Cellular interactions with hydrogels

An efficient means of evaluating potential biomaterials is to use the in vitro fibroblast cell culture model. However, the chemistry which influences cell adhesion on polymer substrates is poorly understood. The work in this thesis aims to rationalise several theories of current opinion and introduce new chemical techniques that may predict cellular behaviour. The keratoprosthesis is a typical example of the need to be able to manipulate cell adhesion of materials since both adhesive and non adhesive sections are needed for proper integration and optical function. Calcein AM/ethidium homodimer-1 and DAPI assays were carried out using 3T3 and EKl.BR cells. Poly(HEMA) was found to be the most cell adhesive hydrogel tested. The reactivity of monomers and the resulting sequence distribution were found to affect surface properties and this may explain the poor levels of cell adhesion seen on NVP/MMA copolymers. Surface free energy is shown to be dependent on the polar and non polar groups present along the backbone chain of the polymers. Dehydrated and hydrated contact angle measurements show the effect of rotation of surface groups around the backbone chain. This effect is most apparent on hydrogels containing methacrylic acid. Dynamic contact angle measurements confirm sequence distribution irregularities and demonstrate the mobility of surface groups. Incorporation of NVI or DEAEMA into the hydrogels does not affect the mobility of the surface groups despite their bulkiness. Foetal calf serum was used for the first time as a test solution in an attempt to mimic a biological environment during surface experiments. A Vroman effect may be present, and may involve different surface proteins for each material tested. This interdisciplinary study combines surface characterisation and biological testing to further the knowledge of the biomaterial/host interface. Surface chemistry techniques appear to be insufficiently sensitive to predict cellular behaviour. The degree of ionisation of hydrogels containing ionic groups depends on the nature of the functional groups as well as the concentration and this is an important parameter to consider when comparing charged materials.

Cellular responses to potential biomaterials

The aim of this study was to systematically investigate the factors considered to be responsible for anchorage-dependent cell behaviour to determine which, if any, of these factors exerts greater influence. An efficient means of doing so is the in vitro fibroblast cell culture model. The interaction of fibroblasts with novel substrata gives information about how a biological system reacts to a foreign material. The may ultimately lead to the development of improved biomaterials. This interdisciplinary study combines the elements of surface characterisation and biological testing to determine the nature of the biomaterial/host interface. Polarity and surface charge were found to have an important influence on fibroblast adhesion to hydrogel polymers, by virtue of their water-structuring effects. The same factors were found to affect cell adhesion on undegraded PHB-HV copolymers and their blends with polysaccharides. On degraded PHB-HV copolymers, the degradation process itself played the greatest role in influencing cell response. Increasing surface charge and mechanical instability in these polymers inhibited cell adhesion. Based on the observations of hydrogels and PHB-copolymers a novel material, gel-spun PHB was designed for use as a wound scaffold. In vitro tests using human and mammalian fibroblasts accentuated the importance of polarity and surface charge in determining cellular response. The overall view of cellular behaviour on a broad spectrum of materials highlighted the effects that polarity and surface charge have on water-structuring, and how this affects interfacial conversion. In degradable systems, mechanical stability also plays an inportant role in determining anchorage-dependent cell behaviour.

Delivery and stability of antisense oligonucleotide conjugates in vitro

The efficacy of antisense oligonucleotide (ODN) therapy is dependent on four major parameters: delivery to cells, intracellular stability and localisation and efficient action at the target site.The aim of this project was to study the delivery of ODNs to macrophages and to assess the stability of two ODN conjugates, in vitro. The first conjugate aimed to improve uptake of ODNs via mannose receptor mediated delivery, the second investigated the improved delivery of ODN conjugates via non-specific lipophilic interaction with the cell membrane. A mono-mannose phosphoramidite derivative was designed and synthesised and a mono-mannose ODN conjugate synthesised by standard phosphoramidite chemistry. Delivery of this conjugate was enhanced to RAW264.7 and J774 macrophage cell lines via a mechanism of receptor mediated endocytosis. The delivery of three lipophilic ODN conjugates, cholesterol (cholhex), 16-carbon alkyl chain (C16) and hexa-ethylene glycol (HEG) moieties and an unconjugated ODN were assessed in RAW264.7 macrophages. All three conjugates increased the lipophilicity of the ODN as assessed from partition coefficient data. Both the cholhex and unconjugated ODNs were found to have higher degrees of cellular association than the C16 and HEG conjugates. Cellular uptake studies implicated internalisation of these ODNs by an adsorptive endocytosis mechanism. Following endocytosis, ODNs must remain stable during their residence in endosomal/lysosomal compartments prior to exiting and exerting their biological action in either the cytosol or nucleus. Assessment of in vitro stability in a lysosomal extract revealed the cholhex conjugate and unconjugated ODNs to have a longer half-life than the C16 and HEG conjugated ODNs, highlighting the influence of conjugate moieties on lysosomal stability. The effects of base composition and length on stability in a lysosomal extract revealed the longest half-life for homo-cytidine ODNs and ODNs over 20 nucleotides in length. These studies suggest that the above conjugates can enhance cellular association and delivery of antisense ODNs to cultured macrophages. This may lead to their use in treating disorders such as HIV infection, which affects this cell type.

Biodegradable polymers for the sustained release of antisense nucleic acids

Antisense oligodeoxynucleotides can selectively inhibit individual gene expression provided they remain stable at the target site for a sufficient period of time. Thus, the efficacy of antisense oligodeoxynucleotides may be improved by employing a sustained release delivery system which would protect from degradation by nucleases whilst delivering the nucleic acid in a controlled manner to the site of action. Biodegradable polymer films and micro spheres were evaluated as delivery devices for the oligodeoxynucleotides and ribozymes. Polymers such as polylactide, polyglycolide, polyhydroxybutyrate and polyhydroxyvalerate were used due to their biocompatability and non toxic degradation products. Release profiles of antisense nucleic acids from films over 28 days was biphasic, characterised by an initial burst release during the first 48 hours followed by a more sustained release. Release from films of longer antisense nucleic acids was slower compared to shorter nucleic acids. Backbone type also affected release, although to a lesser extent than length. Total release of the nucleic acids is dependent upon polymer degradation, no degradation of the polymer films was evident over the 28 day period, due to the high molecular weight and crystallinity of the polymers required to make solvent cast films. Backbone length and type did not affect release from microspheres, release was generally faster than from films, due to the increased surface area, and low molecular weight polymers which showed signs of degradation over the release period, resulting in a triphasic release profile. An increase in release was observed when sphere size and polymer molecular weight were decreased. The polymer entrapped phosphodiester oligodeoxynucleotides and ribozymes had enhanced stability compared to free oligodeoxynucleotides and ribozymes when incubated in serum. The released nucleic acids were still capable of hybridising to their target sequence, indicating that the fabrication processes did not adversely effect the properties of the antisense nucleic acids. Oligodeoxynucleotides loaded in 2μm spheres had a 10 fold increase in macrophage association compared to free oligodeoxynucleotides. Fluorescent microscopy indicates that the polymer entrapped oligodeoxynucleotide is concentrated inside the cell, whereas free oligodeoxynucleotides are concentrated at the cell membrane. Biodegradable polymers can reduce the limitations of antisense therapy and thus offer a potential therapeutic advantage.

Sensitivity of staphylococcus epidermidis to chlorhexidine and associated resistance properties

Staphylococcus epidermidis are common Gram-positive bacteria and are responsible for a number of life-threatening nosocomial infections. Treatment of S. epidermidis infection is problematic because the organism is usually resistant to many antibiotics. The high degree of resistance of this organism to a range of antibiotics and disinfectants is widely known. The aims of this thesis were to investigate and evaluate the susceptibility of isolates of S. epidermidis from various infections to chlorhexidine (CHX) and to other disinfectants such as benzalkonium chloride (BKC), triclosan (TLN) and povidone-iodine (PI). In addition, the mechanisms of resistance of S. epidermidis to chlorhexidine (the original isolates and strains adapted to chlorhexidine by serial passage) were examined and co-resistance to clinically relevant antibiotics investigated. In 3 of the 11 S. epidermidis strains passaged in increasing concentrations of chlorhexidine, resistance to the disinfectant arose (16-fold). These strains were examined further, each showing stable chlorhexidine resistance. Co-resistance to other disinfectants such as BKC, TLN and PI and changes in cell surface hydrophobicity were observed. Increases in resistance were accompanied by an increase in the proportion of neutral lipids and phospholipids in the cell membrane. This increase was most marked in diphosphatidylglycerol. These observations suggest that some strains of S. epidermidis can become resistant to chlorhexidine and related disinfectants/antiseptics by continual exposure. The mechanisms of resistance appear to be related to changes in membrane lipid compositions.

Serodiagnosis of staphylococcal sepsis

The coagulase-negative staphylococci are the most frequent cause of sepsis associated with indwelling intravascular catheters. Current microbiological investigations to support the diagnosis of catheter-related sepsis (CRS) include the culture of blood and catheter tips, however positive results may reflect specimen contamination, or colonisation of the catheter rather than true sepsis. Previous serological approaches to assist in the diagnosis of CRS based on cellular staphylococcal antigens have been of limited value. In this current study, the serodiagnostic potential of an exocellular antigen produced by 7 strains of coagulase-negative staphylococci cultured in brain heart infusion broth was investigated. Antigenic material isolated by gel permeation from liquid culture was characterised by immunological techniques and chemical analysis. Characterisation of the exocellular antigen revealed a novel glycerophosphoglycolipid, termed lipid S. which shared antigenic determinants with lipoteichoic acid, but differed by comprising a glycerophosphate chain length of only 6 units. In addition, lipid S was immunologically distinct from diphosphatidyl glycerol, a constituent cell membrane phospho lipid. An indirect enzyme linked immunosorbent assay (ELISA) based on lipid S was subsequently developed and used to determine serum antibody levels (IgM and IgG) in 67 patients with CRS due to staphylococci, and 67 patients with a central venous catheter (CVC) in situ who exhibited no evidence of sepsis. The sensitivity and specificity of the lipid S IgG ELISA was 75% and 90% respectively whilst the IgM assay had sensitivity and specificity of 52% and 85%. The addition of GullSORereagent to the EL1SA procedure to remove competing serum IgG and rheumatoid factor did not significantly improve the performance of the IgM assay. The serological response in serial serum samples of 13 patients with CRS due to staphylococci was investigated. Elevated levels of antibody were detected at an early stage of infection, prior to the isolation of microorganisms by standard culture methods, and before the clinical presentation of sepsis in 3 patients. The lipid S ELISA was later optimised and a rapid 4-hour assay developed for the serodiagnosis of CRS. Serum IgG levels were determined in 40 patients with CRS due to staphylococci and 40 patients with a CVC in situ who exhibited no evidence of sepsis. The sensitivity and specificity of the rapid IgG assay was 70% and 100% respectively. Elevated serum antibody levels in patients with endocarditis, prosthetic joint infection and pyogenic spondylodiscitis due to Gram-positive cocci were also detected with the lipid S ELISA suggesting that the assay may facilitate the diagnosis of these infections. Unexpected increased levels of anti-lipid S IgG in 31% of control patients with sciatica suggested a possible microbial aetiology of this condition. Further investigation of some of these patients by culture of microdiscectomy tissue removed at operation, revealed the presence of low-virulent microorganisms in 37% of patients of which Propionibacterium aeries accounted for 85% of the positive culture isolates. The results suggested a previously unrecognised association between P. acnes and sciatica, which may have implications for the future management of the condition.

Lipophilic antifolates as potential antipsoriatic agents

The lipophilic dihydrofolate reductase (DHFR) inhibitor m-azidopyrimethamine (MZP) was investigated for suitability for development as a topical antipsoriatic agent. The clinical features and treatments for psoriasis were reviewed. High performance liquid chromatography (HPLC) was employed as the main analytical method, with UV spectroscopy being used in some cases. Reduction of the azido-group was proposed as a potential detoxification mechanism for MZP. The rates of reduction of a series of substituted phenyl azide compounds by dithiothreitol were investigated and found to depend on the substitution pattern of the aryl azide molecular, with electron deficient azides exhibiting faster rates of reduction in the system studied. The rates of reduction of MZP and analogous compounds were also studied using this model. The skin penetration of MZP was assessed using an in vitro hairless mouse skin model. The rate of permeation (flux) of MZP across hairless mouse skin was found to be dependent on the quantity of propylene glycol used as cosolvent in the vehicle and the pH. The use of a pretreatment regime of oleic acid in propylene glycol was shown to greatly increase the penetration of MZP through the hairless mouse skin as compared to application without pretreatment, or pretreatment with other penetration enhancers. The metabolism of MZP was studied in in vitro models comprising skin homogenates, SV-K14 human keratinocyte cell cultures and skin commensal bacterial cultures. No conversion of MZP to the corresponding amine was detected in any of the models. The growth inhibitory properties of MZP were investigated in an in vitro SV-K14 human keratinocyte cell culture model and compared with those of other DHFR inhibitors. [14C]-pyrimethamine was shown to be taken up by the SV-K14 keratinocytes.

Rapid aquaporin translocation regulates cellular water flow:the mechanism of hypotonicity-induced sub-cellular localization of the aquaporin 1 water channel

The control of cellular water flow is mediated by the aquaporin (AQP) family of membrane proteins. The family's structural features and the mechanism of selective water passage through the AQP pore are established, but there remains a gap in our knowledge of how water transport is regulated. Two broad possibilities exist. One is controlling the passage of water through the AQP pore, but this has only been observed as a phenomenon in some plant and microbial AQPs. An alternative is controlling the number of AQPs in the cell membrane. Here we describe a novel pathway in mammalian cells whereby a hypotonic stimulus directly induces intracellular calcium elevations, through transient receptor potential channels, that trigger AQP1 translocation. This translocation, which has a direct role in cell volume regulation, occurs within 30s and is dependent on calmodulin activation and phosphorylation of AQP1 at two threonine residues by protein kinase C. This direct mechanism provides a rationale for the changes in water transport that are required in response to constantly-changing local cellular water availability. Moreover, since calcium is a pluripotent and ubiquitous second messenger in biological systems, the discovery of its role in the regulation of AQP translocation has ramifications for diverse physiological and pathophysiological processes, as well as providing an explanation for the rapid regulation of water flow that is necessary for cell homeostasis.

GP130/OSMR is the only LIF/IL-6 family receptor complex to promote osteoblast differentiation of calvaria progenitors

Leukemia inhibitory factor (LIF) and its receptor (LIFR) are "twins" of Oncostatin M (OSM) and OSMR, respectively, likely having arisen through gene duplications. We compared their effects in a bone nodule-forming model of in vitro osteogenesis, rat calvaria (RC) cell cultures. Using a dominant-negative LIF mutant (hLIF-05), we showed that in RC cell cultures mouse OSM (mOSM) activates exclusively glycoprotein 130 (gp130)/OSMR. In treatments starting at early nodule formation stage, LIF, mOSM, IL-11, and IL-6 + sIL-6R inhibit bone nodule formation, that is, osteoprogenitor differentiation. Treatment with mOSM, and no other cytokine of the family, in early cultures (day 1-3 or 1-4) increases bone colony numbers. hLIF-05 also dose dependently stimulates bone nodule formation, confirming the inhibitory action of gp130/LIFR on osteogenesis. In pulse treatments at successive stages of bone nodule formation and maturation, LIF blocks osteocalcin (OCN) expression by differentiated osteoblasts, but has no effect on bonesialoprotein (BSP) expression. Mouse OSM inhibits OCN and BSP expression in preconfluent cultures with no or progressively reduced effects at later stages, reflecting the disruption of early nodules, possibly due to the strong apoptotic action of mOSM in RC cell cultures. In summary, LIFR and OSMR display differential effects on differentiation and phenotypic expression of osteogenic cells, most likely through different signal transduction pathways. In particular, gp130/OSMR is the only receptor complex of the family to stimulate osteoprogenitor differentiation in the RC cell culture model. © 2005 Wiley-Liss, Inc.

Detection of phosphatidylserine with a modified polar head group in human keratinocytes exposed to the radical generator AAPH

Phosphatidylserine (PS) is preferentially located in the inner leaflet of the cell membrane, and translocation of PS oxidized in fatty acyl chains to the outside of membrane has been reported as signaling to macrophage receptors to clear apoptotic cells. It was recently shown that PS can be oxidized in serine moiety of polar head-group. In the present work, a targeted lipidomic approach was applied to detecting OxPS modified at the polar head-group in keratinocytes that were exposed to the radical generator AAPH. Glycerophosphoacetic acid derivatives (GPAA) were found to be the major oxidation products of OxPS modified at the polar head-group during oxidation induced by AAPH-generated radicals, similarly to previous observations for the oxidation induced by OH radical. The neutral loss scan of 58Da and a novel precursor ion scan of m/z 137.1 (HOPO3CH2COOH) allowed the recognition of GPAA derivatives in the total lipid extracts obtained from HaCaT cells treated with AAPH. The positive identification of serine head group oxidation products in cells under controlled oxidative conditions opens new perspectives and justifies further studies in other cellular environments in order to understand fully the role of PS polar head-group oxidation in cell homeostasis and disease.

Adipose atrophy in cancer cachexia:morphologic and molecular analysis of adipose tissue in tumour-bearing mice

Extensive loss of adipose tissue is a hallmark of cancer cachexia but the cellular and molecular basis remains unclear. This study has examined morphologic and molecular characteristics of white adipose tissue in mice bearing a cachexia-inducing tumour, MAC16. Adipose tissue from tumour-bearing mice contained shrunken adipocytes that were heterogeneous in size. Increased fibrosis was evident by strong collagen-fibril staining in the tissue matrix. Ultrastructure of 'slimmed' adipocytes revealed severe delipidation and modifications in cell membrane conformation. There were major reductions in mRNA levels of adipogenic transcription factors including CCAAT/enhancer binding protein alpha (C/EBPα), CCAAT/enhancer binding protein beta, peroxisome proliferator-activated receptor gamma, and sterol regulatory element binding protein-1c (SREBP-1c) in adipose tissue, which was accompanied by reduced protein content of C/EBPα and SREBP-1. mRNA levels of SREBP-1c targets, fatty acid synthase, acetyl CoA carboxylase, stearoyl CoA desaturase 1 and glycerol-3-phosphate acyl transferase, also fell as did glucose transporter-4 and leptin. In contrast, mRNA levels of peroxisome proliferators-activated receptor gamma coactivator-1alpha and uncoupling protein-2 were increased in white fat of tumour-bearing mice. These results suggest that the tumour-induced impairment in the formation and lipid storing capacity of adipose tissue occurs in mice with cancer cachexia. © 2006 Cancer Research UK.

The interaction of sodium chlorite with phospholipids and glutathione:a comparison of effects in vitro, in mammalian and in microbial cells

In this study the interaction of the preservative sodium chlorite with unsaturated lipids and glutathione was investigated, in comparison with peroxides, sodium hypochlorite, and benzalkonium chloride. The aim was to determine whether the action of sodium chlorite could involve membrane lipid damage or antioxidant depletion, and how this related to toxicity in both mammalian and microbial cells. The treatment of phospholipids with chlorite yielded low levels of hydroperoxides, but sodium chlorite oxidized the thiol-containing antioxidant glutathione to its disulfide form very readily in vitro, with a 1:4 oxidant:GSH stoichiometry. In cultured cells, sodium chlorite also caused a substantial depletion of intracellular glutathione, whereas lipid oxidation was not very prominent. Sodium chlorite had a lower toxicity to ocular mammalian cells than benzalkonium chloride, which could be responsible for the different effects of long-term application in the eye. The fungal cells, which were most resistant to sodium chlorite, maintained higher percentage levels of intracellular glutathione during treatment than the mammalian cells. The results show that sodium chlorite can cause oxidative stress in cells, and suggest that cell damage is more likely to be due to interaction with thiol compounds than with cell membrane lipids. The study also provides important information about the differential resistance of ocular cells and microbes to various preservatives and oxidants.

Receptor for activated C Kinase 1 protein binds to and activates the human estrogen receptor α

The classical concept of estrogen receptor (ER) activation is that steroid passes the cell membrane, binds to its specific protein receptor in the cell's cytoplasm and the steroid-receptor complex travels to the nucleus where it activates responsive genes. This basic idea has been challenged by results of experiments demonstrating insulin-like growth factor 1 (IGF-1) activation of the ER in the complete absence of estrogen suggesting at least one other mechanism of ER activation not involving steroid. One explanation is that activation of the cell surface IGF-1 receptor leads to synthesis of an intracellular protein(s) able to bind to and stimulate the ER. Based on results using the two-hybrid system, coimmunoprecipitation and transfection-luciferase assays, we herein show that one of these proteins could well be receptor for activated C kinase 1 (RACK-1). Using the human ER type α (ER-α) as bait, a cloned complementary deoxyribonucleic acid (cDNA) library from IGF-1 treated human breast cancer MCF-7 cells was screened for ER-α - protein interactions. Many positive clones were obtained which contained the RACK-1 cDNA sequence. Coimmunoprecipitation of in-vitro translation products of the ER-α and RACK-1 confirmed the interaction between the two proteins. Transfection studies using the estrogen response element spliced to a luciferase reporter gene revealed that constitutive RACK-1 expression was able to powerfully stimulate ER-α activity under estrogen-free conditions. This effect could be enhanced by 17β-estradiol (E2) and blocked by tamoxifen, an E2 antagonist. These results show that RACK-1 is able to activate the ER-α in the absence of E2, although together with the latter, enhanced effects occur. Since RACK-1 gene expression is stimulated by IGF-1, it is distinctly possible that RACK-1 is the mediator of the stimulatory effects of IGF-1 on ER-α. © 2014 JMS.

Characterization of microvesicles released from human red blood cells

Background/Aims: Extracellular vesicles (EVs) are spherical fragments of cell membrane released from various cell types under physiological as well as pathological conditions. Based on their size and origin, EVs are classified as exosome, microvesicles (MVs) and apoptotic bodies. Recently, the release of MVs from human red blood cells (RBCs) under different conditions has been reported. MVs are released by outward budding and fission of the plasma membrane. However, the outward budding process itself, the release of MVs and the physical properties of these MVs have not been well investigated. The aim of this study is to investigate the formation process, isolation and characterization of MVs released from RBCs under conditions of stimulating Ca2+ uptake and activation of protein kinase C. Methods: Experiments were performed based on single cell fluorescence imaging, fluorescence activated cell sorter/flow cytometer (FACS), scanning electron microscopy (SEM), atomic force microscopy (AFM) and dynamic light scattering (DLS). The released MVs were collected by differential centrifugation and characterized in both their size and zeta potential. Results: Treatment of RBCs with 4-bromo-A23187 (positive control), lysophosphatidic acid (LPA), or phorbol-12 myristate-13 acetate (PMA) in the presence of 2 mM extracellular Ca2+ led to an alteration of cell volume and cell morphology. In stimulated RBCs, exposure of phosphatidylserine (PS) and formation of MVs were observed by using annexin V-FITC. The shedding of MVs was also observed in the case of PMA treatment in the absence of Ca2+, especially under the transmitted bright field illumination. By using SEM, AFM and DLS the morphology and size of stimulated RBCs, MVs were characterized. The sizes of the two populations of MVs were 205.8 ± 51.4 nm and 125.6 ± 31.4 nm, respectively. Adhesion of stimulated RBCs and MVs was observed. The zeta potential of MVs was determined in the range from - 40 mV to - 10 mV depended on the solutions and buffers used. Conclusion: An increase of intracellular Ca2+ or an activation of protein kinase C leads to the formation and release of MVs in human RBCs.

A co-cultured skin model based on cell support membranes

Tissue engineering of skin based on collagen:PCL biocomposites using a designed co-culture system is reported. The collagen:PCL biocomposites having collagen:PCL (w/w) ratios of 1:4, 1:8, and 1:20 have been proven to be biocompatible materials to support both adult normal human epidermal Keratinocyte (NHEK) and mouse 3T3 fibroblast growth in cell culture, respectively, by Dai, Coombes, et al. in 2004. Films of collagen:PCL biocomposites were prepared using non-crosslinking method by impregnation of lyophilized collagen mats with PCL/dichloromethane solutions followed by solvent evaporation. To mimic the dermal/epidermal structure of skin, the 1:20 collagen:PCL biocomposites were selected for a feasibility study of a designed co-culture technique that would subsequently be used for preparing fibroblast/biocomposite/keratinocyte skin models. A 55.3% increase in cell number was measured in the designed co-culture system when fibroblasts were seeded on both sides of a biocomposite film compared with cell culture on one surface of the biocomposite in the feasibility study. The co-culture of human keratinocytes and 3T3 fibroblasts on each side of the membrane was therefore studied using the same co-culture system by growing keratinocytes on the top surface of membrane for 3 days and 3T3 fibroblasts underneath the membrane for 6 days. Scanning electron microscopy (SEM) and immunohistochemistry assay revealed good cell attachment and proliferation of both human keratinocytes and 3T3 fibroblasts with these two types of cells isolated well on each side of the membrane. Using a modified co-culture technique, a co-cultured skin model presenting a confluent epidermal sheet on one side of the biocomposite film and fibroblasts populated on the other side of the film was developed successfully in co-culture system for 28 days under investigations by SEM and immunohistochemistry assay. Thus, the design of a co-culture system based on 1:20 (w/w) collagen:PCL biocomposite membranes for preparation of a bi-layered skin model with differentiated epidermal sheet was proven in principle. The approach to skin modeling reported here may find application in tissue engineering and screening of new pharmaceuticals. © 2005 Elsevier Inc. All rights reserved.