Toxicity and Biomedical Imaging of Layered Nanohybrids in the Mouse

Abstract

Layered nanohybrids (LNH) are a promising nonviral system allowing controlled drug and DNA delivery. In order to test the toxicity of LNH consisting of a magnesium/aluminum core, mice were subjected to subcutaneous, intraperitoneal, and intravenous injections of these nanoparticles at three doses. Intravenous injections resulted in 8% (1 out of 12) lethality at doses 100 μl and 200 μl of 6.96 × 10⁻⁴ M solution, while all mice survived after LNH administration by any other routes. Histopathological alterations were limited to mild localized inflammatory lesions in the lungs and the dermis after intravenous and subcutaneous administration, respectively. LNH labeled with Lucifer Yellow were readily detectable in both locations by fluorescent microscopy. To test their potential for intravital imaging, LNH-Lucifer Yellow were injected into the ovarian bursa and successfully visualized by multiphoton microscopy within the ovarian surface epithelial cells. In similar experiments, the ovary and the ovarian bursa were readily detectable by magnetic resonance imaging after administration of modified LNH, where aluminum was substituted for gadolinium. Taken together, these results demonstrate minimal in vivo toxicity of LNH and illuminate their potential as multifunctional nanoscale particles suitable for combination of intravital biomedical imaging with controlled drug release.

Introduction

Nanotechnology is expected to be one of the leading technologies of the future. The reduction of material size from micro- to nanoscale offers benefits to diverse scientific fields and has the potential to revolutionize medical diagnostics and care (Stix, 2001; Roco, 2003). The development of multifunctional nanoparticles for biomedical and biotechnological applications may improve cancer therapy, DNA transfection, intravital imaging, targeted drug delivery, and enzyme immobilization (La Van et al., 2002; Vijayanathan et al., 2002; Hirsch et al., 2003; Ferrari, 2005; Wagner et al., 2006). Though there is a growing literature on application of nanoparticles and nanotechnology, only limited information on the biological effects of nanoparticles on cells and tissues and their potential risks to humans and the environment is available (Colvin, 2003; Peter et al., 2004; Service, 2004; Chen and von Mikecz, 2005; Holsapple et al., 2005; Oberdoerster et al., 2005; Donaldson et al., 2006).

Layered nanohybrids (LNH) represent a promising class of therapeutic delivery systems. These nanoscale platform particles are based on a layered inorganic host that can intercalate various biological molecules into the nanometer size galleries between the layers. The host consists of positively charged layers of a mixed divalent/trivalent hydroxide (e.g., magnesium/aluminum as well as other cation combinations described later). The inorganic core of an LNH is like a nanoscale deck of cards. Various molecules can be incorporated between the cards while the outer surface of the deck can be treated and conjugated to different molecules for targeting. LNH are based on self-assembly, a robust, common approach in biological systems. Traditional uses for LNH have focused on both medical and nonmedical applications, for example LNH based on Mg/Al have been used for years as oral antacids.

More recently, several reports describe the use of LNH in other pharmaceutical/medical applications including gene and drug storage, drug and gene delivery and enzyme immobilization (Ambrogi et al., 2001, 2003; Choy et al., 2000a, 2000b; Khan et al., 2001; Hussein et al., 2002; Kwak et al., 2002; Tyner et al., 2004a, 2004b). Recently, LNH were developed for delivering the nonionic, poorly water-soluble drug camptothecin or whole gene including promoter using magnesium-aluminum layered double hydroxide (Tyner et al., 2004a, 2004b). These studies on glioma and choriocarcinoma cell lines and primary culture of cardiac myocytes showed that the nanohybrids were well tolerated, caused no pronounced toxicity and delivered the cytostatic drug and the functioning gene in all cell lines tested. Furthermore, they also demonstrated that LNH could be successfully functionalized and coupled with an antibody. However, biological toxicity of LNH in vivo remained unclear.

In this communication we perform initial testing to evaluate LNH toxicity in vivo. We further demonstrate that, in addition to their value as vehicles for controlled drug release, LNH may be suitable agents for biomedical imaging, including such advanced technologies as multiphoton microscopy and magnetic resonance imaging.

Materials and Methods

Experimental Animals

FVB/N inbred mice and mice with floxed copies of p53 and Rb genes were maintained in our laboratory animal facility under 12 hours light/dark cycle at a temperature of 20°C and relative humidity of 20–50% and were monitored daily following recommendations of the Institutional Laboratory Animal Use and Care Committee of Cornell University.

Preparation of LNH

LNH were synthesized according to a previously described method (Tyner et al., 2004a, 2004b). In brief, positively charged magnesium-aluminum (Mg/Al) double hydroxide forming the layered nanoparticle cores was mixed with the potassium salt of Lucifer Yellow fluorescence dye. Ion exchange allowed the incorporation of the negatively charged dye into the nanometer size galleries between the layers. Nanoparticles with inorganic gadolinium/magnesium (Gd/Mg) double hydroxide core were synthesized similarly.

Administration of LNH

LNH were administered to two to four months old mice intravenously (IV), intraperitoneally (IP) or subcutaneously (SC) at 6.96 × 10⁻⁴ M in 50, 100, and 200 μl of distilled water and mice were euthanized 3, 7, 21, and 35 days postinjection. For each concentration at least 3 animals were used per each time point in each treatment group. As a control, mice were injected with distilled water and collected over the same time schedule.

Pathological Analysis

Moribund and scheduled for material collection mice were anesthetized with avertin (2.5% v/v in 0.85% NaCl, 0.020 ml/g body weight) followed by CO₂ euthanasia and subjected to careful pathological evaluation. Brain, lungs, liver, kidney, spleen, pancreas, tight muscle, eye, and skin were fixed in phosphate-buffered 4% paraformaldehyde and representative specimens were processed routinely and embedded in paraffin as described previously (Flesken-Nikitin et al., 2003; Zhou et al., 2006). Our preliminary results demonstrated that LNH containing Lucifer Yellow (LNH-LY) can be equally well detected both in frozen and paraffin material. Thus, parallel paraffin sections were used for both hematoxylin and eosin (H&E) staining and LNH detection. Distribution of LNH was assessed in either unstained or DAPI counterstained sections mounted with GEL/MOUNT (Biomeda Corp. Foster City, CA), sealed with Clarion Mounting Medium (Biomeda Corp. Foster City, CA), and imaged by fluorescence microscopy (Carl Zeiss Axioskop 2).

Primary Culture of the Ovarian Surface Epithelium (OSE)

Individual ovaries were dissected from mice with floxed copies of p53 and Rb genes, placed in DMEM/F12 (Ham’s) medium containing Collagenase-Dispase at 5% CO₂ for 1 hour and expanded as described previously (Flesken-Nikitin et al., 2003). Inactivation of p53 and Rb was achieved by recombinant adenovirus-mediated expression of Cre recombinase and gene excision was monitored by PCR as in (Flesken-Nikitin et al., 2003). Primary cell cultures of OSE deficient for p53 and Rb were incubated with 6.96 × 10⁻⁴ mol of LNH-Lucifer Yellow for 3 to 6 hours, washed 3 times with PBS, fixed with 4% paraformaldehyde, mounted with GEL/MOUNT and imaged by fluorescence microscopy 4 days after LNH administration.

Administration of Nanohybrids to Mouse OSE in vivo

Ten μl of LNH was delivered into the ovarian bursa by transinfundibular injection with a Hamilton syringe and a 30-gauge beveled needle under the control of a dissection microscope after deep anesthesia with intraperitoneal avertin as described previously (Flesken-Nikitin et al., 2003). As a control, a vehicle solution was injected into the contralateral ovarian bursa.

Multiphoton and Magnetic Resonance Imaging

For multiphoton imaging freshly dissected ovaries were placed in saline and imaged within 10 minutes after dissection as described previously (Flesken-Nikitin et al., 2004; Zipfel et al., 2003a). Multiphoton imaging was accomplished with a water immersion Olympus XLUMPlanFl 20x/0.95NA objective and excitation at 780 nm. The tissue emission was spectrally resolved into 2 channels; a 360–500 nm band (pseudocolored yellow) was used to image tissue structure: cellular autofluorescence and second harmonic generation from collagen, and a 500–650 nm band (pseudocolored green) was used to detect the Lucifer yellow fluorescence. For magnetic resonance imaging (MRI) mice were anesthetized with avertin, placed on a heating pad and imaged in a 2 Tesla MRI scanner at the Cornell University Hospital for Animals 3 days after injection of Gd/Mg LNH.

Results

LNH Toxicity and Biodistribution

To assess the potential in vivo toxicity of LNH, groups of mice were subjected to single IV, IP, and SC injections with 3 different doses of nanoparticles as described in the Materials and Methods. One mouse per each group of 12 (8%) died within 3 days after IV administration of 100 μl and 200 μl at concentration 6.96 × 10⁻⁴ M of LNH. Necropsy findings included extension of alveolar capillaries by blood, alveolar septal edema, and focal intraalveolar hemorrhage. These features were consistent with acute pulmonary congestion, most likely due to the capillary obstruction by nanoparticles. No abnormal clinical signs or behaviors were detected in the remaining mice of the same groups, as well as in all mice of 50 μl IV group and all groups subjected to IP or SC injections.

Mice were euthanized 3, 7, 21, and 35 days postinjection and subjected to careful gross and microscopic pathological evaluation. Most organs and tissues from the treated mice showed no significant changes compared to corresponding samples from the control animals exposed to aqueous vehicle. Exceptions were the presence of inflammatory lesions at sites of subcutaneous injections and limited inflammatory lesions detected in the lungs of mice with IV administration of LNH.

Solitary infiltrates of mixed inflammatory cells including neutrophils, macrophages, lymphocytes and plasma cells were located in the dermis (Figure 1A) 3 days after SC injection. At 7 and 21 days postinjection some lesions became somewhat larger and contained necrotic center surrounded by infiltrating cells and fibrous connective tissue. The lesions become smaller and more compact by 35 days after injection. In agreement with these observations, LNH-specific fluorescence was observed within inflammatory lesions 3 (Figure 1B), 7 and 21 days after LNH administration.

Three days after IV administration of LNH the lungs contained small perivascular inflammatory lesions consisting of neutrophils, macrophages and lymphocytes (Figure 1C). By 35 days postinjection these lesions were smaller in size and mainly composed of macrophages (Figure 1D, 1E).The number of lesions decreased in a dose-dependent manner, and only a few of them were detected in the lungs of mice exposed to 50 μl of LNH. Bright and particulate fluorescence of LNH was observed within inflammatory lesions of the lung at all time points of observation including 35 days postinjection (Figure 1F). Punctate fluorescence was also detected in single liver Kupffer cells of IV-treated mice, although no pathological changes were observed. No specific nanohybrid fluorescence was detected in any other tissues at the time points and routes of application studied.

Biomedical Imaging of LNH

For assessment of potential clinical applications of LNH, recently established approaches for targeting ovarian surface epithelium (Flesken-Nikitin et al., 2003; Nikitin and Hamilton, 2005) were undertaken.

Since prior LNH toxicity and delivery studies on epithelial cells mainly utilized established cell lines, such as 9L glioma cells and JEG3 choriocarcinoma (Tyner et al., 2004a, 2004b), we tested the efficacy of LNH targeting in primary culture of mouse OSE cells acutely transformed by p53 and Rb inactivation. Four days after administration over 80% of OSE retained LNH with no detectable change in their morphology (Figure 2A–2D), extent of cell death or rate of proliferation (not shown).

To extend these studies to animal models, LNH were injected into the ovarian bursa of living mouse resulting in exposure of OSE in situ (Flesken-Nikitin et al., 2003). Using multiphoton microscopy (Zipfel et al., 2003a, 2003b), LNH were easily detected in the OSE 4 days after administration (Figure 2E–2F). Serial optical sections of the OSE confirmed intracellular location of LNH. Parallel histological evaluation did not demonstrate any significant toxicological effects of LNH within this time frame after their administration.

In order to test potential application of LNH in conjunction with MRI, the aluminum of the Al/Mg double hydroxide core was substituted with gadolinium, which is commonly used for clinical MRI as a contrasting agent. The intrabursal administration of Gd-based nanoparticles (n = 4) was tolerated well, resulted in increased contrast of the mouse ovary and demonstrated periovarian retention of nanopartocles 3 days after injection (Figure 3).

Discussion

With the exponential increase of new types and uses of nanoparticles there is clearly a call for more research on toxicology of nanomaterials in order to determine potential health risks (Service 2005; Maynard et al., 2006). The risk assessment studies on graphite, metal oxide and quartz nanoparticles revealed complex relationship between adverse biologic effects and particle composition, size, and other characteristics and their toxicity and adverse biologic effects (Tsuji et al., 2006). The toxicity of airborne nanoparticles is relatively well studied, but other routes of exposure, such as subcutaneous, intravenous and intraperitoneal administrations remain largely terra incognita.

The present study on the systemic and localized administration of LNH represents an initial evaluation of possible in vivo toxicological effects of these nanoparticles, as they have great potential for application as tools for delivery of anti-tumor and anti-inflammatory drugs and genes (Tyner et al., 2004a, 2004b).

The SC, IP and IV administration of LNH did not cause pronounced toxicity in the FVB/N mice, as demonstrated by their high survival and lack of clinical and behavior signs. This is in accord with the decade-long good safety record of aluminum and magnesium oxides and salts as vaccine adjuvants, antacids and laxatives (Verdier et al., 2005).

In the SC-treated mice LNH remained close to the injection site for 21 days in cutaneous and subcutaneous tissue. The microscopic changes in the site of SC injection of LNH resemble the local reaction after administration of aluminium-containing vaccines (Valtulini et al., 2005).

In the IP-injected mice, no specific fluorescence was detected on the peritoneum covering liver, spleen, pancreas, or kidney, as well as in the adjacent mesenteric lymph nodes and additional studies are needed to determine the fate of LNH administered in this fashion.

After IV application specific nanohybrid fluorescence was observed in chronic inflammatory lesions in the lungs and in macrophages in the liver. The lesions in the lungs were comparable in principle to those of rat lung exposed to inhaled nanoparticles reported by Tsuji et al. (2006), who established that macrophage-mediated immunological responses occurred following a transient (24-hour postexposure) neutrophil-associated inflammation. Further studies would be necessary for comprehensive examination of the uptake, metabolism and excretion of the layered nanohybrids similar to those done by Flarend et al. (1997).

The minor pathologic changes caused in lungs of mice treated IV with layered nanohybrids, as compared to the control tissues, speaks in favor of the relative safety for their use. It should be noted that tested doses of LNH administration far exceed those likely to be required for targeted delivery of therapeutic compounds. Toxic properties of LNH could be further attenuated by substituting aluminum for more biologically safe metals, such as iron.

Since LNH can incorporate fluorescent molecules, we used this approach to monitor LNH distribution within the body and to co-localize their presence with potential pathological lesions. After the initial experiments showed that LNH can be readily detected in tissues we extended these studies to evaluate the clinical potential of LNH as imaging and reporting enablers. Our work indicates that LNH are well tolerated both in primary OSE cell culture and OSE in situ. As compared to our previous work with adenoviral vectors (Riley et al., 1996; Flesken-Nikitin et al., 2003; Manor et al., 2003;) and lipid-polycation-DNA (LPD) formulations (Nikitin et al., 1999), LNH provide efficiency exceeding and approaching those of adenoviral vectors and LPD, respectively. Importantly, LNH fluorescence was detected even 35 days after IV administration indicating a possibility of longer intracellular retention of the vector than that of LPD (up to 2 weeks) and on par with that of adenovirus.

Fluorescent photon emission-based and magnetic resonance imaging are among the most common and practical methods for biological assessment of nanoparticles. The fluorescence approach provides an inherent subcellular resolution. But biological specimens are notoriously scattering within distances ~50 μm so that a relatively thick specimen will appear as a blur. Generally one obtains information histologically by fixing the tissue, sectioning it to ~1–5 μm slices and staining those slices. Multiphoton microscopy achieves subcellular resolution in optically thick specimens without having to section them. It works by a quantum mechanical trick in which the photoexcitation (and any phototoxicity or photobleaching) is confined to the focal plane. It enables high-resolution imaging of live mice with minimal photodamage (Zipfel et al., 2003a; 2003b). The availability of highly specific targeting vectors with bright and specific reporters allows imaging of cell-based processes with the promise of minimizing phototoxicity to clinically acceptable levels.

MRI is currently used in clinical practice. It provides high-resolution images of anatomical structures and allows functional assessment of organs and tissues. For example, it is used for estimating the partial pressure of oxygen (and determining the presence of hypoxic tumor cells (Kachur et al., 1999; Seddon et al., 2002), detection of apoptosis in tumors (Schellenberger et al., 2004), and evaluation of angiogenesis, tumor blood volume, and microvessel permeability (Marzola et al., 2003; Kiessling et al., 2004). Although the rate of image acquisition is significantly slower than photon emission-based methods, and structural resolution does not reach subcellular levels comparable to that of fluorescence-based imaging, MRI is one of the best methods for noninvasive imaging. For a comprehensive analysis, probes that can be detected by both strategies are ideal, such as the multimodal proteins described in Schellenberger et al. (2004).

Our current study demonstrated that layered magnesium/aluminum and magnesium/gadolinium-based nanoparticles can be successfully detected in the ovary by multiphoton and magnetic resonance imaging. Further functionalization of LNH with specific peptides and antibodies should allow imaging and monitoring ovarian and other tumors while subjecting them to selective anticancer drug delivery.

The results of the current study on biologic effects and biomedical applications of LNH demonstrate only minor toxicity and are encouraging with respect of their use as imaging enablers. Taken together with established properties of LNH as effective drug and nucleic acid delivery vehicles, this work provides a basis for further development of LNH as multifunctional nanoparticles suitable for a broad variety of biomedical applications.