Magnetic Nanoparticle-based Approaches to Locally Target Therapy and Enhance Tissue Regeneration in vivo

Tissue Engineering & Regenerative Medicine

Tissue engineering or regenerative medicine offers new possibilities for the functional and structural restoration of damaged or lost tissue. Tissue engineering involves either seeding cells into a 3D structure called a scaffold, to which the cells can attach and grow, or a bottom-up approach, which involves patterning cells according to a predefined organization that will guide the maturation of the tissue engineered construct.^[49] Growth factors (GFs) are frequently added to enhance the proliferation and differentiation of the cells that have been seeded into the scaffold. Mechanical stimulation can also be applied for these purposes. MNPs and magnetic fields have been investigated for their applicability to all of these aspects of tissue engineering.

Scaffolds

The general requirements for scaffolds used in tissue engineering and the various techniques available for their manufacture are described elsewhere.^[49–52] Magnetic scaffolds can provide unique capabilities not available with other methods and materials. Magnetic scaffolds can provide controlled release or redosing of GFs, mechanical stimulation of the seeded cells, improved cell seeding and the means of assembling a scaffold in the desired configuration.

The challenge of developing magnetic scaffolds extends beyond the technical problems of scaffold design and the development of an appropriate external magnetic system. Although the general requirements for scaffolds have been well described, the detailed requirements for many specific applications have not yet been determined. For example, mechanical stimulation, discussed in detail below, is known to increase cell differentiation. However, a detailed description of the optimum force required (amplitude, frequency, duration and direction) remains unknown for many applications. If the required forces were well characterized, the problem could be reduced to the still difficult task of designing a magnetic system to deliver this force to cells in a magnetically responsive scaffold.

GF Delivery & Release In most tissue engineering approaches, GFs are preloaded into the scaffold prior to implantation. A general overview of GF delivery in tissue engineering, including an extensive list of commonly used GFs, is provided in.^[53] Non-magnetic approaches for delivery or controlled release of GFs include the use of gelatin^[54] and PLGA^[55] microparticles incorporated within the scaffold pores, polymer microspheres or nanospheres produced from a number of degradable and nondegradable polymers of synthetic or natural origin,^[56,57] and surrounding the scaffold with a gelatin hydrogel loaded with GF.^[58] Hydrogels have also been used in a more controlled approach through electrostatic binding of proteins and their subsequent sequential release at rates reflected by their equilibrium binding constants.^[59,60]

A review of the use of nanoparticles for GF delivery shows that although a large number of investigations of nanoparticle systems for GF delivery have been conducted, very few of these involved MNPs.^[61] For some applications such as bone graft substitution, the optimum approach would be to continuously add GF to the engineered tissue in a way that mimics natural growth conditions. It might also be desirable in some situations to modify the release profile based on measurements of the growth rate. Although the non-magnetic approaches for delivery of GFs often provide controlled release, the release profile is preprogrammed prior to implantation. Magnetic approaches have the capability to modify the release profile as needed in vivo.

Possible methods for magnetic delivery and release of GFs include: attraction of MNP loaded with GFs to a magnetically responsive scaffold; magnetically mediated heating of a thermally responsive polymer through the application of an external magnetic field of high frequency; and exertion of a mechanical force to a magnetically responsive scaffold through the application of an external time-varying or continuous field.

Bock et al. have developed a magnetic scaffold to which GFs or other biologically active molecules bound to magnetic particles can be delivered on demand by means of an externally applied uniform magnetic field.^[62] Magnetically responsive scaffold exposed to a uniform magnetic field alters the distribution of magnetic flux and leads to higher field gradients near/inside the scaffold in much the same manner as described for the magnetic stents. Commercial scaffolds made of hydroxyapatite and collagen are transformed into magnetic scaffolds by dip-coating the scaffolds in aqueous ferrofluids containing iron oxide nanoparticles stabilized by various macromolecules. Mesenchymal stem cells (MSCs) isolated from human bone marrow were seeded inside the scaffolds and cultured under static magnetic conditions. The continuous magnetization of scaffolds did not pose significant adverse effects on cell viability. Computer simulations performed for a spherical magnetic scaffold 1 cm in diameter, and 150-nm diameter MNPs showed the attractive force exerted by the scaffold on the MNP exceeded the weight of the particle when the field gradient reaches 13 Oe cm⁻¹.

Another approach to making GFs available in an engineered tissue is to preload a magnetic scaffold with the appropriate GFs and utilize an external magnetic field to mechanically release the GFs to the cells as needed. It has been suggested that a ferroscaffold developed for drug release applications could be utilized as a scaffold for engineered tissues.^[63] These scaffolds were fabricated using an in situ synthesis of iron oxide nanoparticles in the presence of various concentrations of biodegradable gelatin. Drug release was demonstrated using vitamin B12 and an external magnetic field of approximately 400 Oe that was switched on and off. No attempt was made to seed this scaffold with cells or to deliver GFs, but the authors state that the drug release properties demonstrated make this scaffold a good candidate for tissue engineering.

A macroporous ferrogel scaffold has been developed that can be remotely controlled by a magnetic field to deliver various biological agents on demand.^[64] The active porous scaffold gives a large deformation and volume change of over 70% under a moderate magnetic field. Under applied magnetic fields, the macroporous ferrogel can give large and prompt deformation, causing water flow through the interconnected pores. The resulting deformation and water convection was shown to trigger and enhance the release of biological agents in a mouse model.

Considerable work has been done in the development of thermally responsive polymers that could release a drug or GF with the application of energy from external sources, such as ultrasound, near infrared, UV, visible wavelength light and magnetic fields.^[65] Thermo-responsive materials have a sharp transition temperature at which they become either soluble or insoluble. When the transition is from a more soluble to a less soluble state, this temperature is known as the lower critical solution temperature. Conversely, if the transition is from a less soluble to a more soluble state, this temperature is known as the upper critical solution temperature.^[66] The combination of MNPs and thermo-responsive polymers is unique because MNPs exposed to an alternating magnetic field of high frequency exhibit an increase in temperature due to magnetic hysteresis loss and Brownian relaxation.^[67,68] This change in temperature can be conducted to a thermoresponsive polymer leading to a polymer phase transition and a consequent drug burst release.

Extensive research has been performed in using this technology for controlled drug release^[69–72] but little work on extending this to controlled GF release has been published. Controlled GF release could be achieved either through direct heating of thermally responsive polymeric particles, which contain both MNPs, and the GF, or through secondary heating in which a thermally responsive particle containing the GF is surrounded by magnetic particles that are heated and the resulting increase in local temperature causes release of the GF from the thermally responsive polymeric particle.

Mechano–magnetic Stimulation of Cells Within Scaffolds In tissue regeneration, it has been shown that in addition to molecular signals (e.g., GFs), physical cues, such as electrical signaling, mechanical stimulation of constructs and medium perfusion, may be essential for appropriate tissue formation.^[73–76] These signals aim to either mimic signals found in vivo or induce beneficial cellular processes for tissue formation.

Mechanical stimulation of cells has been broadly investigated in the last couple decades. The most common examples for application of mechanical stimulation are bioreactors, developed to apply mechanical forces via piston/compression systems, substrate bending, hydrodynamic compression and fluid shear.^[77–80] Although this approach was found to have a positive impact for many tissue types, it has its drawbacks. The forces are mainly applied to the scaffold rather than directly to the cell membrane or cytoskeleton where they are required, limiting their applicability to 3D cultured cells. In addition, its implementation is limited to in vitro application and cannot be extended to in vivo tissue engineering.

A major advantage of magnetically mediated stimulation is that nanomagnetic actuation allows 'action at a distance' (thus enabling actuation both in vitro and in vivo). The magnetic field can be coupled to the particle to actuate a process within a target cell regardless of whether there are intervening structures, such as tissue. In addition, stress parameters can also be varied dynamically, simply by changing the strength and frequency of the applied field. The ability to precisely target, manipulate and activate individual ion channels or targets within the cells is probably the biggest advantage of this approach.

Mechanotransduction, while being a very fundamental and important initiator of many biological processes, is still not fully understood in terms of mechanisms. Briefly, stretch-activated ion channels, responsible for the activation of the mechnotransduction pathway, are found within the cell membranes of almost every cell type. There is a large amount of evidence to suggest different kinetic activation patterns, including stretch,^[81–84] state of phosphorylation^[85–87] and the presence of specific ligand.^[88] Nonetheless, the majority of structural studies have shown that these channels can sense membrane tension directly. In order to test whether magnetically tunable particles could induce mechanical stimulation within cells, different application methods were used (broadly reviewed in ^[89]).

The feasibility of mechanical stimulation induced by magnetic forces has been shown in several previous works, mainly by binding magnetic particles to cellular targets. The idea to deform the cytoskeleton and test cell response through magnetically responsive particles bound to integrin receptor on the cell membrane led to numerous experiments already in the 1990s.^[90–92] In this model, magnetic particles are specifically bound to cell surface integrins and actuated from outside by magnetic field application, leading to an overall membrane and cytoskeleton deformation, thus causing activation of the mechanosensitive channels and cell response.

Investigators in this field have aimed to increase binding specificity to ion channels only, thus enabling actuation of the targeted channel without interrupting normal cell function. This could be achieved by particle modification with ion channel-specific antibodies. Bound particles were later actuated by high gradient magnetic fields leading to channel opening and appropriate cellular response.^[93] These and other examples proved the ability of externally applied magnetic fields to induce cellular response.

One of the existing drawbacks of the receptor stimulation technique is its inability to stimulate cells in the long-term due to internalization of the magnetic particles. Hence, the idea to stimulate cells when the particles attached on their surface was further developed by investigating the effect of internalized particles within the stimulated cells. Almost all studied cell types showed the ability to internalize nano- and submicron particles. The ability to internalize the particle is dependent on various factors such as cell type, particle size, the hydrophobicity and surface charge of the particle polymer, the nature of the particle surface coating and the proliferation rate of the cells, broadly reviewed by Hughes et al..^[89] These properties have recently been exploited for use in transfections and internal manipulations within the cell of interest.^[94–96] Several studies investigated various conditions of magnetic cell loading for use in targeted cell delivery.^[97,98] MacDonald et al. had shown that internalization of particles is an active (cytoskeleton reorganization-dependent) and magnetic force-dependent process.^[98]

The authors of this review recently implemented a mechanical stimulation approach in 3D cultivation systems within polymeric scaffolds.^[99] A new alginate-based composite biomaterial with tunable and externally controlled properties was explored in its ability to provide means of physical stimulation to endothelial cells (Figure 3). We created magnetite-impregnated alginate scaffolds proven to be magnetically responsive under exposure to an alternating magnetic field. These scaffolds were seeded with bovine aortic endothelial cells and stimulated by an alternating magnetic field (10–15 Gauss, 40 Hz) during the first 7 days of a 14-day experimental course. Cells within stimulated constructs showed significantly elevated metabolic activity during the stimulation period, implying a migration and reorganization processes within the cells. Immunostaining and confocal microscopy analyses further confirmed this observation showing that on day 14, in magnetically stimulated scaffolds without the addition of any GFs or other supplements, cellular vessel-like (loop) structures, known as indicators of vasculogenesis and angiogenesis were formed compared with cell sheets or aggregates observed in the nonstimulated (control) scaffolds (Figure 4). Accurate control of cellular organization to form tissue-engineered constructs together with additional molecular signals could lead to the creation of an efficient prevascularized tissue construct with potential applicability for transplantation.

The exact mechanism of cell stimulation within the magnetically responsive matrix is yet to be determined. Because MNPs and their aggregates are anisotropic in terms of their geometrical shape (i.e., not of ideally spherical geometry) they might generate local torque forces applied directly on cells adhered to the MNP-decorated scaffold wall surface and result in a local magneto–mechanical effect applied on cells. Alternatively, when the magnetic particle density within the scaffold is relatively high, enabling nanoparticles to experience magnetic attraction; a magnetostrictive mechanism could be employed.

Magnetostriction is a property of ferro- and ferri-magnetic materials that causes material to change its shape or dimensions during the process of magnetization. Individual magnetite crystals in the size range of 5–20 nm are superparamagnetic due to their small size. However in the alginate-composite material, they aggregate into larger structures of 776 ± 416 nm, displaying slight hysteresis, which is indicative of a slow magnetic relaxation process, resulting in a remnant magnetization or ferrimagnetism.^[99] When the composite magnetic material is exposed to a magnetic field, magnetization on particles generates magnetostrictive strain due to particle attraction, leading to overall scaffold deformation and change in dimensions mimicking the behavior of domains in bulk ferro- or ferri-magnetic materials. For example, such alternating deformation can be scaffold contraction leading to a direct mechanical effect applied to cells. These hypotheses should be corroborated experimentally.

Scaffold Production Another application for MNPs in tissue engineering scaffolds is the use of MNPs and an applied magnetic field in the initial fabrication of the scaffold.^[100–103] The ordered structure of the extracellular matrix of tissues in living organisms plays a key role in cellular response. Duplicating these structures at the nanoscale in the laboratory for tissue engineering applications has proved to be a difficult endeavor.

Alsberg et al. have developed a method to spatially control the self-assembly of fibrin lattices.^[100] Fibrin has been investigated extensively as a biological scaffold for bone cartilage, neural, adipose and blood vessel regeneration. The structure and morphology of fibrin networks (i.e., fiber size, branching and fiber spacing) influence their physical properties as a scaffold, and the ability to control the structure is essential. In this technique superparamagnetic microbeads were first coated with thrombin and positioned by a controlled magnetic field into hexagonal arrays. When a fibrinogen solution was added to the magnetically aligned array, the fibrin nano-fibrils that subsequently polymerized from the beads preferentially oriented along the main bead–bead axes in a triangulated geodesic pattern. The authors demonstrated biocompatibility of human microvascular endothelial cells that were cultured on the resultant fibrin matrices.

Control of scaffold porosity in addition to layout is another important factor in developing scaffolds for tissue engineering. Hu et al. combined particulate leaching technology using sugar with magnetic microparticles and a magnetic field to fabricate 2D and 3D porous biodegradable scaffolds made of poly(L-lactide-co-ε-caprolactone).^[104] Ferrite micro-/nano-particles were encapsulated in sugar microspheres to enable their magnetization. A magnetic apparatus consisting of a block-type neodymium magnet underneath a grid of steel wires was magnetized by the magnet and was then used to form an assembled template for polymer. After polymer casting and removal of the sugar template, spherical pores were generated inside the scaffold. The authors demonstrated that this approach could be extended to 3D scaffolds, such as those needed for vascular tissue engineering by winding the 2D porous sheets on sacrificial molds. The biocompatibility of the developed scaffold was confirmed by viable cells after 4-day culture.

Hydrogels are widely used as tissue-engineering scaffolds. One of the challenges has been to use hydrogels in a 'bottom-up' assembly approaches that attempt to replicate nature's use of repeating structures to build constructs by assembling well characterized building blocks. Yuet et al. have approached this problem through the microfluidic synthesis and field-driven self-assembly of monodisperse, multifunctional Janus hydrogel particles with anisotropic superparamagnetic susceptibility and chemical composition.^[103] Janus particles have the property that surfaces of the two hemispheres exhibit different chemical properties. In this case, one hemisphere is superparamagnetic and the other is non-magnetic. This results in an anisotropic magnetic susceptibility and, under an applied field, permits one hemisphere to interact with satellite particles or nearby chains while preserving the chain's symmetry in a lateral field. Under varying conditions the authors were able to demonstrate self-assembly into a stationary, semiregular array and mesh-like superstructures formed as parallel chains zippered together. MNPs have been shown to enhance osteoinduction even without the presence of a magnetic field.^[105] This finding has been used as a basis for development of magnetic biodegradable fibrous materials with potential applications in bone regeneration.^[106] Nanofibrous membranes were fabricated by electrospinning Fe₃O₄/chitosan/polyvinyl alcohol. MG63 human osteoblast-like cells were seeded on the membranes and showed good cell adhesion and proliferation based on scanning electron microscopy observation and MTT assay. Using tissue culture plates as controls, cells cultured on these membranes had increased proliferation on days 3, 5 and 7 and this improvement increased with higher Fe₃O₄ nanoparticle loading. The authors conclude that the results suggest that the magnetic biodegradable nanofibrous membranes can be a promising biomaterial for enhancement of osteogenesis. Because cell adhesion and proliferation correlated with the nanoparticle loading, the authors also suggested the possibility of further controlling cell function through regulation of the Fe₃O₄ nanoparticle loading content in the membranes.

Cell Patterning

The use of scaffolds for tissue engineering presents some limitations that could limit their effectiveness in certain tissue engineering applications. Scaffolds can slow or delay the organization of cells and the establishment of cell–cell interactions.^[107] Scaffolds could be poor substitutes for the extracellular matrix due to a number of factors including insufficient biological activity, immunogenicity and elevated inflammatory reactions, fluctuating degradation rate and uncontrollable cell–biomaterial interactions.^[108] A detailed list of the advantages and disadvantages of scaffold and scaffold-free tissue engineering approaches for different applications is provided in.^[109] A review of non-magnetic approaches for cell patterning including cell sheets, cell-laden hydrogels, 3D printing, inkjet printing and laser-assisted bioprinting is provided in.^[49] MNPs and magnetic fields can be used to position cells in a pattern suitable for tissue engineering without the use of artificial scaffolds.

A technique named magnetic force-based tissue engineering (Mag-TE) has been developed^[110] and applied to a number of applications including preparation of artificial skeletal muscles,^[111] bone tissue for repair of defects,^[112,113] small-diameter vascular tissue for graft survival^[114] and retinal pigment epithelium for choroidal neovascularization.^[115] Although there are variations in the details for the various applications, the basic concept involves using magnetite cationic liposomes containing magnetite nanoparticles that electrostatically interact with cell membranes, and can therefore be used for magnetically labeling live cells. These magnetically labeled cells are accumulated in a desired pattern through an applied magnetic field and steel structure under a cell culture surface. The authors report that both patterned lines of single cells and complex cell patterns (curved, parallel or crossing patterns) were successfully fabricated.

The Mag-TE technique has also been used to generate a MSC sheet to treat severe ischemic diseases.^[116] Using the Mag-TE techniques, magnetized MSCs were formed into multilayered cell sheets. These sheets were placed into nude mice subjected to unilateral hind limb ischemia and compared with both saline and injected MSCs. The MSC sheet group had a greater angiogenesis in ischemic tissues compared with the control and MSC-injected groups as measured by capillary density and arteriole density in histological sections harvested from the ischemic adductor and gastrocnemius muscles.

A nonuniform applied magnetic field has been used to create a 3D cell assembly of magnetically labeled cells^[117] based, in part, on earlier work by Wilhelm et al..^[118] Experiments were conducted using human endothelial progenitor cells and mouse macrophages magnetically labeled using anionic citrate-coated iron oxide nanoparticles. Magnetic field gradients were applied to suspensions of these cells using either a cylindrical tip or a truncated tip placed on a permanent magnet. For the cylindrical tip, the gradient was approximately 1000 T/m at 500 µm resulting in a magnetic force several orders of magnitude higher than other forces experienced by a cell in suspension, including Brownian motion or buoyancy. Cells progressively stacked near the tip to form a 3D aggregate. The authors were able to control the packing density of cells by tuning the magnetic field gradient geometry and intensity, the magnetic cellular load and the number of cells. From the packing density the authors made some structural inferences based on a comparison of packing density to that of a simple cubic crystal. The authors believe that this ability to control cell density and distribution is essential for the formation of tissues in vitro.

Magnetic levitation has been evaluated to address the challenge of 3D tissue culture.^[107] The authors demonstrated that the structure of the tissue culture can be manipulated, and multicellular clustering of different cell types in co-culture can be accomplished through control of the field configuration. Using a variety of ring-shaped magnets, cells were levitated and the resultant structures evaluated. Using a large-radius magnet, it was observed that the shape of the cell pattern generated was ring-shaped and this pattern was maintained after the magnet was removed. The concept of magnetic levitation was also used in development of 3D tumor spheroids that mimic in vivo tumors for potential anticancer drug screening.^[119]

One of the challenges in engineering 3D cell-dense tissues is the conflicting objectives of increasing porosity within fibrous scaffolds to improve cellular infiltration and the negative effect that this has on fiber alignment.^[120] To address these issues an electrospinning technique was used to fabricate fibrous bundles consisting of composite fibers of poly(L-lactic-co-glycolic) acid and MNPs. C2C12 myoblasts were seeded on the bundles and were grown along the direction of the underlying fibers. When treated with the differentiating medium heat-inactivated horse serum, the myoblasts fused together and formed multinucleated myotubes. After exposure to an external magnetic field the resultant cell rods responded by self-assembling into 3D tissues with a highly ordered architecture. After 3 days, 3D cell-dense tissue architecture was retained when the magnetic field was removed.

Microscale cell-laden hydrogels fabricated using the photopatterning method can be useful building blocks for tissue engineering applications. However, 3D assembly of these microgels to form larger 3D complex constructs is still a challenge. To address this challenge an approach has been developed in which MNP-loaded cell-encapsulating microscale hydrogels were fabricated and assembled into 3D multilayer constructs using magnetic fields.^[121] By spatially controlling the magnetic field, the authors demonstrated that 3D construct geometry can be manipulated, and multilayer assembly of multiple microgel layers can be achieved.

A low-cost method termed magnetic hydrogel-based cell patterning has been developed in an attempt to perform cell patterning in a way that is not dependent on cell labeling.^[122] Using hydrogel blocks and a simple magnet, the authors were able to produce complicated cell patterns. The blocks were fabricated by mixing magnetic particles with the hydrogel and then generating the desired patterns through photolithography. Cells were seeded into a culture plate into which the hydrogel blocks had been placed. The hydrogel blocks prevented cell adhesion in those areas onto which they had been inserted. The blocks were subsequently removed using a magnet and the cell pattern maintained. Heterotypic cell patterning can be achieved by seeding a second type of cell, which preferentially adheres to areas not already seeded.

Cell Seeding

Cell seeding into scaffolds for tissue engineering presents several challenges that can be addressed through the use MNPs and applied magnetic fields. Proper distribution of the cells within the scaffold is often difficult to achieve due to the hydrophobic nature of most scaffold materials. In addition, with 3D scaffolds it is relatively easy to seed cells onto the surface layers but much more difficult to seed cells at the proper density into the interior of the scaffold. Non-magnetic approaches to solve these problems include various dynamic seeding methods and bioreactors.^[123–125] These methods may have limitations in that they may either destroy the porous scaffold architecture or fail to produce tissue of adequate thickness.^[126] None of the studies cited below directly compared magnetic cell seeding effectiveness with these other approaches. One potential advantage of magnetic seeding is that it could possibly control the distribution of cells, perhaps in a nonhomogenous configuration, within the scaffold through appropriate design of the magnetic field. Several of the papers below mention this possibility, but it has not been demonstrated experimentally.

A magnetic tweezer system has been developed to apply controlled forces to a large number of magnetic beads or magnetically labeled cells inside a scaffold.^[127] Although the primary purpose of this research is to use the magnetic force acting on magnetic objects of various sizes to determine local physical parameters of the scaffold, the technology can also allow optimization of cell seeding in the construct and induce a defined 3D cellular organization.

The combined use of MNP-seeded cells and magnetic force was shown to increase the infiltration and distribution of cells into PLGA salt-leached scaffolds compared with controls.^[126] 3T3 fibroblast cells containing magnetic iron/platinum nanoparticles were seeded along one side of the scaffold. A cylindrical neodymium magnet (magnetic field, 4000 Gauss) was placed underneath the well plate containing the seeded scaffold. After 3 days of incubation, measurements were taken of cell penetration into the scaffold. Magnetic seeding resulted in a more than tenfold increase in density in the center of the scaffold compared with non-magnetic controls.

The Mag-TE technology described above has been applied to the problem of cell seeding into a scaffold using a technique termed 'mag-seeding'.^[128] NIH/3T3 fibroblasts, magnetically labeled with magnetite cationic liposomes (described above), were seeded onto six types of commercially available scaffolds. A magnet (4000 Gauss) was placed under the scaffold. The cell-seeding efficiency for all scaffolds was enhanced by Mag-seeding relative to static seeding (58.9 vs 10.8%).

The potential of magnetically mediated multilayered cell seeding in a tubular architecture for generation of vascular grafts was demonstrated by Perea and coworkers.^[129] In this work, a radial magnetic force generated around the collagen-based tubular scaffold guided the sequential seeding of five layers of magnetically labeled human smooth muscle cells followed by the deposition of one layer of human umbilical vein endothelial cells. Co-cultured tubular graft incubated over a 5-day period demonstrated densely packed multilayers of smooth muscle cells coated with one layer of endothelial cells resembling the natural blood vessel architecture. Magnetically mediated cell seeding in tubular geometry enables rapid cell deposition, avoiding cell settling effects due to gravity, leading to accelerated cell–substrate adhesion. Creation of tubular constructs using this methodology is immediate as compared with the dynamic rotational seeding, which occurs at a much longer timescale.

Nanomedicine. 2012;7(9):1425-1442. © 2012  Future Medicine Ltd.