Bioconjugate Chem. 2006, 17, 1270−1279
1270
Synthesis and Characterization of AP5346, a Novel Polymer-Linked Diaminocyclohexyl Platinum Chemotherapeutic Agent Paul Sood,* K. Bruce Thurmond, II, Jeremy E. Jacob, Lynda K. Waller, George O. Silva, Donald R. Stewart, and David P. Nowotnik Access Pharmaceuticals, Inc., 2600 Stemmons Freeway, Suite 176, Dallas, Texas 75207. Received March 1, 2006; Revised Manuscript Received June 28, 2006
Syntheses of the novel polymer-bound platinum-based chemotherapeutic agent poly(HPMA)-GGG-AmadPtd (1R,2R)-DACH, AP5346, and its precursors are reported. The method utilized in preclinical development of AP5346 is described herein. Additionally, an improved synthesis, which has shown that ion exchange resins can be removed, significantly less platinum can be used when the reaction mixture is pH-stated, and the synthesis can be performed at a higher concentration, is reported. These combined improvements result in a more cost-effective, scaleable procedure. Various methods of analysis of the drug substance are also discussed. Specifically, 1H NMR spectroscopy is used for identity and can also distinguish small molecule impurities to below 0.1%. 195Pt NMR determines the coordination environment of the platinum and also identity and purity in relation to platinum chelation of the construct. Size exclusion chromatography is used to establish the molecular weight of AP5346 while ICP-AES determines platinum content and platinum release rates in phosphate-buffered saline. The cumulative results of this work have yielded an efficient syntheses of a polymer-based chemotherapeutic agent with subsequent detailed characterization methods.
INTRODUCTION Since the serendipitous discovery of the anticancer properties of cisplatin by Rosenberg (1), platinum-containing drugs have been extensively used in the treatment of solid tumors (2). When administered with other agents, cisplatin has exhibited remarkable activity against refractive disease states. In combination chemotherapy with paclitaxel survival rates of patients with advanced ovarian cancer are significantly improved, while coadministering the drug with vinblastine and bleomycin is extremely effective against testicular cancer (3). This effectiveness has resulted in cisplatin now being one of the three most utilized chemotherapeutics in the world. However, clinical use of cisplatin has been limited by the severe side effects which accompany its use, the most prominent being nephrotoxicity, neurotoxicity, nausea, and vomiting (4-7). Additionally, intrinsic or acquired tumor resistance is a major issue. Such limitations, coupled with a narrow therapeutic index and poor solubility, have been the driving force behind a sustained research effort into the discovery of novel platinum agents or novel formulations and delivery methods of existing platinum agents. The resulting compounds would preferably have better toxicity profiles, improved efficacies, and a broader spectrum of activity. Many new platinum molecules have undergone preclinical and clinical testing; however, only carboplatin and oxaliplatin (Figure 1) have been approved worldwide for routine use in the clinic. Additionally, nedaplatin has been approved for use against ovarian and cervical cancers in Japan (8). The cis-NH3 ligands in cisplatin are also present in both carboplatin and nedaplatin, whereas the two chlorides have been replaced by the more stable cyclobutanedicarboxylate and glycolate ligands, respectively. In oxaliplatin the cis-NH3 groups have been replaced by a diaminocyclohexyl (DACH) chelate ligand and the chlorides by an oxalato chelate. The influence of structure on the activity and toxicity of Pt anti-cancer drugs * Corresponding author. E-mail:
[email protected]. Phone: 214 905 5100. Fax: 214 905 5101.
Figure 1. Structures of cisplatin, carboplatin, oxaliplatin, and nedaplatin.
has been comprehensively described by Hambley (9). Carboplatin has been shown to less toxic than cisplatin and therefore can be administered at much higher doses (10) with the doselimiting toxicity being myelosuppression (11). Oxaliplatin is active in colorectal cancer and has supplanted carboplatin and cisplatin as the largest selling platinum therapeutic. Despite its success, the most frequently encountered dose-limiting toxicity of this agent is acute onset neurotoxicity which is observed in approximately 90% of patients, while 10-15% of patients suffer from cumulative sensory neuropathy (12). In order to improve the therapeutic index of platinum agents, Access Pharmaceuticals has used a rational drug design approach, based on the principles of polymer drug delivery, to develop a polymer-linked DACH-platinum chemotherapeutic, AP5346. The biodistribution of soluble macromolecules is governed extensively by their ability to penetrate endothelial layers. As growing tumors establish their own blood supply, they develop pathological neovasculature whose blood vessels are frequently hyperpermeable to circulating macromolecules and small particulates. In addition to this enhanced permeability, tumor tissue often has limited lymphatic and/or capillary drainage, so the macromolecules can be trapped and concentrated in tumors (extensively reviewed in ref 13). If a chemotherapeutic agent is coupled to a suitable polymer or other
10.1021/bc0600517 CCC: $33.50 © 2006 American Chemical Society Published on Web 08/15/2006
Synthesis and Characterization of AP5346
macromolecular carrier Via a biodegradable linker, then such carriers have the potential of increasing the concentration of the chemotherapeutic agent within the tumor tissue. As a result of these characteristics, concentrations of polymer-drug conjugates in tumor tissue can reach levels some 10-100 times higher than that produced by administration of the free drug (14). This effect has been termed the “enhanced permeability and retention” (EPR) effect (15). In addition, the intrinsic cytotoxicity of the native drug is often substantially diminished when the drug is chemically linked to the polymer. This also reduces the systemic toxicity of the polymer-drug conjugate relative to the native drug. The concept of linking a platinum agent to the water-soluble, biocompatible copolymer, N-(2-hydroxypropyl)methacrylamide (HPMA), was initially developed at the School of Pharmacy, University of London (16). The first HPMA-based chemotherapeutic drug conjugate to enter clinical trials was an HPMAdoxorubicin conjugate named PK1 (17). An analogue of PK1, named PK2, is also under clinical evaluation (18). Additional poly(HPMA)-based drug (camptothecin and paclitaxel) conjugates have also been investigated (19, 20). Generally for HPMA polymer platinum constructs, conjugates are synthesized by copolymerization of HPMA monomer with another monomer unit containing an oligopeptide side chain. The latter moiety contains an activated group such as a paranitrophenoxy unit, which can be substituted with a chelating ligand. This terminus to the oligopeptide is then further reacted with a platinum reagent to form the polymer-platinum construct (21). In other synthetic methodologies, cyclic polyphosphazene is polymerized to form a linear polyphosphazene backbone unit to which a linker-glutamate chelator moiety is then attached (22). This is subsequently reacted with a platinum agent to form the drugpolymer entity. A review of current polymer-platinum drug syntheses was described by Siegmann-Louda and Carraher (23). The polymer-linked DACH-platinum agent AP5346 is a further example of an HPMA-based polymer therapeutic which was designed to deliver more platinum-containing drug to the tumor than can be achieved by conventional small molecule platinum agents. It consists of a hydrophilic polymer backbone linked by a tripeptide to a chelated platinum (II) species. The structure is denoted as poly(HPMA)-GGG-AmadPtd1R,2RDACH, where HPMA denotes the polymer backbone, GGG denotes the tripeptide glycyl-glycyl-glycine, and the AmadPtd 1R,2R-DACH denotes the N,O-amidomalonatedPt-(1R,2R)diaminoclyclohexane chelate. Preclinical data for AP5346 shows that it repeatedly exhibits superior tumor growth inhibition in numerous cell lines than oxaliplatin (24, 25). This paper describes the AP5346 manufacturing procedure and characterization methods developed prior to the initiation of clinical trials, and details of an improved synthetic procedure.
EXPERIMENTAL PROCEDURES Materials. Commercial reagents and solvents were purchased from the following sources: Pyridine (Fisher), ethanol (absolute, Aaper Alcohol), ethyl acetate (HPLC grade, Fisher), diethyl ether (ACS, Fisher), acetone (HPLC grade, Fisher), AIBN (Aldrich, recrystallized from dichloromethane), triethylamine (Aldrich), Bio-Rex MSZ 501 D resin (Bio-Rad), DACHPt(NO3)2 (Strem). Instrumentation. NMR spectra were recorded on a Bruker AVANCE300 multinuclear NMR spectrometer resonating at 300.131 MHz for 1H and 64.354 MHz for 195Pt. A Bruker Variable Temperature unit held samples at a constant temperature of 300 K. 1H NMR samples were prepared by dissolving 20 ( 5 mg of material in 0.7 mL of D2O and then filtering the sample into a 5 mm NMR tube. 1H resonances were referenced to TMSP at 0 ppm. 195Pt NMR samples were prepared by
Bioconjugate Chem., Vol. 17, No. 5, 2006 1271
dissolving 100 ( 5 mg of AP5346 in 0.7 mL of 93/7 H2O/D2O and filtering the sample into a 5 mm NMR tube. 195Pt resonances were externally referenced to Na2PtCl6 at 0 ppm. Spectra were obtained using a broadband observe 5 mm probe with z-axis gradient capability. 1H spectra were acquired using the standard Bruker pulse program zg30 with 16 scans in 25 K data points over a 3.592 kHz spectral width (3.5 s acquisition time). For 195Pt NMR experiments the 90° pulse width was calculated from the 360° pulse width using the Na2PtCl6 sample with the transmitter on resonance. The transmitter power was set to give a 90° pulse of ca. 10 µs. For samples of AP5346 the transmitter offset (o1p) was placed midway between the N,O- and N,Nchelates at -2550 ppm. The Bruker pulse program zgmultiscan was modified to execute a very short delay time (set at d1 ) 2 ms) followed by a hard 90° pulse. The experiment was run in increments of 20 000 scans (ns ) 20 000) over a 130 kHz sweep width (9 ms acquisition time). A loop counter parameter l3 ) 100 was incorporated such that the initial iteration of 20 000 scans was repeated 90 times to give an accumulated number of 2 000 000 scans (the FIDs from each iteration were automatically combined and Fourier transformed to produce the frequency domain spectrum). The FIDs of AP5346 samples were processed to determine what peaks were present and the percent of the N,O-chelate relative to the total area of all integrated peaks. To these ends the following steps were taken with the exponential line broadening set at 150 Hz. The line broadening was selected prior to Fourier transform of the FID and may be increased or decreased appropriately to obtain the best resolved spectrum. First, the FID was left-shifted, Fourier transformed, and phased. This step was repeated until the baseline was flat. Second, a polynomial baseline correction was applied to the spectrum. Finally, the integral values were set such that the total of the integrated peaks was 100 which produced a convenient percent ratio of chelates. 1H and 195Pt NMR were used to determine both identity and purity of AP5346. For identity, the former indicates the presence of HPMA and MA-Gly-Gly-Glyderived subunits while the latter shows the chelation environment of the platinum complexes and the chelate ratio. 1H NMR was used to determine the percent of impurities with fixed or nonexchangeable hydrogen atoms such as ethanol and isopropyl alcohol. 195Pt NMR was used to verify purity in that no less than 90% of the integrated peaks correspond to the amidomalonato-N,O-chelate. A Viscotek TriSec Model 302 gel permeation chromatography (GPC) instrument was used to determine molecular weight distribution as reported by the Mw, Mn, and polydispersity index (PDI) of AP5346. The system consists of an HPLC pump with an autosampler, a refractive index detector, and a column oven set at 40 °C. Three PL Aquagel-OH mixed bed columns (one 70 × 7.5 mm guard column and two 300 × 7.5 mm analytical columns with 8 µm particles) were used. The mobile phase consists of 35/65 MeOH/H2O with 10 mM LiClO4 and was pumped at 1.0 mL/min. The columns were calibrated with PEO/ PEG standards (∼1.5 to 205 kD) and results fit to a fourth order polynomial of log(Mp) vs RT-1. The sample was dissolved in mobile phase to give a ∼1 mg/mL solution of which 100 µL was injected. The standards and sample were run in triplicate. The reported values for Mw, Mn, and PDI represent the average of the three determinations. Inductively coupled plasma-atomic emission spectroscopy (ICP-AES) was used for the determination of platinum in two identity analyses and one purity analysis. The identity tests were the determination of %Pt in AP5346 and the %Pt released from the polymer system at 3 and 24 h in PBS at 38 °C. The purity test was the determination of the %free Pt in pure water at ambient temperature. While release of low molecular weight platinum species within the tumor extracellular fluid or the
1272 Bioconjugate Chem., Vol. 17, No. 5, 2006
malignant cell itself is probably required for anticancer activity, the pharmaceutical constraints of a drug product mandate that the platinum complex remains bound to the polymer prior to administration and that release of platinum from AP5346 be minimal while in systemic circulation. The release of platinum species from the polymer carrier of AP5346 is of interest for both pharmaceutical and physiological reasons. From a pharmaceutical perspective, the release of platinum from AP5346 is a fundamental chemical characteristic, and it is a requirement of a pharmaceutical product that this property remain within a narrow range for all batches of the manufactured product. The %Pt was determined by accurately weighing 15 ( 0.1 mg of AP5346 into each of three 20 mL scintillation vials. An accurately weighed quantity of 20 g of 3% HNO3 was added to each vial and the mixture was vortexed to give a homogeneous sample. Platinum standards at 0, 30, and 60 ppm were prepared from a dilution of a 1000 ppm certified Pt standard in 3% HNO3. The emission at 214.42 nm of the samples and standards were detected. From these standards a calibration curve was generated from which the amount of platinum in the sample was calculated. The results from the three samples were averaged and reported as the %Pt. The %Pt released at 3 and 24 h from a PBS solution at 38 °C was determined by weighing 42 ( 0.1 mg of AP5346 into a 50 mL Corning centrifuge tube. Sufficient PBS solution (Sigma Chemicals: 120 mM NaCl, 2.7 mM KCl, 10 mM sodium phosphate salts, pH ∼7.4) was added to make a ∼2 mg/mL solution, and the mixture was vortexed to homogeneity. The solution was placed in an oven at 38 °C. After 3 h a 2.5 mL aliquot was transferred to each of three prewashed Centricon YM-3 (Millipore Corporation, 3 kDa NMWCO) ultracentrifugation devices. The charged devices were spun down at 38 °C and 4100 rpm for 90 min. After 24 h at 38 °C three additional prewashed Centricon YM-3 devices were similarly charged and centrifuged. A sample for the total Pt content is prepared by diluting 200 µL of the stock solution with 4 mL of the HNO3/ PBS solution. The six filtrates were transferred into 7 mL scintillation vials containing 4.0 mL of a 3% HNO3/PBS solution from which a calibration curve was generated. The Pt concentration of the stock solution and the 3 and 24 h samples were determined from the emission at 214.42 nm. The %Pt released at 3 and 24 h is calculated from the ratio of the Pt content in the filtrates relative to the stock solution. The average of the three samples at each timepoint was reported. The %free Pt of AP5346 was determined in pure water at ambient temperature. The method used was similar to that described above for 3 and 24 h release rates in PBS except the stock solution was allowed to stand for 1 h before being centrifuged. In this case, however, the calibration curve is generated using 0, 5, and 10 ppm Pt standards. The free Pt was a combination of (i) unbound small molecule Pt species, (ii) small molecule Pt species released from the polymer, and (iii) Pt attached to lower molecular weight polymer chains which have breached the membrane. The sodium, chloride and phosphate (as phosphorus) contents were determined by elemental analysis by Desert Analytics, Tucson, AZ. These values represent the extent to which salts used in the synthesis have been removed. The sodium analysis also provides details regarding the purity of AP5346, which is a sodium salt and should contain one sodium per amidomalonato-bound platinum. This equates to a theoretical %Na of around 1%. Tangential Flow Filtration (TFF). AP5346 was purified by an ultrafiltration technique called tangential flow filtration (TFF). This technique, where the filtered solution is pumped tangentially to the membrane, utilizes the large difference in molecular weights for the separation between AP5346 and any small
Sood et al.
molecules or ions. The TFF membranes retain large molecules and allow the small molecule or salt impurities to pass through. The retained solution is called the retentate, and the filtered solution is called the permeate or filtrate. Water for injection (WFI) quality water is used to make up the water that appears in the permeate. When seven or more volumes of permeate relative to the initial retentate volume have been collected, the concentration of small impurities have been reduced to
