Synthesis and characterization of an MRI Gd‐based probe designed to target the translocator protein

DPA‐713 is the lead compound of a recently reported pyrazolo[1,5‐a]pyrimidineacetamide series, targeting the translocator protein (TSPO 18 kDa), and as such, this structure, as well as closely related derivatives, have been already successfully used as positron emission tomography radioligands. On the basis of the pharmacological core of this ligands series, a new magnetic resonance imaging probe, coded DPA‐C6‐(Gd)DOTAMA was designed and successfully synthesized in six steps and 13% overall yield from DPA‐713. The Gd‐DOTA monoamide cage (DOTA = 1,4,7,10‐tetraazacyclododecane‐1,4,7,10‐tetraacetic acid) represents the magnetic resonance imaging reporter, which is spaced from the phenylpyrazolo[1,5‐a]pyrimidineacetamide moiety (DPA‐713 motif) by a six carbon‐atom chain. DPA‐C6‐(Gd)DOTAMA relaxometric characterization showed the typical behavior of a small‐sized molecule (relaxivity value: 6.02 mM−1 s−1 at 20 MHz). The good hydrophilicity of the metal chelate makes DPA‐C6‐(Gd)DOTAMA soluble in water, affecting thus its biodistribution with respect to the parent lipophilic DPA‐713 molecule. For this reason, it was deemed of interest to load the probe to a large carrier in order to increase its residence lifetime in blood. Whereas DPA‐C6‐(Gd)DOTAMA binds to serum albumin with a low affinity constant, it can be entrapped into liposomes (both in the membrane and in the inner aqueous cavity). The stability of the supramolecular adduct formed by the Gd‐complex and liposomes was assessed by a competition test with albumin. Copyright © 2013 John Wiley & Sons, Ltd.


Introduction
The translocator protein (TSPO-18 kDa), formerly known as peripheral benzodiazepine receptor, [1] is a five-transmembrane protein [2] localized primarily on the outer mitochondrial membrane. [1,3] TSPO is expressed in many organs, but it is expressed at the highest level in steroidogenic cells, where TSPO is involved in the translocation of cholesterol from the outer to the inner mitochondrial membranes. TSPO is also located in the central nervous system where it is usually expressed in microglia [4,5] and in some neuronal cell types. [6][7][8][9] Following neuronal injury, TSPO expression is dramatically increased, in particular, in the case of Alzheimer disease and multiple sclerosis. [10,11] Microglia are extremely sensitive toward alterations due to neuronal injury. Until now, positron emission tomography has been the most used medical imaging technique in the study of how microglia are involved in degenerative diseases. Different TSPO targeting molecules have been synthesized over the past two decades. They belong to different chemical classes, [12][13][14] but the most promising one is the 2-phenylpyrazolo[1,5-a]pyrimidineacetamide series [15,16] , which includes, among others, the two following radiotracers: [ 11 C]DPA-713 [17][18][19] and [ 18 F]DPA-714. [20][21][22][23][24][25]27,28 Magnetic resonance imaging (MRI) is a non-invasive technique that allows morphology, physiology and metabolism visualization in vivo. MRI images deal with water 1 H proton spatial localization, whose intensity is sufficiently high, thanks to the high concentration of water in living organisms. The main advantage of this technique is its superb spatial resolution (in ordinary image acquisition, a resolution of 100 mm is attained), but its major drawback is represented by the low sensitivity of its probes. [28] A novel MRI Gd-based probe has therefore been designed, combining the pharmacological properties of the DPA-713 core with a gadolinium-dedicated (DOTA) chelator. The conjugation site was selected on the basis of previous SAR studies having shown that chemical flexibility was permitted at the para position of the phenyl ring. [29] A spacer of six carbon atoms was moreover introduced to minimize the steric hindrance of the DOTA cage at the interaction site of the molecule with TSPO. As the development of an MRI targeting agent requires systems that remain in circulation for a long time, the interaction of DPA-C 6 -(Gd) DOTAMA with human serum albumin (HSA) and its incorporation into liposomes have been explored. Moreover, the latter systems can act as carriers for a great number of imaging reporter units thus offering a route to overcome the sensitivity issue associated to MR applications.

Results and Discussion
DPA-C 6 -(Gd)DOTAMA (1) was synthesized in six steps from DPA-713 and obtained in 13% overall and non-optimized yield (Scheme 1). Briefly, DPA-713, as starting material, was synthesized according to literature procedures. [17] O-demethylation of DPA-713 was carried out with a 1 M solution of BBr 3 in dichloromethane at low temperature (from À60 to À20 C) affording the corresponding phenol 2 (DPA-OH) in high yield (95%). [20] The aminohexyl spacer was introduced into the structure by coupling the DPA-skeleton 2 with 6-(Boc-amino)hexan-1-ol activated as a sulfonate (synthesized as described in the experimental part) in DMSO at 50 C in the presence of NaOH as a base, to afford product 3 in moderate yield (63%). The Boc protecting group was removed using trifluoroacetic acid in dichloromethane, giving the free amine (4) in 80% yield. Coupling of 4 with DOTA(tBu) 3 [30] was carried out in the presence of TBTU and DIPEA in DMF yielding the DOTAMA derivative 5 in 80% yield. Tert-Butyl groups were then removed using trifluoroacetic acid to afford 6 (DPA-C 6 -DOTAMA) in 33% yield after chromatography purification. The ligand complexation was carried out by stoichiometric addition of GdCl 3 maintaining the pH at 6.5 by addition of a 0.1 N aqueous NaOH solution.
Relaxometric characterization of DPA-C 6 -(Gd)DOTAMA (1) The relaxivity of 1 (i.e. the change of the water proton relaxation rate in the presence of the paramagnetic complex at 1 mM concentration) measured at 20 MHz and 20 C was 6.02 mM À1 s À1 . This value was slightly higher than that reported in literature [31] for the parent Gd-DOTAMA complex (r 1p = 4.7 mM À1 s À1 ). The nuclear magnetic resonance dispersion (NMRD) profile was also acquired ( Figure 1) and found to show the typical behavior of small gadolinium complexes.
The NMRD data were fitted to the values calculated on the basis of the established paramagnetic relaxation theory, [32] and the obtained values are reported in Table 1 together with the corresponding ones of the parent Gd-DOTAMA-C 6 -OH. Going from Gd-DOTAMA-C 6 -OH to 1, the increased molecular weight is responsible for the markedly longer molecular re-orientational time t r , which in turn is the main determinant of the increase of the observed relaxivity.
Interaction of DPA-C 6 -(Gd)DOTAMA (1) with HSA Human serum albumin is the most abundant protein in blood and often acts as carrier for metabolites, nutrients and drugs. The lipophilicity of the DPA-713 moiety anticipates the possible binding of 1 to HSA. Binding parameters (the affinity constant K A and the relaxivity of supramolecular adduct r 1p bound ) were determined using the proton relaxation enhancement (PRE) method. [33] The experimental procedure consists of carrying out a titration in which a fixed quantity of Gd-complex (0.19 mM) is titrated with increasing concentrations of the macromolecular host ( Figure 2). The observed relaxation rates increase according to K A and r 1 bound values. In this case, a K A of 1100 M À1 and an r 1 bound of 20 mM À1 s À1 were determined.
The obtained K A value demonstrates that there is a relatively weak interaction between 1 and albumin. The calculated r 1 bound is markedly higher than the r 1p of the free DPA-C 6 -(Gd)DOTAMA (1), in accordance with the formation of the adduct with albumin that is characterized by a markedly slower tumbling rate.
The NMRD profile of a solution containing DPA-C 6 -(Gd) DOTAMA (1) and HSA in ca. 1:1 ratio (0.19 mM of 1 and 0.2 mM of HSA) was acquired (data not shown). On the basis of the previously mentioned reported results, compound 1 is only in part bound to HSA. From the knowledge of the affinity constant (K A ) and the millimolar relaxivity of the bound fraction (r 1 bound ), it   Table 1. Best fitting parameters for from the analysis of the nuclear magnetic resonance dispersion profile of 1 has been possible to quantify as 68% the percentage of compound 1 bound to HSA under the applied 1:1 ratio. From these data, the NMRD profile for the fraction of DPA-C 6 -(Gd)DOTAMA bound to HSA has been calculated ( Figure 1), and the high field relaxivity values fitted according to the theory of paramagnetic relaxation enhancement. The values obtained for the relevant parameters are reported in Table 2.
On going from the free complex (1) to the adduct with HSA, a high relaxivity peak centered at ca. 40 MHz is observed for the latter species indicative of a relaxivity domain determined by the molecular re-orientational time (t r ). In fact, from the NMRD analysis, t r is markedly increased if compared with free DPA-C 6 -(Gd)DOTAMA, as expected. The relaxivity obtained at 20 MHz for the macromolecular adduct with HSA is 17.5 mM À1 s À1 , i.e. almost three times larger than the value obtained for free DPA-C 6 -(Gd)DOTAMA.

Entrapment of DPA-C 6 -(Gd)DOTAMA (1) in liposomes
Entrapment into liposomes of DPA-C 6 -(Gd)DOTAMA (1) has been taken into account to increase the circulation lifetime and to pursue a slow release of the compound of interest. Two types of liposome have been prepared, namely LP1 and LP2. Both liposomes were prepared by means of the thin film hydration method. [34] The latter procedure consists of two key-steps, namely a first step that leads to the lipidic film that is hydratated (second step) with a saline buffer solution (HEPES/NaCl mole ratio 1/6.75). Vesicles were formulated by using 72% (for LP1) or 87% (for LP2) of POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine), 10% of cholesterol and 3% of DSPE-PEG-methoxy-2000 (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy (polyethyleneglycol)-2000] (ammonium salt)). For LP1, 15% of 1 was mixed with the lipidic components to yield the lipidic film. In LP2, 1 was dissolved in the hydration solution (20 mM). By the latter procedure, DPA-C 6 -(Gd)DOTAMA is encapsulated in the aqueous cavity of the liposome. The mean hydrodynamic diameters of the liposomes were determined by dynamic light scattering. They were found to be very similar thus showing that the methodology of preparation is not as relevant as the formulation in determining the final size of the vesicles ( Table 3).
The relaxivity values, measured at 25 C at 20 MHz, were found to be 9.69 and 7.85 mM À1 s À1 for LP1 and LP2, respectively. These values are higher than the value obtained for free 1, as consequence of the interaction with the liposome membrane. Overall, the observed relaxation enhancement is lower than the one usually observed for amphiphilic Gd complexes interacting with a liposomial membrane. For both types of liposomes, one may think that an equilibrium is occurring in the inner cavity for which DPA-C 6 -(Gd)DOTAMA distributes between the inner membrane layer and the aqueous solution. Thus, the NMRD profiles of both liposomes were analyzed taking into account either the contribution due to 1 'free' (in the inner cavity) and the one due to 1 incorporated in the membrane. Equation (1) has been used: where X bound and X free are the molar fraction of 1 that is bound or not bound to the liposome's membrane, r 1p free is the relaxivity value of 1 (6.02 mM À1 s À1 ) and r 1p bound is the relaxivity value of the adduct DPA-C 6 -(Gd)DOTAMA-liposome. We fixed the latter value to 16 mM À1 s À1 , that is the relaxivity value for analogous Gd-DOTAMA system incorporated in liposome's membrane previously investigated in our laboratory. [35] 1/T 1 NMRD profiles for both liposome suspensions were acquired in the high field region, namely from 20 to 70 MHz, as it is known that for paramagnetic macromolecular systems, this is the field strength where the acquired T 1 data are less prone to yield erroneous results in the fitting procedure. On the basis of X free and X bound values (obtained from K A ) the NMRD profiles of the r 1p bound Figure 2. Determination of binding association constants K A between DPA-C 6 -(Gd)DOTAMA (1) and human serum albumin at 25 C and pH = 7.4.  values for the two types of liposomes have been extracted ( Figure 3). The calculated NMRD profiles of the bound fraction were then fitted as earlier recalled, [32] and the best fitting values for the relevant parameters are reported in Table 4.
The 'humps' at high magnetic field observed for both liposomes are an indication that part of the Gd complex is interacting with the liposomial membrane. As discussed earlier, the detection of a relaxivity enhancement at 30-40 MHz is an indication of the occurrence of an increased t r with respect to the value found for the free complex. Also, the exchange lifetime of the coordinated water molecule (t m ) appears slightly increased upon entrapment of the Gd complex in the liposome. This behavior may be accounted for in terms of a slight distortion in the co-ordination cage geometry upon the binding of the lipophilic moiety to the liposome's membrane.
The relaxivity of DPA-C 6 -(Gd)DOTAMA (1) loaded liposomes was then measured in the presence of increasing amounts of albumin (HSA) to evaluate whether some amount of 1 could transfer from the liposome to HSA. The Gd-loaded liposomes were challenged with increasing concentrations of albumin (from 0.1 to 0.8 mM), but the relaxation rate (R 1oss ) values remained unchanged. Although the relaxivity of 1 bound to HSA is similar to that of 1 incorporated in the liposome's membrane, one would expect that the release of 1 from liposomes to HSA causes an overall shift of concentration of 1 either in the aqueous cavity or in the phospholipide bilayer. Such a shift would be related to the amount of added HSA, and it would result in a change of the observed water proton relaxation rate. The observed constancy of R 1 clearly indicates that the Gd-complex is not leaving the liposomes in the presence of HSA in the suspending medium. This result is encouraging in the light of future in vivo experiments as DPA-C 6 -(Gd)DOTAMA loaded liposomes appear to be a stable carrier for improving the blood lifetime of this imaging agent.

Conclusion
The 2-phenylpyrazolo[1,5-a]pyrimidineacetamide DPA-713 was successfully derivatized and coupled to the DOTA cage. The resulting ligand was complexed with gadolinium. The relaxivity of the obtained Gd-complex was slightly higher (6.02 mM À1 s À1 at 20 MHz) than the previously reported ones for analogous Gd-DOTAMA complexes, which is in accordance with its increased molecular weight. The NMRD profile is consistent with the one expected for a small, fast tumbling gadolinium complex. Interaction with HSA was assessed and a K A of 1100 M -1 was calculated. Incorporation into liposomes displayed a relaxivity 'hump' in the NMRD profile, as a demonstration of the formation of a supramolecular adduct. Interestingly, it has been found that the amphiphilic DPA-C 6 -(Gd)DOTAMA complex can be either incorporated in the liposomial membrane and in the inner aqueous cavity. Overall, the availability of liposomes loaded with this complex is promising for future in vivo studies with increased half-life in blood stream and encourages pursuing the task of the slow release of MRI targeting agents.