|Home | About | Journals | Submit | Contact Us | Français|
A variety of polymer nanoparticles (NP) are under development for imaging and therapeutic use. However, little is known about their behavior. This study examined pharmacokinetics, distribution and elimination of stable polyacrylamide (PAA) nanoparticles (~31 nm average diameter). PAA-NPs and polyethylene glycol-coated PAA-NPs were injected into the tail veins of healthy male rats. Blood, tissues and excreta were collected at times ranging from 5 minutes to 120 hours and their radioactive content was quantified. A mathematical model was then applied to analyze the distribution dynamics of both NPs. Elimination from the blood could be accounted for by a quick but finite relocation to the major organs (about 20%, 0.6 to 1.3h half-lives), and a slower distribution to the carcass (about 70%, 35 to 43h half-lives). Excreted urinary levels correlated with blood concentrations. Combined cumulative urinary and fecal output accounted for less than 6% of the dose at 120h. Compared to five other polymeric nanoparticles, the studied particles are at the highest half-lives and Area Under the Curve (4000 to 5000 %-h). These two parameters decrease by three orders of magnitude when nanoparticle size increases from the 30 nm range up to 250 nm. For similar sizes, pegylated nanoparticles are more persistent in the blood than non pegylated ones, but this difference is much smaller in the 30 nm and relatively high dose range than above 100 nm. Persistence of PAA NPs is not associated with acute toxicity signs as measured by typical serum markers of inflammation and cellular damage.
Polyacrylamide-based (PAA) nanoparticles (NPs) have been engineered for the intracellular delivery of sensors and/or pharmaceutically active agents. Use of this flexible polymer platform minimizes perturbations to the cell while providing protection to the embedded chemicals from the intracellular environment. Among other characteristics, PAA NPs combine the advantages of a core that can be loaded with a wide variety of chemicals and a surface that can be covalently functionalized with targeting molecules.
Given this high versatility, multifunctional PAA-based nanosensors have been developed for a number of practical applications such as the quantification of intracellular analytes including several cations (Clark et al., 1999b; Sumner et al., 2002; Park et al., 2003; Sumner et al., 2006), reactive oxygen species (Poulsen et al., 2007), glucose (Xu et al., 2002), and pH (Clark et al., 1999a; Clark et al., 1999b). Recently, in-vivo applications such as the simultaneous magnetic resonance imaging (MRI) enhancement and delivery of photodynamic molecules for the early detection and treatment of cancer and other diseases have been investigated (Moffat et al., 2003a; Kopelman et al., 2005; Reddy et al., 2006; Koo et al., 2007). While these studies demonstrate the promising potential of PAA NPs as targeted imaging and therapeutic agents, their behaviors with respect to pharmacokinetics, whole-body distribution and toxicity in vivo have not been fully characterized. The present study addresses this point by providing a dynamic description of the distribution of non-functionalized and polyethylene glycol coated polyacrylamide nanoparticles in-vivo. Furthermore, we propose a mathematical framework that can be applied to toxicokinetic data from experiments with other types of nanoparticles to obtain a complete description of their bioaccumulation dynamics.
As demonstrated in a number of in vivo experiments, the reticuloendothelial system (RES) comprises a major clearance route for micrometer and nanometer-scale materials of biologic or anthropogenic origin (Moghimi et al. 2001). The rate of removal from the systemic circulation by macrophages comprising the RES is influenced by the binding of opsonins on their surface, while the addition of polyethylene glycol (peg) chains to the surface of NPs has been shown to influence this binding and extend circulation times by attenuating the opsonization process (Moghimi and Hunter, 2001; Zahr et al., 2006). The rate at which NPs are removed from the body depends on their rates of excretion and/or degradation, the latter predominantly occuring in the macrophage (Briley-Saebo et al. 2004; Fernandez-Urrusuno et al. 1996; Moghimi and Szebeni 2003).
Large differences in the biodistribution and pharmacokinetic of polymeric nanoparticles have been reported in the literature (see table S5, supporting information). For PLGA (poly(lactide-co-glycolide)) nanoparticles with sizes above 100 nm (154 nm), Panagi et al. (2001) reported a very rapid elimination from the blood, with the half life ranging from 13 to 35 s, whereas pegylated PLGA (PLGA-mPEG) nanoparticles of 113 nm had much longer half lives in blood, which were close to 7 hours for all tested doses. Gaucher et al. (2009) also observed an increase of half-lives in blood between non-PEG (0.03 h) and PEG (0.3 h) nanoparticles. Gratton et al. (2007) also found a relatively long half-life in blood for 200 nm Polymer Pegylated PRINT nanoparticles, observing a fast distribution, mainly to the liver and the spleen, with an apparent half-life of approximately 17 min., which was followed by slow distribution with a half-life of 3.3 h. The increased persistence of pegylated NPs is also confirmed in the reviews of Owens and Peppas (2006) and of Li and Huang (2008).
Owens and Peppas (2006) indicated that larger nanoparticles tend to be removed more quickly than smaller nanoparticles as long as their molecular weight exceeds 5000; for lower molecular weights, renal excretion could dominate. For such large NPs, 20–30% of the dose throughout the time period studied was found in the carcass, and 70–80% of the PLGA dose was distributed between blood, liver and spleen, suggesting that the removal of the particles from blood was due to their capture in the mononuclear phagocyte system. However, few in vitro experiments have been carried out for polymeric nanoparticles range below 100 nm. For smaller nanoparticles, such as 20, 30 and 60 nm pegylated nanogel and nanolatex nanoparticles, Yang et al. (2009) reported a longer blood half-life, between 15 and 29 hours. The nanoparticles demonstrate biphasic kinetics, but additional information is needed to understand the relative magnitude of possible subsequent transfers.
To complement these studies, there is therefore a need to:
To address these needs, this paper presents an innovative biodistribution study of stable polymeric NPs after intravenous injection in rats. The distribution of radioactive materials related to two different formulations (original “naked” PAA and pegylated PAA NPs) are compared at multiple time-points after intravenous injection, and in a variety of tissues, organs, and excreta. In addition, a comprehensive mass balance analysis is presented, aimed at helping to identify the source and target compartments while controlling for the dose difference between both experiments, such as difference in the dose injected.
Synthesis of PAA and PAApeg NPs has been previously described in detail (Clark et al. 1999a) and has been adapted here for incorporation of [14C]-acrylamide. A water:oil microemulsion was generated by stirring phosphate buffer (pH 7.4) and hexanes under argon. The amine-functionalized NPs were polymerized by addition of tetramethylethylenediamine (TEMED) and ammonium persulfate (APS) to catalyze the free radical mediated cross-linking of acrylamide (17% w/v) with N-(3-amino-propyl) methacrylamide (4% w/v). [14C]-Acrylamide (American Radiolabeled Chemicals, St. Louis, MO) was incorporated into the reaction mixture and the specific activities reached were 0.92 and 1.81 μCi/mg NP for naked and pegylated NP, respectively. Coating of NPs was achieved by attachment of polyethylene glycol (MW 2,000) to available amine moieties on the surface of the NP (Moffat et al., 2003b). Residual surfactants were eliminated by repeated washing with ethanol and water, and the final suspension filtered and stored at 4°C until use. The size-distribution of the nanoparticles was measured by dynamic light scattering (DLS) on nanoparticles produced with the method described above, except for the fact that non-radiolabeled Acrylamide was used. The NPs size distribution is log-normal, with an average measured size of 30.9 nm and a standard deviation of 14.2 nm.
Male Crl CD®(SD)IGS BR rats (weighing 218-241 g, Ch arles River, Portage, Michigan) were acclimatized for seven days. During this time, the rats were observed daily and clinical signs of disease were monitored. All animals were examined visually and weighed prior to dosing. During the study, rats were housed individually in metabolic cages and monitored for morbidity, mortality, and injury. Fluorescent lighting was provided for 12h per day. Temperature and humidity were maintained between 21-23 °C and 48 to 70%, respectively. Diet (block Lab Diet® certified Rodent Diet® #5003, PMI Nutrition International, Inc.) and tap water were available ad libitum.
Five groups of three rats per group received a single iv dose of non-pegylated radioactive NPs ([14C]-PAA NPs) at an average dose level of 11.3 mg-eq./animal (corresponding to 44.6 μg-eq./g) and five additional groups of three rats per group received pegylated radioactive NPs ([14C]-PAApeg NPs) at an average dose level of 7.0 mg-eq./animal (28.2 μg-eq./g). Both formulations we re administered via the tail vein and resulted in a radioactivity content of approximately 10 uCi/animal, with a constant dose volume of 1 ml/kg. Urine, feces, cage residues, and blood samples were collected at designated times and tissue samples of the liver, spleen, kidney, heart, lungs, brain, lymph nodes (mesenteric, inguinal, and popliteal), and bone marrow (both femurs) were collected following euthanasia by CO2 (for exact times of collection see supporting material table S1).
Urine and feces were collected over ice, frozen at −20°C and protected from light until analyzed. Blood samples were collected from all animals prior (0.4 ml) and following (2 ml) dosing via the jugular vein except for the final interval, where collection was achieved by post mortem cardiac puncture. Selected tissues, organs and the rest of the body were collected, blotted dry, weighed, and stored at −20 °C until analyzed. Cages were washed daily with deionized water, and analysis indicated that only negligible traces of radioactivity were present in cage residues and were thus not included in the mathematical analysis.
The concentration of radioactivity in whole blood, excreta, tissues, and cage residues was determined by Liquid Scintillation counting (LSC). Whole blood, urine, and cage rinses were analyzed directly by mixing weighed aliquots with the scintillation cocktail (Ultima Gold, PerkinElmer). Feces (25 to 50% homogenates), tissue samples and carcass were analyzed for total radioactivity following digestion in an organic tissue solubilizer (Solvable, PerkinElmer). Radioanalysis was performed for at least five minutes or 100'000 counts using QuantSmart (version 3.01, PerkinElmer) on a PerkinElmer Tricarb 2900 scintillation counter. Chemiluminescence was corrected using the following standard for determining background. A blank sample (scintillation fluid only) was run in the LSC with every set of samples. The LSC then automatically subtracted that blank sample as a background from each of the subsequent samples. Typical background number were between 12-18 dpm. As a standard samples (blood, feces, and tissue homogenates) that could potentially have chemiluminescence were oxidized. Additionally the LSC has a chemiluminescence correction feature built-in.
Fluorescent PAA and PAApeg NPs were prepared as described above with the difference that no radiolabel was incorporated. Instead, dextran (10'000MW) linked AlexaFluor® 488 and 543 (Molecular Probes, Eugene,O R) were encapsulated into the polyacrylamide matrix at the cross-linking step. This incorporation is achieved because the dyes are hydrophilic and therefore migrate naturally into the polyacrylamide matrix, which is hydrophilic as well. After preparation, NPs were rinsed 5x with 50 ml ethanol and 50 ml sterile H2O, and the solutions were filtered (0.22 um). Doses of 5, 50, and 500 μg/g of fluorescent PAA and PAApeg were injected to healthy male Fisher 344 rats (8-12 weeks), Charles River Laboratories, Wilminton, MA) via the tail vein, and typical markers of inflammation were measured in serum 14 days post-injection.
The NPs suspensions were incubated in isolated plasma for 1h and 24h at 37°C. After incubation periods, NPs were pelleted at 15,000g for 1h and re-suspended with a sterile 0.9% NaCl solution to prevent non-specific binding. The post incubation steps were repeated 3x and proteins were solubilized off the NPs using a 10% SDS and 2.32% DTE solution. 1:100 dilutions of the isolated proteins were loaded per lane after a 5 minute incubation at 95°C min and separated by SDS- PAGE. Two identical gels were run in parallel, one stained with Sypro Ruby Red (Molecular Probes, Eugene, OR) and the other with Coomassie blue. Protein identification was achieved by band excision from the Coomassie gel, protein digestion with trypsin, MALDI-TOF Mass spectrometry and peptide mapping of the most abundant proteins.
Mass balance modeling was performed to investigate the consistency between the measured fluxes, the observed nanoparticle masses and to generate hypotheses on the transfer mechanisms involved. The exploratory model contained four peripheral compartments connected to a central blood compartment and was built to gain kinetic information on the observed transfer of NPs between compartments in the body of rats.
The four peripheral compartments describe the cumulative urine excreted, the cumulative feces excreted, and the accumulation in the organs (fast kinetic) and carcass (slow kinetic) (Figure 1). The final mathematical analysis consisted of a dynamic mass balance equation for which the initial mass injected of a given nanoparticle (Minjected) is equal to the sum of the mass in blood (Mblood), the cumulated masses excreted in urine (Murine) and feces (Mfeces) and the mass deposited in tissues with fast (Mfastd) and slow kinetics (Mslowd), all masses in mg-eq./animal:
More specifically, available data allowed the transfer of material between the blood to the urine and the feces to be analyzed in detail, and assumptions concerning the loading of the carcass and the organs were then made to reconstitute the full mass balance. The blood was first modeled by a bi-exponential function characterizing the concentration of radioactive material over time. The function was fitted against the concentrations measured in blood averaged by groups, and a value of 6.5% (weightblood/weightbody mass) (Chisolm, 1911; Mitruka and Rawnsley, 1981) was used to convert the concentrations to the total amounts in blood where needed (as well as the average weight of all the animals tested in the study: 253.4 gr). The kinetics of loading of other compartments were described by a set of ordinary differential equations (ODE) describing the transfers between the blood to the urine, the feces, the tissues with fast kinetic of deposition comprising the organs, and the tissues with slow kinetic of deposition which corresponds to an undefined deposition site within the remaining carcass. For each compartment, several model formats were assessed and their parameters optimized against measured concentrations and flows using the non-linear Ordinary Least Square Regression method (OLS). The guiding principle for the selection of the final model was that the most parsimonious format structure was held when multiple models fitted the data to similar extents (as measured by the R2 value). Most of the modeling part has been done with Berkeley Madonna 8 (University of California Berkeley, http://www.berkeleymadonna.com/).
When needed, concentrations measured in tissues, organs, and excreta, where converted to total equivalent masses by assuming that full organs, tissues, and excreta were collected, and by multiplying measured concentrations by their respective collected amounts (see supplementary tables S2 and S3 for raw masses and concentrations retrieved).
Results of PAA and PAA-PEG nanoparticles are compared to other experiments, by expressing all Areas Under the Curve in common units (%-h).
Serum clinical chemistry values were assessed 14 days after exposure to doses of 5, 50, and 500 μg/g of PAA or PAApeg NPs as one indica tion of toxicity. With exception of modest but significant elevation in alkaline phosphatase in the 500 μg/g groups for the two NPs formulations, all other parameters were within normal ranges and/or not significantly different from controls (table 1). No signs of toxicity as indicated by overt changes in behavior or by significant weight loss were noted. Histopathologic evaluation of the lungs, liver, kidneys, heart and spleen did not indicate toxicity attributed to NPs at any of the administered doses. Confocal micrographs of sectioned tissues show that uptake of the NPs were apparent in the liver and spleen, suggesting that organs of the RES are recognizing the NPs after vascular exposure. Imaged sections of the kidneys did not show any signs of fluorescence at the site of glomerular filtration or throughout the rest of the kidney (data not shown).
Numerous past studies have suggested an association between the addition of surface coating agents such as polyethylene glycol and increased plasma residence tims (Panagi et al., 2001; Owens Iii and Peppas, 2006). The alteration of PAA NPs surfaces by adding polyethylene glycol to inhibit adsorption of serum/plasma proteins from their surface was determined after in vitro incubation with plasma, elution of the bound proteins from the NPs, and SDS-PAGE (figure 2). More details on the proteins identified can be obtained in the supplementary table S4.
Quantitatively, incubations of 1h gave similar results when compared to 24h incubations for each NP preparation. Surface modification with peg caused a decrease of the quantity of adsorbed protein compared to PAA NPs, however the pattern of the SDS-PAGE was not significantly altered. The identification of selected peptides reveals the binding of classic opsonizing proteins, such as fibronectin and apolipoproteins (figure 3 and table S4) (Peracchia et al., 1999; Kim et al., 2007).
The concentrations of radioactive material remaining in blood for PAA or PAApeg NPs intravenously administered were determined after blood collection and radioactivity quantification from 0 to 120h (figure 3, see table S1 for details on collection times).
In spite of the distinct dosage levels, the concentrations of radioactive material in blood related to both formulations exhibited similar patterns and were characterized by a rapid decrease of about 20% of the initial mass during the first hours (apparent half life of 0.6h – PAA NP, to 1.3h – PAApeg NP). This initial phase is followed by a slower decline (apparent half life of 35h – PAA NP, to 43h – PAApeg NP) down to 5% of the initial dose at the end of the experiment. Blood concentrations corresponding to the non–pegylated formulation varied from 7.03 mg-eq./g at 5 minutes postdose, to 0.034 mg-eq./g at 120h postdose. At identical times, measurements related to the pegylated experiment varied from 0.447 to 0.031 mg-eq./g. The observed trends showed a bi-exponential behavior (eq. 2).
Where Cbloodi (t)(μmg − eq./gblood) is the function describing the concentration of radioactive material in blood over time, A1 and A2 are parameters with units of (μmg − eq./gblood), and are apparent initial and terminal half lives (h), and t is the time postdose (h). Optimized parameters estimates and 95% confidence intervals are given in table 2 and the fitted models are depicted in figure 3. The sum of A1 and A2 equals the injected mass for each animal, whereas A1 represent the initial decreases of 21% and 19% of this injected mass. The resulting Area Under the Curve was calculated by extrapolation up to infinite.
The amounts of radioactive material excreted in urine over the predetermined periods were monitored at regular intervals for individual animals and cumulative profiles for both formulations were obtained (figure 4).
Totals of 0.13 and 0.04 mg-eq. (PAA and PAApeg, respectively) were excreted through urine during the first 120h, representing only 1.2% (PAA) and 0.5% (PAApeg) of the injected doses. Although the urine excretion exhibited a biphasic pattern, a model assuming that the mass of material excreted over time was proportional to the concentration of radioactive material in blood failed to describe the data. An alternative model capable of adequately reflecting the excretion of radioactive material through urine (dMurine (t)/dt mg-eq/h) is described by equation 3.
This model structure represents a flow of radioactive material responsible for the depletion of an initial reservoir of readily excreted material (Mpool(0), mg-eq.), supplemented by a flow proportional to the concentration in blood. Parameter estimates for Mpool(0), kpu (1/h) and kbu (1/h) are presented in table 3. Radioactive materials excreted from Mpool(0) were rapidly depleted and were estimated to represent 0.6% and 0.1% of the injected doses for the PAA and PAApeg formulation, respectively. The rate of depletion of this initial pool is governed by a common rate constant (kpu) governing the rate of depletion of Mpool(0), adequately fitting both formulations. The second process that was proportional to the mass of nanoparticle in blood was responsible for the loss of 0.7% and 0.4% of the initial doses after 120h. While these estimates represent small fractions of the total injected material, the consistency between individual measurements over time allowed them to be precisely determined and revealed that the rates of transfers were slightly higher for PAA when compared to PAApeg.
A hypothesis generated by this sub-model is that NPs from the low end of the size distribution were undergoing glomerular filtration. This hypothesis is supported by the observation that a smaller fraction of the pegylated formulation was present in this initial pool of material undergoing excretion through urine (pegylated NPs are thought to be slightly larger and due to the peg addition, therefore, the size distribution is shifted toward higher sizes). However, it is plausible as well that small fractions of the radioactive material injected was not firmly attached to the NPs, and were thus of sizes compatible with glomerular filtration. Finally, the formulation dependent rate constants describing the excretion that was proportional to the blood concentration might indicate that degradation/erosion was occurring in the blood, with a slightly slower rate associated with the PAApeg NPs when compared to the PAA formulation.
At 120h post injection, cumulative radioactivity recovered in feces represented 4.1% and 4.9% of the initial doses for PAA and PAApeg, respectively. A delay of approximately 8h between the dose administration and the collection of the first samples of feces containing significant amounts of radioactive material was observed (figure 5). The cumulative amount of radioactive material measured in feces (Mfeces [mgeq.]) was best characterized by a delayed elimination that is proportional to the blood content, supplemented by another flux that was constant over the considered period:
Where kbf (1/h) is a first order rate constant corresponding to the transfer of radioactive material from the blood to the feces, and ffix (mg-eq./h) is a zero order constant (i.e. a constant flux). t* (h) is the delay observed from the removal from the blood until the actual sampling and has been set to equal 8h, consistently with the delay observed between the dose administration and the initial recovery of radioactive material. Both rate constants were fitted without any further constraints.
At 120h post injection, Cumulative radioactivity recovered in urine and feces represented only 5.4% for both PAA and PAApeg. Therefore neither the urine, nor the feces excretions are able to explain the rapid decrease of 20% in the blood nor the more important but slower clearance from 80% down to 5% of the injected mass.
The concentration of PAA and PAApeg NPs was monitored in select tissues, including spleen, liver, lungs, brain, heart, kidney, bone marrow, and lymph nodes (mesenteric, inguinal, and popliteal). Tissues of animals that received non-pegylated PAA NPs were monitored after 1,2,4, and 8h (groups 1 to 4), whereas tissues of animals from the pegylated PAA (PAApeg) NPs experiment were monitored after 4, 8, 24, and 48h (groups 6-9). Residual radioactive materials in carcasses were measured at 120h postdose from groups 5 and 10. Detailed concentrations and corresponding masses are presented in figure 6 and in supplementary tables S1 and S2). Time intervals chosen for each NP suspension (PAA vs. PAApeg) were pre-determined based on past reports, indicating increased plasma residence times with the addition of PEG (Moghimi et al., 2001; Panagi et al., 2001; Avgoustakis et al., 2002; Gbadamosi et al., 2002).
The mass of radioactive material recovered in all organs combined (figures 6A and 6B) was quantitatively and kinetically consistent with the mass disappearing from the blood during the first phase of loss from the circulation. After this phase, the masses and concentrations measured in organs varied only slightly. Given this low variation, specific and combined organs radioactivity contents measured across samples (i.e at 1, 2, 4, and 8h for PAA, and 4, 8, 24, and 48h for PAApeg) were averaged, and calculated values were compared between both experiments (table 4). The amounts of radioactive material recovered in all organs combined were similar for the two formulations (average of 1.5 ±0.2 mg-eq., student t-test for the average between the two formulations: p=0.58), however, this value represents 12.7% ±1.7 % and 21.7% ±2.3% of the initial doses p <0.001) of PAA and PAApeg, respectively, suggesting that a non linear mechanism takes place during the distribution to the organs. The liver was the major site of deposition in the organs and its content represented 8.8% ±1.4% (PAA) and 16.9% ±2.1% (PAApeg) of the initial doses. Lungs and kidneys were secondary sites of deposition with less than 1.5% of the administered dose in each of these organs for both formulations. All the other organs taken together contained less than 3% of the initial dose. In term of concentrations, the majority of the organs contained levels of radiolabel that were roughly analogous (see figures 6C and 6D for measured concentrations), with the exception of the brain in which an approximate 10-fold reduction was observed.
For non-pegylated NPs, slight increases in concentrations over time were noted for the spleen, the heart, and the inguinal and popliteal lymph nodes whereas in the PAApeg NP experiment, the statistical analysis (t-test with alternative hypothesis being that the slope is different from 0) and visual inspection revealed no significant trends in the liver, the inguinal and the popliteal lymph nodes and bone marrow from 4 to 48h. However, lymph nodes and bone marrow exhibited an important inter-animal variability that possibly prevented significant trends from being observed. Other organs exposed to PAApeg had small but significant trends over time with decreasing concentrations in lungs, brain, heart, kidney and increases in the lymphatic system, spleen and mesenteric lymph nodes (see supplementary tables S2 and S3).
The largest fractions of radioactive material recovered 120h post-injection was found in carcasses, with 76% ±10% (PAA) and 64% ±20% (PAApeg ) of the initial doses recovered (table 4). Statistical testing failed to show a statistically significant difference between these values (student t-test, p=0.11). However, carcasses contents expressed in masses (8.2 ± 2.1 mg-eq. (PAA) and 4.6 ±1.3 mg-e q. (PAApeg)) were statistically different (p<0.01), suggesting that the extent of the accumulation in the carcass was proportional to the amounts of NPs initially injected.
Overall, the sites of depositions and trends exhibited during these experiments are consistent with published data. Particularly, deposition occurred predominantly in organs of the RES. However, the observed lack of a demonstrable effect on the distribution and pharmacokinetics of PAApeg NPs may simply reflect the need for longer polyethylene glycol chains to evade the RES (Hamidi et al., 2006).
At 120 hours postdose, 12.5 and 16.8% of the initial doses (for PAA and PAApeg, respectively) were accounted for by cumulating blood content, material excreted through urine and through feces. At this time, more than 80% of the dose was thus distributed in the organs and/or in the carcass.
Based on the pattern exhibited by the blood concentrations and the measurements in organs and carcasses, it was determined that two major mechanisms of distributions were taking place simultaneously in addition to the minor fluxes excreted in urine and feces described by equations 3 and 4. During the first phase, the rapid loss of radioactive material from the blood is consistent with the masses recovered in the organs of the RES at 1h for PAA and 4h for PAApeg (first respective sampling times). The slower kinetic observed in the second phase was assumed to be related to the distribution to tissues, in which deposition was slow (an unspecified location within the carcass), and by a flux that was proportional to the blood concentration. Accordingly, the following equations governing the distribution to organs, characterized by a rapid but saturable deposition mechanism, and other tissues exhibiting a slow deposition rate were postulated:
Where Mfastd (t) (mg-eq.) and Mslowd (t) are the formulation dependent cumulated masses in all the tissues with fast and slow deposition, kfastd (1/h) and kslowd (1/h) are the intrinsic rate constant governing the distribution in those tissues, Mmax (mg-eq.) is the maximum mass of radioactive materials the rapidly depositing tissues can accumulate. . This modeled mechanism of saturation preserves the critical features of a mathematical model previously proposed by Wilhelm et al. (Wilhelm et al., 2002) on the basis of in-vitro experiments carried out with anionic superparamagnetic nanoparticles on mouse macrophages (RAW 264.7 cell line) and HeLa cells, which is consistent with a Langmuir adsorption process associated with a finite capacity of the cell for internalizing its membrane surface. An additional constant flow originating from the fast deposition compartment has been linked to the excretion in feces (ffix) to preserve the consistency of the overall mass balance.
The exact origin of the apparent constant elimination described by ffix could not be identified but is likely to result of a first order process occurring in a compartment with stable concentrations of NPs. The low variation of the content accumulated in the organs of the RES over time might represent a potential source.
The values of kfastd, kslowd, and Mmax were calibrated based on a mass conservation assumption by minimizing the difference (OLS) between the average initial dose injected (11.3 and 7.0 mg-eq. for PAA and PAApeg, respectively) and the sum of the materials recovered in all the compartments (equation 1). The modeled cumulated dynamic distribution (Figure 7A and 7B) illustrates the rapid saturated increase in the organs followed by the slow but important distribution to the carcass. As illustrated by the constant total mass over time in these figures, the structures of the equations proposed for tissues with fast and slow deposition (equations 5 and 6) are consistent with the conservation of a stable cumulated mass in whole animals (including excreta).
The simulated time profiles for the tissues with fast and slow kinetics of deposition are in good agreement with the masses recovered in organs and carcasses, respectively. A stable mass in the organs of 2.1 mg of radioactive material was predicted for the overall organ contents for the PAA formulation slightly higher than experimental recovered masses of organs summed yielded 1.45 ±0.19 mg-eq. The model predicted a value of 1.4 mg-eq. for the final organ content in the pegylated experiment, which is within the range of the experimentally recovered value of 1.52 ±0.17 mg-eq. Published literature confirms that organs of the RES, particularly the liver and the spleen, are responsible for the rapid sequestration of polymeric nanoparticles after iv administration (Gref et al., 1994; Panagi et al., 2001). It is thus likely that this compartment represents the population of phagocytic cells in contact with the blood, whose largest population is Kupffer cells in the liver (Wake et al., 1989).
The mathematical model predicts that 68.9% (PAA) and 62.7% (PAApeg) of the administrated dose would be sequestered in slowly depositing tissues after 120h, in agreement with the measured values of 75.8 ±10.4% (PAA) and 64.2 ±20.0% (PAApeg).
Apart from the initial pool in urine, there is no remarkable dissimilarity between the parameter estimates describing the biodistribution of PAA and PAApeg NPs, suggesting similar dynamic behaviors in-vivo. Some rate constant estimates however, are slightly smaller for the pegylated formulation but it is unclear whether these differences are biologically significant, attributable to a slight increase in NP diameter due to the addition of the peg chains themselves, or related to altered surface properties and consequent variations in protein binding.
An apparent saturation of the organ capacity, mainly driven by the materials accumulating in the liver, is observed in the biodistribution study and subsequently suggested by the model. Such saturation implies a non-linear distribution with respect to the dose administered. This non-linear behavior has important implications when comparing experiments carried out at different dose levels, because it creates apparent differences in NP's distribution. To illustrate this effect and to propose a further testable hypothesis, the final model has been modified by explicitly modeling the concentration in blood, instead of using the empirical relationship presented in equation 2. In this prospective model, the mass in blood is specified by the difference between the initial mass of nanoparticles injected and the masses departing from the blood compartment (specified by the solutions of equations 2, 3, 4, and 5). The modeled concentrations in blood were then used as inputs of equations 2, 3, 4, and 5 under the constraint that the total mass of material was being constant (equation 1), to obtain a speculative model predicting the tissue distribution that would have been observed at 120h in case of different initial doses (figure 8). At low doses most of the nanoparticles are rapidly captured by organs. At higher doses, the quick saturation of the organs leads to a much higher fraction of nanoparticles slowly distributed to the rest of the carcass (figure 8).
The toxicity, biodistribution, pharmacokinetics and excretion pathways of PAA and PAA NPs in rats were investigated. The dose of nanoparticle administered to each rat was based on the specific (radioactive) activity and the need to achieve a minimum dose of 10 uCi per animal. This resulted in a 61% higher dose for the PAA cohort when compared to PAApeg, however, the compartmental analysis indicated that once the saturation of certain organs were accounted for, there was no major kinetic difference between the two formulations.
Although the acrylamide monomer used in the synthesis of PAA NPs is a neurotoxin (Dearfield et al. 1988; Friedman 2003), its polymerized form is non-toxic and does not degrade back into its monomeric chemical form (Caulfield et al. 2002). Previous experiments with PAA NPs up to a dose of 500 μg/g over 4 weeks showed no visible alterations in histopathology or clinical chemistry values indicating acute toxicity by the NPs. Because of PAA long-term stability, high biocompatibility (Lemperle et al. 2003), and flexibility for incorporating targeting moieties, it is a promising material for medical applications. Consistently, our acute toxicity study indicates that PAA nanoparticles do not induce detectable organ damages or other measurable signs of toxicity. The modest elevation in alkaline phosphate could suggest impairment of liver, intestine or bone dysfunction although histopathological studies do not indicate toxicity at these targets. In addition, other clinical chemistry markers in the study do not support toxicity of these systems and the modest elevation in alkaline phosphatase is just beyond the published reference range (200-300IU/L) for Fisher344 rats at this age. Future, chronic toxicity studies and multiple exposure studies will be important due to the slow degradative process of the polyacrylamide matrix.
The SDS-PAGE analysis of the proteins bound to either the PAA or PAApeg NPs revealed similar patterns, although proteins associated with naked PAA were more abundant. Apolipoproteins were detected, and it has been previously reported that this class of proteins is involved in the translocation of poly(butyl-cyanoacrylate) (PBCA) NPs across the blood brain barrier (Kreuter et al., 2002). In this study from Kreuter et al, translocation was specifically triggered by apolipoprotein-E and B, which were not detected in our experiments. However, the presence of other detectable proteins and their concurrence with published studies suggests that failure to detect other key proteins may be a product of the limits of detection inherent in the technology employed.
The pharmacokinetic analysis shows an initial rapid transfer from blood to organs with half-lives of 0.6 (PAA) and 1.3h (PAApeg). The liver was the major deposition site, followed by the kidneys and the lungs. The concentration in the popliteal lymph nodes increased steadily and was the most concentrated tissue at 48h postdose, whereas inguinal and popliteal lymph nodes contained modest concentrations. The lowest concentrations were measured in the brain. Part of an explanation for the important liver confiscation might be that it receives approximately 30% of the cardiac output and contains the largest fraction of the total fixed macrophages (Kupffer cells) in the body (Wake et al., 1989). The total recovery in organs accounted only for less than 22% of the total dose administered.
The initial transfer from blood to organs was followed by a slower distribution to the carcass representing more than 75% of the initial dose with a kinetic of appearance of radioactive material at this location corresponding to the kinetic of disappearance observed in the blood (half-life of approximately 40h). Figure 9 compares the maximum half-life in blood and the Area Under the Curve (AUC) obtained in the present study to results for five other polymeric nanoparticles reported in the literature (see table S5, supporting information). The residence times observed in our study were typically longer than those retrieved in experiments with other polymeric NPs such as poly(lactic-coglycolic acid) (PLGA or PLGA-mPEG) nanoparticles (half-lives of 35 sec and 7.0 h, respectively)(Panagi et al., 2001), PLA (Gaucher at al., 2009; Liu, 2008), PRINT-PEG (Gratton et al., 2007), and Nanogel and Nanolatex (Yang et al., 2009). The studied PAA and PAA-PEG particles are at the highest half-lives and AUC. These two parameters decrease by two to three orders of magnitude when nanoparticle size increases from the 20 to 30 nm range up to 250 nm. For similar sizes, pegylated nanoparticles are always more persistent in the blood than non pegylated ones, but this difference is much smaller in the 30 nm and relatively high dose range of the present experiment than in the range above 100 nm.
Neither PAA nanoparticle formulation resulted in significant amounts of recovery in urine and feces after 120h (less than 6% excreted from the body after 120h). This long residence time in the body is not surprising, based on studies reporting the low biodegradability of cross-linked polyacrylamide (Smith et al., 1996; Smith et al., 1997), and the fact that most of the NPs used where too large to undergo glomerular filtration. However, the synthesis of most NPs results in a suspension of NPs of slightly varying diameters that can be defined by a Gaussian distribution. As a result, NPs parts of a single suspension, but having different diameters, may not follow identical kinetics or tissue distribution profiles. Particles with a diameter smaller than 10nm are typically capable of being filtered by the renal system. Based on the mean diameter of the PAA NPs used here (mean diameter 30 nm), renal filtration of the particles may filter a small percentage of the particle suspension in the low end distribution (<10nm), consistently with what was observed.
On a therapeutical level, the rapid saturation observed in the organs of the RES suggests that a pre-treatment with a type of NPs that can be rapidly degraded, immediately followed by the injection of long-lasting PAA NPs would allow the latter to escape first pass capture by the macrophages, and thus extend their circulation time. Finally, the confiscation of PAA and PAApeg NPs by the organs of the RES, probably involving the saturation of their resident macrophages, raises the question of a potential increased susceptibility of treated subjects to superinfections, as macrophages are believed to be major players of body's defense against foreign pathogens.
The authors gratefully acknowledge the technical assistance of Jennifer Fernandez, Rhonda Lightle and Yong-Eun Lee Koo.
This work was supported by grants from the National Institutes of Health/National Cancer Institute [grant number N01-CO-37123 to R.K.], National Institute of Environmental Health Sciences [grant number R01 ES08846 to M.A.P] and National Institute of General Medical Sciences [grant number R01 EB007997 to R.K.].
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest statement
The authors declare that they do not have any competing interests regarding the toxicology issue of the manuscript. Dr Martin Philbert and Ram Reddy nevertheless disclose that a patent related to PEBBLEs (probes encapsulated by biologically localized embedding) is pending under their name.