The effect of surface demineralization of cortical bone allograft on the properties of recombinant adeno-associated virus coatings

Abstract

Freeze-dried recombinant adeno-associated virus (rAAV) coated structural allografts have emerged as an approach to engender necrotic cortical bone with host factors that will persist for weeks following surgery to facilitate revascularization, osteointegration, and remodeling. However, one major limitation is the nonporous cortical surface that prohibits uniform distribution of the rAAV coating prior to freeze-drying. To overcome this we have developed a demineralization method to increase surface absorbance while retaining the structural integrity of the allograft. Demineralized bone wafers (DBW) made from human femoral allograft rings demonstrated a significant 21.1% (73.6+/-3.9% versus 52.5+/-2.6%; p<0.001) increase in percent surface area coating versus mineralized controls. Co-incubation of rAAV-luciferase (rAAV-Luc) coated DBW with a monolayer of C3H10T1/2 cells in culture led to peak luciferase levels that were not significantly different from soluble rAAV-Luc controls (p>0.05), although the peaks occurred at 60h and 12h, respectively. To assess the transduction efficiency of rAAV-Luc coated DBW in vivo, we first performed a dose response with allografts containing 10(7), 10(9) or 10(10) particles that were surgically implanted into the quadriceps of mice, and assayed by in vivo bioluminescence imaging (BLI) on days 1, 3, 5, 7, 10, 14, and 21. The results demonstrated a dose response in which the DBW coated with 10(10) rAAV-Luc particles achieved peak gene expression levels on day 3, which persisted until day 21, and was significantly greater than the 10(7) dose throughout this time period (p<0.01). A direct comparison of mineralized versus DBW coated with 10(10) rAAV-Luc particles failed to demonstrate any significant differences in transduction kinetics or efficiency in vivo. Thus, surface demineralization of human cortical bone allograft increases its absorbance for uniform rAAV coating, without affecting vector transduction efficiency.

Figure 1. Surface demineralization of cortical allograft bone increases its porosity and absorptive distribution of rAAV-Luc

Scanning electron microscopy (SEM) images of non-demineralized bone (NDB) (left) and demineralized cortical bone wafers (DBW) (right) prior to rAAV coating are shown at 3000 X magnification to demonstrate the marked increase in porosity of the demineralized bone (A). 30μl of a 1% sorbitol solution of PBS with red food coloring containing 10⁸ particles of rAAV-Luc was pipetted onto a 10mm NDB (left) or DBW (right), and photographed before (B) and after (C) freeze-drying. Note that the solution pools in the center of the mineralized bone leading to poor coating after freeze-drying, while the solution is more evenly distributed on the demineralized bone leading to a significant increase in surface coating after freeze-drying (52.5 ± 2.6 % vs. 73.6 ± 3.9 %; p<0.001). Further evidence of improved rAAV-coating uniformity is provided by 9 X SEM images of the freeze-dried wafers (D) in which the border of the rAAV coating is clearly visible on the NDB (arrows), while the even distribution on the DBW renders this boarder more transparent.

Figure 2. Surface demineralization does not alter rAAV release kinetics and transduction efficiency in vitro

The in vitro transduction efficiency of the rAAV-Luc vector was determined by infecting a monolayer C3H10T1/2 (10⁵ cells per well) at the indicated multiplicity of infection (MOI) and assaying the cells for luciferase activity (relative light units, RLU) at 12hr or 24hr. All transduction conditions lead to a significant increase in RLU versus the uninfected control (asterisk indicates p<0.05) (A). The effects of the demineralized bone wafer (DBW) on rAAV transduction and freeze-dried rAAV-Luc release kinetics were determined in C3H10T1/2 transduction experiments in vitro (B). Cells were incubated: alone (white bars), with soluble rAAV-Luc (MOI = 100) only (light gray bars), with uncoated DBW and soluble rAAV-Luc (MOI = 100) (dark gray bars), or with rAAV-Luc (MOI = 100) coated-DBW (black bars). RLU assays were performed at the indicated time following treatment. The peak transduction time is indicated by the significant increase over the 12hr control for each group (*; p<0.05). No significant effects of the DBW were observed on soluble rAAV-Luc transduction (p>0.05). However, the peak of soluble rAAV-Luc transduction occurred at 12hr with or without DBW, while the peak of freeze-dried rAAV transduction occurred at 60hr, but the RLU values of these peaks in the DBW containing cultures were not significantly different (p>0.05).

Figure 3. In vivo transduction of rAAV-Luc coated demineralized bone wafers

Demineralized bone wafers (DBW) were coated with 10⁷, 10⁹, or 10¹⁰ particles of rAAV-Luc as described in Figure 1, surgically implanted into the right quadriceps of mice (n=6), and temporal bioluminescent imaging (BLI) was performed at days 1, 3, 5, 7, 10, 14, 21. A longitudinal series of BLI photographs of the experimental leg from a representative mouse in each is shown at the indicated time (A). The mean BLI ± SD for each dose of rAAV-Luc determined at each time point is presented on a log-scale vertical axis (B). + indicates a significant increase in peak intensity on Day 10 vs. Day 1 for both the 10⁹ and 10¹⁰ groups (p<0.001). # indicates significant differences in BLI for the 10¹⁰ group versus the 10⁷ group (p<0.01) at the indicated time point. * indicates significant differences in BLI for the 10¹⁰ group versus the 10⁹ group at the indicated time point (p<0.05).

Figure 4. Demineralization of cortical allograft bone does not affect the rAAV-coating release kinetics

Non-demineralized bone (NDB) and demineralized bone wafers (DBW) were coated with 10¹⁰ rAAV-Luc particles and surgically implanted into mice (n=5) as described in Figure 4. Longitudinal BLI images of a representative mouse from each are shown (A). The BLI signal on days 1, 3, 5, 7, 10, 14, 21, and 28 of each mouse is presented (B). No significant differences in the BLI signal could be observed between the two groups at any time point (p>0.05). Note that the peak BLI for both groups occurs on day 7 with no significant changes between day 7 and 14 for either group (p>0.05), but drops to levels below Day 1 by Day 28 for both groups.