LaPO4 Nanoparticles Doped with Actinium-225 that Partially

ARTICLE pubs.acs.org/bc

LaPO4 Nanoparticles Doped with Actinium-225 that Partially Sequester Daughter Radionuclides Jonathan Woodward,† Stephen J. Kennel,‡ Alan Stuckey,‡ Dustin Osborne,‡ Jonathan Wall,‡ Adam J. Rondinone,† Robert F. Standaert,† and Saed Mirzadeh*,† † ‡

Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States Graduate School of Medicine, University of Tennessee, Knoxville, Tennessee 37920, United States

bS Supporting Information ABSTRACT: Nanoscale materials have been envisioned as carriers for various therapeutic drugs, including radioisotopes. Inorganic nanoparticles (NPs) are particularly appealing vehicles for targeted radiotherapy because they can package several radioactive atoms into a single carrier and can potentially retain daughter radioisotopes produced by in vivo generators such as actinium-225 (225Ac, t1/2 = 10 d). Decay of this radioisotope to stable bismuth-209 proceeds through a chain of short-lived daughters accompanied by the emission of four R-particles that release >27 MeV of energy. The challenge in realizing the enhanced cytotoxic potential of in vivo generators lies in retaining the daughter nuclei at the therapy site. When 225Ac is attached to targeting agents via standard chelate conjugation methods, all of the daughter radionuclides are released after the initial R-decay occurs. In this work, 225Ac was incorporated into lanthanum phosphate NPs to determine whether the radioisotope and its daughters would be retained within the dense mineral lattice. Further, the 225Ac-doped NPs were conjugated to the monoclonal antibody mAb 201B, which targets mouse lung endothelium through the vasculature, to ascertain the targeting efficacy and in vivo retention of radioisotopes. Standard biodistribution techniques and microSPECT/CT imaging of 225Ac as well as the daughter radioisotopes showed that the NPs accumulated rapidly in mouse lung after intravenous injection. By showing that excess, competing, uncoupled antibodies or NPs coupled to control mAbs are deposited primarily in the liver and spleen, specific targeting of NP-mAb 201B conjugates was demonstrated. Biodistribution analysis showed that ∼30% of the total injected dose of La(225Ac)PO4 NPs accumulated in mouse lungs 1 h postinjection, yielding a value of % ID/g >200. Furthermore, after 24 h, 80% of the 213Bi daughter produced from 225Ac decay was retained within the target organ and 213Bi retention increased to ∼87% at 120 h. In vitro analyses, conducted over a 1 month interval, demonstrated that ∼50% of the daughters were retained within the La(225Ac)PO4 NPs at any point over that time frame. Although most of the γ-rays from radionuclides in the 225Ac decay chain are too energetic to be captured efficiently by SPECT detectors, appropriate energy windows were found that provided dramatic microSPECT images of the NP distribution in vivo. We conclude that La(225Ac)PO4
mAb 201B conjugates can be targeted efficiently to mouse lung while partially retaining daughter products and that targeting can be monitored by biodistribution techniques and microSPECT imaging.

’ INTRODUCTION Recently, there has been widespread interest in biomedical applications of nanomaterials for diagnostic and therapeutic purposes. Various inorganic nanoparticles (NPs) have been coupled to antibodies, and the resulting conjugates have been shown to act as targeted drug delivery systems or imaging agents.1
8 Although monitoring the biodistribution of these NPs has been dominated by fluorescence imaging,9
17 a few quantitative studies using radioactive NPs have been published.9,18
21 Radionuclides with specific emission properties can be incorporated into NPs and used for radioimmunotherapy and radioimaging. Nanoparticles are particularly well-suited for the delivery of R- or β
-emitters for radioimmunotherapy. One advantage is the potential for containment of several radioactive atoms within a single carrier. This feature is useful when a target tissue has only a few receptors per cell, limiting the dose that can be delivered.22 A second potential r 2011 American Chemical Society

advantage is the sequestration of daughter radioisotopes from in vivo generators such as 225Ac.22
28 In vivo generators utilize parent radioisotopes that decay to radioactive daughters which themselves decay quickly, potentially amplifying the dose at a target site. One such generator is 225Ac, which has four short-lived R-emitting daughters (221Fr, t1/2 = 4.9 min; 217At, t1/2 = 32.3 ms; 213 Bi, t1/2 = 46.8 min; 209Pb, t1/2 =3.25 h) in two decay chains leading to stable 209Bi (Figure 1).23,24 Decay through these chains releases >27 MeV of energy, ∼90% of which is associated with four sequentially emitted R-particles with individual energies ranging from 5.7 to 8.4 MeV.23,29,30 Retaining the decay daughters of in vivo radioisotope generator systems at the site of targeting would Received: December 14, 2010 Revised: February 28, 2011 Published: March 24, 2011 766

dx.doi.org/10.1021/bc100574f | Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

Figure 1. 225Ac decay scheme. Decay of 225Ac to stable 209Bi releases >27 MeV of energy, ∼90% of which is associated with four sequentially emitted Rparticles with individual energies ranging from 5.7 to 8.4 MeV. Adapted from refs 23 and 24.

improve the specificity of the delivered absorbed dose by limiting nontarget toxicity.29
34 In this paper, we detail the synthesis and biofunctionalization of LaPO4 (monazite) NP doped with 225Ac [La(225Ac)PO4] to study the ability of metal phosphate nanomaterials to target biological epitopes and sequester radionuclides and daughter products, minimizing their release in biological systems. A model system of vascular targeting was used to evaluate NPs performance in vivo. This system has numerous advantages for testing the stability of radioconjugates and the retention of daughter radioisotopes from in vivo generators.35,36 First, the radiolabeled monoclonal antibody 201B to murine thrombomodulin (TM) efficiently targets and delivers >30% of the total dose to lung endothelium, the major site of TM expression.37,38 Targets on the surface of vasculature or circulating cells are necessary for efficient delivery of large constructs such as NPs 18,19 or liposomes,39 since they do not effectively exit the vascular space.21,40
42 Second, targeting with this antibody occurs within minutes of injection and depends on antibody specificity, not trapping of aggregated particles.18,19 Third, radio-iodinated (125I) mAb 201B is retained in lung with a half-life of 2
4 days,38,39,43 allowing long-term studies. These properties permitted the first demonstration of specific targeting of the 225Ac-doped NPs and determination of the fate of daughter radioisotopes (primarily 213Bi) at various times after delivery. Through both standard biodistribution studies and microSPECT/CT (single photon emission computed tomography/X-ray computed tomography) imaging, we found that La(225Ac)PO4 NPs attached to a specific antibody can be targeted rapidly and efficiently to lung endothelium and that a significant fraction (∼80%) of the daughter radionuclides are retained at the targeting site within 24 h. We conclude that LaPO4 NPs have the potential to improve radioimmunotherapeutic efficacy of 225Ac by effectively targeting the appropriate site while partially retaining daughter nuclides and that the distribution of 225 Ac and daughter radionuclides can be monitored by biodistribution techniques and microSPECT imaging.

weight cutoff of 10 (or 25) kDa. The dialysis tubing was washed thoroughly of all preservatives before use. Centrifugation was performed using an Eppendorf MiniSpin Plus (radius = 6 cm) in 1.5 mL polypropylene microcentrifuge tubes. Sonication was performed with a Branson microprobe sonicator at power setting 1. BALB/c female mice were obtained from Taconic, Germantown, NY. The monoclonal antibodies mAb 201B and mAb 33 were prepared and purified as previously described.38,39 Clodronate-containing liposomes were obtained from the Department of Molecular Cell Biology VUmc, FdG, PO Box 7057, 1007 MB Amsterdam, The Netherlands. Radioactivity Measurements. Gamma-spectrometry for in vitro measurements was performed using a calibrated intrinsic Ge detector (crystal active volume ∼100 cm3) and PC-based Multichannel Analyzer (MCA, Canberra). The detector has a resolution of 0.8 at 5.9 keV, 1.0 at 123 keV, and 1.9 at 1332 keV. Energy and efficiency calibrations were determined with γ-ray sources traceable to the National Institute of Standards and Technology. The γ-ray energies and their absolute intensities in parentheses used for determination of 225Ac were 218.0 keV (11.58%) and 440.4 keV (26.1%) from the decay of 221Fr (t1/2 = 4.9 min) and 213 Bi (t1/2 = 46.8 min), respectively. An automated γ-ray scintillation counter consisting of a welltype NaI (Tl) detector (Perkin-Elmer Wizard model 1480) was used for the biodistribution studies detailed below. The twochannel discriminator windows were set to bracket the 218 keV (from 221Fr) in the lower and 440 keV (from 213Bi) in the upper energy windows. Organs (lungs, liver, spleen, and kidneys) harvested for the in vivo release of 213Bi experiment were counted ∼20 min after collection to ensure 99% equilibrium between 221 Fr and 225Ac. The count rates in the lower window (221Fr) and upper window (213Bi) were indicative of 225Ac and 213Bi activities in the samples, respectively. Accordingly, the 213Bi activities were corrected for decay during the 20 min waiting period. Alternatively, to ensure equilibrium between 213Bi and 225Ac, the same samples were recounted after 3 h and the count rates in the upper window were used for the equilibrium value of 213Bi activity. The two approaches yielded similar results. Preparation of 225AcCl3. Actinium-225 was separated from a mixture of 225Ra and 229Th as described earlier.44,45 The purified 225 Ac was evaporated to near-dryness on a hot-plate under an IR heat lamp, followed by addition of 0.5 mL of conc. HNO3 and 3 drops of 30% H2O2 to the residue, and gentle evaporation of the solution to dryness. The unobservable residue was dissolved in 0.2 mL of 0.1 M HCl to afford 225AcCl3 (with a negligible content of nitrate). Assay of a 5 μL aliquot of the purified 225 Ac sample (5 mCi) indicated the presence of 98%, TCI), hexadecane (g99%, Aldrich), hexane (HPLC grade, 99.9%, Fisher), anhydrous tetrahydrofuran (g99.9%, Aldrich), 6-aminohexanoic acid (g99%, Aldrich), NaOH (99.99%, Aldrich), Na2HPO4 3 2H2O (g99%, Fluka), NaH2PO4 3 H2O (g99.5%, Fluka), and tris(hydroxymethyl)aminomethane hydrochloride (tris-HCl, g99%, Fluka). All other chemical reagents used (HNO3, HCl, and H2O2) were of analytical grade or better. Nanoparticle purification was performed by dialysis using Spectra/Por regenerated cellulose membranes with a molecular 767

dx.doi.org/10.1021/bc100574f |Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

dissolved in 200 μL of 0.1 M HCl containing 5.4 mCi (0.41 nmol) of 225AcCl3, and the solution was evaporated to dryness. An additional 200 μL aliquot of 0.1 M HCl was added to redissolve the La3þ/225Ac3þ mixture, and again the solution was evaporated to dryness. The solid was subsequently dissolved in a mixture of 75 μL of tri-n-butyl phosphate and 75 μL of phenyl ether and heated to 100 °C under vacuum with stirring. Residual water was removed by three cycles of evacuating for 30 min followed by backfilling with argon gas at this temperature. Concurrently, a solution was prepared containing 17.0 mg (0.17 mmol) of crystalline H3PO4 and 75 μL (0.17 mmol) of tri-n-octylamine dissolved in a mixture of 230 μL of tri-nbutylphosphate and 460 μL of n-hexyl ether by stirring at room temperature. Again, residual water was removed by three cycles of evacuation for 30 min followed by backfilling with argon gas at room temperature. A 60 μL aliquot of the H3PO4 solution (containing 14 μmol of H3PO4) was then added to the La3þ/225Ac3þ solution at 100 °C and the resulting reaction mixture was heated to 200 °C under an argon atmosphere and held at this temperature with stirring for 2 h. The initially colorless reaction mixture turned slightly brown during this time. After cooling to room temperature, the product was precipitated from the colloid by adding a mixture of 0.5 mL of hexane and 1.0 mL of hexadecane and then separated by centrifuging at 14 000  g for 20 min. The product was washed with two cycles of redispersion in a mixture of 0.5 mL of hexane and 1.0 mL of hexadecane followed by centrifugation at 14 000  g for 2 min. The activity of the product was 3.3 mCi, and the radiochemical yield for this step was 61%. The supernatant from each step was saved for radioactivity assay. Surface Modification and Purification of La(225Ac)PO4 NPs. Freshly synthesized La(225Ac)PO4 NPs (containing 3.3 mCi of 225Ac) were dispersed in 0.5 mL of tetrahydrofuran, added to a solution containing 200 mg of 6-aminohexanoic acid (AHA, 1.5 mmol) dissolved in 400 μL of 0.1 M NaOH, and stirred for 10 min at 90 °C. A 0.6 mL aliquot of 0.1 M tris-HCl buffer (pH = 9) was added to the mixture and stirred for an additional 5 min, whereby the solution clarified slightly. The surface-modified NPs were dialyzed (total sample volume ∼1.5 mL) three times for 2 h each against 500 mL of 0.02 M phosphate buffer (pH = 6.5). Eventually, the NPs settled to the bottom of the membrane and a slightly discolored residue partitioned to the top of the dialysis solution. A 5 mL portion of each dialysate solution was saved for radioactivity assay. The activity of the final purified product was 2.4 mCi with an overall radiochemical yield of 44%. A portion of the surface-modified NPs (containing ∼600 μCi of 225Ac) was tested for in vitro leaching of radionuclides (detailed below), while the remainder (∼1.7 mCi) was used for in vivo biodistribution and SPECT/CT experiments. In Vitro Release of 225Ac, 221Fr, and 213Bi from La225 ( Ac)PO4 NPs. A sample of La(225Ac)PO4 NPs (∼600 μCi) in 450 μL of phosphate buffer solution was dialyzed (total sample volume ∼0.5 mL) against 500 mL of 0.02 M phosphate buffer for a period of 1 month. Periodically, a 5 mL aliquot of the dialysate was withdrawn and assayed for radioactivity. Samples were counted at 1 cm distance from the surface of a calibrated highpurity Ge detector (see Radioactivity Measurements section above). To determine 221Fr activity, the time duration between withdrawal of the sample from the dialysis bag and start of the measurement was about 1 min, and the sample was counted for 2 min. This same measurement also provided quantitative

measurements of 213Bi activity. The quantity of 225Ac was determined by recounting the same sample for 1 h using the same geometry on the following day. Antibody Conjugation. The AHA surface-modified La(225Ac)PO4 NPs, containing ∼1.6 mCi of activity and dispersed in ∼1.5 mL of 0.02 M phosphate buffer (pH = 6.5), were sonicated at lowest power for 30 s. The NPs were then coupled to 300 μg of either mAb 201B or control antibody mAb 33 per 1 mg of NPs using conventional carbodiimide chemistry.18 For this reaction, 2 μL each of 0.1 M 3-sulfo-N-hydroxysuccinimide (as sodium salt, sulfo-NHS) and 0.4 M 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC) aqueous solutions were added to the NP suspension and incubated for 10 min before addition of the mAbs. After overnight reaction at room temperature on a rotating disk to promote mixing, the solution was quenched by addition of glycine to a final concentration of 0.1 M. The suspension was sonicated as described above, and the mixture was separated on a 5 mL Aca34 Ultragel filtration column (Sigma) equilibrated in 0.01 M phosphate buffer, pH = 7.6, and 0.15 M NaCl (PBS) containing 5 mg/mL bovine serum albumin (BSA). The majority of the radioactivity (>90%) eluted in the void volume. The mAb-conjugated NPs were injected into mice tail veins within 1 h of preparation (see below). In Vivo Mouse Biodistribution Experiments. All experiments involving mice were performed according to the Institutional Animal Care and Use Committee of the University of Tennessee approved protocol 1502. Five-week-old female BALB/c mice (body mass ∼15 g, 2 weeks postdelivery) were used for all biodistribution and imaging experiments. Three groups, consisting of five mice per group, were each injected intravenously (tail vein) with 0.2 mL of NP-mAb conjugates containing 5 μCi of 225Ac. Groups 1 and 2 were injected with La(225Ac)PO4-mAb 201B conjugates, while group 3 was treated with La(225Ac)PO4-mAb 33, the control antibodies. Additionally, group 2 mice were injected with 0.1 mL of clodronatecontaining liposome suspension 24 h prior to injection with radioactive NP-mAb conjugates. This procedure has been shown to enhance targeting of ZnS/Cd125mTe NPs to endothelial tissue in the vascular model system by temporary depletion of phagocytic cells in the liver and spleen.19,46,47 An additional separate group of three mice were also injected with 0.2 mL of La(225Ac)PO4-mAb 201B conjugates containing 5 μCi of 225Ac for in vivo retention studies. Mice were housed with food and water ad libitum in a light/dark cycle environment before sacrificing at 1, 4, and 24 h postinjection for imaging and biodistribution experiments and 1, 4, 24, 48, and 120 h postinjection for in vivo retention studies.18,48 Biodistribution studies were performed on lungs, liver, spleen, and kidneys to evaluate the amount of both 225 Ac and 213Bi in target organs by measuring weighed tissue samples in a γ-ray scintillation counter at a specific time postsacrifice and again after the radioisotopes had achieved isotopic equilibrium (>3 h). Quantities of 225Ac and 213Bi present at the time of animal sacrifice were determined by appropriate crossover and decay corrections described above. microSPECT/CT Imaging. Small animal imaging48 was performed using a microCAT IIþ SPECT dual modality platform (Siemens Preclinical Imaging, Knoxville, TN). Mice, treated 24 h prior with 100 μL of clodronate-containing liposomes, were injected intravenously with 0.2 mL containing 45 μCi of La(225Ac)PO4-mAb 201B conjugates (group 1, 2 mice per group); 45 μCi of La(225Ac)PO4-mAb 201B conjugates mixed with 300 μg of cold, uncoupled mAb 201B (competitor group 2, 1 mouse 768

dx.doi.org/10.1021/bc100574f |Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

per group); or 45 μCi of La(225Ac)PO4 mAb 33 conjugates (control group 3, 2 mice per group). Animals were sacrificed by overdose of isoflurane at 1 h postinjection and imaged via microSPECT/CT 3 h later when the 225Ac and its daughters had reached equilibrium. SPECT data for the final images were acquired within energy windows of 12
268 keV at a detector operating bias of 760 V. Energy windows of 230
344 keV on detector 1 and 206
308 keV on detector 2, both operating at 700 V, were also examined. Data were collected using 2 mm pinhole collimators positioned ∼60 mm from the center of rotation. Forty-five projections were acquired at 8° azimuths, each for a period of 60 s. Each projection was normalized using a corrected flood image to improve overall image quality and accuracy. The flood image was obtained by removing the collimator and detector housings. An 225Ac source at equilibrium with daughter radioisotopes (221Fr and 213Bi) was used to uniformly flood both detectors over a 12 h time frame. The resulting data were normalized to unity and used to correct the individual SPECT projection data for nonlinear detector response. The data were reconstructed postacquisition using an 84  84  114 matrix, with isotropic voxels, and an ordered subset expectation maximization algorithm with eight subsets. Computed tomography (CT) data were collected following SPECT data acquisition49 with 360° of rotation and one projection per degree. Data were acquired using an exposure time of 210 ms per projection and X-ray settings of 80 kVp and 500 μA. CT images were reconstructed in real time from a 512  512  768 matrix with 77 μm isotropic voxels using an implementation of the Feldkamp filtered back-projection algorithm (Cobra; Exxim Computing, Pleasanton, CA). CT and SPECT images were manually coregistered and visualized using the Amira image analysis software package (v 4.0; TGS, San Diego, CA). The CT data were rendered as a volume to aid in showing the location of the 225Ac uptake. The SPECT data were also volume rendered in this fashion and registered to the CT data to better demonstrate the uptake of the different NPmAb conjugates (or uncoupled mAbs). These volume renderings may not result in a linear scale that could reliably show relative uptake of the radioisotope.

Figure 2. Schematic depiction of synthesis and surface modification of La(225Ac)PO4 NPs. Equimolar quantities of LaCl3 (14 μmol) and H3PO4 (14 μmol), along with 5.4 mCi of 225AcCl3 and tri-n-butylphosphate and tri-n-octylamine surfactants, react at 200 °C for 2 h within a phenyl ether/n-hexyl ether solvent mixture to form La(225Ac)PO4 NPs. The surfaces of the NPs were subsequently modified with 6-aminohexanoic acid. The amine moieties electrostatically bond to the phosphate groups on NP, leaving carboxylate functionalities at the surface.

’ RESULTS

Figure 3. In vitro release of 225Ac, 221Fr, and 213Bi from La(225Ac)PO4 NPs. While 225Ac is nearly quantitatively retained, ∼50% of 221Fr and 213 Bi are released from the LaPO4 NPs over a period of 30 days. Activity of 225Ac = 600 μCi, pH = 6.5 phosphate buffer. For further details and determination of 221Fr and 213Bi activity, see Experimental Section.

Nanoparticle Synthesis and Surface Modification. The La(225Ac)PO4 NPs were synthesized by the reaction of LaCl3 (14 μmol) with H3PO4 (14 μmol) in high-boiling organic solvents.50 Solubility of the reagents in the solvents (n-hexyl and phenyl ether) is achieved by coordination of La3þ/225Ac3þ and H3PO4 by tri-n-butylphosphate and tri-n-octylamine surfactants, respectively. Addition of 5.4 mCi (0.41 nmol) of 225Ac as the chloride salt to the reaction mixture ensured sufficient radioactivity for biodistribution and SPECT/CT imaging experiments while leaving the bulk chemical composition of the NPs unchanged. The synthesis is reproducible, and repeat experiments provided an average decay-corrected radiochemical yield of 66 ( 4% (n = 3). The NP surfaces were subsequently derivatized with 6-aminohexanoic acid (aminocaproic acid) using an adaptation of a reported procedure51 and dialyzed against pH = 6.5 phosphate buffer solution for purification. Surface modification improved the aqueous dispersibility, although the product settles out of solution within 24 h, and the discolored residue at the top of the dialysis solution was likely displaced hydrophobic surfactants and solvents. The entire process, including synthesis, surface modification, and purification, took 24 h, with an overall average

decay-corrected yield of 47 ( 5% (n = 3). During this time, 7% of the initial radioactivity was lost to decay (t1/2 = 10 d), and the remaining radioactivity was distributed among the residual organic supernatants and solvents in addition to the reaction vessel, stir bar, pipet tips, aqueous washes, and dialysis membranes. The NP synthesis and surface modification steps are depicted in Figure 2. Transmission electron microscopy of cold LaPO4 NPs synthesized in an analogous protocol (no 225AcCl3 added to the LaCl3-tri-n-butylphosphate solution) confirmed the presence of unaggregated, 3
5-nm-diameter NPs.52 In-Vitro Release of 225Ac, 221Fr, and 213Bi from LaPO4 NPs. To probe the release of 225Ac as well as 221Fr and 213Bi daughters from the monazite host in vitro, a small portion of the surfacemodified and purified La(225Ac)PO4 NPs was further dialyzed against pH = 6.5 phosphate buffer solution over a period of 1 month. Aliquots of the dialysate were assayed periodically to determine the amount of each radioisotope present in solution 769

dx.doi.org/10.1021/bc100574f |Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

Figure 4. Schematic depiction of La(225Ac)PO4 NPs
mAb conjugation process. The 6-aminohexanoic acid modified La(225Ac)PO4 NPs with carboxylate moieties on the surfaces connect to free amine groups on the antibody via carbodiimide coupling to form the NP
mAb conjugate.

92 ( 10% (n = 3) of the NP
mAb 201B conjugates eluting from the column in two separate experiments. Since the LaPO4 NPs have molecular weights of ∼200 kDa, the 1
5 MDa gel fraction likely contains multiple NPs conjugated to multiple antibodies, and not aggregates of NPs.18,19 A small portion of the remaining radioactivity consisting of large aggregated material filtered out on top of the resin. The specific activities of the NP preparations were ∼1.7 mCi of 225Ac/mg of unconjugated LaPO4 NPs, ∼1.2 mCi/mg of La(225Ac)PO4-mAb 201B conjugates, and ∼1.1 mCi/mg of La(225Ac)PO4-mAb 33-conjugates. (Values were corrected for 42 h of decay.) Once the NP
antibody conjugates were size-selected, they were diluted in BSA-containing buffer, where no significant aggregation over time was observed. Test samples were analyzed on SDS polyacrylamide gel electrophoresis. Greater than 95% of the radioactivity moved through the stacker gel (4% acrylamide) but did not enter the separation gel (10% acrylamide) confirming the gel filtration results that the NP-mAb conjugates were about 1
5 MDa in size. Biodistribution of Targeted and Control NP
mAb conjugates in Normal and Clodronate-Treated Mice. La(225Ac)PO4 NP
antibody conjugates were injected intravenously into normal and clodronate-treated BALB/c mice within 1 h after column elution to avoid any aggregation or radiolysis that might occur, as well as minimize the loss of specific activity. Clodronate liposomes were injected 24 h prior to injection of NP
mAb conjugates. Groups of five mice were sacrificed at various periods (1, 24, and 48 h) postinjection and the organs collected for biodistribution analysis of radioisotopes in tissue by examining the γ-ray emissions of 213Bi once in equilibrium with 225Ac. Within 1 h postinjection, the NPs coupled to mAb 201B (group 1 mice) primarily targeted the lungs (235% ID/g and ∼30% ID/lung), and while uptake measured in the liver (23.8% ID/g) and spleen (29.1% ID/g) was significant, the concentrations were approximately 10- and 8-fold less, respectively, compared to the concentration in the lungs (Table 1 and Supporting Information Figure S1). In contrast, the NPs coupled to the control mAb 33 (group 3 mice) did not accumulate

(Figure 3 and Supporting Information Table S1). The fraction of 225 Ac which was found in the dialysate increased from 0.03% of the initial quantity to 0.1% over the first 5
7 d and then remained constant over the next 25 d. Within the first 4 h, 24% of the initial quantity of 221Fr was detected in the dialysate, which increased rapidly to ∼50% within 6 d and then remained relatively constant at ∼60% over the following 25 d. The fraction of 213Bi in the dialysate was similar to that of 221Fr, except for the first time point (3.8 h), at which time only 4% of 213Bi was found in the dialysate. Note that, in Figure 3, the y-axis data are plotted on a logarithmic scale since the measured activity of 225Ac in solution was vastly smaller compared to the activity of 221Fr and 213 Bi. The higher initial fraction of 221Fr measured in the dialysate is an indication that transport of 213Bi through the NP lattice, the phosphate buffer, and/or the dialysis membrane is much slower than that of 221Fr. Note that the majority of the 213 Bi detected in the dialysate originates from the decay of 221Fr and is not directly released from the NPs. Eventually, an equilibrium was reached between 221Fr and 213Bi in the dialysate, since 221Fr decays with a half-life ∼1/10 that of 213Bi (see Discussion). These results indicate that, while the 221Fr and 213Bi daughters were only partially sequestered, the 225Ac parent was nearly quantitatively retained in the LaPO4 NPs. Nanoparticle
Antibody Conjugation. The La(225Ac)PO4 NPs were coupled to mAb 201B and the control antibody, mAb 33, by linking the carboxylate moieties on the NP surfaces to amine groups on the antibodies through carbodiimide chemistry (Figure 4) as previously reported.18,19 The molecular weight of a 5 nm diameter LaPO4 NP with an estimated density of 5.12 g/ mL53 is 200 kDa. Using this estimate, the molecular ratio of mAb to NP in the reaction mix was ∼1:2.5. After coupling, the NP
mAb conjugates were briefly probe-sonicated at low power to break up any large aggregates and size-selected by gel filtration. The fraction of NP-mAb 201B or NP-mAb 33 that fractionated in the column void volume, corresponding to 1
5 MDa in size, comprised 95% of the initial radioactivity loaded on the column. Yields in excess of 50% were obtained, with 59 ( 9% (n = 2) and 770

dx.doi.org/10.1021/bc100574f |Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

Table 1. Biodistribution of Targeted La(225Ac)PO4-mAb Conjugates in BALB/c Micea injected dose per gram of tissue (ID/g) (%) (χ ( σ, n = 5) time postinjection (h) tissue

1

24

48

Targeted with mAb 201B (Group 1) Liver

24 ( 6

42 ( 12

48 ( 9

Spleen Kidney

29 ( 8 2.9 ( 0.9

162 ( 33 2.7 ( 0.3

297 ( 71 2.7 ( 0.6

Lung

235 ( 29

184 ( 25

212 ( 28

Clodronate-treated mice targeted with mAb 201B (Group 2) Liver

16 ( 7

20 ( 2

Figure 5. Whole body microSPECT/CT images of La(225Ac)PO4 targeted NPs of mice. NPs conjugated with mAb 201B target the lungs (A) but redistribute to the liver and spleen while in the presence of additional competing, unconjugated mAb 201B (B) as do NPs coupled to the control, mAb 33 (C). BALB/c mice treated 24 h prior with 100 μL of clodronate-containing liposomes were injected i.v. with 45 μCi of La(225Ac)PO4-mAb 201B (groups 1), 45 μCi of La(225Ac)PO4-mAb 201B mixed with 300 μg cold mAb 201B (competitor-group 2) or with 45 μCi of La(225Ac)PO4 mAb 33 (control-group 3). Animals were sacrificed by overdose of isoflurane at 1 h postinjection and imaged via microSPECT/CT 3 h later when the 225Ac and daughters had reached equilibrium. Only the higher quality images per group of mice are shown. For further details, see Experimental Section.

23 ( 3

Spleen

7.2 ( 2.1

18 ( 3

40 ( 3

Kidney

2.8 ( 1.6

1.6 ( 0.4

2.1 ( 0.4

Lung

279 ( 40

261 ( 20

276 ( 16

Liver

66 ( 3

74 ( 7

78 ( 4

38 ( 8 0.44 ( 0.03

66 ( 14 0.38 ( 0.05

95 ( 19 0.49 ( 0.04

2.0 ( 1.0

0.68 (0.34

1.1 ( 0.2

Targeted with mAb 33 (Group 3) Spleen Kidney Lung a

Number of mice per group = 5; Activity injected (i.v.) in each mouse = 5 μCi. Groups 1 and 2: La(225Ac)PO4 NPs-mAb 201B. Group 3 (control): La(225Ac)PO4 NPs-mAb 33. Group 2 mice were injected 24 h prior with 100 μL of clodronate-containing liposomes. Organs were analyzed in a γ-ray scintillation counter after 225Ac and its daughters had reached equilibrium (∼3 h).

conjugates in the liver and spleen of clodronate-treated mice increased slowly, but steadily, with time. Thus, temporary depletion of phagocytic cells enhanced both the accumulation and retention times in the lungs similarly to a previously reported study.19 microSPECT/CT Imaging. To confirm and illustrate the biodistribution data presented earlier, three-dimensional, wholebody microSPECT/CT images of mice treated with the following conjugates were collected: (A) La(225Ac)PO4-mAb 201B conjugates (NP-mAb 201B), (B) ∼100-fold molar excess of uncoupled, cold, competing mAb 201B added prior to injection of NP-mAb 201B, and (C) control NP-mAb 33 conjugates. Mice were sacrificed 1 h postinjection and imaged 3 h postsacrifice when 225Ac was in equilibrium with 213Bi. The SPECT data, obtained by measuring the primary γ-ray emission of 221Fr at 218 keV (the primary γ-ray of 213Bi at 440 keV is too energetic to be collected efficiently by the thin microSPECT detectors), were superimposed onto the images of mouse skeletons from subsequent CT scans. The images clearly indicate that the radioactivity from NP-mAb 201B conjugates was present in the lungs while only slight levels of radioactivity were detected in the spleen and liver (Figure 5A). In contrast, when the NP-mAb 201B conjugates were mixed with an excess of competing mAb 201B, the majority of the radioactivity was found in the spleen and liver (Figure 5B). NPs conjugated with control mAb 33 likewise localized in the spleen and liver (Figure 5C). These results together show that the observed localization of NP-mAb 201B conjugates to the lung represents specific, antibody-mediated interaction with a saturable receptor.38 Note that only the highest quality SPECT/CT images per group of mice are depicted in Figure 5. In Vivo Release of 213Bi following La(225Ac)PO4 NP Targeting to Mouse Lungs. Data in Table 2 show the activity of 213 Bi daughter in the lungs and kidneys as a function of time

significantly in the lungs (1.99% ID/g) but were predominately deposited in the liver (65.9% ID/g) and spleen (38.0% ID/g) (Table 1 and Supporting Information Figure S1). These results clearly indicate that the NPs bound to mAb 201B exclusively target the lungs and NP-mAb 33 conjugates do not. Over the course of 24 h, approximately 30% of the radioactivity in the NPmAb 201B conjugates was lost from the lungs. By 5 d, the activity in the liver almost doubled from 23.8% to 42.1% ID/g. The concentration in the spleen increased dramatically from 29.1% to 161.8% ID/g, which is likely due to the atrophy of the spleen due to radiotoxicity, resulting in a much smaller organ weight (data not shown). After 24 h, activity in the liver for NP-mAb 33 conjugates increased by only 10% while the radioactivity in the spleen almost doubled. These trends continued in a similar fashion over the subsequent 24 h period for both NP-mAb 201B and NP-mAb 33 conjugates. Note that a small but measurable quantity of radioactivity was present in the kidneys and remained relatively constant over time due to the excretion of the 213Bi daughter. For clodronate-treated mice (group 2 mice), the NP-mAb 201B activity in the lung (279% ID/g) was higher compared to that in untreated mice (235% ID/g) at 1 h postinjection, while accumulation in the liver (15.9% versus 23.8% ID/g, respectively) and spleen (7.15% versus 29.1% ID/g, respectively) was lower (Table 1 and Supporting Information Figure S1). After 24 h, the activity of NP-mAb 201B conjugates in the lungs was reduced by only 10% in clodronate-treated mice and remained relatively constant thereafter compared to a 30% loss of radioactivity from the lungs of untreated mice. Accumulation of NP-mAb 201B 771

dx.doi.org/10.1021/bc100574f |Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

Table 2. In Vivo Release of 213Bi following Targeting of La(225Ac)PO4-mAb 201B Conjugates to Lung of Micea time postinjection

Bi from lung

Ac is coordinated to DOTA (1,4,7,10-tetraazacyclododecane1,4,7,10-tetraacetic acid), purified, and coupled to antibodies. The efficiency of the first (coordination) step is fairly high and the reaction can be forced to completion by increasing temperature and pH. However, the efficiency of the second (coupling) step at best is ∼10%.25 The second problem is the potential loss of daughter radioisotopes (221Fr and 213Bi) from the chelator after decay of 225Ac, which adds significantly to the in vivo global toxicity of 225Ac.55 The major advantage of incorporating 225Ac within the NPs is to sequester the daughter radionuclides 221Fr, 217 At, and 213Bi at the site of targeting to mitigate the unintentional, nonspecific radiotoxicity.54 Monodisperse LaPO4 NPs have been synthesized in both organic and aqueous media and investigated for their fluorescent and scintillation properties.50,56
59 Rare earth phosphates with the monazite structure, such as bulk LaPO4, have been studied extensively in the context of nuclear waste treatment for longterm sequestration of radioactive lanthanides and actinides. In the monazite crystal lattice, the lanthanide ion is bound to nine surrounding oxygen atoms53 with sufficient energy to reduce leaching from energetic nuclear decay processes. Numerous studies have demonstrated that monazite is resistant to radiation damage, even under heavy metal ion bombardment.60
62 The radiation-tolerant crystal structure of monazite, combined with its low aqueous solubility, makes it an ideal NP system for sequestering radionuclides and daughter products. Additionally, the small size of the NPs (∼5 nm) is much less than the average range (60
100 μm) of the 5
8 MeV R-particles in water, allowing facile escape of R-particles from the NP lattice.23,44 We have successfully and reproducibly synthesized 3
5-nmdiameter LaPO4 NPs doped with 225Ac with an overall radiochemical yield of 47 ( 5% (n = 3) following surface modification and purification. For in vitro and in vivo studies, the specific activity was ∼1.7 mCi of 225Ac per mg of LaPO4 (La/Ac ≈ 30 000). A 5-nm-diameter LaPO4 NP contains on the order of 1000 La3þ ions, and thus, at La/Ac ≈ 30 000, there is approximately 1 Ac3þ ion per 30 NPs, leaving ample capacity for increased specific activity for therapy applications. Due to similar chemical properties and comparable ionic radii, the La3þ and 225Ac3þ were expected to form isomorphous mixed crystals; i.e., 225Ac3þ ions may replace La3þ ions in the crystal lattice and thus be “carried”. Moreover, competition of common impurities in lanthanum (Fe3þ, Ca2þ, and Zn2þ) with 225Ac3þ for cation sites within the monazite lattice is expected to be minimal. These hypotheses were confirmed in two ways. First, ∼66% of the initial 225Ac3þ was captured in La(225Ac)PO4 NPs before surface modification and dialysis. Second, the observed loss of 27 MeV per decay,23,29,44 making it a potent radioisotope if all the energy is deposited at the tumor site. Two major issues limit this approach. The first problem is the overall low labeling yield of coupling antibodies to 225Ac. In the best conventional method, 772

dx.doi.org/10.1021/bc100574f |Bioconjugate Chem. 2011, 22, 766–776

Bioconjugate Chemistry

ARTICLE

retention of the daughters in the NP suggests that a fraction of the recoil energy may partition into translation of the whole NP rather than being dissipated entirely through atomic displacements, as occurs in the bulk, resulting in a shorter recoil range. Because momentum (p) is conserved in nuclear recoil (pR = precoil), the corresponding recoil energy depends on the mass of the recoiling particles. The minimum energy (Ermin) available for bond breaking within the NP can be calculated as Ermin = Ermax 3 M1/(M1 þ M2), where Ermax is the maximum recoil energy, M1 is the atomic weight of the recoiling 221Fr nucleus (221 amu), and M2 is the molecular weight of the NP (∼2  105 amu). At this limit, the result is ∼103-fold reduction in the recoil energy and hence much shorter recoil range in the NPs versus bulk.30 The actual recoil energy imparted to the 221Fr nuclei most likely has a value between the two extremes. We are currently preparing 225Ac-doped LaPO4 NPs capped with a shell of cold LaPO4 in an effort to reduce daughter escape through both direct ejection and leaching. The dialysis experiments show a higher fraction of 221Fr than 213 Bi in the dialysate after 4 h but essentially equal fractions at later time points over the span of 1 month. This observation suggests that transport of 213Bi through the NP lattice, phosphate buffer, and/or dialysis membrane may be slower than that of 221Fr. Because it is a singly charged alkali metal ion, and its phosphates exhibit considerable aqueous solubility, Frþ should not be strongly bound in the monazite lattice, nor should it be strongly adsorbed to monazite surfaces. Therefore, it is reasonable to assume that the 221Fr concentration in the dialysate is representative of that inside the membrane. In contrast, Bi3þ is multiply charged, its phosphates exhibit poor aqueous solubility, and it may either be retained within or adsorbed to monazite surfaces. In solution, Bi3þ may further form colloidal hydroxides or phosphates that impede transport of Bi3þ ions through solution and the dialysis membrane. Very little activity (