Pharmaceutical Development and Technology, 5(1), 123–130 (2000)
`
`RESEARCH ARTICLE
`
`Assessment of the Myotoxicity of
`Pharmaceutical Buffers Using an In Vitro
`Muscle Model: Effect of pH, Capacity,
`Tonicity, and Buffer Type
`
`Jintana Napaporn,1 Maike Thomas,1 Kari A. Svetic,1 Zahra
`Shahrokh,2 Gayle A. Brazeau1
`1DepartmentofPharmaceutics,CollegeofPharmacy,UniversityofFlorida,Box
`100494JHMHC,Gainesville,Florida32610
`2GenentechInc.,DepartmentofPharmaceuticalR&D,1DNAWay,SouthSan
`Francisco,California94080
`
`Received March 1, 1999; Accepted August 4, 1999
`
`ABSTRACT
`
`The purpose of the present study was to investigate the myotoxicity of three buffers containing carbox-
`ylic acid groups (i.e., acetate, succinate, and citrate) as a function of their pH, capacity, and tonicity.
`The myotoxicity of these buffers in the range of pH 2–6 and 0.001–0.1 M buffer capacity was assessed
`using cumulative creatine kinase (CK) release from an isolated rodent muscle model following injec-
`tion. Phenytoin and 0.9% NaCl injection were used as positive and negative controls, respectively.
`Buffer solutions were prepared. A lower pH and higher buffer capacity was linked to increased myo-
`toxicity for the acetate buffers. However, for succinate and citrate buffers, pH appeared to influence
`the extent of myotoxicity, whereas buffer capacity did not seem to have an effect. When either NaCl
`or trehalose was used as a tonicity-adjusting agent at pH 6, isotonic 0.01 M buffer solutions dramati-
`cally lowered the cumulative CK release compared to those that were not isotonic. Isotonic succinate
`buffers displayed the lowest myotoxicity, whereas citrate buffers displayed the highest values. Citrate
`buffers containing three carboxylic acid groups showed higher myotoxicity than succinate buffers and
`acetate buffers at 0.001 and 0.01 M buffer capacities, whereas acetate buffer produced higher cumula-
`tive CK release than citrate and succinate buffers at 0.1 M buffer capacity. The myotoxicity of pharma-
`ceutical buffers containing carboxylic acid groups appears to be directly affected by lowering the pH
`of the solution.
`KEY WORDS: Buffer; Capacity; Creatine kinase; Intramuscular injection; Myotoxicity; pH; Tonicity.
`
`Address correspondence to Jintana Napaporn. Fax: (352) 392-4447. E-mail: jintana@grove.ufl.edu
`
`Copyright (cid:211)
`
`2000 by Marcel Dekker, Inc.
`
`www.dekker.com
`
`123
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 1 of 8
`
`

`

`124
`
`Napaporn et al.
`
`INTRODUCTION
`
`The intramuscular injection of drugs is a well-estab-
`lished and frequently used procedure of drug administra-
`tion because it can provide relatively fast drug systemic
`availability when oral or intravenous administration is or
`cannot be achieved because of the biopharmaceutical
`characteristics of the drug (e.g., gastrointestinal instabil-
`ity or low aqueous solubility). Furthermore, the intramus-
`cular route provides a means for sustained drug delivery
`and is conceivably an important route for the delivery of
`new therapeutic peptides and/or gene products (1). Un-
`fortunately, the incidence of injection site complications
`following single or multiple intramuscular injections in
`humans can be as high as 20% depending upon the for-
`mulation and the site of injection (2). These complica-
`tions range from minor discomfort and pain to skeletal
`muscle damage (defined as myotoxicity), which will po-
`tentially limit clinical acceptance. Myotoxicity following
`the intramuscular injection of a parenteral formulation
`could be a function of the therapeutic agent (e.g., phenyt-
`oin, cephalosporins, local anesthetics) or the presence of
`other excipients (e.g., solvents, buffer) in the formulation
`(1,2).
`formulations use pharmaceuti-
`Many commercial
`cal buffer systems (primarily acetate, succinate, citrate,
`phosphate) for one or all of the following reasons: (a) to
`reduce discomfort to the patients, (b) to ensure requisite
`drug stability and solubility, and (c) to control the thera-
`peutic activity of the drug substance (3). Parenteral for-
`mulations using these buffers and other excipients have
`the potential to cause damage and/or pain following in-
`jection. Citrate buffer has been reported to cause signifi-
`cantly more pain than 0.9% NaCl injection after subcu-
`taneous administration in patients (4). Furthermore, the
`degree of pain was related to the injected concentration of
`citrate buffer. Also, buffers containing carboxylic acids
`(acetate, succinate, citrate) have been shown to cause red
`blood cell agglutination and/or hemolysis in patients (5).
`Although these isolated studies and anecdotal clinical re-
`ports indicate that buffers used in parenteral formulations
`may be responsible for the pain and/or damage following
`injection, a systematic investigation of buffers and buffer
`properties with respect to their potential to cause tissue
`damage has not been conducted to date.
`On the basis of limited data on parenteral formulations
`and their excipients, solutions to be applied are likely to
`cause toxicity if their pH is extremely high or low com-
`pared to the pH of the relevant body fluid or tissue. Of
`possible greater significance than the actual pH of the
`
`solution is the buffer capacity and the volume to be used
`relative to the volume of body fluid in which the solution
`will come into contact following injection (6). It can,
`therefore, be hypothesized that myotoxicity of a formula-
`tion with a large pH difference between the solution be-
`ing administered and the physiological environment in
`which it is applied will be minimized by lowering the
`buffer capacity of the solution. For parenteral products
`using buffers, the desired pH range, capacity, tonicity,
`buffer type, and tissue damage are critical factors that
`formulation scientists must consider in the development
`and optimization of these formulations.
`Creatine kinase (CK) release is often used as a marker
`of damage to skeletal muscle both in vitro and in vivo
`(7). In the present studies, an in vitro technique that mea-
`sures the release of cytosolic enzyme CK from an isolated
`rat muscle model was employed to screen a series of buff-
`ers for their potential to cause acute skeletal muscle dam-
`age following an intramuscular injection (8). The specific
`purpose of the present study was to investigate the myo-
`toxicity of buffers containing a carboxylic acid group
`(i.e., acetate, succinate, and citrate) as a function of their
`pH, capacity, tonicity, and buffer type. These investiga-
`tions provide the first myotoxicity data for three buffer
`systems commonly used in intramuscular formulations.
`These studies also illustrate a rational approach for test-
`ing of intramuscular injection buffer systems, accounting
`for buffer-induced myotoxicity.
`
`MATERIALS AND METHODS
`
`Materials
`
`The muscle incubation medium, balanced salt solution
`(BSS), contains 116 mM NaCl, 5.4 mM KCl, 5.6 mM
`dextrose, 26.2 mM NaHCO3, and 0.001 mM sodium phe-
`nol red in Sterile Water for Irrigation (Baxter, Deerfield,
`IL). The pH was adjusted to 7.4 with 1 M HCl. Acetic
`acid, succinic acid, citric acid, and their salts were ob-
`tained from Fisher Scientific Co. (Fair Lawn, NJ) and
`Sigma Chemical Co. (St. Louis, MO), respectively. Ster-
`ile Water for Irrigation was used in preparation of all
`buffer solutions. Phenytoin (Dilantin(cid:210) ) and 0.9% NaCl
`injection, the positive and negative controls, were manu-
`factured by Parke-Davis (Morris Plains, NJ), and Abbott
`Laboratories (North Chicago, IL), respectively. All other
`chemicals were at least reagent grade, the highest grade
`commercially available, and were obtained from Sigma
`Chemical Co.
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 2 of 8
`
`

`

`Myotoxicity of Pharmaceutical Buffers
`
`Preparation of Test Formulation
`
`The buffer solutions to be tested at 0.001, 0.01,
`and 0.1 M buffer capacity were prepared by using the
`Henderson–Hasselbalch equation and the Van Slyke equa-
`tion according to their pKa values. To evaluate the effect
`of tonicity, 0.01 M buffer solutions at muscle physiologi-
`cal pH (pH 6) were prepared isotonically using either
`NaCl or trehalose as tonicity-adjusting agents. The sec-
`ond tonicity study was conducted using NaCl in the range
`of 0–2.7% w/v. To compare buffer types and to test the
`possible relationship between buffer capacity, number of
`carboxylic acids, and myotoxicity, buffer solutions were
`prepared by varying buffer capacity from 0.001 to 0.1 M
`at pH 6.
`
`125
`
`tional Institutes of Health Guidelines. The EDL muscles
`were injected with the test solution (15 m
`l) using a 100
`l Hamilton syringe (Reno, NV) equipped with a needle
`guard to control the depth and angle of injection. The
`injected muscles were placed into a Teflon(cid:210) -coated plas-
`tic basket and immersed in 9 ml of carbogenated (95%
`O2/5% CO2) BSS. The viability of this isolated muscle
`in this system was previously shown to be 2 hr (8).
`The solutions were drained and fresh BSS added at 30-
`min intervals. These drained solutions were analyzed for
`CK activity at 30, 60, 90, and 120 min. The assay was
`run at 30(cid:176) C and validated using a standard (Accutrol Nor-
`mal). Myotoxicity was assessed by the cumulative re-
`lease of CK into the incubation medium over a 2-hr pe-
`riod (11).
`
`CK Activity in In Vitro Interference Assay
`
`Data Analysis
`
`All of the test formulations were evaluated to deter-
`mine whether they stimulated or inhibited CK activity.
`Briefly, as described previously (9), rabbit muscle CK
`Type I (lyophilized form, Sigma Chemical Co.) was pre-
`pared by dissolving approximately 1 mg of the enzyme
`in 10 ml of BSS at pH 7.4. A given aliquot of this solution
`was spiked into incubation vessels containing the BSS at
`37(cid:176) C and bubbled with 95% O2/5% CO2. The test solu-
`tion was added to the test incubation vessel, while the
`same volume of 0.9% NaCl injection served as control.
`The amount of CK, approximately 500 U/L, was the
`same in both the test and control incubation vessels. All
`studies were conducted for 30 min. CK was determined
`spectrophotometrically at 340 nm using a commercially
`available kit (no. 47-UV, Sigma Chemical Co.). This
`assay is based upon the change in the absorbance caused
`by a reduction of nicotinamide adenine dinucleotide
`(NAD1) to reduced nicotinamide adenine dinucleotide
`(NADH). The degree of interference with CK activity
`was evaluated by the ratio of CK activity in the presence
`of the test solution to the activity measured in the absence
`of the test solution.
`
`In Vitro Myotoxicity Protocol
`
`Rodent extensor digitorum longus (EDL) muscles
`(approximately 100–200 mg) were isolated from male
`Sprague Dawley rats as previously described (10).
`Briefly, rodents were administered an anesthetic dose of
`sodium pentobarbital and sacrificed via cervical disloca-
`tion as approved by the Animal Care and Use Committee
`at the University of Florida in accordance with the Na-
`
`Data were presented as the mean and standard error
`of mean with n 5 12 samples per test solution for in vitro
`interference assay and n 5 4 muscles per treatment for
`in vitro myotoxicity studies. Statistical analysis of cumu-
`lative CK activity in the different treatments was per-
`formed using the SAS program for completely random-
`ized factorial design to screen for the main effect and
`the interaction effect of pH and buffer capacity for each
`buffer type. A Duncan’s post-hoc test with p , 0.05 was
`considered to be statistically significant in the different
`treatments. A subsequent breakdown analysis was ap-
`plied in order to analyze for the simple effects of pH and
`buffer capacity at the particular level of each factor (12).
`The differences between the positive control (Dilantin)
`and negative control (0.9% NaCl injection) were ana-
`lyzed using analysis of variance (ANOVA) with p , 0.05
`considered to be statistically significant.
`
`RESULTS
`
`Buffer Studies
`
`In all treatment groups, the muscles were similar in
`length and weight, ranging from 3.5 to 4.5 cm in length
`and 90 to 125 mg in weight. The cumulative CK release
`of the positive (Dilantin) and negative (0.9% NaCl injec-
`tion) controls is shown in Fig. 1. The myotoxicity of Di-
`lantin was 14 times higher ( p , 0.05) than that of the
`0.9% NaCl injection.
`The second series of studies was conducted to deter-
`mine if the buffer solutions altered the activity of CK
`(Table 1). The interference assay showed that only the
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 3 of 8
`
`m
`

`

`126
`
`Napaporn et al.
`
`Table 1
`
`In Vitro Interference Assay of CK Activity of Tested Buffer
`Solutions at 30 min (37(cid:176) C) (n 5 12)
`
`Treatment
`
`Acetate buffer solutions
`0.001 M pH 4
`0.01 M pH 4
`0.1 M pH 4
`0.001 M pH 6
`0.01 M pH 6
`0.1 M pH 6
`0.01 M pH 6 isotonic with NaCl
`0.01 M pH 6 isotonic with trehalose
`Succinate buffer solutions
`0.001 M pH 2
`0.01 M pH 2
`0.1 M pH 2
`0.001 M pH 5
`0.01 M pH 5
`0.01 M pH 5
`0.001 M pH 6
`0.01 M pH 6
`0.1 M pH 6
`0.01 M pH 6 isotonic with NaCl
`0.01 M pH 6 isotonic with trehalose
`Citrate buffer solutions
`0.001 M pH 2
`0.01 M pH 2
`0.1 M pH 2
`0.001 M pH 4
`0.01 M pH 4
`0.1 M pH 4
`0.001 M pH 6
`0.01 M pH 6
`0.1 M pH 6
`0.01 M pH 6 isotonic with NaCl
`0.01 M pH 6 isotonic with trehalose
`
`CK Activity Ratio,
`Mean (SD)
`
`0.86 (0.01)
`1.07 (0.11)
`0.93 (0.04)
`1.07 (0.11)
`1.06 (0.09)
`1.06 (0.08)
`0.90 (0.05)
`0.96 (0.05)
`
`1.83 (0.34)a
`0.99 (0.03)
`1.16 (0.11)
`1.15 (0.15)
`1.08 (0.09)
`1.01 (0.06)
`0.97 (0.07)
`1.01 (0.05)
`0.96 (0.04)
`0.97 (0.06)
`0.99 (0.07)
`
`0.99 (0.11)
`0.96 (0.07)
`1.23 (0.21)a
`0.95 (0.13)
`0.90 (0.07)
`1.07 (0.06)
`0.93 (0.07)
`1.05 (0.10)
`1.18 (0.12)
`0.98 (0.05)
`0.90 (0.06)
`
`a Statistically higher than enzyme activity ratio 5 1.0 (0.9% NaCl injec-
`tion), p , 0.05.
`
`Tonicity Studies
`
`Previous reports from the literature indicate that there
`is no hemolysis when red blood cells (RBCs) are exposed
`to NaCl between 0.7 and 1.4% (w/v) (13). To investigate
`if this observation is consistent in muscle, NaCl solutions
`of various concentrations ranging from hypotonic to
`hypertonic, including 0, 0.2, 0.45, 0.7, 0.9, 1.4, 1.8, and
`2.7%, were tested for their myotoxic potential (Fig. 3).
`
`Figure 1. Myotoxicity of positive control (Dilantin) com-
`pared to negative control (0.9% NaCl injection). *Statistically
`significant compared to 0.9% NaCl injection (p , 0.05, n 5
`4, mean 6 SEM).
`
`0.1 M pH 2 citrate and 0.001 M pH 2 succinate solutions
`significantly increased the activity of CK enzyme by 23
`and 83%, respectively (p , 0.05).
`The cumulative CK release of acetate buffers ranged
`from 350 to 1100 U/l [Fig. 2(a)]. Only cumulative CK
`release at 0.1 M solutions for pH 4 and 6 was significantly
`higher than that for 0.9% NaCl injection ( p , 0.05).
`There was a trend of increasing CK release with increas-
`ing buffer capacity and decreasing pH, although values
`were not statistically different from 0.9% NaCl injection
`or from each other. This suggests that the acetate buffers
`with lower pH and higher buffer capacity are more myo-
`toxic (p , 0.05).
`For succinate buffers, the maximum cumulative CK
`release was similar to that of the acetate buffers [Fig.
`2(b)]. A comparison of pH 2 and 5 buffer solutions at a
`constant buffer capacity suggests that pH influences the
`myotoxicity more than does buffer capacity. At all three
`buffer capacities, the myotoxicities of pH 2 buffers were
`statistically higher from 0.9% NaCl injection, while pH
`5 buffers were not different ( p , 0.05).
`The myotoxicity of citrate buffers, in the range of
`500–1750 U/l, was greater than those of acetate and suc-
`cinate buffers [Fig. 2(c)]. Most citrate buffers tested (8/
`9) exhibited significantly higher cumulative CK release
`than did 0.9% NaCl injection. Only 0.1 M pH 6 buffer
`was not different from 0.9% NaCl injection ( p , 0.05).
`Statistical analysis revealed that pH is the important con-
`sideration with respect to the myotoxicity of this buffer
`species. In addition, a higher buffer capacity (0.1 M) is
`associated with less myotoxicity of this buffer at muscle
`pH 6 (e.g., pH 6).
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 4 of 8
`
`

`

`Myotoxicity of Pharmaceutical Buffers
`
`127
`
`Figure 3. Myotoxicity of saline solutions in concentrations
`varying from 0 to 2.7% w/v. The corrected cumulative CK
`was calculated from the degree of interference of test solution
`compared to 0.9% NaCl injection obtained during interference
`assay. *Statistically significant from 0.9% NaCl injection ( p ,
`0.05, n 5 4, mean 6 SEM).
`
`higher toxicity compared to lower buffer capacities.
`However, it was unclear whether myotoxicity could be
`reduced when hypotonic buffers were made isotonic with
`appropriate tonicity-adjusting agents. Because 0.1 M
`buffer capacity is hypertonic, 0.01 M buffers were made
`isotonic using either NaCl or trehalose at pH 6. Isotonic
`buffer solutions showed dramatically lower cumulative
`CK than the corresponding hypotonic, pH 6 solutions
`( p , 0.05), as shown in Fig. 4. Furthermore, isotonic
`
`Figure 4. Myotoxicity of 0.01 M pH 6 buffer solutions to
`screen the effect of buffer tonicity. Solutions were prepared iso-
`tonically using either NaCl or trehalose as tonicity agent (n 5
`4, mean 6 SEM).
`
`Figure 2. Myotoxicity of buffers at 0.001, 0.01, 0.1 M buffer
`capacities. (a) Acetate buffer; (b) succinate buffer; (c) citrate
`buffer. *Statistically significant compared to 0.9% NaCl injec-
`tion (p , 0.05, n 5 4, mean 6 SEM).
`
`CK release following the injection of each solution was
`linear over the 2-hr period. In the absence of NaCl, myo-
`toxicity was seven times higher than that with 0.9%
`NaCl. Myotoxicity was dramatically reduced with 0.2%
`NaCl, but was significantly higher by threefold compared
`to that with 0.9% NaCl injection. There was no signifi-
`cant difference in myotoxicity observed in the range of
`0.45–2.7% NaCl.
`
`Buffer Tonicity Studies
`
`Previous tonicity studies suggested that higher buffer
`capacities (and most often hypertonic solutions) cause
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 5 of 8
`
`

`

`128
`
`Napaporn et al.
`
`tain buffers such as acetate, succinate, citrate, and phos-
`phate. To test whether skeletal muscle damage caused by
`buffers is related to their pH, buffer capacity, tonicity, or
`buffer types, a series of buffers containing varying num-
`ber of carboxylic acid groups was examined in the pres-
`ent study. Three possible explanations for myotoxicity
`can be proposed to account for the differences observed
`in this work: (a) the concentration of undissociated acids,
`(b) the concentration of hydrogen ions in solution, and
`(c) the fraction of dissociated acid species.
`For acetate buffers, there was a trend of increased my-
`otoxicity with increasing buffer capacity and decreasing
`pH at 0.001 and 0.01 M, all still comparable to saline
`myotoxicity. However at 0.1 M, the myotoxicity was sig-
`nificantly greater than that observed for 0.9% NaCl injec-
`tion. Acetate buffer (a monoprotic acid with a pKa 5
`4.76) at an acidic pH of 2 and 4 accepts hydrogen ions
`and converts some of the acetate anions to the undissoci-
`ated acid. When pH decreases below muscle pH (pH 6)
`(14), the molar concentration of undissociated acid in-
`creases. It appears that the myotoxicity of this buffer
`showed a relationship to the increasing molar concentra-
`tion of the undissociated acid. A higher buffer capacity
`caused increased myotoxicity, perhaps due to the increas-
`ing molar concentration of the undissociated acid. Be-
`cause the buffer capacity is affected not only by the salt/
`acid ratio, but also by the total concentrations of acid and
`salt, an increase in the buffer capacity of a solution results
`in a greater total concentration of buffer constituent at
`the injection site (6). Of the buffer components, the un-
`dissociated acid is predicted to be more damaging to tis-
`sues than the salt because it can dissociate to generate
`further H1. The increase in undissociated acid concentra-
`tion with increasing buffer capacity and decreasing buffer
`pH may be the cause of myotoxicity in this buffer.
`Succinate and citrate buffer solutions also showed the
`influence of pH on the extent of myotoxicity, whereas
`buffer capacity did not affect their myotoxicity. At acidic
`pH, the undissociated form prevails, affecting toxicity in
`a manner similar to the monoprotic acetate buffer. As
`polyprotic buffers, succinate (a diprotic acid with pKa1 5
`4.19 and pKa2 5 5.63) and citrate (a triprotic acid with
`pKa1 5 3.14, pKa2 5 4.80, and pKa3 5 6.40) have multi-
`ple species; the undissociated and partially dissociated
`acids which could serve as reservoirs of hydrogen ions
`resist changes in pH attempted by the tissue. Because the
`buffer capacity of polyprotic buffers is the sum of the
`individual buffer capacities of the multiple monoprotic
`groups (15), polyprotic buffer solutions would be able to
`better resist changes in pH caused by extracellular and
`intracellular water compared to monoprotic buffers. Con-
`
`Figure 5. Myotoxicity of buffer solutions at pH 6. Buffer
`capacity was varied in the range of 0.001 to 0.1 M in order to
`compare the effect of buffer type, and the number of carboxylic
`acid groups in molecules was varied. These data compare new
`succinate solutions at pH 6 and include data from Figs. 2(a)
`and (c), acetate and citrate solutions at pH 6, respectively.
`
`buffer solutions showed no significant differences in my-
`otoxicity compared to those with 0.9% NaCl injection
`(p , 0.05). Isotonic succinate buffers displayed the low-
`est myotoxicity, whereas citrate buffers displayed the
`highest myotoxicity values.
`
`Role of Carboxylic Acids
`
`To determine the relationship between myotoxicity
`and the number of carboxylic acids, buffer solutions were
`prepared at muscle pH (pH 6) with varying buffer capaci-
`ties in the range of 0.001–0.1 M. The cumulative CK
`release of these buffer solutions is shown in Fig. 5. Myo-
`toxicity of acetate buffers significantly increased with
`increasing buffer capacity, whereas myotoxicity of suc-
`cinate and citrate buffers showed the opposite trend. Cit-
`rate buffers containing three carboxylic acid groups (3-
`COOH) showed a trend toward higher myotoxicity than
`succinate buffers (2-COOH) and acetate buffers (1-
`COOH) at 0.001 and 0.01 M ( p , 0.05). Acetate buffers
`produced higher cumulative CK release than citrate
`buffer and succinate buffer at 0.1 M buffer capacity (p ,
`0.05).
`
`DISCUSSION
`
`Even though intramuscular injection is often used as
`a route for drug delivery, pain and/or muscle damage fol-
`lowing injection have been reported. Pain and/or skeletal
`muscle damage (myotoxicity) can be caused by the drug
`substance or other excipients (e.g., solvents or buffers).
`Many commercial products reported to cause pain con-
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 6 of 8
`
`

`

`Myotoxicity of Pharmaceutical Buffers
`
`129
`
`sequently, the pH of injected site remains approximately
`the same as that of the initial injected buffer pH for buff-
`ers that contain polyprotic carboxylic acids. Because my-
`otoxicity is already high at low buffer concentration for
`polyprotic buffers, further increases in buffer capacity do
`not further increase myotoxicity. Thus, there appears to
`be no direct relationship between buffer capacity and my-
`otoxicity for polyprotic buffers.
`According to our results, a lower pH is always associ-
`ated with higher myotoxicity. For buffered solutions, any
`changes in pH, and thereby changes in the electrochemi-
`cal potential of protons, are resisted (16). When protons
`are entering or leaving the solution at the tissue boundary,
`some of the diffusing protons are bound by the buffer
`and these protons diffuse as protonated buffer. Therefore,
`in addition to the undissociated acid, the free protons may
`also contribute to myotoxicity at the injection site.
`The tonicity studies were consistent with previous
`studies of saline-induced pain in humans (17) where 5%
`hypertonic solution caused greater pain than isotonic
`solution. Pain on injection of hypertonic saline solution
`may be related to the distribution pattern of isotonic sa-
`line and hypertonic saline (5%) after injection. The distri-
`bution of isotonic saline is localized proximal to the in-
`jection site, whereas hypertonic saline is distributed to a
`larger area. In the present study, minimal CK release was
`observed up to 2.7% NaCl. A much higher tonicity may
`be required to release CK in muscle tissue. This may be
`the reason that in our study, 2.7% saline still produced
`a result similar to that of the isotonic saline. On the
`lower end of tonicity, minimal CK release was observed
`down to 0.45% NaCl, consistent with no hemolysis in
`this range. Increased CK release at very hypotonic solu-
`tions (#0.2% NaCl) is concordant with hemolytic poten-
`tial.
`In this study, isotonic buffer solutions also showed a
`dramatically lower myotoxicity compared to nonisotonic
`solutions. This would suggest that the need to procure
`isotonic conditions in solutions used intramuscularly is
`considerable because of the localized distribution pattern
`and the osmotic pressure of the isotonic solution.
`Interestingly, succinate buffers showed the lowest cu-
`mulative CK release compared to 0.9% NaCl injection.
`The lower succinate myotoxicity may have been depen-
`dent upon muscle weight of rats used in the study. We
`previously found that when younger rats were used (as
`for the succinate studies), there was a lower potential for
`CK release from muscle. The effect of animal size and
`age appears to be substantial in the measurement of CK
`release, and the succinate data may have been con-
`founded by these parameters.
`
`An additional source of myotoxicity may be the num-
`ber of unprotonated carboxylic acids in buffers serving
`as chelating agents. To evaluate the effect of the different
`number of carboxylic acid groups, pH 6 buffer solutions
`of acetate, succinate, and citrate were compared. In this
`study, pH 6 was chosen because the hydrogen ions are
`below 1024 M, and presumably do not influence myotoxi-
`city. The molar concentrations of undissociated and dis-
`sociated acids were then calculated and the relationship
`to myotoxicity was examined. The more carboxylic acids
`in a molecule, the higher the fraction of dissociated acid
`present in solution. If the dissociated acid is involved in
`myotoxicity, the myotoxicity of citrate with the highest
`fraction of dissociated acid should be the highest relative
`to succinate and acetate. A tendency toward increased
`myotoxicity with increasing number of carboxylic acids
`in the molecule was indeed evident for 0.001 M buffer
`where maximal resolution of the effect was achieved.
`However, when buffer capacity was increased, especially
`for succinate and citrate buffers, the cumulative CK re-
`lease data were opposite to that of the results seen with
`the acetate buffer (Fig. 6). This may be the result of the
`high buffer capacity of polyprotic buffers, and also the
`variation in CK release due to the age and size of muscles
`used in this study.
`In conclusion, the myotoxicity of pharmaceutical buf-
`fers containing carboxylic acid groups (i.e., acetate, suc-
`cinate, and citrate) is affected by the pH of the solution.
`This is particularly true for acetate buffers. Increasing
`buffer capacity can also increase myotoxicity for acetate
`buffers. Succinate and citrate buffers, which are poly-
`
`Figure 6. Plot of mean myotoxicity of pH 6 buffers versus
`log of fraction of dissociated acid/undissociated acid. Buffer
`capacity was varied in the range of 0.001–0.1 M. For polyprotic
`buffers (i.e., succinate and citrate), the dissociated acid is calcu-
`lated as the sum of mono-, di-, and tridissociated species.
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 7 of 8
`
`

`

`130
`
`Napaporn et al.
`
`protic, do not show a significant effect of buffer capacity
`on the already high level of myotoxicity. To reduce myo-
`toxicity, buffer solutions with low concentrations should
`be made isotonic. For acetate buffer, we recommend asolu-
`tion formulated at low buffer capacity and near physiologi-
`cal pH. Although the myotoxicity of acetate buffers is not
`significantly different from 0.9% NaCl injection, a trend
`toward increased myotoxicity was evident at higher buffer
`capacity and lower pH. For succinate and citrate buffers,
`we recommend a solution formulated near physiological
`pH because the myotoxicity is minimized in this range.
`
`ACKNOWLEDGMENTS
`
`The authors would like to thank Adam Persky and Brett
`Houk for comments during the preparation of the manu-
`script; Lorena Barron for some buffer preparations; and
`Dr. Mike Powell
`for early discussions on acid-
`induced pain upon injection. This study was supported
`by the 1997 PHRMA Undergraduate Research Fellowship
`Program in Pharmaceutics to Ms. Kari Svetic.
`
`REFERENCES
`
`1. G. A. Brazeau, M. Sciame, S.A. Al-Suwayeh, and E. Fattal,
`Evaluation of PLGA microsphere size effect on myotoxi-
`city using the isolated rodent skeletal muscle model,
`Pharm. Dev. Technol., 1(3), 279–283 (1996).
`2. G. A. Brazeau and H.-L. Fung, Effect of organic cosolvent-
`induced skeletal muscle damage on the bioavailability of
`intramuscular [14C] diazepam, J. Pharm. Sci., 79(9), 773–
`777 (1990).
`3. H. C. Ansel, N. G. Popovich, and L. V. Allen, Jr., Pharma-
`ceutical Dosage Forms and Drug Delivery Systems, 6th
`ed., Williams & Wilkins, Baltimore, 1995.
`4. L. A. M. Frenken, H. J. J. Van Lier, J. G. M. Jordans,
`K. M. L. Leunissen, R. Van Leusen, V. M. C. Verstappen,
`and R. A. P. Koene, Identification of the component part
`in an epoetin alfa preparation that causes pain after subcu-
`taneousinjection,Am.J.Kidney Dis., 22(4), 553–556(1993).
`
`5. S. E. Howe, C. G. Sciotto, and D. Berkner, The role of
`carboxylic acids in EDTA-dependent panagglutination,
`Transfusion, 22(2), 111–114 (1982).
`6. N. Martin, J. Swarbrick, A. Cammarata, and A. H. Chun,
`Physical Principles in the Pharmaceutical Sciences, 4th
`ed., Lea & Febiger, Philadelphia, 1993.
`7. G. A. Brazeau, S. Al-Suwayeh, J. Peris, B. Hunter, and
`D. W. Walker, Creatine kinase release from isolated EDL
`muscles in chronic ethanol-treated rats, Alcohol, 12(2),
`145–149 (1995).
`8. G. A. Brazeau and H.-L. Fung, Use of an in vitro model
`for the assessment of muscle damage from intramuscular
`injections: in vitro-in vivo correlation and predictability
`with mixed solvent systems, Pharm. Res., 6(9), 766–771
`(1989).
`9. G. A. Brazeau and H.-L. Fung, Interferences with assay
`of creatine kinase activity in vitro, Biochem. J., 257, 619–
`621 (1989).
`10. G. A. Brazeau and H.-L. Fung, An in vitro model to eval-
`uate muscle damage following intramuscular injections,
`Pharm. Res., 6, 167–179 (1989).
`11. S. A. Al-Suwayeh, I. R. Tebbett, D. Wielbo, and G. A.
`Brazeau, In vitro-in vivo myotoxicity of intramuscular li-
`posomal formulations, Pharm. Res., 13(9), 1384–1388
`(1996).
`12. P. V. Rao, Statistical Research Methods in the Life Sci-
`ences, Duxbury, Pacific Grove, CA, 1998.
`13. K. H. Bauer, K. H. Froemming, and C. Fuehrer, Pharma-
`zeutische, 5 Aufl, 248 (1997).
`14. D. M. Needham, Machina Carnis; the Biochemistry of
`Muscle Contraction in Its Historical Development, Cam-
`bridge University Press, Cambridge, 1971.
`15. H. Rilbe, pH and Buffer Theory—A New Approach, Wi-
`ley, New York, 1996.
`I. Arif, I. A. Newman, and N. Keenlyside, Proton flux
`measurements from tissues in buffered solution, Plant
`Cell Environ., 18, 1319–1324 (1995).
`17. T. G. Nielsen, A. McArdle, J. Phoenix, L. A. Nielsen,
`T. S. Jensen, M. J. Jackson, and R. H. T. Edwards, In
`vivo model of muscle pain: quantification of intramuscu-
`lar chemical, electrical, and pressure changes associated
`with saline-induced muscle pain in humans, Pain, 69,
`137–143 (1997).
`
`16.
`
` PFIZER, INC. v. NOVO NORDISK A/S - IPR2020-01252, Ex. 1054, p. 8 of 8
`
`