With nearly one million hospitalizations due to bone fractures each year in the United States, we clearly have a need for effective bone regeneration and fracture healing therapies. Bone grafts are currently the most common treatment for serious fractures, at approximately 700,000 per year. Autogenous grafts of bone transplanted from a donor site to fracture site within the same patient has the best clinical outcome. However, such donor bone tissue is limited, and significant pain and morbidity can result at the donor site. Thus arises a demand for a safe and effective material to facilitate bone healing. We propose a biodegradable, synthetic material that can support cell migration, proliferation, and differentiation.

We have synthesized porous poly(ester urethane)urea (PEUUR) scaffolds from polyester polyols and aliphatic polyisocyanates that are biocompatible, biodegradable, and resorbable. The reactive liquid molding method of synthesis allows them to be injectable, and thus minimally invasive during implantation. The isocyanate is mixed with a hardener, consisting of a poly(e-caprolactone-co-glycolide-co-lactide) polyol, water, triethylenediamine catalyst, sulfated castor oil stabilizer, and calcium stearate pore opener. Water reacts with the isocyanate to form CO2, which acts as a blowing agent to create the pores. The scaffold characteristics can be easily tuned by adjusting the ratios and properties of the constituents.

Here we have investigated the effects of various polyol molecular weights and isocyanates, as well as addition of poly(ethylene glycol), on the chemical, thermal, mechanical, and degradation properties and in vivo compatibility of the PEUUR scaffolds. We prepared 900-MW and 1800-MW polyester triols (P7C3G900, P7C3G1800, P6C3G1L900) with the appropriate ratios of caprolactone/glycolide/lactide monomers (70/30/0 & 60/30/10). 600-MW poly(ethylene glycol) (PEG) was added to the foam formulation as 20 to 60% of the polyol component. Lastly, we tested the following aliphatic isocyanates: Desmadour N3300A (hexamethylene diisocyanate trimer, HDIt), lysine diisocyanate (LDI), and lysine triisocyanate (LTI).

Gel permeation chromatography (GPC) and nuclear magnetic resonance (NMR) showed that the polyol reactions proceeded to completion. FT-IR analysis shows characteristic vibration peaks for the ester (1765, 1303, & 1114 cm-1), urethane (3422 & 1765 cm-1), and urea (1469 cm-1) groups. There is no evident NCO peak at 2285-2250 cm-1, which suggests that most of the isocyanate free NCO has reacted upon foaming. The isocyanate seemed to have a greater effect on material and mechanical properties than did the polyol. For example, the porosities of the HDIt, LDI, and LTI foams were 92.5 ± 0.6%, 78.6 ± 3.5%, and 93.1 ± 1.5%, respectively. Scanning electron micrographs (SEM) showed pore sizes to range in diameter from 200-600 um, but pores in the LDI foams tended to be larger and more irregular. Addition of PEG did not appear to affect the pore structure.

Thermal gravimetric analysis (TGA) shows that the foams begin to lose mass at 200 șC, and only 10% of the material remains at 500 șC. Thermal profiles of the foams from dynamic scanning calorimetry (DSC) show single glass transitions (Tg) of -30 șC to 6 șC, which differs from that of the pure polyols (-50 șC to -40 șC), suggesting that the foams are phase-mixed between the hard (isocyanate) segments and the soft (polyol) segments. Furthermore, increasing the polyol MW or PEG600 content causes the foam Tg to decrease. Temperature sweeps by dynamic mechanical analysis (DMA) show the equivalent Tg trends, but these mechanical Tg's were consistently higher than the thermal Tg's found form DSC. DMA stress relaxation tests showed higher compliance of foams made with 1800-MW polyol than 900-MW polyol, such that the relaxation modulus dropped much more steeply over the 20-min test. The presence of PEG in the foams caused them to have faster stress relaxation. However, PEG did not seem to affect the magnitude of the storage or relaxation modulus. LDI lowered the mechanical strength and resilience of the foams, such that DMA analysis could not be performed on the LDI foams. For this reason, they were not used for the in vivo studies.

8x2mm HDIt and LTI foam discs were implanted both subcutaneously and into 8-mm full-thickness excisional wounds in adult Sprague-Dawley rats, which were stented with stainless steel washers. The foams were held in place by semi-occlusive dressing. Wounds were harvested at days 5, 14, and 21, and processed for Gomori's trichrome histological evaluation. The HDIt and LTI foams exhibited complete re-epithelization by day 21, although the LTI foams showed greater degradation. The foams supported mononuclear cellular infiltration in vivo, showing mature granulation tissue and dense collagen fibers. A giant cell response was associated with, and limited to, material remnants.

Ongoing experiments include implantation into critical size bone defects in rabbit femurs, in order to assess the equivalent efficacy in bone. These foams demonstrate promise for future use as bone healing scaffolds, especially with the demonstrated ability to optimize their material properties using various polyols, isocyanates, and additives while still maintaining a desirable cellular response. A further clinical goal is the incorporation and delivery of biologically active molecules and growth factors to enhance bone healing.