`
`Report
`
`Parenteral Peptide Formulations:
`Chemical and Physical Properties of
`Native Luteinizing Hormone-Releasing
`Hormone (LHRH) and Hydrophobic
`Analogues in Aqueous Solution
`
`Michael F. Powell, 1•2 Lynda M. Sanders, 1
`Alan Rogerson, 1 and Vicki Si1
`
`Received January 16, 1991; accepted May 6, /991
`
`The degradation of native LHRH in aqueous buffers of pH -1-10
`obeyed the rate equation, kobs = kH•aH+ + k0 + k1-1o•(aHo· Y.
`where x at 60--IOO"C was ~0.64 and temperature independent. Ex(cid:173)
`trapolation to 25°C using the Arrhenius equation and secondary rate
`constants showed that native LHRH is reasonably stable at pH 5.4,
`giving a shelf life (190) of approximately 5 years. Regarding physical
`properties, hydrophobic LHRH analogues nafarelin and detirelix
`were found to be surface active as demonstrated by a decrease in
`apparent surface tension with increased peptide concentration. The
`CMC for detirelix at pH 7.4 was determined to be 5.3 x 10- • M
`(0.88 mg/ml), and that for nafarelin, >2 mg/ml. At higher concen(cid:173)
`trations (~4-8 mg/ml), nafarelin and detirelix formed nematic liquid
`crystals of undulose extinction (birefringence, <0.001 ). The thermo(cid:173)
`dynamic stability of these peptide liquid crystals was probed by
`determining their melting points (1cm) in the presence of propylene
`glycol, a solvent which proved to be efficacious at suppressing ge(cid:173)
`lation and at destabilizing liquid crystals as measured by a reduction
`in rem·
`KEY WORDS: peptide stability; liquid crystal formation; birefrin•
`gence; surface tension; luteinizing hormone-releasing hormone
`(LHRH); hydrolysis; aggregation; micelles; surface activity.
`
`properties of nafarelin and detirelix that are pertinent to their
`parenteral formulation, as well as some stability data on na(cid:173)
`tive LHRH for comparison. We have determined the chem(cid:173)
`ical stability of native LHRH with respect to temperature
`and pH, as well as probed the physical stability of hydro(cid:173)
`phobic LHRH analogues, by measuring the surface activity
`of these a mphiphilic peptides and by directly measuring the
`stability of peptide aggregates at higher temperatures. The
`second paper in this series will report on the effect of pH,
`added counterions and excipients on peptide aggregation and
`physical stability (10) .
`
`EXPERIMENT AL
`
`Materials
`
`Nafarelin and detirelix (both as the diacetate salt) were
`synthesized and assayed by the Institute of Organic Chem(cid:173)
`istry (Syntex Research). Native LHRH (as the acetate salt)
`was purchased from Sigma and used without further purifi(cid:173)
`cation. Buffer solutions and added cosolvents were prepared
`from reagent grade or USP-grade c hemicals. HPLC-grade
`acetonitrile and water purified by filtration and ion exchange
`(Nanopure) were used in the mobile phase.
`
`Apparatus
`
`Reversed-phase chromatography of native LHRH was
`carried out using an HPLC system consisting of a Waters
`Model 712 Wisp autoinjector, a Waters Model 45 pump, a
`Waters Model 730 data module, and a Waters Model 450
`spectrophotometric detector. A 250 x 4.6-mm Allex Ultra(cid:173)
`sphere ODS column was used for analysis. pH's were deter(cid:173)
`mined using a Radiometer PHM 64 pH meter and a Radiom(cid:173)
`eter GK2401C combination electrode calibrated at the solu(cid:173)
`tion temperature.
`
`INTRODUCTION
`
`HPLC Conditions
`
`The low oral bioavailability of most peptides and pro(cid:173)
`teins has provided a compelling driving force for parenteral
`delivery. Unfortunately, the parenteral formulation of pep(cid:173)
`tides and proteins is far from trivial. For example, certain
`peptides, such as the hydrophobic LHRH analogues nafare(cid:173)
`lin and detirelix (Fig. l ), adhere tenaciously to glass and
`other surfaces at low concentrations (l). At higher concen(cid:173)
`trations, aqueous formulations of nafarelin or detirelix un(cid:173)
`dergo peptide aggregation and solution gelation resulting in
`compromised physical stability (2). Furthermore, although
`the chemical stability of selected hydrophobic LHRH ana(cid:173)
`logues is well understood (3-9), there exist only limited
`chemical stability data for the parent compound, native
`LHRH.
`Herein we report some of the chemical and physical
`
`1 Institute of Pharmaceutical Sciences, Syntex Research, Palo Alto,
`California.
`2 To whom correspondence should be addressed at Genentech, Inc.,
`460 Pt. San Bruno Boulevard, South San Francisco, California
`94080
`
`A linear response (:!: 1%) was obtained throughout the
`range of 0.005-5 µ,g LHRH injected. The separation of
`LHRH from its degradation products was achieved using the
`following conditions: mobile phase, water: acetonitrile
`(82: 18, v/v) buffered with 0.2 M phosphate (pH 3); flow rate,
`1.2 ml/min; detection, 210 nm; injectio n volume, 50 µ I; and
`typical LHRH retentio n time, 15 min. Stability specificity of
`this method was demonstrated by the following: (i) apparent
`pseudo-first-order decay of LHRH to <5% drug remaining;
`(ii) the spectral similarly obtained for early, middle, a nd late
`peak splices; a nd (iii) the lack of degradation product peaks
`at or near the LHRH peak when the mobile phase was var(cid:173)
`ied.
`
`Reaction Kinetics
`
`The buffer concentration used in these experiments (ex(cid:173)
`cept for HCI solutions) was held at 0.01 M to minimize buffer
`catalysis; the ionic strength was constant at µ = 0. 15 (KCI).
`LHRH acetate was added to make the final drug concentra(cid:173)
`tion 25 µ g/ml (~ 2 x 10- 5 M) a nd the pH's of each solution
`were determined at 25, 60, and 80°C. The pH's at l00°C were
`
`0724-8741/9111000-12.58$06.50/0 C> 1991 Plenum Publishing Corporation
`
`1258
`
`FRESENIUS EXHIBIT 1008
`Page 1 of 6
`
`
`
`Parental Peptide Formulations
`
`OH
`OH
`OH
`OH
`H OH
`II
`I
`II
`I
`II
`I
`II
`I
`N
`II
`I
`0 "(" ')-C•N•CH· C·N•CH·C• N•CH·C•N•CH•C• N
`\_J
`I
`I
`I
`I
`
`HN~ 6<;:H2 NH CH20HO<;:H:
`""'-
`OH
`
`'=N ?' I
`
`::::,,_
`
`HO HO
`O
`HO
`0
`I
`II
`I
`II
`II
`I
`II
`11
`H2N•C-CH2· N•Co· N -C-CH-N- C-CH• N•C ·CH
`I
`I
`I
`1H2 R
`(~H2b
`NH
`CH
`I
`/ '
`~C, CH3CH3
`H N
`NH2
`
`OH
`OH
`OH
`OH
`OH
`OH
`II
`I
`II I
`II
`I
`II
`I
`II
`I
`II
`I
`CH1-C• N-CH •C•N-CH•C•N•CH•C•N·CH·C•N·CH·C•N
`I
`I
`I
`I
`I
`CH2
`CH2
`CH2
`CH2 OH CH2
`
`cP 96"" 9-
`
`ii
`
`HO HO
`O
`HO
`0
`I
`II
`I
`II
`II
`I
`II
`11
`H2N· C•CH• N• CoN·C·CH-N•C · CH· N•C-CH
`I
`I
`I
`I
`CH,
`(CH2h CH2
`(CH2),
`I
`I
`I
`NH
`CH
`NH
`I
`/ ' I
`C
`CH3CH3 C
`~ '
`~ '
`NH2
`N
`HN
`N H
`I
`I
`C2Hs C2Hs
`H ), nafarelin (R =
`Fig. 1. Structure of (i) native LHRH (R
`2-naphthylmethyl), and (ii) detirelix.
`
`extrapolated by linear least-squares analysis of 1/T(K - 1
`)
`versus pH. The reaction solutions were then transferred to
`several I-ml clear glass vials (treated, type A), flame-sealed,
`and incubated at 60, 80, or I00°C. At known time intervals
`samples were removed and immediately frozen at - 20°C.
`Typically 8-10 samples were taken per kinetic run. Upon
`removal of the last sample the stored solutions were allowed
`to warm to room temperature, and then all samples were
`analyzed on the same day. Peak area integration values were
`used directly when fitting the data to first order kinetics;
`typically reactions were followed for two to five half-lives.
`
`Physical Study Methods
`
`The apparent surface tension (-y) of LHRH and its ana(cid:173)
`logues was measured using a Fisher Autotensiomat (Model
`215) and a 6-cm platinum ring. Control experiments demon(cid:173)
`strated that at least 3 ml of solution was required for precise
`'f measurements when using a 6-cm ring. Sample solutions
`containing 0.01 M total buffer were allowed to equilibrate for
`several minutes before taking the final reading in triplicate.
`In experiments where 'f changed with time, 'f was taken at
`known time intervals using the point of ring release as the
`recorded time.
`Peptide liquid crystal melting temperatures (Tern> were
`determined by loading peptide solutions into flat , glass cap(cid:173)
`illary tubes (50 x 3 x 0.3 mm, Vitro Dynamics Inc., NY) and
`examining them microscopically (33x) under crossed polar-
`
`1259
`
`ized light. Typically solutions were prepared at least 24 hr
`prior to examination. Liquid crystals, when observed
`through crossed polars, were detected as light-colored struc(cid:173)
`tures of nondistinct crystal form against a black background
`(nematic liquid crystals of undulose extinction). Both nafare.(cid:173)
`lin and detirelix showed an observed optical retardation of
`~ 50-200 nrn/0.3 mm, corresponding to a birefringence of
`~ 0.001 or less as calculated using the Michel- Levy chart
`(11, 12). The extent of birefringence was verified by viewing
`through a " first-order red" plate compensator; by this
`method the liquid crystals were characterized by the second(cid:173)
`order blue (addition) and first-order yellow-orange (subtrac(cid:173)
`tion) colors. In most cases, photomicrographs were taken
`such that both the air (the concave side of the meniscus) and
`the solution regions were observed so as to provide an op(cid:173)
`tical control for birefringence caused by imperfections in the
`microslide. The effect of temperature was determined by
`heating the microslides at 3°C/min until all traces of liquid
`crystallinity were lost, i.e., until the background turned com(cid:173)
`pletely black. This melting temperature is defined herein as
`the critical melting temperature (Tcm>·
`Freeze-fracture electron micrographs were produced by
`flash-freezing an aqueous solution of detirelix to - 80°C in
`liquid freon and then cooling to - l%°C in liquid nitrogen.
`Samples at - 196°C were then fractured by means of a
`cooled blade and etched and shadowed with platinum and
`carbon. Replicates were obtained by dissolving the sample in
`methanol and were then viewed by transmission electron
`microscopy.
`
`RESULTS AND DISCUSSION
`
`Temperature and pH Dependence on Native LHRH
`Degradation. This was determined at 60, 80, and 100°c by
`reversed-phase HPLC analysis. All reactions obeyed
`pseudo-first-order kinetics over the time courses studied
`(Fig. 2). The observed pseudo-first-order rate constants
`shown in Table I and Fig. 3 indicate that the degradation of
`LHRH is both acid and base catalyzed. The slope of - l in
`the acid region is indicative of specific acid catalysis by hy(cid:173)
`dronium ion. In the base region, the observed rate is not
`stoichiometrically proportional to hydroxide ion concentra(cid:173)
`tion but displays a slope of ~0.64. The pH rate profiles
`obeyed the semiempirical rate equation given by Eq. (1 ),
`where kH + , k0 , and k1-10 - are the secondary rate constants for
`catalysis by hydronium ion, water (or a spontaneous
`
`10
`
`30
`40
`20
`Time (days)
`Fig. 2. Degradation of native LHRH in pH 6.3 buffer at 60, 80, and
`1oo•c. The first order fit of the data was made by nonlinear least(cid:173)
`squares analysis.
`
`50
`
`FRESENIUS EXHIBIT 1008
`Page 2 of 6
`
`
`
`1260
`
`Powell, Sanders, Rogerson, and Si
`
`Table r. Observed Rate Constants for the Degradation of LHRH in
`Aqueous Solution
`
`60°c
`
`80°c
`
`1oo·c
`
`Buffer"
`
`pHb
`
`HCI
`HCI
`Acetate
`Acetate
`Phosphate
`Phosphate
`Carbonate
`
`1.11
`3.80<
`4.45
`5.43
`6.27
`7.64
`9.64
`
`107 kob•
`(sec - 1)
`
`270
`0.58
`0.39
`0.42
`0.71
`2.8
`63
`
`pHb
`
`1.12
`3.SOC
`4.52
`5.51
`6.30
`7.65
`9.48
`
`106 koo,
`(sec- 1)
`
`110
`030
`0.23
`0.30
`0.59
`3.2
`55
`
`pHb
`
`1.12
`3.8if
`4.53
`5.53
`6.30
`7.60
`9.36
`
`I~ kot,,
`(sec- 1)
`
`38
`0.16
`0.12
`0.20
`0.38
`2.8
`24
`
`0 Total buffer concentration (except HCI solution), 0.01 M; ionic
`strength, ,.._ = 0. 15.
`" pH's were determined at the reaction temperature indicated ex(cid:173)
`cept for 1oo•c, where pH's were obtained from linear least(cid:173)
`squares analysis of IIT (K- 1) vs pH.
`c The pH before addition of LHRH acetate salt was ~3.02.
`
`(I)
`
`k00• = kH•aH+ + k0 + kHo -(aHo- Y
`reaction), and hydroxide ion, respectively. The terms aH •
`and (aH0 -Y are the hydronium ion and hydroxide ion activ(cid:173)
`ities, where the exponent x represents the observed depen(cid:173)
`dence of the reaction rate upon hydroxide ion activity. As
`suggested earlier for nafarelin (7) a slope of less than one in
`the basic region indicates that LHRH degrades by several
`different basic pathways, each with its own pH dependence
`and true kH0 - catalytic coefficient. The rate data at 60-100°C
`were fitted by nonlinear least-squares analysis (13) to Eq. (1)
`using a four-parameter fit: kH• , k0 , k"0 -, and x. The best-fit
`value of x did not vary significantly over the temperature
`range studied (x = 0.65 :!: 0.05, 0.65 ± 0.03, and 0.61 ± O.o3
`at 60, 80, and 100°C, respectively), showing that a single
`value of x (x ~ 0.64) can be used without oversimplification
`to obtain the rate constants kn•, k0 , and kno· by Eq. (2).
`kobs = kn•aH• + ko + kHo -(GHo-)°·64
`The secondary rate constants obtained by this method
`are shown in Table II, as are the corresponding activation
`parameters. The rate constants at 25°C were estimated by
`
`(2)
`
`- 4
`
`,::--g
`~ - 6
`f - 8
`
`~
`
`-10
`
`2
`
`4
`
`6
`
`8
`
`pH
`Fig. 3. pH-rate profiles for the degradation of native LHRH at 60-
`I000C. The lower dashed line is the pH-rate profile for 25°C calcu(cid:173)
`lated using Eq. (2) and the activation parameters obtained from the
`data at 60-IOO°C.
`
`linear least-squares extrapolation of the Arrhenius data; the
`lower line in Fig. 3 is the best fit of the calculated 25°C rate
`constants to Eq. (2).
`Two features are readily apparent: (i) the pH of maxi(cid:173)
`mum stability for native LHRH (and probably structurally
`similar LHRH analogues as well) at 25°C is near pH 5.4, and
`(ii) native LHRH exhibits a shelf life of2 years or more from
`pH ~ 4.2 to ~7.2. Comparison of these rate data with those
`reported earlier for nafarelin (7) and RS-26306 (8) shows that
`there is little difference in chemical stability between native
`LHRH and some of its hydrophobic analogues. This broad
`plateau region suggests that the spontaneous (or water(cid:173)
`catalyzed) reaction for LHRH degradation largely deter(cid:173)
`mines drug stability from pH 4 to pH 7. Additionally, reac(cid:173)
`tion solutions of pH ~ 2.5 to ~ 9.5 show less than 10% deg(cid:173)
`radation after 1 month, well within the time required to carry
`out the peptide liquid crystal studies herein.
`Surfactant Nature of Nafarelin and Detirelix. This was
`probed by determining their apparent surface tension (-y) of
`aqueous peptide solutions at 25°C. A cursory glance at the
`structures in Fig. I shows that detirelix and nafarelin should
`be more surfac.e active than native LHRH. At neutral pH
`detirelix has two protonated aminoacids (arginine and di(cid:173)
`ethyl homoarginine) which are close to the N terminus, mak(cid:173)
`ing this end hydrophilic. Near the C terminus, detirelix has
`several large organic aminoacids (naphthylalanine, p-chlo(cid:173)
`rophenylalanine, tryptophan, and tyrosine), making this end
`lipophilic ( 14). The aggregation of amphiphiles is known to
`be driven by '' opposing forces' ' of attraction of the lipophilic
`regions and geometrical packing constraints, the balance of
`which is responsible for the characteristics and maintenance
`of subsequent structures (15), Detirelix and nafarelin have
`such lipophilic and hydrophobic regions, making them both
`surface active and highly susceptible to aggregation in aque(cid:173)
`ous solution.
`The apparent surface tension of detirelix in pH 7 .4 phos(cid:173)
`phate buffer showed a classical (15) -y versus log[peptide]
`profile (Fig. 4), giving from the sharp change in slope a crit(cid:173)
`ical micelle concentration (CMC) of 5.3 x 10- 4 M (0.88
`mg/ml). The limiting slope below the CMC, '3-y/alog[peptide],
`is ~ 6.6 mN m - 1
`, providing via the Gibbs equation [Eq. (3)]
`a surface excess concentration (Y) of 1.16 x 10- 6 mol m2.
`
`I
`8-y
`A = 2 -.-30_3_R_T 8log[peptide]
`
`(3)
`
`(4)
`
`Using Eq. (4), where NA is Avagadros number, the calcu(cid:173)
`lated surface area/molecule (A) is 1.43 nm2•
`A = 1/(NAY)
`Inspection of Fig. 4 shows that -y for nafarelin also de(cid:173)
`creases with increasing nafarelin concentration at pH 7.4.
`, giving Y = 1.7 x 10- 6
`The slope of this plot is 9.8 mN m- 1
`mot m2 and a surface area/molecule of 0.97 nm2
`• The surface
`area per molecule for nafarelin is slightly smaller than for
`detirelix, presumably because of the different head group
`size. It is known that the size of the hydrophilic head group,
`rather than the size or the le ngth of the lipophilic chain,
`affects the value of Y (16-19). Nafarelin has a hydrophilic
`region consisting of the sequence Gly-Pro-Arg, whereas de(cid:173)
`tarelix has a slightly greater hydrophilic region, consisting of
`
`FRESENIUS EXHIBIT 1008
`Page 3 of 6
`
`
`
`Parental Peptide Formulations
`
`1261
`
`Table D. Secondary Rate Constants and Activation Parameters for the Degradation of LHRH in Aqueous Solution"
`
`Rate
`constant
`
`25•c
`
`60°C
`
`so·c
`
`100°c
`
`Temperature
`
`kH- (M- 1 sec- ')
`
`).4 X )0- Sb
`
`3.0 X 10 - 4
`
`J.3 X 10 - 3
`
`5.0 X 10 - 3
`
`k0 (sec - 1)
`
`ktt0 - (M- 0·64 sec - 'Y
`
`5.3 X 10- tOb
`
`2.5 X 10 - 8
`
`1.4 X 10- 7
`
`8.3 X J0- 7
`
`2.5 X IQ - Sb
`
`8.5 X 10- 4
`
`5.0 X 10- 3
`
`2.2 X 10- 2
`
`Activation parameters
`(illf+, kcal mol - 1
`)
`(as;, cat K - 1 mo1 - 1)
`
`am =
`16.6
`as:1: = - 25.0
`aH+ = 20.8
`as+ = - 30.9
`am =
`19.3
`as:t: = - 14.8
`
`° Calculated using Eq. (2) (see text). The values of Kw used in calculating aH0 - from the pH's at 25, 60, 80, and 100°C were 14.0, 13.02,
`12.60,and 12.29 M - 1
`, respectively.
`b Calculated from the activation parameters shown in the last c-0lumn.
`< The secondary rate constants for hydroxide ion catalysis demonstrated a 0.64 exponent dependence on hydroxide ion concentration, and
`so the value of kH0 - has the units of M - 0 -64 sec- 1
`• Fortunately, this does not detract from its usefulness in calculating the pH-rate profile
`at 25°C.
`
`the peptide liquid crystal structures. Herein we have chosen
`the critical melting temperature (TcnJ as an indicator for
`characterizing peptide liquid crystals. In general, Tern was
`independent of liquid crystal maturation (Fig. 6) and in(cid:173)
`creased only slightly with increasing peptide concentration
`(Figs. 7 and 8). In the absence of added electrolyte, detirelix
`and nafarelin did not form liquid crystals at concentrations
`below 4 and 8 mg/ml, respectively. That detirelix formed
`liquid crystals at a lower concentration than nafarelin is in
`keeping with Tanford' s suggestion (15) that hydrophobic
`self-association is the driving force behind aggregate forma(cid:173)
`tion-by this criterion detirelix is more hydrophobic than
`nafarelin (Fig. 1). This is also supported by our observation
`that the hydrophilic parent peptide, native LHRH, did not
`exhibit birefringence or anisotropicity, even at 30 mg/ml
`LHRH in l M NaCl solution. Under these extreme condi(cid:173)
`tions both nafarelin and detirelix gel immediately, indicative
`of their hydrophobic nature as compared with native LHRH.
`Thus, both indirect (surface activity and determination
`
`Gly-Pro-Arg-Leu-(homodiethyl Arg). Further, the CMC
`for nafarelio is greater than 2 mg/ml, somewhat higher than
`for detirelix, presumably because nafarelin has a weaker
`lipophilic attraction than detirelix, or because the larger head
`group size for detirelix prevents close packing of the peptide
`monomers to give micelles and higher ordered stmctures.
`Liquid Crystal Formation. The liquid crystal formation
`of nafarelin and detirelix was studied by freeze-fracture elec(cid:173)
`tron microscopy and by examination under crossed polar(cid:173)
`ized light using optical microscopy. Upon freeze-fracture
`analysis, concentrated solutions of detirelix (for example, 15
`mg/ml as shown in Fig. 5) formed lyotropic mesophases,
`indicating a marked ordering of drug in solution. Similar
`freeze-fracture replicates were also observed for nafarelin
`solutions but were absent in control solutions containing
`only buffer (data not shown). Structural ordering of this type
`has been observed previously, for example, in freeze(cid:173)
`fracture replicates of emulsions and creams (20). Peptide
`aggregation in other peptides and proteins has also been re(cid:173)
`ported, for example, for pentagastrin (2 1) and insulin (22).
`Birefringence due to liquid crystal formation in aqueous
`nafarelin and detirelix solutions was determined by sample
`observation under crossed polars. Although measurement of
`birefringence is a sensitive method for the detection of liquid
`crystals, the extent of birefringence is difficult to quantify,
`especially because of its time-dependent nature and un(cid:173)
`known relation to the extent or thermodynamic stability of
`
`-4
`-5
`-6
`-3
`Log Concentration (M)
`Fig. 4. Effect of detirelix (e ) and nafarelin (0) concentration on the
`apparent surface tension (-y) at 25°C. The sharp break at 5.3 x 10- 4
`Mis the critical micelle concentration for detirelix in 0.01 M phos(cid:173)
`phate buffer (pH 7.4).
`
`·ff)/Jj)
`
`._ . >J;•_ •
`
`- 300 nm - ->
`<-
`Fig. 5. A freeze-fracture replicate of detirelix in aqueous solution
`( 15 mg/ml) as viewed by transmission electron microscopy. Control
`electron photomicrographs (not shown) without added peptide did
`not exhibit the ordered array in this figure.
`
`FRESENIUS EXHIBIT 1008
`Page 4 of 6
`
`
`
`1262
`
`Powell, Sanders, Rogerson, and Si
`
`65 ~
`I
`
`3
`
`e
`15
`
`'
`'
`•
`12
`9
`6
`Time (davs)
`Fig. 6. Effect of reaction time on the critical melting temperature
`(Tcml of nafarelin (18.8 mg/ml) and detirelix (19.9 mg/ml) solutions.
`Generally, a slight increase was noted, however, it is not considered
`large enough to complicate the cosolvent addition· results presented
`herein ( vide infra).
`
`40 -
`
`L .. .. ,
`
`5
`
`1 .5%◊ 20% 0
`5% "' 30% 0
`1S0/o v'
`
`, _ _J
`30
`25
`20
`15
`10
`Concentration (mg / ml)
`Fig. 8. Effect of PG (%, w/v) as added cosolvent on the critical
`melting temperature (fem) of detirelix liquid crystals.
`
`of CMC) and direct measurements (optical observation)
`demonstrate that hydrophobic LHRH analogues aggregate
`readily in aqueous solution. Because the aggregation of dc(cid:173)
`tirelix and nafarelin has important pharmaceutical conse(cid:173)
`quences, particularly in the design of peptide parenteral for(cid:173)
`mulations, the remainder of this paper describes how liquid
`crystal formation may be suppressed by added cosolvent.
`Peptide Liquid Crystals. These were disrupted by add(cid:173)
`ing organic cosolvent where, in general, increasing amounts
`of propylene glycol (PG) and higher temperatures sup(cid:173)
`pressed the formation of peptide liquid crystals. Inspection
`of Fig. 7 shows that added PG reduced Tern• particularly
`above 10-15% PG. [Similar decreases in Yem have been re(cid:173)
`ported previously, for example, by the addition of glycerol
`or ethylene glycol to disodium cromoglycate/water systems
`(23)) . The lack of liquid crystals in nafarelin solutions con(cid:173)
`taining more than 15% PG (w/v) may be due to (i) a reduction
`in Tern to room temperature, thus preventing liquid crystal
`formation on a thermodynamic basis, or (ii) a large kinetic
`barrier, resulting in the slow formation of nafarelin liquid
`crystals in the presence of excess PG. This latter rationale is
`supported by the long and variable lengths of time required
`to induce nafarelin liquid crystal formation , even for nafare(cid:173)
`lio solutions without added PG.
`Detirelix exhibited lower Tem's than nafarelin (Fig. 8),
`indicating that detirelix liquid crystals are Jess stable than
`nafarelin liquid crystals at elevated temperatures (40-80°C).
`This is somewhat surprising inasmuch as (i) detirelix exhibits
`a lower CMC (0.88 mg/ml) than does nafarelin, (ii) detirelix
`at 10 mg/ml or greater readily forms liquid crystals in less
`than 1 hr, whereas nafarelin often takes more than l week,
`
`80
`
`G
`~ 70
`i-!'
`
`60
`
`5% c.
`15% □
`
`25
`20
`15
`10
`5
`Concentration (mg / ml)
`Fig. 7. Effect of PG (%, w/v) as added cosolvent on the critical
`melting temperature (Tcml of nafarelin liquid crystals.
`
`and (iii) detirelix forms liquid crystals at a lower concentra(cid:173)
`tion (~ 4 mg/ml) than nafarelin (~ 8 mg/ml). The higher Tem's
`for nafarelin demonstrates that there exists a balance be(cid:173)
`tween kinetic and thermodynamic forces affecting the for(cid:173)
`mation and stability of peptide liquid crystals. These results
`indicate that, even though nafarelin forms liquid crystals
`slowly, once formed they are more stable. Such an observa(cid:173)
`tion is a breakdown of the Hammond postulate (a positive
`correlation between ilG° and iiG+), a tenet of chemistry
`that, unfortunately, has many exceptions.
`
`CONCLUSIONS
`
`Comparison of the pH-rate data herein for native
`LHRH with data reported earlier for nafarelin (7) and RS-
`26306 (8) demonstrates that hydrophobic modification of
`LHRH has little effect on chemical stability. ln contrast, the
`hydrophobic nature of certain LHRH decapeptides such as
`nafarelin and detirelix gives rise to unusual solution dynam(cid:173)
`ics, including peptide aggregation and liquid crystal forma(cid:173)
`tion. These liquid crystals eventually result in solution gela(cid:173)
`tion, with the ramification of compromised formulation ele(cid:173)
`gance and utility. We have demonstrated herein that a
`pharmaceutically acceptable cosolvent, propylene glycol,
`raises the minimum concentration for peptide aggregation,
`and thus, many of the potential problems associated with
`aggregate formation may be minimized or eradicated by co(cid:173)
`solvent addition.
`
`ACKNOWLEDGMENTS
`
`We are indebted to Nancy Benson, Department of Elec(cid:173)
`tron Microscopy, University of California, Berkeley, for the
`production of freeze -fracture replicates and to Dave
`Johnson, Jeff Fleitman, Peter Mishky, Katrina Herb, and
`John Nestor for their help and comments.
`
`REFERENCES
`
`I. S. T. Anik and J .-Y. Hwang. Adsorption of D-Nal(2)6LHRH, a
`decapeptide, onto glass and other surfaces. Int. J. Pharn1.
`16:795-S0I (1983).
`2. A. Rogerson and L. M. Sanders. Some solution properties of
`two decapeptides. Proc. Int . Symp . Control. Rel. Bioact.
`Maier. 14:97-98 (1987).
`3. J. A. Feldman, M. L. Cohn, and D. Blair. Liquid chromato(cid:173)
`graphic analysis of adriamycin and metabolites in biological flu(cid:173)
`ids. J. Liq. Chromatogr. 1:833 (1978).
`4. K. Nishi, H. Ito, S. Shinagawa, C. Hatanaka, M. Fujino, and
`
`FRESENIUS EXHIBIT 1008
`Page 5 of 6
`
`
`
`Parental Peptide Formulations
`
`1263
`
`M. Hattori. Racemization of L H-RH analogs in alkaline solu(cid:173)
`tion. In H. Yonehara (ed.), Peptide Chemistry, Protein Re(cid:173)
`search Foundation, Osaka, 1980, pp. 175-180.
`5. D. C. Sertl, R. N. Johnson, and B. T. Kho. An accurate spe(cid:173)
`cific HPLC method for the analysis of a decapeptide in a lactose
`matrix. J. Liq. Chroma1ogr. 4:1135 (1981).
`6. J. Winterer, D. Chatterji, F. Comite, M. H. Decker, D. L. Lo(cid:173)
`riaux, J. F. GalleUi, and G. B. Cutler, Jr. Thermal stability of a
`long-acting analogue of leutinizing hormone releasing hormone
`(D-Try6-Pro9-Net-LHRH). Contraception 27:195 (1983).
`7. D. M . Johnson, R. A. Pritchard, W. F. Taylor. D. Conley. G .
`Zuniga, and K. G. McGreevy. Degradation of the hydrophobic
`LHRH analogue nafarelin in aqueous solution. Im. J. Pharm.
`31: 125-129 (1986).
`8. R. G. Strickley, M. Brandl, K. W. Chan, K. Straub, and L. Gu.
`High performance liquid chromatographic (HPLC) and HPLC(cid:173)
`mass specrometric (MS) analysis of the degradation of the leu(cid:173)
`tinizing hormone-releasing hormone (LH-RH) antagonist RS-
`26306 in aqueous solution. Pharm. Res. 7:53~536 (1990).
`9. A. R. Oyler, R. E . Naldi, J . R. Lloyd, D. A. Graden, C. J.
`Shaw, and M. L. Cotter. Characterization of the solution deg(cid:173)
`radation products of histrelin, a gonadotropin releasing hor(cid:173)
`mone (LH/RH) agonist. J. Pharm. Sci. 80:271-275 (1991).
`IO. V. Si, M. F. Powell, B. J. Floy, and J. S. Fleitman. Parenteral
`peptide formulations. 2. Effect of counrerions, pH and excipi(cid:173)
`ents on peptide aggregation in aqueous solution (in preparation)
`(1991).
`II. W. C. McCrone, L.B. McCrone, and J . G. DelJy. Polarized
`Light Microscopy, Mccrone Research Institute, Chicago, 1987.
`12. F . D. Bloss . An Introduction to the Methods ofOp1ical Micros(cid:173)
`copy, Holt Rinehard and Winston, New York, 1961.
`13. P. R. Bevington. Data Reduction and Error Analysis in the
`Physical Sciences, McGraw-Hill New York, 1969.
`
`14 . J. J. Nestor,Jr., T. L. Ho, R. A.Simpson,B. L. Horner,G. H.
`Jones, G . I. McRae, and B. H. Vickery. Synthesis and biolog(cid:173)
`ical activity of some very hydrophobic superagonist analogues
`of leutinizing hormone-releasing hormone. J . Med. Chem .
`2S:795-801 (1982).
`15. C. Tanford. The Hydrophobic Effect, John Wiley & Sons, New
`York, 1973.
`16. D. Attwood and A. T. Florence. Surfacta111 Systems. Their
`Chemistry, Pharmacy and Biology. Chapman and Hall, New
`York 1983.
`17. I. Lo, A. T. Florence, J .-P. Treguier, M. Seiller, and F. Pui(cid:173)
`sieux. The influence of surfactant HLB and the nature of the oil
`phase on the phase diagrams of nonionic surfactant-oil-water
`systems. J. Colloid Interface Sci. 59:3 19 (1977).
`18. J .-P. Treguier, I. Lo, M. Seiller, and F. Puisieux. Emulsions and
`water-surfactant-oil diagrams. A system of water-Brijs 92 and
`96-oil of Vaseline. Effect of surfactant hydrophilicity. Pharm.
`Helv. Acta. 50:421 (1975).
`19. I. Lo, F. Madsen, A. T. Aorence, J.-P. Treguier, M. Seiller,
`and F. Puisieux. In K. L. Mittal (ed.), Micellization, Solubili(cid:173)
`zation and Microemu/sions, Vol. 1, Plenum. New York.
`20. C. Mueller-Goymann. Liquid crystals in emulsions, creams, and
`gels containing ethoxylated sterols as surfactant. Phartn. Res.
`4:154-158 (1984).
`21. D. Attwood, A. T. Florence, R. Greig, and G. A. Smail, Some
`solution properties of pentagastrin and angiotensin: Aggregation
`of pentagastrin. J. Pharm. Pharmacol. 26:847-853 (1974).
`22. E.W. Kraegen, L. Lazarus, H. Meler, L. Campbell, and Y. 0.
`Chia. Carrier solutions for low level intravenous insulin solu(cid:173)
`tions. Br. Med. J. 23:464-466 (1975).
`23. H. Lee and M. M. Labes. Phase diagram and thermodynamic
`properties of disodium cromoglycate lyomesophases. Mo/.
`Crys1. Liq. Crysl. 91:53- 58 (1982).
`
`FRESENIUS EXHIBIT 1008
`Page 6 of 6
`
`