Cross Reference to Related Applications

[0001] This application claims the benefit of U.S. Provisional Application No.

60/647,011, filed January 27, 2005, which is herein incorporated by reference in its entirety.

FIELD OF THE INVENTION

[0002] The invention relates to the lyophilization or freeze-drying of a liquid formulation. More particularly, the invention provides an improved process for lyophilization or freeze-drying a liquid pharmaceutical formulation that includes a protein- based active agent.

BACKGROUND OF THE INVENTION

[0003] The stability and/or potency of many pharmaceutical and food products can be adversely affected during long-term storage. Loss of potency may be attributable to direct chemical degradation, structural and/or physical alteration. Examples of degradative chemical reactions include, but are not limited to, hydrolysis, oxidation, isomerization, deamidation, disulfide scrambling, and racemization. Examples of structural or physical alterations include, but are not limited to, denaturation, aggregation, precipitation and polymerization.

[0004] Lyophilization (also called freeze-drying) refers to a process that uses low temperature and pressure to remove a solvent, typically water, from a liquid formulation by the process of sublimation (i.e., a change in phase from solid to vapor without passing through a liquid phase). Lyophilization helps stabilize pharmaceutical formulations by reducing one or more solvent components to levels that no longer support significant rates of chemical or physical degradation.

[0005] Although freeze-drying can effectively enhance protein stability, freeze-drying can also significantly increase the time and cost of the overall fill-finish process because freeze-drying requires additional handling and processing and requires complex and expensive equipment. In some cases, lyophilization can cost upwards of $100k/day. [0006] One way in which the cost of freeze-drying can be reduced is to reduce the processing time, for example, by exposing the formulation to as high a temperature as

possible during primary drying. For example, a 5°C increase in temperature can decrease the drying time by a factor of two. However, excessive temperature can result in a final product with undesirable qualities, for example, cake collapse. Conventionally, it is thought that the primary drying process should be performed at a temperature below the glass transition temperature (Tg') or collapse onset temperature (To) of the formulation to prevent collapse.

[0007] Consequently, one way in which the processing time of the freeze-drying process has been reduced is by increasing the Tg' of the formulation. As the Tg' is increased, the formulation can be dried at a higher temperature, thereby decreasing the processing time. Conventionally, collapse temperature is increased by including a crystalline bulking agent, such as mannitol or glycine, in the formulation. However, it is not always desirable to include a crystalline bulking agent in a formulation, especially when a formulation already includes high concentrations of other excipients. In other instances, the presence of a disaccharide stabilizing agent in a formulation may interfere with crystallization of the crystalline bulking agent. For example, in an isotonic formulation that includes a concentration of stabilizing agent, for example, a mono-, di- or poly- saccharide stabilizing agent, that is greater than about 3% w/v, the addition of a crystalline bulking agent such as glycine or mannitol may not be effective to increase the Tg' or the collapse temperature of the formulation. However, in some cases the amount of stabilizing agent in the formulation needs to be at least 3% w/v to provide protein stability. Therefore, there still remains a need for a cost-effective freeze-drying process, particularly for formulations that lack a crystalline bulking agent or contain a limited amount of crystalline bulking agent.

SUMMARY OF THE INVENTION

[0008] One aspect of the invention relates to the discovery that a pharmaceutical formulation can be lyophilized at a temperature during the primary drying stage that is above the glass transition temperature (Tg ¹ ) of the formulation, but below the collapse temperature (Tc), while retaining desirable properties of the final freeze-dried product, including, but not limited to visual appearance, stability, and biological activity. Another aspect of the invention relates to the discovery that a pharmaceutical formulation can be lyophilized at a temperature during the primary drying stage that is above the collapse onset temperature (To) but below the collapse temperature.

[0009] Another aspect of the invention relates to the discovery that increasing the protein concentration of a pharmaceutical formulation increases the difference between Tg' and Tc. Therefore, another aspect of the invention provides a liquid pharmaceutical formulation suitable for freeze-drying to form a lyophilized formulation, wherein the liquid formulation includes a protein concentration of at least about 20 mg/mL, 40 mg/mL, 50 mg/mL, 60 mg/mL, 70 mg/mL, 80 mg/mL, 90 mg/mL, 100 mg/mL, 125 mg/mL, 150 mg/mL, or higher.

[0010] In yet another aspect of the invention, the liquid formulation includes little or no crystalline bulking agent (i.e., the bulking agent is present in a minor amount, such that the bulking agent is unable to form a crystalline supporting structural matrix during lyophilization). In one embodiment, the liquid formulation includes crystalline bulking agent and amorphous solute at a weight:weight ratio of less than 1.0, 0.9, 0.75, or 0.50. [0011] In yet another aspect of the invention, the liquid formulation includes at least about 2%, 3%, 4%, 5%, 10%, or 15% w/v stabilizing agent, such as a disaccharide stabilizing agent, including, but not limited to sucrose or trehalose.

BRIEF DESCRIPTION OF THE FIGURES

[0012] FIGURE 1 is a graph showing the change in heat capacity (ΔCp) at the glass transition temperature (Tg).

[0013] FIGURE 2 is a graph showing the exothermic transition observed when a crystalline component melts.

[0014] FIGURE 3 shows the Tg' of a sample monoclonal antibody at varying concentrations, as determined by differential scanning calorimetry (DSC). [0015] FIGURE 4 shows the collapse temperature compared to the Tg' of a sample monoclonal antibody at varying concentrations.

[0016] FIGURE 5 shows a FTIR Spectrum comparing Aggressive and Conservative Cycles.

[0017] FIGURE 6 shows a FTIR FDP-minus-BDS Difference Spectra for monoclonal antibody cycles.

[0018] FIGURE 7 shows a lyophilization cycle for a sample monoclonal antibody at 80 mg/mL.

[0019] FIGURE 8A-D shows a Freeze-Dried Microscopy image (10x magnification) of the monoclonal antibody (80 mg/mL) showing the change in microstructure as

temperature is increased over time at a rate of 1°C per minute under vacuum: (A) shows the microstructure at Temperature = -24°C; (B) Temperature = -19°C; (C) Temperature = -

12°C; and (D) Temperature = -TC.

[0020] Figure 9 is a photograph showing cakes of freeze-dried BSA (0 mg/mL - 100 mg/mL) and sorbitol containing formulations.

[0021] FIGURE 10 is an illustration of the structural support matrix provided by a crystalline bulking agent during primary drying.

DETAILED DESCRIPTION OF THE INVENTION

I. Lyophilization in General

[0022] Lyophilization (also called freeze-drying) refers to a process that uses low temperature and pressure to remove a solvent, typically water, from a liquid formulation by the process of sublimation (i.e., a change in phase from solid to vapor without passing through a liquid phase). Lyophilization helps stabilize pharmaceutical formulations by reducing one or more solvent components to levels that no longer support chemical reactions or biological growth.

[0023] Freeze-drying processes are known. In some instances, freeze-drying is performed in a "manifold" process in which flasks, ampules or vials are individually attached to the ports of a manifold or drying chamber. In other instances, freeze-drying is performed as a "batch" process in which one or more similar sized vessels containing like products are placed together in a tray dryer, hi a "bulk" process, the product is poured into a bulk pan and dried as a single unit. The product is removed from the freeze drying chamber prior to closure and then packaged in air-tight containers. The invention described herein can be used in combination with any of these processes. [0024] Generally, lyophilization takes place in at least three stages: freezing; primary drying; and secondary drying, hi some instances, it may be desirable to include an annealing step between the freezing and primary drying stages. Each of these stages will be discussed in more detail below.

Freezing

[0025] During the freezing process, a liquid solution or formulation that contains at least one solvent and at least one solute is placed in a container, which is then placed in a

freeze-dryer. The primary goal of the freezing process is to solidify at least the solvent component of the formulation.

[0026] During the freezing process, the microstracture of both the solvent crystals and the solute is formed. This microstrucrure can affect both the quality of the final product and its processing characteristics, such as the rates of primary and secondary drying. If both the solute and the solvent crystallize during the freezing process, the temperature at which the formulation becomes solid is called the eutectic temperature (Te). A formulation in which both the solute and solvent crystallize during the freezing process is referred to herein as "a crystalline system."

[0027] A formulation in which at least some of the solute remains in an amorphous state is referred to herein as "an amorphous system." One example of an amorphous system is a liquid formulation that contains protein as an active agent. [0028] If some or all of the solute remains substantially amorphous during the freezing process (e.g., more than 50% of the solute remains substantially amorphous), the temperature at which the solute becomes a glassy or amorphous solid is called the glass transition temperature of the maximally freeze concentrated solution (Tg')- [0029] As the temperature of an "amorphous system" is reduced, the solvent component forms crystals (referred to as "the crystalline component"). The crystalline component may also contain crystalline excipients, for example crystalline bulking agents such as mannitol or glycine. The concentration of the solute that remains amorphous (herein referred to as the "amorphous component") increases as the temperature of the formulation is decreased and the solvent crystallizes out of solution. Typically, the amorphous component includes the amorphous active agent, for example, a protein, and any amorphous excipients, for example, a disaccharide stabilizing agent. It is worthwhile to note that a component of a formulation, such as a bulking agent, can exist in a crystalline or an amorphous state depending upon the formulation and/or the processing parameters.

[0030] As used herein, a "frozen" amorphous system contains a crystalline component, which can include the crystalline solvent and crystalline excipients; and an amorphous component located within the interstitial regions of the crystalline component. The amorphous component can include the amorphous active agent; one or more amorphous excipients; and any remaining unfrozen solvent.

[0031] During the freezing process solvent molecules may spontaneously aggregate to

form a template to which other solvent molecules can attach and ultimately form a crystal. This process is referred to as "nucleation." The temperature at which nucleation occurs ("the nucleation temperature") can affect primary drying rate and morphology. As the temperature decreases, the probability of nucleation temporarily increases. However, as the temperature is decreased further, nucleation tends to decrease due to the increased viscosity of the system. The nucleation temperature can be affected by environmental particulates, freezing method and the presence or absence of nucleating agents. As used herein, the term "heterogeneous" nucleation refers to nucleation that was initiated by foreign particles (also called nucleation sites) in the solution or on the surface of the container in which the solution was placed. The term "homogenous nucleation" refers to nucleation that occurs in the absence of a nucleation site in the solution. Generally, homogeneous nucleation is caused by the aggregation of slow moving molecules. The nucleation observed in pharmaceutical solutions is largely heterogeneous. [0032] The rate and method of cooling can influence the structure and appearance of the matrix and final product. For example, if the solution is frozen quickly, the crystals will tend to be small. This may result in a fine pore structure in the product and a corresponding higher resistance to flow of water vapor during primary drying which may lead to a longer primary drying time. However, small crystal structures may be desirable, for example, to preserve structures for microscopic examination. If the solution is frozen at a slower rate, the crystals will tend to grow from the cooling surface and may be larger. Consequently, the product may have a coarse pore structure, resulting in less restrictive channels in the matrix and perhaps a shorter primary drying time.

[0033] Supercooling is another physical event that is observed during the freezing process (in addition to solvent crystallization; concentration of solutes; and in some cases, crystallization of one or more solutes). The term "supercooling" refers to the reduction of the temperature of a liquid beyond its freezing point. As a result of supercooling, the product temperature may have to be decreased significantly below the actual freezing point of the solution before freezing occurs.

[0034] Although the temperature at which the material is frozen depends on many factors, including the formulation, the freezing process is typically performed at a temperature below at least about 0°C, below at least about -10 ⁰ C, below at least about - 25°C, below at least about -40 ⁰ C, or below at least about -50°C. Typically the freezing process is performed at atmospheric pressure. The rate at which the temperature of the

formulation is reduced is referred to herein as the "freezing rate." Although flash freezing processes are known and can be used in connection with the invention, the freezing rate for a formulation is more typically between about 0.05 ⁰ C to about 5.0°C per minute, about 0.1 ⁰ C to about LO ⁰ C per minute, or about 0.3°C to about 0.4°C per minute in production scale freeze-dry cycles.

Annealing

[0035] Some freeze-dry cycles contain an annealing or thermal treatment step following freezing. In an annealing step, the temperature of the frozen formulation is increased temporarily and then, after a specified period of time, the temperature is again lowered.

[0036] Annealing may be used for different purposes. An annealing step may be included to control the nucleation temperature of the formulation, which may affect the primary drying rate and resulting cake morphology. For example, annealing can increase the primary drying rate of a frozen formulation between about 1 and 5 fold. [0037] An annealing process generally results in the removal of solvent crystals smaller than a critical size and generation of larger solvent crystals. The increased crystal size may result in an increased primary drying rate because pores in the material left by large crystals generally provide less resistance to primary drying than smaller pores left by smaller crystals. Annealing is particularly beneficial for crystallizing excipients and bulking agents present in the formulation. In particular, annealing promotes crystallization of crystalline bulking agents such as mannitol or glycine.

[0038] In general, an annealing process involves maintaining a sample below its freezing point, but above the Tg' of the material, for a predetermined period of time, for example, to increase the rate of the annealing process. Because the solvent is already substantially frozen prior to the annealing process, and because, unlike in primary drying, the solvent is not being removed, the annealing process can be performed at a relatively high temperature without adversely affecting the quality of the final product. [0039] Although the processing parameters of the annealing step depend on the formulation, the annealing step is generally performed at a temperature between about - 5O ⁰ C and about 0 ⁰ C, about -40°C and about -5°C, about -30 ⁰ C and about -5°C, about - 25 ⁰ C and about -10 ⁰ C, or about -20 ⁰ C and about -10 ⁰ C, for a time ranging between a few minutes to a few days, typically between about 15 minutes and about 24 hours, or between

about 15 minutes and about 10 hours.

[0040] Because a crystalline bulking agent is not required in formulations that are freeze-dried according to the process of the invention, an annealing step may not be necessary. However, in some cases it may still be desirable to include an annealing step.

Primary Drying

[0041] During the primary drying process, solvent is removed from the formulation by a process of sublimation. As used herein, the term "sublimation" refers to the transition of a solid to a gas, without passing through a liquid stage.

[0042] One goal of the primary drying process is to remove all of the "mobile" solvent from the formulation. Although a majority (i.e., at least about 50 wt%) of the solvent is removed from the formulation during primary drying, the formulation that remains at the end of the primary drying process includes an amorphous component within a glassy matrix that contains between about 5 wt% and about 49 wt% solvent, more typically between about 10 wt% and about 40 wt% solvent, or between about 10 wt% and about 20 wt% solvent. Although the composition of the liquid formulation can affect the cycle time, as well as the desired amount of solvent in the final product, primary drying can take as little as 1 hour, more typically between about 5 hours and about 10 days, or between about 1 day and about 4 days.

[0043] In general, the primary drying process is performed at a reduced pressure (i.e., vacuum) and at a temperature higher than the temperature at which the system was frozen. The increase in temperature provides energy for sublimation. However, it is important to control the drying rate and the heating rate during primary drying. If the drying proceeds too rapidly, the dried product can be carried out of the container by escaping solvent vapor. If the temperature of the system is raised too high, the microstructure of the amorphous and/or crystalline component may melt, which may cause the structure of the cake to collapse. On the other hand, primary drying at lower temperatures tends to increase cycle length, which usually results in a more expensive process.

Temperature

[0044] The term "temperature" can refer to the "shelf temperature" or the "product temperature." As used herein, the term "shelf temperature" refers to the temperature of the lyophilization equipment. The "product temperature" refers to the actual temperature of

the formulation. Although the shelf temperature influences the product temperature, the shelf temperature and the product temperature can differ for a variety of reasons. For example, because sublimation is an endothermic process, as the solvent sublimes the remaining amorphous component of the formulation tends to cool. Consequently, throughout primary drying, the amorphous solid tends to remain colder than the shelf temperature. At the end of primary drying, when the mobile solvent has been removed by sublimation and the heat of sublimation is no longer needed, the temperature of the amorphous component of the formulation tends to increase sharply toward the shelf temperature. Maximum product primary drying temperature is defined as the maximum temperature that the product encounters before primary drying is complete. It may be measured with product thermocouples as a point of inflection in the thermocouple trace when ice sublimation is complete and the product temperature begins to rise. The asymptotic rise in temperature may be used to approximate the endpoint of primary drying. Shelf temperature is used as the primary control of product temperature during primary drying and it typically ranges from about — 30°C to about 1O ⁰ C for protein formulations although the desired product temperature depends on the formulation. [0045] Important temperature-related properties of a formulation include the "eutectic temperature" (Te), the "glass transition temperature" (Tg), the "collapse temperature" (Tc), the "onset temperature" (To), and the melting temperature (Tm). As used herein, the term "Tm" refers to the melting temperature of the solvent itself (e.g., ice).

Eutectic Temperature (Te)

[0046] The eutectic temperature refers to the temperature in a freezing process at which both the solute and solvent present in a crystalline system become frozen or crystalline or the temperature at which the solute has been freeze-concentrated to saturation. As used herein, the term "crystalline" or "crystal" refers to a solid in which the constituent atoms, molecules or ions are packed in a regularly ordered, repeating pattern extending in all three spatial dimensions. Under some conditions, the solid may include a single crystal, where all of the atoms in the solid fit into the same lattice or crystal structure. However, it is more typical that many crystals form simultaneously during solidification, leading to a polycrystalline solid. As used herein, the term "crystalline system" refers to a mixture in which all components, i.e., both the solvent and solute form crystals. Conventionally, the Te has been used as a guide for determining the maximum

product temperature for a crystalline system during primary drying. It is generally thought that the desirable properties of a freeze-dried product may be lost if the product temperature exceeds the eutectic temperature while the frozen solvent is still present because drying will take place from the liquid state instead of the solid state.

Glass Transition Temperature (Tg)

[0047] For solutions in which the solute does not readily crystallize during freezing (i.e., "amorphous systems"), the temperature at which the viscosity of the amorphous component changes from a viscous liquid to a glass is called the "glass transition temperature" or "Tg". In freeze-drying terminology, the glass transition temperature of the maximally freeze-concentrated solution (prior to drying) is referred to as Tg', whereas the glass transition temperature of the dried product is referred to as Tg (i.e., after primary drying). More specifically, as the temperature of an amorphous system is decreased, a critical concentration is achieved at which point the unfrozen fraction exhibits a reduced molecular mobility, and its physical state changes from an elastic liquid to a brittle but amorphous solid glass. The unfrozen fraction is referred to herein as "the amorphous component." The amorphous components are a solid, but unlike a crystalline solid, there is no distinct order of the positions of the atoms, molecules or ions. An example of an amorphous system is a protein-based formulation. In protein-based pharmaceutical formulations, the amorphous component contains at least the protein active agent. The amorphous component may also contain one or more excipients. Conventionally it is thought that the glass transition temperature represents the maximum allowable temperature during the primary drying of the amorphous system, unless there is a substantial amount of crystalline bulking agent also present to support the cake. [0048] The glass transition temperature can be defined empirically as the temperature at which the viscosity of the liquid exceeds a certain value, for example, 10 ¹³ Pascal seconds. Alternatively, the glass transition temperature can be determined experimentally, for example, by Differential Scanning Calorimetry (DSC). DSC is a technique used to study thermal transitions of a system. As used herein, the term "thermal transition" refers to a change in state, for example, the change from a solid to a liquid, as a consequence of a change in temperature. The heat capacity, or C _p is the amount of heat required to raise the temperature of a specified unit mass of substance by one degree Celsius. At the glass transition temperature, the heat capacity of the amorphous component of the frozen

formulation changes. Generally, as the temperature of the amorphous component of the formulation is increased from below the glass transition temperature to above the glass transition temperature, the heat capacity increases (See Figure 1). Typically, the glass transition temperature for an amorphous system is not a single temperature point, but rather a range of temperatures, usually within a range of 1°C to 2 ⁰ C. Li many instances, the viscosity of the amorphous component changes by three or four orders of magnitude over a temperature range of a few degrees at temperatures around the glass transition temperature.

[0049] Other methods by which the glass transition temperature can be determined are known to those of skill in the art. Suitable examples include, but are not limited to Differential Thermal Analysis (DTA), in which the transition from solid to liquid is determined by detecting a temperature difference between the sample and a reference, and Electrical Resistance (ER) measurements, in which the phase change from solid to liquid is determined by measuring the change in the relative conductivity of an electrical current through a sample.

Collapse Temperature (Tc)

[0050] In general, the collapse temperature refers to the temperature at which an increase in mobility or viscous flow of the amorphous component within the interstitial regions of the crystalline component is observed. If the collapse temperature is exceeded during primary drying, the microscopic structure of the amorphous component of the formulation that was formed during freezing and sublimation may be lost. [0051] The collapse temperature is a function of all components present in the formulation and can therefore vary depending on the formulation. The collapse temperature may also be affected by the measurement method and the amount of residual unfrozen solvent contained in the amorphous component. Additionally, within a given system or formulation, as solvent is reduced via sublimation, the collapse temperature tends to increase.

[0052] Although the collapse temperature is often equated with the glass transition temperature, they are not actually equivalent. The glass transition temperature is measured in a closed system of constant composition, whereas collapse is a dynamic process that can occur during the drying process. Conventionally it is thought that, Tg' and Tc for a particular formulation may differ by a few degrees (i.e., between about 2 ⁰ C to

about 3°C). However, as discussed below, according to invention, increasing the protein concentration in the liquid formulation can substantially increase the difference between Tg' and Tc. In fact, the difference between Tg' and Tc can be as much as 1O ⁰ C, 15°C, or 20°C, for example, in a formulation containing disaccharide as the principle lyo-/cryo- protectant. The difference may even be higher for other formulations, for example, those containing sorbitol as the principle lyo-/cryo-protectant.

[0053] Typically, collapse is associated with a decreased surface area of the freeze- dried formulation, reduction in cake volume, and/or loss of pharmaceutical elegance. The term "collapse" may refer to collapse of the lyophilized product as observed by freeze- dried microscopy in which the initial structural appearance after freezing the product is lost at the drying front during warming of the microscope stage. The term "collapse" may also refer to collapse of the lyophilized product (or cake) during lyophilization in a freeze dryer (e.g., lab-scale or production-scale). The collapse temperature as determined by freeze-dry microscopy is thought to be the same or similar (within a few degrees) to that of the collapse temperature observed within the freeze-dryer. As used herein, the term "collapse" in the context of lyophilization in a freeze-dryer refers to a reduction of cake volume of at least about 10%, 25%, 50%, 75%, 85%, 95% or 100%. The reduction in cake volume can be measured using known methodologies, including but not limited to, visual inspection or BET surface area analysis. Collapse may also be associated with a glossy or glassy appearance of the cake and/or an increase in reconstitution time. The loss of crystal structure can be observed using differential scanning calorimetry (DSC). The "melting" or collapse of the crystal structure is an endothermic transition that can be observed by a "dip" in a plot of heat flow versus temperature (See Figure 2). Freeze-dried microscopy can be used to visually observe structural changes of the formulation as the temperature is increased. Generally, during primary drying a receding boundary can be observed as the frozen layer decreases in thickness and the thickness of the partially dried solids increases (See Figure 8A-D).

[0054] In some cases, the collapse of a pharmaceutical product can be merely an aesthetic problem. In other cases, collapse can result in rejection of the product. Generally, when a cake collapses, solvent may become trapped within the cake and may not be removed during secondary drying. This is generally undesirable because the additional residual solvent or moisture may reduce the stability of the final freeze-dried product. Collapse may also make the final product visually unappealing and harder to

reconstitute. Difficulty in reconstituting the drug at the user stage may result in partial drug loss to the patient since many reconstituted lyophilized products are pre-filtered before administration and drug product that is not dissolved may become trapped in the filter.

[0055] The desirable properties of a freeze-dried formulation are related to the microstructure formed during the freezing process. It is therefore desirable to maintain the microstructure of the formulation throughout the drying processes. In an amorphous system, this means that the amorphous component should be sufficiently rigid to support its own weight without the structural support of the crystalline component (i.e., the crystallized solvent with or without a crystalline bulking agent). Conventionally, it is thought that the temperature of an amorphous system should be held below the glass transition temperature (Tg') of the formulation or below the collapse onset temperature during primary drying such that the viscosity of the amorphous component is high enough to retain the structure of the amorphous component throughout the drying process. [0056] However, although it remains important to keep product temperature below the collapse temperature (Tc) during primary drying, Applicants have found that primary drying can be performed above the glass transition temperature (Tg ¹ ) and above the collapse onset temperature (To), at a point where the system shows an "initial structural change" in freeze-dry microscopy.

Collapse Onset Temperature (To)

[0057] As used herein, the term "collapse onset temperature" (To) refers to the temperature at which an initial structural change in the microstructure of the freeze-dried product is detected, for example, by freeze-dried microscopy. To may occur at a temperature around or above the Tg' but below the Tc. Conventionally, the maximum product temperature targeted during primary drying is below To-

[0058] Freeze-dried microscopy can be used to observe structural changes of the formulation as the temperature is increased. As mentioned above, during primary drying, a receding boundary can be observed as the frozen layer decreases in thickness and the thickness of the partially dried solid increases. Figure 8 shows a sequence in time (Panels A-D) in which a monoclonal antibody formulation (at a concentration of 80 mg/mL) is freeze-dried while gradually increasing the temperature. Frozen material is on the right side of each panel and dried material is on the left. When panels A-D are viewed in

sequence, the drying front is gradually moving from left to right. As shown in Figure 8 A- D, the "initial structural change" in the microstructure of the freeze-dried product typically appears as a change in the initial appearance. The precise change in appearance that occurs depends on many factors, for example, the components of the formulation. A person experienced with freeze-dry microscopy, particularly with respect to observing the changes that occur during lyophilization, is able to detect and interpret such changes in appearance. As the temperature is increased, the appearance gradually changes further, until a distinctly different appearance is evident, and the original structure is gone (i.e., collapse).

Pressure

[0059] Another important parameter during the primary drying process is pressure. Pressure can affect the drying rate during primary drying. Generally, the rate of sublimation from a frozen solid depends upon the difference in vapor pressure of the solid compared to the vapor pressure of the chamber. The pressure differential is important because molecules migrate from the high-pressure crystalline component to the low pressure chamber. Thus, the primary drying process is typically performed at a reduced pressure (i.e., under a vacuum). More typically, the chamber pressure is reduced to a pressure that is below the vapor pressure of the solvent component. Because the vapor pressure of the solvent component can vary depending on the formulation, the pressure at which the primary drying process is performed can vary, but is typically _. between about 40 mTorr and about 400 mTorr, or between about 50 mTorr and about 250 mTorr. Typical chamber pressures used in primary drying range from about 75 mTorr to about 225 mTorr. As a general rule of thumb, chamber pressure is usually set between about % and about ¹ A of the vapor pressure of the solvent at the desired product temperature. [0060] Generally, at low pressures, the main form of heat transfer is conduction from the shelf through the bottom of the product container. Since the product containers are typically glass and glass can act as an insulator, conduction is not very efficient and drying can be slow. Therefore, it may be desirable to improve the heat transfer mechanism by introducing an inert gas into the drying chamber at a controlled rate. Suitable inert gasses are known and include, for example, nitrogen. The presence of the inert gas molecules facilitates heating of the walls of the container in addition to conduction through the bottom of the container, thereby increasing the amount of heat being supplied to the

product per unit time. This enhances the drying rate, reduces the cycle time and reduces energy and labor costs associated with a lengthy process.

Primary Drying above Tz' or above the collapse onset temperature [0061] One aspect of the invention relates to the discovery that a pharmaceutical formulation can be lyophilized at a temperature during the primary drying stage that is above the glass transition temperature (Tg') of the formulation, but below the collapse temperature (Tc), while retaining desirable properties of the final freeze-dried product, including, but not limited to visual appearance, stability, and biological activity. In general, retention of such desirable properties is related to retention of the microscopic structure of the formulation that is formed during primary drying, for example, good cake structure that has relatively good porosity and retains pharmaceutical elegance, which allows for efficient secondary drying and low residual moisture. Another aspect of the invention relates to the discovery that a pharmaceutical formulation can be lyophilized at a temperature during the primary drying stage that is at or above the collapse onset temperature (To), but below the collapse temperature (Tc).

[0062] Another aspect of the invention relates to the discovery that increasing the protein concentration of a pharmaceutical formulation increases the difference between Tg' and Tc, and the difference between To and Tc. At lower protein concentrations, i.e., less than about 10 mg/ml, the difference between Tg' and Tc is generally in the range of about 0°C to about 2°C. However, when the protein concentration is increased to about 20 mg/mL, the difference between Tg' and Tc may increase to within the range of about 2°C to about 10°C. At higher protein concentrations, i.e., above about 80 mg/mL, the difference between Tg' and Tc can be greater than about 10°C, for example, between about 10°C to about 2O ⁰ C. While not wanting to be bound by theory, it is believed that a high protein concentration results in an amorphous solid phase in which the protein molecules are densely packed and therefore have reduced freedom of movement. Thus, the increased density results in an amorphous solid phase with more structural integrity, higher viscosity, and decreased mobility. This allows for drying above Tg' and To without any appreciable crystalline bulking agent present. Therefore, another aspect of the invention provides a liquid pharmaceutical formulation suitable for freeze-drying to form a lyophilized formulation, wherein the liquid formulation includes a protein concentration of at least about 20 mg/mL, 40 mg/mL, 50 mg/mL, 60 mg/mL, 70 mg/mL, 80 mg/mL, 90

mg/mL, lOO mg/mL, 125 mg/mL, 150 mg/mL, or higher.

[0063] In yet another aspect of the invention, the liquid formulation includes little or no crystalline bulking agent (i.e., the crystalline bulking agent is present in a minor amount such that the crystalline bulking agent does not form a supporting structural matrix during lyophilization). In one embodiment, the liquid formulation includes crystalline bulking agent and amorphous solute at a weightrweight ratio of less than 1.0, 0.9, 0.75, or 0.50.

[0064] In yet another aspect of the invention, the liquid formulation includes at least about 2%, 3%, 4%, 5%, 10%, 15% w/v stabilizing agent, such as a mono-, di-, or polysaccharide stabilizing agent, including, but not limited to sucrose, trehalose, dextran, and glucose. Other stabilizing agents may include sugar alcohols and amino acids, providing that they are retained in an amorphous phase after freeze-drying, including, but not limited to mannitol, sorbitol, and glycine.

Secondary Drying

[0065] Secondary drying includes the removal of non-frozen water that is bound or adsorbed to the amorphous solid matrix (i.e., the protein and other solutes in the formulation). This is mainly a diffusion-based process that is primarily driven by product temperature and not by chamber pressure. Therefore, it is typically performed using a shelf temperature significantly higher than that used during primary drying. [0066] The goal of the secondary drying process is to obtain a porous "freeze-dried formulation," also referred to as a "cake" with a level of residual moisture that no longer supports chemical or physical degradation. As used herein, the term "freeze-dried formulation" or "cake" refers to the dried formulation that remains after the solvent has been removed by the process of lyophilization. The freeze-dried formulation typically includes an amorphous solid matrix and a minor amount of unfrozen solvent. The amorphous solid matrix contains the active agent and excipients. It is worthwhile to note that the amorphous solid matrix can include both crystalline and amorphous excipients. [0067] Because many proteins require a solvent such as water to maintain proper secondary and tertiary structure, it may not be desirable to remove all of the solvent from the cake. Generally, at the end of the secondary drying process, the freeze-dried formulation has a moisture content below about 5 wt%, typically between about 0 wt% and about 3 wt%. Residual moisture of less than 1% in a protein formulation is generally

considered low.

[0068] Since there is very little mobile solvent in the formulation at the end of the primary drying stage, the shelf temperature may be increased during secondary drying without altering the structure of the resulting cake (i.e., causing melting). Additionally, the solvent remaining during secondary drying is typically more strongly bound to the amorphous solid matrix, and consequently may require more energy for its removal. Thus, during the secondary drying process, the shelf temperature is typically increased and the chamber pressure is decreased.

[0069] Generally, a higher temperature results in a faster drying rate. Additionally, as water leaves the solid amorphous matrix during secondary drying, Tg generally increases. This may allow the temperature to be increased further. However, it is still important that the formulation remain below a certain temperature during the secondary drying process (a temperature that is too high may result in collapse of the product). Typically a secondary drying process is performed at a temperature between about — 20°C and about 50°C, or between about 0 ⁰ C and about 40 ⁰ C, or between about 10°C and about 40°C. Typical secondary drying temperatures for protein formulations range between about 20°C and about 4O ⁰ C. Generally, the heating rate (i.e., the rate at which the temperature is raised from the primary drying temperature to the secondary drying temperature) is similar to the freezing rate. Consequently, the heating rate (determined by the change in shelf temperature) from primary drying to secondary drying is between about 0.05°C to about 5.0 ⁰ C, between about 0.1°C to about 1.0°C, or between about 0.3 ⁰ C to about 0.4 ⁰ C per minute. Although most freeze dryers, when occupied with product, are capable of a faster heating rate than freezing rate (generally due to the reduction in amount of solvent present in the dried product as compared to the liquid product), it is not always advantageous to have the most rapid heating rate. For example, a low heating rate may allow some secondary drying to occur prior to reaching the secondary drying shelf temperature. As the product dries, the product Tg gradually increases. Therefore, a reduced heating rate may reduce the likelihood that the product exceeds Tg or collapse temperature during the heating process. Typically, the secondary drying process is performed under a vacuum, at about the same pressure as the primary drying process, typically between about 40 mTorr and about 400 mTorr, or between about 50 mTorr and about 250 mTorr.

II. Formulations

[0070] As described above, lyophilization is a process in which a liquid formulation is subjected to a freeze-dry process to obtain a freeze-dried formulation. The contents of a freeze-dried formulation may vary depending upon the active agent and the intended route of administration. The liquid formulation generally includes a solvent and solute. The solute typically includes an active agent and, optionally, one or more excipients. The resulting freeze-dried formulation includes an amorphous solid matrix and a minor amount of residual unfrozen solvent. The amorphous solid matrix includes the active agent and, optionally, one or more excipients.

[0071J In general, any component in the formulation that is not the solvent or the active agent is referred to as an "excipient." "Excipients" are included in a formulation for many reasons, although the primary function of many excipients is to provide a stable liquid environment for the active ingredient or to protect the active agent during the freezing or drying process. Some excipients may be used to achieve multiple effects in a formulation. For example, a disaccharide such as sucrose may act as a cryoprotectant, lyoprotectant, bulking agent and tonicity modifier. Behavior of an excipient may change when in the presence of other excipients. Some combinations have a positive synergistic effect, others have a negative synergistic effect. Positive synergy occurs when the sum of the effects of excipients acting together is greater than the additive effects of the individual excipients. Negative synergy occurs when the sum of effects of the combination of excipients is less than that of the individual excipients. Examples of active agents, solvents and excipients are provided below.

Active Agent

[0072] As used herein, the term "pharmaceutical formulation" refers to both formulations that include active agents that are small molecule therapeutics and formulations that include a biopharmaceutical as an active agent. As used herein, the term "small molecule therapeutics" refers to natural and synthetic substances that typically have a low molecular weight (i.e., less than about 1000 Daltons). Small molecules can be isolated from natural sources such as plants, fungi or microbes, or they can be synthesized by organic chemistry. Many conventional pharmaceuticals, such as aspirin, penicillin, and chemotherapeutics, are small molecules. The term "biopharmaceutical" refers to formulations containing active agents that generally have a high molecular weight (i.e., at least about 1000 Daltons). Examples of such "high molecular weight" active agents

include carbohydrates, polypeptides, and proteins.

[0073] The term "polypeptide" or "protein" as used herein can refer to both antibody and non-antibody proteins. Non-antibody proteins include, but are not limited to, proteins such as secreted proteins, enzymes, receptors, and fragments or variants thereof. The polypeptide may or may not glycosylated. The term "polypeptide" or "protein" may also include multimeric proteins, such as hetero- or homo-dimers, trimers, etc. The term "antibodies" can include both monoclonal and polyclonal antibodies, antibody fragments, chimeric antibodies, human or humanized antibodies. Antibody fragments are known and include, but are not limited to, single chain antibodies, such as ScFv, Fab fragments, Fab ¹ or F(ab') ₂ fragments, etc. Although antibodies tend to have a higher molecular weight than non-antibody proteins, the discoveries of the invention (e.g., those relating to drying above the Tg', especially for formulations with increased protein concentration as described herein, formulations containing little or no crystalline bulking agent and/or formulations containing at least about 2 % w/v disaccharide stabilizing agent) can be applied to both.

Solvent

[0074] As discussed previously, lyophilization is the process by which solvent is removed from a liquid formulation. As used herein, the term "solvent" refers to the liquid component of a formulation that is capable of dissolving or suspending one or more solutes. The term "solvent" can refer to a single solvent or a mixture of solvents. A commonly used solvent for pharmaceutical formulations is water for injection (WFI). Depending on the formulation or the freeze-drying process, it may be desirable to include one or more organic solvents in the liquid formulation. For example, it may be desirable to include an organic solvent in the formulation to enhance the solubility of one or more active ingredients. Examples of suitable organic solvents include, but are not limited to, acetonitrile, methanol, ethanol, propanol, tert-butyl alcohol, acetone, cyclohexane, and dimethylsulfoxide (DMSO).

Bulking Agents

[0075] The purpose of the bulking agent is to provide bulk to the formulation and enhance cake formation. As used herein, the term "bulking agent" includes both "crystalline" and "non-crystalline" bulking agents. The term "crystalline" bulking agents refer to bulking agents that are capable of forming a crystal structure under typical

lyophilization conditions.

[0076] In general, a crystalline bulking agents refers to a bulking agent that is capable of crystallizing during freezing (e.g., between a temperature of about O ⁰ C and about - 50°C). Although, a crystalline bulking agent may require an annealing or thermal treatment step to promote crystallization during the freezing process. It is worthwhile to note that a bulking agent may or may not crystallize during lyophilization, depending upon the conditions of the lyophilization process and/or the other excipients present in the formulation. Typically, when a sufficient amount of crystalline bulking agent is included in a liquid formulation (e.g., when the ratio of crystalline bulking agent to amorphous solute is at least about 1.0, about 1.25 or about 1.5) and allowed to crystallize during the lyophilization process, the crystalline bulking agent may form a structural support matrix for the amorphous component of the formulation.

[0077] As used herein, the term "structural support matrix" refers to the support that the crystalline structure provides to the formulation (analogous to a "scaffolding"), such that the macrostructure of the cake is largely unaffected by any "microcollapse" of the amorphous solute residing within the interstices of the structural support matrix during primary drying. This crystalline structural support matrix may allow for primary drying with a product temperature above the Tg' of the amorphous component of the product. [0078] With reference to Figure 10, a crystalline bulking agent can provide a structural support matrix (10) for an amorphous solute (15) contained within the interstices (12) of the structural support matrix (10). As used herein, the term "microcollapse" refers to a change in the structure (e.g., collapse) of the amorphous solute residing within the interstices of the crystalline bulking agent structural support matrix that does not result in a change in the structure of the overall freeze-dried formulation or cake. [0079] Although crystalline bulking agents may improve cake structure, for example, by providing a structural support matrix during primary drying, it may not always be desirable to include a crystalline bulking agent in a formulation. For example, crystalline bulking agents may reduce protein stability both during lyophilization and during storage (see discussion of stabilizing agents below). Also, if a relatively high concentration of stabilizing agent, for example, a mono-, di-, or poly- saccharide stabilizing agent, is added to protect the protein active agent, it may not be possible to add a sufficient amount of crystalline bulking agent to allow for crystallization such that the crystalline bulking agent is able to act as a crystalline support. Common crystalline bulking agents include, but are

not limited to, glycine and mannitol. Mannitol is a naturally occurring carbohydrate classified as a sugar alcohol or polyol. Glycine is a neutral amino acid. [0080] The term non-crystalline bulking agent refers to any component of the formulation that provides bulk to the formulation but does not crystallize under typical lyophilization conditions. For example, a "non-crystalline" bulking agent will generally remain amorphous (i.e. not form a crystal structure) when cooled below O ⁰ C. Examples of non-crystalline bulking agents include monosaccharides, disaccharides, dextran, and polysaccharides. Additional non-crystalline bulking agents also may include amino acids, polypeptides, and proteins.

[0081] According to the invention, the primary drying step is carried out at a temperature above Tg' or at or above about T ₀ for the formulation and below Tc. In one embodiment, the rate of primary drying can be increased without including a crystalline bulking agent in the formulation. Therefore, in one embodiment, the formulation does not include a sufficient amount of crystalline bulking agent to form a structural support matrix. The amount of crystalline bulking agent that is "sufficient to form a structural support matrix" can vary depending upon the type of crystalline bulking agent, e.g., glycine versus mannitol (generally glycine forms crystals more readily and therefore will tend to crystallize even when included at a lower concentration). However, a crystalline bulking agent is generally able to provide a "structural support matrix" when the formulation includes a ratio of crystalline bulking agent to amorphous solute of at least about 1.0. When less crystalline bulking agent is included in the formulation, the crystalline bulking agent may not crystallize, only partially crystallize or only form "pockets" of crystals, rather than forming a crystalline support matrix. In one aspect of the invention, the formulation includes a weightweight ratio of crystalline bulking agent to amorphous solute of less than about 1.0, about 0.9, about 0.75 or about 0.50. As discussed below, the reduction in the amount of crystalline bulking agent in the formulation can result in improved lyoprotection and/or cryoprotection of the active agent.

Stabilizing Agents

[0082] Stabilizing agents are typically added to a formulation to improve stability of the protein formulation, for example, by reducing denaturation, aggregation, deamidation and oxidation of the protein during the freeze-drying process as well as during storage. Examples of stabilizing agents include cryoprotectants and lyoprotectants. The term

"cryoprotectant" refers to compounds that protect the active agent during freezing. The term "lyoprotectant" refers to compounds that protect the active agent during drying. [0083] Saccharides, including monosaccharides such as glucose, disaccharides such as sucrose (glucose + fructose), lactose (glucose + galactose), maltose (glucose + glucose) trehalose (alpha-D-glucopyranosyl alpha-D-glucopyranoside), and polysaccharides such as dextran (polysaccharide containing glucose monomers) are commonly used stabilizing agents. Glucose, lactose and maltose are reducing sugars and can reduce proteins by means of the mailard reaction. Disaccharides such as sucrose, trehalose, and polysaccharides, such as dextran, are non-reducing sugars.

[0084] Sugar alcohols such as mannitol and amino acids such as glycine have conventionally been used as crystalline bulking agents to provide a rigid cake support. However, they may be used as a stabilizing agent, if they remain in an amorphous phase following the freeze-dry process.

[0085] A few hypothesis exist to explain the stabilizing effects of non-reducing sugars. The hydrogen-bonding theory postulates that the disaccharide stabilizer is able to form hydrogen bonds with protein (similar to the replaced water) which, in turn, prevents protein denaturation. This is also called the water replacement hypothesis. The preferential exclusion hypothesis postulates that the stabilizing agent is preferentially excluded from protein surface and destabilizes the unfolded state more than the folded state. Thus, the thermodynamics of the system drives the protein towards the folded (native) state. A final hypothesis is the vitrification hypothesis which postulates that disaccharides form sugar glasses of extremely high viscosity. The protein and water molecules are immobilized in the viscous glass, leading to extremely high activation energies required for any reactions to occur. It is believed that stabilizing agents are able to hydrogen bond with the protein and thus prevent denaturation

[0086] However, during lyophilization, a crystalline bulking agent may become separated from the amorphous solute, for example, as the bulking agent crystallizes, the amorphous solute becomes trapped within the interstices of the crystalline structure. Consequently, the crystalline bulking agent is no longer able to stabilize the active agent by hydrogen bonding, preferential exclusion, or vitrification. Additionally, a crystalline bulking agent may form crystals during storage, which can result in reduced protein stability.

Tonicity/Osmostic Pressure

[0087] When two solutions containing different particle concentrations are separated from each other by a semipermeable membrane, solvent will move across the membrane from the solution with the lower concentration to the solution with the higher concentration. The movement of the solvent depends on the difference in the concentration of the particles and the permeability of the membrane. This movement of solvent is termed osmosis and the necessary pressure to halt its movement is called the osmotic pressure. It is important to realize that the osmotic pressure is determined by the total number of particles in solution, regardless of molecular nature. The total number of particles will thus depend on the degree of dissociation of solutes. For example, when added to water, sodium chloride dissociates into two ions per molecule, whereas sucrose does not dissociate.

[0088] Osmolar concentration can be expressed in two ways: osmolality, which is expressed as mmol/kg of solvent and osmolarity, which is expressed as mmol/1 of solution. Osmolality is a thermodynamically more precise expression because solution concentrations expressed on a weight basis are temperature independent while those based on volume will vary with temperature in a manner dependent on the thermal expansion of the solution.

[0089] Although the terms tonicity and osmolality are often used interchangeably, there is a clear distinction. Osmolality is a physical property dependent on the total number of solute particles present in a solution whereas tonicity is a physiological process dependent upon the selectively permeable characteristics of a membrane. As used herein, the term "tonicity" refers to the osmotic pressure of a solution due to concentration of particles that do not penetrate a cell membrane. For example, an isotonic solution has the same osmotic pressure as the interior of a cell, such that the cell volume is not affected when placed in an isotonic solution. In contrast, when placed in a hypotonic solution, a cell will swell, due to the influx of solvent. When placed in a hypertonic solution, water will leave the cell, causing the cell to shrink. Generally, solutes that permeate through or into cells freely may have no effect on tonicity but may increase the measured osmolality. [0090] hi normal humans the osmolality of body fluids is tightly regulated. Normal serum osmolality lies between 285 mOsm and 290 mOsm. Because movement of solvent from solutions having a low osmotic pressure to solutions having a high osmotic pressure can cause severe physiological problems, including cell dehydration (crenation) or

expansion of the cell until it breaks open (lysis), osmotic pressure is an important consideration when preparing a pharmaceutical formulation. Thus, although the tonicity of the formulation may vary depending upon the stability requirements of the formulation or for the route of administration, in many instances, particularly in formulations for subcutaneous administration, it is important that the formulation have approximately the same osmotic pressure (i.e., isotonic) as the cellular fluid (i.e., within approx. 50 mOsm). Therefore, a pharmaceutical formulation, particularly for subcutaneous administration, should have an osmotic pressure between about 250 mOsm and about 350 mOsm. In some instances, the tonicity modifier is added to the liquid formulation before freeze- drying. In other instances, the tonicity modifier is added along with the diluent during reconstitution of the freeze-dried formulation.

[0091] Excipients such as mannitol, sucrose, glycine, glycerol and sodium chloride are good tonicity adjusters.

pH or Buffering Agents

[0092] Buffers are typically included in pharmaceutical formulations to maintain the pH of the formulation at a physiologically acceptable pH. The desirable pH for a formulation may also be affected by the active agent. For example, most biopharmaceutical active agents have a higher activity within a specific pH range. Generally, the pH of the formulation is maintained between about 4.0 and about 8.0, between about 5.5 and about 7.5, or between about 6.0 and about 7.2. Typically the buffer is included in the liquid formulation at a concentration between about 2 mM to about 50 mM, or between about 10 mM and 25 mM.

[0093] Examples of suitable buffers include buffers derived from an acid such as phosphate, aconitic, citric, gluaric, malic, succinic and carbonic acid. Typically, the buffer is employed as an alkali or alkaline earth salt of one of these acids. Frequently the buffer is phosphate or citrate, often citrate, for example sodium citrate or citric acid. Other suitable buffers include acetate, Tris and histidine buffers.

IV. Modes of administration

[0094] The freeze-dried formulation of the invention is suitable for parenteral administration, including intravenous, subcutaneous and intramuscular administration.

Example 1: Preparation of Monoclonal Antibody in Sucrose Formulation.

[0095] A monoclonal IgG antibody (MW 150-160 kDa) was recombinantly produced using a standard mammalian cell line. Bulk drug substance was then obtained by purifying the monoclonal antibody from the cell culture supernatant using standard purification techniques.

[0096] The monoclonal antibody was diluted to the target concentration (i.e., 0 mg/mL, 10 mg/nxL, 20 mg/mL, 40 mg/mL, 60 mg/mL, 80 mg/niL, and 100 mg/mL) using a formulation buffer that included 10 mM citrate, 8% sucrose, and 0.04% Polysorbate-80 at a pH of 6.5. Protein concentration was monitored by measuring absorbance at A ₂₈₀ .

Example 2: Preparation of Bovine Serum Albumin (BSA) in Sucrose Formulation [0097] Bovine Serum Albumin (BSA) was purchased from Sigma (part #A2153). BSA was measured out by weight and dissolved in a formulation buffer containing 10 mM citrate, 8% sucrose, and 0.04% Polysorbate-80, pH 6.5. BSA was diluted with the formulation buffer to varying concentrations: 10 mg/mL, 20 mg/mL, 40 mg/mL, 60 mg/mL, 80 mg/mL, and 100 mg/mL.

Example 3: Preparation of Bovine Serum Albumin (BSA) in Sorbitol Formulation [0098] Bovine Serum Albumin (BSA) was purchased from Sigma (part #A2153). BSA was measured out by weight and dissolved in a formulation buffer containing 10 mM phosphate, 2% sorbitol, and 0.02% Polysorbate-80, pH 7.2. BSA was diluted with the formulation buffer to varying concentrations: 10 mg/mL, 20 mg/mL, 40 mg/mL, 60 mg/mL, 80 mg/mL, and 100 mg/mL.

Example 4: Freezing and Lyophilization Characterization Studies

A. Differential Scanning Calorimetry (DSC)

[0099] Differential Scanning Calorimetry (DSC) was performed using a TA Instruments QlOOO series DSC with an autosampler and refrigerated cooling system (RCS) to evaluate the effect of protein concentration on Tg'. Approximately 15-20 μL of liquid monoclonal antibody sample or BSA sample (from Examples 1 - 3) was placed into an aluminum pan (DSC aluminum pans, TA instruments, Cat # 900793.901) and

hermetically sealed with an aluminum cover (DSC aluminum cover, TA Instruments Cat #900794.901). The reference used for all samples was an empty aluminum pan crimped in the same manner as the corresponding samples. The RCS was set on continuous run and the nitrogen supply flow meter was set at 50 mL/minute. The glass transition temperatures of the lyophilized and liquid samples were analyzed using TA Instruments' Universal Analysis software.

[0100] The formulations included varying protein concentrations to determine the effect of protein concentration on the glass transition temperature (Tg'). The concentrations used were 0 mg/mL, 10 mg/mL, 20 mg/mL, 40 mg/mL, 60 mg/mL, 80 mg/mL, and 100 mg/mL. The samples were run in duplicate, at minimum, and an average was taken.

[0101] There was a general trend between protein concentration and Tg' for the formulations containing sucrose, as shown in Table 1 and Table 2. As protein concentration increased, Tg' increased (although no significant differences were observed between the 10 mg/ml and 20 mg/ml sample or the 80 mg/ml and 100 mg/ml sample for the monoclonal antibody formulation, nor the 10 mg/ml and 20 mg/ml sample for the BSA formulation). The Tg' for the BSA formulation containing sorbitol changed little with protein concentration, as shown in Table 3 and all were lower than those of formulations containing sucrose.

Table 1 Tg' of the Monoclonal Antibody and Sucrose Formulation at Different

Protein Concentrations

Table 2 Tg ^f of the BSA and Sucrose Formulation at Different Protein Concentrations

Table 3 Tg' of the BSA and Sorbitol Formulation at Different Protein Concentrations

B. Freeze Dry Microscopy

[0102] Freeze dry microscopy studies were performed using liquid monoclonal antibody samples and liquid BSA samples, prepared as described in Examples 1 - 3, with varying protein concentrations to determine the effect of protein concentration on collapse temperature. The concentrations used were 0 mg/mL, 10 mg/mL, 20 mg/mL, 40 mg/mL, 60 mg/mL, 80 mg/mL, and 100 mg/mL. Freeze-dry microscopy runs were performed in duplicate.

[0103] All liquid samples were tested using a Nikon Eclipse E600 POL microscope with a Linkam FDCS 196 stage. Images were recorded using a Sony DXC-390 camera attachment. Briefly, a 1 μh sample of liquid monoclonal antibody sample was placed between two cover slips on the thermal conductor and sealed within the stage. The samples were cooled to -45 ⁰ C at a rate of 5°C/mmute, a vacuum was then initiated within the stage, and then heated to 30°C at a rate of l°C/rninute.

[0104] Tables 4 - 6 show the temperature at which an initial structural change (or change in the initial appearance) was observed (onset of collapse), and the temperature at which collapse became clearly evident (i.e., the initial structural appearance was totally absent). Tg' is also shown in these tables for comparison. Figure 8 shows "snapshots" of the freeze-dry microscopy run for a 80 mg/mL sample of monoclonal antibody taken over a sequence of time as the temperature of the product was gradually increased. Frozen material is on the right side of each panel and dried material is on the left. The drying front gradually moves from left to right. As shown in Figure 8: (A) -24°C was a temperature prior to any change, i.e. demonstrating the "initial appearance"; (B) -19°C was observed as the onset temperature; (C) -12°C was observed as the collapse temperature; and (D) -7°C was observed as the solution melting temperature.

[0105] In general, as the protein concentration increased, the collapse temperature increased and the difference between onset temperature and collapse temperature (at a given concentration) increased. In general, Tg' for the sucrose containing formulations were similar to the onset temperature at protein concentrations up to about 40 mg/ml through 60 mg/ml. The 80 mg/ml monoclonal antibody sample showed a collapse temperature (-12 ⁰ C), which was 14 ⁰ C higher than its Tg', and an onset temperature (-19 ⁰ C) that was 7 ⁰ C higher than its Tg'. This suggests that for formulations with high protein concentration, the generalization that Tg' is closely associated with collapse temperature may not be applicable. Thus, a significantly shorter lyophilization cycle may be possible if the actual collapse temperature of the formulation is considered, rather than using Tg' or T ₀ as a working collapse temperature. Differences in collapse temperature and Tg' at different protein concentrations are compared graphically in Figure 4 for the monoclonal antibody and sucrose formulation.

Table 4 Freeze Dry Microscopy Profile of Monoclonal antibody and Sucrose

Formulation at Varying Concentrations

Table 5

Table 6

Example 5: Lyophilization

[0106] 5.4mL of the formulations described in Examples 1 - 3 were placed in 2OmL Type 1 glass vials (Schott Purform, West/68000321) and partially stoppered with a 20mm lyophilization stopper (4432/50 GR S87J; WESTAR Level-3, West/10142922). The partially stoppered 20 niL vials were then placed in a Virtis Genesis 32 Freeze Dryer and freeze-dried. The freezing rate was 0.3°C/rninute and the primary drying shelf temperature was -5 ⁰ C with a chamber pressure of 100 mTorr. Secondary drying was performed at a

shelf temperature of 3O ⁰ C and a chamber pressure of 100 mTorr. The temperatures of the monoclonal antibody formulation and BSA formulations were monitored during the lyophilization process using thermocouples (Omega, Part # 5SRTC-TT-T-30-36). The majority of samples freeze-dried using this cycle showed a product temperature below the collapse temperature (as measured by freeze-dry microscopy) throughout the primary drying, and the cakes showed no visual signs of collapse. However, the sorbitol containing formulation with 0 mg/mL BSA showed a product temperature that exceeded collapse temperature (determined by freeze-dry microscopy) throughout primary drying. The cakes for the 0 mg/mL samples were completely collapsed following freeze-drying as shown in Figure 9.

[0107] Additionally, the product temperature for the sorbitol containing formulation with 10 mg/mL BSA also exceeded the collapse temperature for a portion of primary drying, and the cakes showed partial collapse (Figure 9). The other sorbitol containing formulations (containing BSA from 20 mg/mL to 100 mg/mL) showed little or no signs of collapse. The product temperature for these formulations were below the collapse temperature (or in some cases just at collapse temperature). All of these samples exceeded Tg' and in some cases T ₀ during the primary drying process. This is consistent with increased protein concentration allowing for drying above the Tg' and To-

Example 6: Freeze-drying Monoclonal Antibody Formulation Under Conservative and Aggressive Cycle Conditions

[0108] A study was performed to compare the cake quality and the stability of a monoclonal antibody sample at 80 mg/mL dried under conservative and aggressive cycle conditions. Under the conservative cycle conditions, the freezing rate was 0.3°C/minute and the primary drying shelf temperature was -25 ⁰ C with a chamber pressure of 100 mTorr. The maximum product primary drying temperature stayed below Tg' and To under these conditions. Secondary drying in this conservative cycle was performed at a shelf temperature of 25 ⁰ C and a chamber pressure of 100 mTorr. Under the more aggressive cycle conditions, freezing rate was also 0.3 ⁰ C per minute, but primary drying was performed with higher shelf temperatures ranging from -5 ⁰ C to as high as 3O ⁰ C and chamber pressures ranged from 100 mTorr to 225 mTorr. The maximum product primary drying temperature under the aggressive conditions was above Tg' in all cases and above both Tg' and To in some cases. Secondary drying conditions varied under the more

aggressive cycle conditions, with a shelf temperature from 30 to 35 ⁰ C and chamber pressure ranging from 100 mTorr to 225 mTorr.

z. Appearance and Reconstitution Time

[0109] Regardless of the cycle conditions and whether the product was freeze-dried above or below Tg' (or above or below To) during primary drying, after freeze-drying, there were no differences observed by visual appearance in the cake or solution post reconstitution. No notable cake collapse was observed for any of the samples. Reconstituted samples contained no apparent particulate matter. Reconstitution times were also comparable.

ii. Stability Indicating Assays

[0110] Stability indicating assays were available for the monoclonal antibody. SEC- HPLC was used to monitor protein aggregation. Samples that were dried under conservative conditions had an equivalent percentage of aggregation to those dried under aggressive conditions. The percentage of aggregate for each of these samples was equivalent to the percentage of aggregate in the starting material prior to freeze-drying. That is, regardless of whether the maximum primary drying product temperature was above or below Tg' and To, the stability indicating assays showed an equivalent profile. This included samples that had a maximum product primary drying temperature as high as -16 ⁰ C, which was 3 ⁰ C higher than the To and 1O ⁰ C higher than Tg'. In one study, samples freeze-dried with a primary drying shelf temperature of -25 ⁰ C (yielding a maximum product primary drying temperature of -28 ⁰ C) were compared to samples freeze-dried with a primary drying shelf temperature of O ⁰ C (maximum product primary drying temperature of -2O ⁰ C). Therefore, maximum product primary drying temperature was either below or above Tg' for these cycles, respectively. These samples were compared by multiple stability indicating assays over a period of 3 months. These samples were also compared at 6-, 9-, and 12-months for stability using multiple stability indicating assays. SEC-HPLC was used to monitor aggregation, relative binding of the monoclonal antibody to its antigen was measured by ELISA, and the relative potencies of the monoclonal antibody samples were measured by an inhibition of binding assay. Reduced SDS-PAGE was used to monitor protein fragmentation, and IEC-HPLC was used to monitor charge heterogeneity. Regardless of whether the maximum primary drying product temperature

was above or below Tg', samples were comparable at each time point of the study for all assays tested.

Ui. Fourier-transform Infrared (FTIR) Spectroscopy

[0111] Fourier-transform infrared spectroscopy was used to characterize and compare the secondary structures of the monoclonal antibody in the bulk liquid and lyophilized sample to compare the effects of the aggressive and conservative lyophilization cycles on the freeze-dried protein structure. Because retention of solution-state structure in the freeze-dried powder is thought to provide improved physical stability, the FTIR data provide initial insight into the relative probable physical stability of the FDP (final drug product) as a result of the different lyophilization cycles.

[0112] FTIR spectra were recorded on an ABB Bomem MB 104 series Fourier- transform infrared spectrometer equipped with a DTGS detector. The instrument was purged continuously with dry nitrogen to minimize water vapor. Both liquid and powder samples were collected on a SensIR Technologies DuraSamplIR II attenuated total reflection accessory fitted with a diamond triple-bounce internal reflection element. A total of 32 scans were collected for each spectrum at a resolution of 4 cm ^"1 . Single-beam spectra of both the background (R ₀ ) and sample (R) were recorded, and the attenuated total reflection absorbance (ATR) was calculated as ATR = -log(R/Ro) in GRAMS software (Thermo Galactic GRAMS/32 AI, v. 6.01). Spectra were interactively corrected for liquid water and water vapor, as appropriate, and then baseline- and offset-corrected between 1715-1590 cm ^"1 (amide I). Spectra were corrected for concentration by normalizing the amide I area to 1.0. Difference spectra were calculated by subtracting the spectrum of the final drug product minus the parent bulk liquid spectrum. The areas of the difference spectra were quantified by integration in GRAMS.

[0113] The spectrum of the BDS (bulk drug substance) liquid is shown in Figure 5. The main peak centered at 1637 cm ^"1 is characteristic of β-sheet structures. There is a broad shoulder spanning 1690-1660 cm ^"1 arising from overlap of the high-frequency contributions of β-sheets (1675-1690 cm ^"1 ) and vibrations of random coils/turns (1660- 1675 cm ^"1 ). These features are all expected based on the known secondary structure content of IgG-I antibodies.

[0114] Upon lyophilization, for both the aggressive and conservative cycles, there is a decrease in the amplitude of the main β-sheet peak at 1637 cm ^"1 and increases centered near 1670 and 1695 cm ^"1 (Figure 5) relative to the bulk spectrum. Visually, the changes

appear quite similar in both spectra (aggressive and conservative cycles), which suggests that both cycles affect the secondary structure of monoclonal antibody in the same way and to the same extent. The changes in the secondary structure of the monoclonal Ab caused by lyophilization can be better represented by calculating the difference spectra between the FDP and parent BDS spectra.

[0115] The final drug product-minus bulk drug substance difference spectra are shown in Figure 6 for the aggressive and conservative lyophilization cycles. In the difference spectra, any features lost from the bulk are negative-going bands, while any features gained in the lyophilized samples are positive-going. It is clear that in both the aggressive and conservative cycles, native (liquid-state) β-sheet structure is lost upon lyophilization, based on the negative band at 1630 cm ^"1 . On the other hand, bands at 1670 and 1695 cm ^"1 have increased in amplitude, which can be assigned to an increase in the amount of turns/random coils and intermolecular β-sheet-like contacts, respectively. [0116] The extent of the changes caused by the two lyophilization cycles can be quantified by calculating the area of the negative bands, which, because the spectra are area-normalized to 1.0, represents the percentage of change from native structure. For the conservative cycle, the difference is 8.6%, while for the aggressive cycle, the difference is 8.8%. The error of the measurements was determined to be approximately 0.7%, based on the area of the difference between spectra of multiple preparations of the same sample. Therefore, within error, the aggressive and conservative cycles induce the same secondary structural changes to the protein, relative to the liquid state, and to the same extent. Based on these results, it is expected that both cycles should have the same impact on the physical stability of the lyophilized product.

zv. B. E. T Surface Area Analysis

[0117] Three lyophilized samples from each run were sent to Clear Science (Minneapolis, MN) for multipoint BET (Brunauer, Emmett and Teller) surface area testing using a Micromeritics Gemini 2375 Surface Area Analyzer. Using this analyzer, surface area values were determined and all isotherms in the lyophilized cakes were classified. These data were compared to three samples taken from a confirmation run. There was no significant difference found in any of the samples sent for surface area testing, and BET surface area values for all samples were relatively low (between 0.4 and 1.7 m ² /g). AU isotherms were classified as Type II indicating that the lyophilized cakes had relatively

low pore volume. The relatively low surface area and pore volume is likely a reflection of the high percentage of solids in the formulation, and it explains the relatively long reconstitution times.