DRUGS AND THE PHARMACEUTICAL SCIENCES
`
`VOLUME 175
`
`
`
`
`
`MeUME)
`TT
`NAT aTTT
`
`MYLAN - EXHIBIT 1020
`
`edited by
`Eugene J. McNally
`Jayne E. Hastedt
`
`informa
`healthcare
`
`MYLAN - EXHIBIT 1020
`
`

`

`Protein Formulation
`and Delivery
`Second Edition
`
`Edited by
`Eugene J. McNally
`Gala Biotech, a Catalent Pharma Solutions Company
`Middleton, Wisconsin, USA
`Jayne E. Hastedt
`ALZA Corporation
`Mountain View, California, USA
`
`McNally_978-0849379499_TP.indd 2
`
`9/26/07 11:10:41 AM
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Copyright_page
`
`Informa Healthcare USA, Inc.
`52 Vanderbilt Avenue
`New York, NY 10017
`
`© 2008 by Informa Healthcare USA, Inc.
`Informa Healthcare is an Informa business
`
`No claim to original U.S. Government works
`Printed in the United States of America on acid-free paper
`10 9 8 7 6 5 4 3 2 1
`
`International Standard Book Number-10: 0-8493-7949-0 (Hardcover)
`International Standard Book Number-13: 978-0-8493-7949-9 (Hardcover)
`
`This book contains information obtained from authentic and highly regarded sources. Reprinted material is
`quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts
`have been made to publish reliable data and information, but the author and the publisher cannot assume
`responsibility for the validity of all materials or for the consequence of their use.
`
`No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechan-
`ical, or other means, now known or hereafter invented, including photocopying, microfi lming, and recording, or
`in any information storage or retrieval system, without written permission from the publishers.
`
`For permission to photocopy or use material electronically from this work, please access www.copyright.
`com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood
`Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profi t organization that provides licenses and
`registration for a variety of users. For organizations that have been granted a photocopy license by the CCC,
`a separate system of payment has been arranged.
`
`Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used
`only for identifi cation and explanation without intent to infringe.
`
`Library of Congress Cataloging-in-Publication Data
`
`Protein formulation and delivery / edited by Eugene J. McNally, Jayne E. Hastedt. -- 2nd ed.
`
`
` p. ; cm. --
`(Drugs and the pharmaceutical sciences ; 175)
`
`Includes bibliographical references and index.
`
`ISBN-13: 978-0-8493-7949-9 (hardcover : alk. paper)
`
`ISBN-10: 0-8493-7949-0 (hardcover : alk. paper)
` 1. Protein drugs--Dosage forms.
`I. McNally, Eugene J., 1961- II. Hastedt, Jayne E.
`III. Series: Drugs and the pharmaceutical sciences ; v.175.
`
`[DNLM: 1. Protein Conformation. 2. Drug Delivery Systems. 3. Drug Design.
`4. Drug Stability. 5. Proteins--administration & dosage. W1 DR893B v.175 2007
`/ QU 55.9 P9667 2007]
`
`RS431.P75P77 2007
`615’.19--dc22
`
`Visit the Informa Web site at
`www.informa.com
`
`and the Informa Healthcare Web site at
`www.informahealthcare.com
`
`2007023435
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`9
`Rational Choice of Excipients for Use
`in Lyophilized Formulations
`
`Evgenyi Y. Shalaev
`Parenteral Center of Emphasis, Groton Laboratories, Pfi zer Inc.,
`Groton, Connecticut, U.S.A.
`
`Wei Wang
`Pharmaceutical Sciences—Global Biologics, Pfi zer Inc.,
`Chesterfi eld, Missouri, U.S.A.
`
`Larry A. Gatlin
`Parenteral Center of Emphasis, Groton Laboratories, Pfi zer Inc.,
`Groton, Connecticut, U.S.A.
`
`INTRODUCTION
`
`The majority of protein drugs are delivered by the injection route, although there
`is an increasing interest in alternative delivery routes, e.g., pulmonary. Ready-to-
`use liquid formulations are preferred injectable dosage forms because they are
`considered easier to manufacture and administer. However, the majority of pro-
`teins are not suffi ciently stable in aqueous media to provide adequate commercial
`shelf-life and this limits the development of protein pharmaceuticals as ready-
`to-use injectables. Freeze-drying is an established process to increase long-term
`stability of proteins and achieve an acceptable shelf-life (1). In some cases, as
`with proteins intended for administration by inhalation, spray-drying is used (2).
`It is also possible to simply dry protein solutions slowly at ambient temperatures
`under vacuum (3). This chapter deals with freeze-dried protein formulations as
`they are the most common commercial dosage forms. However, general prin-
`ciples can be applied to other dehydration processes such as spray-drying and
`vacuum-drying.
`
`197
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`198
`
`Shalaev et al.
`
`Essentially all protein formulations contain one or more inactive ingredi-
`ents (excipients). Excipients are used to facilitate the formulation manufactur-
`ing process, ensure stability of the active ingredient during processing, storage,
`and administration, minimize adverse effects upon administration (e.g., mini-
`mize pain upon injection), and ensure desirable bioavailability and (for sustained
`release dosage forms) release profi les. Each excipient in the formulation requires
`justifi cation for its use and an appropriate rationale for the level selected. Only
`excipients that are essential for performance and/or stability of a dosage form and
`suitable for injectable products are allowed to be included.
`The majority of lyophile protein dosage forms contain buffer and a bulk-
`ing agent, the latter often playing a dual role for both pharmaceutical elegancy
`and cryo- and lyoprotection, to achieve stability during processing and the shelf-
`life. In addition, many protein formulations contain additional stabilizers, e.g., a
`surfactant, and occasionally an antioxidant or a chelating agent. In some cases,
`a tonicity modifi er, a solubilizer, a processing aid, or an antimicrobial agent may
`be used. It is notable that the active ingredient level in the formulation can range
`from as high as close to 100% to as low as a few parts per million. Therefore, it
`is also possible to have a large range of excipient levels in the fi nal formulation.
`It should be mentioned also that excipients, which are important for the recon-
`stituted solution, e.g., antimicrobial agents or tonicity modifi ers, can be added
`with the diluent rather than being incorporated into the lyophile cake. Generally,
`selection of a proper excipient should take into account (i) the type of product,
`(ii) the delivery route, dose, and administration frequency, (iii) the chemical and
`physical properties of the excipient, (iv) potential interactions with other product
`components, and (v) the container/closure system.
`It is typically advantageous to choose formulation excipients that will not
`only enable the product to meet its critical quality parameters but also facili-
`tate the freeze-drying process because of the high cost/long processing times
`for this unit operation. This is especially critical when developing formulations
`for unique package systems such as dual chamber syringes, because of the dif-
`fi culties encountered in uniform drying in these packages. Therefore, selection of
`excipients that can potentially increase the collapse or eutectic temperatures of the
`frozen solution can greatly facilitate the drying process, thereby reducing cost and
`processing times. It is also important to select excipients whose vapor pressure is
`suffi ciently low so as not to permit its removal during the lyophilization process.
`There are a number of reviews available on different aspects of protein
`freeze-drying (1,4–10). In particular, it has been recognized that understand-
`ing phase behavior is a key for lyophile formulation and process development.
`Therefore, we start with a discussion of phase behavior of excipients during
`manufacturing and storage. Phase transitions have a major impact on stability
`and performance of protein dosage forms. For example, crystallization of a lyo-
`protector may result in protein destabilization during freeze-drying and storage.
`Description of excipients based on their functional role in protein formulations is
`given in the section titled Role and Properties of Excipients, followed by practical
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`Rational Choice of Excipients for Use in Lyophilized Formulations
`
`199
`
`advice on rationale excipient choice (based on both functional and physical chem-
`ical properties of excipients).
`
`PHASE BEHAVIOR OF EXCIPIENTS DURING LYOPHILIZATION
`AND STORAGE: GENERAL CONSIDERATIONS
`
`Phase Transitions During Lyophilization and Storage
`
`During initial cooling of protein formulation solutions, water is normally the fi rst
`component to crystallize. At this stage, a biphasic system is formed, consisting of
`ice and residual freeze-concentrated solution (FCS), which contains protein drug,
`excipients, and remaining water. The composition of the FCS after initial (also
`known as primary) water crystallization depends on the ratio of solutes and the
`temperature, but is independent of total solid content (11). As cooling (and water
`crystallization) proceeds further, the FCS may either remain in the amorphous
`state or partially crystallize, depending on the composition of the system and the
`cooling rate (11) as described below:
`
`1. The FCS may form a kinetically stable (but thermodynamically unsta-
`ble) amorphous phase. A typical example of such behavior is sucrose-
`rich formulations. In this case, solutes do not usually crystallize during
`freezing, drying, and storage, provided that the freeze-dried cake is
`protected from water uptake and the storage temperature is well below
`the glass transition temperature.
`2. The FCS may form a “doubly unstable” (i.e., both thermodynamically
`and kinetically unstable) state. Mannitol- and glycine-based formula-
`tions are typical examples of such behavior. In these systems, second-
`ary excipient + ice crystallizationa would occur either during cooling (if
`the cooling rate is slower than the critical cooling rate) or subsequent
`heating/annealing of the frozen solution (if the cooling rate was higher
`than the critical cooling rate). The critical cooling rate depends on the
`composition of the solution (e.g., glycine/sucrose ratio) and increases
`with an increase in the fraction of a crystallizable component (12). It
`should be noted that crystallization of an excipient is often incomplete,
`i.e., the maximal FCS contains usually all the components (water and
`all the solutes including protein and the excipients), although the rela-
`tive fraction of the partially crystallized excipient remaining in the FCS
`is signifi cantly reduced.
`
`In addition to the common cases described above, it was proposed that a liquid–
`liquid (amorphous–amorphous) phase separation might take place, resulting in
`two amorphous phases of different chemical composition (3,9,13). For example,
`
`a Note also that the secondary solute+water crystallization is referred in the pharmaceutical literature
`as “eutectic” crystallization, although this is not a strictly correct term to apply to a multicomponent
`system.
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`200
`
`Shalaev et al.
`
` protein–excipient amorphous–amorphous phase separation would result in pro-
`tein-rich and excipient-rich amorphous phases. Such phase separation is expected
`to compromise the stabilizing activity of excipients. It should be stressed, however,
`that there is a lack of experimental reports on such demixing behavior between
`protein and amorphous excipient during lyophilization, and it is not clear if this is
`a common behavior for protein formulations.
`Although the phase state of excipients is usually “fi xed” during freezing
`and annealing, further phase transformations may take place during primary and
`secondary drying as well as during shelf storage, depending on the properties and
`the concentration of excipients as well as the storage temperature. For example,
`if an excipient crystallizes as a crystallohydrate (i.e., a crystal with water in the
`crystal lattice) during freezing, the water of hydration might be removed dur-
`ing either primary or secondary drying. Such removal of water of hydration can
`result in either amorphous [e.g., sodium phosphate (14)], or crystalline anhydrous
`(e.g., mannitol) excipient. In addition, an amorphous excipient [e.g., inositol (15)]
`may crystallize during the shelf-life, especially if the water content in the lyocake
`increased because of water transfer either from the stopper or (if the stopper was
`not properly sealed) from the environment.
`
`Signifi cance of Excipient Crystallization
`
`Phase transitions of excipients during manufacturing and storage have a major
`impact on both stability and performance of protein dosage forms. In particular,
`crystallization of either a buffer or lyo- and a cryoprotector is usually undesir-
`able because of the negative impact on protein stability. Indeed, crystallization of
`buffer components is often accompanied by signifi cant changes in the pH of the
`FCS (16), which often causes destabilization of a protein. Also, crystallization
`of a lyoprotector can compromise its protective function. For example, it was
`shown that inositol stabilized a protein when it existed as an amorphous form,
`whereas loss of protein activity was observed when inositol crystallized during
`storage (15). Another example is crystallization of a lyoprotector, raffi nose, dur-
`ing freezing, causing destabilization of lactate dehydrogenase (LDH) (17). The
`negative impact of a lyoprotector (e.g., sugar) crystallization on protein stability
`can be attributed to two different factors: (i) Crystallization results in a physi-
`cal separation of sugar molecules from protein molecules, i.e., an increase in
`intermolecular sugar/protein distance, from several angstroms [which is a typical
`hydrogen-bond length (18)] in molecular mixtures to micrometers in physical
`mixtures of crystalline sugar and amorphous protein; such separation would be
`expected to eliminate any protection imparted by sugars irrespective of the exact
`mechanism (e.g., water substitution vs. the glass transition hypothesis, or ther-
`modynamics vs. kinetics mechanism). (ii) Crystallization of a sugar in an anhy-
`drous form would result in a redistribution of water and in a signifi cant increase
`in the local water content of the remaining amorphous protein-containing phase,
`which could be detrimental to long-term stability. It should be noted, however,
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`Rational Choice of Excipients for Use in Lyophilized Formulations
`
`201
`
`that a sugar crystallohydrate (e.g., a pentahydrate as with raffi nose) might serve
`as a water “scavenger” during crystallization, thus preventing an increase in local
`water content (19,20).
`From the manufacturing process point of view, however, crystallization
`of an excipient may be benefi cial because it allows primary drying at a higher
`temperature without visible collapse, which results in a shorter and more robust
`freeze-drying cycle. In addition, partially crystalline formulations have higher
`drying rates (i.e., shorter cycle) than amorphous formulations of a similar compo-
`sition, possibly because a higher fraction of water is isolated as ice, with ice easier
`to remove than nonfrozen water associated with the amorphous phase (11).
`A compromise between a desire to improve freeze-drying cycle effi ciency
`and robustness (which is achieved by using a crystalline bulking agent) and
`sustaining protein protection (which needs an amorphous lyoprotector) can be
`achieved by using partially crystalline–partially amorphous formulations (21). It
`has been proposed that the crystalline portion provides a physical support even
`at relatively high product temperatures (i.e., higher than the collapse temperature
`of the amorphous phase) whereas the amorphous portion provides lyoprotection
`for protein, allowing aggressive primary drying conditions (21). Feasibility of
`such crystalline–amorphous formulations was demonstrated using glycine–sugar
`formulations with a lyophilization-sensitive enzyme, LDH (22). In this system,
`freeze-drying at a product temperature more than 10°C above the Tg′ resulted
`in a freeze-dried cake without any evidence of macroscopic collapse and with a
`retained enzymatic activity, when the crystalline/amorphous ratio was higher than
`1.2/1 (raffi nose) or 1.6/1 (trehalose) (22). One should be aware, however, that
`timing of crystallization of a crystalline bulking agent, i.e., whether the crystalli-
`zation takes place during cooling or annealing, may infl uence protein stability. An
`example of the signifi cance of crystallization conditions was given in Ref. (11),
`where the stability of a freeze-dried conjugate of immunoglobulin G and horse-
`radish peroxidase in a partially crystalline glycine/sucrose matrix was reported.
`In this case, the activity recovered was signifi cantly higher in the material that
`crystallized during annealing as compared with material in which crystallization
`occurred during cooling. Therefore, although crystalline–amorphous formula-
`tions may be benefi cial, the phase behavior and protein stability need to be inves-
`tigated in each particular case in order to ensure a stable and robust freeze-dried
`product.
`Amorphous–amorphous (liquid–liquid) phase separation of excipients also
`may cause protein destabilization, probably because of creation of an interface
`(1). In addition, it is also possible that a similar amorphous– amorphous demixing
`may occur between protein and lyoprotectant, with expected loss of protection
`(4,9), although there is a lack of experimental data on such protein–excipient
`amorphous–amorphous phase separation. Note that excipient– excipient demix-
`ing in FCS can be studied by differential scanning calorimetry DSC (23), whereas
`no reliable methods to detect such transitions exist for protein–lyoprotectant
` amorphous–amorphous demixing.
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`202
`
`Shalaev et al.
`
`Phase State of Excipients: Analytical Aspects
`
`Excipients can undergo phase transitions during different stages of freeze-drying
`and over the shelf-life of the freeze-dried product. To develop a robust and stable
`formulation, it is essential to understand and monitor such changes. The most
`common methods to detect crystalline–amorphous changes are DSC and X-ray
`diffraction (XRD), whereas other methods such as polarized light microscopy
`(PLM) and different spectroscopic techniques can also be used, depending on the
`formulation properties and stage in the formulation “life” when such transitions
`may occur.
`Excipient crystallization may usually be expected to occur during the freez-
`ing and annealing stages of freeze-drying. Such events are commonly and con-
`veniently studied by DSC. On a DSC curve, crystallization can be detected as a
`second exotherm during cooling (with a fi rst exotherm being water crystalliza-
`tion), an exotherm during heating, and/or an additional endotherm during heating
`preceding the main ice melting peak. If either of these events is observed on DSC
`cooling–heating curves, one may conclude that a component(s) of the formulation
`would likely crystallize during freeze-drying. The reverse statement, however,
`is not always correct, i.e., lack of a crystallization event in a DSC experiment does
`not necessarily mean that crystallization would not occur in vials during freeze-
`drying. Indeed, relatively high scanning rates and small sample volume in a DSC
`study would provide less favorable crystallization conditions as compared with a
`larger sample volume and slower temperature ramping during a real freeze-drying
`run. Low-temperature XRD is another main method used to study crystalliza-
`tion in aqueous solutions (12,24), and is especially well suited both to distinguish
`between crystalline and amorphous structures and to identify the nature of any
`crystalline phase(s) present.
`X-ray powder diffraction (XRPD) is probably the most common and conve-
`nient method to detect crystalline structures in a freeze-dried cake. PLM can also
`be used to confi rm the amorphous nature of a freeze-dried cake. If birefringence is
`observed, it usually means that the cake is at least partially crystalline. However,
`PLM does not usually allow for the identifi cation of the specifi c crystalline phase.
`Note that a sample is often exposed to an ambient atmosphere during both XRPD
`and PLM experiments, which may result in water uptake followed by crystalliza-
`tion. Therefore, precautions should be taken to minimize sample exposure to ambi-
`ent relative humidity (RH) during measurements, to prevent erroneous conclusions.
`As indicated previously, DSC is another common tool to distinguish between
`crystalline and amorphous formulations. Spectroscopic methods such as Fourier
`transform infrared (FTIR) and Raman, and solid state nuclear magnetic resonance
`(NMR) may also be used for structure characterization of lyophile cakes.
`Confi rmation of the amorphous nature of a formulation is essential during the
`formulation development process, e.g., when a formulator needs to choose a buffer
`and/or lyoprotector. For both buffer and lyoprotector, retention of the excipient in
`an amorphous state is desirable, and can serve as a key criterion for the selection
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`Rational Choice of Excipients for Use in Lyophilized Formulations
`
`203
`
`of an excipient. Usually, a combination of evaluation of the pre-lyo solution using
`DSC and evaluation of the fi nished cake using XRPD is suffi cient to conclude if
`any crystallization would occur during freeze-drying. For example, if no crystal-
`lization is observed by DSC and a freeze-dried cake is amorphous by X-ray, one
`could conclude that no crystallization of either buffer or lyoprotector occurred dur-
`ing lyophilization. However, although the solution–DSC/freeze-dried-cake–XRPD
`combination is usually suffi cient to make a reliable conclusion about excipient(s)
`crystallization, it is not always the case. In particular, an erroneous conclusion
`might be made in cases when the following conditions are met: (i) an excipient
`crystallizes as a hydrate; (ii) water of hydration is removed during drying (e.g.,
`secondary drying); (iii) loss of water of hydration causes crystal-to-amorphous
`conversion. Although it might appear that such a combination of events is unlikely,
`it was shown by in situ freeze-drying XRD that this scenario can take place in real
`systems, such as phosphate buffer (14) and the lyoprotectant raffi nose (17). As
`a result of such fi ndings, use of the in situ XRPD method is attracting increased
`attention in the fi eld. In addition, it is important to note a recent improvement
`in XRPD through the use of a high-intensity synchrotron radiation source, which
`provides an excellent signal-to-noise ratio and superior sensitivity as compared to
`the traditional XRPD method. Use of synchrotron XRPD (sXRPD) is especially
`important when one needs to detect crystallization of a low-concentration excipi-
`ent, e.g., buffer, where sensitivity is a major issue. Application of sXRPD in the
`analysis of both freeze-dried cakes and phase transitions in frozen solutions and
`during freeze-drying can be found elsewhere (25,26).
`
`ROLE AND PROPERTIES OF EXCIPIENTS
`
`Buffers
`
`Control of pH is often needed to ensure optimal solubility and stability of a prod-
`uct during manufacturing, storage, and upon reconstitution. In most cases, an
`appropriate amount of buffer is needed to provide adequate buffering capacity.
`Buffer type and concentration, as well as solution pH before lyophilization, are
`important formulation variables. Buffering capacity and the possibility of buffer
`catalysis are the major buffer properties to be considered in the development of
`liquid pharmaceutical formulations (27). There are additional requirements for
`buffers for freeze-drying, i.e., they should be nonvolatile, have a high collapse
`temperature (Tc or Tg′) in the FCS, remain amorphous during freeze-drying, and
`have a high glass transition temperature in the solid state (28). Several buffers
`that are common for parenteral formulations have unfavorable freeze-drying
`properties. For example, acetate—a common buffer—is not a preferred buffer for
`lyophilization because it is volatile and can be partially lost during freeze-drying,
`resulting in a signifi cant pH change. Hydrochloric acid is another example of a
`pH modifi er that is volatile and should be used with caution. In addition, sev-
`eral common buffers have a high tendency to crystallize during freezing. Buffer
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`204
`
`Shalaev et al.
`
`crystallization is usually undesirable because it can lead to substantial pH shifts
`during freezing and therefore could destabilize the protein. In particular, sodium
`phosphate buffer demonstrated signifi cant pH changes (several pH units) during
`freezing, as a result of freeze concentration and crystallization of the buffer com-
`ponents (16). Similarly, tartrate and succinate buffers crystallize readily whereas
`citrate, glycolate, and malate are more resistant to crystallization (29). It should
`be noted that both crystallization and collapse behavior depend on solution pH.
`For example, collapse temperatures of several common buffers are presented in
`Figure 1 as a function of solution pH. The fi gure illustrates two interesting fea-
`tures of the collapse behavior of buffers, i.e., signifi cant changes in the collapse
`temperature with solution pH, and an infl uence of a counter ion (e.g., sodium
`citrate vs. potassium citrate).
`Overall, buffers with a higher collapse temperature (Tc or Tg′) and a lower
`crystallization potential are preferred for lyophilized formulations. However, it
`should be stressed that buffers with a relatively low collapse temperature and
`relatively high crystallization potential (e.g., phosphate buffer) can still be used
`in lyophilized formulations. Both collapse temperature and crystallization ability
`can be modifi ed using other excipients. For example, if a formulation contains a
`
`Figure 1 Collapse temperature (Tg′) as a function of pH. n: sodium citrate; ®: potassium
`citrate; ∆: sodium tartrate; ´: L-histidine. Error bars represent standard deviation. Lines are
`given as a visual help. Source: From Refs. 29, 30.
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`Rational Choice of Excipients for Use in Lyophilized Formulations
`
`205
`
`signifi cant amount of a solute with a high Tc (e.g., a protein), the collapse tem-
`perature of the formulation would increase and buffer crystallization may be sup-
`pressed. However, the amount of an amorphous component required to suppress
`crystallization and/or to provide an acceptable Tc would be lower in the case of
`a buffer with a higher Tc or with a lower crystallization potential. Citrate buf-
`fer appears to be a good choice for lyophilized formulations that are prepared at
`acidic or near neutral pH. Frozen solutions of citrate buffer have a low crystal-
`lization potential, relatively high collapse temperature, and minimal pH changes
`during freezing. In addition, citrate has a reasonably high Tg′ especially when
`prepared from solutions at higher pH values. Glycine may be a reasonable choice
`for the alkaline pH region based on a relatively low crystallization potential of
`sodium glycinate. Other details on freeze-drying properties of different buffers
`can be found in Ref. (28).
`
`Bulking Agents/Lyoprotectors
`
`In many cases, the dose of a drug is quite small, and a bulking agent (or fi ller) is
`needed to provide a matrix to carry the active ingredient. Common bulking agents
`include mannitol, lactose, sucrose, dextran, trehalose, and glycine. In protein for-
`mulations, bulking agents often play a dual role as both bulking agent and lyopro-
`tector. These bulking agents range both in their ability to crystallize (mannitol and
`glycine) or remain amorphous (e.g., sucrose) and in their impact on the formula-
`tion collapse or eutectic temperature. An appropriate choice of a bulking agent
`results in optimal product quality (e.g., physical and chemical stability, reconstitu-
`tion time, moisture levels) and facilitates freeze-drying and scale-up to production
`size. The level of bulking agent utilized will vary depending on the rationale for
`use, e.g., as a matrix-forming agent, as the collapse temperature modifi er, or stabi-
`lizer. As a “rule of thumb,” use of a bulking agent can be considered if the active
`concentration in the fi ll solution is less than 2 wt%.
`Mannitol is the most common bulking agent used in protein freeze-dried
`formulations. Usually, mannitol crystallizes during cooling or annealing of fro-
`zen solutions, which, combined with a high mannitol–ice eutectic temperate
`of −1.5°C (31), allows one to freeze-dry such formulations at a relatively high
`primary drying temperature, without a macroscopic collapse. Therefore, manni-
`tol-based formulations are known to be easy to lyophilize, with a shorter and
`robust lyophilization cycle. There are two potential complications associated with
` mannitol-based formulations. Mannitol forms a crystallohydrate during freeze-
`drying (32,33) that hinders removal of water of hydration and requires elevated
`secondary drying temperature. Vial breakage is another potential problem a for-
`mulator might encounter while working with mannitol-rich formulations (34,35).
`It has been shown that vial breakage is likely associated with volume expansion
`that occurs during warming of a frozen solution at approximately −25°C to −20°C
`(36) probably because of secondary mannitol + water crystallization. Vial break-
`age is affected by mannitol concentration, cooling rate, and fi lling volume (34),
`
`

`

`McNally
`
` PTR
`
` 09/24/07
`
` Chapter 09
`
`206
`
`Shalaev et al.
`
`as well as presence of amorphous solutes and vial confi guration (35). Glycine is
`another common crystalline bulking agent. Because of ionizable groups of gly-
`cine, it may also serve as a buffer. At neutral pH, glycine crystallizes as anhydrous
`beta polymorph, and its eutectic with ice has a relatively high eutectic temperature
`of −3.6°C (37). Both solution pH and counter ions have a signifi cant impact on
`glycine crystallization behavior (38), by infl uencing both ionization of glycine
`in solution and a precipitating solid form. Usually, the presence of other solutes
`(such as a lyoprotector or the protein itself) hinders crystallization of both glycine
`and mannitol. For example, sucrose inhibited the crystallization of mannitol at a
`sucrose/mannitol ratio of 2:1 (39). Inhibition of glycine crystallization by sugars
`(sucrose, trehalose, and raffi nose) was reported in Refs. (12,40,41). For example,
`a critical sucrose/(glycine + sucrose) ratio above which glycine does not crystal-
`lize during either cooling or annealing was reported to be 0.8 (12).
`Sucrose is probably the most popular lyoprotector used in lyophile pro-
`tein formulations. Because of its low collapse temperature, however, lyophilizing
`sucrose-rich formulations can be a challenge. In addition, the presence of other
`components with low collapse temperatures (e.g., buffers) may lower the Tc even
`further. Possible ways to overcome this challenge are reducing the concentration
`of components with low Tc, addition of an excipient with a high collapse tempera-
`ture (e.g., dextran), or addition of a crystalline bulking agent. A critical factor for
`the selection of a lyoprotector is its impact on physical and chemical stability of
`proteins, both during freeze-drying and the product shelf-life. Sucrose is known to
`substantially increase the stability of proteins both during freeze-drying and dur-
`ing subsequent storage. Lactose is another popular bulking agent although there
`is a potential for chemical interaction with amino groups of a protein forming a
`Schiff base, known as the Maillard reaction, or nonenzymatic browning. Recently,
`another disaccharide lyoprotector was introduced with claims of superior stabili-
`zation properties, i.e., trehalose (42). The most obvious advantage of trehalose is
`its higher glass transition temperature as compared to sucrose (120°C vs. 74°C).
`However, side-by-side comparison of stabilization of proteins by sucrose and tre-
`halose revealed that sucrose might provide a comparative or, in some cases, better
`protection than trehalose. For example, sucrose appeared to be more effective in
`stabilizing the native structure of lysozyme during spray