Bone grafting is of significant clinical importance in orthopaedic, craniomaxillofacial, and dental reconstruction procedures. Bone tissue engineering strategies are currently being developed as alternative mechanisms to address the clinical demand for bioactive and biomechanical graft material. To date, these efforts have been largely restricted by inadequate supply of committed osteoprogenitor cells and loss of osteoblastic phenotype expression following in vitro culture and expansion.

The objective of this thesis research was to address the cell sourcing limitations of tissue-engineered bone grafts by combining the fundamental principles of osteoblast biology and genetic and tissue engineering. Specifically, the osteoblast-specific transcriptional activator Runx2/Cbfa1 was overexpressed in osteogenic target cells using retroviral gene delivery. It was hypothesized that sustained overexpression of Runx2 in marrow-derived primary stromal cells would augment osteoblastic differentiation, including osteoblast-specific gene and protein expression and matrix mineralization, both in vitro and in vivo. Following initial characterization of this Runx2 overexpression system in model non-osteoblastic and osteoblast-like cell lines, the global objective and hypothesis were tested through in vitro evaluation of Runx2-expressing marrow-derived stromal cells in monolayer culture and on three-dimensional (3-D) polymeric, biodegradable scaffolds, including both polylactide-co-glycolide (PLGA) foams and poly(ε-caprolactone) (PCL) fused deposition-modeled scaffolds. Additionally, the in vivo performance of Runx2-expressing stromal cells was evaluated in two systems, subcutaneous and craniotomy defect syngeneic animal models.

Runx2 overexpression in marrow-derived stromal cells enhanced expression of multiple osteoblastic genes, including osteocalcin and collagen type I, as well as alkaline phosphatase biochemical activity. More importantly, matrix mineralization was upregulated nearly two-fold following 2 and 3 weeks of in vitro culture. To evaluate their performance in bone tissue engineering applications, Runx2-expressing cells were seeded onto PLGA and PCL polymeric scaffolds. These two 3-D matrices possessed substantially different physical characteristics including pore size, strut architecture, and physical integrity. Supporting flexibility of this genetic engineering system, Runx2- transduced cells demonstrated significantly higher levels of 3-D mineralization when cultured on either scaffold compared to control cell-seeded scaffolds.

To evaluate in vivo performance, Runx2-expressing cells were seeded onto PLGA scaffolds and subcutaneously implanted following various in vitro culture durations. Minimal quantifiable mineral was present on Runx2 or control constructs implanted following only 1 or 7 days of pre-culture, time points representing more immature scaffolds. Notably, Runx2-transduced cell-seeded constructs pre-cultured to a mineralizing state in vitro (21 days) prior to implantation demonstrated a 10-fold increase in total mineral deposition and a 5-fold increase in the average daily mineral deposition rate upon subsequent in vivo implantation compared to similarly pre-cultured control cells, demonstrating a synergistic effect of Runx2 overexpression and in vitro construct development.

Finally, Runx2 overexpression and in vitro construct maturation was examined in a critical size craniotomy defect model in syngeneic rats. In this study, Runx2 overexpressing and control cells were seeded on PCL fused deposition-modeled scaffolds possessing a macroporous, fully interconnected architecture. Defects treated with cellfree PCL scaffolds demonstrated a significant healing response, suggesting that in this particular bone defect model, the open architecture and interconnected structural network of the fused deposition-modeled scaffolds provided a suitable osteoconductive scaffold capable of supporting significant new bone formation by infiltrating host cells. Further supporting these observations, immature cell-seeded scaffolds implanted 1 day after cell seeding demonstrated healing comparable to cell-free scaffolds independent of treatment. In contrast to observations from the subcutaneous animal model, in vitro construct development for 21 days demonstrated an inhibitory effect on subsequent in vivo bone formation in defects receiving control bone marrow stromal cell-seeded constructs. As observed in the subcutaneous model, however, defects receiving Runx2-modified cell-seeded PCL constructs pre-cultured for 21 days demonstrated significantly more new bone formation than similarly pre-cultured control constructs. The amount of new bone formation in Runx2-modified cell-seeded, 21 day pre-cultured constructs was comparable to cell-free scaffold and 1 day pre-culture construct groups, suggesting that Runx2 overexpression in pre-cultured bone marrow stromal cells enhanced subsequent in vivo implant performance capable of overcoming the inhibitory effects on new bone formation observed in response to extended in vitro pre-culture of control cells.

Collectively, these studies provided a thorough characterization regimen to observe the in vitro and in vivo performance of Runx2-expressing marrow-derived stromal cells and evaluated their potential as a candidate cell source for bone tissue engineering applications. Furthermore, this series of analyses provided a novel combination of tissue and genetic engineering techniques toward the development of a Runx2-modified stromal cell/polymeric scaffold composite tissue-engineered bone graft substitute.