A Radio-Theranostic Nanoparticle with High Specific Drug Loading for Cancer Therapy and Imaging

Abstract

We have developed a theranostic nanoparticle delivering the model radionuclide ¹⁷⁷Lu based on the versatile lipid-calcium-phosphate (LCP) nanoparticle delivery platform. Characterization of ¹⁷⁷Lu-LCP has shown that radionuclide loading can be increased by several orders of magnitude without affecting the encapsulation efficiency or the morphology of ¹⁷⁷Lu-LCP, allowing consistency during fabrication and overcoming scale-up barriers typical of nanotherapeutics. The choice of ¹⁷⁷Lu as a model radionuclide has allowed in vivo anticancer therapy in addition to radiographic imaging via the dual decay modes of ¹⁷⁷Lu. Tumor accumulation of ¹⁷⁷Lu-LCP was measured using both SPECT and Cerenkov imaging modalities in live mice, and treatment with just one dose of ¹⁷⁷Lu-LCP showed significant in vivo tumor inhibition in two subcutaneous xenograft tumor models. Microenvironment and cytotoxicity studies suggest that ¹⁷⁷Lu-LCP inhibits tumor growth by causing apoptotic cell death via double-stranded DNA breaks while causing a remodeling of the tumor microenvironment to a more disordered and less malignant phenotype.

Graphical abstract

Introduction

Standard cancer treatment regimens use combinations of surgery, chemotherapy, and radiation therapy to eradicate the disease. The majority of patients receiving radiation treatments undergo external beam radiation therapy, but systemic internal radiotherapy has also achieved substantial clinical success in the last decade [1]. Unlike external beam therapy, which generally bombards the tumor area with photons or electrons emitted from an external source, systemic radiotherapy targets radionuclides to the tumor site via intravenous injection. The radiation source therefore resides inside the body, and the low radiation emission distance only allows the therapy to affect cells positioned close to particle deposition. This is in contrast to external beam therapy in which x-rays or gamma rays pass through the body from an external source. The current gold standard in systemic radiotherapy, called radioimmunotherapy, chelates the radioisotope to a tumor-targeted monoclonal antibody. The radio-immunotherapeutic Zevalin, which is labeled with the radioisotope ⁹⁰Yttrium, was the first of its kind, approved by the FDA in 2002 for use against B-cell lymphoma [2], but its clinical success has been tempered because of its high cost and antibody-mediated toxicity [3–5].

In this paper we describe a novel method for systemic internal radiotherapy. We have previously developed a targeted and versatile nanoparticle platform termed Lipid Calcium Phosphate (LCP) [6]. LCP is formulated by mixing two reverse microemulsions containing concentrated calcium and phosphate, and can encapsulate phosphorylated small molecules, peptides, and nucleic acids through their co-precipitation into the amorphous calcium-phosphate precipitate [7–9]. In this report we have now extended the application of this LCP platform by developing a theranostic nanoparticle with high specific drug loading for cancer therapy via the encapsulation of the model radioisotope ¹⁷⁷Lutetium (¹⁷⁷Lu). Trivalent cations such as lutetium can be driven into LCP, as their solubilities with phosphate are many orders of magnitude lower than the solubility of calcium phosphate. Phosphate then prefers to precipitate with these radioisotopes and even a trace amount can be encapsulated with high efficiency.

High encapsulation efficiency of a drug is often attainable only when a low amount of drug is loaded into each particle, requiring large doses of the particle to achieve a therapeutic effect. This can cause particle-mediated toxicity and make clinical scale-up very difficult and expensive. Our method of radioisotope loading does not encounter these issues. Because ¹⁷⁷Lu can be driven into LCP due to its low solubility with phosphate, and because only a small mass of ¹⁷⁷Lu is needed to provide a therapeutic effect, the drug loading window for a batch of LCP can span several orders of magnitude without affecting the ¹⁷⁷Lu encapsulation efficiency or LCP morphology, and without requiring a change in any other input parameters.

¹⁷⁷Lu was chosen as a model radionuclide because of its ability to provide simultaneous imaging and therapy: γ-radiation from ¹⁷⁷Lu is detectible using Single Photon Emission Computed Tomography (SPECT), while β-decay from ¹⁷⁷Lu causes DNA damage to nearby cells in addition to inducing Cerenkov radiation that is detectible using optical imaging techniques. Targeted delivery of these radionuclides therefore allows measurement of tumor accumulation in live mice while simultaneously treating the tumor mass. Unlike chemotherapeutics, radioisotopes like ¹⁷⁷Lu decay at a constant rate independent of their environment, providing a low, continuous dose to the tumor consistent with their decay rate. We therefore tested the ability of ¹⁷⁷Lu-LCP to achieve a sustained and quantifiable therapeutic effect in vivo after just a single dose.

Materials and Methods

2.0.1: Materials

¹⁷⁷LuCl₃ was purchased from PerkinElmer (Waltham, MA). 1,2-dioleoyl-sn-glycero-3-phosphate (DOPA), and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) were purchased from Avanti Polar Lipids (Alabaster, AL). N-(Carbonyl-methoxypolyethyleneglycol 2000)-1,2-distearoyl-sn-glycero-3-phosphoethanolamine, sodium salt (DSPE-PEG2000) was purchased by NOF America Corporation (White Plains, NY). DSPE-PEG2000-Anisamide (DSPE-PEG-AA) was synthesized in our lab as described previously [10]. Other chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

2.0.2: Cell Lines

Three different cell lines were used in the experiments described below: NCI-H460 (H460) human non-small cell lung cancer cells, UMUC3 human bladder cancer cells, and NIH/3T3 (3T3) murine fibroblasts. In some experiments, 3T3 cells that had been stably transfected with green fluorescent protein (GFP) were used.

2.0.3: Experimental Animals

6–8 week old female athymic nude mice were used for all experiments presented in this manuscript. These mice were purchased from the National Cancer Institute (Bethesda, MD) and bred at the Division of Laboratory Animal Medicine at the University of North Carolina-Chapel Hill. All work performed on these animals was approved by the Institutional Animal Care and Use Committee at the University of North Carolina-Chapel Hill.

2.1: ¹⁷⁷Lu-LCP Fabrication and Characterization

¹⁷⁷Lu-LCP fabrication has been modified [6, 11] from what is described in previous publications. Two reverse microemulsions were prepared in round-bottom flasks. Both microemulsion oil phases contained 73.7:6.7:9.7:10 Cyclohexane:Hexanol:IgepalCO520:TritonX-100 (v:v:v:v). 25mg/ml DOPA was added in chloroform to one oil phase (1:0.023 oil:DOPA v:v). A Phosphate aqueous phase (1:0.0125 oil:aqueous v:v) containing 100mM Na₂HPO₄ in 0.0125M NaOH was added to the DOPA oil phase with stirring. A Calcium aqueous phase containing 500mM CaCl₂ and trace ¹⁷⁷LuCl₃ in 0.0125M HCl was added to the other oil phase. After five minutes, the oil phase containing phosphate was added to the oil phase containing calcium and stirred for 40 minutes. 100% ethanol was then added to oil (1:1 ethanol:oil v:v) and stirred for 30 minutes. LCP was washed three times via centrifugation at 12,600g, removing the supernatant, replacing the supernatant with 100% ethanol, and vortexing the solution to resuspend the LCP pellet. Lu, Ca, and PO₄ that had not precipitated during mixing were discarded in the supernatant of the initial wash (~15% of input ¹⁷⁷Lu was removed here), and minimal ¹⁷⁷Lu signal was detected in the second and third supernatants. After the final wash, the precipitate was dissolved in chloroform and centrifuged at 10,000g for five minutes. This supernatant then contained the organic-soluble LCP cores, and the pellet contained any precipitate that was not well-coated in DOPA, and therefore not soluble in chloroform (~15% of input ¹⁷⁷Lu was removed here). Purifying the LCP cores in these two ways ensured that ¹⁷⁷Lu was only present inside the LCP. Outer leaflet lipids (OLLs) dissolved in chloroform were then added to the chloroform-soluble cores (40:40:18:2 DOPC:cholesterol:DSPE-PEG2000:DSPE-PEG2000-AA mol:mol:mol) (0.11:1 total OLL:initial oil phase v:v; all lipids at 20mM concentration). In experiments using DiI-labeled LCP, DiI dissolved in chloroform was added to the cores with the OLLs (1:100 DiI:total OLL mol:mol). Chloroform was evaporated until the particles and lipids coated the vial in a lipid film. Particles were then rehydrated with 80°C water to the injection volume. The solution was vortexed and sonicated liberally, then heated at 80°C for >10 min and passively cooled. The anisamide ligand was used in all formulations to target the sigma receptor overexpressed in UMUC3 and H460 cells [12, 13].

In experiments using DiI-labeled Lu-LCP, sucrose gradient ultracentrifugation (50,000g for 4 h) was used to purify DiI-Lu-LCP from excess free DiI and DiI that had incorporated into excess free liposomes. Encapsulation efficiency of ¹⁷⁷Lu into ¹⁷⁷Lu-LCP was measured by a Model AA2010 “Nucleus” Gamma Scintillation Counter or a Capintec Radioisotope Calibrator CRC-127R and calculated as (signal from ¹⁷⁷Lu in LCP cores)/(signal from ¹⁷⁷Lu in input aqueous phase). ¹⁷⁷Lu-LCP hydrodynamic diameter and zeta potential were measured after sucrose gradient ultracentrifugation using a Malvern Nano ZS dynamic light scattering instrument. Three different batches of LCP were measured. Particle size was corroborated using transmission electron microscopy (TEM).

2.2: Loading Capacity

To simulate an increase in ¹⁷⁷Lu loading into ¹⁷⁷Lu-LCP, a trace amount of ¹⁷⁷Lu was supplemented with additional nonradioactive ¹⁷⁵Lu. As the Lu:Ca input ratio was increased from 1:10,000,000 to 1:1, no changes in any inputs were made save for the increase in ¹⁷⁵Lu. ¹⁷⁷Lu EE was measured using gamma scintillation and ¹⁷⁷Lu-LCP morphologies/structural makeups were measured using TEM and energy-dispersive X-ray spectroscopy (EDS) at the Chapel Hill Analytical and Nanofabrication Laboratory at the University of North Carolina at Chapel Hill. TEM images were taken using a TEM JEOL 2010F-FasTEM or a TEM JEOL 100CX II, and EDS measurements were taken using the TEM JEOL 100CX II.

2.3: Pharmacokinetics and Biodistribution

Healthy nude mice were injected with ¹⁷⁷Lu-LCP or free ¹⁷⁷LuCl₃ of known activities. The ¹⁷⁷Lu-LCP or ¹⁷⁷LuCl₃ was injected into the left tail vein and circulating ¹⁷⁷Lu-LCP or ¹⁷⁷LuCl₃ was measured by removing 20–30 μl of blood from the right tail vein at several time points after injection. The signal was read using gamma scintillation and was used to calculate the total signal remaining in the blood. A Microsoft Excel Add-In called PKSolver [14] was used for the pharmacokinetic analysis. In a separate experiment, H460 or UMUC3/3T3 tumor-bearing mice were injected with a therapeutic dose of ¹⁷⁷Lu-LCP or free ¹⁷⁷LuCl₃. Twenty four h after injection, the mice were sacrificed and their organs were dissected and read for ¹⁷⁷Lu activity.

2.4: SPECT and Cerenkov Imaging

UMUC3/3T3 tumor-bearing nude mice were i.v. injected with 2.5 mCi of ¹⁷⁷Lu-LCP or free ¹⁷⁷LuCl₃. Twenty four h later, the mice were imaged using a small animal SPECT/CT system (eXplore speCZT, GE Healthcare) at the Small Animal Imaging facility on the UNC campus. SPECT Images were taken with the energy window to be set from 188–229 keV to receive the main gamma photon from Lu-77 peaked at 208 keV. A mouse slit collimator was used to provide whole body imaging with 1.5mm transaxial and 2.5mm axial resolution, and a pinhole collimator was used for partial body imaging with higher resolution (1mm isotropic resolution). Immediately after SPECT/CT imaging, the mice were taken to an in vivo optical imaging system (IVIS-Kinetic, Perkin Elmer) to measure Cerenkov emissions generated by ¹⁷⁷Lu.

2.5: In Vivo Tumor Inhibition

Six to eight week old nude mice were subcutaneously inoculated with 100ul PBS containing 5×10⁶ H460 cells or 5×10⁶ UMUC3 cells + 2.5×10⁶ 3T3 cells in 25% matrigel (BD matrigel matrix high concentration) on the left flank. After 8–12 days, when tumors had reached a size of 100–150mm³, mice were given a single injection of PBS (Untreated mice), non-radioactive (cold) ¹⁷⁵Lu-LCP, free radioactive ¹⁷⁷LuCl₃, or radioactive ¹⁷⁷Lu-LCP. Mice bearing UMUC3/3T3 tumors received a dose of 200 μCi/mouse and mice bearing H460 tumors received a dose of 250 μCi/mouse. Mouse weight was measured every 2 days until untreated mice reached humane endpoints. Tumor volume was measured with digital calipers and calculated as (L x W x D)/2. n = 5–8. Student’s t-test of the final AUC used for statistical analysis.

2.6 Tumor Accumulation

Nude mice bearing s.c. H460 or UMUC3/3T3 tumors were treated with one dose of 250 μCi or 200 μCi ¹⁷⁷Lu-LCP, respectively, or appropriate controls. The lipid bilayer of the ¹⁷⁷Lu-LCP particles were labeled with the fluorescent small molecule DiI to allow LCP measurement in the tumor. Twenty four h after injection, mice were sacrificed and tumors were dissected and cut in half. One half was frozen in OCT, sectioned, and mounted for DiI Lu-LCP imaging, and the other half was fixed in 4% formalin, embedded in paraffin, and sectioned for CD-31 staining (1:100 dilution rabbit primary Ab, Abcam #ab28364; 1:100 dilution anti-rabbit 647 fluorescent secondary Ab, Cell Signaling #4414) and DAPI counterstain.

2.7 Cytotoxicity Studies

One, 2, and 4 days after treatment with 200 μCi ¹⁷⁷Lu-LCP or controls, UMUC3/3T3 tumor-bearing nude mice were sacrificed and their tumors were dissected, fixed, and sectioned in paraffin. Adjacent sections were stained for p-H2AX immunofluorescence (1:100 dilution Rabbit monoclonal primary Ab, Cell Signaling #9718; 1:100 dilution anti-rabbit 647 fluorescent secondary Ab, Cell Signaling #4414) and Terminal deoxynucleotidyl transferase dUTP nick end labeling (DeadEnd^™ Fluorometric TUNEL System, Promega) with DAPI counterstains.

2.8 Microenvironment Studies

Nude mice were inoculated with UMUC3/3T3-GFP (fluorescent) tumors using the same cell numbers reported above. When tumors reached ~400 mm³, mice were treated with 200 μCi of fluorescently labeled DiI-¹⁷⁷Lu-LCP or controls. After 2 days, mice were sacrificed and their tumors were dissected and prepared for freezing in OCT via a 2-h fixation in 4% formalin followed by overnight incubation in a 30% sucrose solution at 4°C. The frozen tumors were sectioned and imaged with DAPI counterstain to measure GFP and DiI distribution. DiI distribution was quantified using Matlab.

Separate nude mice were inoculated with UMUC3/3T3 (nonfluorescent) tumors. When tumors reached ~300 mm³, mice were treated with 200 μCi of ¹⁷⁷Lu-LCP or controls. Four days after treatment, the tumors were dissected, fixed, and sectioned in paraffin. Adjacent sections were trichrome stained and stained for α-SMA immunofluorescence (1:100 dilution Rabbit monoclonal primary Ab, Abcam #ab5694; 1:100 dilution anti-rabbit 647 fluorescent secondary Ab, Cell Signaling #4414) with DAPI counterstain. The stained sections were digitally scanned using the Aperio Scanscope at the University of North Carolina-Chapel Hill’s Translational Pathology Lab. Quantitative analysis was performed on Imagescope and ImageJ software.

2.9: Toxicity Studies

Nude mice and CD-1 mice were treated with 250 μCi of ¹⁷⁷Lu-LCP or appropriate controls and were sacrificed 12 days after treatment. Blood and organs were dissected. Serum levels of AST, ALT, BUN, and creatinine were measured, and organs were stained with H&E to measure any changes in tissue morphology. Whole blood was measured for changes in white blood cell count (WBC), hematocrit (HCT), mean cell volume (MCV), red blood cell count (RBC), hemoglobin count (HGB), and platelet count (PLT).

Results

3.1: Lu-LCP Fabrication and Characterization

The LCP nanoparticle platform preferentially encapsulates ¹⁷⁷Lu in its solid amorphous calcium phosphate core during fabrication (Scheme shown in Figure 1A). Just a trace amount of ¹⁷⁷Lu can be efficiently encapsulated because of lutetium’s much lower solubility with phosphate (k_sp = 2×10⁻²⁵) [15] when compared to calcium’s solubility with phosphate (k_sp = 1×10⁻⁷). The fabrication protocol for ¹⁷⁷Lu-LCP was modified [6, 11] from previous publications to maximize ¹⁷⁷Lu encapsulation efficiency (EE) at approximately 70%, with the majority of the loss occurring during washing and purification. Particles were measured via dynamic light scattering to have a diameter of 36 ± 9 nm and a polydispersity index of 0.27 ± 0.05, with a zeta potential of −6.7 ± 3.5 mV after separating the particles from excess empty liposomes using sucrose gradient centrifugation (Supplemental Table S1). Particle size was corroborated by transmission electron microscopy (Figure 1B–C). Although cellular entry is not requisite for ¹⁷⁷Lu-LCP to achieve therapeutic effect, because anisamide has been proven as an effective targeting ligand against the sigma receptor overexpressed in epithelial cancers such as UMUC3 and H460 [12, 13], and because receptor-mediated endocytosis draws nanoparticles inside the cell and closer to nuclear DNA, DSPE-PEG-Anisamide was exclusively used in all Lu-LCP formulations as described in section 2.1 to provide the greatest statistical chance for treatment efficacy.

Figure 1.

¹⁷⁷Lu-LCP Formulation. A) Schematic depicting general formulation procedure. B) TEM image of final ¹⁷⁷Lu-LCP. C) TEM image of final ¹⁷⁷Lu-LCP with uranyl acetate negative stain applied. Scale bar represents 100 nm.

3.2: Loading Capacity

Preferential loading of ¹⁷⁷Lu allows ¹⁷⁷Lu-LCP to be fabricated with a very high radioisotope concentration, consequently allowing very high drug loading in a particle that is still overwhelmingly comprised of calcium phosphate. Figure 2 and Table 1 show that the drug loading window in equivalent-sized batches of LCP core spans across several orders of magnitude without affecting the ¹⁷⁷Lu encapsulation efficiency and not requiring a change in any other input parameters. Additionally, the LCP morphology remains unchanged up to a Lu:Ca input ratio of 1:1000 (Figure 3A–F), above which the Lu starts becoming a structural component of the LCP, as measured by EDS (Figure 3G), which introduces heterogeneity to the formulation. As Lu:Ca is increased to 1:1, Lu forces out nearly all Ca from the LCP to form cores of LuPO₄, but the total dose of heavy metal per particle is undesirable due to its inherent toxicity. These data suggest that the maximum Lu:Ca input ratio that maintains desirable encapsulation and morphology is 1:1000. At this input ratio, ~24 mCi of ¹⁷⁷Lu—enough to treat over 100 mice—can be encapsulated in a batch size of only 3.2 total mL of oil phase. Considering that it is not uncommon for other types of LCP formulations to be synthesized in batches of 120ml or more for small mouse studies, one can understand how Lu-LCP is able to avoid scale-up complications generally associated with nanoparticles. Moreover, a very small mass of total lutetium—approximately 70 pmol/mouse, or 6×10⁻⁴ mg/kg—is used here to achieve therapeutic effect. This low total dose is important in minimizing heavy metal-mediated toxicity. Combined, the high drug loading window and low total mass of lutetium in ¹⁷⁷Lu-LCP minimize batch size and toxicity while subsequently allowing the treatment of a large number of animals per batch.

Figure 2.

¹⁷⁷Lu-LCP encapsulation efficiencies at increasing levels of lutetium. ¹⁷⁷Lu EE decrease begins around 1:100 Lu:Ca. n = 2–3.

Table 1.

¹⁷⁷Lu-LCP inputs and output activity based on a specific activity of 20 Ci/mg of lutetium, the minimum standard for PerkinElmer delivery. Ca is always given at a 5x molar excess of PO₄. Highest level of stable LCP formulation is 1:1000 Lu:Ca. A trace amount of ¹⁷⁷Lu was supplemented with nonradioactive ¹⁷⁵Lu to achieve the proper Lu:Ca ratio.

Lu:Ca Input (mols)	PO4 Input (mols)	Total Lu Input (mols)	Input Activity* (mCi)	177Lu EE	Output Activity (mCi)
1:1×107	2.0×10−6	1.0×10−12	0.0035	0.68	0.0024
1:100,000	”	1.0×10−10	0.354	0.70	0.25
1:1000	”	1.0×10−8	35.4	0.69	24.4
1:100	”	1.0×10−7	354	0.61	217
1:9	”	1.1×10−6	3933	0.22	865
1:1	”	1.0×10−5	35400	0.19	6549

Figure 3.

¹⁷⁷Lu-LCP morphologies at different Lu:Ca inputs. A–F) TEM images of ¹⁷⁷Lu-LCP at varying Lu inputs. G) Percent Lu in output ¹⁷⁷Lu-LCP based on Lu:Ca inputs, calculated by energy dispersive X-ray spectroscopy (EDS). *Lu not detected; negligible small LCP population

3.3: Pharmacokinetics and Biodistribution

The pharmacokinetics and biodistribution of ¹⁷⁷Lu-LCP were determined in tumor-bearing nude mice. A two-compartment model [14] was used to calculate the clearance half-lives, which were 0.52 h for the distribution half-life, t_1/2α, and 7.28 h for the elimination half-life, t_1/2β, as shown in Figure 4A. Pharmacokinetic analysis was also performed for free ¹⁷⁷LuCl₃, which showed nearly 80% blood clearance in the first 30 min after injection (Figure S1). It can therefore be concluded that the circulating signal from mice given ¹⁷⁷Lu-LCP is indeed from intact particles. The biodistribution of ¹⁷⁷Lu-LCP and free ¹⁷⁷Lu in tumor-bearing mice is given in Figure 4B. ¹⁷⁷Lu-LCP shows significant accumulation in the tumor, as well as accumulation in the nanoparticle clearing organs, the liver and spleen, as it is too large to clear renally. In contrast, free ¹⁷⁷Lu is an ion that is plenty small to clear through the kidneys, but also selectively accumulates in the bones, as lanthanide elements are known to do [16]. Free ¹⁷⁷Lu also shows some kidney accumulation, which may occur during the quick renal clearance of the element (Figure 4B).

Figure 4.

¹⁷⁷Lu-LCP Pharmacokinetics and Biodistribution. A) The PK curve of i.v. injected ¹⁷⁷Lu-LCP fit to a two-compartment model; n = 5. B) Organ BD at t = 24h after i.v. injection of ¹⁷⁷Lu-LCP or free ¹⁷⁷LuCl₃; n = 3–4. H460 tumor and UMUC3/3T3 (U/T) tumor were subcutaneously inoculated in different mice. *BD of free ¹⁷⁷Lu was not tested in H460 tumor-bearing mice.

3.4: SPECT and Cerenkov Imaging

In addition to allowing simple means of detection using gamma scintillation, ¹⁷⁷Lu can also be used as a SPECT contrast agent, permitting measurement of ¹⁷⁷Lu biodistribution and tumor accumulation in live mice. Figure 5A–E shows the biodistribution of ¹⁷⁷Lu-LCP 24 h after injection, including the visible accumulation of ¹⁷⁷Lu in the tumor. In contrast, free ¹⁷⁷LuCl₃ does not accumulate in the tumor and therefore tumor accumulation is not visible using SPECT/CT imaging (Figure 5F–H).

Figure 5.

SPECT/CT images of nude mouse bearing subcutaneous UMUC3/3T3 tumor on left flank at 24 h after injection of ¹⁷⁷Lu-LCP (Figures 5A–E) or free ¹⁷⁷Lu (Figures 5F–H). Mouse slit collimator was used for whole body imaging: A) left sagittal, B) coronal, C) midsagittal, F) mid-sagittal and G) coronal views. Pinhole collimator was used in D), E), and H) for higher resolution axial and left sagittal views. ¹⁷⁷Lu-LCP accumulated in tumor, liver, and spleen, while free ¹⁷⁷Lu accumulated in the bone and kidneys, but not in the tumor.

Interestingly, ¹⁷⁷Lu also allows imaging via its β-emissions, which are the same decay products (electrons) that induce its therapeutic effect. As the β-particles move through the tissue over their average path length of ~250 μm, they induce photon emissions from the medium with energies in the visible spectrum. This so-called Cerenkov light was measured using an in vivo optical imaging system (Figure 6), also showing tumor uptake of ¹⁷⁷Lu-LCP and corroborating the results obtained during SPECT/CT imaging. Cerenkov imaging is a particularly interesting imaging modality because it offers a less expensive and more convenient alternative to SPECT that is not generally used in this field. Although its spatial resolution and depth of penetration is not as high as SPECT or PET, Cerenkov imaging uses less expensive hardware and requires acquisition times of only a few seconds, if the radioisotope and dose are adequate.

Figure 6.

Images of Cerenkov luminescence in nude mice bearing subcutaneous UMUC3/3T3 tumor on left flank. Luminescence image (A) and plain photograph (B) in mouse 24 h after injection of 2.5 mCi ¹⁷⁷Lu-LCP. Luminescence image (C) and plain photograph (D) in mouse 24 h after injection of 2.5 mCi free ¹⁷⁷LuCl₃. Optical images demonstrated ¹⁷⁷Lu-LCP accumulation in liver, spleen, and tumor, while free ¹⁷⁷LuCl₃ mainly accumulated in the spine. Data was acquired by an IVIS optical system.

After imaging was complete, the mouse organs were dissected and their ¹⁷⁷Lu signal was read using gamma scintillation. These values are shown in Supplementary Table S2.

3.5: In Vivo Tumor Growth Inhibition

To assess the effectiveness of a single treatment of ¹⁷⁷Lu, one dose of ¹⁷⁷Lu-LCP was intravenously injected into two different subcutaneous xenograft tumor models. One of the models was an aggressive stroma-rich human bladder cancer model previously developed by our lab. This model uses both UMUC3 bladder cancer cells and NIH-3T3 fibroblasts to generate a tumor that has been shown to be more similar to patient bladder cancer than UMUC3 tumors alone [17]. As shown in Figure 7A, a single 200 μCi dose of ¹⁷⁷Lu-LCP showed sustained and significant tumor growth inhibition compared to tumors treated with nonradioactive ¹⁷⁵Lu-LCP and tumors receiving only free ¹⁷⁷LuCl₃. ¹⁷⁷Lu-LCP was also tested on the aggressive and radio-resistant [18] H460 human non-small cell lung cancer model; an increased dose of 250 μCi of ¹⁷⁷Lu-LCP provided statistically significant tumor growth inhibition as shown in Figure 7B. This boosted dose was still low enough to avoid changes in mouse body weight (Figure S2), and the accumulated dose over the course of the study did not significantly elevate liver/kidney toxicity markers (Table 2).

Figure 7.

In vivo tumor growth inhibition in two subcutaneous xenograft tumor models. Nonradioactive ¹⁷⁵Lu was loaded into LCP NPs to act as a vehicle control. One dose per mouse was delivered intravenously on Day 0 (arrows). A) Human UMUC3 bladder cancer supplemented with murine 3T3 fibroblasts. B) Human H460 non-small cell lung cancer. Initial tumor sizes ~100–150 mm³; n = 5–8; †p<0.03, ‡p<5e-5, *p<0.015.

Table 2.

Toxicity of ¹⁷⁷Lu in mice, measured 12 days after treatment with 250 μCi ¹⁷⁷Lu-LCP or controls to test cumulative toxicity: A) Serum protein values in healthy, but immunocompromised, nude mice; B) Serum protein values in healthy CD-1 mice; C) Whole blood analysis in healthy CD-1 mice, measuring white blood cell count (WBC), hematocrit (HCT), mean cell volume (MCV), red blood cell count (RBC), hemoglobin count (HGB), and platelet count (PLT). ¹⁷⁷Lu-LCP treatment does not present significant cumulative renal/hepatotoxicity, but causes a decrease in WBC and PLT.

A) Nude Mice	BUN (mg/dL)	Creatinine (mg/dL)	AST (U/L)	ALT (U/L)
Untreated	23 ± 0.5	0.1 ± 0	99 ± 13	37 ± 2
175Lu-LCP	23 ± 1.0	0.1 ± 0	102 ± 2	34 ± 5
250 μCi Free 177Lu	20 ± 1.7	0.1 ± 0	88 ± 16	41 ± 1
250 μCi 177Lu-LCP	24 ± 1.9	0.1 ± 0	73 ± 4	40 ± 4
Normal Range	8–33	0.2–0.9	54–298	17–77

B) CD-1 Mice	BUN (mg/dL)	Creatinine (mg/dL)	AST (U/L)	ALT (U/L)
Untreated	20 ± 0.3	0.1 ± 0	69 ± 14	33 ± 1
175Lu-LCP	21 ± 1.3	0.2 ± 0.03	91 ± 12	35 ± 2
250 μCi Free 177Lu	15 ± 0.7	0.1 ± 0	109 ± 15	41 ± 3
250 μCi 177Lu-LCP	15 ± 0.5	0.1 ± 0	125 ± 9	45 ± 5
Normal Range	8–33	0.2–0.9	54–298	17–77

C) CD-1 Mice	WBC (103/μl)	HCT (%)	MCV (fl)	RBC (106/μl)	HGB (g/dl)	PLT (103/μl)
Untreated	4.5 ± 0.2	52.6 ± 0.2	48.9 ± 0.2	10.7 ± 0.1	16.3 ± 0.1	892 ± 36
175Lu-LCP	5.2 ± 0.7	50.6 ± 1.1	51.0 ± 1.4	10.0 ± 0.5	15.7 ± 0.3	679 ± 54
250 μCi Free 177Lu	2.3 ± 0.4	47.6 ± 1.7	53.8 ± 3.8	8.9 ± 0.4	14.6 ± 0.3	926 ± 34
250 μCi 177Lu-LCP	0.3 ± 0.1	38.8 ± 2.2	50.5 ± 1.6	7.7 ± 0.4	12.0 ± 0.6	217 ± 24
Normal Range	2.6 – 10.1	32.8 – 48.0	42.3 – 55.9	6.5 – 10.1	10.1 – 16.1	780 – 1540

3.6: Tumor Accumulation

To compare the localization of ¹⁷⁷Lu-LCP inside H460 and UMUC3/3T3 tumors, the fluorescent small molecule DiI was incorporated into the outer leaflet of the nanoparticle. DiI has been widely used in liposomal formulations and as a cell membrane marker, and is well known to faithfully remain in the hydrophobic portion of the lipid bilayer [19, 20]. Tumor-bearing mice were sacrificed 24h after injection with DiI-labeled ¹⁷⁷Lu-LCP, and tumor sections were imaged for DiI uptake. Figure S3A shows that the bladder cancer tumor has more homogeneous uptake of particles when compared to the H460 lung cancer tumor. This could be because the bladder cancer tumors contain more CD-31-positive area than the H460 tumors (Figure S3B), corresponding to more vasculature. This could also explain the greater tumor accumulation of ¹⁷⁷Lu-LCP in the UMUC3/3T3 tumors shown in Figure 4B. The UMUC3/3T3 bladder cancer tumor model was therefore chosen for more detailed mechanistic analysis.

3.7: Mechanisms of Tumor Growth Inhibition: Cytotoxicity Studies

While many chemotherapeutic therapies and treatments using external radiation require multiple doses to generate therapeutic effect, just one dose of ¹⁷⁷Lu-LCP provides continuous treatment of the tumor as the ¹⁷⁷Lu slowly decays, leading to sustained tumor growth inhibition for several days. We desired to investigate the mechanisms by which this sustained effect occurs, and hypothesized that the cytotoxic β-emissions from ¹⁷⁷Lu would persistently induce double-stranded DNA breaks. To study this phenomenon, we looked further into the stroma-rich bladder cancer tumor model. Mice were sacrificed one, two, and four days after treatment with 200 μCi of ¹⁷⁷Lu-LCP or appropriate controls. Immunofluorescent staining for p-H2AX, a protein that is phosphorylated in response to DNA double-stranded breaks [21], showed that the maximum induction of the protein occurred two days after treatment and remained elevated even four days after treatment (Figure 8A). This DNA damage led to cell death via apoptotic fragmentation of DNA, which was most induced four days after treatment as measured through the TUNEL assay (Figure 8B). This suggests that the cumulative dose of radiation from ¹⁷⁷Lu-LCP causes maximum DNA damage after ~48 h, and this damage translates to apoptotic cell death shortly thereafter. ¹⁷⁷Lu-LCP induces significantly higher maximum H2AX phosphorylation and cell death than free ¹⁷⁷Lu (p<0.05), and the difference in the pattern of cytotoxic effects between these two treatments may be due to a difference in pharmacokinetics/dynamics, as free ¹⁷⁷Lu quickly clears from circulation and may deposit into the tumor much earlier, more transiently, and with a different intratumoral distribution than ¹⁷⁷Lu-LCP.

Figure 8.

Cytotoxicity of ¹⁷⁷Lu-LCP in vivo: Mice were treated with 200 μCi of ¹⁷⁷Lu-LCP or controls and were killed one, two, or four days after administration. Figure panel A shows p-H2AX-positive cells (red) in UMUC3/3T3 tumor sections with or without DAPI (blue). Figure panel B shows TUNEL-positive cells (green) in tumor sections with and without DAPI (blue). Sections in Figure panel B are taken from an adjacent section to those displayed in Figure Panel A. Five representative fields were quantified for TUNEL or p-H2AX-positive cells. Images appearing in Figure panels A and B represent the field with the third-highest quantification. Average values are shown in Figure C: highest p-H2AX upregulation (red) occurred two days after ¹⁷⁷Lu-LCP treatment, while highest TUNEL induction (green) occurred 4 days after ¹⁷⁷Lu-LCP treatment. *p < 0.007; **p < 0.0005 compared to control. Scale Bar = 100 μm.

3.8: Mechanisms of Tumor Growth Inhibition: Microenvironment Studies

Changes in the tumor microenvironment were observed as well: Figure 9 shows that a diffuse and disordered fibroblast structure was present two days after treatment with ¹⁷⁷Lu-LCP, in contrast to the organized fibroblast structure present in untreated and control-treated tumors of the same size (see higher magnification images in Figure 9 D–E). These images clearly show that the fibroblasts in untreated tumors can organize into a defined morphology to form tumor nests, while those fibroblasts in tumors that have been treated with ¹⁷⁷Lu-LCP are dispersed and without this mature organization. Four days after treatment, the disorganized fibroblast structure persisted (Figure 10 A–D), along with changes in the collagen structure in the tumor, in which the long swaths of collagen in the untreated tumor had been replaced with shorter and more tortuous collagen fibers (Figure 10 E–H). It is known that the organized fibroblast structure and more ordered, linearized, and bundled collagen fibers present in the untreated tumor are associated with a more mature and malignant phenotype, while a disordered fibroblast structure with short, tortuous collagen fibers like those of the treated tumor suggest a tumor microenvironment less structurally and functionally capable of growth and progression [22–26].

Figure 9.

UMUC3/3T3-GFP (green) tumor sections, with tumor nests outlined in white. Row A) Untreated tumors. Row B) t = 2 days after treatment with cold DiI-¹⁷⁵Lu-LCP (red). Row C) t = 2 days after treatment with hot DiI-¹⁷⁷Lu-LCP (red). Row D) High magnification images t = 2 days after treatment with cold DiI-¹⁷⁵Lu-LCP (red). Row E) High magnification images t = 2 days after treatment with hot DiI-¹⁷⁷Lu-LCP (red); disorganized fibroblast structure does not allow tumor nest formation. Each row shows representative photographs from each group, consisting of several tumors and tumor sections. Scale Bar = 300 μm for rows A–C and 50 μm for rows D–E.

Figure 10.

Microenvironment structure in UMUC3/3T3 tumor sections t = 4 days after treatment with ¹⁷⁷Lu-LCP: A–D) Stain for fibroblast marker α-SMA (green) in UMUC3/3T3 tumor sections. Figures A and C show α-SMA expression in an untreated tumor and in a tumor at t = 4 days after treatment with ¹⁷⁷Lu-LCP; scale bars = 200 μm. White dashed boxes are zoomed in in Figures B and D; scale bars = 20 μm. E–H) Trichrome stain in tumors that are untreated (E) or t = 4 days after treatment (F). Inserts G–H on bottom right have been processed to show a more isolated view of the collagen stain. Arrows show the location of the region in the insert. Scale bar = 50 μm.

We also desired to measure the intratumoral localization of Lu-LCP. Unlike traditional chemotherapeutics that must enter the target cell in order to impart their therapeutic effect, internal radiotherapy can damage cells at a distance several cell lengths from the source. β-particles from ¹⁷⁷Lu, for example, have a maximum energy of ~500 keV, allowing an average path length in tissue of 200–300 μm [27] and a maximum path length of ~4x that value [28]. ¹⁷⁷Lu-LCP may then be able to provide therapeutic effect to cells that did not take up the particles, which is especially important given the intratumoral heterogeneity of nanoparticle delivery in general. We supposed that a majority of cells in a tumor would be “in range” for β-particle damage after ¹⁷⁷Lu-LCP distribution, but desired to measure the intratumoral localization of ¹⁷⁷Lu-LCP in order to calculate this important spatial relationship between cells and particles. The lipid membrane of the LCPs used in the experiment described in Figures 10 and 12 were labeled with fluorescent DiI, and excess DiI was purified away using sucrose gradient centrifugation in order to measure the dispersion of the particles throughout the tumor. Tumor sections from several areas in different tumors were then analyzed to quantify distances between cells and nanoparticles. Figure 11 shows that nearly 100% of cells in the field are within just 50 μm of the nearest Lu-LCP, well within the average path length of a β particle, even though the only nanoparticles considered are those present in that 2-D plane—Lu-LCPs residing in the z-direction would only increase the particles in range for a given cell. This substantiates the claim that although not all tumor cells have taken up particles, all are susceptible to the delivered therapeutic payload.

Figure 12.

Healthy nude mouse organs fixed, sectioned, and stained with hematoxylin and eosin 12 days after treatment with 250 μCi ¹⁷⁷Lu-LCP or controls to test cumulative toxicity. Some acellular regions (arrow) are visible in spleens treated with ¹⁷⁷Lu-LCP. Scale bar = 150 μm.

Figure 11.

Quantification of DiI distribution at t = 2 days: A) UMUC3/3T3-GFP tumor given DiI-¹⁷⁷Lu-LCP; scale bar = 500 μm; B) Isolated DiI signal from panel A; C) Heat map showing each cell’s distance from its nearest Lu-LCP, where black and blue areas show areas of close proximity between cells and particles; D) Quantification of relative cell distance from Lu-LCP; n = 1.75×10⁶ cells in 8 sections; n = 4 sections each for ¹⁷⁷Lu-LCP and ¹⁷⁵Lu-LCP.

3.9: Toxicity Studies

To gauge the cumulative toxicity of ¹⁷⁷Lu-LCP in mice, we treated healthy immunocompromised nude mice and healthy immunocompetent CD-1 mice with a single dose of 250 μCi of ¹⁷⁷Lu-LCP, the highest dose used in this study. Mice were sacrificed 12 days after treatment to assess the cumulative effect of the radiation. At the conclusion of the 12-day study, mouse organs were dissected and sectioned, serum markers for kidney and liver toxicity were measured (ALT, AST, BUN, and creatinine), and whole blood from the immunocompetent CD-1 mice was analyzed for changes in white blood cell count (WBC), hematocrit (HCT), mean cell volume (MCV), red blood cell count (RBC), hemoglobin count (HGB), and platelet count (PLT). These results were compared against appropriate controls and against the normal range for these values. Morphologies of kidney, liver, heart, lung, skin, and small intestine sections after H&E staining showed no difference from normal tissue in nude mice (Figure 12) or in CD-1 mice (Figure S4). Serum markers for renal and hepatotoxicity also remained in the normal range in both nude and CD-1 mice (Table 2A–B) after treatment. The data do show some well-tolerated side effects as a result of this treatment, including differences in spleen morphology as well as a decrease in white blood cell and platelet counts outside the normal range. Hematological side effects are common during cancer therapy and are easily handled using immune boosters, etc., which were not used in this study. None of these side effects translated to a decrease in mouse body weight (Figure S2).

Discussion

In this paper, we have described a novel method for systemic internal radiotherapy by extending the anticancer application of our LCP platform to include a theranostic nanoparticle with high specific drug loading. The preferential uptake of ¹⁷⁷Lu into LCP allows a huge amount of ¹⁷⁷Lu to be encapsulated in a very small batch of particles without changing the LCP’s morphology or encapsulation efficiency. The difficulty of nanoparticle scale-up that is pervasive throughout this field is therefore eliminated, as a therapeutic dose for a mouse or a human can be made on the same benchtop.

The theranostic advantage of radioisotope loading into LCP should also not be understated. ¹⁷⁷Lu’s γ emissions permitted live in vivo imaging using SPECT to measure tumor accumulation of ¹⁷⁷Lu-LCP, paving the way for future delivery to orthotopic tumors that will represent a more clinically relevant “diagnostic” procedure. The β decay from ¹⁷⁷Lu also causes significant tumor inhibition via DNA double-stranded breaks while simultaneously inducing Cerenkov radiation for imaging. Because therapeutic β-emissions are also the imaging contrast agent, Cerenkov imaging is a wonderful example of a theranostic application. Compared to SPECT, Cerenkov imaging correctly estimated the general biodistribution of ¹⁷⁷Lu-LCP at a fraction of the time and cost. This opens up opportunity for several applications such as dynamically looking at early stages of biodistribution and clearance by taking many images in quick succession. Because the high loading of Lu into LCP is broadly applicable to several other polycationic radiometals, such as the β-emitter and effective Cerenkov agent ⁹⁰Y, it is worthwhile to discuss Cerenkov imaging as an interesting and useful theranostic application.

As a follow-up to our tumor inhibition data, we were interested in assessing any significant changes in the tumor microenvironment as a result of this treatment, and we evaluated these changes by measuring the fibroblast and collagen structure in the tumor. The changes in the tumor microenvironment may be explained by looking deeper into how β-radiation causes the double-stranded DNA breaks that are shown to be elevated in the treated tumors. Generally, β-radiation is classified as having low linear energy transfer (LET); that is, the β-particles deposit a small fraction of their energy per unit length as they travel along their path [28]. In most cases, the DNA damage does not occur via direct interaction between a DNA strand and the β-particle itself. In the case of low LET emitters like ¹⁷⁷Lu, it is known that most DNA damage is caused by an increase in reactive oxygen species (ROS) generated as the emitted β-particles interact with and radicalize water molecules along their path. The induced ROS then damage DNA and cause single and double-stranded breaks in an “indirect radiation effect” [28].

Oxidative stress is known to cause many changes in the cellular and extracellular environment, especially in rapidly remodeling tissues like cancer, where actin and collagen are just some of the proteins susceptible to oxidative damage [29–33]. Actin is an important structural component of cells and is also a structural player in the adherens junctions between 3T3 fibroblasts [34]. The cysteine residues on actin are susceptible to oxidative damage, and this damage can prevent polymerization, cause depolymerization, and affect the organization of actin in cells, potentially restricting the organization of fibroblasts in the tumor. Collagen stability is also affected by oxidative damage, as scissions in the collagen α-bands can decrease the degradation temperature of the collagen to below body temperature and prevent the formation of collagen fibrils [30, 31]. These results suggest that cumulative ROS damage may cause the observed changes in the overall in vivo collagen and fibroblast structure shown in Figures 10 and 11, prohibiting the dynamic formation of bundled collagen and organized fibroblast structures that have been implicated in the formation of resistant tumor nests found in more advanced and aggressive tumors.

Tumor growth inhibition via ¹⁷⁷Lu-LCP may then occur by taking advantage of both the uncontrolled growth and the excessive microenvironment remodeling observed in cancer. The low, continuous dose of radiation damages tumor cell DNA that can have fewer active repair mechanisms and divide much more rapidly than healthy cells while inhibiting the progression of the microenvironment toward malignancy. This allows sustained cancer treatment at a dose that preserves the health of the subject.

Part of the beauty of this approach to radioisotope loading is its broad application for many trivalent radiometals, such as other therapeutic radionuclides like ⁹⁰Y or ¹⁹²Ir, which are great candidates for future work in this subject. Combination therapy using a radiotherapeutic and a radiosensitizing chemotherapeutic residing in the same LCP particle should also be possible, given that many small molecule drugs have previously been encapsulated in LCP.

Future work should also investigate whether systemic administration of a radiotherapeutic LCP is the most effective dosing method. Macro and micro-dosimetric calculations should be used to determine absorbed doses in the mouse and how the distribution of nanoparticles necessitates different calculations than doses from an external beam. In one potential local delivery scenario, highly loaded ¹⁷⁷Lu-LCP, ¹⁹²Ir-LCP, or ⁹⁰Y-LCP could be directly injected into prostate tumors to create a sort of nano-brachytherapy, in which modified particles would disseminate locally and stably in the prostate but remain out of the systemic circulation.

Conclusions

Radionuclide loading into ¹⁷⁷Lu-LCP can be increased by several orders of magnitude without affecting the ¹⁷⁷Lu encapsulation efficiency or LCP morphology, generating a nanoparticle with high radioisotope concentration that overcomes scale-up barriers typical of nanotherapeutics and minimizes heavy metal-based toxicities. The choice of ¹⁷⁷Lu as a model radionuclide has allowed in vivo anticancer therapy via DNA damage and microenvironment remodeling, in addition to permitting live animal SPECT/CT and Cerenkov imaging. As studies in LCP-mediated radionuclide therapy continue, other nuclides may be considered, perhaps in combination with chemotherapeutics, for several different therapeutic approaches.

Supplementary Material

1

Figure S1: ¹⁷⁷Lu-LCP and Free ¹⁷⁷ Lu-Cl₃ Pharmacokinetics. Data represented as mean ± SEM. n = 5 for ¹⁷⁷Lu-LCP and n = 3 for free ¹⁷⁷LuCl₃.

2

Figure S2: Relative change in body weight in H460 tumor-bearing mice used in tumor growth inhibition experiment shown in Figure 7B. Mice did not significantly lose or gain weight during the experiment.

3

Figure S3: A) 24 h after injection with DiI-¹⁷⁷Lu-LCP, UMUC3/3T3 tumors show more homogeneous tumor accumulation of DiI-labelled NP (red). B) Untreated UMUC3/3T3 tumors have more vasculature than H460 tumors, shown as more CD-31 positive area (orange).

4

Figure S4: Healthy CD-1 mouse organs fixed, sectioned, and stained with hematoxylin and eosin 12 days after treatment with 250 μCi ¹⁷⁷Lu-LCP or controls to test cumulative toxicity. Scale bar = 150 μm.

5

Table S1: ¹⁷⁷Lu-LCP Characterization. Diameter, zeta potential, and PDI measurements taken in triplicate after sucrose gradient purification.

6

Table S2: Biodistribution of ¹⁷⁷Lu in mice shown in Figures 5–7. Mice were sacrificed after SPECT and Cerenkov measurements at t=24 h and organs were harvested. Left columns show organ accumulation in mouse given free ¹⁷⁷LuCl₃. Right columns show organ accumulation in mouse given ¹⁷⁷Lu-LCP. Organ biodistributions measured using gamma scintillation.

Abbreviations

AA

Anisamide

DOPA

1,2-dioleoyl-sn-glycero-3-phosphate

DOPC

1,2-dioleoyl-sn-glycero-3-phosphocholine

DSPE-PEG2000

N-(Carbonyl-methoxypolyethyleneglycol 2000)-1,2-distearoyl-sn-glycero-3-phosphoethanolamine, sodium salt

EDS

Energy-Dispersive X-ray Spectroscopy

EE

Encapsulation Efficiency

K_sp

Solubility Product Constant

LCP

Lipid-Calcium-Phosphate

Lu

Lutetium

SPECT

Single Photon Emission Computed Tomography

TEM

Transmission Electron Microscopy