MR Cholangiography Demonstrates Unsuspected Rapid Biliary Clearance of Nanoparticles in Rodents: Implications for Clinical Translation

Abstract

Due to their small size, lower cost, short reproduction cycle, and genetic manipulation, rodents have been widely used to test the safety and efficacy for pharmaceutical development in human disease. In this report, MRI cholangiography demonstrated an unexpected rapid (<5 min) biliary elimination of gadolinium-perfluorocarbon nanoparticles (approximately 250 nm diameter) into the common bile duct and small intestine of rats, which is notably different from nanoparticle clearance patterns in larger animals and humans. Unawareness of this dissimilarity in nanoparticle clearance mechanisms between small animals and humans may lead to fundamental errors in predicting nanoparticle efficacy, pharmacokinetics, biodistribution, bioelimination, and toxicity.

Background

Many years ago (around 1998), our laboratories began to study the use of 250 nm paramagnetic perfluorocarbon nanoparticles (NP) in rats. In these early studies, biliary excretion of these paramagnetic particles was observed a few minutes after intravenous injection, enabling the creation of MR cholangiograms, perhaps the first ever made using nanotechnology¹. The accumulation of NP in the intestines of rodents was observed innumerable times over subsequent years, but this biodistribution pattern was not appreciated in numerous larger animal models including rabbits^{2, 3}. The objective of this communication is to share these observations with the nanocommunity as well as some of the key scientific literature beginning in the late 1950’s that previously explored the issue of colloidal clearance in rodents, larger animals, and humans.

Materials and Methods

Preparation and characterization of perfluorocarbon NP

Paramagnetic perfluorocarbon NP were prepared as previously reported⁴. The emulsions were comprised of 40% (v/v) of perfluorooctylbromide (PFOB; Elf Atachem, specific gravity 1.93 g/ml), 2% (w/v) of a surfactant commixture, 1.7% (w/v) glycerin and deionized nanopure water representing the balance. The surfactant mixture consisted of 78 mole% high-purity (>95%) egg yolk phosphatidylcholine, 2 mole% phosphatidylethanolamine, and 20 mole% of the lipophilic gadolinium chelate diethylene-triamine-pentaacetic acid-bis-oleate (Gd-DTPA-BOA; Gateway Chemical Technologies). PFOB, glycerin, and the surfactant commixture were pre-blended with sonication and microfluidized at 14,000 PSI for 4 minutes (Microfluidics, Newton, MA, USA). PFOB particles in example 2 were produced with 20% v/v PFOB and excluded Gd-DTPA-BOA from the surfactant mixture, which was replaced with high purity (>95%) egg yolk phosphatidylcholine on an equimolar basis. Nominal particle sizes were 220±80nm using dynamic light scattering (Malvern Instruments, Malvern, PA) with a polydispersity index <0.2.

MR imaging

All animal studies were conducted in accordance with a protocol approved by the Animal Care and Use Committee of our Institutes

In example 1, 200 g female rats (rnu/rnu) were anesthetized by intraperitoneal (IP) injection of 100 mg/kg ketamine and 10 mg/kg acepromazine. Anesthesia was maintained by administering 25 mg/kg ketamine and 2.5 mg/kg acepromazine IP via a 20-gauge catheter (JELCO, Tampa, FL) every 30 minutes or as needed. Gd-DTPA PFOB NP were injected via tail vein at a dose of 0.05 mmole Gd/kg. MR images were taken before and immediately after injection over a time course of 1 hour. To this end, a Signa 1.5T clinical magnet (General Electric Medical Systems, Milwaukee, WI) was used equipped with a 5-inch circular surface coil and a field of view of 13 cm. A 3D time-of-flight (TOF) spoiled gradient echo (SPGR) sequence was used to obtain coronal images with the following parameters: Echo time (TE)=minimum, flip angle=45 degrees, slice thickness=1.0 mm, matrix=256×192, and number of averages=10. In example 2, PFOB NP were injected via tail vein into a 100 g male Harlan rat at 1.0 ml/kg and allowed to circulate for 2 hours prior to imaging. High-resolution ¹H/¹⁹F MR images were acquired at 3T (Philips Achieva) using an in-house, custom dual-tuned open birdcage transmit-receive coil^{5, 6}. Simultaneous 3D ¹H/¹⁹F imaging was used employing a novel steady state ultrashort echo time (UTE) technique (TE/TR=0.1 ms/1.96 ms) with the frequencies set to the resonance of ¹H and the CF₂ groups of the PFOB spectrum (representing 12 of 17 total ¹⁹F nuclei)⁷. Using a highly oversampled 3D radial readout scheme, the reconstructed image datasets have a nominal resolution of 1.25 mm^3,8.

Results

In example 1, T1-weighted TOF MR imaging of rats given Gd-DTPA PFOB NP revealed an unexpected rapid biliary excretion. This was appreciated as an MR biliary cholangiogram at 5 minutes following intravenous injection (Figure 1). At baseline, the heart, liver, spleen, and vasculature were not apparent but the gastrointestinal system could be observed secondary to the presence of digesta. One minute post-injection, the blood pool was opacified with NP, clearly revealing the heart, aorta,, and large peripheral vasculature. Highly vascularized organs, such as the liver and spleen, also became visible. Five minutes post-injection, the biliary flow of the injected NP could be clearly appreciated passing from the liver along the common bile duct into the small intestine. The anatomical lack of a gall bladder in the rat was apparent, as these animals do not possess this organ.

Figure 1.

3D TOF MR cholangiogram in a rat following intravenous injection of gadolinium-functionalized PFOB NP. (A) Baseline image showing no evidence of vasculature or common bile duct. (B) 1 min post-injection, the blood pool as seen in the heart, aorta (arrow), and peripheral vasculature have strong T1-weighted positive contrast from the NP that are still constrained within the vasculature. H=heart, L=liver. (C) Within 5 min, the NP are rapidly excreted through the common bile duct (arrows), reflecting the rapid shunt of contrast from the liver into the small intestine. (D). After 10 min the small intestine contains most of the contrast, that is passed on to the large intestines at 30 min (E) and 60 min (F).

In the second example, PFOB NP were injected via tail vein into a rat and simultaneous dual ¹H/¹⁹F MR images were obtained as a 3D stack. Figure 2 presents 4 sequential coronal images of that acquisition. At two hours, a low-level blood pool signal (i.e., colorized green) was appreciated within the cardiac ventricles. As expected, PFOB particles were heavily biodistributed into the liver, producing intense signal in that organ. Consistent with the gadolinium-enhanced images in Figure 1, an abundance of ¹⁹F rich PFOB was localized in the small intestines. A tubular ¹⁹F-rich reference standard of PFOB NP suspended in agarose is seen on the animal’s right. Collectively, these images strongly suggest that entire NP are passing directly and rapidly into the biliary tree and their constituents are flowing into the intestines after i.v. injection.

Figure 2.

(A–D). Sequential 3D overlayed ¹H/¹⁹F MR images of a rat 120 min following intravenous injection of PFOB NP. NP accumulated in the liver and intestines.

Discussion

In the present report, rapid biliary excretion of relatively large PFOB NP (~250nm) was observed with T1-weighted 3D TOF and dual ¹H/¹⁹F UTE MR imaging techniques. In man, NP less than 5–6 nm can be bioeliminated through the kidneys and urine, but are otherwise they are retained primarily by the clearance organs belonging to the mononuclear phagocyte system (MPS) until degraded into smaller constitutive elements, which may be metabolized further or eliminated directly via bile, urine, or respiration.^9–12 Although the rapid transit of large, otherwise vascular-constrained particles into the bile and gut was never observed with MRI previously, the scientific appreciation of this phenomena occurred over a century ago.

In 1958, Hampton¹³ referenced the 1881 work of Rütimeyer who intravenously administered particles of cinnabar in rodents and reported their distribution into Kupffer cells, hepatic cells, and bile. Hampton further noted that these early results were later corroborated by Müllendorf in 1916 and Weatherford in 1956 using colloidal dyes and India ink, respectively. Hampton found that colloidal particles (8% HgS in 500 μl or 25% ThO₂, Thorotrast) injected intravenously localized to the liver and spleen with transmission electron microscopy (TEM). Within the liver, they were found in Kupffer cells, hepatocytes, and the biliary tree. Particles phagocytosed by Kupffer cells were retained in those cells indefinitely; whereas particles endocytosed by hepatocytes were transported rapidly into the biliary system. Since systemic protein coating could influence the handling of the particles in the liver, Hampton performed biliary retrograde infusions of Thorotrast colloidal particles. TEM revealed that the particles transited in reverse through hepatocytes into the space of Disse. Some of the Thoratrast was sequestered in Kupffer cells while much passed into the circulatory system. Other evidence of NP excretion into rodent bile and feces include silica particles from 50–200 nm¹⁴, citrate-coated silver particles (~8 nm)¹⁵, and iron oxide core high-density lipoproteins (~10 nm)¹⁶.

In contradistinction to rodents, Juhlin reported in 1960¹⁷ that fluorescent spherical hydrophilic particles of methyl methacrylate injected intravenously in rabbits did not transit effectively into the bile. A 192 mg intravenous injection of particles (20 to 110 nm) produced a minimal biliary concentration of 0.05–0.40 mg/ml during the first minutes after i. v. injection. Fluorescent particles in the range of 60–140 nm were not appreciated in the bile. Further, particles 200–800 nm in diameter were not found in the bile despite a ~10-fold increase in injected dose (1,200 mg).

Rodents are frequently used to characterize the pharmacokinetics, pharmacodynamics, biodistribution, and bioelimination of nanomedicines for basic science and IND regulatory purposes. While these models offers a wealth of research opportunity, the data achieved with nanotechnology must be viewed with an appreciation for the impact of rapid biliary clearance. Rapid excretion of particles along with their therapeutic or imaging payloads substantially lowers the toxicity burden imposed on the animal and can leave a better impression of biosafety than exists. Allometric scaling¹⁸ to project pharmacokinetic data from rodents to man may be inappropriate, since the concept of shape scaling implies that the physiological pathways are conserved. Non-metabolizable particles greater than the renal threshold can pass into the feces of rodents leading to potentially false conclusions of high bioelimination and safety.

In summary, we have presented MRI examples of rapid biliary secretion of relatively large PFOB NP. While the MR cholangiogram using NP may have been the first ever produced in 1998, the recognition that particles can be excreted into the feces via bile in rodents dates back to the 1950s, with the early insights published in the late 19^th century. Today, given the rapid expansion of nanomedicine research, it is incumbent upon us to recognize and adjust to the potential translational issues that the use of rodent preclinical models pose on assessments of nanomedicine safety and efficacy studies.