U.S. patent application number 14/297403 was filed with the patent office on 2014-12-25 for low viscosity highly concentrated suspensions.
The applicant listed for this patent is Board of Regents, The University of Texas System. Invention is credited to Joshua Engstrom, Keith P. Johnston, Maria Andrea Mazuski, Miguel Angelo Rodrigues.
Application Number | 20140378370 14/297403 |
Document ID | / |
Family ID | 42170660 |
Filed Date | 2014-12-25 |
United States Patent
Application |
20140378370 |
Kind Code |
A1 |
Johnston; Keith P. ; et
al. |
December 25, 2014 |
LOW VISCOSITY HIGHLY CONCENTRATED SUSPENSIONS
Abstract
The present invent ion also provides a high concentration low
viscosity suspension of an pharmaceutically acceptable solvent with
one or more sub-micron or micron-sized non-crystalline particles
comprising one or more proteins or peptides. Optionally one or more
additives in the pharmaceutically acceptable solvent to form a high
concentration low viscosity suspension with a concentration of at
least 20 mg/ml and a solution viscosity of between 2 and 100
centipoise that is suspendable upon shaking or agitation, wherein
upon delivery the one or more sub-micron or micron-sized peptides
dissolves and do not form peptide aggregates syringeable through a
21 to 27-gauge needle.
Inventors: |
Johnston; Keith P.; (Austin,
TX) ; Mazuski; Maria Andrea; (Austin, TX) ;
Engstrom; Joshua; (Spotswood, NJ) ; Rodrigues; Miguel
Angelo; (Quinta do Conde, PT) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Board of Regents, The University of Texas System |
Austin |
TX |
US |
|
|
Family ID: |
42170660 |
Appl. No.: |
14/297403 |
Filed: |
June 5, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13129283 |
Jul 27, 2011 |
8779094 |
|
|
PCT/US2009/063852 |
Nov 10, 2009 |
|
|
|
14297403 |
|
|
|
|
61115080 |
Nov 16, 2008 |
|
|
|
Current U.S.
Class: |
514/1.1 |
Current CPC
Class: |
A61K 9/0019 20130101;
A61P 37/00 20180101; A61P 43/00 20180101; A61K 38/385 20130101;
C07K 1/113 20130101 |
Class at
Publication: |
514/1.1 |
International
Class: |
A61K 9/00 20060101
A61K009/00; A61K 38/38 20060101 A61K038/38 |
Claims
1.-50. (canceled)
51. A method of delivering a suspension of protein particles to a
subject, the method comprising subcutaneously injecting said
suspension into said subject, wherein the protein concentration
within said suspension is greater than 200 mg/mL and the viscosity
of said suspension is less than 50 centipoise.
52. The method of claim 51, wherein said injection is accomplished
using a syringe comprising a 21-gauge to 27-gauge needle.
53. The method of claim 52, wherein said syringe comprises a
25-gauge to 27-gauge needle.
54. The method of claim 51, wherein said suspension has a volume of
5 mL or less.
55. The method of claim 51, wherein said suspension has a volume of
1.5 mL or less.
56. The method of claim 51, wherein said particles are sub-micron
sized particles.
Description
TECHNICAL FIELD OF THE INVENTION
[0001] The present invention relates in general to the field of
protein storage and delivery, and more particularly, to novel
compositions and methods of making highly concentrated protein
suspension and precursors thereof.
BACKGROUND ART
[0002] Without limiting the scope of the invention, its background
is described in connection with the concentration of proteins. The
use of proteins and other polypeptides for therapeutics is on the
rise in recent years as a way to expand and better treat patients
since they are viewed to be less toxic and behave more predictably
in vivo than other classes of drugs not naturally found in the
body. Delivery of protein therapeutics has been limited primarily
to dilute large volume intravenous injections to deliver the high
dose required (100-1000 mg) and to avoid physical and chemical
instabilities of proteins at high concentrations. A potentially
less invasive method of administration is subcutaneous injection.
Since the injection volume is limited to 1.5 ml, the concentration
of the protein therapeutic is often substantially above 100 mg/ml.
In addition to polypeptide stability, another major concern is the
dramatic increase in viscosity for solution concentrations greater
than 100 to 400 mg/ml due to protein interactions. If the primary
interactions are attractive protein-protein interactions due to
electrostatics, this increase in viscosity can been avoided by
adding sodium chloride to increase the ionic strength of the
solution and by varying the buffer species and pH of the solution.
At these high concentrations, large excipient concentrations are
often needed to protect against denaturation. An alternative
approach would be to form a suspension of an insoluble protein in a
non-aqueous solvent. The viscosity of highly concentrated
suspensions can be much lower than for solutions and require
smaller excipient levels to stabilize the protein. However, for
successful delivery with concentrated suspensions, the particle
size and suspension uniformity must be controlled in order to
administer an accurate and uniform dose.
[0003] To date, there are relatively few examples of suspensions of
proteins in non-aqueous media for medicinal purposes. Highly
viscous suspensions of bovine somatotropin, marketed to increase
milk production in dairy cows, and a bovine growth hormone
releasing factor analog, used to release somatotropin from the
cow's pituitary gland, are formulated in sesame oil and Miglyol
oil, respectively. These viscous suspensions require a large 14-16
gauge needle for injection, whereas the preferred needle size for
humans is between 25-gauge and 27-gauge. In addition, a few
non-aqueous injections have been formulated as extended release
formulations for the peptide insulin and very stable proteins such
as protein C and a proprietary monoclonal antibody with the aid of
viscosity enhancers and gel forming polymers in the presence of
diluents such as benzyl benzoate or benzyl alcohol. However, these
formulations are syringeable only with a larger 21-gauge needle
causing considerable pain upon injection leading to non-compliance
and the high levels of excipients reduce the overall concentration
of the protein in the formulation. Another option is to crystallize
the protein or monoclonal antibody and form an aqueous suspension
of the crystals. This approach has been shown for three monoclonal
antibodies and insulin. However, crystallization of high molecular
weight proteins can be very difficult due to the high degree of
segmental flexibility, and is more feasible for small peptides that
have a much lower degree of flexibility.
DISCLOSURE OF THE INVENTION
[0004] The present invention provides a method of making a high
concentration low viscosity protein or peptide suspension by
forming one or more sub-micron or micron-sized particles comprising
one or more proteins or peptides, adding optionally one or more
additives to the one or more sub-micron or micron-sized particles
and suspending the one or more sub-micron or micron micron-sized
particles in a pharmaceutically acceptable solvent to form a high
concentration low viscosity suspension with a concentration of at
least 20 mg/ml and a solution viscosity of between 2 and 100
centipoise that is suspendable upon shaking or agitation, wherein
upon delivery the one or more sub-micron or micron-sized peptides
dissolves and do not form peptide aggregates or only a small
fraction of aggregates and is syringeable through a 21 to 27-gauge
needle. The pharmaceutically acceptable solvent may be a
pharmaceutically acceptable aqueous solvent, a pharmaceutically
acceptable non-aqueous solvent or combination.
[0005] In addition, the one or more micron-sized peptide particles
are formed in a dosage container and may be delivered directly from
the dosage container that is a vial, an ampule, a syringe or a bulk
container. The one or more micron-sized peptide particles may be
made by milling, precipitation, dialysis, sieving, spray drying,
lyophilization, spray freeze drying, spray freezing into liquids,
thin film freezing, or freezing directly in a dosage container. The
one or more additives may be part of the one or more sub-micron or
micron-sized particles, the a high concentration low viscosity
suspension or both.
[0006] The present invention also provides a high concentration low
viscosity suspension of an pharmaceutically acceptable solvent with
one or more sub-micron or micron-sized non-crystalline particles
comprising one or more proteins or peptides. Optionally one or more
additives in the pharmaceutically acceptable solvent to form a high
concentration low viscosity suspension with a concentration of at
least 20 mg/ml and a solution viscosity of between 2 and 100
centipoise that is suspendable upon shaking or agitation, wherein
upon delivery the one or more sub-micron or micron-sized peptides
dissolves and do not form peptide aggregates or only a small
fraction of aggregates syringeable through a 21 to 27-gauge needle.
The pharmaceutically acceptable solvent may be a pharmaceutically
acceptable aqueous solvent, a pharmaceutically acceptable
non-aqueous solvent or combination. In addition, the one or more
micron-sized peptide particles are formed in a dosage container and
may be delivered directly from the dosage container that is a vial,
an ampule, a syringe or a bulk container. The one or more
micron-sized peptide particles may be made by milling,
precipitation, dialysis, sieving, spray drying, lyophilization,
spray freeze drying, spray freezing into liquids, thin film
freezing, or freezing directly in a dosage container. The one or
more additives may be part of the one or more sub-micron or
micron-sized particles, the a high concentration low viscosity
suspension or both.
[0007] The present invention provides a single dose high
concentration low viscosity suspension in a single dose container.
The single dose container includes a pharmaceutically acceptable
solvent disposed in the single dose container, wherein the
pharmaceutically acceptable solvent is selected from an aqueous
solvent, a non-aqueous solvent or combination thereof and one or
more sub-micron or micron-sized non-crystalline particles disposed
in the single dose container, wherein the one or more sub-micron or
micron-sized non-crystalline particles comprising one or more
proteins or peptides. In addition, one or more additives may be
optionally disposed in the single dose container to form a high
concentration low viscosity suspension with a a concentration of at
least 20 mg/ml and a solution viscosity of between 2 and 100
centipoise syringeable through a 21 to 27-gauge needle.
DESCRIPTION OF THE DRAWINGS
[0008] For a more complete understanding of the features and
advantages of the present invention, reference is now made to the
detailed description of the invention along with the accompanying
figures and in which:
[0009] FIGS. 1A-C are SEM images of particles made by the process
of one embodiment of the present invention.
[0010] FIGS. 2A-2E are images of suspensions before (FIGS. 2A and
B) and after (FIGS. 2C-E) centrifugation (20 min at 3000 rpm) from
left to right 1.5M Ammonium Sulfate (a and c), 30% PEG300 (b and
d), 35% NMP (e) all in 150 mM pH 4.7 acetate buffer with added NaCl
to 154 mM ionic strength.
[0011] FIG. 3 is an image of aqueous PEG 300 suspensions after 20
min of centrifugation at 3000 g. From left to right 30% PEG300, 40%
PEG300, and 50% PEG 300 all in 150 mM pH 4.7 acetate buffer with
added NaCl to 154 mM ionic strength.
[0012] FIG. 4 is a graph of the apparent viscosity of aqueous
suspensions at various pHs all with 50% PEG 300 and added NaCl to
154 mM ionic strength.
[0013] FIG. 5 is an image of aqueous 50% PEG 300 suspensions after
20 min. of centrifugation at 3000 g. From left to right pH 4.7
acetate buffer, pH 5.5 acetate buffer, and pH 7.4 acetate
buffer.
[0014] FIG. 6 is a graph of the apparent viscosity of aqueous
suspensions with PEG 300 and organic additives in 150 mM pH 4.7
acetate buffer with added NaCl to 154 mM ionic strength.
[0015] FIGS. 7A and 7B are graphs that show the apparent
viscosities of milled BSA and trehalose at various ratios in 150 mM
pH 4.7 acetate buffer along with the theoretical viscosity as
calculated from the Krieger-Dougherty equation using the [.eta.] of
the pure milled particles in a) 25% PEG 300 20% Ethanol and b) 35%
PEG 300 and 15% NMP.
[0016] FIGS. 8A-8E are SEM images of various frozen powders of
IgG.
[0017] FIG. 9 is a graph of the optical density of the IgG with
various additives totaling 50% of the solvent by volume to decrease
the solubility of the IgG as measured at 350 nm. The right hand
side has the absorbance of the 5 mg/ml concentration of IgG in the
pH 6.4 20 mM histidine buffer with no additional additive.
[0018] FIG. 10A-10D are images of various suspensions of IgG.
[0019] FIG. 11A-11E are microscope images of various suspensions of
IgG.
[0020] FIG. 12. Volume % of particles versus size measured for the
original milled particles in acetonitrile and ethanol and after 2
months of storage for the suspensions in pure benzyl benzoate and a
mixture of benzyl benzoate and oil both measured immediately after
being diluted in ethanol to 10-15% obscuration by light
scattering.
[0021] FIG. 13A-13C Pictures of the 300 mg/mL Lysozyme suspension
in 50/50 Benzyl Benzoate and Safflower Oil.
[0022] FIG. 14 Viscosity of a solution of benzyl benzoate and
safflower oil at room temperature at varying concentrations.
[0023] FIG. 15 The apparent viscosity as a function of
concentration of particle as suspensions in the non-aqueous
solvents and the theoretical viscosity of an aqueous lysozyme
solution.
[0024] FIG. 16 Karl Fisher moisture content analysis of the
suspensions.
[0025] FIG. 17 Image of a suspension taken immediately after
formulation.
[0026] FIG. 18 Freezing temperature profiles of lysozyme solutions
(10 mg/ml) inside vials.
[0027] FIG. 19A-19D Volume size distributions of the protein
particles produced by Film Freezing in Vials in the conditions
described in Table 2.
[0028] FIG. 20 shows a 4 ml of lysozyme solution (20 mg/ml) in
water were frozen by film freezing inside the vial.
[0029] FIG. 21 shows a 2 ml of hemoglobin solution (150 mg/ml) in
water was frozen by film freezing inside the vial.
DESCRIPTION OF THE INVENTION
[0030] While the making and using of various embodiments of the
present invention are discussed in detail below, it should be
appreciated that the present invention provides many applicable
inventive concepts that can be embodied in a wide variety of
specific contexts. The specific embodiments discussed herein are
merely illustrative of specific ways to make and use the invention
and do not delimit the scope of the invention.
[0031] To facilitate the understanding of this invention, a number
of terms are defined below. Terms defined herein have meanings as
commonly understood by a person of ordinary skill in the areas
relevant to the present invention. Terms such as "a", "an" and
"the" are not intended to refer to only a singular entity, but
include the general class of which a specific example may be used
for illustration. The terminology herein is used to describe
specific embodiments of the invention, but their usage does not
delimit the invention, except as outlined in the claims.
[0032] As used herein, the term "high protein concentration" refers
to liquids, gels, hydrogels or gel-like compositions with a protein
concentration of greater than 100 mg/ml.
[0033] As used herein, the term "non-aggregating" or "not
aggregating" or "not aggregated" refers to protein particles that
remain in suspension despite being provided in the form of a high
protein concentration, e.g., a protein concentration greater than
100 mg/ml.
[0034] As used herein, the term "syringable" refers to a final
composition for delivery to a subject that is sufficiently fluid to
be flowable. For example, a composition that is "syringable" has a
low enough viscosity to load the syringe and inject a subject from
the syringe without undue force.
[0035] As used herein, the term "non-settling" or "redispersible"
refers to a composition that remains in solution phase (i.e., they
do not sediment) after an extended period of time, e.g., 1 hour, 2
hours, 1 day, 3 days, 5 days, 1 week, 1 month, 3 months, 6 months,
1 year or more. For example, a composition is "re-dispersible" is
upon re-dispersion it does not flocculate so quickly as to prevent
reproducible dosing of a drug.
[0036] As used herein, the terms "protein(s)," "polypeptide(s)" and
"peptide(s)" refers to a polymer composition formed from the
linking amino acids into a chain of various lengths.
[0037] As used herein, the term "additive(s)" refers to salts,
sugars, organics, buffers, polymers and other compositions that
include: Disodium edetate, Sodium chloride, Sodium citrate, Sodium
succinate, Sodium hydroxide, Sodium glucoheptonate, Sodium
acetyltryptophanate, Sodium bicarbonate, Sodium caprylate, Sodium
pertechnetate, sodium acetate, sodium dodecyl sulfate, aluminum
hydroxide, aluminum phosphate, ammonium citrate, calcium chloride,
calcium, potassium chloride, potassium sodium tartarate, zinc
oxide, zinc, stannous chloride, magnesium sulfate, magnesium
stearate, titanium dioxide, DL-lactic/glycolic acids, asparagine,
L-arginine, arginine hydrochloride, adenine, histidine, glycine,
glutamine, glutathione, imidazole, protamine, protamine sulfate,
phosphoric acid, Tri-n-butyl phosphate, ascorbic acid, cysteine
hydrochloride, hydrochloric acid, hydrogen citrate, trisodium
citrate, guanidine hydrochloride, mannitol, lactose, sucrose,
agarose, sorbitol, maltose, trehalose, surfactants, polysorbate 80,
polysorbate 20, poloxamer 188, sorbitan monooleate, triton n101,
m-cresol, benyl alcohol, ethanolamine, glycerin,
phosphorylethanolamine, tromethamine, 2-phenyloxyethanol,
chlorobutanol, dimethylsulfoxide, N-methyl-2-pyrrolidone,
propyleneglycol, Polyoxyl 35 castor oil, methyl hydroxybenzoate,
tromethamine, corn oil-mono-di-triglycerides, poloxyl 40
hydrogenated castor oil, tocopherol, n-acetyltryptophan,
octa-fluoropropane, castor oil, polyoxyethylated oleic glycerides,
polyoxytethylated castor oil, phenol (antiseptic), glyclyglycine,
thimerosal (antiseptic, antifungal), Parabens (preservative),
Gelatin, Formaldehyde, Dulbccco's modified eagles medium,
Hydrocortisone, Neomycin, Von Willebrand factor, Gluteraldehyde,
Benzethonium chloride, White petroleum, p-aminopheyl-p-anisate,
monosodium glutamate, beta-propiolactone, Acetate, Citrate,
Glutamate, Glycinate, Histidine, Lactate, Maleate, Phosphate,
Succinate, Tartrate, Tris, Carbomer 1342 (copolymer of acrylic acid
and a long chain alkyl methacrylate cross-linked with allyl ethers
of pentaerythritol), Glucose star polymer, Silicone polymer,
Polydimethylsiloxane, Polyethylene glycol, carboxymethylcellulose,
Poly(glycolic acid), Poly(lactic-co-glycolic acid), Polylactic
acid, Dextran 40, Poloxamers (triblock copolymers of ethylene oxide
and propylene oxide),
[0038] For highly concentrated protein suspensions in non-aqueous
solvents, the Krieger-Dougherty equation can be used to correlate
the relative viscosity of a suspension .eta. over that of a
solution .eta..sub.o to the volume fraction of particles .phi. (Eq.
1).
.eta. .eta. o = [ 1 - ( .phi. .phi. max ) ] - [ .eta. ] .phi. max
Eq . 1 ##EQU00001##
[0039] The intrinsic viscosity, [.eta.] approaches 2.5, the
Einstein value, assuming non-interacting, spherical particles with
only excluded volume interactions. However, [.eta.] increases upon
solvation of the particles, deviation from a spherical shape and
electrostatic interactions that produce primary, secondary and
tertiary electroviscous effects. For the non-aqueous protein
suspensions demonstrated previously, [.eta.] of approximately 2.5
indicated a lack of solvation, shape and electroviscous effects on
the viscosity of lysozyme milled particles. However, non-aqueous
solvents can sometimes cause pain on injection and delayed and
slowed release of the particles. Consequently, aqueous-based
suspensions of highly concentrated and molecularly stable, protein
particles would be an attractive alternative to non-aqueous
suspensions.
[0040] Therapeutic proteins may be designed for high solubilities
in aqueous media in the range of 100 mg/ml, for example, for the
model protein BSA. Thus, the solubilities must often be decreased
significantly, in order to form a suspension of micron-sized
particles. Precipitant that can decrease the solubility of a
protein in water may be separated into three categories; salts,
polymers and water-soluble organics. Salts may decrease the
solubility of a protein by competing for waters of hydration as
well as ion binding to produce stronger interactions between
protein molecules. Salts also reduce electrostatic repulsion by
decreasing the thickness of the double layer. However, a total
ionic strength of around 154 mM, the tonicity of the blood, is
typically recommended to prevent pain upon injection. Polymers,
most commonly polyethylene glycol (PEG), which are preferentially
excluded from the protein surface, produce depletion attraction
causing precipitation. PEG is also known to increase the thermal
stability of a protein. Furthermore, it is an acceptable excipient
for sub-cutaneous injection. Water-soluble organic additives, such
as ethanol and n-methyl-2-pyrrolidone (NMP), decrease protein
solubility by lowering the dielectric constant and by excluded
volume effects resulting from their exclusion from the protein
surfacc.
[0041] The objective of this study was to form low viscosity
(<50 cP), highly concentrated (100 to 350 mg/ml) aqueous
suspensions of sub-micron to micron-sized particles of the model
protein bovine serum albumin (BSA) suitable for subcutaneous
delivery. Precipitants, known to reduce the solubility of proteins
included combinations of ammonium sulfate (a representative salt),
PEG300 (a low molecular weight polymer), and ethanol and
N-methyl-2-pyrrolidone (NMP). Two key factors influence the
viscosity of the suspension at a given volume fraction of protein:
the initial viscosity of the solution without protein and the
intrinsic viscosity with protein present. Milled particles smaller
than 37 .mu.m of BSA were suspended in a variety of aqueous-based
solvents to characterize these competing effects. In many cases,
the concentrations of additives were within pharmaceutically
acceptable limits. The apparent viscosity of the suspension at
various particle concentrations is correlated with the
Krieger-Dougherty equation to determine the intrinsic viscosity.
The intrinsic viscosity was used to characterize interparticle
interactions including electroviscous and solvation interactions. A
variety of low viscosity, highly concentrated (up to 350 mg/ml)
aqueous-based suspensions comprising pharmaceutically relevant
additives are reported for milled BSA particles. The insight gained
from the study of the viscosities and morphologies of suspensions
of the model protein BSA will be useful for the design of
suspensions of therapeutic proteins such as IgG.
[0042] BSA powder or trehalose powder as received was dry milled
with a porcelain mortar and pestle separately for several minutes.
The milled powder was then sieved through a number 400 mesh and
particles smaller than 37 .mu.m were collected. For the mixed
milled BSA and trehalose particles, the necessary ratio of BSA to
trehalose was then mixed with the mortar and pestle and received
through the number 400 mesh to preserve the correct particle ratio
but retain the smaller than 37 .mu.m particle size. Known weights
of powder were suspended in aqueous solutions composed of NMP, PEG
300, and/or ethanol, and sodium chloride or ammonium sulfate salts.
The pH was chosen as 4.7, 5.5 or 7.4 with an acetate (pH 4.7 and
5.5) or phosphate (pH 7.4) buffer. Each vial was then shaken by
hand to disperse the powder evenly through the suspension.
Additional mixing using the tip of the needle was used to ensure
uniformity if necessary.
[0043] BSA was readily soluble at 5 mg/ml in pH 4.7 150 mM acetate
buffer. An aliquot of a 5 mg/ml BSA solution was mixed with an
equal volume of a second aqueous solution containing additives for
the purpose of determining the degree to which the protein
precipitated (either PEG300, NMP or a combination of the two). The
precipitation of the protein was classified as highly turbid (HT),
lowly turbid (LT, slight turbidity), or no change in turbidity (N).
The solutions were all formulated in a pH 4.7 150 mM acetate
buffer.
[0044] The apparent viscosity of the IgG suspensions was measured
as the time to draw 0.25 mL of the suspension into a 25 gauge 1.5''
needle attached to a 1 ml tuberculin slip tip syringe. Typical
times ranged from 5 to 100 s. Each measurement was made at least 3
times and averaged, while maintaining the suction force by holding
the end of the plunger at the 1 ml mark each time. A linear
correlation curve for viscosity and time to draw 1 mL (4 times the
amount measured) was constructed from measuring liquids of known
viscosity (PEG200, PEG 300, PEG 400, water, ethanol, olive oil, and
benzyl benzoate). This correlation, as expected from the
Hagen-Poiseuille equation, gives an r.sup.2 value greater than
0.999 and was reported previously. In most cases the
reproducibility in viscosity was within 5%. In our experiments a
maximum volume of 25% of the cavity in the syringe was filled with
suspension for the uptake. Consequently, the pressure drop was
relatively constant. The error introduced by the small change in
pressure drop was minimized by using the same plunger position each
time and by correlating the data to liquids of known viscosity.
[0045] Following the suspension viscosity measurement, the protein
suspensions were centrifuged for 20 min at 3000 rpm using a
rotating bucket centrifuge rotor (part A-4-62) with a 2 ml
centrifuge tube adapter for an Eppendorf Centrifuge (model 5810,
Wesbury, N.Y.). The centrifuged samples were photographed and the
supernatant was separated by carefully decanting the sample using a
needle and syringe. The remaining protein particles were then
redispersed in .about.15 ml of acetonitrile under gentle bath
sonication for 5 min. The dissolution of BSA in acetonitrile was
negligible. After redispersion, the protein particle size was then
measured for a drop of the sonicated dispersion. The dispersion was
diluted to an obscuration of approximately 10% in acetonitrile in a
small-volume (11 ml) magnetically stirred cell and the particle
size was analyzed by light scattering using the Malvern Instruments
Mastersizer S.
[0046] After the centrifugation described in the Particle Size
Measurement section, the recovered supernatant was then filtered
through a 0.22 .mu.m filter and collected. The filtered sample was
then diluted to a total volume of 0.7 ml in pH 4.7 150 mM acetate
buffer. Three 200 .mu.l aliquots of each sample were then placed on
a UV-transparent 96-well plate and imaged at 280 nm using a
spectrophotometer. A standard curve of BSA concentrations of 3, 2,
1, 0.5, and 0 mg/ml in the same buffer versus absorbance at 280 nm
yielded an r.sup.2 value greater than 0.99. The calibration curve
was used to regress the soluble concentration values for each
sample. If necessary, the sample was subsequently diluted and
remeasured till the concentration fell in the range of 0.5-3
mg/ml.
[0047] A drop of the aqueous suspension was flash frozen onto
frozen aluminum SEM stages maintained at -200.degree. C. with
liquid nitrogen. The frozen droplet was lyophilized with 12 hours
of primary drying at -40.degree. C. that was followed by a 6 hour
ramp to 25.degree. C. and secondary drying for at least 6 hours at
25.degree. C. using a VirTis Advantage Plus XL-70 shelf
lyophilizer. The lyophilization produced a dried powder sample on
the SEM stage. Dry powder samples of the milled particles were
placed on adhesive carbon tape. Each sample was then gold-palladium
sputter coated using a Cressington 208 bench top sputter coater to
a thickness of 15 nm. Micrographs were then taken using a Zeiss
Supra 40 VP scanning electron microscope with an accelerating
voltage of 5 kV.
[0048] Various additives can be added to decrease the solubility of
a protein in aqueous buffer. The turbidimetric studies of
solubility in Table 1 indicate that BSA precipitates with either
pure PEG and NMP or mixtures thereof even at low protein
concentration of 5 mg/ml. Table 2 indicates that PEG is a stronger
antisolvent than NMP as a low turbidity suspension (LT) is formed
at 30%. Furthermore, mixtures of the two antisolvents can produce
synergistic effects on precipitating protein. For example, a 20-20%
mixture causes high turbidity whereas 40% NMP does not produce a
change in turbidity. Finally for a given total weight % of
antisolvent, the turbidity increases as the relative fraction of
PEG increases. These experiments indicate that even a dilute 5
mg/ml protein concentration is well above the solubility limit with
these antisolvents. Therefore, only a small fraction of the protein
will be dissolved with these additives when the overall protein
concentration is on the order of 200 mg/ml, a typical concentration
for the injectable suspensions.
[0049] FIGS. 1A-C are SEM images of particles made by the process
of one embodiment of the present invention. FIG. 1A is a SEM image
of original milled particles. FIG. 1B is a SEM image of 30% PEG300
suspension at 200 mg/ml flash frozen and lyophilized. FIG. 1C is a
SEM image of 25% PEG300 20% ethanol suspension at 350 mg/ml flash
frozen and lyophilized.
[0050] The morphology of the original milled particles is shown in
FIG. 1A. The average particle size was 20 .mu.m. The size was
chosen to be small enough to pass through a 25-27 gauge needle with
an inside diameter of smaller than 241 .mu.m. A drop of the final
suspension was frozen and lyophilized, and SEM was utilized to
determine the particle size in the suspensions for two choices of
precipitants shown in FIGS. 1B and C. The particle sizes were only
modestly larger than that of the original milled particles
indicating little growth from particle aggregation or Ostwald
ripening. In addition, FIG. 1C shows a slight increase in the
amount of smaller sub-1 .mu.m particles. These results indicate
that the particle size could essentially be maintained in the
suspensions, and that dissolution of the protein was minimal.
[0051] The purpose of this section is to describe the mechanisms
for protein precipitation and to provide a brief overview of the
results for the viscosities for each class of precipitant. To form
a suspension of BSA in aqueous media, the media must be designed to
prevent a significant fraction of the BSA from dissolving. The
solubility of BSA in a pure aqueous buffer at pH of 4.5 is greater
than 100 mg/ml. Various additives were introduced to the aqueous
buffer media to decrease the solubility including salts, polymers,
and water soluble organics.
[0052] Table 1 is a table of the precipitation of BSA in PEG and
NMP mixtures at 5 mg/ml. (N indicates transparent solution with no
change in turbidity versus case without additives, LT indicates low
turbidity, HT indicates a high level of turbidity)
TABLE-US-00001 PEG % NMP % 0 5 10 15 20 25 30 35 40 45 50 0 N N N N
N N LT LT HT HT HT 5 N N N N N N LT LT HT HT 10 N N N N N LT HT HT
HT 15 N N N N LT HT HT HT 20 N N N N HT HT HT 25 N N N LT HT HT 30
N N N HT HT 35 N N HT HT 40 N LT HT 45 N HT 50 N
[0053] The mechanisms by which the solubility is lowered is given
for each type of precipitant in Table 2. Formation of an opaque
white concentrated suspension at 200 mg/ml was possible for all
three various additive groups as seen in Table 1. Table 2 is a
comparison of different additives to lower solubility of a protein,
suggested mechanism for the decrease in solubility and the
suspension viscosity measured at 200 mg/ml in 150 mM pH 4.7 acetate
buffer plus each additive.
TABLE-US-00002 Example of Suspension Viscosity for 200 mg/ml
Additive Mechanism to lower protein solubility milled BSA Salts
Competition for waters of hydration 3 cP for 1.5M Ion binding
changing protein-protein (NH.sub.4).sub.2SO.sub.4 interaction
Decrease in electrostatic repulsion with a decrease in the double
layer thickness Polymers Preferentially Preferential exclusion 18
cP 30% from protein surface leading to depletion PEG300 attraction
Lower solvent dielectric constant Water Soluble Lower solvent
dielectric constant 10 cP for Organics Exclusion of solvent from
protein 35% n- surface produces excluded volume methyl-2- effects
pyrrolidone (NMP)
[0054] A summary of select viscosity results measured with a 25 g
1.5'' needle, which are described in greater detail below, is also
presented in Table 2. An extremely low viscosity of 3 cp was
obtained with 1.5M (NH.sub.4).sub.2SO.sub.4. The values were also
quite low for the other PEG and NMP antisolvents. All three of
these examples are well within the limits of what would be
considered easily syringeable, since it would take less than 20
seconds to expel 1 ml from a 26 g needle. These results will be
examined in much greater detail below in context of the
morphologies of the particles and of the suspensions and for a much
wider range of conditions.
[0055] FIGS. 2A-2E are images of suspensions before (FIGS. 2A and
B) and after (FIGS. 2C-E) centrifugation (20 min at 3000 rpm) from
left to right 1.5M Ammonium Sulfate (a and c), 30% PEG300 (b and
d), 35% NMP (e) all in 150 mM pH 4.7 acetate buffer with added NaCl
to 154 mM ionic strength. As shown in FIG. 2, the various
solubility-decreasing additives produced different amounts of
suspended large particles. The relative quantity of suspended
particles is evident qualitatively from the turbidity of the
initial suspension. An opaque white suspension was formed for 30%
PEG300 and 35% NMP as shown in FIGS. 2B and 2E suggesting a large
amount of suspended particles relative to dissolved protein. This
observation is confirmed more quantitatively by the volume fraction
of the precipitant after 20 minutes of centrifugation at 3000 rpm,
indicating very little of the protein was soluble. This force is
sufficient to settle out all particles greater than .about.400 nm.
These results support the direct observation of micron-sized
protein particles in the suspensions by cryo-SEM in FIG. 1C. For
1.5M ammonium sulfate salt as an additive, the original suspension
was only translucent indicating a much smaller degree of
precipitation. After centrifugation, only a small volume of
precipitate was present (FIGS. 2A and C) consistent with the
observation of relatively low turbidity in the original suspension.
The degree of precipitation as characterized by the turbidity of
the initial suspension and the volume fraction of precipitant after
centrifugation will be an important factor for understanding the
viscosity behavior of the suspensions.
[0056] As shown in Table 2, an aqueous suspension with 30% PEG300
gives an apparent viscosity of 18 cP for 200 mg/ml BSA. This low
viscosity may be examined in terms of the presence of protein
particles relative to dissolved protein in solution. As has been
demonstrated previously, wth the addition of 30% PEG300, the
solubility of BSA is reduced from approximate 6 mg/ml at 25%, to 1
mg/ml. This large decrease in solubility was verified in Table 1 by
the increase in turbidity for a level of precipitant between 25%
and 30% (by volume) PEG300 at a protein concentration of 5
mg/ml.
[0057] Suspensions of the pure milled BSA particles were formed for
PEG300 levels between 30 and 50% by volume. Two key factors
influence the viscosity of the suspension at a given volume
fraction of protein: the initial viscosity of the solution without
protein and the intrinsic viscosity with protein present. Table 3
shows the Intrinsic Viscosity data of suspensions to compare
effects of electroviscous and hydration, all samples at pH 4.7
acetate 150 mM ionic strength buffer (unless indicated otherwise).
For intrinsic viscosity measurement, .phi..sub.max assumed to be
0.64. (NMP--N-methyl-2-pyrrolidone, ND--not determined) The average
particle diameter was determined by static light scattering was 20
.mu.m.
TABLE-US-00003 Suspension Apparent Solution Viscosity at 250 mg/ml
Intrinsic Average Soluble Viscosity BSA Viscosity Particle
Concentration Solvent system (cP) (cP) [.eta.] Diameter (.mu.m)
(mg/ml) 30% PEG300 2.6 56 11.4 9.6 ND 40% PEG300 4.6 85 9.6 17.4
2.0 50% PEG300 6.6 49 7.9 18.5 2.1 pH 5.5 50% PEG300 6.2 47 8.4
18.4 ND pH 7.4 50% PEG300 6.0 57 9.0 ND ND 40% PEG300 and 10% 5.1
22 6.0 20.2 2.9 NMP 35% PEG300 and 15% 4.8 20 5.9 19.1 2.9 NMP 25%
PEG300 and 20% 3.6 18 6.5 19.5 2.3 NMP 25% PEG300 and 20% 3.3 14
5.8 7.4 4.8 Ethanol
[0058] As shown in Table 3, the solution viscosity (without
protein) increases from 2.6 for a 30% PEG300 solution to 6.6 for a
50% PEG300 solution. At the same time however, the intrinsic
viscosity decreased from 11.4 to 7.9. Consequently the overall
suspension viscosity went down about 10%, a value twice that of the
experimental uncertainty. Interestingly, the viscosity was higher
at 40% PEG than 30%, as the intrinsic viscosity did not decrease
enough to compensate for the higher solvent viscosity (without
protein). Thus, the amount of precipitant may be optimized to
balance its effect on the initial viscosity of the solution without
protein versus the intrinsic viscosity of the protein
suspension.
[0059] FIG. 3 is an image of aqueous PEG 300 suspensions after 20
min of centrifugation at 3000 g. From left to right 30% PEG300, 40%
PEG300, and 50% PEG 300 all in 150 mM pH 4.7 acetate buffer with
added NaCl to 154 mM ionic strength. All 3 of the PEG-based
suspensions after 20 min of centrifugation at 3000 g exhibited a
high degree of large settled particles (FIG. 3). However, for the
30% PEG300 suspension, the supernatant appears slightly turbid,
indicating the presence of small suspended nanoparticles. Thus, a
soluble concentration was not determined for this case. The soluble
concentration for the 40 and 50% PEG300 suspensions was between
2.0-2.1 mg/ml indicating that 99% of the particles at 200 mg/ml
were suspended (Table 3). The average particle diameters for the 40
and 50% PEG300 suspensions were both near the original 20 .mu.m
value for the milled particles, as shown in table 3. However, for
the 30% PEG300 suspension, the average particle diameter decreased
to .about.10 .mu.m consistent with greater dissolution on the basis
of the turbidity of the supernatant. This decrease in size is
consistent with the observation by SEM in FIG. 1. These smaller
particles with higher surface area, and dissolved protein
contributed to the higher intrinsic viscosity in Table 3.
[0060] For the 30% PEG300 formulation, additional apparent
suspension viscosities were measured at higher ionic strengths,
shown in Table 4. As the ionic strength of the solution increased,
the viscosity decreased from 56 to 19. The decrease in the
electroviscous effects with an increase in ionic strength and a
decrease in the Debye length contributes significantly to the
decrease in intrinsic viscosity and thus the suspension
viscosity.
TABLE-US-00004 Ionic strength (mM) Apparent Viscosity (cP) 150 56
300 23 500 19
[0061] To change the charge of the protein and the solubility, the
solution pH was increased from the pI of BSA (4.7) to a pH of 5.5,
while maintaining the acetate ion as the buffer. In addition, a
phosphate buffer ion was used as opposed to acetate buffer ion to
raise the pH of 7.4. Since the solubility of a protein increases as
the pH moves away from the pI, the 50% PEG300 additive was included
in the media to ensure a low solubility of BSA.
[0062] FIG. 4 is a graph of the apparent viscosity of aqueous
suspensions at various pHs all with 50% PEG 300 and added NaCl to
154 mM ionic strength. As shown in FIG. 4, neither the buffer ion
nor the pH made a significant change in the apparent suspension
viscosity at any concentration. The solution viscosities for the
varying pHs with 50% PEG300 varied only slightly from 6.0-6.6.
Since the apparent viscosities for all cases at all concentrations
were similar, the calculated intrinsic viscosities increased only
slightly from 7.9 to 9.0 for the increase in pH. The theoretical
curves in FIG. 4 were determined by regressing the intrinsic
viscosity (Table 3) with the Krieger-Dougherty equation.
[0063] FIG. 5 is an image of aqueous 50% PEG 300 suspensions after
20 min. of centrifugation at 3000 g. From left to right pH 4.7
acetate buffer, pH 5.5 acetate buffer, and pH 7.4 acetate buffer.
As shown in FIG. 5, the supernatant is still turbid for both the pH
5.5 and pH 7.4 samples after centrifugation. As a result, the
soluble concentration could not be determined and the average
particle size could only be determined for the pH 5.5 sample. While
the average particle diameter at pH 5.5 is very similar to the
average particle diameter at pH 4.7, the increase in turbidity of
the supernatant indicates the presence of some additional smaller
nanoparticles, not present in the pH 4.7 samples.
[0064] FIG. 6 is a graph of the apparent viscosity of aqueous
suspensions with PEG 300 and organic additives in 150 mM pH 4.7
acetate buffer with added NaCl to 154 mM ionic strength. As shown
in FIG. 6, upon addition of 10-20% of an organic additive to at
least 25% PEG300, the apparent viscosity of the suspension remained
below 50 cP even for extremely high BSA concentrations greater than
300 mg/ml. Furthermore, as shown in table 3, all of these PEG300
plus organic suspensions reduce apparent viscosities to a range of
14-22 cP at a BSA concentration of 250 mg/ml. This viscosity range
is approximately the same as for the higher ionic strength samples
(300 and 500 mM) discussed above (Table 4). These PEG300-organic
samples gave the lowest intrinsic viscosity values (5.8-6.5)
measured in this study (Table 3). All three samples with 10-20% NMP
added, gave particle diameters that were very close to the original
20 .mu.m milled particles. In addition, the soluble concentration
for all three NMP samples was between 2.3-2.9 mg/ml, indicating the
suspension of greater than 99% of the BSA added at the 300 mg/ml
level (Table 3). The sample with 20% ethanol added had the lowest
solution viscosity, giving the lowest suspension viscosity at 250
mg/ml of 14 cP (Table 3). The slightly higher soluble concentration
of 4.8 mg/ml indicated that greater than 98% of the BSA added was
suspended. However, this slight decrease in percent suspended was
sufficient to decrease the average particle diameter to 7.4 .mu.m.
The slight decrease in particle size could be seen in the SEM
imaged.
[0065] All of the above samples contained milled model protein,
BSA, without a sugar lyoprotectant that can help stabilize a
therapeutic protein at the molecular level against monomer
aggregation. BSA is a fairly stable protein and thus the molecular
stability was not considered in this study. However, milling can
generate instabilities for many therapeutic proteins. {Maa} On the
other hand, freezing and lyophilizing of protein particles has been
shown to produce molecularly stable protein particles even for
submicron sizes. A very common lyoprotectant is trehalose.
[0066] FIGS. 7A and 7B are graphs that show the apparent
viscosities of milled BSA and trehalose at various ratios in 150 mM
pH 4.7 acetate buffer along with the theoretical viscosity as
calculated from the Krieger-Dougherty equation using the [.eta.] of
the pure milled particles in a) 25% PEG 300 20% Ethanol and b) 35%
PEG 300 and 15% NMP. FIG. 7 shows images of data of protein and BSA
particles at ratios from 1:0.1 to 1:1 by weight BSA to trehalose.
At all points, the higher ratios of trehalose to BSA increase
monotonically the viscosity of a suspension formed in either 25%
PEG300 20% ethanol or 35% PEG300 and 15% NMP. Even though the
trehalose will be soluble in the aqueous solvents with the
additives, it will still contribute additional excluded volume that
is unavailable either to the solvent or the undissolved particles.
This excluded volume will therefore decrease the highest loading of
particles possible to maintain a given viscosity. The theoretical
lines in FIG. 7 were calculated using the same intrinsic viscosity
for the respective suspensions without trehalose and assuming a
maximum volume fraction of 0.64.
[0067] Antibody therapeutics currently constitute a market size of
$15 billion dollars annually, addressing needs in anti-cancer,
anti-infective and anti-inflammatory diseases. As large protein
molecules, these currently require direct injection (intra-venous
or subcutaneous) for delivery. Since the required doses are quite
large and frequently administered, this poses a considerable
obstacle for drug delivery: to achieve subcutaneous delivery with a
small bore syringe, the antibody must be formulated in a high
concentration (>100 mg/ml), low volume (<1.5 ml), low
viscosity (<100 cP) format. These specifications have been very
difficult to achieve with traditional solution formulations (in a
phosphate buffer containing trehalose) but suspensions of
lyophilized antibody in aqueous solvents containing salts and
precipitants to prevent dissolution represent a possible option. We
have previously demonstrated the success of this approach with a
model protein, BSA and here present an extension of the technology
to polyclonal antibodies. Here we present for the first time,
evidence that immunoglublins can be prepared as aqueous suspensions
in concentrations up to 200 mg/ml, with low viscosity and no
aggregation (96% monomeric protein).
[0068] IgG purified from sheep serum (Product No. 15131) was
purchased from Sigma-Aldrich, Inc. .alpha.-.alpha. trehalose,
polyethylene glycol with an average molecular weight of 300 (PEG
300), ammonium sulfate, USP grade ethanol, and n-methyl
2-pyrrolidone (NMP) were purchased from FisherChemicals.
[0069] After dissolving a known amount of the IgG in an appropriate
amount of a 20 mM pH 5.5 histidine buffer with .alpha.-.alpha.
trehalose, samples were slow frozen on a pre-cooled lyophilizer
shelf at -40.degree. C. The samples were then lyophilized for 12
hours at -40.degree. C. at 100 mTorr, followed by a 6 hour ramp to
25.degree. C. at 50 mTorr, and maintained for secondary drying at
25.degree. C. at 50 mTorr for at least an additional 6 hours. 1 mg
of powder was then weighed out and reconstituted at 1 mg/ml in 200
mM pH 7.0 phosphate buffer for stability analysis by size exclusion
chromatography as described below.
[0070] An aliquot of a 5 mg/ml IgG solution was mixed with an equal
volume of a second aqueous solution containing additives for the
purpose of lowering the protein solubility. The precipitation of
the protein was characterized by an increase in turbidity of the
solution after 24 hours at a wavelength of 350 nm. The solutions
were all formulated in a pH 6.4 20 mM histidine buffer with added
PEG300 and NMP. The turbidity of a 100 .mu.l aliquot of the final
formulation was measured on a UV-transparent 96-well plate using a
spectrophotometer.
[0071] Aqueous-based solvent mixtures, without protein, containing
varying volume percents of NMP, PEG 300, and ethanol and various
molarities of sodium chloride or ammonium sulfate salts were mixed
to form uniform transparent solutions. These solutions were
buffered at varying ionic strengths usings either a histidine or
phosphate buffer at pH 6.4 (the isoelectric point of sheep IgG).
Samples of the protein were then compacted into 0.1 ml conical
vials such that the powder weight was within 5% of the desired
weight. The powder weight depended upon the final protein
concentration and the excpient/protein ratio. A measured amount of
the prepared aqueous-based solvent mixture was added to the conical
vial to form a suspension with a total volume of 0.1 ml. The
mixture was stirred the tip of a needle to remove air pockets and
to form the suspension with sufficient uniformity. Sonication was
not used, nor was it needed. A drop of the uniform suspension was
then placed on a microscope slide to image the suspension and 10
.mu.l of the suspension was diluted to 1 mg/ml in a 200 mM pH 7.0
phosphate buffer to measure the protein monomer fraction by size
exclusion chromatography as specified below.
[0072] The apparent viscosity of the IgG suspensions was measured
as the time to draw 50 .mu.l of the suspension into a 25 gauge
1.5'' needle attached to a 1 ml tuberculin slip tip syringe. A
conically shaped vial was used to minimize the sample volume given
the cost of the protein. Videos of the conical vial containing the
suspension were taken and the time to draw from a height 0.4'' from
the bottom of the cone to a height 0.1'' from the bottom of the
cone was measured using Image J software. The uncertainty in the
height was on the order of 1%. The time was measured to within 0.05
seconds as the video was converted to an image stack with 20 images
per second. Each measurement was made at least 3 times and
averaged, while maintaining the suction force constant for each
measurement by holding the end of the plunger at the 1 ml mark each
time. In most cases the reproducibility in viscosity was 10%.
Previous work with suspensions of model proteins found that the
time to draw up a specified amount of the sample in a syringe was
correlated linearly to viscosity. A maximum volume of 10% of the
cavity in the syringe was filled with suspension for the uptake.
Consequently, the pressure was essentially constant and the
viscosity may be obtained from the Poiseulle equation. In this
case, using standard solutions with various viscosities (pure DI
water, ethanol, PEG 200, PEG300, PEG400 and benzyl benzoate) gave a
linear correlation between the time to draw 0.05 ml from the
conical vial to the viscosity with an r.sup.2 value greater than
0.99.
[0073] Samples for scanning electron microscopy (SEM) of the dry
powders after lyophilization were placed on adhesive carbon tape.
Each sample was then gold-palladium sputter coated using a
Cressington 208 bench top sputter coater to a thickness of 15 nm.
Micrographs were then taken using a Zeiss Supra 40 VP scanning
electron microscope with an accelerating voltage of 5 kV. Optical
microscope images of a drop of the final suspensions on a glass
microscope slide were taken using an MTI CCD 72 (Dage-MTI, Michigan
City, Ind.) camera attached to a Nikon Optiphot2-Pol (Nikon
Instruments Inc. Melville, N.Y.) microscope.
[0074] Percent monomer of the initial solution, reconstituted
powder and final diluted suspension was analyzed by using Tosoh
Biosciences G3000SWXL size exclusion column followed by a G2000SWXL
size exclusion column attached to Waters Breeze HPLC system
containing a model 717plus autosampler, 2487 dual wavelength
detector, and 1525 binary pump (Waters Corporation, Milford,
Mass.). The prepared samples, reconstituted or diluted to .about.1
mg/ml in 200 mM pH 7.0 phosphate buffer were filtered through a
0.22 .mu.m Millex-GV filter to remove large aggregates prior to
analysis. The mobile phase consisted of a pH 7.0 200 mM phosphate
buffer and 50 mM sodium chloride at a flow rate of 0.7 ml/min. The
detection wavelength was 214 nm. An injection volume of 20 .mu.l of
the .about.1 mg/ml prepared sample was used. The monomer eluted at
approximately 21.5 minutes, with the higher molecular weight
aggregates eluting in the last few minutes before this, depending
on their size.
[0075] FIGS. 8A-8E are SEM images of various frozen powders of TgG.
FIG. 8A is a SEM image of 40 mg/ml IgG frozen at a 1:1 IgG to
trehalose ratio. FIG. 8B is a SEM image of 55 mg/ml IgG frozen at a
1:0.5 IgG to trehalose ratio. FIG. 8C is a SEM image of 25 mg/ml
IgG frozen at a 1:0.5 IgG to trehalose ratio. FIG. 8D is a SEM
image of 40 mg/ml IgG no trehalose. FIG. 8E is a SEM image of 20
mg/ml IgG at a 1:1 IgG to trehalose ratio. Large micron-sized
particles of IgG stabilized by .alpha.-.alpha. trehalose were made
by lyophilization using a 1:1, 0.5:1, 0.25:1, or 0:1 ratio of
trehalose to IgG in a 20 mM pH 5.5 histidine buffer at various
initial concentrations between 20-80 mg/ml of IgG. SEM micrographs
of the final dried powder show large 10-100 .mu.m particles with
relatively few fine particles (on the order of hundreds of
nanometers) for the particles frozen at higher concentrations (40
to 80 mg/ml) for each trehalose to IgG ratio (FIGS. 8A and 8B),
including the case with no trehalose (FIG. 8D). The large particles
are in contrast to the smaller web-like morphology visible for the
protein frozen at lower concentrations, 20 and 25 mg/ml IgG, with
high ratios of IgG to trehalose, 1:1 and 1:0.5 respectively (FIGS.
8C and 8E). During freezing, a higher concentration of protein
leads to greater growth and thus larger final particles.
[0076] The relative stability from SEC was defined as the
difference in percent area of the monomer peak after reconstitution
of the dry powder in pH 7.0 phosphate buffer relative to the
initial powder diluted in the same pH 7.0 phosphate buffer. This
relative stability was at least 98.6% and often higher. The
stability was high even for the 40 mg/ml IgG powder frozen without
any trehalose, indicating cryoprotectant is not needed to achieve
high stabilities as measured by this technique. However this value
of 98.6 is lower than that of all of the other examples in the
table that included trehalose. Thus, a cryoprotectant can be
beneficial for increasing the stability, and trehaolse was
included.
[0077] Precipitation of 5 mg/ml IgG solution with various
additives. Various additives can be added to decrease the
solubility of the IgG as described in detail above. We have
confirmed this by observation of precipitation in a high molarity
(1.5M) ammonium sulfate solution at an IgG concentration of 5 mg/ml
(optical density not determined).
[0078] FIG. 9 is a graph of the optical density of the IgG with
various additives totaling 50% of the solvent by volume to decrease
the solubility of the IgG as measured at 350 nm. The right hand
side has the absorbance of the 5 mg/ml concentration of IgG in the
pH 6.4 20 mM histidine buffer with no additional additive. As shown
in FIG. 9, at a concentration of 5 mg/ml IgG, the absorbance at 350
nm increases from .about.0.05 for the pure protein solution at pH
6.4 to .about.0.6 for a 50% volume solution of PEG300 at pH 6.4
(FIG. 9). For the IgG at this concentration in a 50% volume
solution of NMP at pH 6.4, the absorbance at 350 nm was
significantly lower at .about.0.2 than for the case of 50% PEG300
(FIG. 9). The absorbance at 350 nm of mixed samples of PEG300 and
NMP totally 50% by volume of the solvent, with at least 25% PEG300
were similar. The absorbance decreased slightly as the % NMP
increased from .about.0.6 to 0.5 for a 25% PEG300 and 25% NMP mixed
solution. However, a much lower absorbance of .about.0.2 was
observed for a 50% NMP solution, without any PEG. These experiments
indicate that proteins precipitate with these additives even at low
protein concentration of 5 mg/ml. Therefore, only a small fraction
of the protein will be dissolved with these additives when the
overall protein concentration is on the order of 200 mg/ml, a
typical concentration for the injectable suspensions.
[0079] Suspension morphology as a function of particle size. In
addition to the ratio of trehalose to protein in the lyophilized
particles, the size and surface area of the particles, will vary
the morphology and viscosity (syringcability) of the suspension.
The optimum particle size contains particles small enough to flow
up the 25 gauge needle however large enough to minimize the
detrimental effects of hydration and electroviscous forces on the
viscosity. Furthermore, a decrease in the surface area of the
particles may decrease denaturation and aggregation of the
protein.
[0080] FIG. 10A-10D are images of various suspensions of IgG. A)
200 mg/ml IgG suspension made of 55 mg/ml IgG 1:0.5 IgG to
trehalose ratio particles in 20 mM pH 6.4 histidine buffer with
1.5M added ammonium sulfate salt. B) 200 mg/ml IgG suspension made
of 55 mg/ml IgG 1:0.5 IgG to trehalose ratio particles in 50 mM pH
6.4 phosphate buffer with 50% PEG300. C) 200 mg/ml IgG suspension
made of 55 mg/ml IgG 1:0.5 IgG to trehalose ratio particles in 50
mM pH 6.4 phosphate buffer with 35% PEG300 and 15% NMP by volume
after 24 hours D) 200 mg/ml IgG suspension made of 20 mg/ml IgG
with 1:1 IgG to trehalose ratio particles in 50 mM pH 6.4 phosphate
buffer with 35% PEG300 and 15% NMP by volume.
[0081] In FIG. 10A and FIG. 10B, a concentrated suspension is shown
for two different additives and an IgG concentration of 200 mg/mL
initially after forming the suspension. The suspensions were white
and opaque. The path length was approximately 0.5 cm at the
mid-point of the cone. The particles in a droplet of the suspension
were further characterized with optical microscopy in FIG. 4.
Micron-sized particles are present in the range of a few microns to
10 micron, consistent with the dry initial particles from the SEMs
in FIG. 8. FIG. 10C illustrates a typical example of a small degree
of settling of these suspensions in FIG. 10A and FIG. 10B after 24
hours. In FIG. 10D, the initial concentration for lyophilization
was much lower, 20 mg/ml, and the particles were much smaller as
evident in SEM and by optical microscopy of the suspension. These
smaller particles did not scatter light as strongly, and the
suspension appeared translucent, instead of white and opaque.
[0082] FIG. 11A-11E are microscope images of various suspensions of
IgG. FIG. 11A is an image of 200 mg/ml IgG suspension made of 55
mg/ml IgG 1:0.5 IgG to trehalose ratio particles in 20 mM pH 6.4
histidine buffer with 1.5M added ammonium sulfate salt. FIG. 11B is
an image of 200 mg/ml IgG suspension made of 55 mg/ml IgG 1:0.5 IgG
to trehalose ratio particles in 50 mM pH 6.4 phosphate buffer 50%
PEG300. FIG. 11C is an image of 200 mg/ml IgG suspension made of 55
mg/ml IgG 1:0.5 IgG to trehalose ratio particles in 50 mM pH 6.4
phosphate buffer 35% PEG300 15% NMP. FIG. 11D is an image of 200
mg/ml IgG suspension made of 20 mg/ml IgG 1:1 IgG to trehalose
ratio particles in 50 mM pH 6.4 phosphate buffer 35% PEG300 15%
NMP. FIG. 11E is an image of 200 mg/ml IgG suspension made of 80
mg/ml IgG 1:1 IgG to trehalose ratio particles in 50 mM pH 6.4
phosphate buffer 35% PEG300 15% NMP. Table 5 illustrates IgG
lyophilized powders made at various protein concentrations and
trehalose ratios in a 20 mM pH 5.5 histidine buffer, characterized
for the stability of the dry powder by size-exclusion HPLC.
TABLE-US-00005 Protein concentration Trehalose:IgG SEC (% (mg/ml)
(wt.) ratio monomer of original solution) 20 1:1 99.1 25 0.5:1
100.1 40 0 98.6 40 1:1 101.6 55 0.5:1 99.8 65 0.25:1 99.9 80 0
100.1
[0083] Table 6 illustrates 200 mg/ml IgG suspensions in a solvent
containing 35% PEG300, 15% N-methyl-2-pyrrolidone (NMP) by volume
added to a 50 mM pH 6.4 phosphate buffer. (ND--not determined;
NM--immeasurable).
TABLE-US-00006 Frozen IgG Concentration Trehalose:IgG (wt.)
Viscosity SEC (% monomer (mg/ml) ratio in frozen powder (cP) of dry
powder) 20 mg/ml IgG 1:1 ND 102.6 40 mg/ml IgG 0 52 79.7 40 mg/ml
IgG 1:1 194 98.0 55 mg/ml IgG 0.5:1 104 97.1 65 mg/ml IgG 0.25:1
144 86.9 80 mg/ml IgG 0 NM ND
[0084] The largest particles were formed with pure IgG particles
frozen at 80 mg/ml as shown in FIG. 11E. The large particle size
was caused by the high starting concentration during
lyophilization. They were suspended in the 35% PEG300 15% NMP
solvent described in Table 2. The particle size reached >50
micron, and therefore, the particles did not flow through a 25
gauge syringe. In contrast, all of the smaller particles in FIG. 11
were syringeable.
[0085] Table 8 illustrates IgG suspensions in various buffers with
various additives screened for their viscosity and % monomer of the
original sample present. (ND--not determined; EtOH--ethanol,
NMP--N-methyl-2-pyrrolidone).
TABLE-US-00007 IgG Trehalose:IgG SEC (% Suspension Frozen IgG (wt.)
ratio monomer Suspension Suspension Concentration Concentraion in
frozen Viscosity of dry Additive Buffer (mg/ml) (mg/ml) powder (cP)
powder) 50% PEG300 20 mM pH 7.4 100 40 .sup. 1:1 46 ND histidine
buffer 50% PEG300 50 mM pH 6.4 170 55 0.5:1 72 102.0 phosphate
buffer 50% PEG300 50 mM pH 6.4 200 55 0.5:1 78 ND 90 mM
(NH.sub.4).sub.2SO.sub.4 phosphate buffer 40% PEG300 40 mM pH 7.4
100 40 .sup. 1:1 43 ND 10% EtOH histidine buffer 35% PEG300 50 mM
pH 6.4 200 55 0.5:1 92 93.6 15% EtOH phosphate buffer 30% NMP 20 mM
pH 7.4 200 40 .sup. 1:1 71 ND 10% PEG300 histidine buffer 1.5M
(NH.sub.4).sub.2SO.sub.4 20 mM pH 6.4 200 55 0.5:1 12 97.8
histidine buffer 1.5M (NH.sub.4).sub.2SO.sub.4 20 mM pH 6.4 300 55
0.5:1 99 101.2 histidine buffer
[0086] Viscosities less than 100 cP are sufficient for
syingeability through a 25 gauge 1.5'' syringe. Syringeable
viscosities were obtained for suspensions at concentrations up to
200 mg/ml IgG (Tables 6-8). The addition of a cryoprotectant, such
as trehalose, as seen previously for BSA powders {Miller aqueous
BSA}, will increase the viscosity of a suspension (Table 6). This
increase is caused primarily by the excluded volume occupied by the
cryoprotectant. For example, in Table 6, the 200 mg/ml IgG
suspension, with no trehalose has a viscosity of 52 cP as opposed
to 104 for a 0.5:1 ratio of trehalose to IgG. The addition of
trehalose increases the total solute (trehalose plus IgG)
concentration to 300 mg/ml. Further increasing the total solute
concentration to 400 mg/ml by increasing to a 1:1 ratio of IgG to
trehalose, raises the viscosity further to 194 cP. Thus, the
potential need for a cryoprotcctant to form stable protein
molecules must be balanced against the increase in viscosity due to
the excluded volume of a cryoprotectant. To examine the
relationship between the precipitation seen in the solubility study
above and the viscosity of the suspension a series of tests were
run using the 55 mg/ml IgG 0.5:1 trehalose:IgG particles. For the
additive conditions in the first three entries in Table 7, high ODs
were obtained at even the low protein concentration of 5 mg/ml in
the solubility determinations in FIG. 9. The vials for these 150
mg/ml protein suspensions were white opaque as in FIG. 3. These
samples had measurable viscosities between 57 and 98 cP. For the
additive compositions in the last two entries with only 15 or 0%
PEG, and the remainder NMP, the OD was much lower for the
precipitation studies at 5 mg/ml protein, indicating higher protein
solubility. For these additive compositions and 150 mg/ml
suspensions, viscosities were either very high, 287 cP, or not
measurable, as a paste-like gel was formed. Thus, the additive
compositions that cause significant protein precipitation at 5
mg/ml in FIG. 9, also are beneficial for producing lower
viscosities. As the ratio of dissolved protein to micron-sized
protein protein particles goes down, the viscosity is decreased.
This decrease may be attributed to a reduction in solvation and
electrosviscous forces, although further characterization would be
needed to more fully describe the mechanism.
[0087] In Table 8, miscellaneous additive conditions beyond those
in Tables 6 and 7. The aqueous solutions contained pure salt, pure
PEG300, and mixtures of salts, PEG300 and water-soluble organic
additives up to a total of 50% of the solution (Table 8).
Syringeable aqueous-based IgG suspensions were obtained in all of
these cases with a 25 G 1.5'' needle. In each of these cases,
micron-sized particles were formed given the IgG freezing
concentrations of 40-55 mg/ml. For all the suspensions in this
table, the presence of micron-sized particles was confirmed with
the optical microscope (select samples shown in FIG. 11). The high
molarity salt (1.5M) additive gave the lowest viscosity of 12.2 cP
at an IgG concentration of 200 mg/ml (Table 8). At an IgG
concentration of 200 mg/ml with trehalose, the next lowest
viscosity was for the 50% PEG300 sample at 79 cP (Table 8). Adding
15% NMP and decreasing the PEG300 to 35% at a pH of 6.4, increases
the suspension viscosity up to 104 cP (Table 8). A different
organic additive with a similar dielectric constant, ethanol, at
the same 15% level, gave a slightly lower viscosity at 92 cP (Table
4). As previously demonstrated for BSA, the various additive
compositions that decrease the solubility of a protein below 5
mg/ml can often give highly precipitated suspensions and
viscosities less than 100 cP at IgG concentrations greater than 100
mg/ml. Certain additive compositions allow such viscosities up to
200 mg/ml protein, and it may be expected that even higher protein
concentrations may be achieved by further optimization.
[0088] As mentioned in Table 5, powders containing protein highly
stable against monomer aggregation were achieved with or even
without trehalose as a cryoprotectant. We also examined the protein
stability after forming the suspensions. For the final suspensions,
10 .mu.l of the suspension were diluted in pH 7.0 phosphate buffer
necessary to give a final TgG concentration of .about.1 mg/ml. The
relative stability was measured as the difference between the %
monomer after redilution when compared to the % monomer for the
initial lyophilized particles upon reconstitution. For all of the
examples in Tables 6-8, where trehalose was present at 0.5:1 or
higher and without organic solvent (NMP or ethanol), the monomer
was high, at least 97%. The % monomer was >100% in some cases,
either because of experimental error or an actual increase in the
monomer fraction relative to the as received starting bulk
material. The high stabilities for systems with high PEG levels was
not unexpected as PEG is known to help maintain the thermal
stability of a protein. {Stevenson} High stabilities are shown in
the last two rows of Table 4 for the two cases with a high salt
concentration (1.5M ammonium sulfate). The high salt concentration
produces very low protein solubilities and favors the presence of
the micron-sized protein particles.
[0089] For the two studies without trehalose in Table 6 the protein
aggregation was significant. After the powder with no trehalose was
suspended, SEC shows approximately a 20% loss in % monomer to small
aggregates (row 2 in Table 6). A smaller decrease in the % monomer
of .about.15% was seen for particles with an insufficient ratio of
trehalose (0.25:1 trehalose to IgG ratio) (Table 6). Thus trehalose
at a ratio of 0.5:1 of trehalose to IgG was required to maintain
stability of the protein in the suspension, despite the fact that
trehalose had a small effect on stability for the initial powers in
Table 5.
[0090] The behavior is more complicated for the systems containing
NMP and PEG. As shown in the first thre rows in Table 7, the
protein stability decreases with an increase in NMP concentration
for a constant overall additive concentration of 50%. However, at a
higher 1:1 IgG to trehalose ratio with the 35% PEG 300, and 15%
NMP, the initial monomer was 98%. Thus, the higher ratio IgG to
trehalose may compensate for the higher degree of organic additive
(NMP) to maintain the stability.
[0091] The protein stability for the suspensions may be compared
with those for protein in buffer without the addition of agents to
lower the solubility. The power formed at 55 mg/ml initial protein
(Table 5) was added to pH 6.4 buffer without any other additives.
The resulting mixture was translucent and less turbid than in any
of the entries in Tables 5-8. Upon centrifugation at 16,100 g, a
precipitate was formed with a volume less than 10% of the total
protein volume. Thus, most of the protein was dissolved. The %
monomer was 70% for the same procedure as for the suspensions
above. In contrast, the monomer % was 97.1 for the same powder (55
mg/ml) at the 200 mg/ml level for an opaque white suspension of
micron sized particles in Table 6. Thus, micron-sized particles of
protein can be far more stable than proteins primarily in the
dissolved state at high concentrations.
[0092] Stable as measured by SEC, highly concentrated aqueous based
suspensions of a model IgG were created using particles that were
frozen and lyophilized at high concentrations (40-55 mg/ml). This
concentration range (40-55 mg/ml) made particles with a diameter of
.about.10-100 .mu.m that were found to be greater than 98% stable
by SEC. By stabilizing the particles with trehalose at a minimum
ratio of 0.5:1 trehalose to IgG, the final suspensions were also
found to preserve at least 92% of the original monomer percent. The
solubility of the IgG was lowered to less than 5 mg/ml in the
aqueous-based solvent by adding high salt (1.5 M ammonium sulfate),
PEG300 (50% of solvent by volume), or a combination of PEG300 and
ethanol or NMP (total 50% of solvent by volume). The apparent
viscosity through a 25 g 1.5'' syringe of the high salt suspension,
where the solvent viscosity is still .about.1 cP, was the lowest at
approximately 12 cP for a 200 mg/ml stable (by SEC) IgG suspension.
Overall, the stability of the model IgG and the low viscosities
(less than 100 cP) through a 25 g 1.5'' needle obtained for highly
concentrated suspensions indicates a potential advancement in the
subcutaneous delivery of protein therapeutics.
[0093] The delivery of concentrated proteins and peptides in the
range of 100 to 400 mg/ml by subcutaneous injection through a 25 to
27-gauge needle becomes feasible for a stable solution or
suspension with a viscosity below about 50 cP. Viscosities below
this limit were achieved for suspensions of milled lysozyme
microparticles in benzyl benzoate or benzyl benzoate mixtures with
vegetable oils for up to 400 mg/ml protein. The protein molecules
were stable against aggregation for at least 2 months and the solid
particles in suspension were resuspendable after being stored at
room temperature for a year. Correlations between the viscosity of
the suspension and the volume fraction of particles indicate that
the main source of interaction between the particles was simply due
to the high concentration of particles with little effect from
additional forces such as electrostatic repulsion, solvation of the
particles or deviations of the particle shape from a spherical
geometry. In contrast these additional forces can cause large
increases in viscosities for colloidal protein molecules in aqueous
solutions. Thus the lower solvent viscosities for highly
concentrated protein suspensions relative to protein solutions may
offer novel opportunities for subcutaneous injection.
[0094] As used herein, "stable proteins" refer to proteins that do
not show instabilities such as denaturation or aggregation of the
individual protein molecules in the dissolved state. These
instabilities can be measured by techniques such as optical
turbidity, dynamic light scattering, size-exclusion chromatography,
analytical ultracentrifugation, and a protein dependant activity
assays.
[0095] Solvents for use with the present invention include those in
which the non-aqueous suspensions that produced stable particles
not change in particle size over 2 months of storage with a protein
solubility of less than 0.03 mg/ml. The solvent must also not cause
an adverse affect on the stability of the protein particles.
Stability of protein particles can be obtained for a non-aqueous
solvent where the absorption of water into the particles in the
suspensions is approximately equal to the absorption of water into
particles exposed to ambient air conditions at the same relative
humidity. In addition, a non-aqueous solvent should have a low
dielectric constant (less than 37.5) to prevent attractive forces
between the particles to cause caking of the particles at the
bottom of the container causing the suspension to be unable to be
resuspended.
[0096] The objective of this study was to produce a highly
concentrated protein suspension for delivery of high dosages of
monoclonal antibody by subcutaneous injection via a 25- to 27-gauge
syringe. Suspensions of lysozyme particles smaller than 37 .mu.m
with concentrations from 50 to 300 mg/ml were formulated with a
50/50 volume mixture of the pharmaceutically acceptable solvents
safflower oil and benzyl benzoate. This solvent mixture was within
the approved range published on the FDA's Inactive Ingredients
List..sup.18 While benzyl benzoate is less viscous than safflower
oil, which facilitates formation and injection of the dispersions,
it is not currently accepted for injection by the FDA as a pure
solvent; however, indications of lack of toxicity suggest it may be
approved in the future. The apparent viscosities for formulations
in the approved solvent mixture as well as pure benzyl benzoate
will be shown to be in an acceptable range up to a concentration of
at least 300 mg/ml and to correlate with theoretical viscosities of
suspensions according to the Krieger-Dougherty equation. The
necessity to obtain a uniform dose is addressed through
measurements of the settling rate and confirmed by concentration
measurements of aliquots of the suspension. Colloidal stability of
the particles and the stability of the protein molecules are
addressed by measurements of particle and protein aggregation over
time and are further confirmed by analyzing the moisture content in
the various suspensions.
[0097] Lysozyme in lyophilized powder form was purchased from Sigma
Chemical Company (St. Louis, Mo.). ACS grade acetonitrile and USP
grade ethanol were used as received from Fisher Chemicals
(Fairlawn, N.J.). Food grade olive oil and safflower oil were
purchased for initial tests from the grocery store. Benzyl benzoate
was obtained from Acros Organics (New Jersey) and N.F. grade ethyl
oleate from Spectrum Chemical Corp. (Gardena, Calif.).
[0098] Lysozyme powder as received was dry milled with a porcelain
mortar and pestle for several minutes. The milled powder was then
sieved through a number 400 mesh and particles smaller than 37
.mu.m were collected. Samples were then weighed and added to a
measured amount of benzyl benzoate or a premixed 50/50 volume
mixture of benzyl benzoate and safflower oil. Each vial was then
shaken by hand to disperse the powder evenly through the
suspension.
[0099] Particle size was measured by multiangle laser light
scattering with a Malvern Mastersizer-S (Malvern Instruments, Ltd.,
Worcestershire, UK). The milled and sieved powder sample size was
measured as a suspension in acetonitrile in a large recirculation
cell (.about.500 ml) and immediately after being added to ethanol
in a small batch cell (Malvern, Worcestershire, UK, .about.15 ml).
In each case the obscuration during the measurement was between
10-15%. After storing the suspensions for 2 months at room
temperature, the particle size was measured again immediately after
shaking and diluting the sample in ethanol in the small batch
cell.
[0100] The viscosity was measured as the time to draw 1 ml of the
sample into a syringe with a 25 g 5/8'' or 27 g 1/2'' needle. Each
measurement was made at least 3 times and averaged. Liu et al.
found this measurement time to be correlated linearly to viscosity.
From known viscosities of each pure liquid, benzyl benzoate,
ethanol, ethyl oleate and olive oil, the correlation between the
time to draw 1 ml of solution and the viscosity was found for each
needle size to give an r.sup.2 value greater than 0.999. The
apparent viscosity of the suspensions in each pure solvent was
calculated from these correlations and the values for the two
separate needle sizes were averaged to give a final average
apparent viscosity of each sample. Additional samples of the
solvent mixture of benzyl benzoate and safflower oil were made at
10, 20, 30, 40, 50, 60, 70, 80, and 90 percent benzyl benzoate by
volume. The viscosity of each sample was calculated as described
above.
[0101] The settling rate of the particles in the solvents was
measured by shaking up the suspension in a test tube 13 mm in
diameter. Pictures were taken with a standard digital camera after
10, 30, 60, 90, 120, 150, 180, 210, 240, 1200, and 1440 minutes. To
measure the final settling volume of the samples, the vials
containing the suspensions were left undisturbed for 4 months and
images of the settled suspension were taken. All images were
analyzed using ImageJ software for the distance from the meniscus
to the settling front. The maximum volume fraction for the settled
suspension was defined by dividing the volume fraction of particles
in the overall suspension by the ratio of the volume containing
particles after settling for 4 months to the overall volume.
[0102] The concentration of lysozyme in an aqueous solution was
measured following the protocols for the Micro BCA Protein Assay.
Each sample was measured in triplicate with relative standard
deviations (% RSD) less than 2% in a General Assay 96 well plate
(see statistical analysis below). The absorbance was measured at
562 nm in a spectrophotometer. A standard curve of untreated
lysosyme was prepared at concentrations between 2 and 30
.mu.g/ml.
[0103] Partitioning and dissolution of lysozyme from the
concentrated suspension was measured in a pH 7.4 potassium
phosphate buffer. The USP paddle method was used with a VanKel
VK6010 Dissolution Tester with a Vanderkamp VK650A
heater/circulator. 900 ml of dissolution media was preheated in
large 1 L capacity dissolution vessels (Varian Inc., Cary, N.C.) to
37.degree. C. A sample of the concentrated suspension giving a
total of 18 mg of lysozyme was added. At time increments of 2, 5,
10, 20, 40, 60, 120 and 240 minutes, 1 ml samples were taken and
analyzed using the Micro BCA protein analysis mentioned earlier.
0.1 ml of the concentrated lysozyme suspension was added to a test
tube with 4 ml of DI water. This mixture was then gently mixed and
left for 3 days for the protein to partition to the water phase.
The water phase was then separated and a sample was diluted and
tested for concentration using the Micro BCA Protein Assay as
mentioned above. The remaining aqueous solution was diluted to a
concentration of 1 mg/ml. This solution was tested for optical
density using at least 3 300 .mu.l aliquots in a 96-well plate and
analyzed using the .mu.Quant spectrophotometer at 350 nm. A
standard lysozyme aqueous solution was made at 1 mg/ml
concentration and exposed to the pure benzyl benzoate solvent and
the benzyl benzoate and safflower oil solvent mixture for 3 days
and measured as the standard for oil-water interface induced
aggregation of the protein. The aqueous solution without being
exposed to any organic solvent was also measured immediately after
it was made and used as the standard absorbance for all
measurements.
[0104] Three separate 0.1 ml aliquots of the resuspended
concentrated lysozyme suspensions were added to test tubes with 8
ml of DI water. These mixtures were then gently mixed and left for
1 day for the protein to partition to the water phase. The aqueous
phase was then separated and diluted to a theoretical concentration
of 20 .mu.g/ml if 100% of the protein partitioned. The actual
concentration was then analyzed using the Micro BCA protein assay
mentioned above.
[0105] Karl Fischer Moisture Analysis. After being stored for four
months, a sample of 0.1 ml of the redispersed concentrated
suspension was inserted using a 19-gauge needle through the septum
of the titration cell of an Aquatest 8 Karl-Fischer Titrator
(Photovolt Instruments, Indianapolis, Ind.). Each suspension, pure
benzyl benzoate and the benzyl benzoate and safflower oil solvent
mixture was measured in triplicate and averaged.
[0106] Polarity Determination. An aliquot of the suspension was
diluted with the solvent until individual particles were visible on
a slide through an optical microscope (Bausch & Lomb, 10.times.
magnification). Microelectrophoresis was used to determine if a
charge was present on the particles. The diluted particle
dispersion was placed between two parallel wire electrodes
(0.01-in. diameter stainless steel 304 wire, California Fine Wire)
spaced 1 mm apart. The electrodes were secured to a glass
microscope slide and observed by optical microscopy. A potential of
10-100 V was applied with the polarity of the electrodes switched
every 15-60 sec.
[0107] Samples for protein concentration, Karl Fisher moisture
analysis, suspension uniformity, optical density, and rate of
lysozyme partitioning to water were measured in triplicate to
determine the mean, standard deviation and the relative standard
deviation (std. dev./mean).
[0108] For lysozyme particles that were milled by mortar and pestle
and sieved through a number 400 sieve, the average particle size
was approximately 20 .mu.m, according to light scattering
measurements (FIG. 12). A minor secondary submicron peak was also
visible in all measurements. However, since the 15 ml small batch
cell is only calibrated for particle sizes down to 500 nm, this
peak was not included in the analysis. The vial containing the
particles and solvent is then shaken by hand and the particles
disperse to form a uniform suspension (FIG. 13B). When the particle
suspensions are allowed to sit undisturbed, they settle slowly
enough to remain partially suspended even after 24 hours (FIG. 13C)
and the highly concentrated suspension takes up a significant
portion of the volume even after 2 months (FIG. 13A).
[0109] Viscosity of Solvent Mixture and Suspensions. Using the
known viscosities of pure solvents, a correlation between the time
to draw 1 ml of the sample and viscosity was generated. This type
of correlation has been described by Shire and coworkers on the
basis of the Hagen-Poiseuille equation.sup.5-7
v = R 2 8 .eta. ( .DELTA. P L ) Eq . 2 ##EQU00002##
where .upsilon. is the velocity, R is the inner radius of the
needle, .eta. is the viscosity, and .DELTA.P/L is the average
pressure drop over the length of the needle. Ensuring that the
average pressure drop over the length of the needle remains
constant for each sample by maintaining the same suction pressure
inside the syringe, the inverse of the velocity of the fluid
multiplied by the cross-sectional area gives the time to draw up a
specified volume of liquid, in this case 1 ml. This time is
proportional to the viscosity as shown by the Hagen-Poiseuille
equation.
[0110] The measured viscosities of the solvent mixtures of benzyl
benzoate and safflower oil are shown in FIG. 14. In this case,
since the minimum and maximum values are the for the pure solvents,
the generalized mixing rule should follow the form
f ( .eta. m ) L = i x i f ( .eta. i ) L Eq . 3 ##EQU00003##
where .eta..sub.m is the viscosity of the mixture, i is the number
of components, x.sub.i is the liquid volume, weight, or mole
fraction, and .eta..sub.i is the viscosity of the i.sup.th
component. f(.eta.).sub.L can be ln(.eta..sub.L), 1/.eta..sub.L, or
another typical equation..sup.20 In this case, the correlation most
closely associated with the experimental results was when
f(.eta.).sub.L was ln(.eta..sub.L). This theoretical result is
shown by the dotted line in FIG. 14.
[0111] The apparent viscosities of the suspensions with increasing
concentration were measured for both the pure benzyl benzoate
system and the solvent mixture of 50/50 benzyl benzoate and
safflower oil. The resulting viscosities, averaged from the
measurements using two syringe sizes (left y-axis), and the time to
draw 1 ml from the 25-gauge syringe (right y-axis) were plotted
against the concentration of lysozyme particles (FIG. 15). The
correlation of the apparent viscosity with the free solvent volume
fraction was modeled using the Kreiger-Dougherty equation
.eta. .eta. o = [ 1 - ( .phi. .phi. max ) ] - [ .eta. ] .phi. max
Eq . 4 ##EQU00004##
where .eta. is the apparent viscosity of the dispersion,
.eta..sub.o the solution viscosity, .phi. the volume fraction of
particles, .phi..sub.max the maximum packing fraction, and [.eta.]
the intrinsic viscosity. .phi..sub.max was approximated after
gravitational settling of the particles over 4 months. It was
approximately 0.50 for the pure benzyl benzoate solvent solution
and 0.52 for the benzyl benzoate and safflower oil solvent for low
shear rates Using these values, the intrinsic viscosity of the
suspension, [.eta.], was determined to be 2.7 for the pure benzyl
benzoate suspensions and 2.3 for the benzyl benzoate and safflower
oil suspensions.
[0112] The stability of the particles in suspension was measured by
numerous different techniques. First the settling rate was
calculated and compared to the theoretical Stokes settling rate
U s = 2 r 2 .DELTA..rho. g 9 .eta. o Eq . 5 ##EQU00005##
where r is the radius of the particles, .DELTA.p is the difference
in densities between the solvent and the particles, and g is
acceleration due to gravity. For a high concentration of particles,
the particle crowding will reduce the settling rate to yield
U=U.sub.s(1-.phi.).sup.6.55 Eq. 5
[0113] This modified Stokes settling rate and the experimentally
measured values were found to be within a factor of two for most
concentrations lower than 300 mg/ml as shown in Table 1. However,
for a concentration of 400 mg/ml the values are an order of
magnitude lower than the predicted rate (Table 9).
TABLE-US-00008 TABLE 9 Comparison of experimental settling rates
and settling rates quantified by the modified Stokes settling
equation accounting for particle interactions (eq. 4, 5). volume
Modified fraction of Experimental Stokes particles settling
settling Formulation (.PHI.v) rate (cm/min) rate (cm/min) 50 mg/mL
Lys 0.03515 0.0308 0.0141 in 50/50 Safflower Oil and Benzyl
Benzoate 100 mg/mL Lys 0.0703 0.0220 0.0111 in 50/50 Safflower Oil
and Benzyl Benzoate 200 mg/mL Lys 0.1406 0.0053 0.0066 in 50/50
Safflower Oil and Benzyl Benzoate 300 mg/mL Lys 0.2109 0.00071
0.0038 in 50/50 Safflower Oil and Benzyl Benzoate 50 mg/mL Lys
0.03515 0.0651 0.0373 in Benzyl Benzoate 100 mg/mL Lys 0.0703
0.0274 0.0293 in Benzyl Benzoate 200 mg/mL Lys 0.1406 0.0087 0.0175
in Benzyl Benzoate 400 mg/mL Lys 0.2812 0.00052 0.0054 in Benzyl
Benzoate
[0114] The suspension uniformity was further quantified by the
percent extracted into an aqueous phase. Initially the rate of
partitioning of the lysozyme into the aqueous phase was determined
to require approximately 60 minutes. Three aliquots from the
resuspended samples were placed in separate test tubes and allowed
to partition to the aqueous phase for 1 day to ensure complete
partitioning. The aqueous phase was then diluted approximately 1000
times, and the concentration of protein was measured. The results
show that even with a small volume (8 ml) of aqueous phase exposed
to 0.1 ml of the concentrated non-aqueous suspension at least 3/4
of the protein partitions into the aqueous phase in 24 hours (Table
10). The % RSD values were typically below 5% indicating reasonable
uniformity of the protein particles within the redispersed
suspension. The % RSD was slightly larger for the highly
concentrated 300 mg/ml sample in the mixed solvent.
TABLE-US-00009 TABLE 10 Percent of sample recovered in aqueous
phase and % relative standard deviation (% RSD) of 3 samples.
Concentration % recovered in Solvent (mg/ml) aqueous % RSD
Safflower Oil and Benzyl 50 77.2% 2.21% Benzoate Safflower Oil and
Benzyl 100 78.5% 6.61% Benzoate Safflower Oil and Benzyl 200 85.8%
3.20% Benzoate Safflower Oil and Benzyl 300 76.2% 9.45% Benzoate
Benzyl Benzoate 50 85.2% 1.52% Benzyl Benzoate 100 96.9% 3.46%
Benzyl Benzoate 200 92.2% 3.65% Benzyl Benzoate 400 81.2% 4.29%
[0115] The aggregation of the particles exposed to the non-aqueous
solvent was also tested to ensure that growth of the particles is
not occurring at a sufficient rate to jeopardize the storage of the
samples. The original particle size was measured via light
scattering immediately after the particles were sieved and
resuspended in both acetonitrile, where lysozyme is very insoluble,
and ethanol, where lysozyme is slightly soluble. The uniformity of
the two measurements ensures that the time scale of the measurement
is much quicker than the time scale of growth of the particles in
ethanol (FIG. 12). Following 2 months of storage, the samples were
diluted in ethanol and immediately tested. The particle size was
found to be essentially constant during storage (FIG. 12). The
visual inspection of one formulated suspension after storage for a
year and redispersion by shaking confirms that the particles could
be redispersed.
[0116] The potential effect of electrostatic repulsion on the
particle stability was tested. However the lysozyme particles did
not display organized movement when the voltage was changed from 10
to 100 for two electrodes spaced 1 mm apart in the benzyl benzoate
solvent.
[0117] Protein aggregation was investigated by measuring the
optical density on sample aliquots that partitioned from the
organic to the water phase. The protein was diluted to a
concentration of 1 mg/ml. Additional lysozyme samples in an aqueous
solution at the same concentration were exposed to the solvent to
measure the effect of the oil-water interface on aggregation. All
these solutions were checked for large protein aggregates by
comparison with a fresh lysozyme solution at the same
concentration. No aggregates were found since the absorbance at 350
nm was within 1% for the standard and all samples and therefore
within the error of the study.
[0118] Quantification of the moisture content may be used to
determine the free and bound water in the suspension. The moisture
content was measured for each suspension after being exposed to
atmospheric conditions for 2 months. The linear correlations found
between the moisture content and suspension concentration indicates
that the moisture is directly associated with the protein (FIG.
16). The benzyl benzoate solvent contains an average of 20 .mu.g of
water per 0.1 nil of solution or approximately 0.02% by weight. The
safflower oil and benzyl benzoate mixture contains approximately 74
.mu.g of water in the same volume sample or approximately 0.074%.
The sample with the highest concentration of protein in benzyl
benzoate, 400 mg/mL, contained the most moisture, an average of
4450 .mu.g of water per 0.1 ml of solution giving an absolute
maximum concentration of 4.5% by weight of the suspension after
being stored for 2 months.
[0119] In addition to protein and particle stability, the other key
criterion is that the suspension's apparent viscosity must be low
enough for injection through a syringe. For subcutaneous delivery,
50 cP is an appropriate maximum viscosity where it will take
approximately 20 seconds for 1 ml of the suspension to be expelled
via a 26-gauge syringe. From FIG. 15, the highest apparent
viscosity measured was approximately 50 cP, where it took
approximately 55 seconds to draw 1 ml into a 25-gauge syringe. The
disparity in the times measured reflects the smaller suction force
to draw the volume into the syringe relative to the force needed to
expel the solution from the syringe. In addition, these
concentrated suspensions are considered to be shear thinning since
the flow of the suspension will produce a more favorable
rearrangement of the particles. For example, as the shear rate
increases, spherical particles that were initially randomly packed
(.phi.max=0.64) become more ordered and pack tighter, giving a
higher maximum packing fraction around 0.71. As a result, at the
higher shear rate associated with expelling the volume, the
measured apparent viscosity of the suspensions will decrease and
will remain below the maximum for potential subcutaneous
delivery.
[0120] Further analysis of the viscosity of the suspensions using
the Kreiger-Dourghety equation can give an indication of the effect
of interparticle forces in the suspension. The initial equation for
the viscosity of a dilute suspension derived by Einstein takes into
account the particles and assumes that the particles are solid
spheres and their concentration is low enough for the particles to
be treated individually (.phi.<0.1). This gives a first order
equation where the volume fraction of the particles is related to
the viscosity ratio of the suspension over the solvent with a slope
of 2.5. In more general terms, this slope signifies the increment
of viscosity due to the addition of dispersed particles and is also
called the intrinsic viscosity, [.eta.]. For more concentrated
suspensions, accounting for particle crowding and the maximum
packing fraction of the suspension (.phi.max) results in the
Krieger-Dougherty Equation (Equation 5).31,32 In this case, the
intrinsic viscosity term can vary from the Einstein coefficient
value of 2.5 depending on the effects of solvation, varying shapes,
and electrostatic forces as well as the shear rate. Since the
values for the intrinsic viscosity of the benzyl benzoate and the
benzyl benzoate and safflower oil mixture suspensions are close to
the original Einstein derived 2.5, the effects of solvation,
varying shapes, and electrostatic forces may be considered to be
negligible assuming the shear rate is low and can be approximated
as zero. The lack of electrostatic effects was not surprising given
the tendency for ion pairing in the solvent with a dielectric
constant of only 4.8. This is further confirmed with the lack of
electrophoretic mobility measured above. If the particles are
solvated by the solvent the volume fraction would increase by
.phi. solvated .phi. dry = [ 1 + ( m 1 , b m 2 ) ( .rho. 2 .rho. 1
) ] Eq . 7 ##EQU00006##
where m.sub.1,b is the mass of the bound solvent, m.sub.2 is the
mass of the particle, .rho..sub.2 is the density of the particle
and .rho..sub.1 is the density of the solvent. In the case of the
Krieger-Dougherty equation, this increase is absorbed into the
[.eta.] term, increasing the measured intrinsic viscosity. The
deviations of the particles from spherical shape has a strong
effect on the maximum packing fraction and on the intrinsic
viscosity. For example, glass fibers of varying axial ratios, 7, 14
and 21 increase in intrinsic viscosity from 3.8 to 5.03 to 6.0,
respectively, and decrease in maximum packing fraction from 0.374
to 0.26 to 0.233.
[0121] Since a large macromolecule, such as a monoclonal antibody
in solution, can be approximated as a small colloid, similar
viscosity analysis can be conducted. In this case, previously
published values for the increase in viscosity of a solution
containing a monoclonal antibody at varying concentrations was used
giving a final value of [.eta.] of 45 from analysis using the
Krieger-Dougherty equation (Table 11). However, the analysis of
protein solutions is typically done using mass concentrations
(g/ml) rather than volume fractions, leading to values of the
intrinsic viscosity in units of cm3/g and slightly different
derived higher order relationships between the viscosity of a
protein solution and the aqueous solvent. A hard quasispherical
model
.eta. .eta. o = exp ( [ .eta. ] c ( 1 - k .upsilon. [ .eta. ] c ) )
Eq . 8 ##EQU00007##
where c is the mass protein concentration, k is a crowding factor
accounting for high concentrations of the protein and .upsilon. is
the Simha parameter accounting for the change in shape from a
sphere, is derived from the same power series as the
Krieger-Dougherty equation; however, using concentration rather
than volume fraction as the x component. As a result, it leads to a
value of 6.9 cm.sup.3/g for the intrinsic viscosity of a monoclonal
antibody and a .kappa./.upsilon. value of 0.533. For various
proteins, the value of the intrinsic viscosity varies from
approximately 2.7 for lysozyme to over 200 cm.sup.3/g. This model
has been shown to accurately predict the viscosities of hemoglobin,
bovine serum albumin, and two various monoclonal antibodies where
long range electrostatic forces were found to play a negligible
role in the viscosity versus concentration. Even for lysozyme, a
protein with a very small axial ratio, 1.5, this model shows a
dramatic increase in viscosity around a concentration of 300 mg/ml
(FIG. 15). Therefore, for various monoclonal antibodies, BSA, and
hemoglobin that can described by the hard quasispherical model and
have higher axial ratio than lysozyme (eq. 8), the rapid increase
in viscosity will be more dramatic and occur at a lower
concentration. For these solutions, the viscosity increase is due
not only to particle crowding but also to the excluded volume
effects of solvation and deviation of the shape from a sphere. This
strong deviation is not seen for particles in suspension at the
same concentrations because they are fairly spherical, are not
hydrated by the solvents, and the density of a protein is typically
around 1.35 g/cm.sup.3 leading to a lower volume fraction for the
respective concentration.
TABLE-US-00010 TABLE 11 Comparison of the two solvent systems of
suspensions and the high concentrated solution monoclonal antibody
for the exponents for the Krieger-Dougherty equation, the
experimental maximum packing fraction, and intrinsic viscosity.
Exponent for Krieger- Maximum volume Intrinsic Dougherty equation
packing fraction viscosity Solvent system -[.eta.].PHI..sub.max
.PHI..sub.max [.eta.] Benzyl benzoate -1.362 .+-. 0.09 0.50 2.7
suspension 50/50 Safflower -1.149 .+-. 0.06 0.52 2.3 Oil and Benzyl
Benzoate suspension Mab1 solution -45.27 .+-. 0.61 1 45.3 in
aqueous solution
[0122] The viscosities of concentrated suspensions up to 300-400
mg/ml of milled particles of the model protein, lysozyme, were
small enough for subcutaneous injection through a 27-gauge needle.
The protein molecules were stable against aggregation and the
particle size did not vary for at least 2 months when stored at
atmospheric conditions. The apparent viscosity was correlated with
volume fraction at all conditions according to the
Krieger-Dougherty equation. Full settling of the particles was
found to take well over 24 hours, which gave sufficient time for a
uniform aliquot to be taken and analyzed. The results indicated
sufficient suspension stability to allow for accurate dosing,
either for, single injection or multiple injection applications.
Overall, the colloidal stability and dose uniformity of the
lysozyme microparticle suspensions, along with the acceptably low
viscosities, indicates a potential advance for subcutaneous
delivery of therapeutic proteins. For highly concentrated proteins
in solution, various forces including electrostatic repulsion,
solvation forces, and deviations of the particle shape from a
spherical geometry can cause large increases in viscosity, whereas
these forces have almost negligible effects for the current protein
suspensions, resulting in much lower viscosities.
[0123] The objectives of this example are to: (1) use various
particle engineering techniques to produce protein particles to
form suspensions in nonaqueous and aqueous solvents, (2) to find
efficient ways to form the particles in vials or transfer the
particles to vials, (3) to determine the particle size, colloidal
stability and viscosity of the suspensions, and (4) determine the
stability of the protein molecules with regard to denaturation and
aggregation. The relevant particle engineering techniques include
milling, spray drying, precipitation, and thin film freezing. The
feasibility of delivering proteins and peptides by subcutaneous
injection depends on formulating a sufficiently low viscosity
product that is syringeable through a 25 to 27-gauge needle but
contains the necessary high concentration of the active protein or
polypeptide to give a full dose in less than 1.5 ml of volume.
Desired viscosities have been made with protein suspensions with up
to 400 mg/mL.
[0124] Of primary importance in the formulation of a stable
suspension system that contains a high concentration of the
monoclonal antibody and a sufficiently low viscosity is the
formation of stable appropriately sized monoclonal antibody
particles. Previously, a suspension was formulated with
.about.10-20 .mu.m milled lysozyme.1
[0125] In addition to milling, protein particles may be formed with
thin film freezing (TFF), previously disclosed to produce stable
nanoparticles of proteins2, to produce similar micron-sized
particles, depending upon the feed concentration. Since the
monoclonal antibody will come in solution form, TFF will be a
significantly less destructive process, causing only potential
freezing stress whereas lyophilization followed by milling and
sieving the particles will expose them to freezing, heating and
mechanical stresses which can lead to denaturation. Proposed work
will include tuning the TFF process to produce the specified
particle size by altering the feed concentration within the
solubility limits of the Mab and adding cryoprotecting sugars such
as trehalose, varying the solvent (currently pure DI water with a
buffer) to include a percentage of ethanol and various buffers and
other excipients necessary for a subcutaneous injection, and
examine freezing directly into appropriate vials for subcutaneous
injection formulation.
[0126] Various techniques will be utilized to transfer the protein
particles into the vials. The goal will be to simplify processing
steps and to maintain sterile conditions. The first method is to
transfer lyophilized powder to the vials as a solid. A second
method is to transfer the protein to the vials while still in the
frozen state. Particle size data was obtained to demonstrate each
of these methods using the TFF process (Table 12).
TABLE-US-00011 TABLE 12 Comparison of particle size distributions,
measured by multi-angle light scattering of protein particles
suspended in acetonitrile, from different concentration
distributions of TFF. The particles were transferred to vials
either after lyophilization or after only freezing. Particle Size
Sample Freezing Method (dV10, dV50, dV90) High concentration BSA
Frozen in vial (18.93, 32.88, 53.07) High concentration BSA
Separated after (0.32, 17.95, 45.07) lyophilization High
concentration LYS Separated after (0.23, 13.60, 31.30) primary
drying lyophilization High concentration LYS Separated after (0.31,
13.17, 28.57) freezing Low concentration LYS Frozen in vial (0.12,
0.39, 2.27)
[0127] Once appropriate stable particles are made, an appropriate
solvent system must be found. In this case, both aqueous and
non-aqueous based systems will be analyzed. A stable, low viscosity
formulation of milled lysozyme in the pure non-aqueous solvent
benzyl benzoate and the solvent system benzyl benzoate and
safflower oil have been previously analyzedl and will be analyzed
further in a suspension containing the Mab particles described
above. To overcome the instabilities of the protein or polypeptide
in a highly concentrated aqueous solution and the increase in
viscosity caused by soluble aggregates and hydration, the main
objective of this study was to formulate a protein suspension in a
non-aqueous solvent or solvent mixture using the model protein
lysozyme that is syringeable through a 27-gauge needle and at a
concentration greater than 100 mg/ml. The suspensions formulated
were found to remain syringeable up to at least a concentration of
300 mg/ml. Protein and particle stability remained for at least 2
months indicating a potentially stable protein product. Protein
particles in suspension were also found to be resuspendable after
being stored at room temperature for a year. Correlations between
the viscosity of the formulation and the increasing volume fraction
of particles indicates that the main source of interaction between
the particles is due to particle crowding and no additional forces
such as electrostatic repulsion, solvation of the particles from
the solvent, or deviations of the particle shape from spheres are
used to maintain the stability of the suspension.
[0128] However, the main focus will be on aqueous solvents, whereas
the nonaqueous solvents will be of secondary interest. An aqueous
solvent system is slightly more complicated to formulate and must
include: (1) an agent to decrease the solubility of the Mab in
water (if solubility is greater than 20 mg/mL); (2) an antifoaming
agent to prevent the concentrated suspension from producing a foam
when the solvent is added; and (3) additional excipients to
maintain the stability of the Mab and the final suspension product
for a sufficient period of time.
[0129] Agents that can be added to decrease the solubility of a
protein in water include high concentrations of salts such as
sodium sulfate and ammonium sulfate, complexing agents including
zinc, water soluble polymers such as PEG, various organic solvents
such as ethanol and other surfactants such as Tween 20. Addition of
salts increases the ionic strength of the solution which decreases
the solubility of the hydrophibic groups of a protein. However for
subcutaneous injectable delivery, a solution that is not isotonic
will increase the pain upon injection, potentially leading to
noncompliance. Therefore, alternatives to decrease the solubility
with other additives can be advantageous. A polyethylene glycol
(PEG) and water mixture has also been demonstrated to that similar
"salting out" effects of salt and are more readily allowed in
injectable formulations. Different molecular weight PEGs will be
tried since lower molecular weight PEGs can be formulated to much
higher concentrations for additional solubility decrease however
higher molecular weight PEG have been found to increase the thermal
stability of a protein. Tween 20 at a 1%(w/v) concentration has
been demonstrated to cause precipitation of the hydrophobic protein
Humicolalanuginosa lipase5 which may prevent solubilization of the
added Mab particles in suspension and create another alternative
way to formulate the suspension. Additional excipients, such as
water soluble non-aqueous solvents, where the Mab is only slightly
soluble will also be tried including ethanol, propylene glycol, and
dimethylacetamide.
[0130] Successful higher concentration formulations of low
viscosities with salt and PEG formulations have been made in
aqueous media (Table 13). Initial particle size measurements as
well components are also disclosed. These formulations have been
found to contain particle stability for at least 1 hr (FIG. 17);
long time stability has not been investigated. Preliminary data
indicate that the viscosities of some of the aqueous suspensions
were sufficiently low for 25 gauge needles.
TABLE-US-00012 TABLE 13 Protein used, concentration, substance
added to decrease solubility, preliminary viscosity measurement,
preliminary particle size for successful aqueous based
formulations. Agent to Protein and decrease protein Viscosity
concentration solubility measurements Particle size 150 mg/ml PEG
300 57.5 cP (0.40, 13.98, 26.82) BSA aqueous solution 250 mg/ml
Salt aqueous 5.7 cP (0.27, 10.90, 26.91) LYS solution
[0131] Foaming of aqueous dispersions of proteins has been observed
in a 1.0M sodium sulfate aqueous solvent suspension when the liquid
was added to the dried TFF particles. Foam stability of an aqueous
suspension of proteins and fat particles was reduced significantly
with the addition of the different Span excipients where it was
hypothesized to increase bubble coalesance. A trial using SPAN.RTM.
80 successfully broke the foam when added to a concentrated
dispersion of BSA in a 1.0M Sodium Sulfate solution, leaving a low
viscosity, highly concentrated suspension. Other surfactants will
be used both in the particle formation and in the final formulation
stage to resolve the foaming issue. In addition, other excipients
must be added to complete the formulation of a pharmaceutically
acceptable suspension to be administered subcutaneously. To
stabilize the protein, surfactants such as Tween 20 and Tween 80
can be added if necessary. To create a final aqueous dispersion,
buffering agents, antioxidants, and antimicrobial agents may be
added.
[0132] Final formulated suspensions will then be tested for
viscosity using the time to draw 1 mL of solution up a syringe with
a 25- then 27-gauge needle. As demonstrated previously, the
viscosity approximated with this measurement is reasonable with by
both theoretical and previous results (above). In addition,
settling rate measurements, images taken by optical microscopy,
particle size measurements by multi-angle laser light scattering
and electrophoretic mobility measurements will be used to
characterize the stability of the suspension. The settling rate can
be useful to determine the uniformity of the suspension over a
period of time, in addition to aliquots of the suspension removed
randomly and analyzed for protein concentration. In addition, the
images taken by optical microscopy can show the dominant forces on
the particles at lower concentrations and whether the particles are
aggregating, flocculating or repulsing. Particle size measurements
over time will show the effects of Oswald ripening and coagulation
of particles. To complete the study, the electrophoretic mobility
of the particles will be measured and quantify the zeta potential,
which will determine the electrostatic stabilization of the
particles in the suspension.
[0133] The final step is to determine the stability of Mab in the
formulated suspension. For aqueous based suspensions, appropriate
dilutions must be run to produce a solution of the Mab at the
necessary concentration for analysis. The Mab in solution will also
be separated from the suspended particles and both will be analyzed
separately for stability. For non-aqueous based suspensions an
aliquot with sufficient Mab concentration must be exposed to pure
water or a suitable buffer to allow the suspended particles to
partition to the aqueous phase in solution without denaturing on
the oil-water interface. As has been done previously, suspensions
will be exposed to the appropriate aqueous buffer for 1-3 days to
allow slow partitioning and minimal exposure to the potentially
denaturing oil-water interface.1 In addition, standards will be run
to ensure that the measurements are an accurate representation of
the suspended particles.1 The stability of the Mab can then be
characterized by a variety of techniques including size-exclusion
chromatography (SEC), dynamic light scattering (DLS), analytical
ultracentrifugation (AUC), and optical turbidity for soluble and
insoluble aggregates.
[0134] Moisture content of both the dried powder and the
reconstituted non-aqueous suspension will be analyzed over time to
determine the effect of excess water on the stability of the
formulation. As has been noted in previous experiments, at low
levels of protein hydration, the water sorption over time in a
non-polar or moderately polar organic solvent is similar to that
from the vapor phase itself. 12 In addition, a Mab specific ELISA
assay can be run to demonstrate the % activity of the Mab after
formulation to analyze misfolding and denaturation of the Mab. This
technique has been used previously to observe the resulting binding
affinity of an IgG after reconstitution. FTIR can also be run to
determine any Mab-excipient interactions with any part of the final
formulation.
[0135] Production of particles by film freezing in vials. In many
particle formation processes, the transfer of solids from various
surfaces to vials for delivery of the final dosage form presents
problems. It is necessary to maintain sterile conditions and it can
be difficult to determine and control the exact amount of particles
transferred. In addition, the particle size may change during
handling. It would be desirable to produce the particles directly
within the vial the final dosage form will be stored in. This
process of direct freezing and particle production within a vial is
practiced in conventional tray lyophilization. However, the slow
freezing rates by heat transfer through the shelf to the vial leads
to particles typically on the order of hundreds of microns. The
current technique provides a means to produce and control
sub-micron and micron sized particle distributions.
[0136] A method is provided for forming particles of substances
inside vials, the method comprising: (a) introducing a liquid
solution of the substances inside a cylindrical vial; (b) immersing
the cylindrical vial into a liquid coolant while rotating it
horizontally until the liquid has frozen as a film in the vial's
internal walls; (c) removal of the solvent by lyophilization or by
extraction of the frozen solvent into a second solvent.
[0137] Thus, the current invention provides an alternative to
freeze much more quickly with a submersion of the vials in a
cryogenic liquid than in the case of tray lyophilization. The more
rapid freezing on the order of 20 s can result in sub-micron
particles. The thickness of the freezing liquid, in the range of
0.2 to 4 mm, facilitates the rapid freezing. The vials may be
transferred directly to a lyophilizer to remove the solvent while
producing the final particles in the vial. The final particles can
also be produced by addition of agents (salts, a second solvent,
polymers, etc.) to decrease the solubility of the particles to
produce a suspension of the particles. Thus the solids never have
to be removed from the vial.
[0138] This method permits the production of particles with fine
control of the size distribution by freezing liquid solutions of
substances in thin uniform films at fast freezing rates inside
vials. The method is implemented in three fundamental steps: (a)
introducing a liquid solution of the substances inside a
cylindrical vial; (b) immersing the cylindrical vial into a liquid
coolant while rotating it horizontally until the liquid has frozen
as a film in the vial's internal walls; (c) removal of the solvent
by lyophilization or by extraction of the frozen solvent into a
second solvent.
[0139] The first step begins by dissolving the active substances in
an aqueous solution in typical concentrations ranging from 1 mg/ml
to 500 mg/ml. This solution may also contain excipients including
cryoprotectives or surfactants as an example. The solution is
introduced into a cylindrical vial, where the liquid volume and the
vial's dimensions determine the thickness of the final frozen film
(important variables in the control of the size distribution).
Table 14 shows an example of different film thicknesses obtained
for vials of two different sizes. Table 14. Film thicknesses
obtained after freezing different volumes of liquid solution in
vials of different dimensions.
TABLE-US-00013 Vial 1 Vial 2 Internal Diameter 15 mm Internal
Diameter 24 mm Length 40 mm Length 48 mm Liquid Volume Film
thickness Liquid Volume Film thickness (ml) (mm) (ml) (mm) 1 0.6 1
0.3 2 1.2 2 0.6 3 1.8 3 0.9 5 3.5 5 1.5 10 3.2
[0140] The second step includes immersing the vial horizontally
inside liquid coolant (e.g. liquid N2) while rotating it. The
rotation causes the liquid solution to freeze as a film of uniform
thickness in the cylindrical vial internal walls. The coolant
temperature (typically ranging from 50 K to 253 K) and the rotation
speed (typically ranging from 15 RPM to 600 RPM) may be adjusted to
control the freezing rate. FIG. 18 show freezing temperature
profiles measured for freezing different liquid volumes inside
vials. The temperatures were measured with a type T thermocouple
while processing with a coolant at 80 K and a rotation speed of 30
RPM.
[0141] The third step is the removal of the solvent by
lyophilization or adding agents to the frozen solvent to create a
poorly-soluble environment producing a suspension. A second
solvent, salts, polymers and other agents can be added to the
aqueous based formulation to produce a poorly-soluble environment
for the protein-based particles. Solvents are typically
water-miscible organic solvents such as acetonitrile and ethanol.
Salts, such as sodium sulphate and ammonium sulphate, and polymers
such as PEG, cause a decrease in the solubility of the proteins in
an aqueous environment, creating a suspension of particles. FIG. 19
shows typical particle size distributions obtained with the present
method measured by a multiangle laser light scattering with a
Malvern Mastersizer-S with the particles suspended in acetonitrile.
At selected conditions, nanometric particles, micrometric or
bimodal distributions of particles of both size scales can be
produced, as shown in FIG. 19. The size distribution is controlled
by solutes concentration, the temperature of the liquid coolant,
and the volume of the liquid and the vial rotating speed. Table 15
shows the process conditions that resulted in the particle size
distributions shown in FIG. 19.
TABLE-US-00014 TABLE 15 Process conditions used in the production
of particles by Film Freezing into Vials, corresponding to the size
distributions shown in FIG. 13. The film thickness is defined as
the maximum thickness of the final frozen mass within a vial in the
horizontal position. The vial inside diameter is 15 mm. Protein
Coolant REF. in Concentration Film Thickness Rotation speed
Temperature FIG. 7 mg/ml mm RPM K (a) 20 2.6 30 80 (b) 10 0.6 30 80
(c) 5 2.6 120 80 (d) 5 0.6 30 210
[0142] The following examples demonstrate protein suspensions made
by film freezing of protein in a vial followed by lyophilization,
and then suspension of the lyophilized material in a solvent with
manual shaking. In FIG. 20, 4 ml of lysozyme solution (20 mg/ml) in
water were frozen by Film Freezing inside the Vial. After
lyophilization, a suspension with 80 mg/mL lysozyme was formed by
adding benzyl benzoate. In FIG. 21, 2 ml of hemoglobin solution
(150 mg/ml) in water was frozen by Film Freezing inside the Vial.
The frozen solution was lyophilized and the particles were
suspended in 2 ml of Benzyl Benzoate to make up a 150 mg/ml
suspension. In both cases, the suspensions did not settle over 1
day, and could resuspended by manual shaking.
[0143] It is contemplated that any embodiment discussed in this
specification can be implemented with respect to any method, kit,
reagent, or composition of the invention, and vice versa.
Furthermore, compositions of the invention can be used to achieve
methods of the invention.
[0144] It will be understood that particular embodiments described
herein are shown by way of illustration and not as limitations of
the invention. The principal features of this invention can be
employed in various embodiments without departing from the scope of
the invention. Those skilled in the art will recognize, or be able
to ascertain using no more than routine experimentation, numerous
equivalents to the specific procedures described herein. Such
equivalents are considered to be within the scope of this invention
and are covered by the claims.
[0145] All publications and patent applications mentioned in the
specification are indicative of the level of skill of those skilled
in the art to which this invention pertains. All publications and
patent applications are herein incorporated by reference to the
same extent as if each individual publication or patent application
was specifically and individually indicated to be incorporated by
reference.
[0146] The use of the word "a" or "an" when used in conjunction
with the term "comprising" in the claims and/or the specification
may mean "one," but it is also consistent with the meaning of "one
or more," "at least one," and "one or more than one." The use of
the term "or" in the claims is used to mean "and/or" unless
explicitly indicated to refer to alternatives only or the
alternatives are mutually exclusive, although the disclosure
supports a definition that refers to only alternatives and
"and/or." Throughout this application, the term "about" is used to
indicate that a value includes the inherent variation of error for
the device, the method being employed to determine the value, or
the variation that exists among the study subjects.
[0147] As used in this specification and claim(s), the words
"comprising" (and any form of comprising, such as "comprise" and
"comprises"), "having" (and any form of having, such as "have" and
"has"), "including" (and any form of including, such as "includes"
and "include") or "containing" (and any form of containing, such as
"contains" and "contain") are inclusive or open-ended and do not
exclude additional, unrecited elements or method steps.
[0148] The term "or combinations thereof" as used herein refers to
all permutations and combinations of the listed items preceding the
term. For example, "A, B, C, or combinations thereof" is intended
to include at least one of: A, B, C, AB, AC, BC, or ABC, and if
order is important in a particular context, also BA, CA, CB, CBA,
BCA, ACB, BAC, or CAB. Continuing with this example, expressly
included are combinations that contain repeats of one or more item
or term, such as BB, AAA, MB, BBC, AAABCCCC, CBBAAA, CABABB, and so
forth. The skilled artisan will understand that typically there is
no limit on the number of items or terms in any combination, unless
otherwise apparent from the context.
[0149] All of the compositions and/or methods disclosed and claimed
herein can be made and executed without undue experimentation in
light of the present disclosure. While the compositions and methods
of this invention have been described in terms of preferred
embodiments, it will be apparent to those of skill in the art that
variations may be applied to the compositions and/or methods and in
the steps or in the sequence of steps of the method described
herein without departing from the concept, spirit and scope of the
invention. All such similar substitutes and modifications apparent
to those skilled in the art are deemed to be within the spirit,
scope and concept of the invention as defined by the appended
claims.
* * * * *