`ft~~ttJ1
`
`M!
`
`January 2009
`
`Chem. Phann. Bull. 57(1) 43-48 (2009)
`
`43
`
`Freeze-Drying of Proteins in Glass Solids Formed by Basic Amino Acids
`and Dicarboxylic Acids
`
`Ken-ichi lZUTsu,*· 0 Saori KADoYA,b Chikalm YoMarA, 0 Tom KAWANISHI,0 Etsuo YoNEMOCHI,h and
`Katsuhide TERADAb
`
`0 National Institute of Health Sciences; 1-18-1 Kamiyoga, Setagaya, Tokyo 158-8501, Japan: and b Faculty of
`Pharmaceutical Sciences, Toho University; 2-2-1 Miyama, Funabashi, Chiba 274-8510, Japan.
`Received August 1, 2008; accepted October 11, 2008; published online October 15, 2008
`
`The purpose of this study was to produce and characterize glass-state amorphous solids containing amino
`acids and organic acids that protect co-Iyophilized proteins. Thermal analysis of frozen solutions containing a
`basic amino acid (e.g., L-arginine, L-lysine, L-histidine) and a hydroxy di- or tricarboxylic acid (e.g., citric acid, L(cid:173)
`tartaric acid, DL-malic acid) showed glass transition of maximally freeze-concentrated solute at temperatures
`(T~) significantly higher than those of the individual solute solutions. Mixing of the amino acid with some dicar(cid:173)
`boxylic acids (e.g., oxalic acid) also suggested an upward shift of the transition temperature. Contrarily, combi(cid:173)
`nations of the amino acid with monocarboxylic acids (e.g., acetic acid) had T~s between those of the individual
`solute solutions. Co-lyophilization of the basic amino acids and citric acid or L-tartaric acid resulted in amor(cid:173)
`phous solids that have glass transition temperatures (Tg) higher than the individual components. Mid- and near(cid:173)
`infrared analysis indicated altered environment around the functional groups of the consisting molecules. Some
`of the glass-state excipient combinations protected an enzyme (lactate dehydrogenase, LDH) from inactivation
`during freeze-drying. The glass-state excipient combinations formed by hydrogen-bonding and electrostatic in(cid:173)
`teraction network would be potent alternative to stabilize therapeutic proteins in freeze-dried formulations.
`
`Key words
`
`freeze-drying; protein formulation; amorphous; stabilization; glass
`
`Freeze-drying is a popular method of ensuring the stability
`of proteins that are not stable enough in aqueous solutions
`during the period required for storage and distribution. I,2)
`Various freeze-dried protein formulations contain excipients
`(e.g., sugars, polymers, and amino acids) that protect proteins
`from physical and chemical changes. Disaccharides (e.g., su(cid:173)
`crose, trehalose) are the most popular among them because
`they stabilize proteins both thermodynamically and kineti(cid:173)
`cally in aqueous solutions and freeze-dried solids. 3---S)
`The development of freeze-dried protein formulations con(cid:173)
`taining amino acids is often more challenging than the devel(cid:173)
`opment of formulations with saccharides because of the var(cid:173)
`ied physical and chemical properties (e.g., crystallinity, glass
`transition temperature) of the freeze-dried amino acids, as
`well as their tendency to form complexes with other ingredi(cid:173)
`ents. 6> Many amino acids are considered to protect proteins
`basically in similar mechanisms with disaccharides. They
`thermodynamically stabilize protein conformation in aque(cid:173)
`ous solutions and probably in frozen solutions by being pref(cid:173)
`erentially excluded from the immediate surface of proteins. 1>
`Glass-state amorphous solids formed by freeze-drying of the
`disaccharides or some amino acids protect proteins from
`structural changes thermodynamically by substituting sur(cid:173)
`rounding water molecules. 8> They also reduce chemical
`degradation of freeze-dried proteins kinetically by reducing
`8> In addition, some amino acids
`the molecular mobility.2
`•
`(e.g., L-arginine) also prevent protein aggregation in aqueous
`solutions prior to the drying process and after reconstitu(cid:173)
`tion.9> Choosing appropriate counterions that form glass-state
`solid should be one of the key factors in designing amino
`acid-based amorphous freeze-dried formulations. 10
`11> For ex(cid:173)
`•
`ample, glass transition temperatures (Tg) of freeze-dried L(cid:173)
`
`histidine salts depend largely on the counterions. I2) Co(cid:173)
`lyophilization of L-arginine and multivalent inorganic acids
`(e.g., H3P04, H2S04) results in glass-state amorphous solids
`* To whom correspondence should be addressed. e-mail: Izutsu@nihs.go.jp
`CFAD Exhibit 1048
`CFAD v. NPS
`IPR2015-01093
`
`that protect proteins during the process and storage (e.g., tis(cid:173)
`sue plasminogen activator formulation, PDR 2003). IJ) Some
`organic acid and inorganic cation combinations (e.g., sodium
`citrates) also form high glass transition temperature amor(cid:173)
`phous solids_I4
`) Various functional groups (e.g., amino, car(cid:173)
`boxyl, hydroxyl) in the constituting molecules contributes
`significantly to form the glass-state amorphous salt solids. Is)
`Producing glass-state amorphous solids by freeze-drying of
`amino acid and organic acid combinations, and their applica(cid:173)
`tion in pharmaceutical formulations are interesting topics to
`explore. IS)
`The purpose of this study was to produce stable amor(cid:173)
`phous solids that protect proteins by freeze-drying combina(cid:173)
`tions of amino acids and organic acids. The physical proper(cid:173)
`ties of frozen solutions and freeze-dried solids containing the
`popular excipients and model chemicals were studied. The
`effect of the excipient combinations on the freeze-drying of
`lactate dehydrogenase (LDH) was also examined.
`
`Experimental
`Materials LOH (rabbit muscle) was obtained from Sigma Chemical (St.
`Louis, MO, U.S.A.). Succinic acid was produced by Kanto Chemical Co.
`(Tokyo, Japan). L-( +)-Tartaric acid, DL-malic acid, and other chemicals were
`of analytical grade and were purchased from Wako Pure Chemical (Osaka,
`Japan). The protein solutions were dialyzed against 50 mM sodium phos(cid:173)
`phate buffer (pH 7.0), and then centrifuged (1500gX5min) and filtered
`(0.45 µm, polyvinyliden difluoride (PVDF), Millipore) to remove insoluble
`aggregates before the freeze-drying study.
`Freeze-Drying A pH meter (HM-60G, TOA-DKK Co., Tokyo, Japan)
`was used to determine the pH of the aqueous solutions at 25 °C. A freeze(cid:173)
`drier (Freezvac IC, Tozai-Tsusho, Tokyo, Japan) was used to lyophilize the
`aqueous solutions. Aliquots of aqueous solutions (250 µI) in flat-bottom
`glass vials (!Omm diameter) were frozen by immersion in liquid nitrogen.
`The solutions were freeze-dried without shelf temperature control (20 h),
`and then at 35 °C (8 h). Solid samples for diffuse-reflection near-infrared
`analysis were prepared by freeze-drying the aqueous solutions (2 ml) in
`glass vials (21 mm diameter).
`Thermal Analysis Thermal analysis of frozen solutions and dried solids
`NPS EX. 2061
`CFAD v. NPS
`IPR2015-00990
`
`© 2009 Pharmaceutical s99~~t ;apan
`
`
`
`44
`
`Vol. 57,No. l
`
`was performed using a differential scanning calorimeter (DSC) (Q-10, TA
`Instruments, New Castle, DE, U.S.A.) and software (Universal Analysis
`2000, TA Instruments). Aliquots of aqueous solutions (10 µ1) in aluminum
`cells were cooled from room temperature at 10 °C/min, and then scanned
`from -70 °C at 5 °C/min. The effect of heat-treatment (annealing) on the
`thermal properties of the frozen solutions was studied after the initial heat(cid:173)
`ing scan paused at -10 °C, then the samples were maintained at this temper(cid:173)
`ature for I 0 min. Thermal data were acquired in the subsequent heating from
`-70°C at 5°C/min. Freeze-dried solids (l-2mg) in hermetic aluminum
`cells were subjected to the thermal analysis from -20 °C at 5 °C/min under
`nitrogen gas flow. Melted organic acids (approx. 5 mg, 200 °C) in aluminum
`cells were rapidly cooled to -50 °C, and then scanned at 5 °C/min to obtain
`the glass transition temperatures. Glass transition temperatures were deter(cid:173)
`mined as the midpoint (maximum inflection) of the discontinuities in the
`heat flow curves.
`Powder-X-Ray Diffraction (XRD) The powder X-ray diffraction pat(cid:173)
`terns were measured at various temperatures by using a Rint-Altima diffrac(cid:173)
`tometer (Rigaku, Tokyo, Japan) with CuKa radiation at 40kV/40mA. The
`samples were scanned in the area of 5°<20<35° at an angle speed of
`15°/min by heating at 2 °C/min from room temperature.
`Mid- and Near-Infrared Analysis A Fourier-transform infrared spec(cid:173)
`trophotometer (MB-104, Bomen, Quebec, Canada) with a gas generator
`(Balston, Haverhill, MA, U.S.A.) and Grams/32 software were used to ob(cid:173)
`tain mid-infrared spectra of freeze-fried solids. Approximately 0.5 mg of the
`solid was mixed with dried KBr powder (250mg) and made into tablets by
`compression. The KBr tablets were scanned 128 times to obtain the spectra
`in the 40(}-4000 cm- 1 region. Near-infrared spectroscopy was performed
`by using a Bruker MPA system with a diffuse-reflectance integrating-sphere
`probe (PbS detector) and OPUS software (Ettlingen, Germany). Near-in(cid:173)
`frared light was directed upward from the bottom of the glass vials contain(cid:173)
`ing freeze-dried solids to obtain the reflected signal over a range of 4000--
`12000 cm-1 with a resolution of 4cm-1 in128 scans. The freeze-dried solids
`,were measured twice by rotating the sample vials between measurements.
`Activity of Lactate Dehydrogenase in Freeze-Dried Solids Aqueous
`solutions (250 µI) containing LDH (0.05 mg/ml) and excipients were freeze(cid:173)
`dried in fiat-bottom glass vials (10 mm diameter). One of the enzyme solu(cid:173)
`tions was freeze-dried at a higher sodium phosphate buffer concentration
`(50mM, pH 7.0). Other enzyme solutions contained the added excipients and
`lower concentration buffer components (<1 mM) diluted from the dialyzed
`protein solutions. Activity of LDH was obtained spectrophotometrically at
`25 °C. Each 1.0 ml of assay mixture contained 0.35 mM pyruvic acid and
`0.07mM reduced nicotinamide-adenine dinucleotide (NADH) in 50mM
`sodium phosphate buffer (pH 7 .5). The enzyme reaction was started by the
`addition of LDH solution (50 µI), and the decrease in the absorbance at
`340nm was monitored. The enzyme activity (%) relative to that before
`freezing was shown.
`
`Results
`Physical Property of Frozen Solutions The thermal
`profiles of frozen solutions containing L-histidine and citric
`acid at various concentration ratios (total 200mM) are shown
`in Fig. I. The single-solute frozen L-histidine solution
`(200mM) showed a T~ (glass transition temperature of maxi(cid:173)
`mally freeze-concentrated solute) at - 33 .5 °C, and an
`exothenn peak that suggests eutectic crystallization at around
`- 8 °c.12> Freeze-drying of solutions at above their T~ often
`induces physical collapse because of the significantly re(cid:173)
`duced local viscosity in the freeze-concentrated phase. 1> The
`second scan of the 200mM L-histidine solutions after the
`heat-treatment ( -10 °C, 10 min) gave flat thennograms that
`indicate crystallized solute up to the ice melting temperature
`(data not shown). The citric acid solution (200mM) had a Tg'
`at -55.1 °C, indicating that the solute remained amorphous
`in the freeze-concentrated phase surrounding ice crystals.
`The L-histidine crystallization peak disappeared in the pres(cid:173)
`ence of citric acid. The two-solute frozen solutions showed
`transitions ( Tg' s) at temperatures as high as - 22.8 °C at the
`equal (IOOIDM) L-histidine and citric acid concentrations.
`Figure 2 shows transition temperatures (T~) of frozen solu-
`
`L-Histidine, Citric Acid (mM)
`
`§
`
`" ~
`"
`t
`,.
`u:
`" I
`
`'
`
`'0
`20
`40
`60
`80
`100
`120
`140
`160
`180
`200
`
`200
`
`'180
`160
`140
`120
`100
`80
`60
`40
`20
`0
`
`-60
`
`-50
`
`-40
`
`-30
`Temp, c•c)
`
`-20
`
`-10
`
`0
`
`Fig. 1. Thermal Profiles of Frozen Solutions Containing L-Histidine and
`Citric Acid
`Aliquots (10 µI) of solutions in hermetic aluminum cells were scanned from -70 °C
`at 5 °C/min. Glass transition temperatures of maximally freeze-concentrated solutes
`(T;) are indicated by inverted triangles (T).
`
`o Citric
`• L-Tartaric
`A Oxalic
`A DL-Malic
`+ Succinic
`
`-30
`
`-50
`
`A) L-Arginine and Organic Acids
`
`:~g f-~--'--~--'-~--'---~~-~-~-1
`o Citric
`• L-Tartaric
`A Dl-Malic
`
`~~~: .~
`
`B) L-Histidine and Organic Acids
`
`p
`!::'.' -30
`i!?
`~
`~ -40
`E
`~
`c
`0
`~ -50
`~ -60
`
`-20
`
`C) Amino Acids and Citric Acid
`
`o L-Lysine
`• L-Glutamine
`
`0,2
`
`0,6
`OA
`Organic Acid Fraction Ratio
`
`0,8
`
`Fig. 2. Glass Transition Temperatures of Maximally Freeze-Concentrated
`Solute (T~) in Frozen Solutions Containing an Amino Acid and an Organic
`Acid at Varied Concentration Ratios (Total 200mM, Average±S.D., n=3)
`
`tions containing amino acids and organic acids at various
`concentration ratios. Some single-solute frozen amino acid
`or organic acid solutions (200mM) had apparent T~ transi(cid:173)
`tions at -44.2 °C (L-arginine), -55.l °C (citric acid), and
`
`Page 2
`
`
`
`January 2009
`
`45
`
`-57.1 °C (L-tartaric acid). The frozen L-glutamine solution
`showed both T~ (-42.8 °C) and the subsequent eutectic crys(cid:173)
`tallization peak (approx. -25 °C) in the heating scan (data
`not shown). Thermograms of the frozen L-lysine and DL(cid:173)
`malic acid solutions inclined gradually without apparent
`transition up to the ice melting endotherm, which suggested
`T~s lower than -60 °C. Exotherm peaks either in the cooling
`process (glycine, acetic acid) or in the heating scan (oxalic
`acid) indicated eutectic crystallization in the frozen solu(cid:173)
`tion.16) Potential T~ transitions of some frozen solutions that
`also showed eutectic crystallization peaks (e.g., 200ITIM L(cid:173)
`histidine or L-glutamine) were not included in the figure. The
`limited solubility of some amino acids and organic acids
`(e.g., L-glutamic acid, fumaric acid, maleic acid) prevented
`them from undergoing thermal analysis at 200 ITIM. A lower
`concentration glutamic acid solution (100 ITIM) showed a T~
`at - 32.2 °C and an exotherm peak that suggests eutectic
`crystallization at around -11.0 °C (data not shown).
`Mixing of the solutes induced some unique physical prop(cid:173)
`erties in the frozen solutions that depend on the number of
`functional groups in the consisting molecules. The transition
`temperatures (T~s) of frozen solutions containing a basic or
`neutral amino acid (L-histidine, L-arginine, L-lysine, L-gluta(cid:173)
`mine, glycine) and a hydroxy di- or tricarboxylic acid (citric
`acid, L-tartaric acid, DL-malic acid) showed bell-shaped pro(cid:173)
`files. The frozen solutions containing a hydroxy di- or tricar(cid:173)
`boxylic acid (citric acid, L-tartaric acid) and an acidic amino
`acid (L-glutamic acid) did not show the mixing-induced up(cid:173)
`ward T~ shift. Citric acid also effectively prevented the crys(cid:173)
`tallization of glycine in the frozen solutions. Dicarboxylic
`acids (succinic acid, maleic acid, fumaric acid, oxalic acid)
`showed a high tendency to crystallize in the single-solute
`frozen solutions and in some mixture frozen solutions. 15·17>
`The frozen solutions containing L-arginine and oxalic acid or
`succinic acid also presented the high transition temperature
`(T~) by mixing. A mono-carboxylic acid (acetic acid), a hy(cid:173)
`droxy mono-carboxylic acid (glycolic acid), and HCl did not
`show the upward T~ shift in the mixture with the basic amino
`acids. 13>
`Physical Property of Freeze-Dried Solids Freeze-dry(cid:173)
`ing of the single-solute amino acid solutions resulted in
`cylindrical cakes that showed varied crystallinity in the pow(cid:173)
`der X-ray diffraction (XRD) and thermal analyses (Figs. 3,
`4). Freeze-dried L-arginine showed the typical harrow XRD
`pattern of amorphous solids. Thermal scan of the solid
`showed the glass transition (52.6 °C) and subsequent crystal(cid:173)
`lization exotherm (105-110°C). Freeze-dried L-histidine
`showed largely amorphous XRD pattern (30 °C) with the
`broad glass transition ( 65-100 °C) and crystallization at
`varied temperatures (120--150 °C). The L-arginine and L-his(cid:173)
`tidine solids showed apparent crystallization peaks in the
`XRD patterns at the elevated temperature ( 150 °C). The dried
`L-glutamine (200 ITIM) solids showed features of both crys(cid:173)
`talline (e.g., peaks in the XRD pattern) and amorphous (e.g.,
`glass transitions and heat-induced crystallization exotherm)
`solids. The solute concentration in the initial solution and
`thermal history in the freeze-drying process should deter(cid:173)
`mine the crystallinity of the freeze-dried L-histidine and L(cid:173)
`glutamine.12> Glycine was freeze-dried as f3 polymorph crys(cid:173)
`tal.18) Freeze-drying of citric acid or L-tartaric acid solutions
`(200ITIM) resulted in unstructured or particulate solids that
`
`200 mM Arg
`
`30°C
`
`150°C
`
`..... --~~~-~~~~--
`100 mM Arg, 100 mM Citric acid 30°C
`_.,.,....,lln•' r
`,...,,.
`150°C
`
`200 mM His
`
`30°C
`
`.....
`
`30°C
`100 mM His, 100 mM Cttric acid
`""'"'tfbc
`4'WA:t 11'' I...:''\ al
`J';•1to'
`150°C
`1!1'•rt.- tft I\> 0111 '•'t
`.. 0'1t"rll''f
`30°~- ;..>;
`200 mM Gin
`,, d ~ :!-..
`"-' ...........
`100 mM Gin, 100 m Citric acid
`30°C
`
`I,..,.
`
`200mMGly
`
`30°C
`
`30°C
`140 mM Gly, 60 mM Citric acid
`.....,..,.,,,~Lt.i.t';' .. ''' .,,,.,1 • tk ta 1t .JllJ~ll'Jll
`10
`20
`30
`
`28
`
`Fig. 3. Powder X-Ray Diffraction Patterns of Freeze-Dried Solids Con(cid:173)
`taining Amino Acids and Citric Acid
`
`200 mMArg
`
`100 mM Arg
`
`'
`100 mM Citric acid '
`
`200 mM His
`
`E
`~ a 'C
`
`<:
`Q)
`
`";i
`0
`u::
`
`'" Q)
`
`I
`
`100 mM His
`100 mM Citric acid
`
`200 mM Gin
`
`100 mM Gin
`100 mM Citric acid
`
`'
`
`200 mM Gly
`
`r
`
`'
`
`y
`
`100 mM Gly, 100 mM Citric acid
`
`'
`
`-
`
`-20
`
`20
`
`60
`
`100
`Temp. (°C)
`
`140
`
`180
`
`Fig. 4. DSC Thermograms of Freeze-Dried Solids Containing Amino
`Acids and Citric Acid
`Freeze-dried solids (l-2mg) in hermetic aluminum cells were scanned from
`-20 °C at 5 °C/min.
`
`indicate physical collapse in the primary during process.
`Amorphous solids of the organic acids prepared by rapid(cid:173)
`cooling of the melt liquid showed glass transition at 9.2 °C
`(citric acid) and 68.1 °C (L-tartaric acid) in the thermal scan
`(n=3).19)
`Co-lyophilizing the basic or neutral amino acids (L-argi(cid:173)
`nine, L-histidine, L-glutamine, glycine) and the organic acid
`(citric acid, L-tartaric acid) produced cylindrical non-crys(cid:173)
`talline cake solids at wide initial concentration ratios (Figs.
`3-5). The solids obtained by freeze-drying the basic amino
`acids (L-arginine, L-histidine) with citric or L-tartaric acid
`showed glass transition at temperatures (Tgs) much higher
`
`Page 3
`
`
`
`46
`
`than those of the individual components. The trans1t10ns
`were observed at temperatures as high as 89.5 °C (140mM L(cid:173)
`arginine, 60lllM citric acid) or 98.5 °C (160mM L-histidine,
`40 mM citric acid). Shrinking of some solids containing
`higher ratio of organic acid during the freeze-drying process
`suggested their low glass transition temperatures. The XRD
`and thermal analysis also indicated that the co-lyophilized
`solids remained amorphous up to 150 °C. Some binary
`freeze-dried solids showed a broad endotherm that suggests
`component decomposition at the elevated temperatures. The
`mixing of L-arginine with citric acid and with L-tartaric acid
`showed similar Tg profiles, in spite of the large difference in
`their transition temperatures of the cooled-melt solids. The
`bell-shaped profiles of the transition temperatures were sig(cid:173)
`nificantly different from the reported transitions of binary
`nonionic molecule systems that follow Gordon-Taylor equa(cid:173)
`tion.20> Glass transition temperatures of amorphous solids
`containing ideally mixed nonionic molecules without partic(cid:173)
`ular attractive or repulsive interactions shift between those of
`the individual components. Contrarily, the glass transition
`temperatures of co-lyophilized L-glutarnine and citric acid
`
`Vol. 57, No. 1
`
`combination solids shifted linearly between those of the indi(cid:173)
`vidual components, which suggested absence of the particu(cid:173)
`lar attractive interactions between the heterogeneous mole(cid:173)
`cules in the solids. Co-lyophilization of glycine and citric
`acid resulted in amorphous cake solids only at limited molar
`ratios.
`Transition temperatures (T~, T J of the excipient combina(cid:173)
`tions obtained at a fixed (0.1) molar ratio interval were plot(cid:173)
`ted against the pH of the initial solutions (25 °C, Fig. 6).
`Some mixtures (e.g., L-arginine and citric acid, L-histidine
`and citric acid) yielded high T~ frozen solutions and high Tg
`freeze-dried solids from weakly acidic initial solutions
`(-35°C<T~, 80°C<Tv pH 4-6), which are preferable in
`parenteral protein formulations. Small changes in the L-argi(cid:173)
`nine and organic acid compositions (0.1 molar fraction) sig(cid:173)
`nificantly shifted pH at the neutral region.
`The mid- and near-infrared spectra of the freeze-dried L(cid:173)
`arginine and citric acid combinations showed broad absorp(cid:173)
`tion bands that are typical of amorphous solids (Figs. 7, 8).21>
`Co-lyophilization with citric acid reduced an amino group
`absorption band of L-arginine at 1550cm- 1 in the mid-IR
`spectra (K.Br method), indicating altered environment of the
`functional group. Similar reduction of the amino group band
`has been reported in L-arginine-HCl salt crystal and L-argi-
`
`0
`
`0.4
`0.8
`0.6
`0.2
`Otganlc Acid Matar Fraction
`
`Fig. 5. Glass Transition Temperatures of Freeze-Dried Binary Solids
`Each symbol denotes transition of solids containing L-arginine and citric acid (0 ), L(cid:173)
`arginine and tartaric acid (.6.), L-histidioe and citric acid (•), L-glutamine and citric
`acid (.A), or glycine and citric acid(•) (total: 200tnM, average±S.D., n=3).
`
`-20
`
`-30
`
`E -40
`~ .a
`~ -50
`Q)
`CL
`E
`Q)
`-60
`I-
`c 2 100
`·u;
`c
`~
`I-
`
`80
`
`60
`
`40
`
`20
`
`10
`
`Frozen Solution, T9
`
`'
`
`--0- Arg I Gillie Acid
`_,,_ Arg I Tartaric Acid
`__._ His I Citric Acid
`__.._ Gin I Citric Acid
`------ Gly I Citric Acid
`
`Freeze-dried Solid, T9
`
`0
`
`2
`
`4
`
`8
`
`10
`
`12
`
`Solution pH
`
`Fig. 6. Effect of Initial Solution pH (25 °C) on the Transition Tempera(cid:173)
`tures of Frozen Solutions (T~ and Freeze-Dried Solids (T•) Containing an
`Amino Acid and an Organic Acid at a Fixed (OJ) Molar Concentration
`Ratio Intervals (200 mM Total, n = 3)
`
`L-Arginine, Citric Acid
`(mM)
`(mM)
`
`200
`
`180
`
`160
`
`140
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`40
`
`60
`
`80
`
`100
`
`120
`
`140
`
`160
`
`2000
`
`1600
`1800
`Wavenumbers (cm-1
`)
`
`1400
`
`1200
`
`Fig. 7. Mid-Infrared Spectra of Freeze-Dried L-Arginine and Citric Acid
`Combinations Obtained by a K.Br Tablet Method (128 Scans)
`
`L-Arginine, Citric Acid
`(mM)
`(mMJ
`200
`0
`
`180
`
`160
`
`140
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`40
`
`60
`
`80
`
`100
`
`120
`
`140
`
`160
`
`9000
`
`8000
`
`7000
`6000
`Wavenumber (cm.1
`
`)
`
`5000
`
`4000
`
`Fig. 8. Diffuse-Reflection Near-Infrared Spectra of Freeze-Dried L-Argi(cid:173)
`nine and Citric Acid Combinations Obtained at the Bottom of the Glass
`Vials (128 Scans)
`
`Page 4
`
`
`
`--I
`
`J
`
`~
`
`January 2009
`
`Mexcipienl
`50 mM Sodium phosphate buffer
`
`300 mM sucrose
`200mMAtg
`160 mM Alg. 40 mM Citric acid
`160 mM Arg, 40 mM Tartaric acid
`200mMHi$
`160 mM His. 40 mM Citric acid
`160 mM His. 40 mM Tartaric acid
`200mMGln
`160 mM G!n, 40 mM Citric acid
`200mMGty
`160 mM Gly, 40 mM Cilri<; acid
`200 mM Citric acid
`
`0
`
`20
`
`60
`40
`Residual Activity(%)
`
`so
`
`100
`
`Fig. 9. Effect of Amino Acid and Organic Acid Combinations on the Ac(cid:173)
`tivity of Freeze-Dried Lactate Dehydrogenase (50 µg/ml, Average±S.D.,
`n=3)
`
`nine freeze-dried with inorganic acids (e.g., HCI, H3P04).13l
`A carboxyl group band at 1725 cm-1 appeared when the cit(cid:173)
`ric acid ratio was increased. Diffuse-reflection near-infrared
`spectra obtained non-destructively at the bottom of the glass
`vials also indicated the altered local environment of the func(cid:173)
`tional groups. A large amino band oh-arginine (6505cm- 1,
`N-H stretching 1st overtone) disappeared in the presence of
`lower molar concentration ratio of citric acid in the initial so(cid:173)
`lution ( 140 mM L-arginine, 60 mM citric acid). Increasing the
`citric acid ratio also reduced the large absorption band at
`4920cm-1, and concomitantly induced band at 5030cm-1 in
`the co-lyophilized solids. Assignment of these bands remains
`to be elucidated. The results strongly suggested hydrogen(cid:173)
`bonding and/or electrostatic interactions between L-arginine
`and citric acid in the lyophilized solids.
`Effect of Excipients on Inactivation of Freeze-Dried
`LDH Freeze-drying of LDH in the absence of the stabiliz(cid:173)
`ing excipients resulted in significant reduction of the activity
`(approximately 15% of the initial solution) (Fig. 9). Higher
`enzyme activity was retained in freeze-drying at a higher
`phosphate buffer concentration (50ITIM). Some amino acid
`and organic acid combinations that provide neutral to weakly
`acidic initial solution (pH 5-8) and amorphous dried solids
`also retained the enzyme activity. The enzyme lost most of
`the activity in freeze-drying from extreme pH solutions (e.g.,
`200 mM L-arginine, pH 10.6). Addition of citric acid or L-tar(cid:173)
`taric acid slightly reduced the effect of L-histidine to retain
`the activity of LDH during freeze-drying. Crystallization of
`glycine in the single-solute frozen solution, and concomitant
`loss of the protecting effect, should explain the lower remain(cid:173)
`ing enzyme activity. 1·2·22)
`
`Discussion ·
`The freeze-drying of aqueous solutions containing some
`basic or neutral amino acid (e.g., L-arginine, L-histidine) and
`hydroxy di- or tricarboxylic acid (e.g., citric acid, L-tartaric
`acid) combinations resulted in the glass-state amorphous
`solid cakes that protect proteins from dehydration stresses.
`Some of the solids showed glass transition temperatures
`comparable to those of disaccharides (e.g., sucrose, tre(cid:173)
`halose).4l The data and recent literature on the properties of
`related substances in different physical states (e.g., complex
`
`47
`
`crystals, ionic liquids) strongly suggested contribution of the
`multiple functional groups of the consisting molecules to
`form the interaction (e.g., electrostatic, hydrogen-bonding)
`networks required for the glass-state amorphous solids.23- 25l
`Multiple amino, carboxyl, and hydroxyl groups in the solute
`molecules raise transition temperatures of the mixture frozen
`solutions (T;J and the freeze-dried solids (Tg).15) The ammo(cid:173)
`nium carbohydrate ion pairs form multiple hydrogen-bond(cid:173)
`ings in some non-polar solvents.23,24l Differently protonated
`carboxyl and carboxylate groups also form an intermolecular
`hydrogen-bonding network. 25>
`The amino acids and organic acids containing plural
`amino or carboxyl groups should have large chance to form
`the interactions with multiple counterpart molecules. The
`contribution of the multiple functional groups should explain
`the high transition temperatures (T~, Tg) of the L-arginine and
`citric acid combination. L-Arginine also forms stable amor(cid:173)
`phous freeze-dried solids with multivalent inorganic acids
`(e.g., H3P04).11·13> Frozen sodium citrate and tartrate buffer
`solutions exhibit the highest T~ at certain sodium concentra(cid:173)
`tion ratios.17) Supramolecular interactions (e.g., peptide-like
`periodic interactions) reported in some complex crystals of
`amino acid and dicarboxylic acid (e.g., L-arginine and adipic
`acid, X-ray analysis)26) should support the possible multi-mo(cid:173)
`lecular interaction network in the less-ordered amorphous
`phase.
`Hydroxyl groups in the citric acid, L-tartaric acid, and DL(cid:173)
`malic acid should introduce additional hydrogen bonding to
`the amorphous phase. The number of hydroxyl groups in the
`component, and the accompanying change in the molecular
`interactions are major factors in determining the glass transi(cid:173)
`tion temperature of some ionic liquids composed of an amino
`acid and a 1-allylimidazolium cation. 21> The intense interac(cid:173)
`tions and resulting reduction of the molecular mobility may
`prevent the crystallization of amino acids (e.g., glycine, glut(cid:173)
`amine) at concentration ratios much lower than those of
`"inert" nonionic solutes (e.g., sucrose) or inorganic salts
`(e.g., N aCI). 17,3()-32)
`The high glass transition temperature amorphous solids
`formed by combinations of popular excipients would be a
`practical alternative to disaccharides in the design of freeze(cid:173)
`dried protein formulations. The excipient combinations
`would satisfy the two major protein-stabilizing mechanisms
`postulated on saccharides, namely substitution of the sur(cid:173)
`rounding water molecules by hydrogen-bonding and reduc(cid:173)
`tion of the chemical reaction by embedding in the glass-state
`solids.6--8) Additional effects of some amino acids (e.g., re(cid:173)
`duced aggregation in aqueous solution by L-arginine) prefer(cid:173)
`able in protein formulations are also anticipated.9l The lim(cid:173)
`ited crystallinity and low volatility of the amino acid and or(cid:173)
`ganic acid should reduce the risk of pH change and the re(cid:173)
`sulting protein inactivation in the freeze-drying process re(cid:173)
`ported in some buffer systems.28)
`Various proteins degrade during the freeze-drying process
`and subsequent storage through several chemical and physi(cid:173)
`cal mechanisms.3·29
`) The low concentration LDH solution is
`often used as a model system for studying the effect of co(cid:173)
`solutes in the freeze-thawing and freeze-drying processes be(cid:173)
`cause of its apparent tendency to lose its activity due to irre(cid:173)
`versible subunit dissociation and conformation change. 30>
`The ability of excipient combinations to retain the enzyme
`
`Page 5
`
`
`
`48
`
`Vol.57, No. I
`
`activity in the freeze-drying process should indicate the sta(cid:173)
`bilization of the quarterly structure against freeze-concentra(cid:173)
`tion and dehydration stress. Different molecular mobility,
`local pH, water content, and crystallinity of the excipients
`may affect the chemical degradation rate of the freeze-dried
`enzyme in the subsequent storage. The freeze-dried basic
`amino acid and organic acid combination solids should pro(cid:173)
`vide the embedded proteins with unique local environments
`that are significantly different from those of the nonionic ex(cid:173)
`cipients (e.g., saccharides). The structural and chemical sta(cid:173)
`bility of proteins in these solids during the freeze-drying
`process and storage is an intriguing topic that needs further
`study through various model protein and stress systems.
`
`References
`1) Nail S. L., Jiang S., Chongprasert S., Knopp S. A., Pharm. Bio(cid:173)
`technol., 14, 281-360 (2002).
`2) Tang X., Pikal M. J., Pharm. Res., 21, 191-200 (2004).
`3) Carpenter J. F., Arakawa T., Crowe J. H., Dev. Biol. Stand., 74, 225-
`238 (1992).
`4) Franks F.,Dev. Biol. Stand., 74, 9-18 (1992).
`5) Lee J.C., TnnashetTS. N.,J Biol. Chem., 256, 7139---7201 (1981).
`6) Chang B. S., Randall C., Cryobiology, 29, 632--656 (1992).
`7) Arakawa T., Timashetf S. N., Arch. Biochem. Biophys. , 224, 169---177
`(1983).
`8) Sane S. U., Wong R., Hsu C. C., J. Pharm. Sci., 93, 1005-1018
`(2004).
`9) Tsumoto K., Umetsu M., Kwnagai I., Ejima D., Philo J. S., Arakawa
`T., Biotechnol. Prog., 20, 1301-1308 (2004).
`10) Osterberg T., Fatouros A., Mikaelsson M., Pharm. Res., 14, 892-'--898
`(1997).
`II) Mattern M., Winter G., Kohnert U., Lee G., Pharm. Dev. Technol., 4,
`
`199-208 (1999).
`12) Osterberg T., Wadsten T., Eur. J. Pharm. Sci., 8, 301-308 (1999).
`Izutsu K., Fujimaki Y., Kuwahara A., Aoyagi N., Int. J. Pharm., 301,
`13)
`161-169 (2005).
`14) Li J., Chatterjee K., Medek A., Shalaev E., Zografi G., J. Pharm. Sci.,
`93, 697-712 (2004).
`15) Kadoya S., Izutsu K., Yonemochi E., Terada K., Yomota C., Kawanishi
`T., Chem. Pharm. Bull., 56, 821- 826 (2008).
`16) Akers M. J., Milton N., Byrn S. R., Nail S. L., Pharm. Res., 12,
`1457-1461 (1995).
`17) Shalaev E. Y., Johnson-Elton T. D., Chang L., Pikal M. J., Pharm.
`Res., 19, 195-201 (2002).
`18) Chongprasert S., Knopp S. A., Nail S. L., J. Phorm. Sci., 90, 1720-
`1728 (2001).
`19) Lu Q., Zografi G., J. Pharm. Sci., 86, 1374-1378 (1997).
`20) Shamblin S. L., Taylor L. S., Zografi G., J. Pharm. Sci., 87, 694-701
`(1998).
`21) Yonemochi E., Inoue Y., Buckton G., Moffat A., Oguchi T., Yamamoto
`K., Pharm. Res., 16, 835-840 (1999).
`22) Anchordoquy T. J., Carpenter J. F., Arch. Biochem. Biophys., 332,
`231-238 (1996).
`23) Sada K., Tani T., Shin