12632
J. Phys. Chem. B 2010, 114, 12632–12640
Thermodynamic and Kinetic Stability of DSPE-PEG(2000) Micelles in the Presence of Bovine Serum Albumin Mark Kastantin,† Dimitris Missirlis,† Matthew Black,† Badriprasad Ananthanarayanan,† David Peters,‡,| and Matthew Tirrell*,†,§,⊥ Department of Chemical Engineering, Vascular Mapping Center, Burnham Institute for Medical Research, and Materials Research Laboratory, UniVersity of California, Santa Barbara, California 93106, Biomedical Sciences Graduate Group, UniVersity of California, San Diego, California 92037, and Department of Bioengineering, UniVersity of California, Berkeley, California 94720 ReceiVed: January 8, 2010; ReVised Manuscript ReceiVed: July 21, 2010
This work investigated the stability of DSPE-PEG(2000) micelles in the presence of bovine serum albumin (BSA). DSPE-PEG(2000) was found to exist in equilibrium among monomeric, micellar, and BSA-bound states, and this equilibrium shifted toward the BSA-bound state when the temperature increased from 20 to 37 °C. The micellar state is thermodynamically unstable at both temperatures when the concentration of BSA approaches that of DSPE-PEG(2000), and micelle breakup occurs with a first-order time constant of 130 ( 9 min at 20 °C and 7.8 ( 1.6 min at 37 °C. Thus, previous targeting experiments that demonstrate synergistic effects in multiply functionalized DSPE-PEG(2000) micelles are likely due to targeting that occurs on a timescale faster than that of micelle breakup. Micelle breakup was limited by diffusion at 20 °C whereas at 37 °C monomer desorption from the micelle was the rate-limiting step. These findings give clear guidance concerning the lifetimes of micelles that may be used as diagnostic and therapeutic nanoparticles. Introduction A goal of nanomedicine is the facile creation of nanoparticles capable of sophisticated function once introduced into systemic circulation. In this spirit, mixed micelles of peptide amphiphiles have recently demonstrated promise for the targeted delivery of therapeutic and imaging moieties to pathological tissue. In one example, a tumor-targeting peptide-PEG-lipid conjugate was used to direct the delivery of fluorescent micelle components to tumor tissue and to facilitate cellular internalization.1 In another, a mixed micelle containing a plaque-targeting peptidePEG-lipid conjugate was able to increase the efficacy of thrombin-inhibiting hirulog peptide-PEG-lipid conjugates at the surface of vulnerable atherosclerotic plaques.2 In these examples, the observation of both targeting and biofunctional (i.e., imaging or therapeutic) micelle properties demonstrates that mixed micelles were delivered intact to the target tissue. While the examples above used different targeting peptides to reach different tissue and deliver different payloads, they shared a common structural foundation: the PEG-lipid micelle built from self-assembled PEG-lipid monomers. PEG-lipids used in the above work consisted of PEG(2000) covalently linked to a distearoyl lipid tail to form the DSPE-PEG(2000) monomer. Functional groups at the free end of PEG were used to conjugate peptides or dyes to the PEG-lipid. The DSPE lipid tail provides a strong hydrophobic driving force for self-assembly, and the bulky PEG(2000) headgroup imparts sufficient curvature to the core-corona interface to form oblate spheroidal micelles with * Corresponding author. E-mail:
[email protected]. † Department of Chemical Engineering, University of California, Santa Barbara. ‡ Vascular Mapping Center, Burnham Institute for Medical Research, University of California, Santa Barbara. § Materials Research Laboratory, University of California, Santa Barbara. | University of California, San Diego. ⊥ University of California, Berkeley.
a maximum diameter of 18 nm.3 Aggregation numbers near 90 allow for a high-density multivalent display of peptides that can increase the biological response relative to that of the presentation of a single peptide.4-6 In addition, this construct has thus far proven to decouple micelle geometry from the type of functional molecules conjugated to the PEG-lipid.1,2 That is, the micellar structure is independent of the peptide-functionalized PEG-lipid, or mixtures of functionalized PEG-lipids, employed over a wide range. This modularity provides a pathway to small spheroidal micelles with controllable, arbitrary compositions simply by mixing different monomers prior to selfassembly. The ability to create a wide array of micelles easily, which display multiple biologically relevant molecules in high density, is a useful property of self-assembled nanoparticles. However, no covalent bonds form after self-assembly to stabilize the resulting PEG-lipid micelle. On one hand, this may provide favorable toxicity properties because disassembly may minimize entrapment in untargeted tissue. On the other hand, micelle breakup would eliminate multivalency and some of the synergistic benefits created by the self-assembly process. One of the primary functions of albumin, which is prevalent in blood serum, is to transport lipids and fatty acids so that these amphiphilic molecules do not form undesirable self-assembled structures. Indeed, human serum albumin (HSA) contains up to 11 binding sites for fatty acids.7,8 Although an equilibrium exists among monomers in solution, micelles, and albumin-bound monomers, it is unclear whether equilibrium favors the presence of micelles in the previously cited in vivo experiments. Given that the typical postinjection concentration of PEG-lipid (∼30-50 µM) is much less than that of serum albumin in blood (∼600 µM), nonspecific adsorption to albumin poses a significant concern for the stability of DSPE-PEG(2000) micelles.
10.1021/jp1001786 2010 American Chemical Society Published on Web 09/09/2010
Stability of DSPE-PEG(2000) Micelles Previous work by Castelletto and colleagues has interpreted dynamic light scattering and small-angle X-ray scattering data on complexes of DSPE-PEG(2000) and bovine serum albumin (BSA) to conclude that micelles of the PEG-lipid do not form in the presence of BSA.9,10 This conclusion indicates the extremely strong preference of a DSPE-PEG(2000) monomer for the BSA-bound state relative to the micellized state and, by extension, implies that mixed micelles are unsuitable for targeted drug delivery. This statement is seemingly at odds with the previously discussed successful examples of tissue-targeting behavior in suspensions of mixed micelles in vivo. There are several explanations that can reconcile these two observations. First, the transport of naturally occurring lipids by albumins in the bloodstream may inhibit PEG-lipid binding to albumin. Second, micelle breakup may occur, and the observed results1,2 may be the result of multifunctionalized serum albumins. Third, Castelletto and colleagues may have overlooked the presence of micelles coexisting with BSA because the scattering techniques they employ do not explicitly rule out a heterogeneous population of separate BSA and PEG-lipid micelles. Given that BSA, like HSA, has a finite number of binding sites11-13 for lipids, it is unlikely that BSA prevents micelle formation at all ratios of BSA to DSPE-PEG(2000). Fourth, targeting may occur on a timescale that is faster than that for micelle breakup. That is, the kinetics of micelle breakup may be sufficiently slow that micelles can reach targeted tissue intact. To evaluate these possibilities, this work uses fluorescence quenching studies to distinguish quantitatively between micellized and monomeric or BSA-bound, fluorescently labeled PEGlipids. This technique entails labeling the free PEG end of PEGlipids with fluorophores that self-quench when in close proximity to one another as was done in previous work to quantify the activation barrier for monomer desorption from a micelle.14 Fluorescence from PEG-lipids in the micellar state is quenched but recovers as the micelle breaks apart. From these experiments, it is possible to extract kinetic and equilibrium parameters regarding the preference of a PEG-lipid monomer to micellize or to complex with BSA, which enables a better understanding of the processes that affect the survival of circulating micelles in vivo. Materials and Methods Materials. 1,2-Distearoyl-sn-glycero-3-phosphoethanolamineN-[methoxy(poly(ethylene glycol))-2000] and 1,2-distearoyl-snglycero-3-phosphoethanolamine-N-[amino(poly(ethylene glycol))2000] (DSPE-PEG(2000) and DSPE-PEG(2000)-amine, respectively) were purchased from Avanti Polar Lipids and used without further purification. NHS-fluorescein and NHS-rhodamine were purchased from Pierce Protein Research Products. Bovine serum albumin (essentially fatty-acid-free) was purchased from Sigma-Aldrich. Phosphate-buffered saline (PBS) (10×, without calcium chloride or magnesium chloride) was purchased from Invitrogen and diluted 10-fold using water purified by reverse osmosis to 18.2 MΩ cm. Human plasma with sodium citrate was obtained through Innovative Research from healthy donors aged 18-65 at FDA licensed facilities. Experimental Details. Synthesis of Dye-Conjugated DSPEPEG(2000) Lipids. Fluorescein-labeled and rhodamine-labeled DSPE-PEG(2000), denoted by DSPE-PEG(2000)-FAM and DSPE-PEG(2000)-Rhod, respectively, were created by conjugating the dyes to DSPE-PEG(2000)-amine via a peptide bond. A 3-fold molar excess of NHS-fluorescein or NHS-rhodamine was added to the DSPE-PEG(2000)-amine and dissolved in a
J. Phys. Chem. B, Vol. 114, No. 39, 2010 12633 10 mM aqueous phosphate buffer (pH 7.4) containing 10% methanol by volume. After reaction at 4 °C for 8 h, the mixture was purified by reverse-phase high-performance liquid chromatography using 0.1% TFA in acetonitrile-water mixtures on a C8 preparatory column at 60 °C and characterized by electrospray ionization time-of-flight mass spectrometry. DSPEPEG(2000)-FAM was used for experiments in PBS or BSA, but DSPE-PEG(2000)-Rhod was used for experiments in human plasma because the longer excitation wavelength is absorbed and scattered less by proteins in blood and its greater molar extinction coefficient provides a brighter signal relative to that of DSPE-PEG(2000)-FAM. Micelle Preparation. Micelles were prepared in a glass culture tube by dissolving the monomers in chloroform and evaporating the solution under nitrogen. To vary the labeling fraction in the PEG-lipid population, unlabeled DSPE-PEG(2000) was mixed in with fluorescent DSPE-PEG(2000) before solvent evaporation. The resulting film was dried under vacuum for 8 h and rehydrated in 1× PBS at 40 °C (pH 7.4). PEG-lipid samples were incubated at 40 °C for 60 min and transferred to a sonicating bath at 40 °C for 60 min. Samples were allowed to cool to room temperature for 3 h and subsequently filtered through a 220 nm syringe filter. To verify the presence of micelles, dynamic light scattering was performed at 20 °C with a measurement angle of 90° using an avalanche photodiode detector attached to a goniometer (Brookhaven Instruments) and a 632.8 nm HeNe laser (Melles Griot). At this temperature, the micelle hydrodynamic diameter was reproducibly 13.5 ( 0.3 nm, independent of the fluorophore on the micelle. Kinetics of Micelle Breakup. Stock solutions of DSPEPEG(2000)-FAM, DSPE-PEG(2000)-Rhod, and DSPEPEG(2000) micelles were prepared alongside a stock of BSA in PBS or human plasma, and all solutions were incubated at a measurement temperature of either 20 or 37 °C for 60 min. Three different mixing experiments were performed. The first experiment was similar to that in previous work,14 where unlabeled micelles were added to DSPE-PEG(2000)-FAM micelles in a final acceptor/ donor ratio of 36:1 and the fluorescence was tracked over time. In the second experiment, BSA replaced DSPE-PEG(2000) micelles as the acceptor species for DSPE-PEG(2000)-FAM monomers. BSA was added to labeled micelles to give a final BSA concentration of 2.04 mM and a final DSPE-PEG(2000)-FAM concentration of 20 µM. In the third experiment, one part of the DSPE-PEG(2000)-Rhod stock was added to 99 parts of human plasma to give a final DSPE-PEG(2000)-Rhod concentration of 20 µM. The transient increase in fluorescence intensity in each experiment was used to obtain the first-order time constant for micelle breakup. The fluorescence intensity was recorded every 0.1 s, and under this condition, photobleaching was found to occur with a time constant greater than 1000 min for both fluorescein and rhodamine. Equilibrium Partitioning Measurement. Mixtures were made of BSA, DSPE-PEG(2000)-FAM, and DSPE-PEG(2000) in PBS and incubated at either 20 °C for 8 h or 37 °C for 2 h. Different incubation times were used because DSPE-PEG(2000) micelles equilibrate slowly at 20 °C14 whereas visible BSA aggregation was observed after 8 h at 37 °C. Thus these equilibration times were chosen to ensure that each system was at equilibrium while minimizing BSA aggregation. The DSPE-PEG(2000)-FAM concentration was kept constant in all solutions at 30 µM so that the fluorescence intensity could be directly compared between solutions. The BSA concentration was varied from 0 to 300 µM. Above this maximum BSA
12634
J. Phys. Chem. B, Vol. 114, No. 39, 2010
Kastantin et al.
concentration, solution turbidity became a strong effect in the intensity measurement with the high-molecular-weight protein absorbing and scattering both excitation and emission light. This masked the signal change due to DSPE-PEG(2000)-FAM partitioning between micelle and BSA-bound states. BSA autofluorescence was also found to be negligible relative to that of quenched micelles and was consequently neglected. To vary the stoichiometry of the PEG-lipid interaction with BSA at a constant DSPE-PEG(2000)-FAM concentration, DSPE-PEG(2000) was added at a concentration of 0, 15, or 30 µM. This increased the fluorescence intensity of the micellar state, Imicelle, by providing more space between fluorophores on a micelle and decreasing self-quenching. The increase in micelle fluorescence intensity decreased the contrast between micellebound fluorophores and BSA-bound fluorophores or free monomers, but not to a degree that inhibited data modeling. One set of solutions was also prepared with a DSPEPEG(2000) concentration of 300 µM to create a fluorescent micelle with effectively no quenching. In this case, there should be no difference in fluorescence among the free monomer, the BSA-bound fluorophore, and micelle-bound fluorophore, providing a way to calibrate (using only eq 1) the system for BSA turbidity at all concentrations. This calibration provided only a minor (
