Introduction Stem cells possess great potential in treating disk degeneration. MSC is the most clinically relevant cell source but its functional outcomes are still far from satisfaction. One of the primary reasons for the suboptimal performance of MSCs is that the majority of injected cells leaked out from the injection path. Existing attempts addressing this issue focus on further increasing the viscosity of carrier and use of BioGlue. Nevertheless, from others and our own experiences, the high intradiscal pressure in the IVD would extrude the injected cells out immediately even with solid carriers while the blocking effect of glues is only transient as large osteophyte structure was still seen in 6 months time. As a result, we hypothesize that effectively blocking the injection portal with an injectable biomaterial-based annulus plug will solve the cell leakage problem. Our group has previously developed a photochemical cross-linking technology to improve the physicochemical properties including mechanical strength, chemical stability, and swelling property of collagen-based scaffolds. Our ultimate goal is to further enhance the functional outcomes of MSC-based IVD regeneration by effectively solving the cell leakage problem. In this study, we aim to (1) fabricate a photochemically cross-linked collagen annulus plug; (2) evaluate the performance of the plug ex vivo by conducting mechanical push-out test with different loading protocols and cell leakage test; (3) design a clinically applicable delivery device for the plug and (4) evaluate the performance of the plug in vivo by using a rabbit IVD degeneration model. Materials and Methods Acid soluble rat tail type I collagen at 4 mg/mL was poured into cylindrical-shaped container and placed in an alkaline vapor chamber containing ammonium hydroxide to initiate the gelation process for 1 hour. The reconstituted collagen gel was equilibrated with rose Bengal solution (0.0005%) overnight before irradiating the gel Argon laser at 514 nm for 100 seconds at laser power of 0.2 W. The cross-linked collagen gel was then shaped into thin needles with length around 5 to 7 mm and diameter either less than 0.5 mm or around 1 mm by controlled dehydration. The plug was inserted into rabbit lumbar disk either through direct puncture or through a 21G-syringe needle. Histoacryl glue was used to seal the insertion site. The lumbar discs were then subjected to axial loading of 40–50N at 1 Hz for 7 days with either a protocol of continuous cyclic compression or a protocol simulating the physiological daily activities with cyclic and static loading patterns to evaluate the mechanical durability of the plug. In a separate experiment, quantum dots-labeled MSCs were injected into the disk and injection portal was immediately blocked with the annulus plug with and without Histoacryl glue. The discs were subjected to the physiological loading protocol and the culture medium was collected after 7 days to evaluate the cell leakage by visual inspection under a fluorescent microscopy. Results Gently pressing the plug-inserted disk immediately after insertion with fingers resulted in pushing out of the plug. This is because the dried plug is usually slightly smaller than the inner diameter of the needle and takes time to rehydrate to press-fit the injection portal. Sealing the injection site with Histoacryl glue prevents the ejection of plug immediately after insertion and allows rehydration of the plug before subjecting to physiological loading. Preliminary results on the push-out tests demonstrated that all collagen plugs, two using continuous cyclic compression and four using physiological loading patterns, were kept in their original position, that is, were not pushed out after 7 days of loading at a force comparable to the body weight of rabbits. Moreover, the culture medium collected from the quantum-dot labeled MSCs injection experiment showed no cell leakage after 7 days compression. Conclusion BioGlue is necessary to provide an initial sealing of the injection portal. The photochemically cross-linked collagen annulus plug are mechanically durable under ex vivo physiologically relevant loading but in vivo evaluation is necessary. The plug may be able to effectively block cell leakage but more ex vivo tests and in vivo evaluation are necessary. Finally, a custom-made special needle is being designed, fabricated and, evaluated for easy delivery of the plug in vivo. I confirm having declared any potential conflict of interest for all authors listed on this abstract No Disclosure of Interest None declared