Downloaded from journa|.pda.org on December 4, 2014
`
`RESEARCH ARTICLE
`
`Pharmaceutical Development of a Parenteral Lyophilized Formulation
`
`of the Novel Antitumor AgentAp|idine
`
`B. Nuijen“, M. Bouma‘, R. E. C. Henrar2, P. Florianos, J. M. Jimeno3, H. Ta|sma4,
`J. J. Kettenes—van den Bosch5, A. J. R .Heck5, A. Bu|t4 and J. H. Beijnen”
`
`‘Department of Pharmacy and Pharmacology, 8 otervaart Hospi tal /The Netherlands Cancer I nsti tute, Amsterdam,
`
`The Netherlands; ZNDDO-Oncology, Amsterdam, The Netherlands; 3Pharma Mar s.a., Tres Cantos, Madrid, Soai n;
`
`4Faculty of Pharmacy, Utrecht University, Utrecht, The Netherlands; 5Biomolecular Mass Soectrometry, Utrecht
`
`University, Utrecht, The Netherlands
`
`ABSTRACT: Aplidine is a naturally occurring cyclic depsipeptide isolated from the Mediterranean tunicate
`
`Aplidi um al bi cans. Aplidine displays promising in vitro and in vi V0 antitumor activities against various solid
`
`human tumor xenografts and is therefore developed now for clinical testing. The aim of this study was to
`
`develop a stable parenteral pharmaceutical dosage form for clinical Phase Itesting. Aplidine raw material was
`
`characterized by using several chromatographic and spectrometric techniques. These experiments showed that
`
`aplidine exists as two isomers. A stability-indicating HPLC assay was developed. Solubility testing showed
`
`that aplidine exhibits very poor aqueous solubility. Because solubilized aplidine showed substantial degradation
`
`under heat and light stress testing conditions, it was decided to develop a lyophilized dosage form. Freeze-
`
`drying was carried out with a 500 ug/mL solution of aplidine in 40% (V/V) tert-butanol in Water for Injection
`
`(Wfl) containing 25 mg/mL D-mannitol as a bulking agent. Differential scanning calorimetry was applied to
`
`determine the optimal freeze-drying cycle parameters. The prototype, containing 500 ug aplidine and 25 mg
`
`D-mannitol per vial, was found to be the optimal formulation in terms of solubility, length of lyophilization
`
`cycle, and dosage requirements in the forthcoming Phase I clinical studies. Quality control of the freeze-dried
`
`formulation demonstrates that the manufacturing process does not affect the integrity of aplidine. The optimal
`
`reconstitution solution was found to be l5/l5/70% (V/V/V) Cremophor EL/ethanol/Wfl (CEW). Both recon-
`
`stituted product and dilutions of the reconstituted product with normal saline (up to 1:100 V/V) appeared to be
`
`stable for at least 24 hours after preparation. Shelf-life data, available thus far, show that the lyophilized formu-
`
`lation is stable for at least 1 year when stored at +2—8°C in the dark.
`
`Introduction
`
`Aplidine (dehydrodidemnin B (DDB), MW ll09,
`
`(l, 2). This naturally occurring cyclic depsipeptide
`
`Fig. l) is a novel representative of an evolving group
`
`is isolated from the Mediterranean tunicate Apl i di um
`
`of anticancer agents derived from marine sources
`
`al bi cans and belongs to the didemnin family, a class
`
`* Author to whom correspondence should be
`
`addressed: Slotervaart Hospital/The Netherlands
`
`Cancer Institute, Department of Pharmacy &
`
`Pharmacology, Louwesweg 6,
`
`l066EC Amsterdam,
`
`The Netherlands, Tel: (31) 20 512 4737, Fax: (31) 20
`
`512 4753/4824, e-mail: apbnu@slz.nl
`
`of marine-derived compounds which exhibit
`
`antiviral, antitumor, and immunosuppressive
`
`activity. All didemnins share a common macrocy-
`
`clic peptide structure and differ only in the side-
`
`chain attached to the backbone by the amino group
`
`of threonine (3). Aplidine exists in two conformers
`
`or rotamers referring to the cis and trans isomers of
`
`the pyruvoyl-proline amide bond (Fig. 2) (4).
`
`Vol.54, No.3, May/June 2000
`
`193
`
`FRESENIUS KABI 1027-0001
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`Didemnin B (DB), the most potent representative
`
`ofthe didemnin class up to now, was the first marine-
`
`derived anticancer compound to enter clinical trials
`
`in the early 1980’s. Although interesting results were
`
`Materials and Methods
`
`Chemicals and Materials
`
`seen in this phase I/II programme sponsored by the
`
`Aplidine was obtained from natural didemnin A by
`
`National Cancer Institute (NCI), the occurrence of
`
`a three-step synthesis under the responsibility of
`
`dose-limiting neuromuscular and cardiac toxicity
`
`Pharma Mar s.a. (Tres Cantos, Madrid, Spain) and
`
`hindered further dose-escalation and repeated cycles
`
`provided through the New Drug Development
`
`oftherapy (2). The dehydro-derivative of DB, dehy-
`
`Office-Oncology (NDDO-Oncology, Amsterdam,
`
`drodidemnin B (DDB) or aplidine, displays even
`
`The Netherlands). All chemicals used were of
`
`more potent in vitro and in vivo antitumor activity
`
`analytical grade and were used without further
`
`against various solid human tumor xenografts. As
`
`purification. Distilled water was used throughout.
`
`its parent compound, the antitumor effect of aplidine
`
`Excipients, including tert-butanol, and primary
`
`is believed to be primarily mediated through inhibiti-
`
`packaging materials used in the manufacturing of
`
`on of the cell cycle progression in the G1 phase by
`
`aplidine lyophilized product and reconstitution
`
`binding to elongation factor lot in the presence of
`
`solution were of European Pharmacopeia III (Ph.
`
`GTP, thus interfering with protein synthesis (5, 6).
`
`In vitro experiments revealed that aplidine exerts
`
`less neurotoxicity and cardiotoxicity at antitumor
`concentrations than didemnin B. On the basis of
`
`these results, aplidine has been identified as a new
`
`generation of the didemnin class, with a possibly
`
`significantly higher therapeutic index (2, 7, 8, 9).
`
`In animal toxicology studies a maximum tolerated
`
`dose (MTD) of 1250 ug/kg body weight in mice
`
`and a MTD of 570 ug/kg body weight in rats were
`
`determined, respectively. Taking the 1/10 MTD
`
`mouse-equivalent, the initial dosage level in Phase
`
`I clinical trials was therefore defined at 350 ug/m2
`
`(10, 11). A suitable parenteral formulation contai-
`
`ning 500 ug aplidine per dosage unit was required
`
`to start early clinical studies. Because of its very
`
`poor aqueous solubility (< 0.1 mg/mL), an adequa-
`
`Eur. III) or United States Pharmacopeia 24 (USP24)
`
`grade and provided by the supplier with a Certificate
`
`ofAnalysis. Substances were approved on the basis
`
`of in-house quality controls carried out according
`
`to monographs in the mentioned pharrnacopeias.
`
`Normal saline (0.9% w/v sodium chloride in Wfl)
`
`was manufactured in-house at the Department of
`
`Pharmacy of the Slotervaart Hospital (Amsterdam,
`
`The Netherlands).
`
`Characterization of Aplidi ne Bulk Drug
`
`An interim reference standard of the aplidine bulk
`
`drug material, i.e., batch of highest purity available,
`
`was defined (Lot APL-297) and structurally
`
`characterized by fast-atom bombardment mass
`
`spectrometry (FAB-MS), nuclear magnetic
`
`resonance (NMR), and infra-red (IR) spectroscopy,
`
`and analytically characterized by high performance
`
`te vehicle had to be found to administer aplidine to
`
`liquid chromatography (HPLC) and ultraviolet/
`
`the patient. The work described here was directed
`
`visible (UV/VIS) spectrophotometry. On the basis
`
`towards developing a suitable parenteral formulati-
`
`of these results, specifications were drawn up for
`
`on for toxicological and clinical evaluation
`
`the aplidine raw material.
`
`according to the EORTC/CRC/NCI
`
`Joint
`
`Formulation Working Party (JFWP) guidelines (12,
`
`13, 14). Aplidine bulk drug substance was fully
`
`FAB-MS: The FAB mass spectrum was obtained with
`a Model JMS-SX/SX 102A tandem mass
`
`structurally and analytically characterized. The
`
`spectrometer (BEBE, JEOL, Tokyo, Japan). The
`
`formulation approach involved the use of a
`
`acceleration voltage was 10 kV. A xenon source with
`
`cosolvent-surfactant system to enhance the
`
`an energy of approximately 6 keV was used, and
`
`solubility, and lyophilization to improve the stability
`
`the matrix was glycerol. Positive ion spectra were
`
`of the compound.
`
`recorded over a mass range of 10-1500 D.
`
`194
`
`PDA Journal of Pharmaceutical Science & Technology
`
`FRESENIUS KABI 1027-0002
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`NM R-spectroscopy: 1H NMR spectra were recorded
`
`methanol/H20) were recorded with a Model UV/VIS
`
`with a Gemini 300 BB instrument (Varian Assoc.,
`
`918 spectrophotometer (GBC Scientific Equipment,
`
`Palo Alto, CA USA) at 300.1 MHz. The sample
`
`Victoria, Australia) equipped with an LEO personal
`
`(1 mg) was dissolved in deuterochloroform (CDCI3)
`
`computer and an Epson LX-400 plotter.
`
`or hexadeuterodimethylsulfoxide (DMSO-d6). In
`CDCl,, TMS was used as an internal reference; in
`
`DMSO-d, the central DMSO line was set at 2.50
`ppm.
`
`Infra-red spectroscopy: Infra-red (IR) spectra
`
`(4000—600 cm'1) were recorded on a Model PU 9706
`
`IR spectrophotometer (Philips Nederland B.V.,
`
`Eindhoven, The Netherlands) with the potassium
`
`bromide (KBr) pellet technique. The pellet consisted
`
`of 1 mg of aplidine bulk drug and 200 mg of KBr.
`
`The ratio recording mode was set on autosmooth
`and the scan time was 8 minutes.
`
`Formulation of Aplidi ne
`
`Solubility studies: The solubility of aplidine in
`
`various solvents at ambient temperature (20—25°C)
`
`was
`
`examined
`
`by
`
`accurately weighing
`
`approximately 1 mg of aplidine in a glass test tube
`
`and adding subsequent solvent volumina of 100 uL,
`
`1 mL, and 10 mL to the bulk drug. After each
`
`addition the mixture was vigorously shaken for 30
`
`seconds, placed in an ultrasonic bath for 15 minutes
`
`and examined visually under polarized light for
`
`complete dissolution of the aplidine drug substance.
`
`In this way, the solubilities of aplidine in the various
`
`UV/VIS spectroscopy: UV/VIS spectra (800—200
`
`solvents were selectively distributed over four
`
`nm) of aplidine bulk drug (50 ug/mL in 79% (v/v)
`
`solubility ranges (s < 0.1 mg/mL, 0.1 mg/mL S s <
`
`Figure 1: Chemical structure of aplidine (Hip: hydroxyisovalerylproprionylg Ist: isostatineg Leu: leucine; Pro:
`proline; Pyr: pyruvoylg Thr: thre0nine; Tyr: tyrosine).
`
`[33, 4R, 5S-lst]
`
`[23, 4S-Hip]
`
`
`
`N
`
`A
`
`[(3)-|-ell]
`
`[(3)-Pro]
`
`Q
`
`[(3)'Me2'TV|']
`
`Z
`
`o 0*“:-;~'”*-
`‘*n’LLN,;j_)
`
`[Pyr]
`
`O
`
`~ [
`
`(3)-Pr0]
`
`Vol.54, No.3, May/June 2000
`
`195
`
`FRESENIUS KABI 1027-0003
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`1mg/mL, 1mg/mL S s < 10 mg/mL and s > 10 mg/
`
`Formulation process: Aplidine lyophilized product
`
`mL, respectively). In addition, solvents in which
`
`was aseptically prepared from a 500 ug/mL aplidine
`
`aplidine dissolved were examined by diluting the
`
`solution in 40% (v/v) tert-butanol in Wfl contai-
`
`solutions 1:1, 1:5, 1:10, 1:50, and 1:100 (v/v) with
`
`ning 25 mg/mL D-mannitol as a bulking agent. The
`
`normal saline for l1’1fi1SlO1’1 in glass test tubes. After
`
`formulation solution was prepared by weighing
`
`gentle agitation, each of the dilutions was examined
`
`aplidine and D-mannitol and dissolving the
`
`visually under polarized light over a one day period
`
`for any sign of precipitation.
`
`Reconstitution of aplidine lyophilized product with
`
`solutions composed of 40/ 1 0/5 0% (v/v/v) propylene
`
`glycol 400/ethanol/polysorbate 80 (PET), 40/10/5-
`
`0% (v/v/v) propylene glycol 400/ethanol/WfI
`
`(PEW), and 5/5/90% (v/v/v), 15/15/70% (v/v/v), and
`
`30/3 0/40% (v/v/v) Cremophor EL/ethanol/Wfl was
`
`studied by adding increasing volumes of the
`
`reconstitution solutions to 500 ug aplidine lyophili-
`
`zed cake. After each addition, the resulting mixture
`
`was agitated and subsequently examined visually
`
`under polarized light. Quantitative analysis was
`
`carried out by diluting 50 uL samples of reconstitu-
`
`ted product with 950 uL of acetonitrile and
`
`subsequent injection onto the HPLC system.
`
`Furthermore, in case of complete dissolution, the
`reconstituted solution was diluted 1:10 or 1:100
`
`(v/v) with normal saline for infusion in glass test
`
`tubes. Stability of the reconstituted solution and the
`
`infusion solutions were examined visually under
`
`polarized light and by HPLC without further
`
`dilution, over a 24-hour period at room tempera-
`
`ture (20—25°C) and a normal day-night light cycle.
`
`substances by magnetic stirring in 40% (v/v) tert-
`butanol in Wfl. To make a final concentration of
`
`500 ug/mL aplidine, 40% (v/v) tert-butanol in Wfl
`was added. The formulation solution was sterile
`
`filtered through a 0.2 mm Midisart 2000 filter
`
`(Sartorius, Nieuwegein, The Netherlands). Subse-
`
`quently, 1 mL aliquots of the formulation solution
`
`were filled into 10 mL type 1 glass vials with a
`
`Model 501Dz peristaltic pump (Watson Marlow,
`
`UK). After filling, vials were partially closed with
`
`grey butyl rubber stoppers, placed in a Model
`
`Lyovac GT4 freeze-dryer (AMSCO/FinnAqua,
`
`Germany), and lyophilized. The freeze-dryer was
`
`equipped with a cold trap filled with liquid nitrogen
`
`attached to the condensor to ensure complete
`
`condensation of tert-butanol. After completion of
`
`the freeze-drying cycle, sterile filtered medical grade
`
`nitrogen gas was leaked into the freeze-drying
`chamber to reach a final vacuum of 100 mbar. Subse-
`
`quently, the vials were pneumatically closed,
`
`capped, and labeled. In-process-controls consisted
`
`of integrity testing of the filter unit, weight variation
`
`of the filling volume, and determination of the
`
`aplidine concentration of the formulation solution
`
`before and after filtration. Only clean, sterile inert
`
`materials and glassware were used throughout the
`
`manufacturing process. All critical manipulations
`
`Differential scanning calorimetry: Transition
`
`took place under a class 100 (A) down-flow conditi-
`
`temperatures and freezing characteristics of tert-
`
`on with a class 100 (B) background (Interflow,
`
`butanol, D-mannitol, and aplidine solutions were
`
`Wieringerwerf, The Netherlands). Air particle
`
`examined by differential scanning calorimetry
`
`counts in the critical areas as well as microbiological
`
`(DSC). The DSC experiments were carried out with
`
`contamination of the area and personnel were
`
`a TA Instruments DSC 2920 (TA Instruments, New
`
`monitored at operating state.
`
`Castle (DE), USA) equipped with an LNCA for low
`
`temperatures. Samples were sealed in a closed
`
`Reconstitution solution 15/ 1 5/70% (v/v/v)
`
`aluminium pan with an empty pan as reference.
`
`Cremophor EL/ethanol/Wfl (CEW) was prepared
`
`Temperature scale and heat flux were calibrated with
`
`by mixing the appropriate volumes of excipients by
`
`indium. Samples were cooled to -50°C at a rate of
`
`magnetic stirring. The solution obtained was filtered
`
`5°C/min. The DSC heating rate was 2.5°C/min.
`
`through a 0.2 mm Midisart 2000 filter unit (Sartorius,
`
`Analyses were performed under a helium purge.
`
`Goettingen, Germany) and subsequently 2 mL
`
`196
`
`PDA Journal of Pharmaceutical Science & Technology
`
`FRESENIUS KABI 1027-0004
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`aliquots were filled into type 1 glass ampoules with
`
`HPLC analysis: Aplidine was assayed by a
`
`a Model R910 ampoule fill-and-seal machine (Rota
`
`validated, stability-indicating reversed phase-HPLC
`
`den Boer B.V., The Netherlands). After filling, the
`
`method. The HPLC system consisted of a Model
`
`solution was sterilized by autoclaving (Model 6.6.15
`
`SP8800 pump (Thermo Separation Products, USA),
`
`autoclave, Koninklijke Ad Linden B.V., The
`
`a Model Spectra 200 UV-VIS detector (Spectra-
`
`Netherlands) for 20 minutes at 120°C. All
`
`Physics, San Jose, USA), and a Model SP8880
`
`manipulations involving an “open” solution or
`
`autosampler (Thermo Separation Products, USA).
`
`excipient were conducted under class 100 (A)
`
`Analyses were carried out with a Zorbax SB-C18
`
`conditions. The manufacturing of both aplidine
`
`analytical column (4.6 mm ID X 150 mm, particle
`
`lyophilized product and its reconstituted solution
`
`size 3.5 mm, Waters, USA) held at a constant
`
`were performed according to the Good Manufac-
`
`temperature of 80°C with a Model 7971 column
`
`turing Practice (GMP) guidelines (15).
`
`Figure 2: cis/trans isomerism of the pyruvoyl-proline
`amide bond.
`
`o
`
`J
`N
`
`O‘
`\rr&N+D
`
`‘L
`
`O
`
`H-R
`
`0 o
`
`H
`==;-=”-
`'
`
`R
`
`_M__N+
`
`O-
`
`,
`
`P
`
`Quality Control of Aplidi ne Lyophilized Product
`
`Quality control of aplidine lyophilized product
`
`consisted of identification by visual inspection of
`
`appearance and colour of the pharmaceutical
`
`product; determination of
`
`reconstitution
`
`characteristics and pH of the reconstituted product;
`
`content,
`
`content uniformity,
`
`and purity
`
`determination by HPLC analysis, residual moisture
`determination with the Karl-Fischer titration
`
`method, and residual tert-butanol content by gas
`
`chromatographic (GC) analysis. Furthermore,
`
`sterility of the pharmaceutical product was checked
`
`by the filtration method and the presence of bacterial
`
`endotoxins with the limulus amoebocyte lysate
`
`(LAL) test, both carried out according to the
`
`European Pharrnacopeia III.
`
`heater (Jones Chromatography, USA). The mobile
`
`phase at a flow of 0.6 mL/min consisted of a linear
`
`gradient of acetonitrile (ACN) containing 0.04%
`
`trifluoroacetic acid (TFA) 35% to 70% in 15 minutes
`
`and water containing 0.04% TFA. An injection
`
`volume of 20 uL and a total run time of 30 minutes
`
`were used. UV detection was performed at 225 nm.
`
`Under these conditions the chromatogram of
`
`aplidine consisted of a single peak eluting at 21
`
`minutes. A series of standard solutions of aplidine
`
`in ACN in the concentration range of 5 ug/mL to
`
`300 ug/mL were prepared in duplicate from a stock
`
`solution of 1 mg/mL aplidine in ACN and injected
`
`into the HPLC system. Quality control samples at
`
`concentrations of 7.5, 100, 250, and 275 ug/mL of
`
`aplidine in ACN were prepared in quadruplicate
`
`from another stock solution with separate weighing
`
`of aplidine and injected into the HPLC system.
`
`Aplidine lyophilized product was diluted to a test
`
`concentration of 250 ug/mL by dilution with a
`
`solution of 1:1 (v/v) ACN/water. Least squares
`
`regression analysis was used to calculate the slope
`
`and intercept for the standard calibration curve from
`
`measured peak areas versus concentration. Sample
`concentrations were calculated from the
`
`corresponding peak areas using the regression
`
`equation. All aplidine chromatograms were
`
`electronically stored in the computer system
`
`LABNET (Spectra-Physics, San Jose, USA).
`
`Reprocessing of aplidine chromatograms was
`
`performed using PC1000 software (Thermo
`
`Separation Products, USA).
`
`Vol.54, No.3, May/June 2000
`
`197
`
`FRESENIUS KABI 1027-0005
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`Resi dual moisture content: Residual moisture levels
`
`in aplidine lyophilized product were determined
`with the Karl-Fischer titration method. The content
`
`of a vial was quantitatively transferred to the titration
`
`Quality Control of Apl idi ne Reconstitution Solution
`
`Quality control of 15/ 15/70% (v/v/v) Cremophor
`EL/ethanol/Wfl reconstitution solution consisted of
`
`unit of a Model 65 8KF titrino apparatus (Metrohm
`
`visual
`
`inspection of the appearance of the
`
`Herisau, Switzerland) with previously dried metha-
`
`pharmaceutical solution and presence of particulate
`
`nol, and subsequently titrated using Hydranal®
`
`matter, determination of the pH, refractive index,
`
`Titrant 2.0 mg H20/mL (Riedel-de Haen, The
`
`relative density, and ethanol content. Also, as
`
`Netherlands). The end-point of the titration was
`
`aplidine lyophilized product, sterility and the
`
`determined biamperometrically.
`
`Residual tert-butanol content: Residual tert-butanol
`
`levels in aplidine lyophilized product were
`
`presence of bacterial endotoxins were examined
`
`according to the European Pharmacopeia III using
`
`the filtration method and LAL test, respectively.
`
`determined by gas chromatographic (GC) analysis.
`
`The specificity of the tests for the refractive index,
`
`The GC system consisted of a Model 5890 gas
`
`relative density, and ethanol content were examined
`
`chromatograph (Hewlett Packard, USA), equipped
`
`by comparing the results of the tests performed on
`
`with a flame ionization detector (FID), a HP-5 cross-
`
`linked phenyl-methyl silicone capillary column
`
`(25 m x 0.32 mm x 0.52 mm film thickness, Hewlett
`
`Packard, USA) and a Model 6890 series autosampler
`
`(Hewlett Packard, USA). The column temperature
`
`was held at 60°C for 2 minutes and was subsequently
`
`linearly ramped to 180°C at a heating rate of
`
`30°C/min and kept at this temperature for 2 minu-
`
`solutions composed of varying ratios of Cremophor
`EL, ethanol and Wfl. The refractive indices and
`
`relative densities of solutions composed of 5/5/90%,
`10/10/80%, 15/15/70%, 20/20/60%, 25/25/50%
`
`(v/v/v) Cremophor EL/ethanol/Wfl were measured
`and the ethanol content of 10/ 10/80%, 15/ 15/70%,
`
`20/20/60% (v/v/v) Cremophor EL/ethanol/Wfl
`solutions was determined. Also, to simulate
`
`tes. The injector temperature was kept at 275°C and
`
`erroneous compounding of the excipients, the
`
`the detector temperature at 250°C. An injection
`
`refractive index and relative density of solutions
`
`volume of 1 uL and a run time of 12 minutes were
`
`employed. tert-Butanol standard solutions were
`
`prepared from a 2 mg/mL stock solution in 40% (v/
`
`v) propylene glycol in water in a concentration range
`
`of 20-100 ug/mL. Least square regression analysis
`
`was used to calculate the slope and intercept of the
`
`composed of 15/70/15% and 70/15/15% (v/v/v)
`
`Cremophor EL/ethanol/Wfl were examined.
`
`Refractive index: The refractive indices of the
`
`various Cremophor EL/ethanol/Wfl solutions were
`determined on a Model Abbe’ 302 refractometer
`
`standard calibration curve from the measured peak
`
`(Atago, Japan) at 20°C.
`
`areas versus tert-butanol concentration. Sample
`
`solutions were prepared by reconstituting a vial of
`
`Rel ative densi ty: The relative densities ofthe various
`
`aplidine lyophilized product with 3 mL of 40% (v/
`
`Cremophor EL/ethanol/Wfl solutions were
`
`v) propylene glycol in water and injecting 1 uL onto
`
`determined by accurately weighing 1.00 mL of the
`
`the GC system. The tert-butanol concentrations in
`
`respective solutions on a Model 440 balance
`
`the sample solutions were calculated from the
`
`(Mettler Toledo) at 20°C.
`
`corresponding peak areas using the regression
`
`equation. All tert-butanol chromatograms were
`
`Ethanol content: Ethanol content of Cremophor EL/
`
`electronically stored in the computer system
`
`ethanol/Wfl solutions was determined by GC
`
`LABNET (Spectra-Physics, San Jose, USA).
`
`analysis. The GC system consisted of a Model
`
`Reprocessing of tert-butanol chromatograms was
`
`HP5710A gas chromatograph (Hewlett Packard,
`
`performed using PC1000 software (Thermo
`
`USA) equipped with a flame ionization detector
`
`Separation Products, USA).
`
`(FID) and a Tenax GC 60-80 Mesh 1 m x 2 mm
`
`198
`
`PDA Journal of Pharmaceutical Science & Technology
`
`FRESENIUS KABI 1027-0006
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`column. The column temperature was held at 105°C.
`
`analysis. 10/10/80%, 15/15/70%, 20/20/60%
`
`The injector and detector temperatures were kept at
`
`200°C. An injection Volume of 1 uL and a run-time
`
`(V/V/V) Cremophor EL/ethanol/Wfl solutions were
`diluted to a theoretical ethanol concentration of
`
`of 5 minutes were employed. Standard solutions
`
`1.6%o with distilled water. Before injection onto the
`
`containing approximately 0.75%o, 2.5%o and 4%o
`
`GC system, each sample was diluted 1:1 (V/V) with
`
`ethanol were prepared by weighing accurately
`
`the internal standard solution. Subsequently, 1 uL
`
`approximately 75 mg, 250 mg and 400 mg of
`
`of the resulting solution was injected into the GC
`
`absolute ethanol, respectively, into a 100 mL
`
`system. From the obtained ethanol/2-propanol peak
`
`Volumetric flask and adding distilled water to reach
`a final Volume of 100.0 mL. As an internal standard,
`
`ratios, the ethanol concentration (%o) was calculated
`with the standard calibration curve. The ethanol
`
`a solution containing approximately 3.5%o
`
`content of the Cremophor EL/ethanol/Wfl solution
`
`2-propanol was
`
`prepared
`
`by weighing
`
`was calculated by dividing the obtained ethanol
`
`approximately 350 mg of2-propanol into a 100 mL
`
`concentration (%o) by the theoretical ethanol
`
`Volumetric flask and adding distilled water to reach
`
`concentration (%o), times 100%.
`
`a final Volume of 100.0 mL. Before injection into
`
`the GC system, each ethanol standard was diluted
`
`Shelf-life Studies
`
`1:1 (V/V) with the internal standard solution.
`
`Subsequently,
`
`1 uL of the resulting solution was
`
`injected into the GC system. From the obtained
`
`Long-terrn stability of aplidine lyophilized product
`was studied at +2—8°C in the dark. Content and
`
`ethanol/2-propanol peak ratios, a calibration curve
`
`purity of the lyophilized product were determined
`
`was calculated using least squares regression
`
`by HPLC analysis.
`
`Figure 3: Chromatogram obtained with an isocratic HPLC system showing the double-peak for aplidine due to
`cis/trans isomerism of the pyruvoyl-proline amide bond (concentration of 60 ug/mL aplidine in ACN).
`
`0.025
`
`0.02
`
`0.015
`
`mAU
`
`0.01
`
`0.005
`
`Vol.54, No.3, May/June 2000
`
`199
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`Time (minutes)
`
`FRESENIUS KABI 1027-0007
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`Figure 4: Representative chromatogram of aplidine (Rt of 21 minutes) obtained with stability-indicating gradi-
`
`ent HPLC system (concentration of 250 pg/mL aplidine in ACN).
`
`0.2
`
`0.18
`
`0.16
`
`0.14
`
`0.12
`
`0.1
`
`0.08
`
`0.06
`
`0.04
`
`0.02
`
`mAU
`
`0.00
`
`5.00
`
`10.00
`
`15.00
`
`20.00
`
`25.00
`
`30.00
`
`Time (minutes)
`
`Figure 5: DSC thermogram of aplidine formulation solution (500pLg/mL aplidine and 25 mg/mL D-mannitol in
`40% v/v tert-butanol/water).
`
`0.0
`
`-0.5
`
`5E
`
`.
`E -1.0
`LI.
`‘is’(D
`:|:
`
`-1.5
`
`-2.0
`/\
`exo
`
`200
`
`PDA Journal of Pharmaceutical Science & Technology
`
`Temperature (YC)
`
`FRESENIUS KABI 1027-0008
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`Results and Discussion
`
`Characterization of Apl i di ne
`
`Infra-red spectroscopy: Major assignments are
`
`presented in Table 1.
`
`FAB-MS: The FAB-MS shows the protonated
`
`molecular ion [M+H]* at m/z 1110.7. The major
`
`fragmentations are represented in Table 1 and are
`
`indicative of the structure of aplidine.
`
`H PLC analysis: The existence of multiple
`
`conformations of aplidine as described in the NMR
`
`section complicated the development of a stability-
`
`indicating HPLC method. The initial assays, both
`
`‘H-NM R: Table 1 gives the partial assignments for
`
`normal phase and isocratic and gradient reverse
`
`the 1H-NMR spectrum in CDCL3, based on 1D
`
`phase high-perforrnance liquid chromatographic
`
`proton and 2D HH-COSY spectra. Two conformers
`
`(HPLC) methods, showed the presence of two
`
`appear to be present in an approximately 1:1 ratio
`
`(NH protons and methyl protons of the pyruvoyl
`
`group). The existence of two conformers relating to
`
`cis/trans isomerism of the pyruvoyl-proline amide
`
`bond was also shown with “C NMR (4). The
`
`spectrum in DMSO-d6 shows the presence of even
`
`four conformations in approximately a 35 :35: 15: 15
`
`ratio (methyl protons of the pyruvoyl group). The
`
`additional signals probably originate from Ci ytr ans
`
`isomerism of the second proline amide bond. This
`
`hypothesis was examined by carrying out
`
`temperature experiments in DMS0-d6 at 105°C
`
`which showed that only the signals for the pyruvoyl
`
`methyl group are still present at this elevated
`
`temperature. In the CDCL3 spectrum, the threonine
`
`NH signal of one conformer is more shielded than
`that of the other conformer. This indicates that the
`
`two conformers arise, indeed, as a result of Ci ytr ans
`
`isomerism at the pyruvoylproline residue (Fig. 2).
`From the literature the occurrence of more than one
`
`conformation in proline-containing peptides is well-
`
`known (16,17). The proline residue and Ci ytrans
`
`isomerization is extremely important in the folding,
`
`denaturation, and renaturation of polypeptides and
`
`proteins, and might also be a factor in the biological
`
`activity of aplidine.
`
`aplidine peaks in the chromatograms referring to the
`
`equilibrium between the cis and trans isomers of
`
`the pyruvoyl-proline amide bond (18). At ambient
`
`temperature (20—25°C) this equilibrium appears too
`
`slow to elute both isomers as one peak but also too
`
`fast to elute them as two baseline-separated peaks
`
`(Fig. 3). By lowering the column temperature this
`
`equilibrium is frozen, thus resulting in improved
`resolution. Collection of one of the individual
`
`conformations at the detector outlet and reinj ection
`
`of the component results again in a double-peak,
`
`indicating a continuous interconversion of both
`
`conformers. Although quantification of aplidine
`
`concentrations on the total area of both isomer peaks
`
`was satisfactory, the stability-indicating capability
`
`of the assay was not. It was proven that upon stress-
`
`testing of aplidine solutions degradation products
`
`co-eluted with the aplidine isomer peaks. Therefore,
`
`at present a gradient HPLC method with a column
`
`temperature of 80°C is employed. Due to this high
`
`column temperature aplidine elutes as a single peak.
`
`Furthermore, this method shows good selectivity for
`
`impurities and degradation products. A represen-
`
`tative chromatogram of aplidine is given in Fig. 4.
`
`Aplidine elutes at a retention time (R) of approxi-
`
`mately 21 minutes. The peaks eluting at
`
`UV/VIS spectrophotometry: The UV spectrum of
`
`approximately 14 and 20 minutes, respectively, are
`
`aplidine shows absorption maxima at 212 and 273
`
`impurities of the aplidine bulk drug. Identification
`
`nm, respectively (Table 1).
`
`of these impurities is ongoing.
`
`Vol.54, No.3, May/June 2000
`
`201
`
`FRESENIUS KABI 1027-0009
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`Table 1: Identification and characterization of aplidine bulk drug (Lot APL-297)
`
`Analytical Method
`
`Results
`
`Appearance
`
`FAB-MS
`
`‘H NMR (CDCI3)
`
`UV/VIS
`spectrophotometry
`
`IR spectroscopy
`
`Pale beige, amorphous powder
`
`Molecular formula: C37H37N7O33; Protonated molecular ion [M+H]+
`1110.7; 1040.7 [M+H]+-70, CH3COCO; 985.6 [M+H]+-125, Pro-CO;
`943.6 [M+H]+-167, pyruvoyl-Pro; 859.5 [M+H]+-251, Leu-diMe-keto-
`OH-hexanoic acid-H3O; 816.0 [M+H]+-294, pyruvoyl-Pro-N-Me-Leu;
`701.4 [M+H]+-409, Leu-diMe-keto-OH-hexanoic acid-NH3-OH-Me-
`heptanoic acid-NH; 295.2, pyruvoyl-Pro-N-Me-Leu; 177.1, Tyr-CH3
`(N-CH3); 169.1, pyruvoyl-Pro
`
`Proposed assignments:
`Amino acid residue: Pro/ine (1): 6 4.70 H-oc; 2.10 H-B3; 1.95 H-B3;
`3.66 H-63; 3.84 H-63; 2.54/2.52 CH3CO; N-Me-Leucine: 6 3.15/3.11
`N-CH3; 4.62 H-oc; 2.15 H-B3; 1.78 H-B3; 1.28 H-y; 0.90 CH3; Threonine:
`7.61/7.05 NH3; 4.57/4.65 H-oc; 5.29/5.17 H-B; 1.41/1.40 y-CH3;
`4-Amino-3-hydroxy-5-methylheptanoic acid: 7.20/7.19 NH3; 3.25 H-oc3;
`2.62 H-on 3; 4.08 H-8; 4.08 H-y; 1.80 H-5; 1.64 6-CH3; 0.90 H-5; 0.90 e-
`CH3; 2, 5-diMe-hydroxy-3-ketohexanoic acid: 4.23/4.19 H-oc; 1.33/1.32
`on-CH3; 5.40 H-y; 1.70 H-6; 0.90 6-CH3; Leucine: 7.86/7.81 NH3; 4.80
`H-oc; 1.60 H-I31; 1.20 H-B3; 0.90 CH3; Pro/ine (2): 5.10 H-oc; 2.90 H-B3;
`1.90 H-I32; N,O-diMe-tyrosine: 3.80 O—CH3; 2.54 N-CH3; 3.57 H-oc; 3.36
`H431; 3.16 H432; 7.08 H-1,4; 6.85 H-2,3
`
`Absorption maxima at 212 and 273 nm
`
`Characteristic absorption bands (approximately):
`3340 cm‘: N-H stretching (amide); 2960/2880 cm‘: C-H stretching
`(aliphatic); 1740 cm‘: C=O stretching; 1640 cm‘: N-C=O stretching
`(amide); 1520 cm‘: C=C stretching (aromatic); 1450 cm:‘: C-H
`deformation; 1250 cm:‘: phenolic C-O; 1170/1070 cm‘: C-O
`stretching;
`820 cm:‘: C-H bending of paradisubstituted benzene
`
`Solubility Studies
`
`Cremophor EL/ethanol did not precipitate over a
`
`24-hour period. Apparently, a cosolvent in combi-
`
`Table 2 gives the approximate solubilities of aplidine
`
`nation with a micelle-forming agent is necessary to
`
`in various solvent systems. Solvents were chosen
`
`on the basis of current use and experience in clinical
`
`practice (l4, l9). Aplidine is very poorly soluble in
`
`solubilize aplidine and make it suitable for further
`dilution in infusion solution.
`
`water, but dissolves well in a number of organic
`
`Initial stability studies showed that aplidine in
`
`solvents. Furthermore, aplidine has high solubilities
`
`solution degrades under the influence of light and
`
`in 60/40/10% (v/v/v) polyethylene glycol 400/
`
`heat. Therefore, to increase the stability, it was
`
`ethanol/polysorbate 80 (PET) and 50/50% (v/v) Cre-
`
`decided to develop a lyophilized dosage form. As
`
`mophor EL/ethanol cosolvent/surfactant systems.
`
`the aqueous solubility of aplidine is very poor,
`
`However, upon dilution with normal saline up to l: 100
`
`another suitable freeze-drying medium had to be
`
`(v/v), only aplidine solubilized in 50/5 0% (v/v)
`
`found. On the basis of previously reported studies,
`
`202
`
`PDA Journal of Pharmaceutical Science & Technology
`
`FRESENIUS KABI 1027-0010
`
`

`
`Downloaded from journa|.pda.org on December 4, 2014
`
`50% (v/v) polyethylene glycol 4000 in water and
`various concentrations of tert-butanol in water were
`
`consisting of 500 ug/mL aplidine and 25 mg/mL
`
`D-mannitol as bulking agent in 40% (v/v) tert-
`
`selected as alter