Modification of biomaterials for the immobilization and release of chemokines

The presentation of biologically active molecules by modified biomacromolecules is of scientific interest for the understanding and treatment of inflammatory processes. The focus of this work lies on the presentation of chemokines to analyze chemokine-induced cell migration and on the characterization of photodynamically modified collagen biomaterials. In regenerative medicine, controlled drug release by modified biomaterials represents an important technique, as side effects due to systemic or excessive local drug applications can be avoided. Collagen is a commonly used biomaterial in regenerative medecine, as it provides a substrate for the adhesion of cells and is degraded to non-toxic products. Modifications of the biomaterial collagen by rose bengal and green light cross-linking (RGX) have already been described for different clinical applications, such as for orthopedics or ophthalmology. In this work RGX was used to bond stacks of different collagen materials to laminates with tailored properties for a biomedical application, e.g. a controlled drug release to avoid surgical site infections (SSIs). SSIs are the most common complications in orthopedic surgery and can be prevented by antimicrobial prophylaxis. The aim of this work was to characterize multilayered collagen laminates in terms of their swelling behavior and antibiotic release. For the fabrication of multilayered collagen laminates, two types of collagen materials were selected. The results indicated that homogeneous laminates composed of sponge-like collagen showed a lower swelling degree than a single RGX-treated, sponge-like collagen sheet. This was explained by the additional layer of rose bengal at the interface between the piled sheets. In contrast, homogeneous laminates composed of a thin collagen membrane did not show any change in their swelling degree, independent of the number of collagen layers. This was explained by the compact structure of the material. Heterogeneous collagen laminates composed of both materials reached swelling degree values in-between. For homogeneous sponge-like laminates and heterogeneous laminates the experimental swelling degrees were significantly smaller then the theoretical ones. These findings were explained by the different number of available swelling interfaces in laminates compared to individual sheets. To test the release of the model antibiotic vancomycin, an additively manufactured sample holder was developed, which allows to quantify the release of vancomycin into opposite directions. The sample holder elongated the time until half-maximal release was reached. This can be attributed to the decreased size of release areas compared to a sample in solution. Release experiments with heterogeneous, bi-layer collagen laminates under physiological conditions showed that the release of vancomycin preferentially takes place at the surface of a thin collagen film instead of sponge-like collagen, independent of the laminate’s orientation or loading. Furthermore, loading of vancomycin into the sponge-like collagen layer of heterogeneous, bi-layer collagen laminates led to an elongated time of half-maximal release. Studies with triple-layer collagen laminates led to similar results, as the time for half-maximal release increased with the number of sponge-like collagen layers. Furthermore, vancomycin was again preferentially released at the surface of the thin collagen film layer. These findings can be explained by shorter diffusion pathways and a negligible effect of re-swelling for the thin collagen film in contrast to the sponge-like collagen. In detail, the higher porosity results in a higher swelling degree. Compression by insertion of the sample into the sample holder therefore has a stronger effect on the sponge-like material and leads to a larger uptake of fluid during re-swelling that is opposed to the diffusion of vancomycin out of the collagen sheet. Consequently, vancomycin is released more slowly. In contrast to bi-layer laminates, the orientation of triple-layer heterogeneous laminates determined to which extent vancomycin was released by the thin collagen film layer. This can be explained by an increased swelling of sponge-like collagen at the bottom side of the sample holder. Similarly, an increased swelling and release at the bottom side of a triple-layer, sponge-like, homogeneous collagen laminate was observed. This can be attributed to the laminate's thickness. Overall, these findings reveal mechanisms that must be considered for the composition of collagen laminates. Since the pH can change during wound healing or infections, release studies under alkaline (pH 8.5) or acidic (pH 5.5) conditions reported for these processes were carried out. The pH did not have any effect on the total amount of released vancomycin from single sheets of RGX-modified, sponge-like collagen or heterogeneous, bi-layer laminates. The latter showed again a preferential release at the side of the thin collagen film layer, as observed for physiological pH (pH 7.4). In addition, the release from RGX-modified single sheets and bi-layer laminates was retarded at pH 5.5, which was explained by an increased swelling degree at this pH. At pH 8.5, the release was also delayed but no changes in the swelling degree were observed between pH 8.5 and pH 7.4. These results might be explained by electrostatic interactions, as negatively charged vancomycin might interact with positively charged areas of collagen at pH 8.5. The findings of this work can be used in regenerative medicine, as sponge-like collagen would be more suitable for a delayed release of active substances that support tissue regeneration. In summary, the results showed that the biomaterial collagen can be modularly assembled to laminates that allow a controlled and directed release of antibiotics. These findings are of biomedical interest, as sponge-like collagen may be more suitable for the delayed release of active substances at later stages of wound healing than for the fast release of antibiotics at infected sites. Further studies might include the release of proteins from collagen laminates or the cross-linking of proteins within collagen matrices with RGX for a delayed release. The second part of this work dealt with the presentation of biologically active molecules, such as chemokines on surfaces. These small secreted signaling proteins induce and steer the migration of cells in homeostatic and inflammatory processes. Chemokines interact with corresponding receptors on the surfaces of cells, induce different signaling cascades and lead to cellular responses, such as cell migration. The latter can either occur along soluble chemokine concentration gradients, which is specified as chemotaxis, or along surface-bound gradients, a process termed haptotaxis. As chemokines are also involved in different diseases, understanding the migration of cells upon chemokine stimulation contributes to understand and treat inflammatory processes. Interleukin-8 (CXCL8) was used as a model inflammatory chemokine in this work. In previous works, a dopamine-heparin coating had been developed that allowed a reversible immobilization of CXCL8 in microfluidic channels. In this work, reversibly immobilized CXCL8 gradients in microfluidic channels should be further characterized with respect to their suitability for different migration experiments. The results indicated that the dopamine-heparin coating led to a homogeneous distribution of CXCL8. Furthermore, the gradient stability at 37 °C was not different from previous experiments at room temperature. Cell migration experiments confirmed a directed and reproducible migration of THP-1 cells that express receptors for CXCL8 along the reversibly immobilized CXCL8 gradient towards high concentrations of CXCL8. Control experiments did not show any directed cell migration only in the presence of buffer or homogeneously distributed chemokine. Since chemokine gradients naturally occur as a mixture of soluble and immobilized gradients, the influence of overlaid, soluble chemokine on the gradient and cell migration was examined. An overlay of the gradient with soluble chemokine that has a concentration in the same order of magnitude as initial CXCL8 used for gradient formation increased the gradient's steepness within the first 20 h of incubation and led to a delay of cell migration. In contrast, very low concentrations of soluble CXCL8 overlaid over the immobilized gradient did not influence directed cell migration. In summary, the method represents a simple setup to analyze to which extent non-covalently immobilized chemokine gradients are influenced by soluble chemokine and enables a reproducible analysis of cell migration along these gradients. Since the coating method with dopamine and heparin only enables a non-covalent immobilization of CXCL8 that changed over the time of the migration experiment, it was attempted to covalently immobilize CXCL8 in microfluidic channels. The development of a suitable method was motivated by the fact that RGX worked well in the former project for cross-linking collagen layers and that collagen is a commonly used substrate in migration studies. However, no CXCL8 immobilization could be detected under the tested conditions. Interaction studies could not confirm an interaction of CXCL8 with collagen, while weak, non-specific interactions of RB with CXCL8 and BSA were observed. However, the interactions were apparently not sufficient to activate the protein for binding to collagen. As the immobilization of native proteins by a photodynamic process is attractive, future studies might involve the use of higher light intensities to develop an immobilization method