Control of crosslinking for tailoring collagen-based scaffolds stability and mechanics

Abstract

We provide evidence to show that the standard reactant concentrations used in tissue engineering to cross-link collagen-based scaffolds are up to 100 times higher than required for mechanical integrity in service, and stability against degradation in an aqueous environment. We demonstrate this with a detailed and systematic study by comparing scaffolds made from (a) collagen from two different suppliers, (b) gelatin (a partially denatured collagen) and (c) 50% collagen-50% gelatin mixtures. The materials were processed, using lyophilisation, to produce homogeneous, highly porous scaffolds with isotropic architectures and pore diameters ranging from 130 to 260 μm. Scaffolds were cross-linked using a carbodiimide treatment, to establish the effect of the variations in crosslinking conditions (down to very low concentrations) on the morphology, swelling, degradation and mechanical properties of the scaffolds. Carbodiimide concentration of 11.5mg/ml was defined as the standard (100%) and was progressively diluted down to 0.1%. It was found that 10-fold reduction in the carbodiimide content led to the significant increase (almost 4-fold) in the amount of free amine groups (primarily on collagen lysine residues) without compromising mechanics and stability in water of all resultant scaffolds. The importance of this finding is that, by reducing cross-linking, the corresponding cell-reactive carboxylate anions (collagen glutamate or aspartate residues) that are essential for integrin-mediated binding remain intact. Indeed, a 10-fold reduction in carbodiimide crosslinking resulted in near native-like cell attachment to collagen scaffolds. We have demonstrated that controlling the degree of cross-linking, and hence retaining native scaffold chemistry, offers a major step forward in the biological performance of collagen- and gelatin-based tissue engineering scaffolds.

Statement of significance: This work developed collagen and gelatine-based scaffolds with structural, material and biological properties suitable for use in myocardial tissue regeneration. The novelty and significance of this research consist in elucidating the effect of the composition, origin of collagen and crosslinking concentration on the scaffold physical and cell-binding characteristics. We demonstrate that the standard carbodiimide concentrations used to crosslink collagenous scaffolds are up to 100 times higher than required for mechanical integrity in service, and stability against dissolution. The importance of this finding is that, by reducing crosslinking, the corresponding cell-reactive carboxylate anions (essential for integrin-mediated binding) remain intact and the native scaffold chemistry is retained. This offers a major step forward in the biological performance of tissue engineered scaffolds.

Keywords: Collagen; Crosslinking; Gelatin; Scaffolds; Tissue engineering.

Graphical abstract

Fig. 1

Dependence of the apparent viscosity (Pa s) of Col 1% (w/v) suspensions on shear rate (s⁻¹) for Sigma and Devro Type-I collagen at 20 °C.

Fig. 2

Visualisation of internal structure of collagen suspensions after attachment of GFOGER-FL. Confocal images of different Col samples: (A) monomeric (negative control); (B) Devro; (C) Sigma and (D) Ethicon (positive control). The scale bars are 100 μm.

Fig. 3

SEM images of scaffolds of different compositions and XL status: (A) non XL scaffolds (different longitudinal cross-sections), (B) influence of cross-linking and Col type (transversal cross-sections).

Fig. 4

Influence of EDC concentration on the free amine group content on lysine residues of scaffolds of different compositions: (a) Sigma Col-based and Gel samples; (b) Devro Col-based and Gel samples; (c) Col(S) vs Col(D) scaffolds. Results are mean values of three parallel measurements. Error bars represent the standard error.

Fig. 5

Influence of EDC concentration and composition on dissolution behaviour of Col(D)-based and Gel scaffolds. Results are mean values of four parallel measurements. Error bars represent the standard error.

Fig. 6

Dissolution profiles of Col(S)-based scaffolds cross-linked with different EDC concentration. Results are mean values of four parallel measurements. Error bars represent the standard error.

Fig. 7

Mass loss: (a) Col(S vs D) and (b) mixed Col(S vs D)-Gel and Gel scaffolds after 14d in water. Results are mean values of four parallel measurements. Error bars represent the standard error.

Fig. 8

Influence of EDC concentration on the compressive modulus (E^∗) of the scaffolds of different compositions (a) and typical profiles of compressive stress–strain curves observed for hydrated Col(S) scaffolds after repeated compressions of the same specimen over the entire strain range (e: 0–0.8) (b). Results are mean values of five parallel measurements. Error bars represent the standard error.

Fig. 9

Integrin mediated platelet (a) and HT1080 cell (b) attachment to films prepared from sigma collagen (Col(S) – black bars), 50:50 sigma collagen:gelatin (Col(S):Gel – dark grey bars) and gelatin (Gel – light grey bars) treated with increasing concentrations of EDC where 100% equates to 2:5:1 M ratio of EDC:NHS:collagen COOH groups. Results are mean values of quadruplicate measurements. Error bars represent standard deviation from the mean.