Surgeons treating fractures with many small osteochondral fragments have often expressed the clinical need for an adhesive to join such fragments, as an adjunct to standard implants. If an adhesive would maintain alignment of the articular surfaces and subsequently heal it could result in improved clinical outcomes. However, there are no bone adhesives available for clinical indications and few pre-clinical models to assess safety and efficacy of adhesive biomaterial candidates. A bone adhesive candidate based on water, α-TCP and an amino acid phosphoserine was evaluated in-vivo in a novel murine bone core model (preliminary results presented EORS 2019) in which excised bone cores were glued back in place and harvested @ 0, 3, 7, 14, 28 and 42days. Adhesive pull-out strength was demonstrated 0–28 days, with a dip at 14 days increasing to 11.3N maximum. Histology 0–42 days showed the adhesive progressively remodelling to bone in both cancellous and cortical compartments with no signs of either undesirable inflammation or peripheral ectopic bone formation. These favourable results suggested translation to a large animal model.

A porcine dental extraction socket model was subsequently developed where dental implants were affixed only with the adhesive. Biomechanical data was collected @ 1, 14, 28 and 56 days, and histology at 1,14,28 and 56 days. Adhesive strength assessed by implant pull-out force increased out to 28 days and maintained out to 56 days (282N maximum) with failure only occurring at the adhesive bone interface. Histology confirmed the adhesive's biocompatibility and osteoconductive behavior. Additionally, remodelling was demonstrated at the adhesive-bone interface with resorption by osteoclast-like cells and followed by new bone apposition and substitution by bone. Whilst the in-vivo dental implant data is encouraging, a large animal preclinical model is needed (under development) to confirm the adhesive is capable of healing, for example, loaded osteochondral bone fragments.

Acknowledgements: The murine study was supported, in part, by the Swedish Foundation for Strategic Research (#RMA15-0110).

An ex vivo biomechanical test model for evaluating a novel bone adhesive has been developed. However, at day 1 in the in vivo pilot, high blood flow forced the study to halt until the solution presented here was developed.

The profuse bleeding after bone core removal affected the bond strength and was reflected in the lower mean peak value 1.53N. After considering several options, we were successful in sealing the source of blood flow by pressing adhesive into place after bone core removal. After the initial adhesive had cured additional adhesive was used to secure the bone core in place. The animals were sacrificed after 24 h and a tensile test was undertaken on the bone core to failure.

The ex vivo study produced mean peak tensile loads of 7.63N SD 2.39N (n=8, 4 rats 8 femurs). Whilst the mean peak tensile loads in the day 1 in vivo pilot were significantly lower 1.53N SD1.57 (n=8, 6 rats 8 femurs − 4 used for other tests). The subsequent layered adhesive bone cores showed a mean peak tensile force of 6.79N SD =3.13 (n=8, 4 rats 8 femurs). 7/8 failed at the bone to glue interface. This is the first successful demonstration of bonding bone in vivo for this class of adhesives.

The development of a double adhesive method of fixing a bone core in the distal femur enabled mean peak tensile forces to be achieved in vivo at 24 hours that were comparable with the ex vivo results previously demonstrated. This method supports application in further animal series and over longer time scales. Biomaterials researchers that intend to use gel or paste like preparations in distal femur defects in the rat should be aware of the risks of biomaterial displacement by local blood flow.

There are clinical situations in fracture repair, e.g. osteochondral fragments, where current implant hardware is insufficient. The proposition of an adhesive enabling fixation and healing has been considered but no successful candidate has emerged thus far. The many preclinical and few clinical attempts include fibrin glue, mussel adhesive and even “Kryptonite” (US bone void filler). The most promising recent attempts are based on phosphorylating amino acids, part of a common cellular adhesion mechanism linking mussels, caddis fly larvae, and mammals. Rapid high bond strength development in the wetted fatty environment of fractured bone, that is sustained during biological healing, is challenging to prove both safety and efficacy. Additionally, there are no “predicate” preclinical animal and human models which led the authors to develop novel evaluations for an adhesive candidate “OsStic^tm” based on calcium salts and amino acids. Adhesive formulations were evaluated in both soft (6/12 weeks) and hard tissue (3,7,10,14 & 42 days) safety studies in murine models. The feasibility of a novel adhesiveness test, initially proven in murine cadaver femoral bone, is being assessed in-vivo (3,7,10,14 & 42 days) in bilateral implantations with a standard tissue glue as the control. In parallel an ex-vivo human bone model using freshly harvested human donor bone is under development to underwrite the eventual clinical application of such an adhesive. This is part of a risk mitigation project bridging between laboratory biomaterial characterisation and a commercial biomaterial development where safety and effectiveness have to meet today´s new medical device requirements.