Chitosan fibrous scaffolds for cartilage tissue engineering

The biocompatibility of chitosan and its similarity with glycosaminoglycans make it attractive for cartilage engineering despite its limited cell adhesion properties. Structural and chemical characteristics of chitosan scaffolds may be improved for cartilage engineering application. We planned to evaluate chitosan meshes produced by a novel technique and the effect of chitosan structure on mesenchymal stem cells (MSCs) chondrogenesis. Another objective was to improve cell adhesion and chondrogenesis on chitosan by modifying the chemical composition of the scaffold (reacetylation, collagen II, or hyaluronic acid (HA) coating). A replica molding technique was developed to produce chitosan meshes of different fiber-width. A polyglycolic acid (PGA) mesh served as a reference. Constructs were analyzed at two and 21 days after seeding chondrocytes with confocal microscopy, scanning electron microscopy, histology, and quantitative analysis (weights, DNA, glycosaminoglycans, collagen II). Chondrocytes maintained their phenotypic appearance and a high viability but attached preferentially to PGA. Matrix production per chondrocyte was superior on chitosan. Chitosan meshes and sponges were analyzed after seeding and culture of MSCs under chondrogenic condition for 21 days. The cellularity was similar between groups but matrix production was greater on meshes. Chitosan and reacetylated-chitosan scaffolds were coated with collagen II or HA. Scaffolds were characterized prior to seeding MSCs. Chitosan meshes were then coated with collagen at two densities. PGA served as a reference. Constructs were evaluated after seeding or culture of MSCs for 21 days in chondrogenic medium. MSCs adhered less to reacetylated-chitosan despite collagen coating. HA did not affect cell adhesion. The cell attachment on chitosan correlated with collagen density. The cell number and matrix production were improved after culture in collagen coated meshes. The differences between PGA and chitosan are likely to result from the chemical composition. Chondrogenesis is superior on chitosan meshes compared to sponges. Collagen II coating is an efficient way to overcome poor cell adhesion on chitosan. These findings encourage the use of chitosan meshes coated with collagen II and confirm the importance of biomimetic scaffolds for tissue engineering. The decreased cell adhesion on reacetylated chitosan and the poor mechanical stability of PGA limit their use for tissue engineering.

Evaluation of Adoptive T Cell Therapy and Oncolytic Virotherapy for Treatment of Brain Tumors

Adoptive immunotherapy and oncolytic virotherapy are two promising strategies for treating primary and metastatic malignant brain tumors. We demonstrate the ability of adoptively transferred tumor-specific T cells to rapidly mediate the clearance of established brain tumors in several mouse models. Similar to the clinical situation, tumor recurrences are frequent and result from immune editing of tumors. T cells can eliminate antigen-expressing tumor cells but are not effective against antigen loss variant (ALV) cancer cells that multiply and repopulate a tumor. We show that the level of tumor antigen present affects the success of adoptive T cell therapy. When high levels of antigen are present, tumor stromal cells such as microglia and macrophages present tumor peptide on their surface. As a result, T cells directly eliminate cancer cells and cross-presenting stromal cells and indirectly eliminate ALV cells. We were able to show the first direct evidence of tumor antigen cross-presentation by CD11b+ stromal cells in the brain using soluble, high-affinity T cell receptor monomers. Strategies that target brain tumor stroma or increase antigen shedding from tumor cells leading to increased crosspresentation by stromal cells may improve the clinical success of T cell adoptive therapies. We evaluated one potential strategy to complement adoptive T cell therapy by characterizing the oncolytic effects of myxoma virus (MYXV) in a syngeneic mouse brain tumor model of metastatic melanoma. MYXV is a rabbit poxvirus with strict species tropism for European rabbits. MYXV can also infect mouse and human cancer cell lines due to signaling defects in innate antiviral mechanisms and hyperphosphorylation of Akt. MYXV kills B16.SIY melanoma cells in vitro, and intratumoral injection of virus leads to robust, selective and transient infection of the tumor. We observed that virus treatment recruits innate immune cells iii to the tumor, induces TNFα and IFNβ production in the brain, and results in limited oncolytic effects in vivo. To overcome this, we evaluated the safety and efficacy of co-administering 2C T cells, MYXV, and neutralizing antibodies against IFNβ. Mice that received the triple combination therapy survived significantly longer with no apparent side effects, but eventually relapsed. Based on these findings, methods to enhance viral replication in the tumor and limit immune clearance of the virus will be pursued. We conclude that myxoma virus should be further explored as a vector for transient delivery of therapeutic genes to a tumor to enhance T cell responses.

Cell-cell communication in three dimensional microenvironments

Cellular behavior is dependent on a variety of extracellular cues required for normal tissue function, wound healing, and activation of the immune system. Removed from their in vivo microenvironment and cultured in vitro, cells lose many environmental cues and that may result in abberant behavior, making it difficult to study cellular processes. In order to mimic native tissue environments, optical tweezer and microfluidic technologies were used to place cells within defined areas of the culture environment. To provide three dimensional supports found in natural tissues, hydrogel scaffolds of poly (ethylene glycol) diacrylate and the basement membrane matrix Matrigel were used. Optical tweezer technology allowed precision placement and formation of homotypic and heterotypic arrays of human U937, HEK 293, and porcine mesenchymal stem cells. Alternatively, two microfluidic devices were designed to pattern Matrigel scaffolds. The first microfluidic device utilized laminar flow to spatially pattern multiple cell types within the device. Gradients of soluble molecules were then be formed and manipulated across the Matrigel scaffolds. Patterning Matrigel using laminar flow techniques require microfluidic expertise and do not produce consistent patterning conditions, limiting their use difficult in most cell culture laboratories. Thus, a buried Matrigel polydimethylsiloxane (PDMS) device was developed for spatial patterning of biological scaffolds. Matrigel is injected into micron sized channels of PDMS fabricated by soft lithography and allowed to thermally cure. Following curing, a second PDMS device was placed on top of the buried Matrigel channels to support media flow. In order to validate these systems, a cell-cell communication model system was developed utilizing LPS and TNFα signaling with fluorescent reporter systems to monitor communication in real time. We demonstrated the utility of microfluidic devices to support the cell-cell communication model system by co culturing three cell types within Matrigel scaffolds and monitoring signaling activity via fluorescent reporters.