Carbohydrate Antigen-bearing Nanoparticles for Antitumor Therapy

An established hallmark of tumorigenesis is the biosynthesis of aberrant glycan chains due to changes in the expression of glycoprocessing enzymes in tumor tissue. These aberrations become more marked as the tumor acquires a more aggressive phenotype. Tumor cell-surface carbohydrates play important roles in the motility and metastasis of many different cancer cells. In addition, many of these aberrant glycans are tumor-associated carbohydrate antigens (TACA) and have been used in the development of tumor vaccines. Since most of the cellular interactions with TACAs are not well understood, there is an urgent need to better characterize the specific molecular interactions that occur during these events. One feature of carbohydrate binding to macromolecules that is well understood is the concept of multivalency: Monomer carbohydrates bind to proteins very weakly while clustering of a monomer raises this affinity as much as a million-fold. We have prepared the important Thomsen-Friedenreich (Tf) antigen (Gal(beta)1-3GalNAc(alpha)-O-Ser/Thr) on very specific templates to take advantage of this so-called cluster glycoside effect. As mentioned in the last report, we have prepared gold self-assembled nanospheres and quantum dots containing sugar derivative and reported preliminary details on their function. The in vivo experiments with our gold nanospheres in mice were conflicting, so we retreated to basics and performed more rigorous characterization and explored a host of new syntheses that allowed for production of more uniform particles. We proceeded to systematically study the optimum procedure, from several related methods, that offered the highest quality particles with regards to stability and uniformity. We have now a new, optimized and reproducible synthesis for gold nanoparticles bearing the TF antigen attached to both eth threonine and serine amino acid residues with a capping acetate at the N-terminal amine group. Examination of these particles in various tumor cell lines for cytotoxicity has been performed and compared against each construct. These cells were chose to be either positive or negative for expression of the anti-apoptotic protein Galectin-3, an in vivo receptor for the TF antigen. Cells that are negative for Gal-3 do not respond to our particles, which suggest that cytotoxicity does go through a Gal-3 mediated pathway. We put a heavy emphasis on preparing particles that encompassed what we consider the best antigen, a glycopeptide from tumor associated cell-surface mucins, and combined that with various concentrations of linker and B-cell epitopes to construct particles that may act as novel immunogens. We prepared at least seven separate particles with various placements of the disaccharide on the peptide, and along with linker and a 28-residue portion of C3d, a domain of complement component 3 and a ligand of CD21, a B-cell surface protein that, when engaged, lowers the threshold of B-cell activation.. These particles were injected into mice and the sera were analyzed for immune responses. A statistically significant immune response was observed in at least two test groups, and animals we boosted a second time with fresh particles. Tumors were implanted and survival was followed. Although one specific antigen group did better than the others, they did not do better than the group that received only PBS. There are several parameters that could have led to a lower than desired response, and we are looking into these now. A full paper in Bioconjugate Chemistry was recently published on this work, and the antitumor activity of some of these constructs has been evaluated by several methods and this paper is written with Kate Rittenhouse-Olson and close to submission. Further work in this area for this research cycle has been to take the aforementioned best construct, called MUC4-5TF, and prepare polyclonal antibody sera. This construct as prepared by us was conjugated to KLH and mice were vaccinated by Precision Antibodies Inc. Titers against our specific glycopeptides antigen are in the tens-of-thousand range and the sera were given to Professor Pinku Mukherjee, at the University of North Carolina at Charlotte. This sera did not stain any of the tissue that was available, but we are looking at more appropriate tumor samples with distinct expression of aberrant MUC4. However, in this cycle, we tested the polyclonal sera for binding to our glycopeptide constructs that we put on a glycopeptide array developed by our collaborator Jeff Gildersleeve. This sera stained ONLY the glycopeptides we printed, and not to the TF antigen in ANY OTHER CONTEXT (there are about 30 on Jeff's array). Thus we have a company preparing a monoclonal antibody for us that we feel will be an extremely useful reagent for detection of aberrant MUC4 on tumor cells. The new study started in collaboration with Howard Young mentioned in the last annual report, continues to explore the modulation in cytokine profiles that is elicited by particles with varying antigens in different chemical guises. Initial data showed that levels of several cytokines from activated murine macrophages are either potentiated or attenuated with particles containing different surface chemistries. This was reexamined and refined to show that specifically, TNF-alpha expression was turned on much more with very specific glycopeptide constructs than others. We have prepared three new sets of particles of various sizes coated with our important antigens. These were examined in the macrophage system and showed a dramatic increase in cytokine expression with particle size. Thus particles from sizes of 3, 16, 25 and 40 nm have dramatically different effects on cytokine gene expression, which is also dependent on the ligand for activity. In addition, we tested al the constructs in a mutant MyD88 knockout macrophage cell line. MyD88 is an adapter protein involved in the signal transduction cascade for the expression of toll-like receptor proteins. The cytokine expression induced by the nanoparticles was greatly reduced in the knockout cell line, suggesting that toll-like receptors are involved in the process of cytokine gene expression by the nanoparticles. To complete this study, we are focusing on the MUC4-5TF construct (vide supra) to hone in on the actual chemical requirements for eliciting gene expression of inflammatory cytokines. The manuscript mentioned in the last report on the evaluation of the optimum precursors for the preparation gold nanoparticles with a variety of ligands was published in the Journal of Colloid and Interface Science this year. We have shown that some of the particles are cytotoxic through an apoptotic mechanism, and we now want to define why the multivalent platforms work in cells but not on simple proteins. This is probably due to the aggregation state of gal-3 being very different in cells and in while in a simple buffer solution, and we are designing experiments to help unravel this mystery. We are looking at various cell types that either express or do not express Gal-3 with Chand Khann of the CCR as well as performing antimetastaic studies in vivo in collaboration with the Small Animal Imaging Program of Leidos Biomedical. We have found that nano-constructs where the TF antigen is displayed on a threonine residue are are active than those where it is attached to a serine residue. This was a fascinating result and we are now trying to reproduce this in mice with a variety of different nanoparticles. A new mouse study using a statistica *TRUNCATED*