B.C. Toolan, P. Bolt, A.N. Clerk, H.H. Luu, Q. Kang, J.L. Kummer, Z.-L. Deng, K. Olson, F. Primus, A.G. Montag, T.-C. He, and R.C. Haydon reply:
We appreciate the insightful comments from Dr. Rickert regarding our recent article on the use of adenovirus-mediated transfer of BMP-14 (GDF-5) for Achilles tendon lacerations in a rat model, as well as the opportunity to respond to his concerns.
First, our manuscript was originally submitted for publication in February 2006, only a couple of months after the publication of the article by Rickert et al. in Connective Tissue Research1. We were unaware of that study during the preparation of our manuscript; however, we referenced many other articles from Dr. Rickert's laboratory at the University of Heidelberg, recognizing his important contributions to this area of research. As such, it was not our intention to ignore his study, and we welcome the chance to discuss the two articles.
Second, there are three important differences between the model of Achilles tendon injury that we used and that used by Rickert et al. Most notably, we repaired the Achilles tendon with a single stitch to appose the ends of the lacerated tendon, thereby permitting tensile loading across the repair site. We did not lacerate the adjacent plantaris tendon, to ensure that functional walking would not be altered after the procedure. In Dr. Rickert's model, the Achilles and plantaris tendons were lacerated and not repaired. Lastly, we used a substantially lower titer of virus for transduction of the ends of the lacerated tendon (108 plaque-forming units [pfu] as opposed to 3 × 1010 pfu).
These differences may account for the differences in results between the two studies. Dr. Rickert asserts that previous reports from the University of Manchester2,3 have suggested that "cartilage and bone formation is essential during rat Achilles tendon healing." This hypothesis is debatable and worthy of further study; however, it should be emphasized that cartilage and bone formation occurred mainly in models in which no repair was performed. Forslund and Aspenberg4 found that heterotopic ossification was infrequent and extremely limited in tendons that were exposed to tensile loads, whereas robust heterotopic bone formation was found in nearly all unloaded tendons. This was one reason why we chose to repair the tendon after infection with the adenovirus. Although it would be of interest to look at later time-points, the main point of this study was to examine tensile properties at early time-points that would be of clinical interest, not to examine the impact of mechanical loading on lineage commitment in cells that mediate tendon repair. Nonetheless, on the basis of our previous analysis of the fourteen types of BMPs for their osteogenic activity, we believe that BMP-14 is among the least osteogenic BMPs both in vitro and in vivo5,6.
Dr. Rickert also questions how tendon healing and mechanical strength were assessed in our study. As clinicians, we are most concerned about rerupture of the Achilles tendon early after repair. Our study was more narrowly focused in this regard compared with his study, and we did not include measurements of the cross-sectional area of the repair tissue and Young's modulus. Assessment of the cross-sectional area can be highly subjective, since values can vary depending on which part of the repair tissue is selected for analysis. For that reason, it was not included. Instead, we used tensile load to failure and gap formation as, perhaps, more clinically relevant methods to determine the feasibility of this approach in patients. Dr. Rickert mentions that measurements of gap formation are "imprecise and unconventional"; however, in a study by Gelberman et al.7, gaps in the repair tendons were found to correlate with "the ultimate force, repair-site rigidity, and repair-site strain" in a canine model of tendon healing. For this reason, we felt that gap formation was a more clinically meaningful measurement to determine the risk of rerupture after repair.
Lastly, Dr. Rickert questioned the justification behind our conclusion that there was "no adverse immunological response to the adenoviral vector." In a previous study, we reported on the dose-response relationship between different titers of adenoviral vectors and inflammation in a rabbit flexor-tendon model8. Using standard hematoxylin and eosin stains, we could identify histologic evidence of acute inflammation at titers over 109. These parameters were used to determine the doses which were used in the present study, and the histologic data presented in the study suggest that this titer does not cause excessive inflammation in the rat Achilles tendon model. There is no mention made of inflammation in Dr. Rickert's study, and it would be interesting to compare results, as the doses in his study were considerably higher. He also mentions that standard hematoxylin and eosin stains may not be sufficient to rule out the presence of inflammatory cells in healing tendons, suggesting that we consider approaches that specifically identify CD4+ or CD8+ T-cells. Although this is a valid point, histologic evaluation with hematoxylin and eosin stains only has been used by others9 to assess inflammation in tendons after adenoviral gene transfer, without the need for additional tests. Given that inflammation is an important part of the healing process in tendons, some inflammatory cells would be expected at the repair site, regardless of whether an adenovirus was present or not. Hematoxylin and eosin stains, however, represent a reasonable method to determine the presence of excessive inflammation that might impact clinical outcomes.
We believe that the differences cited above between the two studies highlight some of the important factors that affect tendon healing and the utility of adenoviral delivery of BMP-14 for Achilles tendon repair. We believe the findings of our two studies, instead of being contradictory, should be viewed as complementary because both contribute meaningful data to the debate over the role of gene therapy in the future treatment of Achilles tendon ruptures.
These letters originally appeared, in slightly different form, on . They are still available on the web site in conjunction with the article to which they refer.