Nanoconjugation of PSMA-targeting ligands enhances perinuclear localization and improves efficacy of delivered alpha-particle emitters against tumor endothelial analogues

Abstract

This study aims to evaluate the effect on killing efficacy of the intracellular trafficking patterns of alpha-particle emitters by using different radionuclide carriers in the setting of targeted antivascular alpha-radiotherapy. Nanocarriers (lipid vesicles) targeted to the prostate-specific-membrane-antigen (PSMA), which is unique to human neovasculature for a variety of solid tumors, were loaded with the alpha-particle generator actinium-225 and were compared to a PSMA-targeted radiolabeled antibody. Actinium-225 emits a total of four alpha-particles per decay, providing highly lethal and localized irradiation of targeted cells with minimal exposure to surrounding healthy tissues.

Lipid vesicles were derivatized with two types of PSMA-targeting ligands: a fully human PSMA antibody (mAb), and a urea-based, low-molecular-weight agent. Target selectivity and extent of internalization were evaluated on monolayers of human endothelial cells (HUVEC) induced to express PSMA in static incubation conditions and in a flow field. Both types of radiolabeled PSMA-targeted vesicles exhibit similar killing efficacy, which is greater than the efficacy of the radiolabeled control mAb when compared on the basis of delivered radioactivity per cell. Fluorescence confocal microscopy demonstrates that targeted vesicles localize closer to the nucleus, unlike antibodies which localize near the plasma membrane. In addition, targeted vesicles cause larger numbers of DNA double strand breaks per nucleus of treated cells compared to the radiolabeled mAb.

These findings demonstrate that radionuclide carriers, such as PSMA-targeted lipid-nanocarriers, which localize close to the nucleus increase the probability of alpha-particle trajectories crossing the nuclei, and, therefore, enhance the killing efficacy of alpha-particle emitters.

INTRODUCTION

The importance of antivascular therapy in the adjuvant treatment of cancer is well recognized (1). Critical prerequisites in this scenario, however, include the selective targeting of the tumor vasculature and the targeted delivery of highly lethal therapeutics. Among numerous anti-vascular agents developed and studied (1-3), alpha-particle emitters are identified for their exceptional suitability (3, 4). This is due to the high linear energy transfer (LET) (of the order of 80 keV/μm) and short range (50-100 μm) of alpha-particles resulting in highly lethal and localized irradiation of the tumor vasculature. To increase the killing efficacy of delivered radioactivity further, although not traditionally considered for alpha-particle emitters (5), different radionuclide carriers could be evaluated to explore potentially favorable spatiotemporal intracellular distributions (intracellular trafficking) of the alpha-emitters which could increase the probability of nuclear hits.

The design of preclinical studies, which aim to evaluate experimental neovasculature-targeting constructs, faces at least two major technical limitations. Human tumor endothelial cells expressing human antigens of targeting interest are practically still not available in culture (6), and in animal models the neovasculature and its antigens are of host-origin. To emulate tumor endothelium analogues in vitro, we utilize a parallel-plate flow chamber with a controlled flow field containing the targeted therapeutics and with walls coated with monolayers of model human endothelial cells (HUVEC) induced to express the prostate specific membrane antigen (PSMA). PSMA is a homodimeric type II integral membrane glycoprotein, is selectively found in the neovasculature of patients with several PSMA-negative tumors, and is absent in the healthy endothelium (7, 8).

In this study, we hypothesize that the patterns of intracellular trafficking of delivered alpha-particle emitters may significantly affect the efficacy of the delivered radioactivity. In order to explore this hypothesis, we designed lipid-based nanocarriers (lipid vesicles) loaded with the alpha-particle generator Actinium-225 (²²⁵Ac) and labeled the vesicles with two different types of PSMA-targeting ligands, which appear to target similar epitopes of PSMA: a fully human PSMA antibody (mAb), and a urea-based low-molecular-weight agent (9). The therapeutic generator ²²⁵Ac emits a total of four alpha-particles per decay (10). We evaluate both vesicle constructs and compare to the radiolabeled antibody in terms of targeting selectivity and killing efficacy, which are then compared to the intracellular trafficking patterns and any resulting DNA double strand breaks (dsDNA) for all constructs.

MATERIALS AND METHODS

Materials

The lipids 2-diheneicosanoyl-sn-glycero-3-phosphocholine (21PC), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[Methoxy(Polyethylene glycol)-2000] (Ammonium Salt) (DSPE-PEG), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[PDP (Polyethylene Glycol) 2000] (Ammonium Salt) (PDP-PEG-lipid), 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(LissamineRhodamine B Sulfonyl) (Ammonium Salt) (DPPE-Rhodamine) were purchased from Avanti Polar Lipids (Alabaster, AL) and were used without further purification (all lipids at purity > 99%). 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) and p-SCN-Bn-DOTA (DOTA-SCN) were purchased from Macrocyclics (Dallas, TX). Cholesterol (Chol), phosphate buffered saline (PBS), Sephadex G-50, Sepharose 4B, sodium carbonate, tetramethylammonium acetate (TMAA), sodium chloride (NaCl), glycine, sucrose, Diethylenetriaminepentaacetic acid (DTPA), calcium ionophore A23187, dithiothreitol (DTT), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), Endothelial Cell Growth Supplement (ECGS), heparin sodium salt, and N,N-dimethylformamide (DMF) were purchased from Sigma-Aldrich Chemical (Atlanta, GA). Ethylenediaminetetraacetic Acid, Disodium Salt Dihydrate (EDTA) was purchased from Fisher Scientific (Pittsburgh, PA). Fetal bovine serum (FBS) was purchased from Omega Scientific (Tarzana, CA). CellTiter 96^® Non-Radioactive cell proliferation assay (MTT) was purchased from Promega Corporation (Madison, WI). Matrigel^TM was purchased from BD Biosciences (San Jose, CA). 10DG and PD10 desalting columns were obtained from BioRad (Hercules, CA). For isotype control antibody, a human IgG1K (catalog number 0151K-01) was purchased from Southern Biotech (Birmingham, AL). Actinium-225 (²²⁵Ac, actinium chloride) was provided by the Institute for Transuranium Elements, Germany.

For the synthesis of lysine-glutamate urea conjugated to the free polymer chain(s) of 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N- [amino(polyethylene glycol)-2000] (DSPE-PEG(2000)), a solution of N-hydroxysuccinimide ester of suberate lysine-glutamate urea (11) (5 mg, 8.98 μmol, in 500 μL DMF) was added to DSPE-PEG(2000) (10 mg, 3.6 μmol in 900 μL DMF) followed by N,N-diisopropylethylamine (DIEA) (3 μL, 18 μmol). The reaction mixture was stirred at room temperature for two hours, and was then concentrated under reduced pressure at approximately 40°C. The semi-solid residue thus obtained was washed with 5 × 1 mL 80/20 water/acetonitrile solution to remove unreacted lysine-glutamate urea. The colorless product was dried under high vacuum for 20 hours and was used for experiments. The synthesis yield was approximately 90% and with m/Z: 1594.5 [M+1]2+ by ElectroSpray Mass Spectrometry. The molecular structure of the functionalized lipid with a non active analogue (negative control) of the urea ligand is shown in Figure S1C. The targeting urea molecule is a small molecule (not a peptide), therefore, for the design of the negative control, a linker was attached to the lipid.

Lipid vesicle preparation and loading of constructs with ²²⁵Ac

Vesicles composed of 21PC:Chol:DSPE-PEG:DPPE-Rhodamine at the mole ratio of 66.0:28.3:4.7:1.0 were formed using the thin film hydration method. All vesicles were labeled with 1 mole % of DPPE-Rhodamine lipid. For all targeted vesicles, either 0.5 mole % of the lysine-glutamate urea-based lipid (urea-lipid (9)) conjugate (resulting in a mole ratio of 65.7:28.2:4.7:0.9:0.5 21PC:chol:DSPE-PEG:DPPE-Rhodamine:lipid-urea) was included during the vesicle preparation step or the PSMA targeting antibody (Progenics Pharmaceuticals, Inc., Tarrytown, NY) was conjugated on the PDP-modified free ends of PEG-chains using standard click-chemistry (12). Vesicles were suspended in HEPES buffer (20 mM HEPES, 250 mM sucrose at pH = 7.4) and were prepared to encapsulate citrate buffer (140 mM citrate buffer with 5 mg/mL DOTA and 2.1 mg/mL ascorbic acid, pH 5.0). Vesicle size was measured using a Zetasizer NanoSeries (Malvern Instruments Ltd, Worcestershire, UK).

To load ²²⁵Ac in vesicles, 1 mL (5 mM total lipid) of vesicle suspension was incubated for 60 min at 80°C in a dry heating bath with 0.08 mL of a solution containing the following: 0.03 mL of ²²⁵Actinium in 3 mM HCl combined with 0.05 mL suspension of A23187 in an equimolar mixture of ethanol and water and at a final concentration in the vesicle suspension of 0.37 mg A23187/mL. After one hour of incubation, 0.05 mL of 10 mM DTPA was added to the vesicle suspension to complex any unentrapped ²²⁵Ac or ²²⁵Ac adsorbed on the vesicle surface (12). Purification and determination of loading efficiency was performed using size exclusion chromatography with a 10 cm Sephadex G-50 column eluted with phosphate buffer (PBS, pH = 7.4, 300 mOsm) and by measuring radioactivity associated with the eluted vesicle suspension.

Antibody radiolabeling was performed using a two-step process which involved radiolabeling of a isothiocyanate-functionalized derivative of DOTA at 60°C in 2 M tetramethylammonium acetate followed by labeling of antibodies with the radioactive chelate at lower temperature (13). The final eluted sample was evaluated for radiochemical purity and radiolabeling efficiency using ITLC.

Characterization of targeted constructs

Immunoreactivity of the urea-based ligand (11) and of the antibody (before and) after conjugation to vesicles was evaluated on the stably PSMA-positive LnCaP cell line in monolayers using antigens in excess of the ligands. The apparent binding affinity K_D of targeted vesicles was evaluated on fixed, but not permeabilized, PSMA⁺ HUVEC by measuring the cell-associated radioactivity on cells incubated with targeted ²²⁵Ac-loaded vesicles using serial dilutions, and on cells incubated in addition with 50 times excess of targeted non-radiolabeled vesicles (14).

Cell lines and preparation of cell conditioned media (CCM)

All cell lines (HUVEC and MDA-MB-231) were acquired from ATCC (Manassas, VA) in 2013 and 2014, and were cultured in media suggested by ATCC supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin in an incubator at 37°C and 5% CO₂. F12-K medium, used for HUVEC, was additionally supplemented with Heparin sodium salt (0.1 mg/mL) and ECGS (0.03 mg/mL). ATCC uses the following characterization tests: post-freeze viability, confirmation that cells are adherent, cell morphology, mycoplasma contamination, confirmation that cells are human in origin, and confirmation that cells do not contain pathogenic viruses. CCM was generated from the MDA-MB-231 breast cancer cell line (15).

Induction of PSMA expression on HUVEC cells

HUVEC cells were plated on Matrigel^TM at a density of 10⁶ cells per mL of CCM and were incubated in a humidified incubator at 37°C and 5% CO₂ for 18 hours according to a published method (15). Upon completion of incubation, PSMA expression was evaluated fluorometrically by immunofluorescence measured in a FACS Calibur flow-cytometer and by Western blot.

In particular, suspensions of fixed (with paraformaldehyde) HUVEC which previously had been carefully scraped and had been separated from Matrigel^TM by centrifugation (277 × g for 5 minutes) were incubated for 60 minutes with a primary human anti-PSMA mouse monoclonal antibody (LifeSpan BioSciences, Inc. (Seattle, WA)) followed by an additional 60 minute incubation with a FITC-labeled goat anti-mouse Fc secondary antibody (Jackson ImmunoResearch Laboratories (West Grove, PA)). PSMA positive LnCaP cells and PSMA negative (not activated) HUVEC cells were also evaluated.

For Western blot, cell lysate proteins were processed through SDS-PAGE under reducing conditions, followed by transfer onto nitrocellulose membrane. Membrane was blocked with 5% non-fat dry milk in TBST (20 mM Tris, 150 mM NaCl, 0.05% Tween-20) for 1 hour at room temperature, probed with primary mouse anti-PSMA antibody (ab19071, Abcam) in 0.5% milk in TBST for 1 hour at room temperature, and washed three times with TBST. The membrane was then probed with goat anti-mouse IgG-HRP conjugate (A24518, Life Technologies) for 1 hour at room temperature and then washed 3 times with TBST. Color was developed by adding HRP substrate (1706431, Bio-Rad) and incubating for 30 minutes.

Cell association of delivered radioactivity and cell viability studies

Binding of constructs under flow was determined using a rectangular parallel plate flow chamber (Glycotech, Gaithersburg, MD). For PSMA⁺ HUVEC, studies were initiated by introduction of ²²⁵Ac-loaded constructs in CCM at a lipid concentration of 83.33 μM and activity of 37 kBq/mL at a constant flow rate of 0.1 mL/min using a PHD Ultra Syringe Pump (Harvard Apparatus, South Natick, MA) at 37°C and 5% CO₂. Parallel experiments were performed under static conditions. For studies on PSMA⁻ HUVEC, cells were directly plated on fibrinogen-covered slides in regular F12-K media as also were all constructs. At the end of incubation, cells were washed twice using ice cold PBS and were re-suspended in F12-K media with 10% FBS, were counted using a hemocytometer, and were then used to determine (a) the cell associated radioactivity by counting the γ-emissions of Bismuth-213 upon reaching secular equilibrium; (b) the extent of cell internalized radioactivity; and (c) the viability of cells using the MTT assay. The extent of ²²⁵Ac retention by all constructs was measured on a fraction of cell-exposed suspension using size exclusion chromatography.

Intracellular localization of constructs imaged by confocal microscopy

HUVECs were incubated under static conditions with non-radioactive, rhodamine-tagged vesicles or FITC-labeled antibody in the presence (see below) or absence of endocytosis inhibitors. After completion of incubation (6 hours), HUVECs were fixed with paraformaldehyde (4% in PBS for 10 minutes) prior to Hoechst 33342 staining, and then washed (3×) with PBS. Monolayers were imaged using a Leica TCS SP2 confocal laser scanning microscope under a 40× and an oil immersion 100× objective. For quantitative intracellular distributions of all constructs, a dilation protocol was applied. The fluorescence image of each cell nucleus was thresholded to yield a binary space that represented the boundaries of the nucleus. A circular morphological structuring element was used to follow the edges of the thresholded nucleus image forming multiple concentric rings that expanded outward from the nucleus to the plasma membrane of the cell. The sum of pixel intensities within each ring were normalized to the sum of pixel intensities over the entire cell and graphed relative to the distance from the nucleus edge.

Characterization of cellular uptake mechanism

PSMA+ HUVECs were incubated for 30 minutes with chlorpromazine (10 μg/mL) or with genistein (100 μg/mL) before incubation with fluorescently labeled constructs. Upon completion of incubation, cells were washed and evaluated for fluorescence uptake using flow cytometry. In pre-saturation experiments, cells were incubated for 60 minutes with 1000× excess of free PSMA antibody (25 μg/ml) or the free lysine glutamate urea agent (0.1 μg/ml) (16) before incubation with fluorescently labeled constructs.

Immunofluorescent staining of γ-H2AX foci

Phosphorylation of histone γ-H2AX was imaged by immunofluorescence using the same confocal microscope as above. The γ-H2AX foci were treated as biomarkers of DNA double strand (dsDNA) breaks induced by the emitted alpha-particles. Upon completion of incubation with radioactivity, cells were washed and fixed in 4% paraformaldehyde for 10 minutes at room temperature, and were permeabilized using 0.1% Triton X-100 in PBS. After permeabilizing for 5 minutes at room temperature, cells were stained for γ-H2AX using OxiSelect^TM DNA Double Strand Break kit from CellBioLabs per kit instructions.

Statistical analysis

Results are reported as the arithmetic mean of n independent measurements ± the standard deviation. Student’s t-test was used to calculate significant differences in killing efficacy between the various constructs. p-values less than 0.01 are considered to be significant.

RESULTS

Lipid vesicle characterization

The average vesicle size was 107 ± 5 nm (n = 12); PDI = 0.06 ± 0.04 (n = 12). Antibody conjugation resulted in 31 ± 9 antibodies per vesicle (n = 4), and the optimization studies on urea-based targeted vesicles indicated that a density of approximately 368 urea-based ligands per vesicle exhibit best uptake (Figure S1). The immunoreactivity of the radiolabeled antibody was 88.6 ± 0.8%. The apparent immunoreactivity of the antibody-labeled vesicles was 18.1 ± 1.5% (n = 3), and of the urea-labeled vesicles was 15.8 ± 1.6% (n = 3). Targeted vesicles exhibited limited association with PSMA-positive cells upon receptor blocking with the targeting antibody (0.1 ± 0.0% and 3.9 ± 3.7% for antibody- and urea-targeted vesicles, respectively).

The equilibrium binding affinity (K_D) was estimated from nonlinear regression curve fit analysis (0.97<R²<1) for the antibody-labeled vesicles (35.6 ± 1.5 μM), the urea-based labeled vesicles (147 ± 9 μM), and the radiolabeled antibody (22.9 ± 2.1 nM) (Figures S2A, S2B and S2C, respectively).

The average loading efficiency of ²²⁵Ac in all targeted and non-targeted vesicles was 47.1 ± 16.6% (n = 16). In the presence of cells, vesicles retained 78.5 ± 3.6% (n = 16) of encapsulated radioactivity for the entire length of studies. The antibody radiolabeling efficiency was 3.4 ± 0.3% (n = 2). Antibody radiolabeling was stable (86.3 ± 2.3% of radioactivity retained, n = 2) for the length of the reported studies.

PSMA expression by HUVEC

HUVEC cells incubated with CCM on Matrigel^TM expressed PSMA at significant levels (63 mean shift with respect to a shift of 3 for untreated cells, Figure S3A; and positive staining by Western blot, Figure S3B) relative to the stably PSMA-positive cell line LnCaP (255 mean shift) corresponding to a reported 1.8 × 10⁵ PSMA receptors per cell (17). Since removal of CCM - upon completion of the initial 18 hour incubation period - resulted in decrease of PSMA expression by HUVEC over time (Figure S4), all reported studies were performed in the presence of CCM to ensure stable PSMA expression. Finally, introduction of a flow field (at 15 s⁻¹ shear rate) overlaying HUVEC decreased the PSMA expression in HUVEC by a factor of almost three relative to static culture conditions (Table S1).

Killing efficacy and cell associated radioactivity delivered by different constructs

Specific cell association of delivered radioactivity significantly increased with incubation time for all targeted constructs (p<0.01) and depended on incubation conditions (static and under flow) (Table S2). Both types of targeted vesicles delivered comparable radioactivities per cell with similar extents of internalization. For comparison, the ²²⁵Ac-labeled antibody delivered significantly greater levels of radioactivity per cell relative to targeted vesicles at the longer incubation times (4 and 6 hours; p<0.01).

Across all targeted constructs loaded with ²²⁵Ac, viabilities of PSMA-positive HUVEC exhibited significant decrease with treatment time for each construct (p<0.01) (Figures 1A and 1B). Initiating the MTT assay after three doubling times resulted in approximately 5-10% reduced viability relative to performing the assay after only one doubling time (Figure S5) (one doubling time of PSMA-positive HUVEC was 36 hours).

Figure 1.

Viability of PSMA-positive (A-D) and PSMA-negative HUVEC (E,F) in monolayers after treatment by various constructs loaded with ²²⁵Ac for different incubation times in static incubation conditions (A, C, E), and in the presence of a flow field (15 s⁻¹ shear rate) (B, D, F). White bars: Ab-labeled vesicles; Gray bars: urea-based labeled vesicles; Black bars: radiolabeled Ab; White bars with tilted pattern: vesicles containing no targeting group loaded with ²²⁵Ac; Gray bars with tilted pattern: ²²⁵Ac-DOTA chelate at activity levels corresponding to the values released from lipid vesicles; Black bars with tilted pattern: non-radiolabeled antibody. White bars with horizontal pattern: ²²⁵Ac-loaded vesicles labeled with an non-specific antibody; Gray bars with horizontal pattern: ²²⁵Ac-loaded vesicles labeled with an non-active urea analog. Viability evaluated after three doubling times. Error bars correspond to standard deviations of repeated measurements (2 and 3 independent preparations for antibodies and for vesicles, respectively, 2 measurements per preparation) (* p=0.044; ** p=0.063; † p=0.039; ‡ p=0.031; *** p=0.012; **** p=0.007; all other cases p>0.12).

Both types of targeted vesicles (white and grey bars in Figures 1A and 1B) resulted in significantly greater kill (p≤0.002) of PSMA-positive HUVEC compared to vesicles containing no targeting group (white bars with tilted pattern) at 6 hours of incubation (Figures 1C and 1D), for both static and flow conditions. The observed cell kill mediated by both types of targeted vesicles should not be attributed to leaked radioactivity (²²⁵Ac-DOTA at 7.4 kBq/mL or 20% of 37 kBq/mL) (at 6 hours; p<0.01) (Figures 1C-1F). Vesicles loaded with ²²⁵Ac and labeled with an non-specific antibody (white bars with horizontal pattern in Figure 1C, D) or with an non-active urea analog (grey bars with horizontal pattern in Figure 1C, D) affected cell viability similarly to the ²²⁵Ac-loaded vesicles containing no targeting group (white bars with tilted pattern).

Selectivity in affecting the viability of PSMA-positive HUVEC (Figures 1A and 1B) relative to normal endothelium (PSMA-negative HUVEC) (Figures 1E and 1F) was indicated by the corresponding different cell viabilities (p≤0.01) at 6 hours of incubation for all radiolabeled targeted constructs, and at 10 hrs of incubation for the radiolabeled antibody (Table S3), both at the static and flow conditions.

On PSMA-negative HUVEC in static conditions, after one hour of incubation, Ab-targeted vesicles were more lethal than urea-targeted vesicles (p=0.007). This finding cannot be supported by different radioactivity uptake by cells (Table S2).

Interestingly, Figure 2 (and Figure S6) - which depicts the viability of cells (shown in Figure 1) vs. the corresponding uptake of radioactivity by cells (shown in Table S2 as a function of time) for each of the constructs that were evaluated - shows that any level of associated radioactivity per cell resulted in greater killing efficacy when delivered by PSMA-targeting vesicles instead of the radiolabeled PSMA-targeting antibody. The presence of flow (Figure S7) did not affect efficacy.

Figure 2.

Viability of PSMA-positive HUVEC as a function of delivered radioactivity per cell mediated by the Ab-targeted radiolabeled vesicles (white symbols), the urea-based targeted radiolabeled vesicles (gray symbols), and radiolabeled antibody (black symbols), in (A) static incubation conditions, and in (B) the presence of a flow field. Data re-ploted from Figure 1 and Table S2.

Intracellular localization of constructs

PSMA-positive HUVEC incubated with PSMA-targeting antibody-labeled (red) and urea-labeled (purple) vesicles (Figure 3A, first and second panel, respectively) exhibited punctate fluorescence within the cytoplasm and - in the case of the urea-targeted vesicles - a pronounced perinuclear localization. The PSMA-targeting antibody (Figure 3A, third panel, green) localized mainly in the region closer to the plasma membrane far from the nuclear envelop. Quantitative processing of images of all constructs (Figure 3B), in the absence and presence of inhibitors, support these observations. Chlorpromazine, an inhibitor of clathrin-mediated endocytosis, did not affect the cell uptake (Figure 3A, thin continuous lines) and the intracellular spatial distributions (Figure 3B) of any construct. Genistein, an inhibitor of caveolae-mediated endocytosis, decreased the cell uptake of both types of targeted vesicles and of the antibody (Figure 3A, dashed lines).

Figure 3.

A. Intracellular spatial distribution (fused fluorescence and bright field images) and flow cytometry shifts in the absence and presence of endocytosis inhibitors: (first panel) Ab-targeted vesicles (red); (second panel) urea-targeted vesicles (purple); (third panel) fluorescently labeled antibody (green), in PSMA+ HUVEC. Cell nuclei are stained in blue. Scale bar is 40 μm.

B. Corresponding quantitative intracellular distributions of Ab-targeted vesicles (red symbols), urea-based targeted vesicles (purple symbols), and the antibody (green symbols) in above incubating conditions. n indicates total number of analyzed cells. Error bars correspond to standard deviations of the means of analyzed cells.

C. Cell uptake (fused fluorescence and bright field images) and flow cytometry shifts in the absence and presence of pre-saturation of cells with the free PSMA antibody or the free lysine glutamate urea agent at 1000× excess relative to cell receptors.

Pre-saturation of cells with the free PSMA antibody or the free lysine glutamate urea agent (Figure 3C) significantly decreased the association of all constructs with cells demonstrating PSMA-specific receptor-mediated uptake.

DISCUSSION

The major finding of this study is that radionuclide carriers (and in particular targeted lipid vesicles) that promote localization of ²²⁵Ac close to the cell nucleus, further enhance the already high killing efficacy of the delivered activity of alpha-particle emitters compared to radionuclide carriers that localize close to the cell’s plasma membrane (such as radiolabeled antibodies). Nuclear localization has been extensively explored for very short range emitters (such as of Auger electrons) (5), but not for alpha-particle emitters probably with the exception in the context of boron neutron capture therapy (18) and one study on using a modular transporter of astatine-211 (19). We utilized lipid vesicles loaded with ²²⁵Ac and labeled with a urea-based low-molecular-weight agent (11) and an antibody (20), both targeting the PSMA on human endothelial cells. Both types of vesicles improve the killing efficacy of delivered activity per cell by almost three fold relative to the killing efficacy of the same levels of activity when delivered by the PSMA-targeting antibody (Figure 2). The increase in killing efficacy, which is also accompanied by increased levels of DNA double strand breaks (beyond the arbitrary threshold shown by the red line, Figure 4) (21), strongly correlates with intracellular patterns of vesicles exhibiting localization close to the cell nucleus, unlike the antibody which preferentially localizes near the plasma membrane during the period of observation (Figure 3B). Other nanoparticles, not necessarily lipid vesicles, may result in similar perinuclear trafficking (22).

Figure 4.

Characteristic images obtained by immunofluorescent staining of γ-H2AX foci (green) in cell nuclei (blue), and distribution of the number of foci per cell nucleus upon exposure to the Ab-targeted vesicles (A), urea-targeted vesicles (B), and the radiolabeled Ab (C). Corresponding distributions of γ-H2AX foci per cell for Ab-targeted vesicles (white bars) (D), urea-based targeted vesicles (gray bars) (E), and the antibody (black bars) (F) were obtained by counting at least 50 cells per case. The red line serves as a threshold to indicate the maximum limit of foci mediated by the radiolabeled antibody. Scale bar is 40μm.

The observed spatial intracellular distributions of vesicles are expected to result in an increase of the effective solid angle of the (at least first) emitted alpha-particle (from the carrier-associated parent nuclide) with respect to the nucleus increasing, therefore, the probability of its trajectory crossing the nucleus. An approximate dosimetric analysis using the MIRDcell.v2.0 software (23) supports the above claims (Table S4 for ²²⁵Ac without daughters and Table S5 for ²²⁵Ac including radioactive daughters).

The choice of targeting ligand (in the sense of its role on altering the spatiotemporal intracelular distributions of the delivered nanoparticle) may affect the efficacy of nanoparticles as is suggested by a previous study on vesicles labeled with the PSMA targeting antibody J591 or the A10 aptamer (12) which demonstrated that only vesicles targeted with the J591 antibody perform similarly well as the radiolabeled antibody when compared in terms of radioactivity delivered per cell.

In the present studies, although the extent of endocytosis of both types of targeted vesicles and of the antibody is significantly decreased by the same inhibitor of caveolae-mediated endocytosis, their intracellular spatial distributions are dramatically different (Figure 3B) suggesting the existence of potentially additional factors that govern the internalization mechanism(s) (22).

To enable further development of targeted vesicles as carriers of ²²⁵Ac against easily accessibly cells (e.g. neovasculature) with relatively low copies of targeted receptors, at least two aspects of their design should be improved: the apparent K_D and the loading ratio of emitter to carrier. We have shown that the latter can in principle be improved by introducing greater radioactivity levels during loading which is equal to or greater than 47% of introduced radioactivity (for radioactivity levels up to at least 7.4 MBq). Decreasing the apparent K_D of targeted vesicles may be enabled by potentially optimizing the conjugation density of targeting ligands on vesicles (24) or the size of nanoparticles (25).

In summary, nanoconjugation of PSMA-targeting ligands using lipid vesicles loaded with ²²⁵Ac improves the killing efficacy of delivered activity per cell by almost three fold relative to the killing efficacy of the same levels of activity per cell when delivered by the PSMA-targeting antibody. We attribute this finding to the perinuclear localization of the emitter when delivered by targeting vesicles that enhances the probability of emitted alpha-particles traversing through the cell nucleus.