`Freeze-Drying
`
`ANAS AL-HUSSEIN,1 HENNING GIESELER1,2
`
`1 Division of Pharmaceutics, Freeze Drying Focus Group, University of Erlangen, Erlangen 91058, Germany
`
`2 GILYOS GmbH, Wuerzburg 97076, Germany
`
`Received 2 September 2012; revised 29 November 2012; accepted 30 November 2012
`
`Published online 20 December 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23427
`
`ABSTRACT: The objective of this study was to investigate the effect ofhistidine on the stability
`of the model protein lactate dehydrogenase (LDH) during freeze-drying. Several parameters
`were varied including pH of the bulk solution, histidine concentration, and performance of an
`annealing step during freezing. First, histidine was used as a buffer in the protein formulations
`and compared with "conventional" potassium phosphate and citrate buffer systems. For this
`purpose, sucrose or mannitol was used as stabilizers. Second, the possibility of using histidine
`as both buffer and stabilizer (cryoprotectant and lyoprotectant) in the protein formulations was
`evaluated with focus on protein stability and the physical state of histidine in the final product,
`in addition to cake elegance. Protein stability was evaluated both functionally by measuring the
`activity recovery of the model protein LDH after freeze-drying and structurally by analyzing the
`protein secondary structure. LDH showed improved stability in histidine buffer in comparison
`with other buffers. Protein stability and the tendency of histidine to crystallize during freeze(cid:173)
`drying were pH dependent. Annealing destabilized LDH and resulted in a decrease of the
`activity recovery. However, the extent of protein destabilization caused by annealing appears to
`be also pH dependent.© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
`J Pharm Sci 102:813-826, 2013
`Keywords: freeze drying/lyophilization; stability; proteins; excipients; physical characteriza(cid:173)
`tion; buffers
`
`INTRODUCTION
`
`Advances in biotechnology over the last decades re(cid:173)
`sulted in an increasing number of therapeutic pro(cid:173)
`teins. Protein-based products are likely to represent
`four of the five top-selling drugs globally by 2013. 1 The
`maintenance of protein stability and efficacy in the
`dosage form presents a great challenge to the phar(cid:173)
`maceutical industry. Because of their limited stability
`in an aqueous environment, proteins often need to be
`converted into solid state to achieve an acceptable
`shelflife as pharmaceutical products.2 The most com(cid:173)
`monly used method for manufacture of solid protein
`pharmaceuticals is freeze-drying (lyophilization).3•4
`However, the freeze-drying process generates many
`stresses during both freezing and drying, which may
`cause the loss of protein bioactivity. Therefore, a range
`
`of excipients can be added to the protein formulation
`to overcome these stresses and to improve protein sta(cid:173)
`bility during freeze-drying and storage. This goal is
`commonly achieved by using amorphous excipients
`to serve as protein stabilizer during both freezing
`(cryoprotectant) and drying (lyoprotectant).3 •4 One of
`the most widely accepted protein stabilization mecha(cid:173)
`nisms during freezing is preferential exclusion, which
`means that the excipient is preferentially excluded
`from the surface of the protein. Thereby the free
`energy required for denaturation is increased and
`the native structure of the protein is stabilized.5
`During the drying phase, the protein is stabilized
`by the water replacement mechanism and by the
`formation of a viscous glassy state. The water re(cid:173)
`placement mechanism involves the formation of hy(cid:173)
`drogen bonds between a protein and an excipient
`to satisfy the hydrogen-bonding requirement of po(cid:173)
`lar groups on the protein sFrf""'"' 6 'T'h<> fnrm<:ttion
`-ying
`of an amorphous viscous glaiNPS EX. 2060
`and the corresponding ext CF AD v. NPS
`osity
`IPR2015-0l 093
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 102, NO. 3, M~~Q~o1J
`
`Correspondence to: Henning Gieseler (Telephone: +49-931-
`90705678; Fax: +49-931-90705679; E-mail: info@gilyos.com)
`Journal of Pharmaceutical Sciences, Vol. 102, 813-826 (2013)
`© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
`CFAD Exhibit 1047
`CFAD v. NPS
`IPR2015-00990
`
`813
`
`
`
`Investigation of Histidine Stabilizing Effects on LOH During
`Freeze-Drying
`
`ANAS AL-HUSSEIN,1 HENNING GIESELER1
`
`,
`
`2
`
`1 Division of Pharmaceutics, Freeze Drying Focus Group, University of Erlangen, Erlangen 91058, Germany
`
`2 GILYOS GmbH, Wuerzburg 97076, Germany
`
`Received 2 September 2012; revised 29 November 2012; accepted 30 November 2012
`
`Published online 20 December 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23427
`
`ABSTRACT: The objective of this study was to investigate the effect ofhistidine on the stability
`of the model protein lactate dehydrogenase (LDH) during freeze-drying. Several parameters
`were varied including pH of the bulk solution, histidine concentration, and performance of an
`annealing step during freezing. First, histidine was used as a buffer in the protein formulations
`and compared with "conventional" potassium phosphate and citrate buffer systems. For this
`purpose, sucrose or mannitol was used as stabilizers. Second, the possibility of using histidine
`as both buffer and stabilizer (cryoprotectant and lyoprotectant) in the protein formulations was
`evaluated with focus on protein stability and the physical state of histidine in the final product,
`in addition to cake elegance. Protein stability was evaluated both functionally by measuring the
`activity recovery of the model protein LDH after freeze-drying and structurally by analyzing the
`protein secondary structure. LDH showed improved stability in histidine buffer in comparison
`with other buffers. Protein stability and the tendency of histidine to crystallize during freeze(cid:173)
`drying were pH dependent. Annealing destabilized LDH and resulted in a decrease of the
`activity recovery. However, the extent of protein destabilization caused by annealing appears to
`be also pH dependent.© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
`J Pharm Sci 102:813-826, 2013
`Keywords: freeze drying/lyophilization; stability; proteins; excipients; physical characteriza(cid:173)
`tion; buffers
`
`INTRODUCTION
`
`Advances in biotechnology over the last decades re(cid:173)
`sulted in an increasing number of therapeutic pro(cid:173)
`teins. Protein-based products are likely to represent
`four of the five top-selling drugs globally by 2013. 1 The
`maintenance of protein stability and efficacy in the
`dosage form presents a great challenge to the phar(cid:173)
`maceutical industry. Because of their limited stability
`in an aqueous environment, proteins often need to be
`converted into solid state to achieve an acceptable
`shelflife as pharmaceutical products.2 The most com(cid:173)
`monly used method for manufacture of solid protein
`pharmaceuticals is freeze-drying (lyophilization).3•4
`However, the freeze-drying process generates many
`stresses during both freezing and drying, which may
`cause the loss of protein bioactivity. Therefore, a range
`
`Correspondence to: Henning Gieseler (Telephone: +49-931-
`90705678; Fax: +49-931-90705679; E-mail: info@gilyos.com)
`Journal of Pharmaceutical Sciences, Vol. 102, 813-826 (2013)
`© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
`
`of excipients can be added to the protein formulation
`to overcome these stresses and to improve protein sta(cid:173)
`bility during freeze-drying and storage. This goal is
`commonly achieved by using amorphous excipients
`to serve as protein stabilizer during both freezing
`(cryoprotectant) and drying (lyoprotectant).3•4 One of
`the most widely accepted protein stabilization mecha(cid:173)
`nisms during freezing is preferential exclusion, which
`means that the excipient is preferentially excluded
`from the surface of the protein. Thereby the free
`energy required for denaturation is increased and
`the native structure of the protein is stabilized. 5
`During the drying phase, the protein is stabilized
`by the water replacement mechanism and by the
`formation of a viscous glassy state. The water re(cid:173)
`placement mechanism involves the formation of hy(cid:173)
`drogen bonds between a protein and an excipient
`to satisfy the hydrogen-bonding requirement of po(cid:173)
`lar groups on the protein surface.6 The formation
`of an amorphous viscous glass during freeze-drying
`and the corresponding extremely high viscosity
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 102, NO. 3, MJR<'!Qeo13
`
`813
`
`
`
`814
`
`AL-HUSSEIN AND GIESELER
`
`also increase protein stability by retarding protein
`denaturation. 7 •8
`Disaccharides such as sucrose and trehalose are
`widely used as protein stabilizers, and these sug(cid:173)
`ars have been extensively studied in the literature
`to investigate their stabilizing effect on proteins dur(cid:173)
`ing freeze-drying. Several amino acids are frequently
`cited as being suitable excipients for freeze-drying of
`proteins. Glycine, for example, is widely used as crys(cid:173)
`talline bulking agent in freeze-dried formulations, 9
`whereas other amino acids such as lysine and argi(cid:173)
`nine have been described as possible buffers in protein
`formulations. 10 Arginine in combination with phos(cid:173)
`phoric acid was reported to exert a stabilizing effect on
`lactate dehydrogenase (LDH) during freeze-drying. 11
`In contrast to sugars, amino acids can also func(cid:173)
`tion as buffers, and therefore, provide more choices/
`flexibility for the design of proteins formulations. His(cid:173)
`tidine is one of the amino acids that can be used in
`protein formulations to function as both buffer and
`protein stabilizer. Histidine, which has three ioniza(cid:173)
`tion sites on the molecule's carboxyl, imidazole, and
`aminogroupwithpK1 ofl.9,pK2 of6.1, andpK3 of9.1,
`has been used as a buffer especially in the pH range
`5_7_12
`Several studies have already referred to a stabi(cid:173)
`lizing effect of histidine on proteins during freeze(cid:173)
`drying. Although most of the stability studies deal
`with in-process instability of proteins, there were
`some (long-term) storage-stability studies. Osterberg
`et al. 10 described the development of a stable freeze(cid:173)
`dried formulation for recombinant factor VIII-SQ (r(cid:173)
`VIII SQ) without the addition of albumin. The au(cid:173)
`thors found that a combination of sucrose, nonionic
`surfactant (polysorbate 80), crystalline bulking agent
`(sodium chloride), and L-histidine preserve factor(cid:173)
`VIII activity during freeze-drying and storage. It was
`also reported that L-histidine, besides functioning as a
`buffer, also had a stabilizing effect on r-VIII SQ during
`freeze-drying and storage. However, it is important to
`underline that the stabilizing effect of histidine was
`not studied in depth and not delineated and differen(cid:173)
`tiated from the stabilizing effect of other stabilizers
`used in the same formulation. Cleland et al.13 used
`histidine as a buffer for the freeze-drying of a mon(cid:173)
`oclonal antibody, rhuMAb HER2. The authors com(cid:173)
`pared the stability profile of rhuMAb HER2 formu(cid:173)
`lated at 25 mg/mL in either 5 mM succinate (pH 5)
`or 5 mM histidine (pH 6) in the presence of other
`excipients. They found that in the absence of sugar,
`a greater extent of aggregation was observed in the
`histidine formulation than in the succinate formu(cid:173)
`lation. Chen et al.14 found that the increase of the
`histidine concentration from 4 to 6 mM reduced the
`soluble-aggregate levels of a human anti-IL8 mon(cid:173)
`oclonal antibody (ABX-IL8) upon freeze-drying. The
`authors used multiple excipients systems consist-
`
`ing of glycine, glutamic acid, mannitol, and polysor(cid:173)
`bate 20. Furthermore, the freeze-dried monoclonal
`antibody trastuzumab (Herceptin®) produced by
`F. Hoffmann-La Roche (Basel, Switzerland) is formu(cid:173)
`lated with histidine.
`Overall, the reported studies about the stabilizing
`effect of histidine on proteins during freeze-drying are
`still limited, and all of these studies did not differen(cid:173)
`tiate between the buffering effect of histidine and the
`other possible stabilizing effects of this amino acid on
`the proteins. Furthermore, literature does not provide
`data regarding the influence of histidine on proteins
`in the absence of other excipients that are regularly
`included in protein formulations. The presence of such
`excipients might complicate the identification of the
`effect of histidine on protein stability during freeze(cid:173)
`drying.
`The scope of the present study was to investi(cid:173)
`gate the influence of histidine on the stability of a
`model protein LDH for concentrations relevant for
`use as a buffer and as a stabilizer. Histidine buffer
`was compared with other common buffers (potassium
`phosphate and citrate). Furthermore, the ability to
`use histidine as a sole excipient in protein formu(cid:173)
`lation was investigated. Moreover, the effect of pH
`and histidine concentration in the pre-freeze-dried
`bulk solution on both protein stability and the ele(cid:173)
`gance of the final freeze-dried product was investi(cid:173)
`gated. Because the amorphous state of the stabilizer
`is an essential property for the stabilization of pro(cid:173)
`tein during freeze-drying, the physical state of his(cid:173)
`tidine in freeze-dried samples with and without an(cid:173)
`nealing was analyzed and correlated with the protein
`stability. LDH was selected for this study because of
`its well-documented labile nature and sensitivity to
`the stresses generated during freeze-drying. 15 A com(cid:173)
`parably low protein concentration of 15 µg/mL was
`employed to avoid protein self-protection, which is
`present at high concentrations.6·16
`
`MATERIALS
`I-Lactate dehydrogenase type II from rabbit mus(cid:173)
`cle (11.4 mg protein/mL; 1150 units/mg) was used as
`aqueous suspension in ammonium sulfate and pur(cid:173)
`chased from Sigma-Aldrich (Munich, Germany). L(cid:173)
`Histidine, sodium pyruvate, and P-nicotinamide ade(cid:173)
`nine dinucleotide (NADH) were also obtained from
`Sigma-Aldrich at analytical grade. Sucrose was ob(cid:173)
`tained from Fluka Analytical (Buchs, Switzerland)
`and D-mannitol was purchased from Riedel-de Haen
`(Seelze, Germany). Potassium dihydrogen phosphate
`(KH2P04) and citric acid were obtained from Carl
`Roth (Karlsruhe, Germany) to prepare phosphate and
`citrate buffers, respectively. The pH of formulations
`containing histidine was adjusted to 4, 5, 6, and 7
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 102, NO. 3, MARCH 2013
`
`~~00 002/jps
`
`
`
`Table I. Overview of the Studied Formulations and the Buffers in Each Part of this Study
`
`HISTIDINE STABILIZING EFFECTS ON LDH DURING FREEZE-DRYING
`
`815
`
`Studied Aspect
`
`Histidine as a buffer (comparison
`with other buffers)
`
`Histidine as a stabilizer
`(comparison with other solutes)
`
`Histidine as buffer and stabilizer
`(histidine-concentration effect)
`
`Histidine as buffer and stabilizer
`(effect of solution pH)
`
`LDH
`(µg/mL)
`
`15
`
`15
`
`15
`
`15
`
`15
`15
`15
`15
`
`15
`15
`15
`15
`15
`
`15
`15
`15
`15
`
`Buffer Concentration
`
`pH
`
`Nonbuffer Solute Concentration
`
`Citrate (10, 50, and 150 mM)
`
`7.3 (Only buffer) Sucrose, S (lOmg/mL) Mannitol, M
`(lOmg/mL)
`Histidine (10, 50, and 150 mM) 7.3 (Only buffer) Sucrose, S (lOmg/mL) Mannitol, M
`(lOmg/mL)
`Phosphate (10, 50, and 150 mM) 7.3 (Only buffer) Sucrose, S (lOmg/mL) Mannitol, M
`(lOmg/mL)
`
`Potassium phosphate (10 mM)
`
`7.3
`
`Sucrose, S (100 mM)
`
`Histidine
`
`Histidine
`
`7.3
`7.3
`7.3
`7.3
`
`7.3
`7.3
`7.3
`7.3
`4
`
`5
`6
`7
`8
`
`Mannitol, M (100 mM)
`Histidine, H (100 mM)
`(Only buffer)
`Histidine (2 mg/mL)
`
`Histidine (5 mg/mL)
`Histidine (10 mg/mL)
`Histidine (20 mg/mL)
`Histidine (35 mg/mL)
`Histidine (20 mg/mL)
`
`Histidine (20mg/mL)
`Histidine (20 mg/mL)
`Histidine (20 mg/mL)
`Histidine (20 mg/mL)
`
`with hydrochloric acid and to 8 with sodium hydrox(cid:173)
`ide, whereas potassium hydroxide was used to adjust
`the pH in the formulations containing KH2P04 as a
`buffer. Sodium hydroxide was used also to adjust the
`pH when citrate buffer was used. Potassium hydrox(cid:173)
`ide, sodium hydroxide, and hydrochloric acid were
`purchased from Sigma-Aldrich. Vials (2 mL) were
`purchased from SCHOTT forma vitrum (Miillheim,
`Germany). FluroTec® 13-mm stoppers were obtained
`from West Pharmaceutical Services (Eschweiler,
`Germany). To detect the possible pH shifts in
`the studied formulations during freezing, a univer(cid:173)
`sal indicator solution was purchased from Sigma(cid:173)
`Aldrich.
`The range of formulations that were freeze-dried
`and analyzed is illustrated in Table 1. The experi(cid:173)
`ments were structured into four parts to evaluate the
`effect of histidine on LDH when applied as a buffer
`and as a stabilizer. First, histidine was used as a
`buffer in three different concentrations and compared
`with two other common buffer systems. The buffer
`formulations were also freeze-dried with addition of
`sucrose and mannitol. Second, histidine was used as
`a stabilizer and compared with sucrose and mannitol.
`The same buffer system was used for all combinations
`and was also freeze-dried without addition of stabiliz(cid:173)
`ers for reference. In the third and fourth part, formu(cid:173)
`lations containing only LHD and histidine in different
`concentrations and pH levels were investigated under
`consideration of the physical state.
`
`METHODS
`
`Preparation of Enzyme Solutions
`
`The LDH suspension was dialyzed with a spe(cid:173)
`cial membrane (Spectra/Por® membrane, Spectrum
`Laboratories, Rancho Dominguez, California) with
`a molecular weight cut off (MWCO) of 12-14kDa
`(molecular weight ofLDH: 140 kDa). Dialysis was per(cid:173)
`formed against potassium phosphate buffer (pH 7.3)
`at 5°C overnight. The obtained enzyme solution was
`concentrated with an Amicon Ultra-15 centrifugal fil(cid:173)
`ter device (MWCO 30 kDa; Millipore Corporation, Bil(cid:173)
`lerica, Massachusetts) in a centrifuge (Minifuge RF,
`Heraeus Sepatech GmbH, Osterode, Germany). Pro(cid:173)
`tein concentration was determined spectrophotomet(cid:173)
`rically at 280 nm. Aliquots of dialyzed LDH and ex(cid:173)
`cipients solution were mixed in glass vials to obtain
`0.5-mL samples with a final concentration of 15 µg/
`mLenzyme.
`
`Freeze-Drying Process Conditions
`
`Individual formulation containing LDH (0.5 mL) was
`filled into 2-mL vials, and the vials were subsequently
`semistoppered. Freeze-drying experiments were per(cid:173)
`formed using a FTS Lyostar™ II (SP Scientific, Gar(cid:173)
`diner, New York). Samples were loaded onto the
`shelves at room temperature, frozen using a 1°C/min
`shelf cooling rate down to-40°C, and maintained at
`this temperature for 1 h. Annealing (when applied)
`
`DOI 1 0.1 002/jps
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 102, FJ~~tJARCH 2013
`
`
`
`816
`
`AL-HUSSEIN AND GIESELER
`
`Table 2. Obtained Glass Transition Temperature (T 'g) for the
`Studied Formulations
`
`Formulationsa
`
`Only histidine formulationsc
`
`Sucrose formulationsd
`
`Mannitol formulationsd
`
`pH 4
`pH5
`pH6
`pH7
`pH8
`In phosphate buffer
`In citrate buffer
`In histidine buffer
`In phosphate buffer
`In citrate buffer
`In histidine buffer
`
`Tg (OC)b
`
`--49.66
`--49.12
`--40.80
`-35.50
`-36.85
`-34.07
`-33.25
`-30.80
`-32.35
`-33.27
`-38.58
`
`• LDH concentration in all studied formulations is 15 µ g/mL.
`bMeasurements were performed in triplicate (n = 3). All standard
`deviations were <l.4'C.
`cHistidine concentration is 20mg/mL.
`dpH 7.3.
`
`was conducted by ramping the shelf temperature from
`-40°C to -20°C at 1°C/min and keeping this shelf
`temperature for 4 h for thermal treatment. Then, the
`product was frozen back to -40°C and kept at this
`temperature for 1 h. For all formulations, primary
`drying was performed by controlling the shelf temper(cid:173)
`ature at -30°C and the chamber pressure at 80 mTorr
`for 25 h. The selected shelf temperatures for anneal(cid:173)
`ing and primary drying were selected on the basis of
`the glass transition temperatures (Tg) of the studied
`formulations, which were determined by differential
`scanning calorimetry (DSC). It is well known that an(cid:173)
`nealing should be carried out at temperature above
`T'g of the formulation. 3 The selected annealing tem(cid:173)
`perature in this study (-20°C) was higher than the
`measured glass transition temperatures of all studied
`formulation. The glass transition temperatures (T g)
`of the studied formulations are reported in Table 2.
`The conditions used are very conservative for conven(cid:173)
`tional freeze-drying, and the rationale behind the use
`of such experimental parameters is (1) to avoid any
`impact of the primary drying phase on the physico(cid:173)
`chemical state of the mixture and (2) to verify that
`differences in protein stability between different for(cid:173)
`mulations did not arise from differences in drying
`conditions.
`Lastly, secondary drying was carried out at the
`same chamber pressure applied during primary
`drying but increasing the shelf temperature to +40°C
`at a ramp rate of O.l°C/min and maintaining this
`temperature for 4 h. The shelf temperature and dura(cid:173)
`tion of secondary drying were selected to obtain final
`products with relatively low and comparable residual
`moisture content to ensure that differences in protein
`stability between different formulations are not
`attributed to different residual moisture contents.
`Throughout this study, product temperatures during
`the cycle were measured using calibrated 36-gauge
`
`from
`thermocouples
`copper/constantan
`T-Type
`Omega (Omega Engineering, Stamford, Connecti(cid:173)
`cut). Each thermocouple was introduced through a
`stopper and positioned bottom-center in the vial to
`achieve both representative temperature monitoring
`as well as accurate endpoint detection of the time
`point when no ice was left in the product.
`
`Assay to Assess the Enzymatic Activity
`
`Enzymatic-activity recovery was used in this study
`as an indicator for the functional stability of LDH.
`LDH catalyzes the interconversion ofpyruvate to lac(cid:173)
`tate with concomitant interconversion of NADH to
`NAD+. The decomposition of NADH was measured
`by the decrease in absorption at 340 nm. A 10-mm
`quartz cuvette was placed into a Lambda 25-UVNis
`spectrometer (PerkinElmer, Rodgau, Germany) con(cid:173)
`nected to a computer system (WinLab V 5.0 software,
`PerkinElmer). The rate of absorbance decrease is di(cid:173)
`rectly proportional to LDH activity in the sample. Ac(cid:173)
`tivity was measured before freeze-drying and after re(cid:173)
`constitution of the freeze-dried product. The enzyme
`activity was defined as 100% in the solution prior to
`freezing, and the remaining activity of LDH in the
`reconstituted freeze-dried samples was expressed as
`percentage of the original activity before freezing. All
`assays were performed in triplicate. Both the mean
`value and the standard deviation were calculated.
`The variation coefficient of this assay ranged between
`4.3% and 4.7%.
`
`Turbidity Measurements
`
`The turbidity of solutions was measured to quantify
`protein denaturation that led to the formation of in(cid:173)
`soluble protein aggregates. Turbidity values were ob(cid:173)
`tained by UV-spectroscopy measurements at 350 nm.
`
`Fourier Transform Infrared Spectroscopy
`
`Fourier transform infrared spectroscopy (FTIR) was
`employed to evaluate potential changes in pro(cid:173)
`tein secondary structure after freeze-drying. The
`evaluation was executed through comparison of
`second-derivative spectrum of untreated LDH be(cid:173)
`fore freeze-drying with second-derivative spectra of
`LDH obtained after the reconstitution of freeze(cid:173)
`dried samples. FTIR spectra of samples containing
`LDH were obtained using a Nicolet Magna IR 550
`FTIR spectrometer (Thermo Fisher Scientific Inc.,
`Waltham, Massachusetts). The apparatus was con(cid:173)
`stantly purged with dry air. Samples were measured
`in a temperature-controlled CaF2 window with a fixed
`sample-layer-thickness of 5.6 µm. The water spec(cid:173)
`trum was subtracted from the sample spectrum us(cid:173)
`ing the Nicolet Omnic software (Thermo Scientific,
`Waltham, MA, USA). A subtraction of the background
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 102, NO. 3, MARCH 2013
`
`~ecA 002/jps
`
`
`
`HISTIDINE STABILIZING EFFECTS ON LDH DURING FREEZE-DRYING
`
`817
`
`signal of the excipients was performed manually. The
`amide I band range between 1580 and 1720 cm- 1 was
`isolated from the spectrum and a baseline correction
`was executed. Individual peak positions were identi(cid:173)
`fied for the second-derivative spectrum of this range.
`
`Differential Scanning Calorimetry
`
`Thermal analysis was conducted using a DSC822e
`(Mettler Toledo, Greifensee, Switzerland) to evalu(cid:173)
`ate the thermal properties of the solutions and the
`freeze-dried cakes. The samples were prepared in
`40-µ L Al pans in a glove box (relative humidity < 1 % )
`and hermetically sealed at room temperature. Pans
`were weighed before (empty) and after filling on a
`calibrated microbalance (AT 261 Delta Range, Met(cid:173)
`tler Toledo) to determine the exact weight of each in(cid:173)
`dividual sample. After placing the pans in the DSC
`measurement cell, nitrogen was used for purging and
`drying (30 and 150 mUmin, respectively). The freeze(cid:173)
`dried samples were heated from room temperature to
`300°C at a heating rate of 5°C/min, whereas the liquid
`samples were cooled from room temperature to -80°C
`at cooling rate 5°C/min and held at this temperature
`for 5 min before being heated to 5°C at a heating rate
`of2°C/min.
`All thermal events including the determination of
`the glass transition temperature (T' g) were evaluated
`using the Mettler STARe Software V 9.01 (Mettler
`Toledo).
`
`X-Ray Powder Diffraction
`
`To characterize the crystallinity of the freeze-dried
`products, samples were examined by X-ray powder
`diffraction (XRPD) using a Philips X'pert MPD with
`Cu Ka radiation at 40 kV/40 mA (Philips Analytical
`Technology, Munich, Germany). Powder samples were
`filled into an Al sample holder and compressed using
`a simple cover glass. All scans were measured in the
`range 28 = 0.5°-40°, with a step size of 0.02° (time
`per step = 1 s).
`
`Residual Moisture Content (Karl-Fischer Titration)
`
`The residual moisture content in the freeze-dried
`samples was determined by Karl-Fischer titration
`using a Metrohm Coulometric KF titrator (831 KF
`Coulometer, Metrohm, Filderstadt, Germany) with an
`oven system. All measurements were performed in
`triplicate at an oven temperature of 120°C.
`
`RESULTS AND DISCUSSION
`
`Residual Moisture Contents
`
`The obtained residual moisture contents in the freeze(cid:173)
`dried cakes were comparable (about 1 %), and no
`substantial differences were observed in the water
`
`content between the freeze-dried samples. Therefore,
`the different protein stabilities in the studied formu(cid:173)
`lations (see below) did not result from different resid(cid:173)
`ual moisture contents.
`
`Comparison of Histidine as a Buffer System with Citrate
`and Phosphate Buffers
`
`Lactate dehydrogenase was freeze-dried without an(cid:173)
`nealing in three different buffers (histidine, citrate,
`and potassium phosphate) at three different buffer
`concentrations (10, 50, and 150 mM). The pH for all
`buffers was adjusted to 7.3 (see Table 1). For this part
`of the study, LDH was freeze-dried either in formula(cid:173)
`tions containing solely the studied buffers or in excip(cid:173)
`ient systems containing sucrose or mannitol (10 mg/
`mL) in addition to the buffer. The rationale for this
`combination of experiments is to obtain comprehen(cid:173)
`sive information about the effect of histidine buffer
`on protein stability in different environments, that is,
`when no other solutes are present in the formulation,
`compared with inclusion of common excipients such
`as sucrose and mannitol. Figure 1 illustrates the ac(cid:173)
`tivity recovery of LDH after the reconstitution of the
`studied formulations. Irrespective of the formulation
`composition, remaining activities ofLDH in histidine
`buffer were higher than those observed with citrate
`or phosphate buffer at all buffer concentrations. The
`addition of sucrose or mannitol to the formulations
`resulted in a significant increase in the protein ac(cid:173)
`tivity, and the higher residual activities ofLDH were
`obtained in sucrose formulations. However, remain(cid:173)
`ing activity ofLDH in histidine buffer compared with
`that in the other buffers was also higher in the pres(cid:173)
`ence of sucrose or mannitol, indicating a higher de(cid:173)
`gree of protein stability in the presence of histidine. It
`is noteworthy that remaining activity ofLDH is lower
`in phosphate buffer than in citrate and histidine, and
`this difference is very distinct in formulations con(cid:173)
`taining only protein and buffer, where protein activ(cid:173)
`ity was dramatically improved in citrate and histidine
`samples. It is well known that pH shifts can occur dur(cid:173)
`ing freezing because of the selective crystallization of
`the buffer. This change of pH might affect protein
`stability and result in protein denaturation. The pos(cid:173)
`sible pH changes in the studied formulations were in(cid:173)
`vestigated using a universal pH indicator, which was
`added to the formulations prior to freezing. The pH
`shifts were observed in all formulations formulated
`with phosphate buffer at concentrations of 50 and
`150 mM, whereas no pH shifts were observed with
`citrate or histidine formulations. Note that the selec(cid:173)
`tive crystallization of phosphate buffer during freez(cid:173)
`ing, particularly at elevated concentrations, was also
`reported by Pikal. 17
`As a side note, mannitol is rarely used as protein
`stabilizer during freeze-drying because of its strong
`
`DOI 10.1002/jps
`
`JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 102, ~EIOO ,{iARCH 2013
`
`
`
`H
`
`H
`
`818
`
`AL-HUSSEIN AND GIESELER
`
`90
`
`80 -~ 0
`70 - 60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`10mM
`
`SOmM
`
`150mM
`
`Figure 1. Activity recovery of freeze-dried LDH (15 µg/mL) in different buffers (pH 7.3) at
`10, 50, and 150 mM. Dashed columns, only buffer formulations; empty columns, sucrose formu(cid:173)
`lations; and filled columns, mannitol formulations. C, citrate; H, histidine; and P, potassium
`phosphate. Solid lines in each bar represent standard deviations.
`
`tendency to crystallize, and therefore, mannitol is nor(cid:173)
`mally used as crystalline bulking agent.2 •4 However,
`in the presented results, mannitol also showed a sta(cid:173)
`bilizing effect on LDH, which might be ascribed to
`the fact that freeze-drying was performed without
`annealing and therefore, a fraction of mannitol re(cid:173)
`mained amorphous and exerted the stabilizing effect.
`To gain further insight into protein denaturation
`mechanisms in the studied formulations, the turbid(cid:173)
`ity of protein solutions was measured after reconsti(cid:173)
`tution of the freeze-dried samples. Turbidity values
`of formulations to which only buffer was added are
`listed in Table 3. Note that very low turbidity results
`(between 0.002 and 0.004) were found in the solutions
`prior to freeze-drying. Turbidity values of freeze-dried
`histidine buffer formulations were lower compared
`with freeze-dried formulations of the other buffers at
`all used concentrations. In contrast, the turbidity val(cid:173)
`ues of the formulations containing phosphate buffer
`were significantly elevated compared with those ob(cid:173)
`tained with citrate and histidine buffers. The results,
`therefore, indicate that protein denaturation mani(cid:173)
`fested by the formation of insoluble aggregates oc(cid:173)
`curred to a lower degree in formulations containing
`histidine. Protein denaturation might be manifested
`also by the formation of soluble aggregates, but these
`aggregates play no role in the turbidity of a protein so(cid:173)
`lution and therefore, they are neglected by measuring
`the turbidity of the protein solution. Note that turbid-
`
`Table 3. Turbidity Values for LDH at 15 µg/mL of Reconstituted
`Samples Freeze-Dried in Different Buffers
`
`Fonnulationsa
`
`Turbidity A;..350
`(Optical Density at 350nm)b
`
`C10mM
`H10mM
`P1omM
`CsomM
`HsomM
`PsomM
`C1somM
`H15omM
`P1sOmM
`
`0.015
`0.009
`0.032
`0.022
`0.010
`0.038
`0.010
`0.0