Neural differentiation regulated by biomimetic surfaces presenting motifs of extracellular matrix proteins

The interaction between cells and the extracellular matrix (ECM) is essential during development. To elucidate the function of ECM proteins on cell differentiation, we developed biomimetic surfaces that display specific ECM peptide motifs in a controlled manner. Presentation of ECM domains for collagen, fibronectin, and laminin influenced the formation of neurites by differentiating PC12 cells. The effect of these peptide sequences was also tested on the development of adult neural stem/progenitor cells. In this system, collagen I and fibronectin induced the formation of beta-III-tubulin positive cells, whereas collagen IV reduced such differentiation. Biomimetic surfaces composed of multiple peptide types enabled the combinatorial effects of various ECM motifs to be studied. Surfaces displaying combined motifs were often predictable as a result of the synergistic effects of ECM peptides studied in isolation. For example, the additive effects of fibronectin and laminin resulted in greater expression of beta-III-tubulin positive cells, whereas the negative effect of the collagen IV domain was canceled out by coexpression of collagen I. However, simultaneous expression of certain ECM domains was less predictable. These data highlight the complexity of the cellular response to combined ECM signals and the need to study the function of ECM domains individually and in combination.

Transplantation of porous tubes following spinal cord transection improves hindlimb function in the rat

STUDY DESIGN

Experimental.

OBJECTIVE

To determine the effects of a porous tube transplant in spinal cord transected rats.

SETTING

Acadia University, Wolfville, Nova Scotia, Canada.

METHODS

Female rats were randomly assigned to three experimental groups: control (Con, n=8), spinal cord transected (Tx, n=5) and spinal cord transected with transplant (TxTp, n=7). The rats in the TxTp and Tx groups received a complete spinal cord transection at the T10 level and the TxTp group immediately received a porous tube transplant.

RESULTS

Locomotor activity rated on the Basso, Beattie, Bresnahan scale improved significantly in the TxTp animals over the 4 weeks such that final scores were 21, 1.4 and 7.1 for the Con, Tx and TxTp groups, respectively. As expected, the muscle to body mass ratios of the hindlimb skeletal muscles of the Tx group were decreased (soleus 35%, plantaris 29% and gastrocnemius 29%) and this was also observed in the TxTp group (soleus 33%, plantaris 23% and gastrocnemius 30%). Cytochrome c oxidase (CYTOX) activity in the plantaris was decreased by Tx but maintained in the TxTp group (Con=82.2, Tx=44.8 and TxTp=72.8 U/min/g).

CONCLUSION

Four weeks after the spinal cord transection, plantaris CYTOX activity and locomotor function improved with porous tube implantation.

SPONSORSHIP

Natural Sciences and Engineering Research Council.

Use of a biomimetic strategy to engineer bone

Engineering trabecular-like, three-dimensional bone tissue throughout biodegradable polymer scaffolds is a significant challenge. Using a novel processing technique, we have created a biodegradable scaffold with geometry similar to that of trabecular bone. When seeded with bone-marrow cells, new bone tissue, the geometry of which reflected that of the scaffold, was evident throughout the scaffold volume and to a depth of 10 mm. Preseeded scaffolds implanted in non-healing rabbit segmental bone defects allowed new functional bone formation and bony union to be achieved throughout the defects within 8 weeks. This marks the first report of successful three-dimensional bone-tissue engineering repair using autologous marrow cells without the use of supplementary growth factors. We attribute our success to the novel scaffold morphology.

Photoimmobilization of biomolecules within a 3-dimensional hydrogel matrix

It has been recognized that a three-dimensional cell invasive scaffold that provides both topographical and chemical cues is desirable in regenerative tissue engineering to encourage cell attachment, migration, regrowth and ultimately tissue repair. Carbohydrate hydrogels are attractive for such applications because they are generally biocompatible and able to match the mechanical properties of most soft tissues. Although carbohydrate hydrogels have been previously modified with cell adhesive peptides and proteins, complicated hydrogel matrix activation was required prior to biomolecule coupling and, perhaps more importantly, the overall immobilization yield was low at approximately 1%. In this study, we report the photo-immobilization of a model biomolecule, ovalbumin (OVA), to agarose gel. We describe two methods of modification where the photoactive moiety is coupled to either the protein (i.e. OVA) or the matrix (i.e. agarose) prior to immobilization. We found that the photo-immobilization yield depends on the location of the photoactive moiety. Using photoactive OVA, 1.8% of the OVA initially incorporated into the agarose gel is immobilized; using photoactive agarose, 9.3% of the OVA initially mixed with the agarose is immobilized. The latter is a significant improvement over previous yields and may be useful in attaining our goal of immobilizing a biomolecule gradient for guided tissue regeneration.

Investigating the synergistic effect of combined neurotrophic factor concentration gradients to guide axonal growth

Neurotrophic factors direct axonal growth toward the target tissue by a concentration gradient, which is mediated through different tyrosine kinase cell surface receptors. In this study, well-defined concentration gradients of neurotrophic factors (NFs) allowed us to study the synergistic effect of different NFs (e.g. nerve growth factor [NGF], neurotrophin-3 [NT-3] and brain-derived neurotrophic factor [BDNF]) for axonal guidance of embryonic lumbar dorsal root ganglion cells (DRGs). Effective guidance of DRG axons was achieved with a minimum NGF concentration gradient of 133 ng/ml/mm alone, or combined NGF and NT-3 concentration gradients of 80 ng/ml/mm each. Interestingly, the combined concentration gradients of NGF and BDNF did not show any significant synergism at the concentration gradients studied. The synergism observed between NGF and NT-3 indicates that axons may be guided over a 12.5 mm distance, which is significantly greater than that of 7.5 mm calculated by us for NGF alone or that of 2 mm observed by others.

Creating porous tubes by centrifugal forces for soft tissue application

Chemically crosslinked poly(2-hydroxyethyl methacrylate) (PHEMA) tubes were synthesized by applying centrifugal forces to propagating polymer chains in solution. Initiated monomer solutions, with a composition typical for PHEMA sponges, were placed into a cylindrical mold that was rotated about its long axis. As polymerization proceeded, phase separated PHEMA formed a sediment at the periphery under centrifugal action. The solvent remained in the center of the mold while the PHEMA phase gelled, resulting in a tube. By controlling the rotational speed and the formulation chemistry (i.e., monomer, initiator and crosslinking agent concentrations), the tube dimensions and wall morphology were manipulated. Tube manufacture was limited by a critical casting concentration [M]c, above which only rods formed. All tubes had an outer diameter of 2.4 mm, reflecting the internal diameter of the mold and a wall thickness of approximately 40-400 microm. Wall morphologies varied from interconnecting polymer and water phases to a closed cell, gel-like, structure. Concentric tubes were successfully prepared by using formulations that enhanced phase separation over gelation/network formation. This was achieved by using formulations with lower concentrations of monomer and crosslinking agent and higher concentrations of initiator. This technique offers a new approach to the synthesis of polymeric tubes for use in soft tissue applications, such as nerve guidance channels.

Defining the concentration gradient of nerve growth factor for guided neurite outgrowth

The developing axon is believed to navigate towards its target tissue in response to a concentration gradient of neurotrophic factors, among other diffusible and surface-bound stimuli. However, the minimum concentration gradient required for guidance over the maximum distance is still unknown, largely because well-defined systems have not been utilized to address this question. In this study, a linear concentration gradient of nerve growth factor was achieved across a 5-mm agarose membrane that separated a nerve growth factor source compartment from a sink compartment. The concentrations in both compartments were maintained constant (and different). Both concentration and concentration gradient were well defined across the membrane, allowing us to study the relative importance of concentration gradient vs concentration for neurite guidance. The orientation of PC12 cell neurites was studied in response to a series of nerve growth factor concentration gradients in vitro. For effective guidance of PC12 cell neurite outgrowth, a minimum concentration gradient of 133ng/ml per mm was required, below which guidance was ineffective. Higher gradients were effective for guidance yet were limited by the concentration of nerve growth factor in the source compartment. At a nerve growth factor concentration of 995ng/ml, the PC12 cells' receptors were saturated, thereby limiting the maximum effective distance for guidance to less than 7.5mm in response to a diffusible nerve growth factor cue. This distance exceeds the 0.5-2mm distance observed by others for effective neurite guidance. Using this model system, we propose that the minimum concentration gradient can be defined for other cells and growth factors. Ultimately, it is anticipated that such concentration gradients could be included in a device to promote regeneration.

Enhancing the neuronal interaction on fluoropolymer surfaces with mixed peptides or spacer group linkers

Embryonic hippocampal neurons cultured on surface modified fluoropolymers showed enhanced interaction and neurite extension. Poly(tetrafluoroethylene-co-hexafluoropropylene) (FEP) film surfaces were aminated by reaction with a UV-activated mercury ammonia system yielding FEP-[N/O]. Laminin-derived cell-adhesive peptides (YIGSR and IKVAV) were coupled to FEP surface functional groups using tresyl chloride activation. Embryonic (E18) hippocampal neurons were cultured in serum-free medium for up to 1 week on FEP film surfaces that were modified with either one or both of GYIGSR and SIKVAV or GGGGGGYIGSR and compared to control surfaces of FEP-[N/O] and poly(L-lysine)/laminin-coated tissue culture polystyrene. Neuron-surface interactions were analyzed over time in terms of neurite outgrowth (number and length of neurites), cell adhesion and viability. Neurite outgrowth and adhesion were significantly better on peptide-modified surfaces than on either FEP or FEP-[N/O]. Cells on the mixed peptide (GYIGSR/SIKVAV) and the spacer group peptide (GGGGGGYIGSR) surfaces demonstrated similar behavior to those on the positive PLL/laminin control. The specificity of the cell-peptide interaction was demonstrated with a competitive assay where dissociated neurons were incubated in media containing peptides prior to plating. Cell adhesion and neurite outgrowth diminished on all surfaces when hippocampal neurons were pre-incubated with dissolved peptides prior to plating.

Optimizing the sterilization of PLGA scaffolds for use in tissue engineering

There are few suitable techniques available to sterilize biodegradable polyester three-dimensional tissue engineering scaffolds because they are susceptible to degradation and/or morphological degeneration by high temperature and pressure. We used a novel polyllactide-co-glycolide) scaffold (Osteofoam) to determine the optimal sterilization procedure--i.e. a sterile product with minimal degradation and deformation. Initial studies, found that an argon plasma created at 100W for 4min was optimal for sterilizing Osteofoam scaffolds without affecting their morphology. The RFGD plasma sterilization method was compared to two well-established techniques--ethylene oxide (ETO) and gamma-irradiation (gamma)--which were in turn compared to disinfection in 70% ethanol. Disinfection in 70% ethanol serves as a useful control because it affects neither the morphology nor the molecular weight of the polymer: yet, ethanol is unsuitable as a sterilization method because it does not adequately eliminate hydrophilic viruses and bacterial spores. The three sterilization techniques, ETO, gamma and RFGD plasma, were compared in terms of their immediate and long-term effects on the dimensions, morphology, molecular weight and degradation profile of the scaffolds. Scaffolds shrank to 60% of their initial volume after ETO sterilization whereas their molecular weight (Mw) decreased by approximately 50% after gamma-irradiation. Thus, both ETO and gamma-irradiation posed immediate problems as sterilization techniques for 3-D biodegradable polyester scaffolds. During the in vitro degradation study, all sterilized samples showed advanced morphological and volume changes over time relative to ethanol (EtOH) disinfected samples, with the greatest changes observed for gamma-irradiated samples. ETO, RFGD plasma sterilized and EtOH disinfected samples showed similar changes in Mw and mass over the 8-week time frame. Overall, of the three sterilization techniques studied, RFGD plasma was the best.

Engineering three-dimensional bone tissue in vitro using biodegradable scaffolds: investigating initial cell-seeding density and culture period

New three-dimensional (3D) scaffolds for bone tissue engineering have been developed throughout which bone cells grow, differentiate, and produce mineralized matrix. In this study, the percentage of cells anchoring to our polymer scaffolds as a function of initial cell seeding density was established; we then investigated bone tissue formation throughout our scaffolds as a function of initial cell seeding density and time in culture. Initial cell seeding densities ranging from 0.5 to 10 x 10(6) cells/cm(3) were seeded onto 3D scaffolds. After 1 h in culture, we determined that 25% of initial seeded cells had adhered to the scaffolds in static culture conditions. The cell-seeded scaffolds remained in culture for 3 and 6 weeks, to investigate the effect of initial cell seeding density on bone tissue formation in vitro. Further cultures using 1 x 10(6) cells/cm(3) were maintained for 1 h and 1, 2, 4, and 6 weeks to study bone tissue formation as a function of culture period. After 3 and 6 weeks in culture, scaffolds seeded with 1 x 10(6) cells/cm(3) showed similar tissue formation as those seeded with higher initial cell seeding densities. When initial cell seeding densities of 1 x 10(6) cells/cm(3) were used, osteocalcin immunolabeling indicative of osteoblast differentiation was seen throughout the scaffolds after only 2 weeks of culture. Von Kossa and tetracycline labeling, indicative of mineralization, occurred after 3 weeks. These results demonstrated that differentiated bone tissue was formed throughout 3D scaffolds after 2 weeks in culture using an optimized initial cell density, whereas mineralization of the tissue only occurred after 3 weeks. Furthermore, after 6 weeks in culture, newly formed bone tissue had replaced degrading polymer.

Cell delivery to the central nervous system

A dysfunctional central nervous system (CNS) resulting from neurological disorders and diseases impacts all of humanity. The outcome presents a staggering health care issue with a tremendous potential for developing interventive therapies. The delivery of therapeutic molecules to the CNS has been hampered by the presence of the blood-brain barrier (BBB). To circumvent this barrier, putative therapeutic molecules have been delivered to the CNS by such methods as pumps/osmotic pumps, osmotic opening of the BBB, sustained polymer release systems and cell delivery via site-specific transplantation of cells. This review presents an overview of some of the CNS delivery technologies with special emphasis on transplantation of cells with and without the use of polymer encapsulation technology.

Patterned poly(chlorotrifluoroethylene) guides primary nerve cell adhesion and neurite outgrowth

Central nervous system (CNS) neurons, unlike those of the peripheral nervous system, do not spontaneously regenerate following injury. Recently it has been shown that in the developing CNS, a combination of cell-adhesive and cell-repulsive cues guide growing axons to their targets. We hypothesized that by mimicking these guidance signals, we could guide nerve cell adhesion and neurite outgrowth in vitro. Our objective was to direct primary nerve cell adhesion and neurite outgrowth on poly(chlorotrifluoroethylene) (PCTFE) surfaces by incorporating alternating patterns of cell-adhesive (peptide) and nonadhesive (polyethylene glycol; PEG) regions. PCTFE was surface-modified with lithium PEG-alkoxide, demonstrating the first report of metal-halogen exchange with an alkoxide and PCTFE. Titanium and then gold were sputtered onto PEG-modified films, using a shadow-masking technique that creates alternating patterns on the micrometer scale. PCTFE-Au regions then were modified with one of two cysteine-terminated laminin-derived peptides, C-GYIGSR or C-SIKVAV. Hippocampal neuron cell-surface interactions on homogeneously modified surfaces showed that neuron adhesion was decreased significantly on PEG-modified surfaces and was increased significantly on peptide-modified surfaces. Cell adhesion was greatest on CGYIGSR surfaces while neurite length was greatest on CSIKVAV surfaces and PLL/laminin positive controls, indicating the promise of peptides for enhanced cellular interactions. On patterned surfaces, hippocampal neurons adhered and extended neurites preferentially on peptide regions. By incorporating PEG and peptide molecules on the surface, we were able to simultaneously mimic cell-repulsive and cell-adhesive cues, respectively, and maintain the biopatterning of primary CNS neurons for over 1 week in culture.

In vitro degradation of a novel poly(lactide-co-glycolide) 75/25 foam

Macroporous poly(lactide-co-glycolide) PLGA 75/25 foams were prepared for application in bone tissue engineering. Their in vitro degradation behaviour was followed over a 30 week period at 37 degrees C and at one of three pHs: (1) pH 5.0, which mimics the acidic environment produced by activated macrophages, (2) pH 7.4, which reproduces normal physiological conditions and (3) an intermediate pH 6.4. The degradation of the PLGA 75/25 foams was studied by measuring changes in mass, molecular weight and morphology. The degradation profile of foams maintained at pH 5.0, 6.4 and 7.4 was similar until week 16, after which foams maintained at pH 6.4 and 7.4 had comparable degradation patterns whereas foams maintained at pH 5.0 degraded faster. For example, mass loss was less than 3% for foams maintained at all three pHs until week 16; however, by week 30, foams maintained at pH 6.4 and 7.4 had lost 30% of their mass whereas foams maintained at pH 5.0 had lost 90% of their mass. Foams maintained at pH 6.4 and 7.4 showed a similar constant decrease in molecular weight over the entire degradation study. Foams maintained at pH 5.0 had a similar rate of molecular weight loss as those maintained at pH 6.4 and 7.4 until week 16, after which the rate of molecular weight loss of foams maintained at pH 5.0 was accelerated. The morphology of the foams maintained at pH 6.4 and 7.4 was unchanged for 25 weeks. Foams maintained at pH 5.0 collapsed after week 18. Thus the PLGA 75/25 foams, described herein, maintained their 3-D morphology at physiological pH for over 6 months, which is an important feature for tissue engineering applications.

Patterned glass surfaces direct cell adhesion and process outgrowth of primary neurons of the central nervous system

Glass surfaces were patterned with cell-adhesive regions of laminin adhesive peptides YIGSR, RGD, and IKVAV, and cell-repulsive regions of poly(ethylene glycol) (PEG). The patterns were created by sputter-coating titanium and then gold onto glass coverslips through electron microscope grids. Gold surfaces were modified with cysteine-terminated peptides to have approximately 450 fmol/ cm2 of peptide incorporated on the glass coverslips as determined with radiolabeled CGYIGSR. Amine-functionalized glass coverslips were prepared using an amine-functionalized silane and then further modified with PEG-aldehyde by a Schiff base reduction. All surfaces were characterized by X-ray photoelectron spectroscopy and water contact angles. Hippocampal neurons, plated from a serum-free medium, adhered preferentially to peptide-functionalized surfaces over PEG-modified surfaces. Cell adhesion and neurite outgrowth were limited to the peptide region, demonstrating that neurite outgrowth could be directed by a combination of cell-adhesive and cell-repulsive cues.

Peptide surface modification of poly(tetrafluoroethylene-co-hexafluoropropylene) enhances its interaction with central nervous system neurons

Poly(tetrafluoroethylene-co-hexafluoropropylene) (FEP) film surfaces were chemically surface modified to introduce one of three laminin adhesive peptides: GYIGSR, GRGDS, or SIKVAV. FEP film surfaces were first reduced with sodium naphthalide to introduce surface carbon-carbon double bonds at two reaction conditions: 20 min at -78 degrees C, and 3 h at 25 degrees C. Scanning electron microscopy and atomic force microscopy indicated that surface topography was unaffected by the reaction conditions. Reduced FEP film surfaces were further modified to introduce hydroxyl groups via hydroboration/oxidation or carboxylic acid groups via oxidation. The hydroxyl (FEP-CHxOH) and carboxylic acid (FEP-COOH) functionalized surfaces provided reactive handles for peptide coupling using tresyl chloride. Surface elemental composition data, determined from X-ray protoelectron spectroscopy, indicated that equivalent amounts of GYIGSR, GRGDS, and SIKVAV were introduced. Two additional coupling reagents, SMCC and TSU, were compared to tresyl chloride for the coupling of radio-labeled tyrosine of GYIGSR. Between 8 and 150 fmol/cm2 of peptide was introduced to the hydroxyl and carboxylic acid functionalized surfaces, with the tresyl coupling reagent showing the greatest amount of peptide incorporated. The tresyl-coupled peptide-modified surfaces were compared in terms of the response of primary, embryonic hippocampal neurons plated from serum-free medium for 4 days. The number and length of neurites extending from the cell bodies were averaged over 50 cells after 1 and 4 days FEP-CHxO-peptide surfaces had either a greater or equivalent hippocampal neuron interaction than the corresponding FEP-COO-peptide surfaces. All peptide-functionalized surfaces had a greater hippocampal neuron interaction than the corresponding FEP-CHxOH, FEP-COOH, and FEP controls after 4 days underlying the importance of the peptides over hydrophilic or hydrophobic surfaces. After 4 days differences in neurite extension were evident among the peptide-functionalized surfaces, with the longest neurites observed on the SIKVAV-functionalized surfaces.

Enhancing the interaction of central nervous system neurons with poly(tetrafluoroethylene-co-hexafluoropropylene) via a novel surface amine-functionalization reaction followed by peptide modification

Poly(tetrafluoroethylene-co-hexafluoropropylene) (FEP) surfaces were modified with cell adhesive peptides, via a novel amination reaction, to enhance the neuron-substrate interaction. Amination of FEP surfaces was achieved by exposing FEP film samples to a UV-activated mercury/ammonia system for either 3 or 24 h, yielding nitrogen compositions of 3.5 and 13.2%, respectively. By labeling the nitrogen functionality with trichlorobenzaldehyde, the surface amine compositions were calculated to be 14 and 4.3% for the 3 and 24 h amination reactions, respectively. Three oligopeptide sequences derived from laminin (GYIGSR, GRGDS, and SIKVAV) were coupled to the aminated FEP (FEP-NH2) surfaces and found to have almost identical surface concentrations as determined by XPS. Using radiolabeled GYIGSR, three coupling agents were compared and the concentration of peptide per surface area was calculated to be 3 and 6 fmol cm-2 for surfaces aminated for 3 and 24 h, respectively, regardless of the coupling agent. The interaction of embryonic hippocampal neurons with the modified surfaces was compared to that with the positive poly(L-lysine)/laminin control in terms of number and length of extended neurites. After 1 day incubation, neurite extension on the GYIGSR- and SIKVAV-coupled surfaces was similar to that on the positive control but significantly greater than that on FEP and FEP-NH2 control surfaces. These peptide-coupled fluoropolymer surfaces enhance the neuron-fluoropolymer interaction, similar to that observed with PLL/laminin.

In vivo biostability of a polymeric hollow fibre membrane for cell encapsulation

The biostability of poly(acrylonitrile-co-vinyl chloride) (P(AN/VC)) hollow fibre membranes was assessed in the rat peritoneal cavity over a 12 month period. The mechanical and chemical stabilities of the hollow fibre membrane (HFM) were characterized by measuring its tensile strength and molecular weight (by gel permeation chromatography) pre-implantation and post-explantation. The stability of the HFM transport properties was determined by molecular weight cut-off (MWCO) and hydraulic permeability (HP). Explanted HFMs were treated with 4 M NaOH to remove adsorbed protein before measuring mechanical, chemical and transport properties. The HFM was stable in vivo for at least 12 months: (i) weight average molecular weight (Mw) at t = 0 was 143,000 g mol-1 (with a polydispersity index (PDI) of 2.3) and at t = 12 months it was 128,000 g mol-1 (with a PDI of 2.8); and (ii) tensile strength at t = 0, 52 +/- 2 mdyne, did not change significantly over time and was 46 +/- 7 mdyne at t = 15 months (P > 0.05 by a two-tailed Student's t-test); and (iii) no significant differences, with respect to standard deviation, were observed in the transport properties: HP was 7.4 +/- 1.5 ml min-1 m-2 mmHg-1 at t = 0 and 7.5 +/- 1.5 ml min-1 m-2 mmHg-1 at t = 12 months, while MWCO (at 90% rejection) was initially 40,000 +/- 8000 g mol-1 and then 54,000 +/- 10,000 g mol-1 at t = 12 months.