In phase 1, a suitable subcutaneous implantation technique was evaluated. For this purpose, six male Lewis rats were divided into two identical groups (A and B). In animals from group A, two teflon isolation chambers containing fibrin gel matrices were implanted subcutaneously in conventional implantation technique with the opening of the chamber facing the dorsal skin of the rat and without any fixation of the chambers [19, 20] (Fig. 1A and B). In group B, the chambers were implanted in an upside-down fashion on the back of the rats with the opening of the chamber facing the dorsal muscle and with fixation of the chambers (Fig. 1C and D). In both groups, the matrices consisted of 500 μl fibrin gel with identical fibrinogen and thrombin concentrations (10 mg/ml fibrinogen, 2 I.U./ml thrombin). The explantation interval was 10 days after the initial operation for all groups (including 6 matrices per group and time-point).

In phase 2, the angiogenetic effects of VEGF and bFGF immobilized in a fibrin-based drug delivery system were quantitatively assessed in the subcutaneous rat model (Fig. 2). Four teflon isolation chambers containing fibrin gel matrices were implanted subcutaneously in an upside-down fashion on the back of 30 Lewis rats. The matrices consisted of 500 μl fibrin gel with two different fibrinogen concentrations (10 mg/ml or 40 mg/ml fibrinogen, 2 I.U./ml thrombin), loaded with VEGF165 and bFGF in five different concentrations (0 to 250 ng/ml each). The explantation intervals were 3, 7 and 14 days after the initial operation for all groups (including 4 matrices per group and time-point).

The cylindrical chamber (inner diameter 10 mm, height 6 mm) was made of heat-resistant medical grade teflon (P. Greil, Department of Materials Science, Glass and Ceramics, University of Erlangen). In this study, the matrix was prepared from a fibrin sealant clinically approved in the United States and some European countries (Tisseel®, Baxter, Vienna, Austria). A final fibrinogen concentration of 10 mg/ml or 40 mg/ml, a thrombin concentration of 2 I.U./ml and an aprotinin concentration of 1500 K.I.E./ml to delay fibrinolysis were used. In phase 2 of this study, the matrix was loaded with VEGF165 and bFGF in five different concentrations (0, 25, 100 and 250 ng/ml bFGF and VEGF each and VEGF alone at 100 ng/ml) before implantation. Recombinant human VEGF165 (R&D Systems, Minneapolis, MN, USA) and recombinant human bFGF (Roche Diagnostics, Indianapolis, IN, USA) were used.

Syngenic male Lewis rats (Charles River Laboratories, Sulzfeld, Germany) were used in this study. German regulations for the care and use of laboratory animals were observed at all times. All experiments were approved by the animal care committee of the University of Erlangen and the Government of Mittelfranken, Germany. The animals were housed in the veterinary care facility of the University of Erlangen Medical Center and submitted to a12 hrs dark/light cycle with free access to standard chow (Altromin, Hamburg, Germany) and water.

All operations were performed by the same microsurgeon using an operative microscope (Karl Zeiss, Jena, Germany) under general anaesthesia with Isoflurane (Baxter, Unterschleiflheim, Germany).

For the conventional implantation technique, two small subcutaneous pockets were created on the back of the rats by two longitudinal skin incisions. Chambers were placed without any fixation in the pockets containing 500 μl fibrin gel with the opening of the chamber facing the dorsal skin. Haemostasis was assured and the wound was closed using Vicryl 5-0 (Ethicon, Norderstedt, Germany).

For the upside-down implanted chambers, one large subcutaneous pocket was created by a mid-line skin incision on the back of the rats. In the first phase, two, and in the second phase, four chambers containing 500 μl fibrin gel were implanted in the subcutaneous pocket with the opening of the chamber facing the dorsal muscle of the rats. In phase two, the fibrinogen component was loaded with VEGF and bFGF as described in detail in Figure 2. The chambers were fixed using Prolene 3-0 (Ethicon, Norderstedt, Germany) sutures. Haemostasis was assured and the wound was closed using Vicryl 5-0 (Ethicon, Norderstedt, Germany). All animals received 0.2 ml Benzylpenicillin-Benzathin (Tardomycel Comp®, Bayer, Leverkusen, Germany) and buprenorphin (0.3 mg/kg rat weight, Temgesic®, Essex Chemie AG, Luzern, Switzerland) postoperatively.

Specimens were explanted in toto with the surrounding tissue and fixed in 3.5% formalin for histological processing. Afterwards, the constructs were dehydrated in graded ethanol and embedded in paraffin. Three μm cross-sections were obtained from two standardized planes (1 mm proximal and 1 mm distal of the central plane) using a Leica microtome (Leica Microsystems, Bensheim, Germany). Cross-sections were stained using haematoxylin and eosin according to standard protocols in phase 1 and 2. In phase 2, additional immuno-histochemical staining was performed using the lectin Bandeiraea Simplicifolia agglutinin (BS-1), alpha-smooth muscle actin and ED1. The images were evaluated by two independent and blinded observers.

All images of haematoxylin and eosin-stained cross-sections were generated with a Leica Microscope and Digital Camera under 25× magnification. In phase one, images were evaluated morphologically. In phase two, the matrix height and volume and the height and volume of the fibrovascular tissue adjacent to the matrix were calculated for each group and each time-point (Fig. 3).

The fibrovascular tissue adjacent to the matrix of all lectin-stained cross-sections was divided into 3 equal sectors, 2 at the periphery and 1 at the central part of the section (Fig. 4) and all images of each sector were acquired with a Leica Microscope and Digital Camera under 200× magnification. The images were rendered bimodal (standardized threshold) (WinQ, Leica Microsystems, Bensheim, Germany). The number of blood vessels, standard deviation of blood vessel size and average size of blood vessels was calculated for each group and each time. Results are given as mean ± standard deviation. Statistical analysis was performed using GraphPad Prism software (GraphPad Software, San Diego, CA, USA). Normal distribution was checked positive using the Kolmogorov-Smirnov test for all groups. Two-tailed unpaired Student’s t-test was applied for statistical analysis. The critical level of statistical significance chosen was P < 0.05.

All 36 animals tolerated the surgical procedure well. There were no major postoperative complications such as infection, haematoma or wound dehiscence. No extrusion of the implants occurred over the observation period.

After opening the back of the rats, there were newly formed capsules of fibrous tissue around the chambers containing the matrices. After opening the capsules and removal of the chambers, all matrices implanted using the upside-down implantation technique were adherent to the subjacent tissue whereas matrices implanted using the conventional implantation technique were detached of the surrounding tissue.

Histologically, matrices of group A, which were implanted in a conventional manner, showed an inhomogeneous clot structure. Also, there was only a small area of contact between the clot and the adjacent fibrovascular tissue. In contrast, matrices implanted using the upside-down implantation technique displayed high adhesion of the fibrin clot to the subjacent tissue and a highly homogeneous structure (Fig. 5).

Different concentrations of fibrinogen did not influence tissue formation and vascularization patterns in this experimental setting. Therefore, only results using the diluted fibrinogen are presented in detail.

Macroscopically, all chambers in all groups were surrounded by a capsular layer of connective tissue. Upon day 3, all fibrin clots in all groups were still present without any fibrinolysis or degradation. At 1 week, volume of the fibrin clots was considerably smaller in the 100 and 250 ng/ml VEGF and bFGF groups compared to lower growth factor concentrations (Fig. 6). Almost complete resorption of the matrices had occurred upon day 14 after implantation in all groups.

Histologically, during the 14 days of implantation the fibrin gel was not invaded by blood vessels in any group. The fibrovascular tissue subjacent to the fibrin gel was composed of inflammatory cells, fibroblasts and blood vessels in all groups. Using ED1 immuno histochemical staining, some macrophages invading the fibrovascular tissue were visible independent of added growth factors. No significant foreign body reaction with presence of multinucleated giant cells was detected within the fibrin clots and the fibrovascular tissue. Vascularization of this tissue could be confirmed by immunohistochemical staining using the lectin Bandeiraea Simplicifolia agglutinin and alpha-smooth muscle actin (Fig. 7). Except for blood vessels there was no significant staining by ASMA in the fibrovascular tissue.

In phase two, morphometric analysis of the constructs demonstrated a significant decrease of construct height following application of 100 and 250 ng/ml VEGF and bFGF with 1.75 ± 0.71 mm and 1.76 ± 0.9 mm in comparison to 25 ng/ml VEGF and bFGF with 3.65 ± 0.89 mm and 100 ng/ml VEGF with 5.65 ± 1.27 mm at 1 week (Fig. 8). In all groups, the matrix height did not signifi cantly differ 3 days after implantation. At day 14, all fibrin clots were almost completely degraded in all groups.

The use of growth factors with a concentration of 25 ng/ml VEGF and bFGF and 100 ng/ml VEGF significantly increased height of fibrovascular tissue at the matrix–host interface in comparison to growth factor-free controls at 1 week (0.87 ± 0.36 mm and 1.19 ± 0.58 mm versus 0.1 ± 0.12 mm) and at 2 weeks (1.11 ± 0.59 mm and 1.37 ± 0.66 mm versus 0.29 ± 0.12 mm). Higher concentrations did not further induce tissue formation (Fig. 9).

The total number of blood vessels counted in the fibrovascular tissue adjacent to the matrices of lectin stained cross-sections was significantly increased when using of 25 ng/ml VEGF and bFGF and 100 ng/ml VEGF in comparison to growth factor-free fibrin clots at 1 week (185 ± 45 vessels per field of view [FOV] and 168 ± 38 vessels per FOV versus 110 ± 21 vessels per FOV) and at 2 weeks (179 ± 64 vessels per FOV and 168 ± 42 vessels per FOV versus 99 ± 28 vessels per FOV). There was no further increase of blood vessel numbers when higher concentrations of growth factors were used (Fig. 10).

Similar results could be obtained regarding average fraction of blood vessels and total area of blood vessels (data not shown).

There was no difference regarding average blood vessel size and standard deviation of blood vessel size between the different growth factor concentrations (data not shown).