Biomimetic Tympanic Membrane Grafts

Abstract

The tympanic membrane (TM), commonly known as the eardrum, is a thin tissue that captures sound waves from the environment and transmits them to the ossicles of the middle ear. Importantly, its circumferential and radial collagen fiber architecture enables efficient sound-induced motion over a wide range of frequencies that match the human cochlea’s dynamic range (20 – 20,000 Hz). While its delicate structure serves a crucial acousto-mechanical role in sound conduction, the fragile TM can be damaged by exposure to impulse blast waves, chronic ear infections, or trauma. While small TM perforations may heal spontaneously, many perforations do not heal on their own. When damaged, the TM is no longer efficient at transducing sound pressure waves into mechanical motion and hearing loss results. Moreover, perforated TMs permit pathogens to enter the middle ear space, resulting in infection, pain and dizziness. Surgical repair of the TM, also known as tympanoplasty, uses grafted materials harvested from the patient to close the perforation and restore proper sound conduction to the ossicular chain. Unfortunately, commonly used graft materials, such as temporalis fascia, cartilage, fat, and dermis, do not degrade nor remodel after implantation. These reconstructed TMs also lack the native circumferential and radial collagenous architecture and, thus, are less efficient at sound conduction. As a result, healing and hearing outcomes in tympanoplasty patients are often poor.

The overarching goal of this Ph.D. thesis is to create biomimetic TM grafts with anisotropic architecture that overcome the current tympanoplasty challenges described above. Specifically, this research focuses on novel TM designs, synthesis of biodegradable, biocompatible materials, 3D printing, and in vitro and in vivo characterization of their performance. In this dissertation, the parameters for an “ideal TM graft” are first described followed by a discussion of the research carried out to achieve the stated goal of improved TM grafts and tympanoplasty outcomes.

We first investigated the TM structure via histopathology to determine the circumferential and radial fiber arrangement in the human TM. Inspired by the native TM, we created synthetic TM grafts via 3D printing of three polymeric inks: polydimethylsiloxane (PDMS), poly(lactic acid) (PLA), and polycaprolactone (PCL). Their printed circular/radial architectures were infilled with a hydrogel solution to create a confluent graft capable of transmitting sound. Next, we carried out in vitro acoustic testing via laser Doppler vibrometry (LDV) and digital opto-electronic holography (DOEH) to obtain high frequency and spatial resolution of sound-induced motion. We find that these first-generation biomimetic TM grafts exhibited high fidelity in their sound-induced motion compared to inconsistent motion patterns in harvested autologous fascia grafts. Through simple modal motion patterns at low frequencies and complex, organized motion patterns at high frequencies, biomimetic composite grafts were able to transmit sound in a manner similar to the unloaded native TM. We also find that their sound-induced motion varies with graft material and architecture, demonstrating their tunability via 3D printing. We then explored the mechanical properties of the printed and infilled grafts as well as their propensity to enable cellular ingrowth and remodeling. Biomimetic TM grafts exhibited lower hysteresis compared to fascia grafts under repeated loading. We find that fibroblasts seeded onto these grafts deposit collagen I in the infill region. However, cellular remodeling occurred isotropically due to limitations in both material properties and in the isotropic nature of the infilled region between printed features.

To further optimize the performance of biomimetic TM grafts, we synthesized a biomaterial ink that enables post-implantation anisotropic remodeling in a controlled fashion, guided by the deterministic architecture defined during 3D printing. This biomaterial was designed with a tunable biodegradation rate, mechanical properties that mimic the native TM and facilitate manipulation during placement, and biocompatibility with relevant cell types. Specifically, we created a thermoplastic poly(ester urethane urea) (PEUU) elastomer blended with a fugitive poly(ethylene glycol) (PEG) component to form porous PEUU (P-PEUU) grafts upon leaching PEG species from the graft material. We characterized the rheological and printing behavior of this ink, alongside control PEUU, PCL, and porous PCL (P-PCL) inks, and optimized these parameters for high operating temperature-direct ink writing (HOT-DIW). After fabricating these second-generation TM grafts, we measured their Young’s Moduli, in vitro degradation, in vitro proliferation of fibroblast and keratinocyte cells, topography, and alignment of fibroblasts along the print path under various print conditions. Specifically, the mean Young’s modulus of P-PEUU specimens that were printed parallel to the tension testing direction (E|| = 33.7 ± 2.8 MPa) is 142% that of those printed orthogonal to the tension testing direction the (E⊥ = 23.8 ± 1.5 MPa). Grafts from P-PEUU inks aligned fibroblasts and extracellular deposited collagen I to a greater extent than grafts from both PEUU and P-PCL inks printed with the same speed and filament width, while PCL grafts and melted P-PEUU grafts did not show any degree of fibroblast alignment or extracellular collagen alignment. Notably, increasing the print speed from 5 mm/s up to 20 mm/s while maintaining identical filament width leads to higher cellular and collagen I alignment in P-PEUU grafts. Together, these observations revealed that our synthetic P-PEUU ink is well suited for creating biomimetic TM grafts that may induce anisotropic remodeling.

Next, we investigated the acoustic and mechanical responses of the biomimetic TM grafts (based on P-PEUU) with the goal of optimizing their architecture to best match native TM properties in vivo. Specifically, printed grafts were patterned in a 50 circular and 50 radial (50C/50R) architecture as well as a 50 circular only (50C) architecture. Control TM grafts composed of human fascia and a porcine small intestinal submucosa, which are commonly used in tympanoplasty, were used as controls. Based on LDV and DOEH in vitro acoustic tests, we find that these biomimetic TM grafts exhibited superior sound-induced motion compared to both control grafts as well as to circular-only P-PEUU grafts and isotropic versions of the same printed grafts (melted to “erase” their 50C/50R architecture) with identical thickness. Biomimetic P-PEUU grafts showed simple modal motion patterns at low frequencies that increase to more complex motion patterns at higher frequencies, similar to isolated human TM tissue. Finally, we carried out in vivo studies on biomimetic P-PEUU grafts and control grafts implanted into chinchilla models of chronic TM perforations. After 3 months, chinchillas implanted with biomimetic P-PEUU 50C/50R grafts exhibit higher healing rates and superior hearing outcomes compared to control grafts (human fascia and Biodesign® porcine small intestinal submucosa). The structure of these biomimetic TM grafts is partially degraded via ingrowth of native cells and host vasculature, as determined by post-mortem histology. The P-PEUU grafts were well tolerated and non-toxic to the surrounding tissue and the delicate structures within the inner ear.

In summary, we have developed biomimetic TM grafts through the integration of design, biomaterial synthesis, and additive manufacturing with improved performance compared to native (control) grafts. We have validated their functionality through a combination of in vitro and in vivo experiments. The work presented in this Ph.D. dissertation provides a strong foundation for translating these biomimetic TM grafts into human patients for use in tympanoplasty and possibly well beyond.