Development of hydrogel-based tissue-engineered salivary glands

Abstract

The salivary gland comprises lumenized acini that produce saliva and interconnected ducts that modify and carry the saliva to the oral cavity. Salivary glands can be irreversibly damaged by high-energy radiation commonly used to treat head and neck cancers. As a result, patients develop xerostomia (or dry mouth syndrome), and their quality of life is severely compromised. Currently, there are no curative solutions, and all the palliative treatments are ineffective. Salivary gland tissue engineering offers an alternative long-term solution to xerostomia. The goal of my dissertation work is to develop implantable tissue-engineered salivary gland mimetics for the restoration of salivary gland function. Towards this goal, human primary adult stem/progenitor cells (hS/PCs) were isolated from patients undergoing parotidectomy. Although these cells exhibit some stem/progenitor markers, the expression of KIT, a receptor tyrosine kinase, is relatively low. During development, KIT positive cells are responsible for the establishment of the entire organ. I hypothesize that reconstituting hS/PCs in a synthetic extracellular matrix (ECM) that can be remodeled by the resident cells will give rise to secretory constructs containing multipotent, KIT+ cells with self-renewal capacities. To this end, I developed hyaluronic acid (HA)-based hydrogels with varying degrees of protease susceptibility (0DEG: nondegradable, 50DEG: 50% degradable, 100DEG: 100% degradable) using thiolated HA (HA-SH), maleimide functionalized cell adhesive peptide (RGD-MI), and bis-maleimide functionalized proteolytically degradable (GIW-bisMI) and/or nondegradable (GIQ-bisMI) peptide crosslinkers. I found that singly dispersed hS/PCs formed multicellular structures in all three types of hydrogels. As the matrix degradability increased, cell proliferation increased and the structures became larger, less compact, and more irregularly shaped. Using quantitative polymerase chain reaction (qPCR) methods, I discovered that the expression of the key stem/progenitor markers, including KIT, KRT5, and ETV4/5, was significantly increased in the degradable matrices as compared to the nondegradable counterparts. Through Western blotting, I found that there was a 4-fold increase in KIT expression by cells grown in degradable gels as compared to those residing in nondegradable gels. Importantly, the degradable matrices also promoted a pro-acinar phenotype as evidenced by the upregulation of acinar markers AMY, AQP5, and SLC12A2. In summary, cell-mediated matrix remodeling is necessary for the development of regenerative pro-acinar progenitor cells from hS/PCs. In the native gland, directional secretion of saliva into the central lumen is possible because cells within the acinus share a common lumen and exhibit defined apicobasal polarity. To achieve correct apicobasal polarity in the engineered gland, I cultured hS/PCs in HA gels established above in the presence of an inhibitor of Rho kinases (ROCKi). I hypothesize that inhibition of ROCK-mediated signaling in hS/PCs will reduce cellular tension to allow for proper organization of individual cells into the multicellular spheroids to achieve apicobasal polarization. I discovered that ROCKi stimulated the development of lumenized structures in 0DEG and 50DEG cultures. While structures developed in 0DEG gels were compact and spherical, those generated in 50DEG gels were lobular and more closely resembled the acinus seen in vivo. Immunocytochemistry analyses showed that GM130, a polarity marker, was localized between the nucleus and the apical membrane in both 0DEG and 50DEG cultures. Moreover, integrin β1, a marker for cell-ECM interaction, and laminin α1 and collagen IV, basement membrane proteins, were all basally localized. To study if cells can secrete α-amylase into the lumen, 50DEG cultures were stimulated with isoproterenol for 1 h before being fixed for immunocytochemistry analyses. Confocal imaging revealed the presence of α-amylase granules in the lumen, confirming that the engineered constructs are functional. Importantly, spheroids established in 0DEG and 50DEG gels with ROCKi had a higher expression of key acinar markers (AQP5, NKCC1) at the mRNA and protein levels. To assess how ROCKi affects the physical state of hS/CPs and contributes to the development of polarized structures, I characterized the stiffness of 3D cultured hS/PCs by particle tracking microrheology using mitochondria as the tracer particles. Interestingly, both matrix degradation and ROCK inhibition affect cell stiffness. hS/PCs in 100DEG constructs were stiffer as compared to those cultured in 0DEG gels. On the other hand, hS/PCs in constructs treated with ROCKi were less stiff as compared to the untreated controls. I conclude that the development of correctly polarized acini-like structures in HA gels requires an optimal combination of cell-mediated matrix remodeling and cellular stiffness. Salivary gland morphogenesis requires input and signaling from the integrated vasculature. To deliver an innovative engineering solution to salivary gland regeneration, I investigated how endothelial cells affect the phenotype and function of salivary gland epithelial cells by culturing hS/PCs with human umbilical vein endothelial cells (HUVECs) in a tissue-mimetic configuration. First, I screened the respective monocultures under various media conditions and found that EGM2 media, designed for endothelial cell culture, is most appropriate for the coculture studies as it permitted rapid cell growth and maintained the correct phenotype for both cell types. Next, hS/PCs were allowed to self-assemble into spheroids in agarose microwells. After growing in suspension for 7 days, the spheroids were encapsulated in cell-adhesive, protease degradable HA gels established via tetrazine ligation with trans-cyclooctene and norbornene. Subsequently, HUVECs were plated on top of the hS/PC-laden construct and the coculture was maintained in EGM2 media for 14 days. hS/PC aggregates grew into lobular structures with or without the cocultured endothelial cells. Under the culture conditions, hS/PCs expressed significantly higher levels of stem/progenitor biomarkers (KRT5, KRT14, ETV5), acinar biomarkers (AMY, AQP3, SLC12A2) and ductal biomarkers (KRT7, TFCP2L1), as compared to the epithelial monoculture. Finally, HA gels loaded with endothelial cues/ growth factors were implanted in partially resected parotid glands in rats. Compared to the resection only controls, animals receiving the gel implants exhibited reduced tissue fibrosis and more organized acinar structures. Collectively, endothelial cues promote pro-acinar differentiation and reduce tissue fibrosis. This is significant because radiation therapy is known to cause tissue fibrosis, which in turn, leads to acinar cell damage, senescence, and apoptosis. In summary, this dissertation work demonstrates a bottom-up approach to the development of salivary gland mimetics using patient-derived stem/progenitor cells. Future work will focus on in-depth in vivo assessment of the engineered gland in an animal model.