Biomimetic Tympanic Membrane Replacement Made by Melt Electrowriting

Abstract

The tympanic membrane (TM) transfers sound waves from the air into mechanical motion for the ossicular chain. This requires a high sensitivity to small dynamic pressure changes and resistance to large quasi-static pressure differences. The TM achieves this by providing a layered structure of about 100µm in thickness, a low flexural stiffness, and a high tensile strength. Chronically infected middle ears require reconstruction of a large area of the TM. However, current clinical treatment can cause a reduction in hearing. With the novel additive manufacturing technique of melt electrowriting (MEW), it is for the first time possible to fabricate highly organized and biodegradable membranes within the dimensions of the TM. Scaffold designs of various fiber composition are analyzed mechanically and acoustically. It can be demonstrated that by customizing fiber orientation, fiber diameter, and number of layers the desired properties of the TM can be met. An applied thin collagen layer seals the micropores of the MEW-printed membrane while keeping the favorable mechanical and acoustical characteristics. The determined properties are beneficial for implantation, closely match those of the human TM, and support the growth of a neo-epithelial layer. This proves the possibilities to create a biomimimetic TM replacement using MEW.

Keywords: MEW; biomimicry; implants; melt electrowriting; tympanic membranes.

Figure 1

Biomimetic approach of designing a melt electrowritten TM implant. A) Stereo microscopic image of a human TM with the collapsed 3D curvature toward the middle ear. B) Top view of the TM from A) with additional sketch (white lines) to indicate the assumed circular and the radial orientation of collagen fibers within the TM. C) Measurement of the mechancical properties of a human TM with an iIllustration of the mechanical fixation and measuring concept. D) Sound transfer function frequency response curve of four human TMs with removed malleus in the test stand, measured with laser Doppler vibrometry at the umbo from the medial side. E) Porous melt electrowritten PCL scaffold mimicking the 3D curvature of the TM. F) Melt electrowritten scaffold mimicking the radial and circular collagen fiber orientation.

Figure 2

A) Photo showing the melt electrowriting process with the used printing parameters. B,C) Microscopic images of the chosen designs (90° and 45° layer‐to‐layer orientation) for investigations upon the acoustical and mechanical interplay of different scaffold parameters. The insets show cutout enlargements.

Figure 3

Investigation of the mechanical response of different scaffold designs. A) Pressure/relative displacement curves of fiber diameters of 10 and 15 µm and of different number of layers (4, 6, 8) at a layer‐to‐layer orientation of 90°. B) The effect of an increase of the fiber spacing from 150 to 250 µm for both layer‐to‐layer orientations and a fiber diameter of 10 µm, additionally for 4, 6, and 8 layers, is demonstrated. C) The bending stiffness of all designs is shown depending on the number of layers. The range of the measured bending stiffness of the investigated TM is highlighted in orange, the dotted line indicates the average bending stiffness of the fully spanned TM. D) The bending stiffness is ordered depending on the average scaffold thickness (n = 5, ± standard deviation).

Figure 4

Acoustic properties of MEW scaffolds, measured with higher fixation forces. A) Comparison of sound transfer function mean curves (for standard deviation see Figure S7 in the Supporting Information) between scaffolds with 10 and 15 µm fiber diameter, 4 to 8 layers, 90° layer orientation, 150 µm fiber spacing. B) Comparison of sound transfer function mean curves (for standard deviation see Figure S7 in the Supporting Information) between 150 and 250 µm fiber spacing scaffolds with 4 to 8 layers, 45° layer orientation and 10 µm fiber diameter, C,D) Box plots of the first resonance frequencies and E) Sound transfer function mean curves (n = 5) and standard deviations for the scaffold designs with similar bending stiffness, as described in 2.3.4.

Figure 5

Application of collagen type I to the scaffold design 4L45°d10w250 and mechanical characterization. Ai) It is demonstrated that collagen covers the complete scaffold surface achieving an airtight structure. Even small areas revealing more complex pore geometries were filled with collagen as shown in (Aii) and (Aiii). B) The change of the mechanical properties through the application of gamma‐ sterilization, ultraviolet radiation w/ and w/o collagen coating. Static (n = 5, ± standard deviation) and cyclic loading measurements were performed for scaffolds w/ and w/o collagen, w/o any sterilization method. C) The response of 4L45d10w250 w/ and w/o collagen toward static and cyclic mechanical loading is depicted. D) The sound transfer function mean curves of collagen coated scaffolds and the mean (n = 3) and standard deviation in comparison to the mean human TM sound transfer function in the test stand is shown. The mean first resonance frequency is at about 370 Hz.

Figure 6

In the upper half of the Figure an image sequence showing the first vibration modes of a human tympanic membrane (n = 1) is depicted. The first mode is prominent at a frequency of 400 Hz, the 2nd mode at 808 Hz, and the 3rd at 1849 Hz. The lower half shows the three visible vibration modes of a flat collagen coated MEW scaffold (n = 1) with the design 4L45d10w250 which are prominent at 340, 800, and 1240 Hz. The first mode is expected to have the biggest contribution to the vibration of the middle ear. The maximum magnitude was decreasing with higher modes (scaled up for better visibility) as observed for the TM. The differences in the mode frequencies were caused by structural differences (e.g., curved vs. flat shape) and natural variations. The higher mode frequencies should provide a bigger variance than the first mode.

Figure 7

Adhesion, viability and proliferation of human keratinocytes, grown on MEW membranes with and without collagen coating for 14 days. A) Cell adhesion and viability shown with fluorescence micrographs after Live/Dead staining; viable cells appear in green and dead cells in red. B) The morphology and density of the cell layers, formed over time, is demonstrated by fluorescence microscopy, applying DAPI/Phalloidin staining (DAPI for cell nuclei/blue and Phalloidin for actin cytoskeleton/green). Scale bars in (A) and (B) represent 250 µm. SEM images confirm the results from fluorescence microscopy and show the typical, polygonal morphology of the keratinocytes, especially on collagen coated scaffolds at day 14. C) magnifications 200×. D) Cell proliferation was quantified by measurement of DNA content and intracellular LDH activity for cells grown on scaffolds w/ and w/o collagen and w/ and w/o UV treatment over 2 weeks; two individual experiments for LDH activity and DNA quantification were performed with each n = 3. Average n = 6 ± standard deviation.