Pharmacokinetics, biodistribution and metabolism of squalenoyl adenosine nanoparticles in mice using dual radio-labeling and radio-HPLC analysis.

Adenosine is a pleiotropic endogenous nucleoside with potential neuroprotective pharmacological activity. However, clinical use of adenosine is hampered by its extremely fast metabolization. To overcome this limitation, we recently developed a new squalenoyl nanomedicine of adenosine [Squalenoyl-Adenosine (SQAd)] by covalent linkage of this nucleoside to the squalene, a natural lipid. The resulting nanoassemblies (NAs) displayed a dramatic pharmacological activity both in cerebral ischemia and spinal cord injury pre-clinical models. The aim of the present study was to investigate the plasma profile and tissue distribution of SQAd NAs using both Squalenoyl-[(3)H]-Adenosine NAs and [(14)C]-Squalenoyl-Adenosine NAs as respective tracers of adenosine and squalene moieties of the SQAd bioconjugate. This study was completed by radio-HPLC analysis allowing to determine the metabolization profile of SQAd. We report here that SQAd NAs allowed a sustained circulation of adenosine under its prodrug form (SQAd) for at least 1h after intravenous administration, when free adenosine was metabolized within seconds after injection. Moreover, the squalenoylation of adenosine and its formulation as NAs also significantly modified biodistribution, as SQAd NAs were mainly captured by the liver and spleen, allowing a significant release of adenosine in the liver parenchyma. Altogether, these results suggest that SQAd NAs provided a reservoir of adenosine into the bloodstream which may explain the previously observed neuroprotective efficacy of SQAd NAs against cerebral ischemia and spinal cord injury.

Introduction

Adenosine is an ubiquitous endogenous nucleoside able to exert a broad spectrum of physiological and pathophysiological functions, thanks to its four receptors (A1, A2A, A2B, A3) widely expressed throughout the body [1]. Targeting adenosine receptors (ARs) showed promises for numerous therapeutic applications including cardiovascular, inflammatory and neurodegenerative diseases. Adenosine itself is rapidly metabolized by adenosine kinase and, to a lesser extent, by adenosine deaminase which explains that medicinal chemistry research focused in the last decades on the development of selective agonists and antagonists of the ARs [2]. However, clinical applications of such adenosine receptors ligands are lacking, due to the advent of side effects and/or the lack of real pharmacological efficiency in clinical trials [1]. For central nervous system (CNS) diseases, the high degree of physiopathological complexity accounts for making success rates in drug development below average [3,4]. The use of multitargeted pleiotropic molecules, such as adenosine, able to modulate several pathways in all the cells of the neurovascular unit, has been suggested as a rational strategy for the treatment of neurological injuries [5,6]. In particular, the four adenosine receptors have been shown to be expressed at significant levels in neurons and glial cells (astrocytes, microglia and oligodendrocytes), as well as at the peripheral level on the cerebral endothelial cells [7]. However, the extremely short plasma half-life of adenosine [8] has hampered its use as a CNS treatment so far.

Nanomedicines have been reported to promote efficacy of CNS drugs, thanks to prolonged blood circulation and specific molecular targeting capabilities, allowing a better neurovascular access [9-11]. Indeed, the pharmacological activity of a CNS drug is directly dependent on its exposure to the Blood–Brain Barrier (BBB), which itself is a function of the compound absorption, distribution, metabolism and excretion [12]. The incorporation of nucleotides, nucleosides and other pharmacologically-active adenosine and purine receptor agonists into nanocarriers has already been described [13-15], but none ever addresses the delivery of adenosine itself. We recently discovered that the conjugation of adenosine to squalene (a natural lipid precursor of the cholesterol biosynthesis) resulted in an amphiphilic prodrug [Squalenoyl-Adenosine (SQAd)] (Fig. 1) able to self-assemble as nanoparticles (NAs) [16], displaying dramatic pharmacological efficacy in pre-clinical models of both cerebral ischemia in mice and spinal cord injury in rats [17]. It is believed that the mechanism behind the observed neuroprotective efficacy of SQAd NAs could result from altered pharmacokinetic and biodistribution profiles of adenosine.

Fig. 1

Chemical structures of adenosine (A), squalenoyl-[3H]-adenosine (✽ site of 3H-radiolabeled proton) (B) and [14C]-squalenoyl-adenosine (✽ site of 14C-radiolabeled carbon) (C).

Thus, in the present study we examined the plasma and tissue profiles of both Squalenoyl-[3H]-Adenosine NAs (tritium labeling on the adenosine moiety of the prodrug, Fig. 1B) and [14C]-Squalenoyl-Adenosine NAs ([14C] labeling on the squalene moiety of the prodrug, Fig. 1C), with comparison to the distribution profile of [3H]-Adenosine injected as a free drug. By using both liquid scintillation counting and radio-HPLC as analytical methods, we were able to assess the fate of the SQAd prodrug. Monitoring both drug (i.e. adenosine) and carrier (i.e. squalene) biodistributions, thanks to dual-labeling strategy combined with the use of analytical methods such as high performance liquid chromatography (HPLC), represents a unique source of information on the nanoformulation's integrity and stability, which may provide a better understanding of the therapeutic efficacy [18-20].

Materials & methods

Materials

Adenosine was purchased from Calbiochem (Merck Millipore). Adenosine 5′-diphosphate (ADP), adenosine 5′-monophosphate (AMP), adenosine 3′,5′-cyclic monophosphate (AMPc), inosine, Erythro-9(-2-hydroxy-3-nonyl)adenine (EHNA), dipyridamol and dextrose were purchased from Sigma-Aldrich (France). FAAH enzymes were from Cayman Chemical. Ethanol was from Carlo-Erba (France). [2,8-3H]-Adenosine and [2,8-3H]-Adenosine 5′-triphosphate tetraammonium salt were obtained from Moravek Biochemicals (Brea, CA). Fluorescent probes CholEsteryl BODIPY® FL C12 and CholEsteryl BODIPY® 542/563 C11 were obtained from Life Technologies. Ultra-pure water was prepared using a MilliQ system (Millipore Corporation, Billerica, MA). Soluene, Ultima Gold and Hionic-Fluor were purchased from Perkin Elmer (France). Reagents and HPLC grade solvents as methanol, ethyl acetate, acetonitrile, and ammonium hydroxyde (NH4OH) were provided by Carlo Erba (France). Potassium hydrogen phosphate (KH2PO4) and tetrabutylammonium chloride (TBAC) were from Fluka, ammonium acetate (CH3COONH4) from VWR and formic acid from Sigma-Aldrich (France). Liquid scintillator cocktails were provided from Perkin Elmer (France). Cartridges for SPE were from Phenomenex (France), and 0.45 μm Millex filters were from Millipore (France).

Chemical synthesis of SQAd, SQ-[3H]-Ad and [14C]-SQAd

Squalenoyl adenosine (SQAd) and radiolabeled SQAd (SQ-[3H]-Ad and [14C]-SQAd) were synthesized as previously reported [17]. Briefly, for radioactive synthesis, SQ-[3H]-Ad was obtained by H/T exchange with HTO. A suspension containing HTO in dioxane-d8 was introduced to an NMR tube containing dry SQAd and heated at 105 °C in the dark for 16 h. The slightly yellow solution obtained was diluted with MeOH and the solvent was evaporated. The crude product was purified by preparative HPLC to give 12 mCi of pure SQ-[3H]-Ad with a specific activity of 1.95 Ci/mmol. The product has been conserved as a solution of 1,4-dioxane (C = 1 mCi/mL) at − 20 °C. For [14C]-SQAd, [14C]-squalenylacetic acid was synthetized. After synthesis of the bromoprecursor derived from squalenylacetic acid, a mixture of the bromoprecursor and K[14C]N was heated at 80 °C for 3 h in anhydrous DMSO. The reaction mixture was then diluted with water and the aqueous layer was extracted with diethyl ether. The residue was purified by flash chromatography on silica gel to give the pure cyano-[14C] derivative. [14C]-squalenylacetic acid was obtained from the cyano-[14C] derivative by heating at 80 °C for 5 h and purification by flash chromatography on silica gel. The synthesis of [14C]-SQAd was then performed according to the experimental procedure described before for cold SQAd [17] to provide pure [14C]-SQAd (10.4 mCi, 69% radioactive yield, 0.056 Ci/mmol).

Preparation and characterization of [3H] or [14C] radiolabeled Squalenoyl-Adenosine Nanoassemblies (SQAd NAs)

Radiolabeled SQAd NAs were prepared by the nanoprecipitation technique, as previously described [17]. Briefly, SQAd and SQ-[3H]-Ad or [14C]-SQAd were dissolved together in ethanol and added dropwise under magnetic stirring to a 5% (w/v) aqueous dextrose solution. The ethanol was then completely evaporated using a Rotavapor® to obtain an aqueous suspension of pure NAs (2 mg/mL). The mean particle size and surface charge of the nanoassemblies were determined at 25 °C using a Malvern Zetasizer Nano ZS 6.12 (Beckman Coulter, Inc., Fullerton, CA). Particle diameters and zeta potential values of the NAs used in this study are presented in Supp. Table 1.

In vitro colloidal stability studies

Colloidal stability of SQAd NAs was investigated in vitro in mouse plasma at 37 °C using dynamic light scattering (DLS) and Förster Resonance Energy Transfer (FRET). For the DLS study, 200 μL of SQAd NAs was added to 1 mL of freshly isolated plasma (final concentration in plasma: 0.4 mg/mL) and incubated at 37 °C to mimic in vivo conditions. NA size (hydrodynamic diameter) was immediately measured by DLS as previously described and compared with the same SQAd NAs diluted in distilled water. NA size was then measured again after 1 h, 2 h, 4 h, 6 h, 9 h and 24 h of incubation in plasma. For the FRET study, FRET SQAd NAs were prepared and characterized as previously described [21]. Briefly, two fluorophores (a green cholesteryl ester BODIPY-FL probe and a red cholesteryl ester BODIPY-542/563 probe) were co-nanoprecipitated with SQAd. When the resulting NAs were excited at 490 nm (donor [i.e. BODIPY-FL] excitation wavelength), an emission signal at 574 nm (acceptor [i.e. BODIPY-542/563] emission wavelength) was measured, highlighting an energy transfer from the donor to the acceptor. It was shown that such energy transfer could occur only if the NAs remained intact; hence an acceptor emission signal under donor excitation could be considered as a validated marker of NA colloidal integrity [21]. Practically, 200 μL of FRET SQAd NAs was added to 1 mL of freshly isolated plasma (final concentration in plasma: 0.4 mg/mL) and incubated at 37 °C. Acceptor emission spectrum (500–600 nm) under donor excitation (490 nm) was immediately measured using a spectrofluorimeter and compared with the same FRET SQAd NAs diluted in distilled water. The acceptor emission spectrum of the NA suspension was then measured after 1 h, 2 h, 4 h, 6 h, 9 h and 24 h of incubation in plasma.

In vitro chemical stability studies

Chemical stability of SQAd was investigated in vitro, either in mouse plasma at 37 °C, or in the presence of FAAH enzymes using high performance liquid chromatography (HPLC). To study the chemical stability of SQAd in plasma, 1 mL of freshly isolated plasma was spiked with 10 μL of SQAd NAs (final concentration in plasma: 25 μg/mL of SQAd) and incubated at 37 °C before aliquots (100 μL) were removed after 30 min, 1 h, 2 h, 4 h, 6 h and 24 h. Supernatants were collected and evaporated to dryness at 40 °C under nitrogen flow, and dried samples were stored at − 20 °C until further analysis. Immediately before analysis, samples were thawed, reconstituted in 100 μL MeOH and vortexed for 1 min. SQAd quantification was performed using a reversed-phase HPLC system equipped with a Halo® C18 column (4.6 × 150 mm, 5 μm, Interchim), a 1525 Binary LC Pump (Waters), a 2707 Auto-sampler (Waters) and a 2998 PDA detector (Waters). Samples were eluted with 100% MeOH, at a flow rate of 0.8 mL/min. Temperature was set to 30 °C and UV detection occurred at 260 nm.

In the case of the FAAH enzymes, SQAd (final concentration: 135 μg/mL) was incubated with FAAH enzymes (final concentration: 9 U/mL) in a 20 mM HEPES buffer (pH 7.4) at 37 °C, before aliquots (200 μL) were removed after 5 min, 15 min, 30 min, 1 h, 2 h, 6 h and 24 h. 20 μL of theophylline (intern standard, final concentration: 5 μg/mL) and 500 μL of MeOH were added to the aliquots before vortexing during 30 s and centrifugation (10,000 g, 10 min). MeOH addition, vortex and centrifugation were repeated two times. Supernatants were collected and evaporated to dryness at 40 °C under nitrogen flow, and dried samples were stored at − 20 °C until further analysis. Immediately before analysis, samples were thawed, reconstituted in 200 μL of KH2PO4 0.01 M pH = 4/methanol (80%/20%, v/v) and vortexed for 1 min. Adenosine quantification was performed with mobile phase KH2PO4 0.01 M pH = 4/methanol (80%/20%, v/v) from t = 0 to t = 7 min, then followed by a 5 min linear gradient to reach 97.5%/2.5% (v/v). This composition was held from t = 12 min to t = 25 min and finally followed by a 5 min linear gradient to reach 80%/20% (v/v) from t = 30 min to t = 40 min.

In vitro blood cell interaction studies

In vitro blood cell interaction studies were conducted using whole blood of untreated mice collected by cardiac puncture. All Eppendorf collection tubes were washed with concentrated sodium citrate before the addition of blood to avoid coagulation. Blood samples (700 μL) were incubated at 37 °C for 15 min to mimic in vivo temperature conditions. 150 μL SQ-[3H]-Ad NAs were then added (final concentration in plasma: 0.35 mg/mL, 1.5 μCi/mL) and incubated at 37 °C for 1 min, 5 min, 15 min, 30 min and 1 h. At the end of the incubation period, plasma was separated by centrifugation (3000 g, 10 min) and [3H] radioactivity was measured using a β-scintillation counter (Beckman Coulter LS6500). The fraction of SQAd in plasma (Fp) was calculated according to Eq. (1) [22]. The estimated hematocrit value (Hc) was 42%. Cp and Cb refer to concentrations in plasma and whole blood, Cb being equal to the initial concentration C0 as the adsorption of SQAd on Eppendorf walls was negligible. The experiment was performed in triplicate using three different NA preparations and results are presented as mean ± SEM.

Animal experiments

The animal experiments were carried out according to the principles of laboratory animal care and European legislation (recommendation 2007/526/EC) after the ethics protocols were institutionally approved. Male SWISS NIH mice purchased from HARLAN Laboratories (5–6 weeks old) weighing about 25–30 g were used for the in vivo studies. Mice were provided with standard mouse food and water ad libitum and maintained under conventional housing conditions in a temperature-controlled room with 12-hour dark–light cycle. For pharmacokinetics and biodistribution studies, mice were divided into three main groups for treatment with free [3H]-Adenosine or with SQ-[3H]-Ad or [14C]-SQAd NAs, with subgroups of six mice for each time points. SQ-[3H]-Ad NAs, [14C]-SQAd NAs and free [3H]-Adenosine were injected at an equivalent dose of 5.55 mg of adenosine/kg, corresponding to a dose of 15 mg/kg in SQAd. Adenosine and SQAd NAs were prepared in 5% dextrose. All the compounds were administered by intravenous (IV) injection through the tail vein in a volume of 200 μL.

Pharmacokinetic study

1 min, 5 min, 15 min, 30 min, 45 min, 60 min, 2 h, 4 h, 8 h, 24 h, 48 h and 5 days after intravenous (IV) injection, following anesthesia with a lethal dose of pentobarbital injected intraperitoneally, blood samples were collected by cardiac puncture using concentrated sodium citrate as anticoagulant. 20 μL of EHNA (1 mg/mL in 0.9% NaCl) and 20 μL of dipyridamol (1 mg/mL in 0.9% NaCl) were added to the samples (700–900 μL of blood) to inhibit adenosine deaminase and adenosine transporter of the erythrocytes respectively. Without delay plasma was separated by centrifugation (3000 g, 10 min) before immediate freezing in liquid nitrogen. The samples were kept at − 80 °C until further analysis. Measurements of sample total radioactivity using liquid scintillation were performed on plasma collected from animals 5 min, 15 min, 1 h, 2 h, 4 h, 8 h, 24 h, 48 h and 5 days following IV administration of SQ-[3H]-Ad NAs (42.4 μCi/kg), [14C]-SQAd NAs (9.2 μCi/kg) or free [3H]-Adenosine (60 μCi/kg). After defrosting of plasma samples, 10 mL of Ultima Gold scintillation liquid was added to each sample, vigorously vortexed for 30 s and [14C] or [3H] radioactivities were counted after 1 h using a β-scintillation counter (Beckman Coulter LS6500). The metabolic profiles of SQ-[3H]-Ad and [3H]-Adenosine were performed on plasma collected from animals 1 min, 5 min, 15 min, 30 min, 60 min and 24 h following IV administration of SQ-[3H]-Ad NAs (600 μCi/kg) or free [3H]-Adenosine (600 μCi/kg), by high-performance liquid chromatography coupled to radioactivity and UV detections (radio-HPLC). The radio-HPLC method was preliminary developed using tritiated radioactive compounds (SQAd, Adenosine and ATP) and cold compound (inosine) as references in order to separate the postulated metabolites of SQAd and adenosine. After defrosting of plasma samples, 300 μL of methanol was added to 300 μL of isolated plasma, vortexed for 15 s and centrifuged (12,000 rpm, 10 min). The supernatant (supernatant 1) was collected and 100 μL of methanol was added to re-suspend the pellet during 15 s vortexing and centrifugation (12,000 rpm, 10 min). Once again, the supernatant (supernatant 2) was collected and pooled with supernatant 1. The resulting solution (supernatants 1 and 2) was then filtered on 0.45 μm (Millex filters, 13 mm). Recovery was checked by measuring radioactivity before and after extraction for each plasma samples. 50 μL was then injected (Perkin Elmer series 200) onto the column for radio-HPLC analysis. The analysis was performed using a reversed-phase HPLC system equipped with a C18 Nucleosil 100-5 of 150 mm × 4.6 mm column (Macherey-Nagel) and a C18 pre-column (Phenomenex), with a flow rate of 1 mL/min (Perkin Elmer series 200). Detection was performed in radioactive channel calibrated for tritium (Flo-one A515, Packard) at a flow rate of 2 mL/min of Ultima-Flo AP scintillator cocktail and with UV detection at 260 nm (Agilent 1100 series). HPLC solvents consisted of a 10 mM KH2PO4, pH 4 (solvent A) and methanol (solvent B). The applied gradient was 0 to 3 min: 95% A; 3 to 11 min: from 95 to 80% A; 11 to 14 min: 80 to 0% A; 14 to 26 min 100% B; 26 to 27 min: 100 to 5% B; and 27 to 32 min: 95% A. In those conditions, the retention times were 2.1, 7.2, 13 and 22.4 min for ATP, inosine, adenosine and SQAd respectively. Representative chromatograms of reference compounds in plasma are presented in Supp. Fig. 1. Results of the pharmacokinetic study were expressed as percentage of injected dose per milliliter of plasma. Radio-HPLC analysis results are presented as representative radio-chromatograms. Each radioactive peak was quantified as cpm peak area and calculated as percent of the sum of all quantified peaks. The limit of quantification was set to 600 cpm which represented a statistical error of 10%. Values between 150 and 600 cpm were nevertheless extrapolated and provided as informative values.

Organ distribution studies

5 min, 15 min, 1 h, 2 h, 4 h, 8 h, 24 h, 48 h and 5 days after IV injection, following anesthesia with a lethal dose of pentobarbital injected intraperitoneally, the organs including the liver, spleen, kidneys, heart, lungs and brain were dissected, rinsed with 0.9% NaCl and blotted using tissue paper to remove adherent blood and fatty matter. Approximately 50 to 100 mg of the tissues was weighted and either placed in scintillation counting vials for [14C] and [3H] radioactivity measurements or immediately frozen in Precellys® tubes (BERTIN Technologies, Montigny-le-Bretonneux, France) using liquid nitrogen for radio-HPLC analysis. In the case of radio-HPLC analysis, 100 μL of EHNA (1 mg/mL in 0.9% NaCl) was added to inhibit adenosine deaminase before freezing, and the samples were kept at − 80 °C until analysis. Measurements of total radioactivity using liquid scintillation were performed on organs collected from animals 5 min, 15 min, 1 h, 2 h, 4 h, 8 h, 24 h, 48 h and 5 days after IV administration of SQ-[3H]-Ad NAs (42.4 μCi/kg), [14C]-SQAd NAs (9.2 μCi/kg) or free [3H]-Adenosine (60 μCi/kg). For total radioactivity measurement, 1 mL of Solvable® was added into the vials containing the samples and maintained overnight at 55 °C to allow complete dissolution. After cooling, 2 × 100 μL of 30% H2O2 was added and further incubated at 55 °C for 30 min. The samples were then allowed to cool down before 10 mL of Hionic-Fluor scintillation liquid was added to the vials, vigorously vortexed for 30 s, and set aside for 1 h. The radioactivity of the samples was determined using a β-scintillation counter (Beckman Coulter LS6500). Radio-HPLC analysis was performed on the liver, spleen and kidney collected from animals 1 h and 24 h after IV administration of SQ-[3H]-Ad NAs (600 μCi/kg) or free [3H]-Adenosine (600 μCi/kg). After defrosting, organs were crushed two times 30 s at 6500 rpm. A weighted aliquot of ca. 10 mg was removed and total radioactivity was determined. After two successive crushing steps using MeOH/KH2PO4 10 mM and MeOH, the supernatants were collected, pooled and filtered on 0.45 μm filters. Total radioactivity was determined in supernatant and individual recoveries determined for each specimen. 50 μL was injected onto the column for radio-HPLC analysis. The analysis was performed as previously described for the pharmacokinetic study. Representative chromatograms of the reference compounds in each matrix and the corresponding retention times are presented in Supp. Fig. 2 and Supp. Table 2 respectively. The biodistribution data were converted and interpreted as percentage of injected dose per gram of tissue. Radio-HPLC analysis results are presented as representative radio-chromatograms. Each radioactive peak was quantified as cpm peak area and calculated as percent of the sum of all quantified peaks since efficiency of counting was constant all along the HPLC gradient and equal to ca. 19%. The limit of quantification was set to 600 cpm which represented a statistical error of 10%. Values between 150 and 600 cpm were nevertheless extrapolated and provided as informative values.

Results and discussion

In vitro colloidal and chemical stability of SQAd NAs

The integrity of the NAs in mouse plasma was first investigated in vitro using DLS and FRET. Colloidal stability is of crucial importance as it may affect biodistribution and metabolization profiles and, in turn, therapeutic efficacy and toxicity [18]. Initially, the mean diameter of SQAd NAs in water as measured by DLS was 110 ± 3 nm (peak corresponding to 100% based on intensity-weighted average diameter). When incubated in water at 37 °C, the mean diameter of the NAs did not significantly change for at least 24 h of incubation (Fig. 2A). But when the NAs were dispersed in freshly collected mouse plasma pre-warmed at 37 °C, the mean diameter significantly increased to reach 175 ± 31 nm, likely due to the formation of a protein corona (Fig. 2A and Supp. Fig. 3). The particle diameter then decreased by 8%, 33% and 52% after 2 h, 4 h and 24 h of incubation respectively. Noteworthy, the intensity peak of the DLS measurement of the SQAd NAs in plasma after 24 h of incubation perfectly overlapped the intensity peak of the plasma without NAs, demonstrating the complete disaggregation of the NAs at this time point (Supp. Fig. 3). Using FRET SQAd NAs, we measured acceptor emission fluorescence under donor excitation used as a marker of NA structure integrity [22]. When NAs were dispersed in mouse plasma, a strong acceptor signal was measured confirming NA integrity in plasma (Fig. 2B). The acceptor signal then decreased by 10%, 17% and 62% after 2 h, 4 h and 24 h of incubation respectively. Altogether, these data suggest that the SQAd NAs were stable during the first 2 h of incubation in mouse plasma at 37 °C, before to progressively disassemble from 2 h to 24 h. This disassembly is probably due to the interaction with plasma proteins such as lipoproteins, as previously described for squalene [23,24] and other squalene-based nanomedicines [25]. The chemical stability of the SQAd prodrug was then investigated up to 24 h incubation in mouse plasma (at 37 °C) using HPLC. As shown in Fig. 2C, more than 80% of the prodrug remained intact until 2 h, and 54% after 24 h of incubation. To be noted, after 4 h of incubation in plasma, 25% of the SQAd prodrug was metabolized, which remained in the same order of magnitude than the SQAd NA diameter decrease (i.e. 33%). Hence, it suggests that the prodrug was rapidly metabolized once released from the nanostructure likely by the amidases present in plasma [26,27]. Finally, the extent of blood cell partitioning of SQAd NAs was assessed in vitro. Fig. 3D displays the corresponding plasma fractions (Fp) over incubation time. As soon as 1 min after addition to blood, a strong interaction of SQAd NAs with blood cells was observed (Fp = 35 ± 1.6%), in accordance with the ability of lipophilic drugs to partition with RBC membranes as previously demonstrated [19]. Hence, the RBC may contribute to the metabolization of the SQAd prodrug and therefore may constitute as a source of adenosine.

Fig. 2

In vitro colloidal and chemical stability of SQAd NAs in mouse plasma. SQAd NAs were dispersed in freshly isolated mouse plasma or water, and NA integrity was assessed over incubation time by measuring the NA mean diameter using DLS (A). Diluting FRET SQAd NAs in mouse plasma, the acceptor emission fluorescence under donor excitation was measured to assess colloidal status (assembled or disassembled, refer to Materials & methods section for detailed analysis of FRET signal) (B). SQAd NAs were dispersed in freshly collected mouse plasma and chemical stability over incubation time was assessed by HPLC. Results are presented as mean ± SEM from N = 3 independent samples (C). Plasma fraction of whole blood spiked with SQ-[3H]-Ad NAs in vitro was measured after different incubation times. Results are presented as mean ± SEM from N = 3 independent samples (D).

Fig. 3

Plasma concentrations of SQAd NAs. (A) [3H]- and [14C]-radioactivity plasma time profiles expressed as the percentage of the injected dose per mL of plasma after intravenous injection of SQ-[3H]-Ad NAs (15 mg/kg, 42.4 μCi/kg) or [14C]-SQAd NAs (15 mg/kg, 9.2 μCi/kg). Results are presented as mean ± SEM from N = 6 animals/time-point/group. Representative radio-chromatograms obtained after radio-HPLC analysis of plasma (B) 1 min, (C) 15 min and (D) 1 h post-injection of SQ-[3H]-Ad NAs (15 mg/kg, 600 μCi/kg).

Plasma profile of SQAd NAs

Dual radiolabeled SQAd NAs were separately intravenously administered through the tail vein of SWISS NIH male mice. The resulting plasma radioactivity concentration profiles for each radiolabel (i.e. [3H] on the adenosine moiety or [14C] on the squalene moiety) are reported in Fig. 3A and expressed as the percentage of injected dose per milliliter of plasma. Plasma profiles for [3H] and [14C] were quite divergent. [14C]-radioactivity found in plasma rapidly decreased and 1 h post-injection only 0.6 ± 0.03% of the injected dose/mL was still detected, when the amount of tritium was 2.8-fold greater. These results clearly suggested that the SQAd prodrug underwent metabolization in systemic circulation in vivo. Nanoassembly disaggregation in vivo was significantly faster than observed in vitro, likely because the body is an open system, with other rapidly occurring events such as tissue distribution. Moreover, the presence of shear stress and other mechanical constraints in the bloodstream may increase the disaggregation rate of SQAd NAs. The in vivo metabolization of the SQAd prodrug has been further investigated by radio-HPLC analysis of the plasma radioactivity 1 min (Fig. 3B), 15 min (Fig. 3C) and 1 h (Fig. 3D) post-injection of SQ-[3H]-Ad NAs. The SQAd peak intensity (retention time (RT) = 22.5 min) significantly decreased when blood was collected 1 min, 15 min and 1 h post-injection respectively, confirming a progressive metabolization of the prodrug in the circulation. To be noted that when injected as a free molecule, no adenosine (RT = 12.7 min) could be detected at those time points, which was in accordance with the extremely fast metabolization of this molecule in systemic circulation [7] by leukocytes, erythrocytes and endothelial cells [28,29]. Thus, SQAd NAs represent a unique approach to significantly improve blood retention of adenosine under the form of SQAd prodrug. Inosine (RT = 7.1 min), the first circulatory metabolite of adenosine, was not detected either, but a peak corresponding to unidentified highly polar metabolites (RT = 2.1 min) strongly increased over time while the SQAd peak decreased (Fig. 3B–D).

Organ distribution of SQAd NAs

Tissue distribution of SQ-[3H]-Ad NAs and [14C]-SQAd NAs was investigated in mice after intravenous injection of SQAd NAs and compared to the distribution of free [3H]-Adenosine. Amount of [3H]- and [14C]-radioactivities resulting from SQ-[3H]-Ad NA and [14C]-SQAd NA administrations rapidly diverged in the heart (Fig. 4A) but remained similar in the lungs up to 15 min post-administration (Fig. 4B). Except in the heart at 5 min and 15 min post-administration, radioactivity levels resulting from SQAd NA administration were lower compared with radioactivity subsequent to free [3H]-Adenosine injection. Notably, in the heart the radioactivity raising from adenosine administration was relatively constant up to 5 day post-injection (ca. 6% of injected dose/g of organ 5 day post-administration), when the one resulting from NAs decreased constantly to reach a very low level (1.5 ± 0.1% of injected dose/g of organ for [3H]-radioactivity and 0.4 ± 0.02% of injected dose/g of organ for [14C]-radioactivity), suggesting the absence of accumulation of SQAd NAs or free SQAd prodrug in this organ. These distinct profiles confirm that the squalenoylation of adenosine allows to modify the drug biodistribution pattern.

Fig. 4

Distribution of the radioactivity in heart and lungs after intravenous injection of free [3H]-Adenosine, SQ-[3H]-Ad NAs or [14C]-SQAd NAs. Radioactivity found in the heart (A) and lungs (B) was expressed as the percentage of injected dose per gram of tissue following intravenous administration of [3H]-Adenosine (5.55 mg/kg, 60 μCi/kg), SQ-[3H]-Ad NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 42.4 μCi/kg) or [14C]-SQAd NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 9.2 μCi/kg). Results are mean ± SEM from N = 6 animals/time-point/group.

Total radioactivity of [3H] and [14C] rapidly diverged in the brain following SQ-[3H]-Ad NA or [14C]-SQAd NA administration (Fig. 5). [3H]-radioactivity content progressively increased with time to reach a plateau 8 h post-administration corresponding to 1.4 ± 0.1% of injected dose/g of organ. Meanwhile, [14C]-radioactivity content stayed constant and equal to ca. 0.1% of injected dose/g of tissue up to 5 day post-injection. These data clearly demonstrated that the NAs did not translocate from the blood circulation to the brain parenchyma. The overall brain distribution of [3H]-radioactivity following SQ-[3H]-Ad NA administration was lower or similar to the one obtained following free [3H]-Adenosine administration at all time-points. Noteworthy, as previously shown [17], no intact adenosine could be detected in the brain following either free [3H]-Adenosine or SQ-[3H]-Ad NA administration, showing that the detected radioactivity corresponded only to metabolization products.

Fig. 5

Distribution of the radioactivity found in the brain after intravenous injection of free [3H]-Adenosine, SQ-[3H]-Ad NAs or [14C]-SQAd NAs. Radioactivity found in the brain was expressed as the percentage of injected dose per gram of tissue following intravenous administration of [3H]-Adenosine (5.55 mg/kg, 60 μCi/kg), SQ-[3H]-Ad NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 42.4 μCi/kg) or [14C]-SQAd NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 9.2 μCi/kg). Results are mean ± SEM from N = 6 animals/time-point/group.

The highest radioactivity content was observed in the liver and in the spleen (Fig. 6) after IV administration of SQAd NAs, while the radioactivity found in the kidneys was more important after injection of Adenosine free (Supp. Fig. 4). This biodistribution pattern was in accordance with previous studies showing that colloids are preferentially captured by the liver [30,31], due to opsonization [32]. On the other hand, [3H]-radioactivity associated with [3H]-Adenosine administration was mainly eliminated by the kidneys (Supp. Fig. 4) as commonly observed for small hydrophilic molecules. Maximum accumulation of NAs in the liver and spleen occurred at 15 min post-injection, and then constantly decreased with time. [3H] and [14C] contents were similar in the liver for the first 8 h post-injection (Fig. 6A) and in the spleen up to 24 h after administration (Fig. 6D), suggesting the uptake of intact SQAd by these tissues. These results were confirmed by radio-HPLC analysis of the same organs. Intact SQAd represented 66 ± 0.8% of the total radioactivity found in the liver at 1 h post-injection (Fig. 6B) but the prodrug was no more detectable after 24 h (Fig. 6C). In the spleen, SQAd accounted for 96 ± 1% of the total radioactivity found in this organ at 1 h post-injection (Fig. 6E) and still for 66 ± 4% after 24 h (Fig. 6F). Interestingly, adenosine released from SQAd NAs was detected in the liver 1 h (accounting for 27 ± 7% of radioactivity detected in this organ) and 24 h (accounting for 47 ± 8% of radioactivity detected in this organ) post-injection, but not in the spleen. SQAd metabolization in the liver was probably due to the activity of amidases, and in particular fatty acid amide hydrolases (FAAHs) [33-35], as the incubation of the SQAd prodrug in the presence of FAAH enzymes in vitro was shown to induce a sustained release of adenosine (Supp. Fig. 5). Moreover, FAAHs were shown to be 100-times more active in the liver compared to the spleen [36], probably explaining the absence of metabolization in this organ 24 h post-administration. Purines and more specifically adenosine have been shown to influence hepatic cell activity such as glycogenolysis, vasoconstriction or collagen expression [37]. Recent studies suggested that the liver could be involved in coordinated endocrine protective mechanisms in response to an ischemic injury in a remote organ, and in particular in the case of cerebral ischemia, via the production of proteins such as trefoil factor 3 (TFF3) [38], fetuin-A [39], or the Insulin-like Growth Factor I (IGF-I) [40]. Interestingly, the plasma level of IGF-1 has also been demonstrated to be an independent predictor of stroke outcome [41]. Since adenosine has been shown to trigger early and late pre-conditioning in liver cells [42,43], the production of adenosine in the liver by SQAd NAs may possibly be triggering the release of endocrine neuroprotective molecules in the circulation participating in the observed pharmacological activity of the NAs in experimental brain ischemia and spinal cord injury [17]. On the other hand, the observed important capture of SQAd NAs by the RES organs may have toxicological repercussions. However, as previously shown by us in a short-term (24 h post-injection) and long-term (7 days and 28 days post-injection) toxicity study, neither hepatic nor hematologic toxicity was observed as assessed from the biochemical and hematological parameters, as well as immunohistological observations of the liver and the spleen [17].

Fig. 6

Distribution of the radioactivity found in liver and spleen after intravenous injection of free [3H]-Adenosine, SQ-[3H]-Ad NAs or [14C]-SQAd NAs and metabolic profiles obtained by radio-HPLC. (A) Radioactivity found in the liver expressed as the percentage of injected dose per gram of tissue following intravenous administration of [3H]-Adenosine (5.55 mg/kg, 60 μCi/kg), SQ-[3H]-Ad NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 42.4 μCi/kg) or [14C]-SQAd NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 9.2 μCi/kg). Results are presented as mean ± SEM from N = 6 animals/time-point/group. Representative radio-chromatograms obtained after radio-HPLC analysis of liver (B) 1 h and (C) 24 h post-injection of SQ-[3H]-Ad NAs (15 mg/kg, 600 μCi/kg). Retention times in liver were RT(SQ-[3H]-Ad) = 22.5 min and RT([3H]-Adenosine) = 12.9 min. (D) Radioactivity found in the spleen expressed as the percentage of injected dose per gram of tissue following intravenous administration of [3H]-Adenosine (5.55 mg/kg, 60 μCi/kg), SQ-[3H]-Ad NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 42.4 μCi/kg) or [14C]-SQAd NAs (15 mg/kg, equiv. 5.55 mg/kg adenosine, 9.2 μCi/kg). Results are mean ± SEM from N = 6 animals/time-point/group. Representative radio-chromatograms obtained after radio-HPLC analysis of spleen (E) 1 h and (F) 24 h post-injection of SQ-[3H]-Ad NAs (15 mg/kg, 600 μCi/kg). Retention times in spleen were RT(SQ-[3H]-Ad) = 22.5 min and RT([3H]-Adenosine) = 13.4 min.

Conclusion

The present study represents an important contribution since it clearly demonstrates that the squalenoylation of a fragile molecule such as adenosine may dramatically modify the circulation residency, the metabolism and the biodistribution of this nucleoside. The use of dual radio-labeling and radio-HPLC as analytical methods provided important information on the NAs and prodrug integrity in vivo. Our results indicate that the greater pharmacological efficacy of SQAd NAs compared to free Adenosine observed both in a cerebral ischemia model and a spinal cord injury model may be explained by (i) a sustained circulation in the vascular compartment allowing an improved interaction of adenosine with the endothelial cells of the neurovascular unit, and (ii) a greater distribution to reticuloendothelial organs possibly improving the treatment of ischemic injuries, thanks to endocrine protective mechanisms which remain to be confirmed. It is important to note that adenosine is already marketed in the United-States as Adenocard® to treat paroxysmal supraventricular tachycardia. Moreover selective agonists/antagonists of adenosine receptors or adenosine itself have been engaged in clinical trials for numerous applications [2] as diverse as perioperative cardioprotection [44], asthma [45], arthritis [46] or diabetes [47] treatments. However, the rapidly fluctuating adenosine concentrations, the complexity and opposing effects of adenosine receptor activation and their widespread distribution leading to severe side-effects have limited clinical success [1]. In this context, the use of SQAd NAs as a sustained delivery system of adenosine may open interesting prospects.