Mimicking the Human Tympanic Membrane: The Significance of Scaffold Geometry

Abstract

The human tympanic membrane (TM) captures sound waves from the environment and transforms them into mechanical motion. The successful transmission of these acoustic vibrations is attributed to the unique architecture of the TM. However, a limited knowledge is available on the contribution of its discrete anatomical features, which is important for fabricating functional TM replacements. This work synergizes theoretical and experimental approaches toward understanding the significance of geometry in tissue-engineered TM scaffolds. Three test designs along with a plain control are chosen to decouple some of the dominant structural elements, such as the radial and circumferential alignment of the collagen fibrils. In silico models suggest a geometrical dependency of their mechanical and acoustical responses, where the presence of radially aligned fibers is observed to have a more prominent effect compared to their circumferential counterparts. Following which, a hybrid fabrication strategy combining electrospinning and additive manufacturing has been optimized to manufacture biomimetic scaffolds within the dimensions of the native TM. The experimental characterizations conducted using macroindentation and laser Doppler vibrometry corroborate the computational findings. Finally, biological studies with human dermal fibroblasts and human mesenchymal stromal cells reveal a favorable influence of scaffold hierarchy on cellular alignment and subsequent collagen deposition.

Keywords: biofabrication; characterization tools; computational modeling; tissue engineering; tympanic membranes.

Figure 1

Schematic diagram illustrating the overall flowchart of the work. 3D: three‐dimensional; ES: electrospinning; FDM: fused deposition modeling; TM: tympanic membrane.

Figure 2

Generation of 3D TM models using a Python script. A) Demonstration of the versatility of the developed script in generating structures with varying fiber arrangement, shape, and dimensions (i‐vi). B) Conical TM models (an example shown with height 2.2 mm) can be created with the script. C) Chosen test cases to decouple the key geometrical features of the native TM.

Figure 3

COMSOL‐based computational models to predict the mechanical and acoustical behavior of the TM scaffolds. A) Simulated mechanical response at a defined scan‐rate of 3 µm s^‐1. The scaffold deformation with time, highlighted at three different stages: (T1) 0 s, (T2) 250 s, and (T3) 1000 s. B) Simulated acoustical response at a constant sound pressure of 0.02 Pa. (S1–S5) Key deformation shapes chosen to depict the different modes of vibrations, occurring at specific resonant frequencies of the TM models: (S1) initial state of the scaffolds, (S2) one circular node, (S3) one circular node and one nodal diameter, (S4) one circular node and two nodal diameters, and (S5) two circular nodes.

Figure 4

Optimization of hybrid biofabrication approach. A) Optimization of the electrospinning parameters: (i–iv) evolution of fiber morphology from beads to nanofibers, followed by thicker microfibers (scale bar = 10 µm), (v) fiber diameter with respect to polymer concentration (12–20%) and applied voltage (12, 20, and 28 kV), (vi) heat map demonstrating the frequency distribution of the fiber diameters (left y‐axis) obtained with different voltages, and (vii) fiber diameter with respect to collecting substrate and ES duration; conditions used for the parametric optimization: Al = aluminum foil, 5 = 5 min, and F = flat collector; conditions used for the final scaffold production: PP = polypropylene sheet, 30 = 30 min, R = rotating mandrel. B) Optimization of operating parameters for the fused deposition modeling based 3D fiber deposition: (i) patterns produced with different needle classes (ID184, ID100, and ID70) at a constant pressure of 750 kPa and screw speed of 70 rpm, (ii) varying pressure for ID70 at a constant screw speed of 50 rpm; * indicates the comparison between screw speeds of 70 rpm and 50 rpm at 750 kPa with ID70, (iii–iv) bar graphs summarizing and comparing the average filament diameters obtained from (i) and (ii), respectively.

Figure 5

Mechanical and acoustical characterization of the fabricated scaffolds. A) A macroindentation approach was implemented for evaluating the mechanical response: (i) images taken during the measurement at time‐points: (T1) 0 s, (T2) 250 s, (T3) 1000 s, and (T4) perforation; scale bar = 10 mm, (ii) schematic representation of the underlying physics, F = force applied, T = tension, (iii) average stress‐strain curve until 60% (normal) and 10% (inset) strain, (iv) average Young's moduli for the fabricated scaffolds (n = 4) compared with that of the native human TM (‡ calculated from select literature, Table 1), and (v) average resilience values summarized for all the cases. B) Acoustical characterization of the TM scaffolds with laser Doppler vibrometery: (i–iv) individual representation of datasets (n = 3) with normalized displacement plotted against the frequency sweep of the applied sound pressure wave for (i) control, (ii) case I, (iii) case II, and (iv) case III; (v) comparative summary of the mean curves for each case along with native TM (n = 2), and (vi) average first resonance frequencies.

Figure 6

Assessment of cell distribution and alignment on the TM scaffolds. A) Phalloidin labeled F‐actin (green) at days 1 and 7 for human dermal fibroblasts cultured on the fabricated scaffolds; scale bar = 3 cm. B) Phalloidin labeled F‐actin (green) and DAPI nuclear staining (blue) at days 1, 3, and 7 highlighting the influence of FDM filaments on guiding cell distribution and alignment; scale bar = 250 µm (lower magnification) and 50 µm (higher magnification). C) Quantification of (i) cell distribution (n > 45, technical replicates per region and time point) and (ii) cell alignment (n > 18, technical replicates per region and time point), with respect to regions: with no filament, along radial filament, along circumferential filament, and junction between two filaments.

Figure 7

Collagen deposition by the cultured TM scaffolds. A) CNA35‐FITC staining at day 7 and day 14 highlights the gradual production of collagen (yellow) by the seeded human dermal fibroblasts; scale bar = 3 cm. B) Higher magnification images of the deposited collagen (yellow) along with F‐actin (magenta) and nuclei (cyan): Phalloidin/DAPI (left) and CNA35‐FITC/DAPI (right); scale bar = 100 µm.