The Relationship Between Tissue Oxygenation and Cell Differentiation During Fracture Healing

Introduction Oxygen is required for proper tissue healing. Oxygen is an important bacteriocidal agent, is required for the activity of a variety of enzymes, and is required for formation of collagen fibrils (reviewed in: [1]). In addition, oxygen has been proposed to act as a morphogen by regulating differentiation of chondrocytes and osteoblasts in vitro (reviewed in: [2]). Chondrocytes are thought to differentiate in areas of relative hypoxia while osteoblasts require oxygenated conditions for differentiation. However, a causal relationship between oxygen level and cell differentiation has not been tested directly during fracture healing. Here we used a method called Electron Paramagnetic Resonance (EPR) spectroscopy to assess the local oxygen tension in the tissue adjacent to the fracture site (reviewed in: [3]). In this technique an oxygen-sensitive paramagnetic crystal, lithium phthalocyanine (LiPc), is placed in the tissue of interest. Oxygen tension at the surface of the crystal can then be determined non-invasively and without interfering with healing or disrupting tissue architecture at the fracture site [4]. In this study, we measured oxygen levels immediately after injury, during stages of cell differentiation and during production of bone and cartilage matrices. We then performed histology to assess the extent to which tissue oxygen levels were related to differentiation of osteoblasts and chondrocytes. Methods In vivo transverse tibial fractures were induced in male, 10 week old, wild type mice by three-point bending (n=26). Immediately after fracture the oxygen sensing LiPc crystals were placed into the facture site with a 21 gauge needle. Using a custom built L-band EPR spectrometer, pO2 was measure in each mouse on the day of fracture, and at days 1, 3, 5, 7, and 10 after injury. After analysis was complete, the fractured limb was dissected, decalcified, dehydrated, embedded in paraffin and sectioned (10μm). Hall-Bryant Quadruple staining was performed to visualize tissues in sections. All protocols were approved by the IACUC (Institutional Animal Care and Use Committee) at Dartmouth Medical School, Hanover, NH. The EPR data were evaluated by fitting raw data into 2 components using MATLAB and a customized multi-fitting program developed at Dartmouth. This software enables us to fix the relationship between linewidth and intensity, allowing pO2 to vary. Thus, the pO2 is estimated directly with fewer degrees of freedom for more robust estimation. This approach is valuable for in vivo data where much lower signal-to-noise ratios are common and where it may be difficult to fit multiple linewidths and intensities directly. The data were then analyzed via ANOVA to determine significance of the data between crystal location and the different time points of pO2 measurements. Results Our data demonstrate that oxygen levels dropped to hypoxic levels immediately after fracture. In some cases oxygen levels rose rapidly during the first 24 hours after injury (n=19/26), but in other animals oxygen levels remained low (n=7/26). By day 10 oxygen levels in all animals had returned to normal levels. Next, we performed histology to determine where the LiPc crystal was located in each animal. The crystals were found in one of three locations in the injured limbs. In some animals, the crystal was located in the muscle adjacent to the tibia (n=12). In other animals the crystal was located at the edge of the fracture callus (n=5). In a third group of animals the crystals were located in the fracture callus within the cartilage matrix (n=9). In 4 of these animals the crystal was embedded in cartilage and in newly formed endochondral bone. Therefore, during the early stages of healing these crystals were in regions that would later form cartilage, and later the crystals were located in the replacement bone. The oxygen levels at the measured times points for each group is displayed on Table 1. We did not detect a statistical difference between the location of the LiPc crystals and the oxygen levels at any of the time points. Discussion There is a large amount of work in the literature that indicates chondrogenesis occurs in regions of relative hypoxia regions (reviewed in: [2]). However, we failed to detect a relationship between tissue oxygen levels and cell differentiation during fracture healing. During fracture healing, the oxygen level dropped immediately after injury either due to disruption of the blood supply or reduced blood flow due to shock. Within 24 hours, oxygen levels had increased dramatically in the muscle and what would later become the outer part of the callus. In regions that would form cartilage and bone, the oxygen levels were lower, but not statistically different from that observed in other regions of the limb. The large variance that we observed in EPR measurements was not related to crystal location or size. Rather, oxygen levels in each animal seem to result from unique responses to the injury. We are currently trying to resolve the cause of the variance in the measurements, in order to determine the extent to which oxygen levels are related to other healing parameters such as callus size, the rate of bone repair, or the extent of the initial injury. Nonetheless, our results indicate that the role of oxygen as a morphogen during the early stages of cell differentiation may need to be re-evaluated. The hypoxia that is apparent in the cartilage matrix may result from the avascularity of the cartilage itself, rather than being a cause of cartilage formation.