Search

Aberrant mucin-type O-linked glycosylation is a common occurrence in cancer where the upregulation of sialyltransferases is often seen leading to the early termination of O-glycan chains. Mucin-type O-linked glycosylation is not limited to mucins and occurs on many cell surface glycoproteins including EGFR, where the number of sites can be limited. Upon EGF ligation, EGFR induces a signaling cascade and may also translocate to the nucleus where it directly regulates gene transcription, a process modulated by Galectin-3 and MUC1 in some cancers. Here, we show that upon EGF binding, breast cancer cells carrying different O-glycans respond by transcribing different gene expression signatures. MMP10, the principal gene upregulated when cells carrying sialylated core 1 glycans were stimulated with EGF, is also upregulated in ER-positive breast carcinoma reported to express high levels of ST3Gal1 and hence mainly core 1 sialylated O-glycans. In contrast, isogenic cells engineered to carry core 2 glycans upregulate CX3CL1 and FGFBP1 and these genes are upregulated in ER-negative breast carcinomas, also known to express longer core 2 O-glycans. Changes in O-glycosylation did not significantly alter signal transduction downstream of EGFR in core 1 or core 2 O-glycan expressing cells. However, striking changes were observed in the formation of an EGFR/galectin-3/MUC1/β-catenin complex at the cell surface that is present in cells carrying short core 1-based O-glycans but absent in core 2 carrying cells.

Carbohydrate–protein interactions govern many crucial processes in biological systems including cell recognition events. We have used the sensitive force probe optical tweezers to quantify the interactions occurring between MGL lectins and MUC1 carrying the cancer-associated glycan antigens mucins Tn and STn. Unbinding forces of 7.6±1.1 pN and 7.1±1.1 pN were determined for the MUC1(Tn)—MGL and MUC1(STn)—MGL interactions, at a force loading rate of ~40 pN/s. The interaction strength increased with increasing force loading rate, to 27.1±4.4 and 36.9±3.6 pN at a force loading rate of ~ 310 pN/s. No interactions were detected between MGL and MUC1(ST), a glycoform of MUC1 also expressed by breast carcinoma cells. Interestingly, this glycan (ST) can be found on proteins expressed by normal cells, although in this case not on MUC1. Additionally, GalNAc decorated polyethylene glycol displayed similar rupture forces as observed for MUC1(Tn) and MUC1(STn) when forced to unbind from MGL, indicating that GalNAc is an essential group in these interactions. Since the STn glycan decoration is more frequently found on the surface of carcinomas than the Tn glycan, the binding of MUC1 carrying STn to MGL may be more physiologically relevant and may be in part responsible for some of the characteristics of STn expressing tumours. © 2017 Hadjialirezaei et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

We read with interest the manuscript by June and colleagues published recently in Immunity in which they describe targeting of aberrantly glycosylated tumor-associated cell membrane mucin MUC1 using chimeric antigen receptor-engineered human T cells (Posey et al., 2016). In that study, the authors used a second generation 4-1BB costimulatory-molecule-based chimeric antigen receptor (CAR) (Imai et al., 2004) in which targeting was achieved using a single-chain variable fragment (scFv) derived from the 5E5 antibody. This CAR selectively binds MUC1 that carries the Tn or sialyl (S)Tn glycan. Both of these truncated glycans are aberrantly expressed on the MUC1 glycoprotein in a spectrum of malignancies and consequently represent attractive targets for immunotherapeutic exploitation.

Siglec-9 is a sialic-acid-binding lectin expressed predominantly on myeloid cells. Aberrant glycosylation occurs in essentially all types of cancers and results in increased sialylation. Thus, when the mucin MUC1 is expressed on cancer cells, it is decorated by multiple short, sialylated O-linked glycans (MUC1-ST). Here we found that this cancer-specific MUC1 glycoform, through engagement of Siglec-9, 'educated' myeloid cells to release factors associated with determination of the tumor microenvironment and disease progression. Moreover, MUC1-ST induced macrophages to display a tumor-associated macrophage (TAM)-like phenotype, with increased expression of the checkpoint ligand PD-L1. Binding of MUC1-ST to Siglec-9 did not activate the phosphatases SHP-1 or SHP-2 but, unexpectedly, induced calcium flux that led to activation of the kinases MEK-ERK. This work defines a critical role for aberrantly glycosylated MUC1 and identifies an activating pathway that follows engagement of Siglec-9.

Mucin glycoproteins are major secreted or membrane-bound molecules that, in cancer, show modifications in both the mucin proteins expression and in the O-glycosylation profile, generating some of the most relevant tumour markers in clinical use for decades. Thus far, the identification of these biomarkers has been based on the detection of either the protein or the O-glycan modifications. We therefore aimed to identify the combined mucin and O-glycan features, that is, specific glycoforms, in an attempt to increase specificity of these cancer biomarkers. Using in situ proximity ligation assays (PLA) based on existing monoclonal antibodies directed to MUC1, MUC2, MUC5AC and MUC6 mucins and to cancer-associated carbohydrate antigens Tn, Sialyl-Tn (STn), T, Sialyl-Le(a) (SLe(a)) and Sialyl-Le(x) (SLe(x)) we screened a series of 28 mucinous adenocarcinomas from different locations (stomach, ampulla of Vater, colon, lung, breast and ovary) to detect specific mucin glycoforms. We detected Tn/STn/SLe(a)/SLe(x)-MUC1 and STn/SLe(a)/SLe(x)-MUC2 glycoforms in ≥50% of the cases, with a variable distribution among organs. Some new glycoforms-T/SLe(a)-MUC2, STn/T/SLe(a) SLe(x)-MUC5AC and STn/T/SLe(a)/SLe(x)-MUC6-were identified for the first time in the present study in a variable percentage of cases from different organs. In conclusion, application of the PLA technique allowed sensitive detection of specific aberrant mucin glycoforms in cancer, increasing specificity to the use of antibodies either to the mucin protein backbone or to the O-glycan haptens alone.

Chinese hamster ovary (CHO) cells are widely used for the production of recombinant proteins for clinical use as well as academic research. They are particularly important for the production of glycoproteins where bacteria cannot be used. TGFβ1 is a potent cytokine highly conserved across species with multiple immunological and non-immunological effects. We have discovered that CHOK1, the CHO clone most commonly used by the pharmaceutical industry, constitutively secretes latent TGFβ1 and that this hamster TGFβ1 is active on human cells inducing profound immunological effects. As far as we are aware, the production of TGFβ1 by CHOK1 cells has not been reported before in the literature. As TGFβ1 exerts powerful and pleiotropic effects on diverse cell types, and as CHO cells are used to produce a large number of clinical and non-clinical products, our findings are highly relevant to studies that rely on recombinant proteins.

Changes in glycosylation are common in malignancy, and as almost all surface proteins are glycosylated, this can dramatically affect the behavior of tumor cells. In breast carcinomas, the O-linked glycans are frequently truncated, often as a result of premature sialylation. The sialyltransferase ST3Gal-I adds sialic acid to the galactose residue of core 1 (Galbeta1,3GalNAc) O-glycans and this enzyme is over-expressed in breast cancer resulting in the expression of sialylated core 1 glycans. In order to study the role of ST3Gal-I in mammary tumor development, we developed transgenic mice that over-express the sialyltransferase under the control of the human membrane-bound mucin 1 promoter. These mice were then crossed with PyMT mice that spontaneously develop mammary tumors. As expected, ST3Gal-I transgenic mice showed increased activity and expression of the enzyme in the pregnant and lactating mammary glands, the stomach, lungs and intestine. Although no obvious defects were observed in the fully developed mammary gland, when these mice were crossed with PyMT mice, a highly significant decrease in tumor latency was observed compared to the PyMT mice on an identical background. These results indicate that ST3Gal-I is acting as a tumor promoter in this model of breast cancer. This, we believe, is the first demonstration that over-expression of a glycosyltransferase involved in mucin-type O-linked glycosylation can promote tumorigenesis.

Changes in the composition of glycans added to glycoproteins and glycolipids are characteristic of the change to malignancy. Sialyl-Tn (STn) is expressed by 25–30% of breast carcinomas but its expression on normal tissue is highly restricted. Sialyl-Tn is an O-linked disaccharide that can be carried on various glycoproteins. One such glycoprotein MUC1 is expressed by the vast majority of breast carcinomas. Both STn and MUC1 have been considered as targets for immunotherapy of breast cancer patients. Here we used different immunogens to target STn in an MUC1 transgenic mouse model of tumour challenge. We show that synthetic STn coupled to keyhole limpet haemocyanin (Theratope), induced antibodies to STn that recognised the glycan carried on a number of glycoproteins and in these mice a significant delay in tumour growth was observed. The protection was dependant on STn being expressed by the tumour and was antibody mediated. Affinity chromatography of the STn-expressing tumour cell line, followed by mass spectrometry, identified osteopontin as a novel STn-carrying glycoprotein which was highly expressed by the tumours. These results suggest that if antibodies can be induced to a number of targets expressed by the tumour cells, a humoral response can be effective in controlling tumour growth.

The outermost layer of all immune cells, the glycocalyx, is composed of a complex mixture of glycoproteins, glycolipids and lectins, which specifically recognize particular glycan epitopes. As the glycocalyx is the cell's primary interface with the external environment many biologically significant events can be attributed to glycan recognition. For this reason the rapidly expanding glycomics field is being increasingly recognized as an important component in our quest to better understand the functioning of the immune system. In this review, we highlight the current status of immune cell glycomics, with particular attention being paid to T- and B-lymphocytes and dendritic cells. We also describe the strategies and methodologies used to define immune cell glycomes.

To fulfil their function as APCs, dendritic cells (DC) and their precursors need to travel from blood to the peripheral tissues and, upon activation, migrate from tissues to draining lymph nodes. Because O-glycans play a role in T cell trafficking, we investigated the O-glycosylation profile of human monocyte-derived DC. Sialyl-Lewisx (sLex), a glycan involved in extravasation via selectin binding, was found to be expressed exclusively on P-selectin glycoprotein ligand-1 in monocytes and immature DC. However, sLex was lost from mature DC even though these cells retained expression of P-selectin glycoprotein ligand-1. Maturation of DC led to a rapid change in the expression of glycosyltransferases involved in O-linked glycosylation. A down-regulation of C2GnT1 mRNA and enzymatic activity was observed with a concurrent up-regulation of ST3Gal I and ST6GalNAc II mRNA resulting in a loss of the core 2 structures required for sLex expression as a P-selectin ligand. Interestingly, the early regulation of these glycosyltransferases was mediated by PGE2, which is known to be required for human DC migration. The pattern of O-glycosylation seen in mature cells was very similar to that expressed by naive T cells, which home to lymph nodes. Our data show that the regulation of O-glycosylation controls sLex expression, and also suggest that O-glycans may have a function in DC migration.

The molecular mechanism(s) underlying the enhanced self-interactions of mucins possessing the Tn (GalNAcα1-Ser/Thr) or STn (NeuNAcα2-6GalNAcα1-Ser/Thr) cancer markers were investigated using optical tweezers (OT). The mucins examined included modified porcine submaxillary mucin containing the Tn epitope (Tn-PSM), ovine submaxillary mucin with the STn epitope (STn-OSM), and recombinant MUC1 analogs with either the Tn and STn epitope. OT experiments in which the mucins were immobilized onto polystyrene beads revealed identical self-interaction characteristics for all mucins. Identical binding strength and energy landscape characteristics were also observed for synthetic polymers displaying multiple GalNAc decorations. Polystyrene beads without immobilized mucins showed no self-interactions and also no interactions with mucin-decorated polystyrene beads. Taken together, the experimental data suggest that in these molecules, the GalNAc residue mediates interactions independent of the anchoring polymer backbone. Furthermore, GalNAc-GalNAc interactions appear to be responsible for self-interactions of mucins decorated with the STn epitope. Hence, Tn-MUC1 and STn-MUC1 undergo self-interactions mediated by the GalNAc residue in both epitopes, suggesting a possible molecular role in cancer. MUC1 possessing the T (Galβ1-3GalNAcα1-Ser/Thr) or ST antigen (NeuNAcα2-3Galβ1-3GalNAcα1-Ser/Thr) failed to show self-interactions. However, in the case of ST-MUC1, self-interactions were observed after subsequent treatment with neuraminidase and β-galactosidase. This enzymatic treatment is expected to introduce Tn-epitopes and these observations thus further strengthen the conclusion that the observed interactions are mediated by the GalNAc groups.

Monocyte-derived dendritic cells (DC) are currently under extensive evaluation as cell vaccines for cancer treatment. The requirement for large-scale cell products demands optimized and standardized protocols. However, the yield of DCs from inoculated monocytes is reported to be always lower than 50%. In this present study we investigated whether this cell loss was caused by the properties of the starting population of inoculated monocytes. CD14 cells were enriched by immunomagnetic-bead selection and analyzed for apoptosis by an annexin V/propidium iodide assay. We found that 37.8+/-11.1% (n=8) of freshly isolated monocytes from buffy coats of healthy donors underwent programmed cell death. Further analysis of the fate of apoptotic cells during differentiation suggested phagocytosis. Monocytes were differentiated with GM-CSF and interleukin-4 into a viable, non-apoptotic population of immature dendritic cells. Addition of tumor necrosis factor-alpha and prostaglandin E2 resulted in fully matured dendritic cells, which were evaluated by phenotypic analysis and by allogeneic and MHC class-I-restricted T-cell responses. About 90.2+/-16.7% of the non-apoptotic monocyte population differentiated to viable matured dendritic cells. These results indicate that the yield of dendritic cells is mainly influenced by the percentage of apoptotic cells in the inoculum, and this has implications for DC generation in clinical applications.

We have developed an expression system for the production of large quantities of recombinant MUC1 mucin in CHO-K1 (Chinese-hamster ovary K1) cells. The extracellular part of human MUC1, including 16 MUC1 tandem repeats, was produced as a fusion protein with murine IgG Fc, with an intervening enterokinase cleavage site for the removal of the Fc tail. Stable MUC1–IgG-producing CHO-K1 clones were generated and were found to secrete MUC1–IgG into the culture medium. After adaptation to suspension culture in protein-free medium in a bioreactor, the fusion protein was secreted in large quantities (100 mg/l per day) into the culture supernatant. From there, MUC1 could be purified to homogeneity using a two-step procedure including enterokinase cleavage and ion-exchange chromatography. Capillary liquid chromatography MS of released oligosaccharides from CHO-K1-produced MUC1 identified the main O-glycans as Galβ1-3GalNAc (core 1) and mono- and di-sialylated core 1. The glycans occupied on average 4.3 of the five potential O-glycosylation sites in the tandem repeats, as determined by nano-liquid chromatography MS of partially deglycosylated Clostripain-digested protein. A very similar O-glycan profile and site occupancy was found in MUC1–IgG produced in the breast carcinoma cell line T47D, which has O-glycosylation typical for breast cancer. In contrast, MUC1–IgG produced in another breast cancer cell line, MCF-7, showed a more complex pattern with both core 1- and core 2-based O-glycans. This is the first reported production of large quantities of recombinant MUC1 with a breast cancer-like O-glycosylation that could be used for the immunotherapy of breast cancer.

Mucin glycoproteins on breast cancer cells carry shortened carbohydrate chains. These partially deglycosylated mucin 1 (MUC-1) structures are recognized by the monoclonal antibody SM3, which is being tested for its diagnostic utility. We used NMR spectroscopy to analyze the binding mode and the binding epitope of peptide and glycopeptide antigens to the SM3 antibody. The pentapeptide PDTRP and the glycopentapeptide PDT(O-alpha-D-GalNAc)RP are known ligands of the monoclonal antibody. The 3D structures of the ligands in the bound conformation were determined by analyzing trNOESY build-up rates. The peptide was found to adopt an extended conformation that fits into the binding pocket of the antibody. The binding epitopes of the ligands were determined by saturation transfer difference (STD) NMR spectroscopy. The peptide's epitope is predominantly located in the N-terminal PDT segment whereas the C-terminal RP segment has fewer interactions with the protein. In contrast, the glycopeptide is interacting with SM3 utilizing all its amino acids. Pro1 shows the strongest binding effect that slightly decays towards Pro5. The GalNAc residue interacts mainly via the N-acetyl residue while the other protons show less interactions similar to that of Pro5. The glycopeptide in the bound state also has an extended conformation of the peptide with the carbohydrate oriented towards the N-terminus. Docking studies showed that peptide and glycopeptide fit the binding pocket of the mAb SM3 very well.

In breast cancer, the O-glycans added to the MUC1 mucin are core 1- rather than core 2-based. We have analyzed whether competition by the glycosyltransferase, ST3Gal-I, which transfers sialic acid to galactose in the core 1 substrate, is key to this switch in MUC1 glycosylation that results in the expression of the cancer-associated SM3 epitope. Of the three enzymes known to convert core 1 to core 2, by the addition of GlcNAc to GalNAc in core1 C2GnT1 is the dominant enzyme expressed in normal breast tissue. Expression of C2GnT1 is low or absent in around 50% of breast cancers, whereas expression of ST3Gal-I is consistently increased. Mapping of ST3Gal-I and C2GnT1 within the Golgi pathway showed some overlap. To examine functional competition, the enzymes were overexpressed in T47D cells, which normally make core 1-based structures, have no detectable C2GnT1 activity and express the SM3 epitope. Overexpression of C2GnT1 resulted in loss of binding of SM3 to MUC1, accompanied by a decrease in the GalNAc/GlcNAc ratio, indicative of a switch to core 2 structures. Transfection of a C2GnT1 expressing line with ST3Gal-I restored SM3 binding and reduced GlcNAc incorporation into MUC1 O-glycans. Thus, even when C2GnT1 is expressed, the O-glycans added to MUC1 become core 1-dominated structures, provided expression of ST3Gal-I is increased as it is in breast cancer.

Glycosylation is a very important posttranslational modification of many biologically relevant molecules. A change in the structure of glycans added to glycoproteins and glycolipids is a common feature of the change to malignancy. With the cloning of many of the glycosyltransferases and the identification of specific target molecules, it is now possible to define these changes at the molecular level and to dissect the mechanisms involved. Within the mammary gland, mucin-type O-linked glycosylation has been studied most extensively. In normal resting, pregnant and lactating breast, mucin O-glycans are largely extended (core 2 type) structures. In contrast, mucin O-glycans found in breast carcinomas are often truncated (core 1 type). One mechanism that is responsible for this increase in core 1 structures is a change in the expression of glycosyltransferases, particularly an increase in the expression of the sialyltransferase, ST3Gal-I. The loss, at least to some degree, of core 2 based glycans is a consistent feature of MUC1 mucin when it is expressed by mammary tumours as demonstrated by the unmasking of the SM3 epitope in greater than 90% of breast carcinomas.

The anti-breast tumour antibody SM3 has a high selectivity in reacting specifically with carcinoma-associated mucin. SM3 recognises the core repeating motif (Pro-Asp-Thr-Arg-Pro) of aberrantly glycosylated epithelial mucin MUC1, and has potential as a therapeutic and diagnostic tool. Here we report the crystal structure of the Fab fragment of SM3 in complex with a 13-residue MUC1 peptide antigen (Thr1P-Ser2P-Ala3P-Pro4P-Asp5P-Thr6P -Arg7P-Pro8P-Ala9P-Pro10P-Gly11P- Ser12P-Thr13P). The SM3-MUC1 peptide structure was solved by molecular replacement, and the current model is refined at 1.95 A resolution with an R-factor of 21.3% and R-free 28.3%. The MUC1 peptide is bound both by non-polar interactions and hydrogen bonds in an elongated groove in the antibody-combining site through interactions with Complimentarity Determining Regions (CDRs), three of the light chain (L1, L2, L3) and two of the heavy chain (H1 and H3). The conformation of the peptide is mainly extended with no discernable standard secondary structure. There is a single non-proline cis-peptide bond in H3 (Val95H-Gly96H-Gln97H-Phe98H-Ala101H-Ty r102H) between Gly96H and Gln97H, which appears to play a role in SM3-peptide antigen interactions, and represents the first such example within an antibody hypervariable loop. The SM3-MUC1 peptide structure has implications for rational therapeutic and diagnostic antibody engineering.

A fourth human UDP-GalNAc:polypeptideN-acetylgalactosaminyltransferase, designated GalNAc-T4, was cloned and expressed. The genomic organization of GalNAc-T4 is distinct from GalNAc-T1, -T2, and -T3, which contain multiple coding exons, in that the coding region is contained in a single exon. GalNAc-T4 was placed at human chromosome 12q21.3-q22 by in situ hybridization and linkage analysis. GalNAc-T4 expressed in Sf9 cells or in a stably transfected Chinese hamster ovary cell line exhibited a unique acceptor substrate specificity. GalNAc-T4 transferred GalNAc to two sites in the MUC1 tandem repeat sequence (Ser in GVTSA and Thr in PDTR) using a 24-mer glycopeptide with GalNAc residues attached at sites utilized by GalNAc-T1, -T2, and -T3 (TAPPAHGVTSAPDTRPAPGSTAPPA, GalNAc attachment sites underlined). Furthermore, GalNAc-T4 showed the best kinetic properties with an O-glycosylation site in the P-selectin glycoprotein ligand-1 molecule. Northern analysis of human organs revealed a wide expression pattern. Immunohistology with a monoclonal antibody showed the expected Golgi-like localization in salivary glands. A single base polymorphism, G1516A (Val to Ile), was identified (allele frequency 34%). The function of GalNAc-T4 complements other GalNAc-transferases in O-glycosylation of MUC1 showing that glycosylation of MUC1 is a highly ordered process and changes in the repertoire or topology of GalNAc-transferases will result in altered pattern of O-glycan attachments.

Mucin-type O-glycosylation is initiated by UDP-N-acetylgalactosamine:polypeptide N-acetylgalactosaminyltransferases (GalNAc-transferases). The role each GalNAc-transferase plays in O-glycosylation is unclear. In this report we characterized the specificity and kinetic properties of three purified recombinant GalNAc-transferases. GalNAc-T1, -T2, and -T3 were expressed as soluble proteins in insect cells and purified to near homogeneity. The enzymes have distinct but partly overlapping specificities with short peptide acceptor substrates. Peptides specifically utilized by GalNAc-T2 or -T3, or preferentially by GalNAc-T1 were identified. GalNAc-T1 and -T3 showed strict donor substrate specificities for UDP-GalNAc, whereas GalNAc-T2 also utilized UDP-Gal with one peptide acceptor substrate. Glycosylation of peptides based on MUC1 tandem repeat showed that three of five potential sites in the tandem repeat were glycosylated by all three enzymes when one or five repeat peptides were analyzed. However, analysis of enzyme kinetics by capillary electrophoresis and mass spectrometry demonstrated that the three enzymes react at different rates with individual sites in the MUC1 repeat. The results demonstrate that individual GalNAc-transferases have distinct activities and the initiation of O-glycosylation in a cell is regulated by a repertoire of GalNAc-transferases.

The alpha2,3 sialyltransferase, alpha2,3 SAT (O), catalyzes the transfer of sialic acid to Galbeta1,3 N-acetyl-D-galactosamine (GalNAc) (core-1) in mucin type O-glycosylation, and thus terminates chain extension. A Core-2 branch can also be formed from core-1 by the core-2 beta1,6 N-acetyl-d-glucosamine transferase (beta1,6 GlcNAc T) that leads to chain extension. Increased levels of the alpha2,3 SAT (O) and decreased levels of the core-2 beta1,6 GlcNAc T are seen in breast cancer cells and correlate with differences in the structure of the O-glycans synthesized (Brockhausen et al., 1995; Lloyd et al., 1996). Since in mucin type O-glycosylation sugars are added individually and sequentially in the Golgi apparatus, the position of the transferases, as well as their activity, can determine the final structure of the O-glycans synthesized. A cDNA coding for the human alpha2,3 SAT (O) tagged with an immunoreactive epitope from the myc gene has been used to map the position of the glycosyltransferase in nontumorigenic (MTSV1-7) and malignant (T47D) breast epithelial cell lines. Transfectants were analyzed for expression of the enzyme at the level of message and protein, as well as for enzymic activity. In T47D cells, which do not express core-2 beta1,6 GlcNAc T, the increased activity of the sialyltransferase correlated with increased sialylation of core-1 O-glycans on the epithelial mucin MUC1. Furthermore, in MTSV1-7 cells, which do express core-2 beta1,6 GlcNAc T, an increase in sialylated core-1 structures is accompanied by a reduction in the ratio of GlcNAc: GalNAc in the O-glycans attached to MUC1, implying a decrease in branching. Using quantitative immunoelectron microscopy, the sialyltransferase was mapped to the medial- and trans-Golgi cisternae, with some being present in the TGN. The data represent the first fine mapping of a sialyltransferase specifically active in O-glycosylation and demonstrate that the structure of O-glycans synthesized by a cell can be manipulated by transfecting with recombinant glycosyltransferases.

Aberrant glycosylation is a common feature of malignant change. Changes in mucin-type O-linked glycosylation in breast cancer can result in the expression of truncated core 1-based sialylated glycans rather than the core 2-based glycans observed in normal mammary epithelium cells. This has been shown, in part, to be due to changes in the expression of glycosyltransferases, including the up-regulation of some sialyltransferases. Using the breast cancer cell line T47D, we have shown that PGE2, one of the final products of the cyclooxygenase-2 (COX-2) pathway, can induce the mRNA expression of the sialyltransferase α-2,3-sialyltransferase-3 (ST3Gal-I), resulting in increased sialyltransferase activity, demonstrated by a reduction in PNA lectin staining. Induction of COX-2 in the MDA-MB-231 breast cancer cell line also results in the increased expression of ST3Gal-I, leading to increased sialylation of the substrate of ST3Gal-I, core 1 Galβ1,3GalNAc. This effect on sialylation could be reversed by the selective COX-2 inhibitor celecoxib. The use of siRNA to knock down COX-2 and overexpression of COX-2 in MDA-MD-231 cells confirmed the involvement of COX-2 in the up-regulation of ST3Gal-I. Moreover, analysis of the expression of ST3Gal-I and COX-2 by 74 primary breast cancers showed a significant correlation between the two enzymes. COX-2 expression has been associated with a number of tumors, including breast cancer, where its expression is associated with poor prognoses. Thus, these results suggest the intriguing possibility that some of the malignant characteristics associated with COX-2 expression may be via the influence that COX-2 exerts on the glycosylation of tumor cells.

MUC-1 mucin is considered to be aberrantly glycosylated in breast, ovary, and other carcinomas in comparison with mucin from corresponding normal tissues. In order to clarify these differences in glycosylation, we have compared the O-linked carbohydrate chains from MUC-1 immunoprecipitated from [3H]GlcN-labeled breast epithelial cell lines (MMSV1-1, MTSV1-7, and HB-2) derived from cells cultured from human milk, with three breast cancer cell lines (MCF-7, BT-20, and T47D). Analysis by high pH anion chromatography showed that the normal cell lines had a higher ratio of GlcN/GalN and more complex oligosaccharide profiles than the cancer cell lines. Structural analyses were carried out on the oligosaccharides from MTSV1-7 and T47D MUC-1, and the following structures were proposed. MUC-1 from T47D had rather a simple glycosylation pattern, with NeuAcalpha2-3Galbeta1-3GalNAc-ol, Galbeta1-3GalNAc-ol, and GalNAc-ol predominating; in contrast, MUC-1 from MTSV1-7 had more complex structures, including a number of disialo, core 2 species, i.e. NeuAcalpha2-3Galbeta1-4GlcNAcbeta1-6[NeuAcalpha2 -3Galbeta1-3]GalNAc- ol and NeuAcalpha2-3Galbeta1-4GlcNAcbeta1-6[NeuAcalpha2 -3Galbeta1-4GlcNAcbet a1-3Galbeta1-3]GalNAc-ol. Double-labeling experiments with [3H]GlcN and 14C-aminoacids and analysis of GalNAc or GalNAc-ol:protein ratios in MUC-1 showed that there was also a significant difference in the degree of glycosylation of the mucin between the two cell types. We conclude that MUC-1 from breast cancer cell lines has simpler, and fewer, carbohydrate chains than MUC-1 from normal breast epithelial cells, and that these differences, combined or separately, explain the differential tumor specificity of some MUC-1 antibodies and T cells.

The glycome acts as an essential interface between cells and the surrounding microenvironment. However, changes in glycosylation occur in nearly all breast cancers, which can alter this interaction. Here, we report that profiles of glycosylation vary between ER-positive and ER-negative breast cancers. We found that genes involved in the synthesis of sialyl-Lewis x (sLex; FUT3, FUT4, and ST3GAL6) are significantly increased in estrogen receptor alpha-negative (ER-negative) tumors compared with ER-positive ones. SLex expression had no influence on the survival of patients whether they had ER-negative or ER-positive tumors. However, high expression of sLex in ER-positive tumors was correlated with metastasis to the bone where sLex receptor E-selectin is constitutively expressed. The ER-positive ZR-75-1 and the ER-negative BT20 cell lines both express sLex but only ZR-75-1 cells could adhere to activated endothelial cells under dynamic flow conditions in a sLex and E-selectin–dependent manner. Moreover, L/P-selectins bound strongly to ER-negative MDA-MB-231 and BT-20 cell lines in a heparan sulfate (HS)–dependent manner that was independent of sLex expression. Expression of glycosylation genes involved in heparan biosynthesis (EXT1 and HS3ST1) was increased in ER-negative tumors. Taken together, our results suggest that the context of sLex expression is important in determining its functional significance and that selectins may promote metastasis in breast cancer through protein-associated sLex and HS glycosaminoglycans. Cancer Res; 71(24); 7683–93. ©2011 AACR.

The type of interaction between tumor-associated antigens and specialized antigen-presenting cells such as dendritic cells (DCs) is critical for the type of immunity that will be generated. MUC1, a highly O-glycosylated mucin, is overexpressed and aberrantly glycosylated in several tumor histotypes. This results in the expression of tumor-associated glycoforms and in MUC1 carrying the tumor-specific glycan Tn (GalNAcα1-O-Ser/Thr). Glycopeptides corresponding to three tandem repeats of MUC1, enzymatically glycosylated with 9 or 15 mol of GalNAc, were shown to specifically bind and to be internalized by immature monocyte-derived DCs (iDCs). Binding required calcium and the GalNAc residue and was competed out by GalNAc polymer and Tn-MUC1 or Tn-MUC2 glycopeptides. The macrophage galactose-type C-type lectin (MGL) receptor expressed on iDCs was shown to be responsible for the binding. Confocal analysis and ELISA done on subcellular fractions of iDCs showed that the Tn-MUC1 glycopeptides colocalized with HLA class I and II compartments after internalization. Importantly, although Tn-MUC1 recombinant protein was bound and internalized by MGL, the glycoprotein entered the HLA class II compartment, but not the HLA class I pathway. These data indicate that MGL expressed on iDCs is an optimal receptor for the internalization of short GalNAcs carrying immunogens to be delivered into HLA class I and II compartments. Such glycopeptides therefore represent a new way of targeting the HLA class I and II pathways of DCs. These results have possible implications in designing cancer vaccines. [Cancer Res 2007;67(17):8358–67]

The functional properties of glycoproteins are strongly influenced by their profile of glycosylation, and changes in this profile are seen in malignancy. In mucin-type O-linked glycosylation these changes can result in the production of mucins such as MUC1, carrying shorter sialylated O-glycans, and with different site occupancy. Of the tumor-associated sialylated O-glycans, the disaccharide, sialyl-Tn (sialic acid alpha2,6GalNAc), is expressed by 30% of breast carcinomas and is the most tumor-specific. The ST6GalNAc-I glycosyltransferase, which can catalyze the transfer of sialic acid to GalNAc, shows a highly restricted pattern of expression in normal adult tissues, being largely limited to the gastrointestinal tract and absent in mammary gland. In breast carcinomas, however, a complete correlation between the expression of RNA-encoding ST6GalNAc-I and the expression of sialyl-Tn is evident, demonstrating that the expression of sialyl-Tn results from switching on expression of hST6GalNAc-I. Endogenous or exogenous expression of hST6GalNAc-I (but not ST6GalNAc-II) always results in the expression of sialyl-Tn. This ability to override core 1/core 2 pathways of O- linked glycosylation is explained by the localization of ST6GalNAc-I, which is found throughout the Golgi stacks. The development of a Chinese hamster ovary (CHO) cell line expressing MUC1 and ST6GalNAc-I allowed the large scale production of MUC1 carrying 83% sialyl-Tn O-glycans. The presence of ST6GalNAc-I in the CHO cells reduced the number of O-glycosylation sites occupied in MUC1, from an average of 4.3 to 3.8 per tandem repeat. The availability of large quantities of this MUC1 glycoform will allow the evaluation of its efficacy as an immunogen for immunotherapy of MUC1/STn-expressing tumors.

The epithelial mucin MUC1 is one of the few tumour-associated antigens identified for breast cancer. Several MUC1-derived peptides binding HLA-A*0201 molecules have been identified that correspond to sequences outside the tandem repeat. Immunisation with some of these peptides induces protective antitumour immunity in mice. Another HLA-A*0201-binding peptide has been identified in a human system. We have evaluated the CD8+ T-cell responses to all these peptides using peripheral blood lymphocytes from breast cancer patients and normal donors. Specific CD8+ T-cell responses could be generated in vitro against some of these peptides but only after several rounds of in vitro restimulation, and they did not recognise human cells endogenously expressing the antigen. Nevertheless, T cells recognised by HLA-A*0201 tetramers carrying a peptide from the signal sequence (LLLLTVLTV) could be detected in the peripheral blood of some HLA-A*0201+ breast cancer patients but not in healthy adults. This peptide is the only one of those tested which was identified in the human system, and the results emphasize the potential problems involved in translation of data from laboratory animal models to the human system. © 2005 Wiley-Liss, Inc.

A recombinant mucin O-glycosylation reporter protein, containing 1.7 tandem repeats (TRs) from the transmembrane mucin MUC1, was constructed. The reporter protein, MUC1(1.7TR)-IgG2a, was produced in CHO-K1 cells to study the glycosylation of the MUC1 TR and the in vivo role of polypeptide-GalNAc-T4 glycosyltransferase. N-terminal sequencing of MUC1(1.7TR)-IgG2a showed that all five potential O-glycosylation sites within the TR were used, with an average density of 4.5 glycans per repeat. The least occupied site was Thr in the PDTR motif, where 75% of the molecules were glycosylated, compared to 88-97% at the other sites. This glycan density was confirmed by an alternative liquid chromatography-mass spectrometry (LC-MS) based approach. The O-linked oligosaccharides were released from MUC1(1.7TR)-IgG2a and analyzed by nano-LC-MS and LC-MS/MS. Four oligosaccharides were present, NeuAcalpha2-3Galbeta1-3GalNAcol, NeuAcalpha2-3Galbeta1-3(NeuAcalpha2-6)GalNAcol, Galbeta1-3(NeuAcalpha2-6)GalNAcol, and Galbeta1-3GalNAcol, the two first being most abundant. Coexpression of the human polypeptide-GalNAc-T4 transferase with MUC1(1.7TR)-IgG2a increased the glycan occupancy at Thr in PDTR, Ser in VTSA, and Ser in GSTA, supporting the function of GalNAc-T4 proposed from previous in vitro studies. The expression of GalNAc-T4 with a mutation in the first lectin domain (alpha) had no glycosylation effect on PDTR and GSTA but surprisingly gave a dominant negative effect with a decreased glycosylation to around 50% at the Ser in VTSA. The results show that introduction of glycosyltransferases can specifically alter the sites for O-glycosylation in vivo.

The MUC1 mucin is expressed on the luminal surface of most simple epithelial cells but in carcinomas, especially those of the breast and ovary, MUC1 is upregulated and aberrantly glycosylated. MUC1 contains a large amount of O-linked glycans which, in the mucin expressed by normal mammary epithelial cells, consist mainly of core 2 based structures carrying polylactosamine chains. However, the mucin expressed by breast carcinomas has shorter side-chains, often consisting of sialylated core 1 (Galbeta1-3GalNAc). in situ hybridization of primary breast tissue showed that a sialyltransferase (ST3Gal I), responsible for adding sialic acid to core 1 thereby terminating chain extension, is elevated in primary breast carcinomas when compared to normal or benign tissue. Furthermore, the level of mRNA expression encoding ST3Gal I is correlated to the intensity of staining seen with the antibody SM3, which specifically recognises underglycosylated, tumour associated MUC1. Thus, the aberrant glycosylation of MUC1 seen in breast carcinomas appears to be due, at least in part, to the elevation of ST3Gal I.

The product of the MUC1 gene, the polymorphic epithelial mucin (PEM) is aberrantly glycosylated in breast and other carcinomas, resulting in exposure of normally cryptic peptide epitopes. PEM expressed by breast cancer cells contains more sialylated O-glycans and has a lower GlcNAc content than that expressed by normal cells. The exposure of peptide epitopes is thus thought to be due to the sugar side chains being shorter on the tumour-associated mucin. To investigate possible mechanisms underlying the different pattern of glycosylation in breast cancer cells, we analysed the pathways involved in the biosynthesis of O-glycan chains of mucins in normal and cancerous mammary epithelial cells. An immortalized mammary epithelial cells line originating from normal human milk. MTSV1-7, and three human breast cancer cell lines, BT20, MCF-7 and T47D, were studied. Glycosyltransferase activities assembling, elongating and terminating O-glycan core-1 [Gal beta 1-3GalNAc alpha-R] and core-2 [GlcNac beta 1-6 (Gal beta 1-3) GalNAc alpha-R] were present in the normal mammary cell line. Many of the glycosyltransferase activities were also expressed at variable levels in breast cancer cells. However, a sialyltransferase activity (CMP-sialic acid Gal beta 1-3GalNAc alpha 3-sialyltransferase) was increased several fold in all three cancer cell lines. Moreover, mammary cancer cell lines BT20 and T47D have lost the ability to synthesize core-2, as shown by the lack of UDP-GlcNAc: Gal beta 1-3GalNAc (GlcNAc to GalNAc) beta 6-GlcNAc-transferase activity, which corresponded to the absence of the mRNA transcript. However, MCF-7 breast cancer cells expressed this enzyme. Thus, the mechanism for the exposure of peptide epitopes in BT20 and T47D cells is proposed to be the loss of core-2 branching leading to shorter, sialylated O-glycan chains. A different mechanism is proposed for MCF-7 breast cancer cells.

International audience; ABSTRACT: Glycosylation changes that occur in cancer often lead to the expression of tumour-associated carbohydrate antigens. In breast cancer, these antigens are usually associated with a poor prognosis and a reduced overall survival. Cellular models have shown the implication of these antigens in cell adhesion, migration, proliferation and tumour growth. The present review summarizes our current knowledge of glycosylation changes (structures, biosynthesis and occurrence) in breast cancer cell lines and primary tumours, and the consequences on disease progression and aggressiveness. The therapeutic strategies attempted to target tumour-associated carbohydrate antigens in breast cancer are also discussed.

Mucins, present on highly polarized, secretory epithelial cells, have gained prominence in recent years as many monoclonal antibodies selected for their reactivity on differentiated or normal tissues react with epitopes present on these molecules. Although difficult to analyze biochemically because of their large size and large amount of O-linked carbohydrate, recent cloning studies from a number of different laboratories have produced structural information of the core protein. In most cases cDNA clones were obtained from λgt11 expression libraries following the development of antibodies to the stripped core protein. Thus far, three human mucin partial or full-length cDNA clones have been characterized as well as the porcine submaxillary mucin and a Xenopus integumentary mucin. In each case a domain of the core protein was found to consist of tandem repeats of a defined length. Although no homology exists between the tandem repeats of the various mucin genes at either the DNA or protein level, the repeated sequences in each case code for molecules which could be highly O-glycosylated. One characteristic feature of mucins is the presence of between 50 and 90% carbohydrate which is linked to serines or threonines via an O-glycosidic linkage to N-acetylgalactosamine. A second characteristic is the presence of prolines which along with glycosylation help to provide the extended core protein structure characteristic of mucins. All of the predicted proteins coded for by the mucin clones contain these features.

The major epithelial mucin, which is expressed by more than 90% of breast carcinomas and epithelial breast cancer cell lines, is MUC1. The MUC1 membrane mucin has a relatively simple structure, where a tandem repeat (TR) region carrying multiple O-glycans forms the major part of the extracellular domain. The gene shows a size polymorphism, due to variation in the number of tandem repeats, each of which has five potential glycosylation sites. MUC1 has been found to be a good model for studying mucin-type O-glycosylation both in vitro and in vivo. Changes in the pattern of glycosylation of MUC1 have been observed in breast cancer (Burchell et al. 2001) and have led to an interest in MUC1 as a possible target antigen for immunotherapeutic intervention (Taylor-Papadimitriou et al. 2000). The mucin expressed at high levels in the normal lactating mammary gland carries mainly core 2-based O-glycans (2–3 per tandem repeat). In breast cancers, the level of the ST3 Gal 1 enzyme is elevated (Burchell et al. 1999) and this is associated with increased numbers of Sialyl T O-glycans being added, as predicted from in vitro studies, showing competition for the core 1 substrate between this sialyl transferase and the relevant core 2 enzyme (Dalziel et al. 2001). Moreover, the TR domain is more densely glycosylated. While the Sialyl T O-glycan is commonly found on normal cells (e.g. resting T cells), it is cancer associated in the mammary gland. However, several lines of evidence indicate that this MUC1 glycoform, as expressed on breast cancers, can be immunosuppressive, maybe reflecting an inhibitory interaction with a lectin, possibly a siglec (Nath et al. 1999), on immune effector cells. In 25–30% of breast cancers, O-glycosylation is terminated very early so that GalNAc (Tn) or NeuAc α2,6 GalNAc (Sialyl Tn) glycans appear. Sialyl Tn is highly tumour-specific and is detected immunohistochemically only when the enzyme ST6GalNAcI is expressed. The Tn and Sialyl Tn will bind to different lectins, including MGL and could stimulate the immune response. Preliminary studies in mouse models suggest that an immunization protocol which includes MUC1-carrying Sialyl Tn O-glycans can protect against tumour growth, while clinical studies in breast cancer patients using immunogens based on the Sialyl Tn glycan are underway.

The polymorphic epithelial mucin (PEM) appears to be the target molecule for many monoclonal antibodies (MAbs) which react with tumour-associated and epithelium-specific antigens. PEM contains a large domain made up of 20 amino-acid tandem repeats which are highly immunodominant as many of the antibodies reactive with this molecule recognize epitopes within this area. Using overlapping peptide octamers, we have precisely mapped the epitopes of 4 MAbs reactive with the tandem repeats including one, SM-3, which shows enhanced tumour specificity. We report that the core of the SM-3 epitope corresponds to the continuous amino acid sequence Pro-Asp-Thr-Arg-Pro. We also show that the epitopes recognized by 3 other antibodies, which show reactivity with normal and malignant tissues, map to within this area of the tandem repeat. However, none of these epitopes contain the proline found at the amino end of the SM-3 determinant. These results are consistent with the idea that, in the cancer-associated mucin, premature termination of the carbohydrate side-chains results in the exposure of the SM-3 epitope.

Tissue wet weight, nucleic acid content and epithelial and stromal cell numbers were measured in the seminal vesicles of sexually mature male rats. After castration, tissue weight and RNA decreased rapidly and in aprallel to reach, after 14 days, values only 15-20% of those in control (not castrated) animals. During this period, DNA decreased to a much lesser extent (by about 40%), but this change in DNA correlates well with the observed loss of cells from the epithelium. Testosterone in vivo promoted an immediate resynthesis of RNA, the value characteristic of control animals being reached within 80h. Delays occurred in the hormone-induced regain of tissue weight (30h) and DNA (40h), each of which preceded proliferation of the epithelium (40-50h). The cells of the stroma were unaffected by these changes in the androgenic statls of the animal. It is suggested that these proliferative changes in the epithelium cannot account for the previously reported induction by testosterone of basic secretory proteins in this tissue.

The sequence complexity of mRNA from seminal vesicles has been examined in rats of different androgen status. Total cellular poly(A)-rich RNA was isolated from normal rats, rats castrated seven days previously and castrated rats induced with testosterone in vivo for three days. Radioactive DNA complementary to each RNA preparation was synthesised with reverse transcriptase from avian myeloblastosis virus and then hybridised to its own RNA in vast RNA excess. Computer analysis of the results suggests that in each case the data are best explained by assuming the RNA to be composed of three classes distinguished by their relative abundance. In normal and induced rats, approximately 40 % of the mRNA is accounted for by about one species represented up to 240000 times per cell. Castration has little effect on the proportion of total mRNA occupied by the highly abundant mRNA but greatly decreases the cellular concentration (3 1 100 copies per cell). In normal rats the remainder of the mRNA of the seminal vesicle comprises about 17 middle abundant sequences (15800 copies per cell) and at least 6000 scarce sequences (10 copies per cell). Heterologous hybridisations are virtually indistinguishable over their entire rot range. This suggests that androgen manipulation has little effect on the total sequence complexity of seminal vesicle mRNA and is accompanied by no gross differences in the relative abundance of the majority of mRNA species. In this respect, testosterone action in the seminal vesicle differs substantially from those of other steroid hormones in other steroid-responsive systems and even from that of testosterone itself in another accessory sexual tissue of the male rat, the ventral prostate.

Since the development of the first polyclonal antiserum to human milk fat globule (HMFG) membranes (1), this immunogen has been used by numerous workers to produce polyclonal and monoclonal antibodies. These reagents have been extremely useful in identifying markers of breast epithelial differentiation and have had a number of clinical applications. In addition, monoclonal antibodies to HMFG and to many tumour-associated antigens have brought into focus a group of highly immunogenic glycoproteins which have the characteristics of mucins and are expressed by many carcinomas. In this review we will discuss mainly work which has been done with the human milk fat globule, but it should be borne in mind that the same membranes and their components have been extensively studied in rodents and in the cow.

The levels of the human milk fat globule 1 and 2 antigens have been measured in the sera of patients with advanced breast cancer, using a "sandwich" type radioimmune assay which exploits the carbohydrate nature of the antigenic determinants. In the series studied, 30% of sera from advanced breast cancer patients contained elevated levels of the HMFG-I antigen as compared with 6% of sera from healthy control women, whereas 53% of the advanced breast cancer patients showed elevated levels of HMFG-2 antigen compared with 16,6% of the controls. By means of the immune blotting technique, the components carrying the antigenic determinants in sera have been identified and compared for size with those molecules expressing the determinants in primary and secondary breast tumours. Both antibodies react with similar molecular weight components of 320kd and 280kd which are present in serum and tumour samples, although lower molecular weight bands of 230kd and 190kd can be seen in some tumours. These components are much smaller than the glycoprotein (greater than 400kd) present in the human milk fat globule, which carries the antigenic determinants recognized by HMFG-I and 2.

Four hybridomas have been produced that secrete monoclonal antibodies to human beta-casein, a protein synthesized by fully differentiated breast epithelial cells. The antibodies have been characterized on immunoblots and have been shown to react with methanol: acetone-fixed sections of human lactating mammary gland. Immunoblots show that three of these antibodies react with casein whereas one, F20.10, also recognizes an epitope present on human alpha-lactalbumin. Computer analysis of the amino acid sequences of these two milk proteins reveals very little sequence homology, leading to the conclusion that the three-dimensional shape, rather than the primary sequence, is important in this cross-reactivity.

1. Two basic proteins were purified from secretions of rat seminal vesicles by using Sephadex G-200 chromatography and polyacrylamide-gel electrophoresis under denaturing conditions. 2. It is not certain that these two proteins are distinct species and not subunits of a larger protein, but their properties are similar. Highly basic (pI = 9.7), they migrate to the cathode at high pH and their amino acid composition shows them to be rich in basic residues and serine. Threonine and hydrophobic residues are few. Both proteins are glycoproteins and have mol.wts. of 17000 and 18500. 3. Together these two proteins account for 25-30% of the protein synthesized by the vesicles, but they are absent from other tissues. 4. Changes in androgen status of the animal markedly affect these proteins. After castration, a progressive decrease in the basic proteins is observed and the synthesis of the two proteins as measured by [35S]methionine incorporation in vitro is is decreased. Testosterone administration in vivo rapidly restores their rates of synthesis. 5. These effects on specific protein synthesis are also observed for total cellular protein, and it is suggested that testosterone acts generally on the total protein-synthetic capacity of the cell and not specifically on individual proteins. Proliferative responses in the secretory epithelium may also be involved. 6. The extreme steroid specificity of the induction process suggests that the synthesis of these basic proteins is mediated by the androgen-receptor system. 7. The biological function of these proteins is not clear, but they do not appear to be involved in the formation of the copulatory plug.

The MUC1 membrane mucin was first identified as the molecule recognised by mouse monoclonal antibodies directed to epithelial cells, and the cancers which develop from them. Cloning the gene showed that the extracellular domain is made up of highly conserved repeats of 20 amino acids, the actual number varying between 25 and 100 depending on the allele. Each tandem repeat contains five potential glycosylation sites, and between doublets of threonines and serines lies an immunodominant region which contains the epitopes recognised by most of the mouse monoclonal antibodies. The O-glycans added to the mucin produced by the normal breast are core 2 based and can be complex, while the O-glycans added to the breast cancer mucin are mainly core 1 based. This means that some core protein epitopes in the tandem repeat which are masked in the normal mucin are exposed in the cancer associated mucin. Since novel carbohydrate epitopes are also carried on the breast cancer mucin, the molecule is antigenically distinct from the normal breast mucin. (Changes in glycosylation in other epithelial cancers have been observed but are not so well documented.) Immune responses to MUC1 have been seen in breast and ovarian cancer patients and clinical studies have been initiated to evaluate the use of antibodies to MUC1 and of immunogens based on MUC1 for immunotherapy of these patients. The role of the carbohydrates in the immune response and in other interactions with the effector cells of the immune system is of particular interest and is discussed.

The mammary gland is a complex tissue containing several lineages, including the luminal or secretory epithelial cells and the basal or myoepithelial cells which line the ducts and alveoli. These two cell types can be distinguished in tissue sections by immunohistochemical staining with monoclonal antibodies to structural components and to functional products characteristic of each cell type. Such studies have focused attention on the luminal or secretory epithelial cell lineage, since the dominant cell in breast cancers expresses both the simple epithelial keratins and the mucin molecules expressed by the luminal epithelial cells in the normal gland1. In these particulars the breast cancer cell resembles other adenocarcinomas, such as those from the colon, lung and ovary.

Many existing MAbs raised against the human milk fat globule or against carcinoma cells recognise epitopes on high-molecular-weight glycoproteins. In a comparative ELISA assay a number of these antibodies have been shown to react with an extract of the human milk fat globule. In comparative immunoblots of cultured normal milk cells and breast cell lines, all were found to bind to large molecules which show some variation in molecular weight depending on the cell source. The HMFG-2 antibody, which is widely used in cancer diagnosis, also recognises epitopes on lower-molecular-weight components. In T47D cells these may be as small as 80,000 Mr, and with electron microscopy this cell line can be shown to accumulate HMFG-2-reactive components in the Golgi apparatus. Using an HMFG-2 affinity column we have immunopurified HMFG-2-reactive material from the 2 breast cancer cell lines MCF-7 and T47D and shown that all of the above antibodies react in a solid-phase ELISA with the purified material. In addition to high-molecular-weight components, the immunopurified material was found to contain lower-molecular-weight components including a glycoprotein of 68,000 Mr that was not, however, recognised by the HMFG-2 antibody on a Western blot. We have used this immunopurified material to generate new MAbs. All of these recognise the high-molecular-weight bands seen with the other antibodies, but 2 of them also recognise a band at 68,000 Mr in blots of MCF-7 and T47D. The second-generation antibodies show a spectrum of reactivity on tissues similar to HMFG-2 and one reacts at least as strongly as HMFG-2 with methanol-acetone-fixed sections of breast cancers.

Largely because of their complexity, the detailed structure of the mucins has been difficult to analyze. This group of compounds is categorized mainly by the fact that they contain a high level of carbohydrate which is attached in O-linkage to serine and/or threonine via the linkage sugar N-acetylgalactosamine. The mucous secretions produced by some epithelial cells, particularly those lining the gastro-intestinal tract and the lungs, contain mucins along with other products and these components have been studied for some time at the biochemical level. However, other glandular epithelial cells, such as the salivary gland, breast, ovary, endometrium, and sweat glands, also produce mucins, and some of these simpler mucins have recently received much attention. This is because many antibodies selected for epithelial or tumor specificity have been found to react with high molecular weight glycoproteins which are produced by simple epithelial cells and have the properties of mucins.1–9

Three hybridomas producing monoclonal antibodies (IgG), reacting with components of the human mammary milk fat globule have been isolated. When tested for binding to a wide range of human cell lines and strains, all three antibodies show negative reactions with fibroblasts, lymphoblastoid cells, and a large number of epithelial cell lines of non-breast origin. Two of the antibodies (1.10.F3 and 3.14.A3) reacted with seven out of eight breast cancer lines tested, and with epithelial cells cultured from human milk. The other antibody (3.15.C3) reacted with only two of the breast cancer cell lines.

The monoclonal antibody (MAb) SM-3, which was raised to chemically deglycosylated milk mucin, reacts with an epitope present on the core protein of this mucin which we have referred to as PEM (polymorphic epithelial mucin). Although this mucin is abundantly expressed by both the lactating breast and breast carcinomas, the antibody SM-3 shows very little or no reactivity on the former but does react with 92% of breast carcinomas. Furthermore, SM-3 stains primary carcinomas of the lung, colon and ovary, but on the corresponding normal tissue the epitope is expressed at a much reduced level or not at all. These results indicate that an epitope masked in the normal mucin is exposed in the mucin produced by tumour cells, perhaps due to aberrant glycosylation. An extensive immunohistochemical study of other normal tissues reveals that the majority show only weak focal staining with SM-3 or none at all, the distal tubules and collecting ducts of the kidney, and sebaceous glands being the only normal tissues studied to show homogeneously positive staining.

The nucleotide sequences of partial cDNA clones coding for the core protein of a human polymorphic epithelial mucin were determined, and a large domain was found to consist of a 60-base pair tandem repeat sequence. The cDNA clones were originally selected (Gendler, S. J., Burchell, J. M., Duhig, T., Lamport, D., White, R., Parker, M., and Taylor-Papadimitriou, J. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 6060-6064) using three monoclonal antibodies which show differential reactivity with the mucin produced by normal and malignant breast. Two of the epitopes are exposed in the normally processed and cancer-associated mucin, while one epitope is unmasked only in the cancer-associated mucin (Burchell, J. M., Durbin, H., and Taylor-Papadimitriou, J. (1983) J. Immunol. 131, 508-513; Burchell, J., Gendler, S., Taylor-Papadimitriou, J., Girling, A., Lewis, A., Millis, R., and Lamport, D. (1987) Cancer Res. 47, 5476-5482). We show here that all three antibodies react with a synthetic peptide with an amino acid sequence corresponding to that predicted by the tandem repeat. Identification of the epitopes preferentially expressed on the cancer-associated mucin should allow a directed approach to the development of tumor-specific antibodies using synthetic peptides as immunogens.

MUC1 is a highly attractive immunotherapeutic target owing to increased expression, altered glycosylation, and loss of polarity in >80% of human cancers. To exploit this, we have constructed a panel of chimeric Ag receptors (CAR) that bind selectively to tumor-associated MUC1. Two parameters proved crucial in optimizing the CAR ectodomain. First, we observed that the binding of CAR-grafted T cells to anchored MUC1 is subject to steric hindrance, independent of glycosylation status. This was overcome by insertion of the flexible and elongated hinge found in immunoglobulins of the IgD isotype. Second, CAR function was highly dependent upon strong binding capacity across a broad range of tumor-associated MUC1 glycoforms. This was realized by using an Ab-derived single-chain variable fragment (scFv) cloned from the HMFG2 hybridoma. To optimize CAR signaling, tripartite endodomains were constructed. Ultimately, this iterative design process yielded a potent receptor termed HOX that contains a fused CD28/OX40/CD3ζ endodomain. HOX-expressing T cells proliferate vigorously upon repeated encounter with soluble or membrane-associated MUC1, mediate production of proinflammatory cytokines (IFN-γ and IL-17), and elicit brisk killing of MUC1+ tumor cells. To test function in vivo, a tumor xenograft model was derived using MDA-MB-435 cells engineered to coexpress MUC1 and luciferase. Mice bearing an established tumor were treated i.p. with a single dose of engineered T cells. Compared with control mice, this treatment resulted in a significant delay in tumor growth as measured by serial bioluminescence imaging. Together, these data demonstrate for the first time that the near-ubiquitous MUC1 tumor Ag can be targeted using CAR-grafted T cells.