Introduction

There has been increasing interest in respect of biomedical applicability of nanoparticles (NPs) over the recent years. The potential of these compounds is related with their structure that could serve as a platform for the design of drug delivery systems (Crommelin and Florence 2013; Kim et al. 2010; Nicolas et al. 2013). Consequently, the release of a therapeutic drug may be directed against a specific target like a tumour, reducing at the same time the exposure of the healthy organs. Other proposed applications of nanoparticles include their use as contrast or antioxidant agents or gene delivery systems, among others (Hyafil et al. 2007; Martin et al. 2010a, b; Menchon et al. 2012).

Mesoporous silica nanoparticles (MSiNPs) have been used in several areas such as catalysis, photonics, solar cells and others due to their high surface area, stability and porosity. Recently, MSiNPs have been proposed for several biomedical applications such as drug and gene delivery systems, biosensors, biomarker detections and others (Bitar et al. 2012; Tan et al. 2004; Wang et al. 2014; Wu et al. 2013). The high surface area of MSiNPs can be easily modified enabling the covalent addition of molecules or functional groups. In addition, MSiNPs are inert in drastic environments such as acidic/basic conditions and in consequence are resistant to metabolization. Finally, the porosity of MSiNPs can shelter the guest molecules or drugs that are protected from the environment by the structure of the nanoparticles. The incorporation or elimination of the guest molecule from pores occurs simply by diffusion.

To explore the biomedical applicability of a new compound, it is mandatory to carry out in vivo studies to evaluate its pharmacokinetics, pharmacodynamics and toxicological properties (Lavik and von Recum 2011). Several studies have reported an unfavourable pharmacokinetic profile for different nanocompounds. In consequence, considerable effort is focused on the improvement of pharmacokinetic profile of NPs (Duan and Li 2013). Usually, NPs suffer an elevated clearance from blood due to phagocytic activity of the reticuloendothelial system and filtration in pulmonary vessels that in turn could contribute to increase the toxicities of these compounds (Moros et al. 2013; Rojas et al. 2011, 2012). In this way, the in vitro toxicological studies indicated that MSiNPs are non-cytotoxic at low concentrations (Asefa and Tao 2012; Guadagnini et al. 2013; Uboldi et al. 2012). However, it can cause mild-to-moderate hepatotoxic effects in mice, when it was chronically administered for 7 days by accumulation into the Kuppfer’s cells (Lu et al. 2010). So, adjusting the pharmacokinetic profile of MSiNPs could contribute to minimize this problem.

Till date, several studies have evaluated the pharmacokinetic properties of different types of MSiNPs when they were administered intravenously. Some of them used whole animal florescence methods that could provide qualitative in vivo data of the accumulation of the nanoparticles in the different organs, but such methods are not suitable for performing quantitative studies and can only be applied to mice (Botella et al. 2011; Lu et al. 2010; Meng et al. 2011). In contrast, high-sensitivity quantitative pharmacokinetic data of MSiNPs have been obtained using radiolabeled nanoparticles with 14C, 75Se and 125I (Borchardt et al. 1994; Sakai et al. 2012; Xie et al. 2010). An alternative approach to evaluate the pharmacokinetic properties of MSiNPs consists in the ex vivo analysis of the different tissues and organs to quantify the amount of Si by inductively coupled-plasma mass spectroscopy. Although it is less sensitive than radiolabeled methods, this approach has been used by some authors to characterize the biodistribution of MSiNPs in rodents (Huang et al. 2011; Malfatti et al. 2012).

Positron Emission Tomography (PET) has proven to be useful for evaluating in vivo the biodistribution and pharmacokinetics of different nanomaterials (Guerrero et al. 2012; Karmani et al. 2013; Moros et al. 2013; Rojas et al. 2012). With this approach, dynamic quantitative data of the biodistribution can be easily obtained from the whole body with extremely high temporal resolution. In consequence, the biodistribution pattern can be well characterized using small groups of animals. The most common radioisotope used for this type of studies, 18-fluorine (18F), can be incorporated in a great variety of nanoparticles without altering significantly their chemical properties (Guerrero et al. 2012; Moros et al. 2013; Rojas et al. 2012). Previous works used PET to explore the biodistribution of MSiNPs radiolabelled with 64Cu, 124I or 89Zr (Chen et al. 2014; Kumar et al. 2010; Miller et al. 2014). The labelling strategy used in these cases modifies considerably the surface of the nanoparticles. The approach proposed in the present study causes less alteration of the surface of the nanomaterial and uses the more available PET isotope, the 18F.

Here, we sought to present a novel and simple strategy for labelling MSiNPs with the 18F isotope in order to characterize their pharmacokinetic properties by PET. For this purpose, MSiNPs were functionalized in their surface with ω-aminopropyl groups in which 4-[18F]fluorobenzoyl moieties have been attached. The resulting radiolabeled MSiNPs were intravenously administered to mice to study their in vivo biodistribution.

Experimental

Materials and methods

All the reactants for synthesis were purchased in Advanced Biochemical Compounds (Radeberg, Germany) Sigma-Aldrich (Madrid, Spain) and Scharlab (Sentmenat, Spain). QMA, C18 cartridges were obtained from Waters (Cerdanyola del Vallès, Spain), and animals from Charles Rivers (France). The Cyclotron used for 18F production was an 18/9 model from IBA (Louvain-la-Neuve, Belgium). Analytical and purification HPLC were Agilent 1100 series coupled to a diode array UV–Vis Agilent 100 detector (Madrid, Spain) and a Raytest GINA isotopic detector (Barcelona, Spain). Centrifuge used was a Hettich-Zentrifugen (Tuttlingen, Germany).

ω-Aminopropyl-functionalized silica nanoparticles (NH2SiNPs)

NH2SiNPs were prepared by starting from commercial colloidal silica nanoparticles (Aerosil 200, Evonik, 0.5 g) that were suspended in toluene (100 ml) by sonication and treated at reflux temperature with 3-aminopropyltriethoxysilane (100 mg, 0.45 mmol) for 2 h under inert atmosphere. Afterwards, the solid was removed by filtration, washed with fresh toluene (3 × 20 ml) and dried at the open air. The resulting NH2SiNPs was characterized by combustion chemical and surface area analysis, thermogravimetry, Dynamic Light Scattering, electron microscopy, and IR and solid-state 13C NMR spectroscopy.

N-succinimidyl 4-[18F]fluorobenzoate (18F-SFB)

The synthesis was done in Eckert and Ziegler modules following the methodology described by Mading et al. (2005). In brief, the 18F− produced in the cyclotron was trapped in a QMA cartridge. Then, it was extracted towards the reactor by transferring a mixture of 2 mg of K2CO3 and 1.8 mg of Kryptofix 222 into 1 ml of a mixture of H2O/CH3CN 1:1 through the cartridge. The reactor was heated at 100 °C under He flow and, the vacuum was applied for 5 min. Afterwards, 1 ml of anhydrous CH3CN was added, and the mixture was dried as before. A solution of 5 mg of 4-(tert-butoxycarbonylmethyl)phenyl trimethylammonium trifluoromethanesulfonate in 1 ml of anhydrous CH3CN was incorporated, and the mixture was stirred for 10 min at 90 °C. Then, 0.5 ml of 1 M HCl was added, and the reaction mixture was heated at 100 °C for 5 min. After cooling to 25 °C, the mixture was diluted with 9 ml of water and passed through a C18 HELA cartridge. Subsequently, 3 ml of CH3CN was passed through the cartridge, and the solution was sent to a second reactor. This elute was treated with 20 μl of 25 % methanolic Me4NOH in 500 μl of CH3CN and dried at 90 °C under continuous He flow and in vacuum. Drying was completed by the addition of 3 ml of anhydrous CH3CN. Afterwards, a solution of 15 mg of N,N,N′,N′-Tetramethyl-O-(N-succinimidyl)uronium tetrafluoroborate in 500 μl of anhydrous CH3CN was added, and the mixture was heated at 90 °C for 2 min. The mixture was cooled to 25 °C and diluted with 5 % of aqueous acetic acid. The reaction crude was purified in a semi-preparative HPLC, and the desired fraction was reformulated to obtain 18F-SFB in DMSO. Finally, quality control was performed by analytical HPLC using gamma and UV–Vis (254-nm) detectors. 18F-SFB was obtained with a yield of 37 ± 5 %, with chemical and radiochemical purities exceeding 98 % and 102 ± 7 GBq/μmol of specific activity, respectively.

18F-labelled MSiNPs (18F-MSiNPs)

2.2 ± 0.3 mg of NH2SiNPs was stirred at room temperature for 45 min with a solution composed of 200 ± 17 mCi of pure 18F-SFB in a mixture of 20 µl of DMSO and 200 µl of sodium citrate buffer at pH 11 (basificated with NaOH 3 M solution). Afterwards, 800 µl of H2O was added, and the mixture was centrifuged at 4000 r.p.m. for 10 min. The resultant solid was washed three times by centrifugation at 4000 r.p.m. using 1 ml of H2O. After checking the absence for any radioactivity in the supernatant of the last centrifugation, the washed 18F-MSiNPs were dispersed in 2 ml of the physiologic saline +5 % Tween 80 and sonicated for 5 min before injection. Using the described procedure, a radiochemical yield of 14.3 ± 7 % was obtained after washing.

Animals and 18F-MSiNPs administration

BALB/c mice weighting 28.5 ± 9.1 g (mean ± standard deviation; n = 9) were used. Animals were housed under controlled laboratory conditions with the temperature maintained at 21 ± 1 °C and humidity at 55 ± 10 %. Food and water were available ad libitum. Animal procedures were conducted in strict accordance with the guidelines of the European Communities Directive 86/609/EEC regulating animal research and were approved by the local ethical committee (CEEA-PRBB). Animals received and intravenous dose of 0.8 ± 0.2 MBq of 18F-MSiNPs resuspended in 0.15 mL of saline. The dose was administered in the lateral tail vein under light anaesthesia of isoflurane. After injection, four of the animals were immediately scanned in the microPET. The rest remained in their cages for a period of 120 min.

PET image acquisition and analysis

Immediately after injection, animals (n = 4) were placed in an animal-dedicated camera (microPET R4; Concorde, Siemens, Knoxville, TN) for dynamic whole body image acquisition. Whole body data were acquired for 120 min. During the entire acquisition procedure, anaesthesia was maintained with a facial mask and a concentration of 2.5 % of isoflurane vaporized in O2. PET data were corrected for non-uniformity, random coincidences, and radionuclide decay and reconstructed with a filtered back-projection algorithm into a matrix size of 128 × 128, voxel size of 0.85 × 0.85 mm and slice thickness of 1.21 mm. After reconstruction, volumes of interest were manually drawn for the different organs, individual time–activity curves were obtained and the mean organ time–activity curves were calculated. Activity concentration at the end of the acquisition was used to determine the percentage of injected dose for each gram of tissue (%ID/g).

Ex vivo study

At the end of acquisition, animals were euthanized, and the different organs and tissues were collected, weighted, and measured in a gamma ray counter (Wallac 1470 Wizard, Perkin-Elmer, Waltham, MA) for 18F radioactivity determination. The measured radioactivity was corrected for decay, injected dose, and the mass of the tissue sample to obtain the percentage of injected dose for each gram of tissue (%ID/g).

Results and discussion

Preparation of 18F-MSiNPs

NH2SiNPs were obtained starting from commercial colloidal silica (Aerosil 200, 200 m2 × g−1) that was functionalized on the surface by 3-aminopropyltriethioxysilane in toluene in the absence of moisture. The presence of OH groups in the MSiNPs surface allows the covalent anchoring of different molecules using several conventional reactions such as esterification or silylation, among others. So, in order to obtain amino groups on the previously prepared and purified MSiNPs surface, a silylation reaction between the ω-aminopropyl triethoxysilane reagent and the OH groups was performed (see Scheme 1). It is well known that triethoxysilanes are suitable reagents for surface functionalization under mild conditions in non-aqueous media of nanoparticles with OH group in their surface (Etgar et al. 2013). The success of the functionalization and the loading of aminopropyl groups can be determined by chemical analysis, thermogravimetry and spectroscopy. Then, NH2SiNPs were characterized by chemical analysis that shows the presence of C (8.36 wt%) and N (3.26 wt%) in the atomic proportion (2.991) expected for the aminopropyl functionalization. This loading of aminopropyl groups is also responsible in thermogravimmetric analysis for the loss of 12 wt% in NH2SiNPs in the temperature range from 150 to 400 °C (data not shown). The specific surface area of NH2SiNPs was determined by isothermal nitrogen adsorption to be 325 m2 × g−1, significantly larger than that of the starting commercial silica, reflecting the functionalization of the surface. The presence of the aminopropyl groups on the silica can be assessed by the IR and solid-state 13C NMR spectra of NH2SiNPs (Fig. 1). In IR spectroscopy, the presence of amino groups is responsible for the strong vibration at 3100 cm−1. 13C NMR spectra reveals the presence of three broad peaks at 45, 28 and 14 ppms corresponding to the chemical shifts of the CH2 groups at positions 3, 2 and 1, respectively.

Scheme 1
scheme 1

Synthetic methodology used for labelling MSiNPs with 18F

Full size image
Fig. 1
figure 1

FT-IR (a) and solid-state 13C NMR, (b) spectra of NH2SiNPs

Full size image

To determine the size of NH2SiNPs, transmission electron microscopy (TEM) was used, which shows that NH2SiNPs are constituted by a broad distribution of nanoparticles ranging from 30 to 130 nm, with most of them being in the range between 60 and 90 nm. This distribution of particle size is somewhat larger than that of the starting Aerosil silica between 50 and 80 nm and reflects a growth in the size as consequence of the treatment with aminopropyltriethoxysilane. Figure 2 provides a representative TEM image of the sample and statistical particle size distribution.

Fig. 2
figure 2

Representative TEM image of NH2SiNPs (a) and the particle size distribution analysis (b) of the sample based on the measurement of a large number of particles

Full size image

Once MSiNPs and NH2SiNPs were obtained, a sample of each one was submitted to Dynamic Light Scattering study in order to determine the size distribution profile of these nanoparticles in suspension and their zeta potential value. These data provide some information about the modification of the MSiNPs by anchoring the 3-aminopropyl groups. It was found that the hydrodynamic size of NH2SiNPs in water was about 400 ± 80 nm that is higher than the value determined by TEM (see Fig. 2), but this should reflect some partial agglomeration of the nanoparticles or, more likely the fact that there is a cloud of solvent molecules solvating the NH2SiNPs. Regarding stability of these nanoparticles, the zeta potential value at neutral pH was −21.5 ± 5 mV that is within the permissible limit to be considered a persistent colloid.

The rationale for the preparation of these modified MSiNPs with aminopropyl groups on their surface was to add covalently the 18F isotope by using the succinimidyl functionalization of the 18F-SFB synton (see Scheme 1). This strategy was chosen by the well-stabilized premise about the high reactivity of the succinimidyl compounds with amino groups which have been used in radiochemistry in some labelling procedures (Guerrero et al. 2012; Mading et al. 2005; Rojas et al. 2011). Thus, the free amino groups of NH2MSiNPs were reacted with succinimidyl groups to attach the isotopically labelled 4-(18F)fluorobenzoyl by using 18F-SFB that will serve to trace the biodistribution of 18F-MSiNPs.

This new methodology for labeling MSiNPs is more user-friendly and easier than previous methodologies, which used to label them with 18F under drastic conditions attaching them directly to the Si atoms. Thus, the current methodology performs only minor modifications in the NPs scaffold and allows to attach 18F in less steps (Lee et al. 2013; Sarparanta et al. 2011).

Pharmacokinetic assessment

The labelling of MSiNPs with 18F-SFB was achieved satisfactorily, and this strategy enabled the continuous evaluation of the pharmacokinetics of the radiolabeled nanoparticles in living mice for at least 2 h after intravenous administration by in vivo and ex vivo studies.

In vivo study

The PET images revealed that 18F-MSiNPs were not homogeneously distributed throughout the mice body. In contrast, it accumulated in spleen, lung and especially in liver (Fig. 3). In addition, two routes of excretion could be evidenced from the image analysis. High amounts of radioactivity appear in intestine and urinary bladder indicating that 18F-MSiNPs or their radiolabeled metabolic products were excreted through biliary and urinary tracts. Moreover, the radioactivity in liver slightly decreases during the acquisition in parallel with intestinal accumulation of the injected dose. Finally, in some animals, even the gall bladder could be observed as an area of high radioactivity concentration in the hepatic area confirming that the radioactivity content detected in the intestine was secondary to biliary excretion (image not shown). No other organ or tissues presented sufficient radioactivity accumulation to be identified in the PET image indicating that nanoparticle accumulation was nearly negligible.

Fig. 3
figure 3

PET image from mice injected with 18F-MSiNPs. Image was reconstructed from the accumulated data during the 120 min of PET acquisition. 1 Lungs, 2 Liver, 3 Spleen, 4 Intestine, 5 Urinary bladder

Full size image

Time–activity curves (Fig. 4) revealed that, immediately after injection, the 18F-MSiNPs became irreversibly trapped in the vascular system of the lungs and in the reticuloendothelial system in organs such as spleen and liver. This pharmacokinetic behaviour has been observed in other types of nanomaterials such as carbon and ceria nanoparticles. A possible explanation to this fact is that, since 18F-MSiNPs are partially soluble and can form aggregates, some nanoparticles could be recognized as foreign bodies by the macrophages, and in consequence, they could have been actively phagocyted by the reticuloendothelial system. This would explain the uptake observed in the spleen and a part of that in the liver. On the other hand, the uptake observed in the lungs is probably explained by the trapping of the nanomaterial using the pulmonary capillary system via size exclusion. The radioactivity in urinary bladder and intestine increased during the 2 h of the scan. Thus, in contrast with other nanoparticles, 18F-MSiNPs presented two routes of excretion: one through the biliary tract into the intestine and another by renal clearance from the blood.

Fig. 4
figure 4

Average of time–activity curves obtained from mice subjected to PET scanning (n = 4), revealing the pharmacokinetics of 18FMSiNPs during the 2 h after injection

Full size image

At the end of acquisition, the mean percentage of injected dose for cc of tissue (%ID/g) in the parenchymatous organs was 15.4 ± 4.5 for the liver, 4.2 ± 0.7 for the lungs and 9.5 ± 3.1 for the spleen. Higher concentrations of radioactivity were observed in the intestine and bladder with values of 27 ± 7.3 and 58.3 ± 13.2 %ID/g, respectively (Fig. 5a). The concentration of nanoparticles in blood was only of 2.2 ± 0.5 %ID/g. Considering a total blood volume of ten percent of the body weight, this implies that only a 5 ± 1.2 % of the injected dose remained in circulatory system at that time. To sum up, the plasmatic availability of the 18F-MSiNPs was low due to the trapping of nanoparticles in the lungs, the phagocytosis by the reticuloendothelial system and their active excretion into the bile and urine.

Fig. 5
figure 5

a Biodistribution data at the end of the PET acquisition obtained from the image analysis (n = 4). b Biodistribution data obtained during the 2 h after 18F-MSiNPs injection by ex vivo quantification of the radioactivity in the different tissues and organs of mice (n = 9)

Full size image

Ex vivo study

Results of the ex vivo study performed 2 h after 18F-MSiNPs injection were similar to those provided by the in vivo assessment. In fact, the highest concentration of radioactivity was detected in urine, followed by the spleen, liver and lungs. Intestine was not evaluated in this experiment. In this case, the mean %ID/g observed was 17.3 ± 7.2 for lungs, 28.2 ± 10.2 for spleen, 26 ± 8 for the liver and 136 ± 24.3 for the urine. However, the ex vivo study enabled the evaluation of blood, at 2.2 ± 0.5 %ID/g, and other organs such as brain that presented the lowest radioactivity concentration, at only 0.2 ± 0.1 %ID/g (Fig. 5b), thereby confirming that 18F-MSiNPs do not penetrate significantly into the blood–brain barrier. Even though, the in vivo and ex vivo biodistribution patterns were highly comparable, the concentration observed in the in vivo study was systematically lower than that of ex vivo analysis for all the organs. Given that these differences were more evident in small size organs such as spleen and lungs, it is reasonable to assume that the cause of this phenomenon is the partial volume effect. This is a well-known limitation of the use of PET in pharmacokinetic studies of small laboratory animals, which arises from the limited spatial resolution of the method that results in an underestimation of the absolute concentration of radioactivity. However, this limitation can be easily overcome by performing an ex vivo evaluation of the organs at the end of the acquisition. Using this approach, the biodistribution pattern, the routes of excretion and the pharmacokinetics in the different organs could be obtained in vivo with PET, and the precise %ID/g in each organ could be determined from the ex vivo analysis.

Conclusions

A novel and simple methodology to label appropriate amino functionalized MSiNPs with 18F-SFB has been achieved, thus enabling the in vivo pharmacokinetic evaluation of MSiNPs. These nanoparticles accumulate in the organs of reticuloendothelial system by phagocytosis, in the capillary network of the lungs by vascular trapping and were excreted into bile and urine. As a consequence of their excretion, their vascular trapping and the active phagocytosis of the reticuloendothelial system, the plasmatic bioavailability of 18F-MSiNPs is low.