β-catenin [67].

Raina et al. have reported that a MUC1 inhibitor (GO-201) that binds to the MUC1 cytoplasmic domain could target the MUC1 oncoprotein and effectively induce human breast cancer cell death in vitro and in tumour animal models [136]. The mechanism for the GO-201 is that targeting MUC1 CT oligomerization can block nuclear localization of MUC1 CT and induce growth arrest and death of breast cancer cells [136]. Bitler et al. also found that an intracellular protein MUC1 inhibitory peptide blocks both MUC1/β-catenin and MUC1/EGFR interactions; induces a significant reduction in proliferation, migration and invasion of metastatic breast cancer cells in vitro and inhibits tumour growth and recurrence in an established MDA-MB-231 SCID mouse model [137]. These MUC1 CT inhibitor and peptide may be promising for EOC treatment in the future.

6.5 Emerging innovation approaches for MUC1 targeting in EOC

Aptamers are single-strand oligonucleotides that can bind to target molecules with high affinity and specificity. Comparing to MAbs, aptamers possess distinctive advantages as a targeting ligand: high affinity for binding to most molecules, limited synthesis cost, low-immunogenicity and small size that allows it to penetrate solid tumours [138]. Due to these advantages, aptamers have been employed as novel targeting ligands in drug delivery systems against prostate cancer [139] and leukaemia [140]. For the TAA of MUC1, Ferreira et al. have developed several aptamers that could bind to the MUC1+ tumour cells [141]. It has also been shown that the MUC1 aptamers could be employed to selectively deliver phototherapy agent to cancer cells in vitro [142]. Savla et al. demonstrated the design and delivery of a tumour-targeted, pH-responsive quantum dot–mucin1 aptamer–doxorubicin (QD-MUC1-DOX) conjugate for the chemotherapy of EOC. The system demonstrated the ability to preferentially target EOC cells, efficiently released doxorubicin in acidic pH and had higher toxicity in EOC cells when compared with free doxorubicin as well as preferentially accumulated in EOC tumour xenografts [143]. Hu et al. developed an 86-base DNA aptamer (MA3) that bound to a peptide epitope of MUC1 and formulated an aptamer–doxorubicin complex by intercalating doxorubicin into the DNA structure of MA3. The complex can selectively deliver cytotoxic agent to MUC+ lung cancer and breast cancer cells in vitro [144]. Liu et al. recently demonstrated that a chimera that combines MUC1 aptamer and let-7i miRNA can specifically be delivered into OVCAR-3 cells and the released let-7i significantly sensitized the role of paclitaxel in inhibiting cell proliferation, inducing cell apoptosis and decreasing long-term cell survival [145]. These data confirmed that the MUC1 aptamer may be an effective alternative to conventional targeting agents such as peptides and antibodies, and data obtained have a high potential of the proposed conjugate in treatment of multidrug-resistant EOC.

MUC1 has been shown to serve as a natural ligand of galectin-3 in human colon cancer cells (Fig. 1). It was reported that the interaction between circulating galectin-3 and tumour-associated MUC1 enhances cancer cell-endothelial adhesion, promotes metastasis [146]. In the following study, the same research group further confirmed that the interaction between free circulating galectin-3 and tumour-associated MUC1 promotes embolus formation and survival of disseminating tumour cells in the circulation [147]. These data suggest that targeting the interaction between circulating galectin-3 with MUC1 in the circulation may represent an effective therapeutic approach for preventing metastasis. So far, the investigation of the interaction between galectin-3 and MUC1 in EOC has not been reported. It will be very interesting to work on this area in the future.

7 Conclusions and future perspectives

The finding that the over-expression of tumour-associated MUC1 in EOC tissues and its multiple biological functions contribute to cell–cell adhesion, signalling, migration, proliferation and differentiation of cancers indicates that the modulation of MUC1 in malignant cells may alter the pathogenesis of EOC, and the signal-transduction pathways that are regulated by cell surface tumour-associated MUC1 could be manipulated to alter the proliferation and differentiation status of EOC. Ligand receptor interactions for MUC1 and signal transduction events that are mediated by CT can be used as targets for future drug development that block the pathways to prevent EOC development.

CA125 (MUC16) is already being used in clinics as the only serum marker for EOC. MUC1 can also be used for the same purpose or combined with CA125 to improve sensitivity and specificity of the assay for EOC early diagnosis and monitoring cancer progression, especially for chemorefractory EOC.

The broad and stable expression of MUC1 in primary and metastatic EOC indicates that this protein is a promising target for future EOC therapy. The potential MUC1-directed therapy includes antibodies to inactivate MUC1 proteins, vaccination against tumour-specific MUC1, radiolabelled or toxin-conjugated anti- MUC1 MAbs, molecules that block the certain MUC1-related signalling pathways or proteins. Certain epitopes of MUC1 with high positive rate in EOC tissues should be considered for MUC1-targeted therapies. MUC1-aptamer-guided nanocarriers would be useful to selectively attach to EOC via the enhanced permeability and retention effect.

The potential of MAbs to complement current treatment in EOC is encouraging and may bring a significant improvement to the overall therapeutic outcomes currently being achieved in this disease. MAbs are multifunctional molecules that can target tumour cells, stimulate the immune system to attack tumour cells and engage receptor pathways effective in tumour cell destruction. MAbs directly targeting MUC1 combined with other modalities such as chemotherapy, small molecular inhibitors and signalling pathway inhibitors are promising for metastatic EOC. Although preclinical studies with anti-MUC1 MAbs have been promising, the clinical utility of these MAbs warrants further study.

MUC1-based vaccines could stimulate an immune response, in particular a cellular response, which can have a beneficial impact on EOC progression. However, MUC1 vaccination alone is not enough to tackle advanced and progressive EOC, but must be combined with more traditional therapies to have a significant impact. It will be worthwhile to test MUC1-based vaccines with chemotherapy or targeted therapy to control metastatic EOC in future study. Novel vaccine approaches targeting EOC-associated MUC1 such as peptide-based or DNA-based vaccines should be developed and might help eliminate non-resectable peritoneal metastases after surgical removal of primary EOCs.

Short half-life, short range and high linear energy transfer α particles from 213Bi have advantages over β particles. MUC1-specific MAbs or fragments of MUC1 such as scFv molecules could be conjugated with 213Bi to form RICs for targeted therapy. It is worthwhile investigating targeted α therapy or targeted α therapy used with other treatment strategies such as chemotherapy or small molecular inhibitors in future clinical trial for control of metastatic EOC, post-surgical minimal residual disease or micrometastatic lesions in peritoneal cavity with moderate to strong MUC1 over-expression.

It will be critically important to corroborate and clarify the contribution of the interactions between MUC1 and MUC1 CT-mediated signalling pathway proteins in EOC progression. Disruption of these interactions will elucidate the contribution of each process, and the intervention of these pathways may be helpful in controlling EOC progression. The clinical significance and functional roles of MUC1 CT in EOC metastasis and chemoresistance need to be further investigated in the future.