Article pubs.acs.org/IC
Dipicolinate Complexes of Gallium(III) and Lanthanum(III) David M. Weekes, Caterina F. Ramogida, Maria de Guadalupe Jaraquemada-Peláez, Brian O. Patrick, Chirag Apte, Thomas I. Kostelnik, Jacqueline F. Cawthray, Lisa Murphy, and Chris Orvig* Medicinal Inorganic Chemistry Group, Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia V6T 1Z1, Canada S Supporting Information *
ABSTRACT: Three dipicolinic acid amine-derived compounds functionalized with a carboxylate (H3dpaa), phosphonate (H4dppa), and bisphosphonate (H7dpbpa), as well as their nonfunctionalized analogue (H2dpa), were successfully synthesized and characterized. The 1:1 lanthanum(III) complexes of H2dpa, H3dpaa, and H4dppa, the 1:2 lanthanum(III) complex of H2dpa, and the 1:1 gallium(III) complex of H3dpaa were characterized, including via X-ray crystallography for [La4(dppa)4(H2O)2] and [Ga(dpaa)(H2O)]. H2dpa, H3dpaa, and H4dppa were evaluated for their thermodynamic stability with lanthanum(III) via potentiometric and either UV−vis spectrophotometric (H3dpaa) or NMR spectrometric (H2dpa and H4dppa) titrations, which showed that the carboxylate (H3dpaa) and phosphonate (H4dppa) containing ligands enhanced the lanthanum(III) complex stability by 3−4 orders of magnitude relative to the unfunctionalized ligand (comparing log βML and pM values) at physiological pH. In addition, potentiometric titrations with H3dpaa and gallium(III) were performed, which gave significantly (8 orders of magnitude) higher thermodynamic stability constants than with lanthanum(III). This was predicted to be a consequence of better size matching between the dipicolinate cavity and gallium(III), which was also evident in the aforementioned crystal structures. Because of a potential link between lanthanum(III) and osteoporosis, the ligands were tested for their bone-directing properties via a hydroxyapatite (HAP) binding assay, which showed that either a phosphonate or bisphosphonate moiety was necessary in order to elicit a chemical binding interaction with HAP. The oral activity of the ligands and their metal complexes was also assessed by experimentally measuring log Po/w values using the shake-flask method, and these were compared to a currently prescribed osteoporosis drug (alendronate). Because of the potential therapeutic applications of the radionuclides 67/68Ga, radiolabeling studies were performed with 67Ga and H3dpaa. Quantitative radiolabeling was achieved at pH 6.5 in 10 min at room temperature with concentrations as low as 10−5 M, and human serum stability studies were undertaken.
■
La3+ is an approved orally administered treatment for hyperphosphatemia, and the drug owes its effectiveness to its very low intestinal absorbance (estimated to be less than 0.002%).12 It was in this form that an effect of the metal ion on bone histology was first noted;11 however, with such low bioavailability, vastly elevated quantities would need to be administered in order to generate an effective oral dose. Fosrenol is known to have adverse effects on the GI tract,13 so elevating the dose as a method to increase La3+ absorbed would undoubtedly be met with extremely low compliance. By modification of the chemical environment around La3+ (through rational chelator design), our previous studies have shown progress in the quest for a system that improves the oral activity of the metal ion, without impeding its natural tendency for targeting bone tissue upon absorbance.14−16 In 2007, Barta et al. investigated a series of tris-bidentate lanthanide complexes based on 3-hydroxy-4-pyrone and -pyridinone scaffolds and found that the complex tris(1,2-dimethyl-3-oxy-4-pyridinone)-
INTRODUCTION Osteoporosis, characterized by low bone mineral density, is one of the most common diseases of the skeleton, predominantly affecting the elderly, with 1 in 2 women and 1 in 4 men over the age of 50 likely to suffer an osteoporotic fracture.1 The most common medicinal agents designed to prevent bone loss are bisphosphonates (BPs), a class of orally delivered drugs that suppress the proliferation of cells responsible for bone resorption.2,3 In spite of their wide use, BPs suffer from a number of significant drawbacks, including side effects such as musculoskeletal pain and gastrointestinal (GI) upset,4 as well as growing uncertainty over their long-term efficacy.5 Consequently, there is a strong drive to develop alternative treatments. Various studies over the past 15 years have suggested that, at certain concentrations within bone, La3+ (which bears a number of chemical similarities to Ca2+ including ionic radius, preferred coordination number, and preference for hard donor ligands) can invoke a cellular response that enhances bone building, signifying a potential for La3+ to counter the effects of osteoporosis.6−11 As a carbonate salt (known as Fosrenol), © XXXX American Chemical Society
Received: September 27, 2016
A
DOI: 10.1021/acs.inorgchem.6b02357 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry lanthanum(III) [La(dpp)3; Figure 1a] greatly improved transport of the metal ion across the membrane of Caco-2
Chart 1. Structures of the Four dpa-Containing Ligands
Figure 1. Structures of the two lead compounds previously investigated to improve the oral activity of trivalent lanthanum in the development of novel treatments for osteoporosis: (a) La(dpp)3;14−16 (b) La(XT).15−17
administered compounds for delivering lanthanum to bone tissue. The functionalized ligands include either a carboxylate (6,6′-{[(carboxymethyl)azamediyl]dimethylene}dipicolinic acid, H3dpaa), a phosphonate (6,6′-{[(2-phosphonoethyl)azanediyl]dimethylene}dipicolinic acid, H4dppa), or a BP (6,6′-{[(4-hydroxy-4,4-diphosphonobutyl)azanediyl]dimethylene}dipicolinic acid, H7dpbpa) moiety. Also included in these studies is the unfunctionalized version [6,6′(azanediyldimethylene)dipicolinic acid, H2dpa], which enables a calibrated assessment of the effect of each of these functional groups. We aimed to assess the bone-targeting ability of the ligands through HAP binding studies, the contribution to La3+ complex stability from the carboxylate group in H3dpaa or the phosphonate group in H4dppa through potentiometric titration, and the expected cell permeability via lipophilicity measurements (and the effect that lanthanum chelation has on these values). The solid-state crystal structures of the 1:1 La3+ complex of H4dppa is also described. In addition, because of its potential to form a charge-neutral complex, the carboxylate-functionalized H3dpaa was investigated for its coordination properties, thermodynamic stability, and in vitro kinetic inertness with gallium(III). Complexes of H3dpaa with lutetium(III) and gadolinium(III) have been reported previously as potential MRI contrast agents,19 and our interest herein lies with gallium(III) because of the high-quality positron emission tomography (PET) imaging potential of the radionuclide 68Ga.20 The solid-state crystal structure of the 1:1 Ga3+ complex of H3dpaa was also obtained and is fully described.
cells (an in vitro model for human epithelial cells).14 In 2013, a similar result was obtained by Mawani et al. for the phosphinate-containing 1:1 ethylenediaminetetraacetic acid (EDTA)-type complex [bis[[bis(carboxymethyl)amino]methyl]phosphinate]lanthanum(III) [La(XT); Figure 1b],15,17 and in 2015, these two complexes were compared directly to probe the thermodynamic and kinetic interactions between each system and hydroxyapatite (HAP), an ex vivo model for bone mineral, as well as the influence of a chelator on La3+ in vivo biodistribution.16 Some key factors considered in the fundamental design of these systems include synthetic accessibility, bone mineral targeting, thermodynamic stability of the lanthanum complex, and lipophilicity. In this regard, there are advantages and drawbacks to both dpp and XT; in an effort to combine the favorable attributes of each, we propose a ligand scaffold based on a dipicolinic acid (dpa) binding motif (Figure 2). Dipicolinate derivatives have been examined previously for their chemistry with trivalent lanthanides, predominantly in the context of magnetic resonance imaging (MRI) contrast agents.18,19 Here, the ligand scaffold consists of two picolinic acid (pa) groups connected via methylene bridges at the 6 position of each pa to a central nitrogen atom, resulting in a metal binding cavity that is inherently five-coordinate, thereby offering greater thermodynamic stability than bidentate dpp. Furthermore, the central nitrogen atom provides a convenient handle for functionalization, which allows tuning and optimization of the ligand’s properties. Herein, we present a set of N-functionalized dpa derivatives (Chart 1), as well as their corresponding La3+ complexes, which have been synthesized, characterized, and subjected to a number of testing procedures to assess their suitability as orally
■
RESULTS AND DISCUSSION Ligand Synthesis and Characterization. The synthetic routes to all four dpa-based compounds, H2dpa, H3dpaa, H4dppa, and H7dpbpa, rely on the alkylating agent methyl 6-
Figure 2. Solid-state ORTEP structures for crystals of (a) 2[H2dpa·H2O], (b) H3dpaa·H2O, and (c) H4dppa·3H2O. Full experimental details and crystallographic data are presented in the Supporting Information (Table S1). B
DOI: 10.1021/acs.inorgchem.6b02357 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Scheme 3. Synthetic Route to H3dpaa as the HCl Salta
(bromomethyl)picolinate (3), which is generated via a wellestablished protocol from commercially available dimethylpyridine-2,6-dicarboxylate (1; Scheme 1).20,21 The alkyl-protected Scheme 1. Established Synthesis for the Alkylating Agent 320,21
Reagents and conditions: (i) 3 (2.1 equiv), K2CO3, CH3CN, 60 °C, 12h; (ii) 6 M HCl, 110 °C, 12 h.
a
ligands are then generated from 3 via a double N-alkylation of the primary amine synthon that corresponds to the desired compound, followed by deprotection (for all except H7dpbpa, in which the deprotected product is generated during alkylation) in a strong acid to obtain the final ligands as the HCl salts (Schemes 2−5).
Scheme 4. Aza-Michael-Addition-Mediated Synthetic Route to H4dppa as the HCl Salta
Scheme 2. Synthetic Route to H2dpa as the HCl Salta
Reagents and conditions: (i) excess NH4OH(aq), H2O, 0 °C to room temperature, 12 h; (ii) 3 (2.1 equiv), K2CO3, 60 °C, 12 h; (iii) 12 M HCl, 110 °C, 12 h. †Yield over two steps. a
Reagents and conditions: (i) 3 (2.1 equiv), K2CO3, CH3CN, 60 °C, 24 h; (ii) PhSH (1.1 equiv), tetrahydrofuran, 70 °C, 48 h; (iii) 6 M HCl, 110 °C, 12 h. a
unwanted byproducts invariably forms and, given the oily consistency of the alkyl phosphonates and the absence of a chromophore, are very challenging to separate. This was addressed by the direct alkylation of a crude reaction mixture from step i in Scheme 4, followed by column separation to isolate the desired dimethyl picolinate intermediate 11, which was deprotected in strong acid in a manner similar to that for the other compounds. Finally, because of challenges in obtaining the necessary alkyl-protected primary amine synthon, H7dpbpa was accessed in a single step via a pH-controlled aqueous reaction of 3 with alendronic acid (12; Scheme 5). Because of the basicity of the
For H2dpa, because the final product is the secondary amine, an additional 2-nitrobenzenesulfonamide (4)-mediated protection−deprotection step is required (Scheme 2). This chemistry was optimized by Price et al. and enables good yields and straightforward purification compared to other amine protecting groups.21 The synthesis of H3dpaa is more straightforward because it only requires N-alkylation of the commercially available glycine ethyl ester (hydrochloride salt, 7) prior to deprotection (Scheme 3), and a similar procedure has been previously reported.19 The primary amine synthon (10) required to make H4dppa is not commercially available but can be synthesized in a single step and under mild conditions via an aqueous-mediated catalyst-free aza-Michael addition of diethyl vinylphosphonate (9) with ammonium hydroxide (Scheme 4).22 One drawback of this procedure is the tendency for overreaction to occur via subsequent 1,4-additions into the vinylphosphonate Michael acceptor from the more active amines that are formed, resulting in bis- and trisphosphorylated byproducts; however, this can be minimized by the very gradual addition of 9 into a large excess of an ammonium hydroxide solution. In spite of the optimized conditions, such is the activity of the primary amine 10 relative to ammonia, a small quantity of
Scheme 5. Synthetic Route to H7dpbpa as the HCl Salta
a
Reagents and conditions: (i) 3 (3.0 equiv), 3:1 water/tetrahydrofuran, 1 M NaOH (pH 12), 3 days (followed by acid workup).
C
DOI: 10.1021/acs.inorgchem.6b02357 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 3. Solution depletion plots showing the [ligand] remaining in solution (unbound to HAP) over 48 h as % [ligand] at time = 0 (n = 3): (a) H2dpa; (b) H3dpaa; (c) H4dppa; (d) H7dpbpa. Conditions for all experiments: 37 °C, pH 7.4, I = 0.16 M (NaCl), agitation at 220 rpm.
primary amine group of 12 (pKa 12.7),23 a pH between 12 and 13 was maintained throughout the course of the reaction by the regular dropwise addition of 1 M NaOH. This has the simultaneous effect of removing the alkyl-protecting groups of the picolinate arms; however, unwanted bromide hydrolysis obviated good yields despite efforts to optimize the reaction conditions. A similar reaction procedure (with good yield) has been previously reported in the synthesis of bone-seeking chelating ligands for 99mTc, but in those cases the alkylating agent was simply 2-(chloromethyl)pyridine (no carboxylate functionality), which may explain the improved yield in that case.24 The four compounds were characterized by 1H and 13C NMR spectroscopy (as well as 31P{1H} NMR for the phosphorus-containing compounds), low- and high-resolution mass spectrometry, and elemental analysis (see the Supporting Information). Crystals suitable for solid-state X-ray analysis for 2[H2dpa·H2O], H3dpaa·H2O, and H4dppa·3H2O (confirming attainment of the desired products) were obtained as zwitterions, and the structures are presented in Figure 2 (see the Supporting Information, Table S1, for crystallographic data). Bone Targeting. With the dpa-based ligands in hand, solution depletion experiments were performed to determine the influence of the R appendage on bone mineral targeting and binding (in the absence of a chelated metal ion) using HAP as a model for the bone mineral. Ligand solutions (0.1 mM) buffered to pH 7.4 were incubated and agitated (37 °C and 220 rpm) with a suspension of synthetic HAP (in 200-fold excess), and at fixed time points over the course of 48 h, the concentration of ligand remaining in solution (not bound to HAP) was determined by ultraviolet−visible (UV−vis) spectrometry. The results (Figure 3) are plotted as % [ligand] remaining in solution relative to [ligand]t=0 versus time in hours.
Figure 3a shows the disappearance of H2dpa (R = H) from solution to an approximate minimum of [H2dpa] = 80% after 48 h, with the other 20% presumably bound to HAP. Although there is very little error associated with the data, the first few time points are erratic and incoherent from a kinetic point of view, and this, coupled with the fact that the majority of the compound remains in solution, infers that there is little to no chemical attraction between H2dpa and HAP. This is consistent with other compounds that do not contain specific bone binding groups (such as a phosphonate or BP), and the minimal disappearance from solution that is observed is likely due to physisorption in the form of weak van der Waals interactions between solvated H2dpa and suspended HAP particles. A similar argument can be made to explain the results for H3dpaa (Figure 3b), which also does not contain any functionality that specifically binds to bone mineral. As with H2dpa, a slight depletion from solution is observed to a minimum [H3dpaa] of approximately 85% after 48 h, which is again likely only due to physical adsorption onto the surface of HAP particles. The observed depletion profile changes significantly with the introduction of the phosphonate moiety in H4dppa (Figure 3c). In this case, the concentration of the ligand in solution disappears in a smooth and coherent manner to a minimum [H4dppa] in solution of around 55% after 48 h. This is indicative of a process involving chemical absorption, consistent with the ability of phosphonates to bind to HAP with moderate strength.25 With H7dpbpa (Figure 3d), the compound binds rapidly to HAP and is almost undetectable in solution (
