Cummings/Ju - Harvard; Emory

Principal Investigator: Richard D Cummings, Ph.D.
Institution: Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA

Principal Investigator: Tongzhong Ju, M.D., Ph.D.
Institution: Emory University, Atlanta, GA

Our project, The Tumor Antigens Tn and SialylTn in Human Colorectal Carcinoma, is a joint collaboration between our laboratories at Emory and with the laboratory of Dr. Lance Wells at the CCRC/University of Georgia.

In our studies we have found that colorectal cancer cells exhibit increased expression of a specific O-glycan biomarkers termed the Tn antigen, and its sialylated derivative SialylTn, which are found on numerous membrane glycoproteins and in secreted mucins. Expression of these antigens often occurs at an early stage of colon carcinogenesis, and is associated with poor prognosis and tumor metastasis; however, the molecular basis for expression of these antigens is not well understood. We recently found that expression of the Tn/SialylTn antigens can arise from loss-of-function of an enzyme termed the T-synthase, which required in all human cells for normal O-glycans biosynthesis. More importantly, we also found that expression of the T-synthase in turn requires the specific expression of a molecular chaperone we discovered and termed Cosmc. The Cosmc gene is on the X-chromosome and we found that spontaneous, acquired mutations or copy loss of this gene can occur in tumor cells and is associated with decreased expression of the T-synthase and consequent expression of the Tn/SialylTn antigens. We believe that the abnormal expression of these antigens, which are not usually found on any normal glycoproteins in humans, is associated with altered glycoprotein functions, cell signaling, and adhesion, all of which may contribute to tumor progression and metastasis. Our preliminary studies also show that gastrointestinal tract epithelial cell-specific loss of Cosmc in mice causes Tn/SialylTn antigen expression in the small and large intestine.

In our studies supported by the Alliance of Glycobiologists for Detection of Cancer, we are exploring the hypothesis that Tn/SialylTn tumor antigens in colorectal carcinomas are novel glycan biomarkers for human colon cancer and arise from alterations in Cosmc or altered expression of T-synthase. Thus, we are studying the molecular basis for the expression of Tn/SialylTn and other tumor-associated carbohydrate antigens in human primary colon tumors, defining the glycan and glycopeptide profiles of colorectal carcinoma cells compared to normal cells to identify protein carriers of the Tn/Sialyl Tn antigens, as well as defining the molecular mechanism(s) for Tn/SialylTn expression in human primary colon tumors by identifying the alterations in Cosmc, and expression of other perhaps relevant related ?glycogenes?. Finally, we are also developing well-defined monoclonal antibodies that exhibit both high specificity and affinity to Tn or SialylTn antigens, which could be used in both diagnosis and treatment of colorectal carcinoma.

Synopsis of Research and Network Resources

Introduction: A Brief Summary of the State of the Science and Research Needs

A major issue in cancer biology is to understand the many factors that can influence the transformation fand growth rates of tumors, the changes in their environment, and how these contribute to clinical outcomes and therapy. Multiple factors, including genetic, epigenetic, and metabolic, as well as interactions of tumor cells with stroma, can all contribute to tumorigenesis and malignancy. The surfaces of all cells are covered in glycoconjugates, which represent the many receptors, adhesion molecules, transporters, and other molecules that are largely glycoproteins. Changes in glycosylation of proteins underlie key aspects of the genetic and phenotypic changes occurring in cells upon oncogenic transformation and tumor progression. There are many classes of glycans that undergo structural changes, along with changes in glycoprotein expression. One major class of great interest to cancer biology is O-glycans, which are now known to be linked to one or more Ser/Thr residues on proteins of cell surfaces or secreted by cells. It has been observed for many years that a common phenomenon in human and animal carcinomas is the altered expression and truncation of O-glycan structures. Two of the major truncations in O-glycosylation are represented by the Tn antigen (GalNAc?1-Ser/Thr) and the SialylTn antigen (NeuAc?2-6GalNAc?1-Ser/Thr). Neither of these antigens is expressed in normal tissue nor during development, and their expression is correlated with poor prognosis. While earlier studies suggested that expression of the Tn and SialylTn antigen might simply reflect altered metabolism or indirect changes in tumor cells which have little to do with tumorigenesis or metastasis, growing evidence indicates that expression of these antigens has severe biological consequences by altered functions of glycoproteins. These changes can lead to altered cell adhesion, signaling, redox states, interactions with the extracellular matrix, all of which may directly contribute to oncogenic transformation and tumor development.

The Tn and SialylTn antigens have been observed in many human carcinomas, including carcinomas of the pancreas, prostate, lung, cervix, esophagus, liver, ovary, breast, and colorectum. Our focus has been on the underlying contributions of Tn and SialylTn antigen to dysregulate cell adhesion and signaling, with special emphasis on colorectal cancer (CRC). CRC, including colon cancer and rectal carcinoma is one of the most common malignancies worldwide. The majority of CRC occurs spontaneously with only a minority of colon cancer cases having a familial history, such as hereditary nonpolyposis colorectal cancer (HNPCC). The Tn and SialylTn antigens are highly expressed in >70% of human CRC, and their expression is associated with the poor prognosis of the disease. Surprisingly, no animal model has been developed to study tumors in conjunction with expression of these glycan antigens. Therefore, a novel model to induce expression of the Tn/SialylTn antigens in the distal colon is necessary to explore the role of these antigens in tumor development and metastasis. A key finding in our work has been that the Tn and SialylTn antigen can arise from altered expression or localization of the T-synthase and/or its essential chaperone Cosmc, encoded by the X-linked gene on human Xq24 and the murine Xc3. In our studies under way, we have found that in mice engineered to express the Tn antigen in their GI tract due to deletion of Cosmc, the male animals spontaneously develop adenocarcinomas and heterozygous females are highly susceptible to induced tumors. We also recognize that new assays are needed to detect circulating glycoproteins carrying the Tn antigen as a biomarker of disease. We have a suite of monoclonal antibodies against the Tn antigen and are attempting to develop more of these, that will be specific to the Tn antigen, not cross-reactive with circulating IgA1, which is the only serum glycoprotein naturally known to have a glycoforms with the Tn antigen, as we have shown recently. Such monoclonal antibodies to the Tn and SialylTn antigen will provide a better method of detecting these antigens in patients and perhaps be useful in drug delivery and therapy. We are also developing monoclonal antibodies to the T-synthase and Cosmc, which will be useful in immunohistochemical assays and pathological assessments.

Laboratory specific studies to meet the research needs

We are studying the specific glycoproteins in human tumor specimens that express the Tn and SialylTn antigens, using modern glycoproteomic technologies. Our studies on human tumor specimens have shown that the majority of human carcinomas express the Tn and/or SialylTn antigens. Because studies of the O-glycans is of paramount importance, we are also developing new glycomic tools to allow identification of O-glycans in living cells and tissues using metabolic strategies. These new approaches will help to both define the relationship of the Tn and SialylTn antigens to cancer, as well as identify the common O-glycan biosynthetic pathways used in normal and transformed cells. We have a number of cell lines in which the Cosmc gene has been deleted, along with the cellular counterparts expressing normal Cosmc, through our collaborations with Dr. Henrik Clausen and his group in Copenhagen. These cells are highly useful in technology development underway in our laboratory. Our focus is to identify the O-glycosylated proteins in human tumors, as well as define the biosynthetic pathways for O-glycans and disruption in such pathways in cancer. In addition, we have prepared with this program and within the National Center for Functional Glycomics glycopeptide and glycoprotein microarrays, which are unique resources for identifying glycan antigens in the context of peptide and protein expression.

Resources and Reagents for sharing

Our laboratory is also part of the P41 National Institute of General Medical Sciences funded National Center for Functional Glycomics (NCFG). In addition, we are the headquarters for the glycan microarray services of the Consortium for Functional Glycomics. The NCFG is is an outgrowth of the highly successful Consortium for Functional Glycomics (CFG) where the development of glycan microarray technologies was a key resource. The mission of the NCFG is to foster technology development in the area of functional glycomics and array technologies, enhance biomedical research, enlarge collaborations in the field, and provide needed services, training, and data and technology dissemination. The NCFG has a number of glycan microarrays which have been highly useful in defining the glycan-binding specificity of many different anti-glycan antibodies, other types of glycan-binding proteins, including receptors, lectins, adhesins, hemagglutinins, etc. The NCFG has a wide capacity to collaborate and provide services to investigators interested in glycan-binding proteins, glycan microarrays, shotgun glycomics, preparations of glycans from glycoproteins, glycolipids, cells and tissues, and analyses of glycan structures. Emory has sophisticated mass spectrometry capability for glycan characterization and sequencing, including two Thermo Scientific Orbitrap Fusion Tribrid Mass Spectrometer with Electron Transfer Dissociation (ETD), a Thermo Scientific Quadrupole-Orbitrap Hybrid Mass Spectrometer, and a Bruker UltraFlex II MALDI/TOF Mass Spectrometer. In addition, the NCFG has a Biacore T100 for defining kinetics of binding and affinity constants for interactions. In addition, the NCFG has developed software for interpreting glycan-binding data from microarray studies, such as GlycanMotifMiner and GlycoPattern.

Public Health Implication/Advancing the Field of Glycobiology

Our results show that Tn and STn antigens are very specific to pathological states, especially carcinomas, and should be exploited to develop glycomarkers for various types of tumors. Specifically, we surveyed more than 90 cases of CRC samples available through the tumor cell arrays at the Winship Cancer Institute at Emory University School of Medicine, and found that more than 75% were expressing Tn and STn antigens; about ~25% of those adjacent tissues mainly expressed Tn antigen, indicating Tn antigen could be a pre-malignant glycan-tumor marker as well. Using our monoclonal antibody-based technologies, early diagnosis or screening, prognosis and even the targeted therapeutics can be developed. For example, the anti-Tn and STn monoclonal antibodies can be tested as therapeutic reagents in our mouse model with Cosmc-deletion in the GI-tract which spontaneously develop CRC.

Cancer-Specific Relevance: detection, prevention and treatment

In normal tissues or cells, the Tn antigen is always modified, branched majorly by the key glycosyltransferase T-synthase, although it can also be modified by another enzyme expressed exclusively in the GI tract epithelia, the core 3 ß3GnT. As a result, the Tn antigen is never exposed in normal tissues. In pathological situations, including the autoimmune disorders Tn –syndrome and carcinomas, the Tn antigen and its sialylated version STn antigen are abnormally expressed. We and others have shown that the somatic mutation or silencing of X-linked Cosmc is the cause of Tn syndrome, the alterations including the mutations and gene deletions are found in the human and mouse tumor cell lines and cervical cancers and lead to the Tn/STn expression, although there may be other molecular mechanisms underlying the abnormal O-glycan expression in human carcinomas. From extraction of literatures on the Tn expression in human tumors, Tn/STn antigens are often occurring on those solid tumors, mainly carcinomas. Consistent with this, from our own unpublished studies, we observed high expression of Tn antigens in tumors of colon, lung, breast, ovary, endometrium, and cervix, and we did not detect Tn expression in 16 chronic lymphocytic leukemia (CLL) samples. Therefore, Tn/STn antigens could be used for diagnosis, and prognosis for many types of carcinomas, by combination with other tumor markers, such as CA125 for ovarian cancer, CA15-3 for breast cancer, etc. Furthermore, by combination with MUC1 (CA15-3) in stool samples, Tn/STn antigens could be very useful for CRC screening. More excitingly, generating highly specific anti-Tn and anti-STn antibodies with high affinity could open a new era for targeted immunotherapies for multiple types of carcinomas.

Besides of the truncation of O-glycans such as Tn and STn antigens, and altered expression of other glycosyltransferases, the O-glycan structures may be mixed in containing a range of truncated and extended O-glycans, and are possibly altered in other ways, such as expression of Sialyl-Lea (CA19-9 antigen) and certainly other structures yet to be defined. With our newly developed novel technology for profiling and amplifying the O-glycome of any given cultured cells, the O-glycome of different types of tumors could be readily revealed, and the unique aberrant O-glycan structures for a certain type of tumor could be identified, characterized and could serve as potential glyco-biomarkers for human tumors.

Opportunities for Collaboration

The technology in the NCFG and technologies arising from studies in this program are available to all researchers. Please contact Dr. Richard D. Cummings (, Dr. Tongzhong Ju (, or Dr. David F. Smith ( (Technical Director of the NCFG) for information and consultation. Also, see the websites for the National Center for Functional Glycomics and the Consortium for Functional Glycomics.