>> Research Projects >> Textured Hollow Fiber Membranes

1. Fabrication of semipermeable hollow fiber membranes with highly aligned texture for nerve guidance

Among a variety of tubular structures used for guidance channels, a semipermeable hollow fiber membrane (HFM) shows very promising results in promoting axonal regeneration both in vitro and in vivo. HFMs can reduce the infiltration of fibrous tissue, provide a conduit for the diffusion of neurotropic and neurotrophic factors, increase the concentration of endogenous proteins inside the channel, and present a barrier to selectively permit or inhibit the diffusion of macromolecules between the bridging device and the surroundings.However, the HFMs used so far lack guidance cues, because the diameter of the HFMs is significantly larger than the physiological size of the neurites. To overcome this problem, Wen et al. have entubulated thin filaments (5 m in diameter) into the HFMs, which have shown increased orientation and outgrowth rates of regenerating neurites (Full paper). Another approach to increase the guidance role of semipermeable HFMs is to generate aligned grooves on the inner surface of semipermeable HFMs. To this end, semipermeable HFMs with highly aligned textures were fabricated for promoting nerve regeneration. By precisely controlling the fabrication parameters, such as polymer solution flow rate, coagulant solution flow rate, and the air-gap distance, also called drop height, different-sized aligned grooves can be fabricated on the inner surface of HFMs. The working hypothesis of this study is that highly aligned textures on the inner surface of permeable hollow fiber membranes can increase the directionality and outgrowth rate of regenerating neurites. The long-term objective is to engineer nerve grafts in vitro based on a highly aligned 3D scaffolds for the treatment of spinal-cord injury and nerve damage. Preliminary studies using in vitro dorsal root ganglion (DRG) regeneration assay showed that both the alignment and outgrowth rate of regenerating axons increased significantly on HFMs with aligned textures compared to those on HFMs with a smooth inner surface. Studies in progress are evaluating axonal outgrowth and regeneration using in vivo sciatic-nerve and spinal-cord-injury models.

Mathematical modeling of the texture formation
The texture forming mechanism is qualitatively explained using a PU-DMSO-water ternary phase diagram and the process dynamics. The texture forming is modeled as sinusoidal along the perimeter direction as follows:


It is considered that the texture is initialized by the chemical energy gradient such as the surface tension gradient and/or thermal fluctuation in the bulk solution and/or the interface between the solution and coagulant. The high frequency modes decay within a short time, and the low frequency modes are dominant in the final profile of the inner surface because the precipitation is so rapid that these low frequency modes may have no enough time to decay. All these combined effects lead to the groove texture on the inner surface under specific conditions.


Figure 1. Morphology of hollow fiber membranes fabricated with the use of a phase inversion technique. (A)-(C) are conventional single-skin HFMs with smooth inner surface. (D)-(F) are single-skin HFMs with highly aligned grooves on the inner surface. (A) and (D) are lower-magnification images of the cross section; (B) and (E) are higher-magnification images of the cross section; and (C) and (F) are of the inner surface.

Figure 2. Morphology of HFMs with highly aligned textures on the inner surface. These HFMs were fabricated at constant water flow rate of 4 mL/min, constant drop height of 10 cm, and polymer flow rate at (A) 1.2 mL/min, (B) 1.6 mL/min, and (C) 1.8 mL/min. The average groove heights were (A) 38.5 um, (B) 61.6 um, and (C) 91 um. The average full-width half-maximums (FWHM) were (A) 49.2 um, (B) 58.7 um, and (C) 79.8 um, respectively.

2. A tubular scaffold with highly aligned texture on the inner surface for small-diameter blood vessel  tissue engineering: threshold for endothelial cell alignment

In a normal healthy blood vessel, endothelial cells (ECs) are highly aligned in the direction of the vessel long axis or the blood flow through the organization of cytoskeletal filaments and focal adhesion complexes. Such alignment is important for the functionality of blood vessels in vivo, although the responsible factors or pathways have not been fully identified. It was demonstrated that elongated ECs were resistant to atherosclerotic lesions, which were preferentially formed on ECs with round morphology. Under traditional cell culture conditions, ECs exhibit a cobblestone-like morphology, leading to functional and behavioral deviations from the normal ECs in vivo. To reproduce the oriented morphology, therefore, achieve functionality of ECs in tissue-engineered blood vessels, many attempts have been made, however, are associated with problems. For example, exposure of cells to flow is invasive, culture of cells in microchannels is not feasible for tissue engineering purposes, patterned substrates or extracellular matrix have short lifetimes, and creating substrates with waves or grooves using microfabrication techniques is time consuming. To overcome these problems, we have developed a simple fabrication technique based upon a wet phase inversion process that allows the formation of well-controlled aligned textures (Fig. 3) on the inner surface of a tubular scaffold, which has the potential for applications involving small-diameter blood vessel tissue engineering.  

We found that on the smooth control inner surface, ECs assumed spreading cubic morphologies without a preferential direction of alignment (Fig. 4A). Regardless of the form of the aligned textures, among the five sizes (10, 25, 50, 75, and 100 um in the largest cross-dimension) that were examined, EC alignment along the HFM long axis was only observed when the largest cross-dimension of the texture is no greater than 50 um (Fig. 4B-E), which is the spreading size of ECs, suggesting that EC spreading size may be the threshold value in the largest cross-dimension of an aligned texture to induce EC alignment. This threshold value is greater than those reported in similar studies using other cell types, such as neurons, epithelial, macrophages, and neutrophils, perhaps due to much larger surface area of ECs in their spreading form, which enables them to sense large textures. Below 50 um in the largest cross-dimension for each texture, ECs exhibited an increasing degree of alignment along the HFM long axis as a function of decreasing largest cross-dimension of the texture (Fig. 2C-E). Convex and concave sides of the same texture did not show difference in affecting EC alignment. These data were also confirmed on flat substrates with microfabricated aligned textures.
 

Figure 3: The morphologies of the inner surfaces of tubular scaffolds tested. (A) Smooth inner surface, (B) Inner surfaces with aligned textures, (C) convex and (D) concave side of a round wave texture, (E) convex and (F) concave side of a sharp wave texture, (G) convex and (H) concave side of a flat groove texture.


Figure 4: ECs cultured on (A) smooth inner surface, (B) 75 um grooves, (C) 50 um grooves, and (D and E) 30 um sharp wave textures. Green is staining for actin filament; red is staining for von Willebrand factor.