Mechanical behaviour of a membrane made of human umbilical cord for dental bone regenerative medicine

Keywords:

• Mechanical properties
• hyperelasticity
• human umbilical cord
• guided bone regeneration
• Bio-Gide®

1. Introduction

Guided bone regeneration (GBR) is one of the most attractive techniques for restoring oral bone defects following tooth extraction or periodontal disease. To ensure the stability of a dental implant, the GBR technique relies on the use of an occlusive membrane, which is positioned as a barrier over the bone defect, providing space maintenance required for bone growth and preventing the ingrowth of fibrous tissue into the bone defect (Elgali et al. 2017). This membrane has to meet criteria guidelines in terms of mechanical properties, as the membrane has to allow a good handling for dental surgeons, without collapsing into the bone defect. Healthy perinatal tissues are promising biomaterials because they are inexpensive, and universally available (Ferguson and Dodson 2009). Among these tissues, the human umbilical cord, mainly composed of collagen fibres and glycosaminoglycan, especially hyaluronic acid and chondroitin sulphate, is expected to offer outstanding opportunities for tissue engineering by serving as a suitable biocompatible membrane for GBR.

In this study, a novel membrane derived from the human umbilical cord (UC-membrane) was successfully developed following tissue stripping and freeze-drying processes. Ice crystal formation following tissue freezing is known to induce pore formation within the tissue, hampering their mechanical properties. Herein, the mechanical behaviour of UC-membrane was determined and compared to the Bio-Gide® membrane (gold standard membrane)

2. Methods

2.1. Samples

Human umbilical cord harvesting was approved ethically and methodologically by our local Research Institution and was conducted with informed patients (written consent) in accordance with the usual ethical legal regulations (Article R 1243-57, in accordance with our authorization and registration number DC-2014-2262 given by the French institutions). Fresh human umbilical cords, obtained after full-term births, were washed several times with distilled water and dissected for vascular structures removal. Umbilical cord tissue was cut into pieces and preserved in dried condition at −20 °C until freeze-drying process. Hematoxylin-Eosin-Saffron (HES) staining was performed on paraffin embedded UC-membrane and Bio-Gide® membrane (Geistlich Pharma).

2.1. Mechanical test

The mechanical quality of the samples was tested through quasi static tensile tests up to failure. The loading sequence was divided into two parts: (a) a dry test under elastic limits to avoid any damages followed by (b) a hydrated one allowing a full behaviour characterisation of the materials ([NaCl] = 9 g.L⁻¹) at 37 °C. Five minutes were given for the sample to accommodate prior to be tested through cyclic strain loads and eventually up to failure. All loadings were performed at a 0.01 mm.s⁻¹ velocity to remain in the quasi static framework. A Universal Testing Machine Zwicky0.5 equipped with a 10 N load cell was used to measure samples’ response. Specimens were cut to the same dimensions to get hydrated samples of: 13.79 ± 0.59 × 4.65 ± 0.57 × 0.95 ± 0.19 mm³ for the UC-membrane and 13.01 ± 0.01 × 4.39 ± 0.05 × 0.5 ± 0.01 mm³ for the Bio-Gide®. The engineering stress definition: $\sigma =F/S_0$ has been used to characterise the mechanical behaviour where $F$ is the current force and $S_0$ is the initial cross-section. On the other hand, the engineering strain is defined as: $\epsilon =\Delta l/l_0$ where $\Delta l$ stands for the variation of length and $l_0$ for the initial length. An incompressible three terms Ogden’s hyperelastic law (Dorfmann and Ogden 2004) has been also used to fit the results: $$\sigma =\mathrm{ }\sum _{i=1}^3{\mu }_i\left({\lambda }^{{\alpha }_i-1}-{\lambda }^{-\frac{{\alpha }_i}{2}-1}\right)$$ where ${\mu }_i$ and ${\alpha }_i$ are the parameters related to the shear modulus $\mu$ such as: $\mu =\frac{1}{2}\sum _i{\mu }_i{\alpha }_i$ while $\lambda$ stands for the stretch defined as $\lambda =1+\epsilon .$

The data were post-processed thanks to Python libraries to obtain: the stress–strain curves and the optimised effective elastic moduli for toe and linear regions as well as the hyperelastic model parameters.

3. Results and discussion

HES staining, presented in Figure 1, showed that UC-membrane exhibited a fibrous and porous aspect with mainly collagenous matrix while Bio-Gide® membranes has a muscle-like aspect with dense bundles.

Figure 1. HES staining of UC- (left) and Bio-Gide® membranes (right).

From a mechanical point of view, both Bio-Gide® and UC-membranes behave similarly with nonlinear stress strain curves (Figure 2). The cyclic strain loads (not presented on Figure 2) highlighted hysteresis which are consistent with the expected visco-elastic behaviour of these materials. The mass and volume properties of the samples gave their density summed up in the Table 1 confirming the HES staining conclusion regarding the lower density of the UC-membrane.

Figure 2. Stress strain curves with UC-membrane curves in green and Bio-Gide® in blue.

Table 1. Properties of the hydrated UC- (n = 9) and Bio-Gide® (n = 2) membranes (mean ± SD).

	Bio-Gide^®	UC-membrane
Density (kg.m^-3)	173.60 ± 2.09	36.99 ± 3.78
Yield stress (kPa)	589.82 ± 37.01	446.32 ± 86.62
Yield strain (%)	22.29 ± 1.65	36.45 ± 4.72
Toe modulus (kPa)	1167.05 ± 26.99	288.89 ± 42.24
Linear modulus (kPa)	4265.28 ± 161.72	2160.05 ± 646.56
Shear modulus (kPa)	317.91 ± 9.55	72.11 ± 53.12

As the failure of the samples was difficult to detect, occurring either at an end or in the centre, the yield point has been considered for comparison. The mechanical properties extracted from the tests and presented into the Table 1 are consistent with the literature (Raz et al. 2019). Although withstanding high strain loading up to 35% on average, UC-membranes exhibited a yield stress of around 0.4 MPa. Therefore, with elastic moduli three times lower than the Bio-Gide® samples, UC-membranes are more compliant and softer, suggesting a good handling of the membrane while surgical use. However, stiffening UC-membranes could be considered in order to avoid the membrane collapsing into the bone defect. The fitting obtained with the Ogden model reproduces the nonlinear behaviour of the membranes as presented in Figure 2 with the plots using the average parameters’ values from Table 1.

Eventually, from a manufacturing point of view, UC-membranes exhibit comparable properties with respect to Bio-Gide® ones with almost five times lower densities.

4. Conclusions

In conclusion, scaffolds made of freeze-dried umbilical cord outperform existing products regarding their biological and quasi-static mechanical use. The hyperelastic characterization points out bilinear approach limits giving overestimated elastic moduli for numerical simulation. Nonetheless, this study does not consider the viscous behaviour that plays a role on the device wearing life. It is currently under investigation as well as in vivo bone regeneration in calvaria bone defect.