Modulation of epithelial tissue and cell migration by microgrooves

We used a polystyrene substratum to study the response of migrating epithelium to 1- or 5-microm depth microgrooves with groove/ridge widths of 1, 2, 5, or 10 microm. The migration of a tissue sheet was enhanced along the microgrooves, while migration across the microgrooves was inhibited. Changing the depth of the microgrooves had a greater effect on migration than alteration of the groove/ridge width. The migration of epithelial cells from a confluent monolayer culture followed a similar pattern to that of intact epithelial tissue. Cellular extensions generally followed the microgroove direction by tracking along the top of the ridges or following the ridge walls, as revealed by scanning electron microscopy. Actin filaments within the basal cell layer of the tissue were aligned with the microgrooves, unlike filaments in the superficial layers that did not appear to be affected by the presence of underlying microgrooves. The basal cell layer of the tissue conformed to the contours of the microgroove following migration. However, the ultrastructure of the tissue above the ridges resembled that of tissue on a flat surface. We concluded that surface microgrooves have the potential to direct the migration of immediately adjacent epithelial tissue, the effect of which is to guide epithelial tissue on the surface of implanted biomaterials.     

Effect of microtextured surfaces on the performance of percutaneous devices

Along the percutaneous part of implantable devices, like (semi-)permanent catheters, epithelial downgrowth can occur. This process can eventually lead to implant loss. Various treatments for the catheter surface have been proposed, to improve their performance. The purpose of the current study was to investigate the effect of a microgroove pattern on the tube surface, on epithelial downgrowth. Catheterlike implants were made of silicone tubes, with anchoring cuffs made of titanium-fiber mesh. A thin sheet of silicone with microgrooves was applied on the tubes. Two types of texturing were used, a square groove of 10 microm wide and 1 microm deep; or a V-shaped groove of 40 microm wide. The grooves were directed either along the long axis of the catheter tube (grooves perpendicular to the skin surface) or circling around the catheter (grooves directed parallel to the skin surface). As controls, catheters with a smooth outer surface were used. Implants were placed in 30 rats, with a follow-up period of 9 weeks. During this time, animals were inspected biweekly, and catheter exit sites were evaluated using a scoring system. At the end of the 9-week period the implants and surrounding tissues were processed for histological evaluation. For the clinical evaluation of the exit sites, no statistical differences were found between the study groups. Histologically, epithelial downgrowth was observed for all samples. The histomorphometrical measurements showed that there were no differences in downgrowth between the smooth and parallel-grooved catheters. However, there was increased epithelial downgrowth along the catheters with grooves perpendicular to the skin. In conclusion, a grooved microtexture can direct epithelial tissue ingrowth, but this study found no beneficial effects of the guidance phenomenon.     

Transforming growth factor-beta 1, 2, and 3 can inhibit epithelial tissue outgrowth on smooth and microgrooved substrates

In this study, we describe the influence of parallel surface microgrooves, and of TGF-beta, on the outgrowth of corneal epithelial tissue. Microgrooves (depth 1 microm, width 1-10 microm) were made in polystyrene culturing surfaces. These surfaces were left untreated, or loaded with TGF-beta 1, 2, or 3 (6.0 ng/cm(2)). Subsequently, epithelial explants from bovine corneas were placed on the experimental surfaces. After 9 days of culturing, tissue outgrowth was evaluated. Furthermore, the tissue cultures were analyzed histologically. It was shown that epithelial tissue grew from the explants over all experimental surfaces. On microgrooved surfaces outgrowth proceeded in the direction of the grooves, rather than perpendicular to the grooves. The addition of each type of TGF-beta resulted in a reduction of outgrowth. However, outgrowth remained directed by the grooves. Further, the explants had shrunk after TGF treatment. Histology showed that this shrinkage was not related to alpha-smooth muscle actin expression in the explants. We conclude that microgrooves can direct, and TGF-betas can inhibit the outgrowth of epithelial tissue. This finding could be useful in biomaterial applications where the growth of epithelial tissue needs to be discouraged.     

Topographical control of human neutrophil motility on micropatterned materials with various surface chemistry

Controlling cell responses to an implantable material is essential to tissue engineering. Because the surface is in direct contact with cells, both chemical and topographical properties of a material surface can play a crucial role. In this study, parallel ridges/grooves were micropatterned on glass surfaces using photosensitive polyimide to create transparent substrates. The migratory behavior of live human neutrophils on the patterned surfaces was observed using a light microscope with transmitted light source. The width (2 microm) and length (400 microm) of the ridges were kept constant. The height (5 or 3 microm) and the repeat spacing (6-14 microm) of the ridges were systematically changed to investigate the effect of microgeometry on neutrophil migration. In addition, the effect of surface chemistry on neutrophil migration was studied by deposition of a thin layer of "inert", biocompatible metal such as Au-Pd alloy and titanium on patterned substrates. More than 95% of neutrophils moved in the direction of the long axis of ridges/grooves regardless of the topographical geometry and chemistry, consistent with a phenomenon termed "contact guidance". Therefore, cell migration was characterized using a one-dimensional persistent random walk. The rate of cell movement was strongly dependent on the topographical microgeometry of the ridges. The random motility coefficient mu, 9.8 x 10(-9) cm2/s, was the greatest at a ridge height of 5 microm and spacing of 10 microm, about 10 times faster than on smooth glass surface. The Au-Pd coating did not change neutrophil migratory behavior on patterned surfaces, whereas titanium decreased cell motility substantially. The results of this study suggest that optimization of both surface chemistry and topography may be important when designing biomaterials for tissue engineering. In addition, parallel ridges/grooves can be used to control the direction and rate of cell migration on the surface.     

Chondrocyte aggregation on micrometric surface topography: a time-lapse study

Cellular aggregation or mesenchymal "condensation" is a prerequisite in the process of chondrogenesis. It has been observed that during in vitro engineering of cartilage, chondrocytes form aggregates in the initial cell seeding process of polymer scaffolds. However, the exact mechanism behind this aggregation has yet to be elucidated, although cell collision has been implicated. As all polymers have a surface topography, we hypothesized that topography may play a role in chondrocyte aggregation. 1(st) and 2(nd) passage chondrocytes were seeded on micrometric topography ranging from 0.75 to 8 microm in depth and 5 to 12.5 microm in width. Both 1(st) and 2(nd) passage cells formed aggregates as cells collided, and larger aggregates formed as aggregates collided with each other on the grooves. Furthermore, aggregates migrated parallel to the direction of the groove long axis. F-actin organization was altered only in the cellular layer in direct contact with the surface; stress fibers oriented in the direction of the groove long axis. Chondrocytes maintained type II collagen expression on all groove depths. This study shows that micrometric grooves could be an effective means for studying chondrocyte aggregation, and could additionally be utilized in the regeneration of cartilage.     

Nanoscale topography modulates corneal epithelial cell migration

The purpose of this study was to evaluate the effect of surface topographic features that mimic the corneal epithelial basement membrane on cell migration. We used electron-beam and X-ray lithography and reactive ion etching to pattern silicon wafers with pitches (groove width plus ridge width) of nano- and microscale dimensions (pitches ranged from 400 to 4000 nm). Additionally, polyurethane patterned surfaces were created by replication molding techniques to allow for real-time imaging of migrating cells. Individual SV40-transformed human corneal epithelial cells frequently aligned with respect to the underlying surface patterns and migrated almost exclusively along grooves and ridges of all pitches. Direction of migration of individual cells on smooth surfaces was random. In cell dispersion assays, colonies of cells migrated out from initially circular zones predominantly along grooves and ridges, although there was some migration perpendicular to the ridges. On smooth surfaces, cells migrated radially, equally in all directions, maintaining circular colony shapes. We conclude that substratum features resembling the native basement membrane modulate corneal epithelial cell migration. These findings have relevance to the maintenance of corneal homeostasis and wound healing, as well as to the evolution of strategies in tissue engineering, corneal prosthesis development, and cell culture material fabrication.     

Small-diameter porous poly (epsilon-caprolactone) films enhance adhesion and growth of human cultured epidermal keratinocyte and dermal fibroblast cells

Autologous keratinocyte grafts provide clinical benefit by rapidly covering wounded areas, but they are fragile. We therefore developed biocompatible hexagonal-packed porous films with uniform, circular pore sizes to support human keratinocytes and fibroblasts. Cells were cultured on these porous poly (epsilon-calprolactone) films with pore sizes ranging from novel ultra-small 3 microm to 20 microm. These were compared with flat (pore-less) films. Cell growth rates, adhesion, migration, and ultrastructural morphology were examined. Human keratinocytes and fibroblasts attached to all films. Furthermore, small-pore (3-5 microm) films showed the highest levels of cell adhesion and survival and prevented migration into the pores and opposing film surface. Keratinocyte migration over small-pore film surface was inhibited. Keratinocytes optimally attached to 3-microm-pore films due to a combination of greater pore numbers (porosity), a greater circumference of the pore edge per unit surface area, and greater frequency of flat surface areas for attachment, allowing better cell-substrate and cell-cell attachment and growth. The 3-microm pore size allowed cell-cell communication, together with diffusion of soluble nutrients and factors from the culture medium or wound substrate. These characteristics are considered important in developing grafts for use in the treatment of human skin wounds.     

The effect of topographic characteristics on cell migration velocity

The migration of cells on structured surfaces is known to be affected by its surface topography. Although the effects of topography have been extensively investigated the crucial parameters determining the cell-surface reaction are largely unknown. The present study was performed to describe and to define the role of groove/elevation (ridge) dimensions at the micrometre scale on fibroblast cell migration by correlating cell shape, migration angle alpha, cell orientation beta and velocity with these dimensions. For this a quantitative method was developed. We could show that the surface structures significantly influenced migration direction alpha, cell orientation beta and mean velocity, as well as migration speed in the directions parallel and perpendicular to the grooves/elevations in a surface structure dependant way. Cell migration velocity parallel, respectively, perpendicular to the structures was significantly affected by the geometries and dimensions of the substratum. Surface structures were not able to significantly affect distribution patterns of cell shapes. Overall, it could be shown that differently structured surfaces influenced the cells but no crucial feature could be clearly identified, suggesting that the reaction of the surface structure might be far more complex than generally is assumed.     

A PLGA membrane controlling cell behaviour for promoting tissue regeneration

Barrier membranes are used in periodontal applications with the aim of supporting bone regeneration by physically blocking migrating epithelial cells. We report a membrane design that has a surface topography that can inhibit epithelial cell migration and proliferation on one side and a topography that guides osteoblast migration to a desired area. A PLGA copolymer (85:15) blended with MePEG, was cast to have surfaces with smooth, grooved or sandblasted-acid-etched topographies. Epithelial cells spread on smooth surfaces after 24 h, and cell numbers increased after 5 days. Cells on the smooth surface exhibited no preferred direction of migration. On the sandblasted-acid-etched topography epithelial cells spread but the cell number did not significantly increase after 5 days. Cell migration was inhibited on this surface. Osteoblasts spread on both grooved and smooth surfaces and cell number increased after 5 days on all surfaces. The cells that adhered in the grooves migrated preferentially in the direction of the grooves. Positive alkaline phosphatase staining was seen on all surfaces within 4 weeks and positive Von Kossa nodule staining within 6 weeks. These results suggest that surface topographies replicated on opposite sides of a biodegradable polymers membrane can inhibit proliferation and migration of the epithelial cells, and promote proliferation and directional migration of osteoblasts. Addition of appropriate surface topographies to membranes used in guided tissue regeneration has the possibility of improving clinical performance in periodontal tissue regeneration procedures.     

Polymer design for corneal epithelial tissue adhesion: pore density

A porous polymer is required to sustain corneal epithelial tissue on the anterior surface of an implantable contact lens (corneal onlay). Porosity creates topography on the polymer surface and, if defined, this can be manipulated to elicit a particular tissue response. Previous work identified pores of 100-nm diameter to be the critical size of a discontinuity in a polymer surface to facilitate the migration, stratification, and sustained adhesion of corneal epithelial tissue. Now we address the issue of pore density. Corneal epithelial tissue was grown for 21 days on nonporous polycarbonate and polycarbonate track-etched with pores 100-nm diameter with pore densities ranging from 0.27 and 14.4% of the total polymer surface area. Histology was used to score epithelial structure, and electron microscopy was used to quantitate the formation of adhesive structures (basal lamina and hemidesmosomes) at the tissue-polymer interface. Data showed that epithelial tissue stratified and epithelial adhesive structures formed on polycarbonate surfaces with pore densities between 0.52 and 14.4% inclusive. This means that epithelial tissue can be maintained on a polymer where up to 14.4% of the surface is dedicated to small discontinuities in the form of pores 100-nm diameter. These figures can be used to specify the design of a polymer for applications requiring epithelial cover.     

Contact guidance of rat fibroblasts on various implant materials.

Providing a substrate surface with micrometer-sized parallel grooves influences the behavior of cells growing on such substrates in vitro. Cells elongate in the direction of the groove and migrate guided by the grooves. It has been suggested that cellular alignment on microgrooves is predominantly dependent on groove dimensions and that surface chemical variation of the substrate material has little effect. Therefore we seeded primary rat dermal fibroblasts (RDF) on smooth and microgrooved (groove width 1-10 microm, depth 0.5 microm) polystyrene (PS), poly-L-lactic acid (PLA), silicone (SIL), and titanium (Ti) substrates. The production process was found to be more accurate for PS and PLA than for SIL and Ti substrates. A proliferation study, scanning electron microscopy, confocal laser scanning microscopy, and transmission electron microscopy revealed differences between RDF behavior on the materials. Our conclusions are (1) the accuracy of microtexture production by casting depends greatly on the material used; (2) even if no sharp discontinuities are present, microtextures still are potent tools for inducing contact guidance; and (3) besides surface texture, surface chemistry has a definitive influence on cell morphology.     

Titanium-coated micromachined grooves of different dimensions affect epithelial and connective-tissue cells differently in vivo

A desirable feature of an implant surface which penetrates epithelium would be that the surface impedes epithelial downgrowth. Previous experiments have shown that the micromachined, horizontally oriented grooves on the percutaneous implant surface can impede epithelial downgrowth (Chehroudi et al., J. Biomed. Mater. Res., 22, 459 (1988) and 23, 1067 (1989)). However, little is known of the effect of varying groove parameters such as depth, spacing, and orientation on epithelial downgrowth and attachment of epithelial (E)-cells and fibroblasts (F) to percutaneous implants in vivo. Grooves were produced with a 30-micron pitch and depths of 22 microns, 10 microns, or 3 microns. In addition, 10-microns- and 3-microns-deep grooves were made with pitches of 39 microns and 7 microns, respectively. Implants with grooves oriented either horizontally or vertically to the long axis of the implant as well as smooth control surfaces were coated with 50 nm of titanium and placed in the parietal area of rats for a period of 7 days. Close attachment of E-cells was found on the smooth, 10-microns- and 3-microns-deep, horizontally or vertically aligned grooved surfaces; in contrast, E-cells bridged over the 22-microns-deep, horizontally oriented grooves. F formed a capsule on the smooth surface as well as the 10-microns- and 3-microns-deep horizontally oriented grooves, but F inserted obliquely into the 22-microns-deep, horizontally aligned grooved surface. Histomorphometric measurements indicated that the epithelial downgrowth was greatest on the vertically oriented grooved and smooth surfaces and was shortest on the 22-microns-deep and 10-microns-deep horizontally aligned grooved surfaces. These differences indicate that epithelial downgrowth was accelerated on the vertically oriented grooved surfaces and inhibited on the horizontally oriented grooved surfaces. Moreover, the mechanism of inhibition of the epithelial downgrowth may differ among these surfaces. E-cells bridged over the 22-microns-deep grooves and their migration appeared to be inhibited by the F that inserted into the implant surface. In the shallower horizontal grooves, however, epithelial downgrowth was probably inhibited by contact guidance because there was no evidence of F inserting obliquely into the implant surface.     

Neurite bridging across micropatterned grooves

After injury, regenerating axons must navigate complex, three-dimensional (3D) microenvironments. Topographic guidance of neurite outgrowth has been demonstrated in vitro with culture substrates that contain micropatterned features on the nanometer-micron scale. In this study we report the ability of microfabricated biomaterials to support neurite extension across micropatterned grooves with feature sizes on the order of tens of microns, sizes relevant to the design of biomaterials and tissue engineering scaffolds. Neonatal rat dorsal root ganglion (DRG) neurons were cultured on grooved substrates of poly(dimethyl siloxane) coated with poly-L-lysine and laminin. Here we describe an unusual capability of a subpopulation of DRG neurons to extend neurites that spanned across the grooves, with no underlying solid support. Multiple parameters influenced the formation of bridging neurites, with the highest numbers of bridges observed under the following experimental conditions: cell density of 125,000 cells per sample, groove depth of 50 microm, groove width of 30 microm, and plateau width of 200 microm. Bridges were formed as neurites extended from a neuron in a groove, contacted adjacent plateaus, pulled the neuron up to become suspended over the groove, and the soma translocated to the plateau. These studies are of interest to understanding cytoskeletal dynamics and designing biomaterials for 3D axon guidance.     

Effect of porosity and surface hydrophilicity on migration of epithelial tissue over synthetic polymer

The relative effects of porosity and surface chemistry on the migration of epithelial tissue over the surface of a polymer were determined in vitro. These studies compared nonporous polymers with those having 0.1-microm diameter track-etched pores and were conducted on polycarbonate and polyester. Epithelial tissue migration over the polymer surface was stimulated by the presence of these pores. The surface chemistries of the polymers were modified by deposition of various polymer films using radio frequency gas deposition, giving a range of surfaces that varied in air:water sessile contact angle (SCA) of between 26 and 100 degrees. Tissue migration on the nonporous surfaces was affected by the surface chemistry, being generally linear as a function of the SCA and higher on hydrophilic than on hydrophobic surfaces but reduced if the hydrophilic surface had a mobile chemistry. The effects of the 0.1-microm diameter pores and the surface hydrophilicity were additive with the maximal level of epithelial tissue migration occurring on a porous, hydrophilic polymer surface.     

Stimulation of epithelial tissue migration by certain porous topographies is independent of fluid flux

A surface with columnar pores 0.1 or 0.4 microm in diameter is shown to have a novel effect on the migration of corneal epithelial tissue sheets; migration is stimulated in a nondirectional manner with respect to migration over a planar, nonporous surface (Dalton, Evans, McFarland, and Steele, J Biomed Mater Res 1999;45:384-394; Steele, Johnson, McLean, Beumer, and Griesser, J Biomed Mater Res 2000;50:475-482). By blind-ending the pores, we show that this increase in tissue migration is not dependent on fluid flux through the pores and so appears to occur as a result of surface topography. From transmission electron micrographs, the migrating tissue appears to form either close contacts or focal adhesions at the edge of some pore channels; we speculate that this may provide a fulcrum for the enhanced migration. Scanning electron micrographs suggest that within tissue that migrates over the surfaces that contain blind-ended pores, the cells are more extensively spread than those in tissue migrating on a planar surface. The migration of disaggregated epithelial cells is enhanced on surfaces that contain 0.1- or 0.4-microm-diameter pores (compared with a planar surface), and this is similarly independent of fluid flux.     

Guidance of liver and kidney organotypic cultures inside rectangular silicone microchannels

We have studied the effect of rectangular polydimethylsiloxane (PDMS) microchannels on the behavior of embryonic liver and kidney explants maintained in contact with these microchannels. The microchannel widths were varied from 35 to 300 microm and depth from 45 to 135 microm. The growth of these tissue types were compared to the development on flat silicone and plastic control material. At seeding, due to the viscoelastic properties of both organs, "capillary-like filling" was observed inside the narrowest microchannels. In those cases, the tissues grew to a confluent layer joining the microchannels with no cell migration and proliferation inside the microchannels. In the largest microchannels, only a weak migration was observed and the cellular behavior appears quite similar to that of PDMS flat culture conditions. In intermediate geometries, we observed different tissue growth progressed inside those microchannels with an average growth properties inside the microchannels when compared to other sizes. The liver tissues velocity of up to 72 microm/day resulting to form a dense three-dimensional multicellular 'liver-like tissue'. Scanning electron microscopy (SEM) observations demonstrated that the tissue was organized like an epithelial layer with round cells embedded in an extracellular matrix. Liver cell mobility may result primarily from the activity of the marginal cells, whereas the sub-marginal cells appeared passively dragged. Parenchymal organization demonstrating differentiated states was also observed. Kidney grew mainly on the microchannel walls and the tissues never appeared dense and organized as the liver ones.     

A light and electron microscopic study of the effects of surface topography on the behavior of cells attached to titanium-coated percutaneous implants

Previous studies using light microscopy have demonstrated that micromachined grooved surfaces inhibit epithelial (E) downgrowth and affect cell orientation at the tissue/implant interface. This study investigates the ultrastructure of the epithelial and connective-tissue attachment to titanium-coated micromachined grooved, as well as smooth control, implant surfaces. V-shaped grooves, 3, 10, or 22 microns deep, were produced in silicon wafers by micromachining, replicated in epoxy resin, and coated with 50-nm titanium. These grooved, as well as smooth, titanium-coated surfaces were implanted percutaneously in the parietal area of rats and after 7 days processed for electron microscopy. The tissue preparation technique used in this study enabled us to obtain ultrathin sections with few artifacts from the area of epithelial and connective-tissue attachment. The histological observations demonstrated that E cells closely attached to, and interdigitated with, the 3-microns and 10-microns grooves. In contrast, E cells were not found inside the 22-microns-deep grooves and made contact only with the flat ridges between the grooves. As a general rule, fibroblasts (F) were oriented parallel to the long axis of the implants and produced a connective tissue capsule with 3-microns and 10-microns-deep grooved surfaces as well as smooth surfaces. On the 22-microns-deep grooved surfaces, however, F inserted obliquely into the implant. The attachment of F to the titanium surface was mediated by two zones; a thin (approximately 20 nm), amorphous, electron dense zone immediately contacting the titanium surface, and a fine fibrillar zone extending from the amorphous zone to the cell membrane. As oblique orientation of F has been associated with the inhibition of epithelial downgrowth, micromachined grooved surfaces of appropriate dimensions have the potential to improve the performance of percutaneous devices.     

Microfabricated grooves recapitulate neonatal myocyte connexin43 and N-cadherin expression and localization

Our objective is to alter the surface topography on which cardiac myocytes are grown in culture so that they more closely resemble their in vivo counterparts. Microtextured silicone substrata were made using photolithography and microfabrication techniques and then coated with laminin. Primary cardiac myocytes from newborn rats were plated on microgrooved and nontextured substrata. Myocytes were highly oriented on 5 microm grooves (69.8 +/- 2.0%) and significantly different, p < 0.0001, compared with randomly oriented cells grown on nontextured surfaces (2.9 +/- 0.95%; n = 19). Cells on shallower, 2 microm, grooves were slightly less well oriented (46.9 +/- 4.3%, n = 5, p < 0.001). The lateral spacings of the grooves were altered to examine changes in cell-to-cell contact by confocal immunocytochemistry and quantitative protein analysis. Connexin43 and N-cadherin were distributed around the perimeter of the myocytes plated on 10 x 5 x 5 microgrooved surfaces, similar to the localization found in the neonate. Connexin43 expression in cultures on 5 microm deep grooved substrata was equal to the neonatal heart, whereas it differed in nontextured surfaces. We conclude that it is necessary to combine groove depth (5 microm) and lateral ridge dimensions between grooves (5 microm) in order to recapitulate connexin43 and N-cadherin expression levels and subcellular localization to that of the neonate.     

Subcellular imaging of epithelium with time-lapse optical coherence tomography

We present the first experimental result of direct delineation of the nuclei of living rat bladder epithelium with ultrahigh-resolution optical coherence tomography (uOCT). We demonstrate that the cellular details embedded in the speckle noise in a uOCT image can be uncovered by time-lapse frame averaging that takes advantage of the micromotion in living biological tissue. The uOCT measurement of the nuclear size (7.9+/-1.4 microm) closely matches the histological evaluation (7.2+/-0.8 microm). Unlike optical coherence microscopy (OCM), which requires a sophisticated high-NA microscopic objective, this approach uses a commercial-grade single achromatic lens (f/10 mm, NA/0.25) and provides a cross-sectional image over 0.6 mm of depth without focus tracking, thus holding great promise of endoscopic optical biopsy for diagnosis and grading of flat epithelial cancer such as carcinoma in situ in vivo.     

The influence of surface topography of a porous perfluoropolyether polymer on corneal epithelial tissue growth and adhesion

Design principles for corneal implants are challenging and include permeability which inherently involves pore openings on the polymer surface. These topographical cues can be significant to a successful clinical outcome where a stratified epithelium is needed over the device surface, such as with a corneal onlay or corneal repair material. The impact of polymer surface topography on the growth and adhesion of corneal epithelial tissue was assessed using porous perfluoropolyether membranes with a range of surface topography. Surfaces were characterised by AFM and XPS, and the permeability and water content of membranes was measured. Biological testing of membranes involved a 21-day in vitro tissue assay to evaluate migration, stratification and adhesion of corneal epithelium. Similar parameters were monitored in vivo by surgically implanting membranes into feline corneas for up to 5 months. Data showed optimal growth and adhesion of epithelial tissue in vitro when polymer surface features were below a 150 nm RMS value. Normal processes of tissue growth and adhesion were disrupted when RMS values approached 300 nm. Data from the in vivo study confirmed these findings. Together, outcomes demonstrated the importance of surface topography in the design of implantable devices that depend on functional epithelial cover.