Osteoarthritis is a joint disease associated with the irreversible breakdown of articular cartilage in joints, causing pain, impaired mobility, and reduced quality of life in over 27 million Americans and many more worldwide. The tolls by osteoarthritis (OA) on the workforce and healthcare system represent significant economic burdens. An attractive strategy for treating OA is cartilage tissue engineering (CTE). CTE strategies have been promising at producing cellscaffold constructs at small sizes (3-5 mm in largest dimension), but OA often does not present symptoms until lesions reach 25 mm in diameter.

Using bovine chondrocytes seeded in agarose, our lab has produced small CTE constructs with native cartilage levels of compressive stiffness and proteoglycan content. As construct dimensions are increased, however, the resulting tissue suffers from extreme heterogeneity of deposited matrix due to nutrient transport limitations. The ability to successfully scale up constructs to clinically relevant sizes is a major goal in CTE research. Another major and largely unresolved obstacle is the translation of successes from animal cell models to CTE systems with human cells, which is ultimately necessary for clinical treatment of OA. In this dissertation, experiments are placed forth which seek to address the nutrient limitations in large cartilage constructs and to help bridge the gap from animal cells to human cells for CTE.

The growth of CTE constructs is limited by the poor availability of nutrients at construct centers due to consumption by cells at the construct periphery. The first series of studies in this dissertation sought to identify nutrients in culture media that are consumed by cells and are critical for matrix production, and to characterize their transport behavior. Among several candidate nutrients, glucose proved to be the most indispensable;​ little to no growth transpired in constructs when glucose fell below a critical threshold concentration. A subsequent study provided a system-specific glucose consumption rate. These parameters were informative for computational models of construct growth, which helped predict transport and growth phenomena in constructs and suggest improved culture techniques for later experiments.

The cultivation of tissue constructs of increasing size presents logistical challenges, as the constructs’ requirements for nutrients, growth factors, and even sizes of culture vessels increase. The addition of nutrient channels to constructs to improve nutrient transport and tissue growth is a promising strategy, but more sophisticated casting and culture techniques are required for constructs with channels, particularly as construct size is increased. We first designed casting and culture devices for cylindrical 10 mm × 2.3 mm (diameter × height) constructs with 1 mm diameter nutrient channels. With information gleaned from computational models predicting glucose availability in constructs, we refined our culture system and demonstrated beneficial effects of nutrient channels on construct mechanical properties and extracellular matrix contents. This was the most successful instance to date of the use of nutrient channels in CTE, and is highly promising for channels’ ability to mitigate transport limitations in constructs.

We next sought to optimize key parameters for culturing channeled constructs. The addition of channels is an optimization problem:​ greater numbers of closer-packed channels increase nutrient availability within the construct but simultaneously detract from the construct’s initial volume and cell population. Furthermore, we suspected that uneven swelling of 10 mm diameter constructs was a side effect of transient treatment with 10 ng/mL TGF-β, a highly effective and commonly-employed technique for elevating construct functional properties. By increasing channel densities in 10 mm diameter constructs, we identified a channel spacing that yielded optimal construct functional properties. In constructs with this channel spacing, reducing the TGF-β dosage by tenfold resulted in similar or elevated properties by constructs. These experiments supplied us with optimal parameters for further scaling up our constructs to clinically-relevant sizes, a practice that can be adapted for any CTE culture system for large constructs.

The ability to treat severe OA by entirely resurfacing diseased joints with CTE would be highly desirable, yet this ability remains elusive, as efforts to grow constructs of such size have thus far been stymied by nutrient transport limitations. We scaled up our culture system for 10 mm diameter constructs, employing previously optimized culture conditions and channel spacing, and cultured articular surface-sized (40 mm diameter, 2.3 mm thick) constructs. These constructs were 100× the size of our small constructs, yet they still attained similar functional properties, reaching native cartilage levels of compressive stiffness and proteoglycan content. These are the largest CTE constructs to ever achieve such favorable properties. These results demonstrate that with nutrient channels, CTE constructs have the potential to replace entire joint surfaces that have been compromised by OA.

Finally, we began to explore the feasibility of translating techniques from our bovine and canine model systems into human cells. We harvested adult human chondrocytes from expired osteochondral allografts and cast them in small (3 mm diameter) constructs, culturing the constructs under various conditions that have been previously successful for animal constructs. We observed similarities between human versus bovine and canine constructs, most notably that high initial cell seeding density led to marked increases in functional properties, in some cases approaching mechanical and biochemical properties of native human cartilage. Human constructs also exhibited poor GAG retention and long-term growth relative to animal constructs. By establishing successful techniques for human constructs in addition to identifying new challenges, we provided an in-depth characterization of human chondrocytes in agarose that is promising overall for eventual clinical translation.

The body of work presented in this dissertation followed a methodical approach to scaling up CTE constructs to the sizes of entire joint surfaces, through experimentation with nutrient channels in constructs and with the support of predictive computational models. The principle behind nutrient channels is fundamental and therefore can be applied to CTE systems using other scaffold and cell types. By incrementally increasing the scale of bovine chondrocyteladen constructs and by performing initial studies with small human CTE constructs, we have laid down groundwork for future studies seeking to grow articular surface-sized human engineered cartilage.