- Email: [email protected]
The in-flow capture of superparamagnetic nanoparticles for targeting therapeutics
Available online at www.sciencedirect.com
Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19 – 29 www.nanomedjournal.com
Original Article: Clinical Nanomedicine
The in-flow capture of superparamagnetic nanoparticles for targeting therapeutics Nicholas J. Darton, PhD, Bart Hallmark, MEng, PhD, CEng, MIChemE, Xuan Han, MEng, Sarah Palit, MEng, Nigel K.H. Slater, PhD, CEng, FIChemE, FREng,⁎ Malcolm R. Mackley, PhD, FIChemE, FInstP, FREng Department of Chemical Engineering, University of Cambridge, Cambridge, United Kingdom
Abstract
Key words:
Superparamagnetic nanoparticles have been synthesized that could potentially be used to magnetically target therapeutics within the body. The magnetic targeting and successful in-flow capture of 330-nm and 580-nm agglomerates of these magnetite nanoparticles was performed using a 0.5-T magnet. Optical observation of magnetic nanoparticle capture in microcapillary flow provides a useful preliminary way of establishing conditions for the magnetic capture of nanoparticles with direct relevance to blood vessels for magnetically directed therapy. A stable nanoparticle layer of 580-nm agglomerates could be formed at mean capillary flow velocities of up to 2.5 cm s–1 and for the 330-nm agglomerates at velocities up to 4.4 cm s–1. These data show that smaller nanoparticle agglomerates form a layer that is impervious to erosion by fluid shear. Capillary blocking by nanoparticles, analogous to an embolism, was not detected in these experiments. © 2008 Elsevier Inc. All rights reserved. Superparamagnetic nanoparticles; Magnetic capture; In-flow deposition; Targeted therapy
One of the key problems that hampers the advancement of chemotherapeutic and gene therapy strategies is the targeting of the therapeutic to the site of disease. The use of magnetic targeting of therapeutic-tagged superparamagnetic nanoparticles for this purpose is an exciting possibility, because it could potentially provide a noninvasive, precise, and inexpensive solution. Magnetic nanoparticles are already in use clinically as contrast agents in magnetic resonance imaging as well as cell sorting and immunoassays.1 The delivery of cytotoxic drugs
Received 24 August 2007; accepted 13 November 2007. The authors would like to acknowledge both the Engineering and Physical Sciences Research Council, United Kingdom, and the Biotechnology and Biological Sciences Research Council, United Kingdom, for financial support. ⁎ Corresponding author: Department of Chemical Engineering, New Museums Site, Pembroke St., Cambridge CB2 3RA, United Kingdom. E-mail address: [email protected] (N.K.H. Slater).
by magnetic targeting was investigated in early work with magnetic microparticles.2,3 Most recently implanted magnets have been used to target intravenously administered 200-nm paramagnetic nanoparticle–linked doxorubicin to tumors in vivo.4 The use of paramagnetic nanoparticles for targeting of virus-based gene therapy vectors was proposed from an in vitro study by Hughes et al.5 Subsequently in vivo gene delivery by this method has been successfully demonstrated in mice.6 Promising in vivo results have led to three early-stage clinical trials of magnetic targeting of cytotoxic drugs to tumors. In two of these trials, epirubicin-linked 100-nm paramagnetic nanoparticles administered intravenously were successfully targeted to tumors with an externally applied 0.2- to 0.8-T magnet placed beside the tumor.7,8 Wilson et al9 delivered a sample of intra-arterially administered doxorubicin-linked 0.5- 5-μm paramagnetic nanoparticles to tumors in patients with an external 1.5-T magnet placed adjacent to the tumor.
1549-9634/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.nano.2007.11.001 Please cite this article as: N.J. Darton, B. Hallmark, X. Han, S. Palit, N.K.H. Slater, M.R. Mackley, The in-flow capture of superparamagnetic nanoparticles for targeting therapeutics, Nanomedicine: NBM 2008;4:19-29, doi:10.1016/j.nano.2007.11.001
20
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Superparamagnetic magnetite (Fe3O4) nanoparticles can be synthesized cheaply and efficiently by co-precipitation of Fe2+ and Fe3+ aqueous salt solutions by addition of a base under an oxygen-free, nonoxidizing environment. Control of size, shape, and composition of nanoparticles depends on the type of salts, Fe2+ and Fe3+ ratio, pH and ionic strength of the media.10 The addition of a polymer such as dextran during the precipitation process results in coated particles that can be attached to different ligands, such as antibodies,11 that can bind therapeutics. A problem facing this precipitation method of nanoparticle formation is that the 6- to 10-nm magnetite nanoparticles produced have a high propensity to agglomerate. This problem can be reduced with the addition of polymethacrylic acid (PMAA) followed by ultrasonication.12 Some previous observations have been made of magnetic targeting of 500-nm ferromagnetic particles in flow through a microcapillary tube as a model of targeting in blood vessels.13 These have had only limited success and have not used superparamagnetic nanoparticles or realistic flow rates. Thus far little experimental investigation has been done into the effect of fluid flow rate on the dynamics of superparamagnetic nanoparticle capture and buildup of particles at the magnetically targeted site. This work addresses these issues and also studies the stability of the captured nanoparticle layer by increasing the fluid flow rate and observing the patterns of erosion that result. Magnetic capture of superparamagnetic nanoparticles was performed with fluid viscosity, density, and velocities representative of those of blood in medium-sized venules and arterioles in the human body, thus ensuring that the shear stress on the capillary walls in vitro are of a similar order to those in vivo.
Methods FeCl3·6H2O, FeCl2·4H2O, 25% NH3 ·H2O, and HCl were obtained from Fisher Scientific (Loughborough, United Kingdom) and PMAA from Sigma Aldrich (St. Louis, Missouri). The plastic capillary arrays were manufactured in house from a commercially available plastomer (Dow Affinity) using a novel extrusion-based process that has been documented elsewhere.14,15 Synthesis of superparamagnetic nanoparticles Superparamagnetic magnetite nanoparticles with no additional polymer coating were synthesized so that possible surface coat unevenness or interactions would be reduced. The method for fabricating nanoparticles was adapted from Xia et al16 using overhead mixing rather than ultrasonication to mix the reaction. First, 9.6 mL of 1.5 M FeCl3·6H2O was added to 6.4 mL 0.5 M FeCl2 · 4H2O, then made up to 80 mL with degassed MilliQ (Millipore, Watford, United Kingdom) H2O. This mixture was filtered through a 0.22-μm filter membrane then mixed for 15
Figure 1. Transmission electron microscopy image of a cluster of 10-nm superparamagnetic magnetite nanoparticles.
minutes at 60°C under nitrogen. Then, 6 mL 25% NH3 ·H2O was added dropwise over 5 minutes with vigorous stirring (900 rpm) with an overhead stirrer. As the base was added, the solution changed in color from brown to black. This reaction was stirred for a total of 15 minutes at 60°C under nitrogen. The resulting black nanoparticle solution was then dialyzed in 12-kDa cutoff dialysis tubing against three changes of MilliQ H2O. To reduce agglomeration of the superparamagnetic nanoparticles, 3% by weight PMAA was added and the pH adjusted to 7.4.12 This resulted in superparamagnetic nanoparticle agglomerates of 575 ± 8.0 nm in size as measured in a Brookhaven ZetaPALS sizer (Brookhaven, Holtsville, New York). To get a smaller agglomerate size of 328 ± 3.5 nm, a sample of nanoparticles in 3% by weight PMAA was sonicated with a 330 W ultrasonic cell crusher (Heat Systems XL-2020, Farmingdale, New York) on full power for 10 minutes. The morphology of magnetite nanoparticles in MilliQ H2O was analyzed on a Philips CM100 (FEI Europe, Eindhoven, The Netherlands) transmission electron microscope (Figure 1). Roughly spherical nanoparticles with an average size of 10 nm could be observed on transmission electron microscopy (TEM) to be gathered in clusters. Dehydration of the sample in preparation for TEM always resulted in the presence of nanoparticle clusters, so the agglomerates observed by Zeta-sizing could not be distinguished. Measurements of the iron oxide lattice structure in the nanoparticles by high-resolution electron microscopy indicated that the nanoparticles were composed of magnetite, Fe3O4. Measurements of magnetic moment by SQUID magnetometery at 293 K from –1 to 1 T showed that the particles were superparamagnetic with a magnetic susceptibility of 1.57 × 10–4.
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
21
Figure 2. Schematic diagram of the flow loop, observation, and injection system used to image the deposition and erosion of the superparamagnetic magnetite nanoparticle.
Apparatus and protocol for the in-flow capture and erosion of superparamagnetic nanoparticles A key criterion for an experimental apparatus to observe the deposition and erosion of superparamagnetic nanoparticles was that observation of the flow should be inhibited as little as possible by curved-surface refraction effects. To achieve this a novel plastic capillary array was used, termed a microcapillary film or MCF,14,15 containing capillaries of 410-μm diameter. A useful feature of MCFs is that the circular capillaries that they contain are embedded in an essentially flat film, thus facilitating easier optical observation of the capillary internals. The MCFs were fabricated in house from a commercially available polymer resin, Dow Affinity Plastomer, using a novel extrusion process.15 Previous research had already quantified the characteristics of fluid flows in MCFs.17 Only 1 of the 19 available capillaries was used in each film. Other important criteria were precise control of the flow rate of fluid containing the nanoparticle suspension; a reliable, reproducible means of introducing a pulse of nanoparticle-PMAA solution into the flow system; and a magnet with a high field strength to permit nanoparticle capture. These considerations were met by using an HPLC pump (Kontron 422, Kontron Instruments, Milan, Italy), an electromechanical injection valve unit (VICI Valco, Houston, TX), and a 0.5-T magnet (e-magnets UK, Sheffield) held in position such that it was placed within a few millimeters of one side of the MCF. The apparatus diagram is shown in Figure 2.
Once the pump had been calibrated and the microscope (Intel Digital Blue QX3, Intel, Santa Clara, California) and light source positioned for optimal clarity, experiments were carried out on two sizes of nanoparticle aggregate, 330 nm and 580 nm, in two distinct phases. The first phase of experimentation was deposition. It was found that flow rates less than 1 mL min–1 were optimal for deposition; hence flow rates between 0.1 mL min–1 and 0.5 mL min–1 were examined in steps of 0.05 mL min–1. These volumetric flow rates correspond to fluid velocities of 1.3 cm s–1 and 6.3 cm s–1, respectively, with steps of 0.6 cm s–1. The Reynolds numbers corresponding to these flow velocities are 1.7 and 8.0, respectively, using a measured fluid viscosity of 3% by weight PMAA solution of 2.5 × 10–3 Pa s and a density of 1024 kg m –3 . The flow was, as expected, laminar. Additionally, the physical properties of the 3% by weight PMAA solution were close to those of blood, with values of blood viscosity reported about 3.3 × 10-3 Pa s–1 18 and density between 1043 kg m–3 and 1051 kg m–3.19 A pulse, approximately 2 mL in volume of 40 mg mL–1 nanoparticles in 3% by weight PMAA solution, was injected via the electromechanical valve into the established flow. This deposition phase was continued until one of two criteria were met; either a steady thickness of deposited nanoparticles was held for over an hour after the injected pulse, or the initial layer had subsequently been eroded by the flow. Erosion became the dominant behavior at the upper end of the flowrate range explored. The second phase of the experimental protocol was to observe the erosion of an established nanoparticle layer. The
22
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Figure 3. Sequence of images at (A) 0 minutes, (B) 7 minutes, (C) 32 minutes, and (D) 90 minutes showing the in-flow deposition of superparamagnetic magnetite nanoparticles. The cluster size of the nanoparticles was 580 nm, and the flow rate 0.1 mL min–1. The location of magnet is shown schematically by the cross-hatched rectangle in each image.
flow rate of fluid was incremented from the original deposition flow rate in steps of 0.05 mL min–1 such that a new steady-thickness nanoparticle layer was achieved for a period of 10 minutes at each flow rate. This procedure was repeated until the fluid flow completely eroded the nanoparticle layer. Results Deposition The sequence of optical micrographs shown in Figure 3 illustrates the observed nanoparticle deposition behavior where a steady-state nanoparticle layer thickness was achieved—in this case using 580-nm nanoparticles at a flow rate of 0.1 mL min–1. In this sequence of photographs the location of the capillary walls and of the magnet is shown schematically. Sequences of micrographs were taken for both 580-nm and 330-nm nanoparticles over the flow-rate range studied, and subsequent image analysis was used to quantify the thickness of the nanoparticle layer in millimeters. An estimate of the error in these measurements was also obtained by taking into account the uncertainty of the exact border between the nanoparticle layer and the fluid, and of the layer's unevenness. The series of plots shown in Figure 4 illustrate the deposition kinetics for 580-nm nanoparticles for flow rates
between 0.1 mL min–1 and 0.25 mL min–1. In all four of these cases a stable thickness of the nanoparticle layer was quickly reached. For flow rates above 0.3 mL min–1, corresponding to a fluid velocity of 3.8 cm s–1, a stable layer was not attained as the initially formed layer of nanoparticles underwent subsequent erosion. The sequence of plots shown in Figure 5 illustrates the deposition kinetics for the 330-nm nanoparticles. In this case, it was found that stable nanoparticle layers were formed at flow rates up to 0.35 mL min–1 (corresponding to a fluid velocity of 4.4 cm s–1), higher than the maximum flow rate that yielded a stable layer of 580-nm nanoparticles. Erosion The sequence of optical micrographs shown in Figure 6 illustrates the erosion of an established layer of nanoparticles with further increases in flow rate. In this particular case the layer of 580-nm nanoparticles was deposited at a flow rate of 0.1 mL min–1 until steady-state behavior was reached. Then, the flow rate was incremented in steps of 0.05 mL min–1 until a new steady-state thickness was established. It was observed that there was an upper flow-rate limit, dependent on nanoparticle diameter, for which a steady-thickness layer could be achieved before the layer was advected away with the flow. The plots shown in Figures 7 and 8 quantify this behavior for 580-nm and 330-nm nanoparticles, respectively. The plot with open symbols on each figure corresponds to
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
23
Figure 4. Plots showing the deposition kinetics of 580-nm nanoparticles for (A) 0.10 mL min–1 and 0.15 mL min–1, and (B) 0.20 mL min–1 and 0.25 mL min–1. The dashed horizontal line serves as a visual guide for the approximate location of the steady-state region.
the erosion behavior of a layer of nanoparticles deposited at 0.15 mL min–1, whereas the plot with closed symbols corresponds to both the deposition and erosion behavior at 0.5 mL min–1 since a steady-thickness layer was not achieved at this flow rate for either particle diameter. Discussion The results presented for the deposition of the 580-nm and 330-nm nanoparticles demonstrate that it is possible to capture superparamagnetic nanoparticles that are being advected in a fluid flow by using a magnetic field from a 0.5-T magnet adjacent to a plastic capillary. The thickness and stability of the captured layer of nanoparticles is strongly affected by both the fluid flow rate and the mean hydraulic
diameter of the nanoparticle aggregates. If the plots in Figures 4 and 5 are compared, then it is evident that the larger nanoparticles form a thicker layer, typically between 0.26 mm and 0.16 mm, for flow rates between 0.1 mL min–1 and 0.25 mL min–1, respectively. In contrast, the 330-nm nanoparticles tend to form a more consistently, smaller sized layer, typically between 0.12 mm and 0.09 mm for flow rates of 0.1 mL min–1 and 0.35 mL min–1, respectively. From the optical observations reported in this article it is clear that the magnetic field is capable of attracting the nanoparticles to the walls of the capillary and that the thickness of the nanoparticle layer that is formed is sensitive to flow conditions. A possible explanation for the smaller deposited layer of 330-nm nanoparticles is that each particle contains a smaller
24
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Figure 5. Plots showing the deposition kinetics of 330-nm nanoparticles for (A) 0.10 mL min–1 and 0.20 mL min–1, and (B) 0.30 mL min–1 and 0.35 mL min–1. The dashed horizontal line serves as a visual guide for the approximate location of the steady-state region.
mass of magnetic material when compared with a nanoparticle of 580-nm diameter; the smaller particles, therefore, experience a weaker force as a result of the applied magnetic field, hence a smaller amount are likely to be captured from the fluid flow. The 580-nm nanoparticles, however, contain a larger amount of magnetic material, and hence they experience a larger force due to the applied magnetic field, resulting in a greater amount being captured. In terms of the erosion behavior of the nanoparticles at increasing flow rates, the observation that the layer of 580-nm nanoparticles is unable to be maintained at flow rates at which the 330-nm nanoparticle layer is still present could be explained qualitatively in terms of the applied shear stress on the particle layer due to the fluid flow. The thicker layer of
nanoparticles results in a more constricted flow, hence the flow velocity in the capillary is higher for a given volumetric flow rate. This increased flow velocity, in turn, leads to higher shear stresses on the nanoparticle surfaces and hence a greater driving force to erode the particles. The thinner layer of nanoparticles, on the other hand, results in less flow constriction, and therefore lower shear stresses are present on the particles. In this way the layer of 330-nm nanoparticles can exist at higher volumetric flow rates when compared with the layer of 580-nm nanoparticles. The hypothesis, therefore, is twofold. First, the amount of particles captured, hence the final thickness of the layer that forms, is dependent on both the magnetic force and the hydrodynamic force experienced by the particle. When
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
25
Figure 6. Sequence of images showing the erosion of a pre-established nanoparticle layer, deposited at 0.1 mL min–1, with increasing flow rate. Image (A) corresponds to the initial steady-state layer when the flow rate was first increased, (B) 25 minutes later, (C) 27 minutes later, and (D) 30 minutes later. The cluster size of the nanoparticles was 580 nm, and the flow rate 0.1 mL min–1. The location of magnet is shown schematically by the cross-hatched rectangle in each image.
magnetic forces dominate the hydrodynamic forces, then a particle is captured. However, when hydrodynamic forces become dominant when compared with the magnetic forces, the nanoparticle layer is eroded by the flow. This lessens the flow constriction caused by the nanoparticle layer on the capillary wall and hence lowers the hydrodynamic forces; thus a new equilibrium is achieved. These qualitative explanations can also account for the transient behavior that can be observed in Figures 7 and 8. These two figures present, at first sight, seemingly contradictory trends. The thickness of the layer of 580-nm nanoparticles can be seen to decrease with increasing flow rate until either a new steady-state thickness is formed or the layer erodes away completely. Given the proposed hypothesis, the trend of erosion displayed by the 580-nm nanoparticles with increasing flow rate is expected. In contrast, the layer of 330-nm nanoparticles can be seen at first to increase with increasing flow rates, before becoming steady and then finally eroding away. A proposed mechanism for this observation is that the smaller amount of magnetic material present in the 330-nm nanoparticles results in lower magnetic forces and hence fewer particles from the 2-mL pulse being captured when the initial layer was formed. This layer results in only a small amount of constriction of the fluid flow, not enough for hydrodynamic forces to become dominant. If any additional nanoparticles were present in the flow, they would still therefore be
captured. When the flow rate is increased, nanoparticles captured outside the field of view of the microscope, in an area of weaker magnetic field further away from the magnet, are eroded but then subsequently recaptured onto the nanoparticle layer in the field of view. This leads to the observation that the thickness of the nanoparticle layer at first increases with an increase in volumetric flow rate. This increase is enough to reach equilibrium between magnetic and hydrodynamic forces; thus, subsequent increases in flow rate serve to erode the nanoparticle layer away. Further insight into this mechanism can be had by calculating an estimate of the shear stress on the top of the nanoparticle layer for a given flow rate. This has been done by approximating the area available for fluid flow as a large segment of the circular capillary cross section, whereas the remaining, smaller segment of the capillary represents the nanoparticle layer. An equivalent diameter is then found for the segment containing the flow and the shear stress calculated by assuming that the flow is Poiseuille flow of a Newtonian fluid through a circular capillary with solid walls of the equivalent diameter that had been calculated; this is shown schematically in Figure 9. Despite this approach making some first-order approximations about the nature of the boundary condition on the nanoparticle layer and also the shape of the flow geometry, a reasonable estimate of the shear stress on the nanoparticle layer can be obtained; this estimate is of a similar magnitude to shear stresses reported
26
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Figure 7. Plot of erosion kinetics for 580-nm nanoparticles for increasing erosion flow rates where the initial deposition flow rates were 0.15 mL min–1 (open symbols) and 0.5 mL min–1 (closed symbols).
Figure 8. Plot of erosion kinetics for 330-nm nanoparticles for increasing erosion flow rates where the initial deposition flow rates were 0.15 mL min–1 (open symbols) and 0.5 mL min–1 (closed symbols).
for the onset of erosion of biomass in filtration systems.20 Figure 10 shows the nanoparticle layer thickness that was detailed in Figures 7 and 8 replotted as a function of the shear
stress exerted on it by the fluid flow for the case where both 330-nm and 580-nm nanoparticles were initially deposited at 0.15 mL min–1 and then subjected to flow-rate increments of
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
27
Figure 9. Schematic diagram showing the transformation of the real flow domain into one suitable for the estimation of the maximum fluid shear stress on the top surface of the layer of captured nanoparticles.
Figure 10. Plot of the calculated shear stress on the top surface of the nanoparticle layer as a function of layer thickness for 330-nm nanoparticles (open symbols, dotted line) and for 580-nm nanoparticles (closed symbols, dashed line). The initial deposition flow rate was 0.15 mL min–1 in both cases, and the lines marked on the plot are for visual guidance only.
0.05 mL min–1 until the layer had fully eroded away. The shape of the two trends shown on this plot adds credence to the proposed hypothesis; the applied stress on the 330-nm nanoparticle layer first increases markedly before erosion, suggesting that particle recapture may be occurring as a result of poor initial capture. The applied stress on the
580-nm nanoparticle layer, however, is seen to only increase by a small amount, whereas initial particle erosion rather than capture is occurring before rapid erosion of the layer takes place. It is interesting to note that the stress at which rapid erosion occurs is different in both cases, suggesting that its
28
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
underlying mechanism is complex, possibly involving variables such as nanoparticle packing, how the nanoparticles were initially deposited, and their interaction with the surface of the film and with each other. Further investigation is needed to probe this behavior such that fundamental understanding of this phenomenon can be gained. It must be noted that the environment in which these deposition and erosion experiments have been carried out is a highly simplified model of a circulatory system. The regime of flow is one of constant flow rate rather than the constant pressure drop that one would expect to find across a capillary network, and the fluid carrying the nanoparticles does not contain any particulate matter such as blood cells. The physics underlying the interaction between the hydrodynamic and magnetic forces acting on an immobilized layer of nanoparticles, however, is similar, and the experiments detailed in this article allow one to gain fundamental insight into this system. Future work will involve repeating these experiments in an environment of constant pressure drop that will be more analogous to the regime found in a circulatory system. In this work superparamagnetic nanoparticles have been successfully captured in microcapillaries of 410 μm in diameter in a flow of up to 0.3 mL s–1, equivalent to a velocity of 3.8 cm s–1. This is of particular relevance to therapeutic nanoparticle targeting, because this is the blood velocity expected in a medium-sized human venule or arteriole.21 The measured viscosity of the 3% by weight PMAA solution used in our capture experiments of 2.5 × 10–3 Pa s–1 is also closer than water to that of blood, about 3.3 × 10–3 Pa s–1.18 The effectiveness of nanoparticle capture supports the successful results of magnetic targeting of intravenously7,8 and intra-arterially9 administered nanoparticles in initial clinical trials. Interestingly, further analysis of nanoparticle deposition layer thickness and stability indicates that these properties are dependent on nanoparticle size. The larger 580-nm particles deposit to form a thicker layer quicker that is more susceptible to subsequent erosion. The smaller 330-nm particles initially deposit to form a more stable thinner layer that increases in thickness as flow rates are increased until a critical flow rate is reached and erosion begins. It has been hypothesized in this article that this observed behavior is due to the effect of shear stress on the nanoparticle layer. The shear stress increases as the capillary flow is constricted, causing the nanoparticle layer to reach an equilibrium where the force of the magnetic attraction on the nanoparticles, proportional to particle size, balances the shear stress exerted by the fluid. This would also explain why complete blocking of the capillary with nanoparticles was reassuringly not observed. In future clinical work, nanoparticle agglomerates of between 300 and 600 nm in size could be used for delivery of bound drug or gene-therapy vector to sites of disease within the body. Nanoparticle agglomerates in this size range have
the advantage that they can be trapped and manipulated at the site of disease by inexpensive permanent magnets as small as 0.5 T but are still superparamagnetic. Once magnetically delivered to the site of interest, the nanoparticles may cross into the cells as 10-nm nanoparticles by pinocytosis or as larger agglomerates by phagocytosis to deliver their payload. These findings will provide a useful basis for future nanoparticle design for effective magnetic targeting of therapeutics in vivo. References 1. Pankhurst QA, Connolly J, Jones SK, Dobson J. Applications of magnetic nanoparticles in biomedicine. J Phys D Appl Phys 2003;36: R167-81. 2. Senyei A, Widder K, Czerlinski G. Magnetic guidance of drug-carrying microspheres. J Appl Phys 1978;49:3578-83. 3. Widder KJ, Morris RM, Poore GA, Howard DP, Senyei AE. Selective targeting of magnetic albumin microspheres containing low-dose doxorubicin-total remission in Yoshida sarcoma-bearing rats. Eur J Cancer Clin Oncol 1983;19:135-9. 4. Fernandez-Pacheco R, Marquina C, Valdivia JG, Gutiérrez M, Romero MS, Cornudella R. Magnetic nanoparticles for local drug delivery using magnetic implants. J Magn Magn Mater 2007;311:318-22. 5. Hughes C, Galea-Lauri J, Farzaneh F, Darling D. Streptavidin paramagnetic particles provide a choice of three affinity-based capture and magnetic concentration strategies for retroviral vectors. Mol Ther 2001;3:623-30. 6. Morishita N, Nakagami H, Morishita R, et al. Magnetic nanoparticles with surface modification enhanced gene delivery of HVJ-E vector. Biochem Biophys Res Commun 2005;334:1121-6. 7. Lubbe AS, Bergemann C, Brock J, McClure DG. Physiological aspects in magnetic drug-targeting. J Magn Magn Mater 1999;194:149-55. 8. Lubbe AS, Bergemann C, Riess H, et al. Clinical experiences with magnetic drag targeting: a phase I study with 4′-epidoxorubicin in 14 patients with advanced solid tumors. Cancer Res 1996;56:4686-93. 9. Wilson MW, Kerlan RK, Fidelman NA, et al. Hepatocellular carcinoma: regional therapy with a magnetic targeted carrier bound to doxorubicin in a dual MR imaging/conventional angiography suite-initial experience with four patients. Radiology 2004;230:287-93. 10. Gupta AK, Gupta M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 2005;26: 3995-4021. 11. Duan HL, Shen ZQ, Wang XW, Chao FH, Li JW. Preparation of immunomagnetic iron-dextran nanoparticles and application in rapid isolation of E.coli O157:H7 from foods. World J Gastroenterol 2005;11: 3660-4. 12. Mendenhall GD, Geng Y, Hwang J. Optimization of long-term stability of magnetic fluids from magnetite and synthetic polyelectrolytes. J Colloid Interface Sci 1996;184:519-26. 13. Kikura H, Matsushita J, Kakuta N, Aritomi M, Kobayashi Y. Cluster formation of ferromagnetic nano-particles in micro-capillary flow. J Mater Process Tech 2007;181:93-8. 14. Hallmark B, Gadala-Maria F, Mackley MR. The melt processing of polymer microcapillary film (MCF). J Non-Newton Fluid 2005;128: 83-98. 15. Hallmark B, Mackley MR, Gadala-Maria F. Hollow microcapillary arrays in thin plastic films. Adv Eng Mater 2005;7:545-7. 16. Xia ZF, Wang GB, Tao KX, Li JX. Preparation of magnetite-dextran microspheres by ultrasonication. J Magn Magn Mater 2005;293: 182-6. 17. Hornung CH, Hallmark B, Hesketh RP, Mackley MR. The fluid flow and heat transfer performance of thermoplastic microcapillary films. J Micromech Microeng 2006;16:434-47.
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29 18. Lowe G, Rumley A, Norrie J, et al. Blood rheology, cardiovascular risk factors, and cardiovascular disease: The West of Scotland Coronary Prevention Study. Thromb Haemostasis 2000;84:553-8. 19. Hinghofer-Szalkay HG, Greenleaf JE. Continuous monitoring of blood volume changes in humans. J Appl Physiol 1987;63:1003-7.
29
20. Bustnes TE, Kaminski CF, Mackley MR. The capture and release of biomass in a high voidage fibrous microstructure: mechanisms and shear stress levels. J Membrane Sci 2006;276:208-20. 21. Sherwood L. Human physiology from cells to systems. New York: Wadsworth; 1997.
Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19 – 29 www.nanomedjournal.com
Original Article: Clinical Nanomedicine
The in-flow capture of superparamagnetic nanoparticles for targeting therapeutics Nicholas J. Darton, PhD, Bart Hallmark, MEng, PhD, CEng, MIChemE, Xuan Han, MEng, Sarah Palit, MEng, Nigel K.H. Slater, PhD, CEng, FIChemE, FREng,⁎ Malcolm R. Mackley, PhD, FIChemE, FInstP, FREng Department of Chemical Engineering, University of Cambridge, Cambridge, United Kingdom
Abstract
Key words:
Superparamagnetic nanoparticles have been synthesized that could potentially be used to magnetically target therapeutics within the body. The magnetic targeting and successful in-flow capture of 330-nm and 580-nm agglomerates of these magnetite nanoparticles was performed using a 0.5-T magnet. Optical observation of magnetic nanoparticle capture in microcapillary flow provides a useful preliminary way of establishing conditions for the magnetic capture of nanoparticles with direct relevance to blood vessels for magnetically directed therapy. A stable nanoparticle layer of 580-nm agglomerates could be formed at mean capillary flow velocities of up to 2.5 cm s–1 and for the 330-nm agglomerates at velocities up to 4.4 cm s–1. These data show that smaller nanoparticle agglomerates form a layer that is impervious to erosion by fluid shear. Capillary blocking by nanoparticles, analogous to an embolism, was not detected in these experiments. © 2008 Elsevier Inc. All rights reserved. Superparamagnetic nanoparticles; Magnetic capture; In-flow deposition; Targeted therapy
One of the key problems that hampers the advancement of chemotherapeutic and gene therapy strategies is the targeting of the therapeutic to the site of disease. The use of magnetic targeting of therapeutic-tagged superparamagnetic nanoparticles for this purpose is an exciting possibility, because it could potentially provide a noninvasive, precise, and inexpensive solution. Magnetic nanoparticles are already in use clinically as contrast agents in magnetic resonance imaging as well as cell sorting and immunoassays.1 The delivery of cytotoxic drugs
Received 24 August 2007; accepted 13 November 2007. The authors would like to acknowledge both the Engineering and Physical Sciences Research Council, United Kingdom, and the Biotechnology and Biological Sciences Research Council, United Kingdom, for financial support. ⁎ Corresponding author: Department of Chemical Engineering, New Museums Site, Pembroke St., Cambridge CB2 3RA, United Kingdom. E-mail address: [email protected] (N.K.H. Slater).
by magnetic targeting was investigated in early work with magnetic microparticles.2,3 Most recently implanted magnets have been used to target intravenously administered 200-nm paramagnetic nanoparticle–linked doxorubicin to tumors in vivo.4 The use of paramagnetic nanoparticles for targeting of virus-based gene therapy vectors was proposed from an in vitro study by Hughes et al.5 Subsequently in vivo gene delivery by this method has been successfully demonstrated in mice.6 Promising in vivo results have led to three early-stage clinical trials of magnetic targeting of cytotoxic drugs to tumors. In two of these trials, epirubicin-linked 100-nm paramagnetic nanoparticles administered intravenously were successfully targeted to tumors with an externally applied 0.2- to 0.8-T magnet placed beside the tumor.7,8 Wilson et al9 delivered a sample of intra-arterially administered doxorubicin-linked 0.5- 5-μm paramagnetic nanoparticles to tumors in patients with an external 1.5-T magnet placed adjacent to the tumor.
1549-9634/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.nano.2007.11.001 Please cite this article as: N.J. Darton, B. Hallmark, X. Han, S. Palit, N.K.H. Slater, M.R. Mackley, The in-flow capture of superparamagnetic nanoparticles for targeting therapeutics, Nanomedicine: NBM 2008;4:19-29, doi:10.1016/j.nano.2007.11.001
20
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Superparamagnetic magnetite (Fe3O4) nanoparticles can be synthesized cheaply and efficiently by co-precipitation of Fe2+ and Fe3+ aqueous salt solutions by addition of a base under an oxygen-free, nonoxidizing environment. Control of size, shape, and composition of nanoparticles depends on the type of salts, Fe2+ and Fe3+ ratio, pH and ionic strength of the media.10 The addition of a polymer such as dextran during the precipitation process results in coated particles that can be attached to different ligands, such as antibodies,11 that can bind therapeutics. A problem facing this precipitation method of nanoparticle formation is that the 6- to 10-nm magnetite nanoparticles produced have a high propensity to agglomerate. This problem can be reduced with the addition of polymethacrylic acid (PMAA) followed by ultrasonication.12 Some previous observations have been made of magnetic targeting of 500-nm ferromagnetic particles in flow through a microcapillary tube as a model of targeting in blood vessels.13 These have had only limited success and have not used superparamagnetic nanoparticles or realistic flow rates. Thus far little experimental investigation has been done into the effect of fluid flow rate on the dynamics of superparamagnetic nanoparticle capture and buildup of particles at the magnetically targeted site. This work addresses these issues and also studies the stability of the captured nanoparticle layer by increasing the fluid flow rate and observing the patterns of erosion that result. Magnetic capture of superparamagnetic nanoparticles was performed with fluid viscosity, density, and velocities representative of those of blood in medium-sized venules and arterioles in the human body, thus ensuring that the shear stress on the capillary walls in vitro are of a similar order to those in vivo.
Methods FeCl3·6H2O, FeCl2·4H2O, 25% NH3 ·H2O, and HCl were obtained from Fisher Scientific (Loughborough, United Kingdom) and PMAA from Sigma Aldrich (St. Louis, Missouri). The plastic capillary arrays were manufactured in house from a commercially available plastomer (Dow Affinity) using a novel extrusion-based process that has been documented elsewhere.14,15 Synthesis of superparamagnetic nanoparticles Superparamagnetic magnetite nanoparticles with no additional polymer coating were synthesized so that possible surface coat unevenness or interactions would be reduced. The method for fabricating nanoparticles was adapted from Xia et al16 using overhead mixing rather than ultrasonication to mix the reaction. First, 9.6 mL of 1.5 M FeCl3·6H2O was added to 6.4 mL 0.5 M FeCl2 · 4H2O, then made up to 80 mL with degassed MilliQ (Millipore, Watford, United Kingdom) H2O. This mixture was filtered through a 0.22-μm filter membrane then mixed for 15
Figure 1. Transmission electron microscopy image of a cluster of 10-nm superparamagnetic magnetite nanoparticles.
minutes at 60°C under nitrogen. Then, 6 mL 25% NH3 ·H2O was added dropwise over 5 minutes with vigorous stirring (900 rpm) with an overhead stirrer. As the base was added, the solution changed in color from brown to black. This reaction was stirred for a total of 15 minutes at 60°C under nitrogen. The resulting black nanoparticle solution was then dialyzed in 12-kDa cutoff dialysis tubing against three changes of MilliQ H2O. To reduce agglomeration of the superparamagnetic nanoparticles, 3% by weight PMAA was added and the pH adjusted to 7.4.12 This resulted in superparamagnetic nanoparticle agglomerates of 575 ± 8.0 nm in size as measured in a Brookhaven ZetaPALS sizer (Brookhaven, Holtsville, New York). To get a smaller agglomerate size of 328 ± 3.5 nm, a sample of nanoparticles in 3% by weight PMAA was sonicated with a 330 W ultrasonic cell crusher (Heat Systems XL-2020, Farmingdale, New York) on full power for 10 minutes. The morphology of magnetite nanoparticles in MilliQ H2O was analyzed on a Philips CM100 (FEI Europe, Eindhoven, The Netherlands) transmission electron microscope (Figure 1). Roughly spherical nanoparticles with an average size of 10 nm could be observed on transmission electron microscopy (TEM) to be gathered in clusters. Dehydration of the sample in preparation for TEM always resulted in the presence of nanoparticle clusters, so the agglomerates observed by Zeta-sizing could not be distinguished. Measurements of the iron oxide lattice structure in the nanoparticles by high-resolution electron microscopy indicated that the nanoparticles were composed of magnetite, Fe3O4. Measurements of magnetic moment by SQUID magnetometery at 293 K from –1 to 1 T showed that the particles were superparamagnetic with a magnetic susceptibility of 1.57 × 10–4.
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
21
Figure 2. Schematic diagram of the flow loop, observation, and injection system used to image the deposition and erosion of the superparamagnetic magnetite nanoparticle.
Apparatus and protocol for the in-flow capture and erosion of superparamagnetic nanoparticles A key criterion for an experimental apparatus to observe the deposition and erosion of superparamagnetic nanoparticles was that observation of the flow should be inhibited as little as possible by curved-surface refraction effects. To achieve this a novel plastic capillary array was used, termed a microcapillary film or MCF,14,15 containing capillaries of 410-μm diameter. A useful feature of MCFs is that the circular capillaries that they contain are embedded in an essentially flat film, thus facilitating easier optical observation of the capillary internals. The MCFs were fabricated in house from a commercially available polymer resin, Dow Affinity Plastomer, using a novel extrusion process.15 Previous research had already quantified the characteristics of fluid flows in MCFs.17 Only 1 of the 19 available capillaries was used in each film. Other important criteria were precise control of the flow rate of fluid containing the nanoparticle suspension; a reliable, reproducible means of introducing a pulse of nanoparticle-PMAA solution into the flow system; and a magnet with a high field strength to permit nanoparticle capture. These considerations were met by using an HPLC pump (Kontron 422, Kontron Instruments, Milan, Italy), an electromechanical injection valve unit (VICI Valco, Houston, TX), and a 0.5-T magnet (e-magnets UK, Sheffield) held in position such that it was placed within a few millimeters of one side of the MCF. The apparatus diagram is shown in Figure 2.
Once the pump had been calibrated and the microscope (Intel Digital Blue QX3, Intel, Santa Clara, California) and light source positioned for optimal clarity, experiments were carried out on two sizes of nanoparticle aggregate, 330 nm and 580 nm, in two distinct phases. The first phase of experimentation was deposition. It was found that flow rates less than 1 mL min–1 were optimal for deposition; hence flow rates between 0.1 mL min–1 and 0.5 mL min–1 were examined in steps of 0.05 mL min–1. These volumetric flow rates correspond to fluid velocities of 1.3 cm s–1 and 6.3 cm s–1, respectively, with steps of 0.6 cm s–1. The Reynolds numbers corresponding to these flow velocities are 1.7 and 8.0, respectively, using a measured fluid viscosity of 3% by weight PMAA solution of 2.5 × 10–3 Pa s and a density of 1024 kg m –3 . The flow was, as expected, laminar. Additionally, the physical properties of the 3% by weight PMAA solution were close to those of blood, with values of blood viscosity reported about 3.3 × 10-3 Pa s–1 18 and density between 1043 kg m–3 and 1051 kg m–3.19 A pulse, approximately 2 mL in volume of 40 mg mL–1 nanoparticles in 3% by weight PMAA solution, was injected via the electromechanical valve into the established flow. This deposition phase was continued until one of two criteria were met; either a steady thickness of deposited nanoparticles was held for over an hour after the injected pulse, or the initial layer had subsequently been eroded by the flow. Erosion became the dominant behavior at the upper end of the flowrate range explored. The second phase of the experimental protocol was to observe the erosion of an established nanoparticle layer. The
22
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Figure 3. Sequence of images at (A) 0 minutes, (B) 7 minutes, (C) 32 minutes, and (D) 90 minutes showing the in-flow deposition of superparamagnetic magnetite nanoparticles. The cluster size of the nanoparticles was 580 nm, and the flow rate 0.1 mL min–1. The location of magnet is shown schematically by the cross-hatched rectangle in each image.
flow rate of fluid was incremented from the original deposition flow rate in steps of 0.05 mL min–1 such that a new steady-thickness nanoparticle layer was achieved for a period of 10 minutes at each flow rate. This procedure was repeated until the fluid flow completely eroded the nanoparticle layer. Results Deposition The sequence of optical micrographs shown in Figure 3 illustrates the observed nanoparticle deposition behavior where a steady-state nanoparticle layer thickness was achieved—in this case using 580-nm nanoparticles at a flow rate of 0.1 mL min–1. In this sequence of photographs the location of the capillary walls and of the magnet is shown schematically. Sequences of micrographs were taken for both 580-nm and 330-nm nanoparticles over the flow-rate range studied, and subsequent image analysis was used to quantify the thickness of the nanoparticle layer in millimeters. An estimate of the error in these measurements was also obtained by taking into account the uncertainty of the exact border between the nanoparticle layer and the fluid, and of the layer's unevenness. The series of plots shown in Figure 4 illustrate the deposition kinetics for 580-nm nanoparticles for flow rates
between 0.1 mL min–1 and 0.25 mL min–1. In all four of these cases a stable thickness of the nanoparticle layer was quickly reached. For flow rates above 0.3 mL min–1, corresponding to a fluid velocity of 3.8 cm s–1, a stable layer was not attained as the initially formed layer of nanoparticles underwent subsequent erosion. The sequence of plots shown in Figure 5 illustrates the deposition kinetics for the 330-nm nanoparticles. In this case, it was found that stable nanoparticle layers were formed at flow rates up to 0.35 mL min–1 (corresponding to a fluid velocity of 4.4 cm s–1), higher than the maximum flow rate that yielded a stable layer of 580-nm nanoparticles. Erosion The sequence of optical micrographs shown in Figure 6 illustrates the erosion of an established layer of nanoparticles with further increases in flow rate. In this particular case the layer of 580-nm nanoparticles was deposited at a flow rate of 0.1 mL min–1 until steady-state behavior was reached. Then, the flow rate was incremented in steps of 0.05 mL min–1 until a new steady-state thickness was established. It was observed that there was an upper flow-rate limit, dependent on nanoparticle diameter, for which a steady-thickness layer could be achieved before the layer was advected away with the flow. The plots shown in Figures 7 and 8 quantify this behavior for 580-nm and 330-nm nanoparticles, respectively. The plot with open symbols on each figure corresponds to
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
23
Figure 4. Plots showing the deposition kinetics of 580-nm nanoparticles for (A) 0.10 mL min–1 and 0.15 mL min–1, and (B) 0.20 mL min–1 and 0.25 mL min–1. The dashed horizontal line serves as a visual guide for the approximate location of the steady-state region.
the erosion behavior of a layer of nanoparticles deposited at 0.15 mL min–1, whereas the plot with closed symbols corresponds to both the deposition and erosion behavior at 0.5 mL min–1 since a steady-thickness layer was not achieved at this flow rate for either particle diameter. Discussion The results presented for the deposition of the 580-nm and 330-nm nanoparticles demonstrate that it is possible to capture superparamagnetic nanoparticles that are being advected in a fluid flow by using a magnetic field from a 0.5-T magnet adjacent to a plastic capillary. The thickness and stability of the captured layer of nanoparticles is strongly affected by both the fluid flow rate and the mean hydraulic
diameter of the nanoparticle aggregates. If the plots in Figures 4 and 5 are compared, then it is evident that the larger nanoparticles form a thicker layer, typically between 0.26 mm and 0.16 mm, for flow rates between 0.1 mL min–1 and 0.25 mL min–1, respectively. In contrast, the 330-nm nanoparticles tend to form a more consistently, smaller sized layer, typically between 0.12 mm and 0.09 mm for flow rates of 0.1 mL min–1 and 0.35 mL min–1, respectively. From the optical observations reported in this article it is clear that the magnetic field is capable of attracting the nanoparticles to the walls of the capillary and that the thickness of the nanoparticle layer that is formed is sensitive to flow conditions. A possible explanation for the smaller deposited layer of 330-nm nanoparticles is that each particle contains a smaller
24
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Figure 5. Plots showing the deposition kinetics of 330-nm nanoparticles for (A) 0.10 mL min–1 and 0.20 mL min–1, and (B) 0.30 mL min–1 and 0.35 mL min–1. The dashed horizontal line serves as a visual guide for the approximate location of the steady-state region.
mass of magnetic material when compared with a nanoparticle of 580-nm diameter; the smaller particles, therefore, experience a weaker force as a result of the applied magnetic field, hence a smaller amount are likely to be captured from the fluid flow. The 580-nm nanoparticles, however, contain a larger amount of magnetic material, and hence they experience a larger force due to the applied magnetic field, resulting in a greater amount being captured. In terms of the erosion behavior of the nanoparticles at increasing flow rates, the observation that the layer of 580-nm nanoparticles is unable to be maintained at flow rates at which the 330-nm nanoparticle layer is still present could be explained qualitatively in terms of the applied shear stress on the particle layer due to the fluid flow. The thicker layer of
nanoparticles results in a more constricted flow, hence the flow velocity in the capillary is higher for a given volumetric flow rate. This increased flow velocity, in turn, leads to higher shear stresses on the nanoparticle surfaces and hence a greater driving force to erode the particles. The thinner layer of nanoparticles, on the other hand, results in less flow constriction, and therefore lower shear stresses are present on the particles. In this way the layer of 330-nm nanoparticles can exist at higher volumetric flow rates when compared with the layer of 580-nm nanoparticles. The hypothesis, therefore, is twofold. First, the amount of particles captured, hence the final thickness of the layer that forms, is dependent on both the magnetic force and the hydrodynamic force experienced by the particle. When
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
25
Figure 6. Sequence of images showing the erosion of a pre-established nanoparticle layer, deposited at 0.1 mL min–1, with increasing flow rate. Image (A) corresponds to the initial steady-state layer when the flow rate was first increased, (B) 25 minutes later, (C) 27 minutes later, and (D) 30 minutes later. The cluster size of the nanoparticles was 580 nm, and the flow rate 0.1 mL min–1. The location of magnet is shown schematically by the cross-hatched rectangle in each image.
magnetic forces dominate the hydrodynamic forces, then a particle is captured. However, when hydrodynamic forces become dominant when compared with the magnetic forces, the nanoparticle layer is eroded by the flow. This lessens the flow constriction caused by the nanoparticle layer on the capillary wall and hence lowers the hydrodynamic forces; thus a new equilibrium is achieved. These qualitative explanations can also account for the transient behavior that can be observed in Figures 7 and 8. These two figures present, at first sight, seemingly contradictory trends. The thickness of the layer of 580-nm nanoparticles can be seen to decrease with increasing flow rate until either a new steady-state thickness is formed or the layer erodes away completely. Given the proposed hypothesis, the trend of erosion displayed by the 580-nm nanoparticles with increasing flow rate is expected. In contrast, the layer of 330-nm nanoparticles can be seen at first to increase with increasing flow rates, before becoming steady and then finally eroding away. A proposed mechanism for this observation is that the smaller amount of magnetic material present in the 330-nm nanoparticles results in lower magnetic forces and hence fewer particles from the 2-mL pulse being captured when the initial layer was formed. This layer results in only a small amount of constriction of the fluid flow, not enough for hydrodynamic forces to become dominant. If any additional nanoparticles were present in the flow, they would still therefore be
captured. When the flow rate is increased, nanoparticles captured outside the field of view of the microscope, in an area of weaker magnetic field further away from the magnet, are eroded but then subsequently recaptured onto the nanoparticle layer in the field of view. This leads to the observation that the thickness of the nanoparticle layer at first increases with an increase in volumetric flow rate. This increase is enough to reach equilibrium between magnetic and hydrodynamic forces; thus, subsequent increases in flow rate serve to erode the nanoparticle layer away. Further insight into this mechanism can be had by calculating an estimate of the shear stress on the top of the nanoparticle layer for a given flow rate. This has been done by approximating the area available for fluid flow as a large segment of the circular capillary cross section, whereas the remaining, smaller segment of the capillary represents the nanoparticle layer. An equivalent diameter is then found for the segment containing the flow and the shear stress calculated by assuming that the flow is Poiseuille flow of a Newtonian fluid through a circular capillary with solid walls of the equivalent diameter that had been calculated; this is shown schematically in Figure 9. Despite this approach making some first-order approximations about the nature of the boundary condition on the nanoparticle layer and also the shape of the flow geometry, a reasonable estimate of the shear stress on the nanoparticle layer can be obtained; this estimate is of a similar magnitude to shear stresses reported
26
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
Figure 7. Plot of erosion kinetics for 580-nm nanoparticles for increasing erosion flow rates where the initial deposition flow rates were 0.15 mL min–1 (open symbols) and 0.5 mL min–1 (closed symbols).
Figure 8. Plot of erosion kinetics for 330-nm nanoparticles for increasing erosion flow rates where the initial deposition flow rates were 0.15 mL min–1 (open symbols) and 0.5 mL min–1 (closed symbols).
for the onset of erosion of biomass in filtration systems.20 Figure 10 shows the nanoparticle layer thickness that was detailed in Figures 7 and 8 replotted as a function of the shear
stress exerted on it by the fluid flow for the case where both 330-nm and 580-nm nanoparticles were initially deposited at 0.15 mL min–1 and then subjected to flow-rate increments of
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
27
Figure 9. Schematic diagram showing the transformation of the real flow domain into one suitable for the estimation of the maximum fluid shear stress on the top surface of the layer of captured nanoparticles.
Figure 10. Plot of the calculated shear stress on the top surface of the nanoparticle layer as a function of layer thickness for 330-nm nanoparticles (open symbols, dotted line) and for 580-nm nanoparticles (closed symbols, dashed line). The initial deposition flow rate was 0.15 mL min–1 in both cases, and the lines marked on the plot are for visual guidance only.
0.05 mL min–1 until the layer had fully eroded away. The shape of the two trends shown on this plot adds credence to the proposed hypothesis; the applied stress on the 330-nm nanoparticle layer first increases markedly before erosion, suggesting that particle recapture may be occurring as a result of poor initial capture. The applied stress on the
580-nm nanoparticle layer, however, is seen to only increase by a small amount, whereas initial particle erosion rather than capture is occurring before rapid erosion of the layer takes place. It is interesting to note that the stress at which rapid erosion occurs is different in both cases, suggesting that its
28
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29
underlying mechanism is complex, possibly involving variables such as nanoparticle packing, how the nanoparticles were initially deposited, and their interaction with the surface of the film and with each other. Further investigation is needed to probe this behavior such that fundamental understanding of this phenomenon can be gained. It must be noted that the environment in which these deposition and erosion experiments have been carried out is a highly simplified model of a circulatory system. The regime of flow is one of constant flow rate rather than the constant pressure drop that one would expect to find across a capillary network, and the fluid carrying the nanoparticles does not contain any particulate matter such as blood cells. The physics underlying the interaction between the hydrodynamic and magnetic forces acting on an immobilized layer of nanoparticles, however, is similar, and the experiments detailed in this article allow one to gain fundamental insight into this system. Future work will involve repeating these experiments in an environment of constant pressure drop that will be more analogous to the regime found in a circulatory system. In this work superparamagnetic nanoparticles have been successfully captured in microcapillaries of 410 μm in diameter in a flow of up to 0.3 mL s–1, equivalent to a velocity of 3.8 cm s–1. This is of particular relevance to therapeutic nanoparticle targeting, because this is the blood velocity expected in a medium-sized human venule or arteriole.21 The measured viscosity of the 3% by weight PMAA solution used in our capture experiments of 2.5 × 10–3 Pa s–1 is also closer than water to that of blood, about 3.3 × 10–3 Pa s–1.18 The effectiveness of nanoparticle capture supports the successful results of magnetic targeting of intravenously7,8 and intra-arterially9 administered nanoparticles in initial clinical trials. Interestingly, further analysis of nanoparticle deposition layer thickness and stability indicates that these properties are dependent on nanoparticle size. The larger 580-nm particles deposit to form a thicker layer quicker that is more susceptible to subsequent erosion. The smaller 330-nm particles initially deposit to form a more stable thinner layer that increases in thickness as flow rates are increased until a critical flow rate is reached and erosion begins. It has been hypothesized in this article that this observed behavior is due to the effect of shear stress on the nanoparticle layer. The shear stress increases as the capillary flow is constricted, causing the nanoparticle layer to reach an equilibrium where the force of the magnetic attraction on the nanoparticles, proportional to particle size, balances the shear stress exerted by the fluid. This would also explain why complete blocking of the capillary with nanoparticles was reassuringly not observed. In future clinical work, nanoparticle agglomerates of between 300 and 600 nm in size could be used for delivery of bound drug or gene-therapy vector to sites of disease within the body. Nanoparticle agglomerates in this size range have
the advantage that they can be trapped and manipulated at the site of disease by inexpensive permanent magnets as small as 0.5 T but are still superparamagnetic. Once magnetically delivered to the site of interest, the nanoparticles may cross into the cells as 10-nm nanoparticles by pinocytosis or as larger agglomerates by phagocytosis to deliver their payload. These findings will provide a useful basis for future nanoparticle design for effective magnetic targeting of therapeutics in vivo. References 1. Pankhurst QA, Connolly J, Jones SK, Dobson J. Applications of magnetic nanoparticles in biomedicine. J Phys D Appl Phys 2003;36: R167-81. 2. Senyei A, Widder K, Czerlinski G. Magnetic guidance of drug-carrying microspheres. J Appl Phys 1978;49:3578-83. 3. Widder KJ, Morris RM, Poore GA, Howard DP, Senyei AE. Selective targeting of magnetic albumin microspheres containing low-dose doxorubicin-total remission in Yoshida sarcoma-bearing rats. Eur J Cancer Clin Oncol 1983;19:135-9. 4. Fernandez-Pacheco R, Marquina C, Valdivia JG, Gutiérrez M, Romero MS, Cornudella R. Magnetic nanoparticles for local drug delivery using magnetic implants. J Magn Magn Mater 2007;311:318-22. 5. Hughes C, Galea-Lauri J, Farzaneh F, Darling D. Streptavidin paramagnetic particles provide a choice of three affinity-based capture and magnetic concentration strategies for retroviral vectors. Mol Ther 2001;3:623-30. 6. Morishita N, Nakagami H, Morishita R, et al. Magnetic nanoparticles with surface modification enhanced gene delivery of HVJ-E vector. Biochem Biophys Res Commun 2005;334:1121-6. 7. Lubbe AS, Bergemann C, Brock J, McClure DG. Physiological aspects in magnetic drug-targeting. J Magn Magn Mater 1999;194:149-55. 8. Lubbe AS, Bergemann C, Riess H, et al. Clinical experiences with magnetic drag targeting: a phase I study with 4′-epidoxorubicin in 14 patients with advanced solid tumors. Cancer Res 1996;56:4686-93. 9. Wilson MW, Kerlan RK, Fidelman NA, et al. Hepatocellular carcinoma: regional therapy with a magnetic targeted carrier bound to doxorubicin in a dual MR imaging/conventional angiography suite-initial experience with four patients. Radiology 2004;230:287-93. 10. Gupta AK, Gupta M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 2005;26: 3995-4021. 11. Duan HL, Shen ZQ, Wang XW, Chao FH, Li JW. Preparation of immunomagnetic iron-dextran nanoparticles and application in rapid isolation of E.coli O157:H7 from foods. World J Gastroenterol 2005;11: 3660-4. 12. Mendenhall GD, Geng Y, Hwang J. Optimization of long-term stability of magnetic fluids from magnetite and synthetic polyelectrolytes. J Colloid Interface Sci 1996;184:519-26. 13. Kikura H, Matsushita J, Kakuta N, Aritomi M, Kobayashi Y. Cluster formation of ferromagnetic nano-particles in micro-capillary flow. J Mater Process Tech 2007;181:93-8. 14. Hallmark B, Gadala-Maria F, Mackley MR. The melt processing of polymer microcapillary film (MCF). J Non-Newton Fluid 2005;128: 83-98. 15. Hallmark B, Mackley MR, Gadala-Maria F. Hollow microcapillary arrays in thin plastic films. Adv Eng Mater 2005;7:545-7. 16. Xia ZF, Wang GB, Tao KX, Li JX. Preparation of magnetite-dextran microspheres by ultrasonication. J Magn Magn Mater 2005;293: 182-6. 17. Hornung CH, Hallmark B, Hesketh RP, Mackley MR. The fluid flow and heat transfer performance of thermoplastic microcapillary films. J Micromech Microeng 2006;16:434-47.
N.J. Darton et al. / Nanomedicine: Nanotechnology, Biology, and Medicine 4 (2008) 19–29 18. Lowe G, Rumley A, Norrie J, et al. Blood rheology, cardiovascular risk factors, and cardiovascular disease: The West of Scotland Coronary Prevention Study. Thromb Haemostasis 2000;84:553-8. 19. Hinghofer-Szalkay HG, Greenleaf JE. Continuous monitoring of blood volume changes in humans. J Appl Physiol 1987;63:1003-7.
29
20. Bustnes TE, Kaminski CF, Mackley MR. The capture and release of biomass in a high voidage fibrous microstructure: mechanisms and shear stress levels. J Membrane Sci 2006;276:208-20. 21. Sherwood L. Human physiology from cells to systems. New York: Wadsworth; 1997.