Heart valve replacement is a life-saving procedure conducted worldwide approximately 300 000 times per year. Used mechanical devices and bioprosthetic valves, however, exhibit severe drawbacks. They either request for lifelong anticoagulation therapy, or display an early failure due to calcification and stenosis. Furthermore, their use in pediatric surgery requests reoperations since these prosthesis lack the ability to grow. An optimal heart valve substitute that exhibits unlimited durability, perfect hemodynamics (no need for anticoagulation) and the capacity to grow has been identified by implantations of decellularized allogenic heart valve matrices. Such grafts are revitalized in vivo by cells of the recipient, which form a functional endothelium and a living interstitium. However, the remaining obstacle is the limited availability of human donor heart valves. Until now, xenogenic substitutes (porcine origin) failed due to acute or chronic rejection when implanted into humans as well as in the sheep model. Thus, we aim to generate a porcine derived decellularized matrix, free of immunogenic epitopes, for endogenous tissue regeneration and tissue engineering.

In a first step, we developed analytical tests to quantify residual α(1,3)Gal epitopes, known to be responsible for hyper-acute rejection reactions, and other xenoantigens detected by natural human anti-pig antibodies, which are present on matrices treated by our standard decellularization protocol. Since extracellular matrix is insoluble with employed the inhibitory ELISA technique using the commercial available anti-αGal antibody M86 and natural anti-pig antibodies isolated from the serum of healthy volunteers. In brief, porcine kidneys were perfused with human plasma. Unbound antibodies were washed away and bound xenoantibodies were eluted using 3.5 M NaSCN. After dialysis against PBS antibodies were used to perform the inhibitory ELISA as follows. Native as well as decellularized heart valve tissue was crushed in liquid nitrogen and incubated with a defined amount of antibodies. Unbound antibodies were separated from matrix bound antibodies by centrifugation and quantified. M86 ELISA was performed using αGal-BSA (Dextra) as solid antigen and HRP conjugated goat anti-mouse IgM antibody as detection antibody. In the case of natural human anti-pig antibodies, the total amount of human IgG was quantified using a human IgG ELISA (Mabtech).

In our experiments comparing native and decellularized porcine heart valves, we found that standard decellularization removes already 70% of the αGal epitopes. In heart valves from α1,3-galactosyltransferase KO (GalT-KO) pigs the binding of M86 antibody is comparable to the extent as binding of M86 is found in human pulmonary artery tissue. In respect to general xenoantigens reduced binding of xenoantibodies after decellularization was detected. However, human as well as GalT-KO pig derived tissue bind less xenoantibodies as tissue derived from normal pigs. Comparing only GalT-KO and human tissue, the human tissue binds less xenoantibodies indicating the existence of non-Gal antigens on GalT-KO tissue.

In conclusion, we found strong evidence that xenoantigens including αGal and non-Gal xenoantigens are not only present on cells and cell debris but also on decellularized heart valve matrixes that are in fact insoluble extra cellular matrix proteins. Thus, decellularization causes a reduction of the antigenicity, however, the degree of this removal might be dependent on the decellularization protocol. Modification of the current decellularization protocol as well as enzymatic treatment against glycocalyx structures, which are believed to be hard candidates as xenoantigens may result in an additional reduction.