Journal of Pharmaceutical Sciences xxx (2017) 1-7

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences

journal homepage: www.jpharmsci.org

Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Overcoming the Challenges of Low Drug Solubility in the Intravenous

Formulation of Solithromycin
Daniel Evans 1, *, Samuel Yalkowsky 2, Sara Wu 3, David Pereira 3,
Prabha Fernandes 4
1
Pharmacy Practice & Science, University of Arizona College of Pharmacy, Tucson, Arizona 85721
2
Department of Pharmaceutical Sciences, University of Arizona, Tucson, Arizona 85721
3
Drug Product Development, Cempra Pharmaceuticals Inc., Chapel Hill, North Carolina 27517
4
Administration, Cempra Pharmaceuticals Inc., Chapel Hill, North Carolina 27517

a r t i c l e i n f o a b s t r a c t

Article history: Solithromycin is a fluoro-ketolide (a fourth-generation macrolide) antibiotic that has been undergoing
Received 2 September 2017 clinical trials for the treatment of community-acquired bacterial pneumonia. In this study, development
Revised 16 October 2017 of the tri-amino acidebuffered solithromycin intravenous (IV) formulation was performed to minimize
Accepted 17 October 2017
the occurrence of infusion-associated local adverse events (infusion-site pain or phlebitis) observed in
patients who received the tartaric acidebuffered IV formulation with a lower buffered capacity during
phase I clinical trials. Development of the tri-amino acidsebuffered solithromycin IV formulation was
Keywords:
formulation achieved using a dynamic in vitro precipitation model. Computational modeling also supports the
precipitation superiority of the amino acid-buffered formulation over the tartaric aidebuffered formulation.
in vitro models © 2017 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.
acid-base equilibria
injectables

Introduction Background

Solithromycin is a fourth-generation macrolide antibiotic that
completed 2 phase 3 clinical trials (oral and IV to oral switch) for
the treatment of community-acquired bacterial pneumonia
(CABP).1 Solithromycin, the first fluoro-ketolide used in human
clinical trials, shows higher efficacy and has a broader spectrum
than other macrolide antibiotics such as azithromycin and
erythromycin.2
In addition to its oral formulation, solithromycin was formulated
for IV infusion delivery to treat patients in an acute care setting.
During an early phase I study, several patients receiving the IV
formulation experienced infusion-siteerelated pain and phlebitis.
This article describes the development of a solithromycin formu-
lation that was designed to minimize the potential for drug pre-
cipitation on infusion which can cause infusion-siteerelated
injuries.
* Correspondence to: Daniel Evans (Telephone: þ1-5206261289; Fax: þ1-
5206260703).
E-mail address: evans@pharmacy.arizona.edu (D. Evans).
The structure of solithromycin is shown in Figure 1 below. Sol-
ithromycin, which has a low aqueous solubility at pH 7.4, is a dibasic
compound with ionization constants of pKa1 ¼ 3.94, pKa2 ¼ 9.44.
Having the 2 basic ionizable functional groups allows the use of pH
control to increase the aqueous solubility of the drug to clinically
relevant levels.
The total solubility (ST) of a dibasic drug as a function of solution
pH can be estimated using the exponential form of the Henderson-
Hasselbalch equation.3


ST ¼ Su 1 þ 10pKa2 pH þ 10pKa1 þpKa2 2pH (1)
where Su is the aqueous solubility of the unionized form of the
drug. Differences between Henderson-Hasselbalchederived solu-
bility data and data obtained experimentally are often due to ag-
gregation of dug molecules4 as well as the formation of insoluble
salts of the drug and its counter-ion.5
During intravenous administration, the infusion solution of the
formulation is injected into the blood stream which is naturally
buffered at pH 7.4. Thus, on mixing with blood, a formulation

https://doi.org/10.1016/j.xphs.2017.10.030
0022-3549/© 2017 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.
2 D. Evans et al. / Journal of Pharmaceutical Sciences xxx (2017) 1-7
Figure 1. Structure of solithromycin.
containing a basic drug solubilized in a low pH solution will
experience the processes of dilution and titration simultaneously.
The effect of dilution on the concentration of a drug (Cdrug) is linear
and can be determined by the following formula6:
CdrugðdilutedÞ ¼ Cdrug  f form ¼ Cdrug  ð1  f blood Þ
where fform is the volume fraction of formulation and fblood is the
volume fraction of blood. A simple correlation between fform and
fblood during IV administration can be done by considering match-
ing the injection rate to a theoretical blood flow rate (5 mL/min) in
the forearm vein. Assuming complete mixing of the blood and drug
formulation, the relationship of fform and fblood can be expressed in
Equation 3.
f form rinj
z (3)
f blood rbloodflow
where rinj is the injection rate of the formulation and rblood flow is
the blood flow rate.
While the concentration of the basic drug is decreasing line-
arly due to process of dilution, the buffered blood will also titrate
the pH of the formulation back to 7.4. According to Equation 1, a
rapid increase in pH will cause the solubility of the basic drug
(using a low pH to increase solubility) to decrease exponentially.
If the simultaneous processes of dilution and titration exerted by
Figure 2. Illustration of the dynamic in vitro device used by Evans et al (a) and Johnson et al. (b).
the blood on the formulation causes the solubility of the drug to
fall below the drug concentration during mixing, a supersatu-
rated solution is formed. This may ultimately lead to drug
precipitation.
Drug precipitation that occurs when IV formulations are mixed
with blood can cause infusion-site phlebitis.7 In initial IV formu-
lation development studies, the solubility of solithromycin was
determined in many aqueous solutions which contained organic
co-solvents, surfactants, cyclodextrins, and buffers. Solithromycin
showed good solubility in low concentrations of tartaric acid so-
lutions in lower pH range. Therefore, a lyophilized formulation
(250 mg/vial) with tartaric acid and mannitol at pH 3.8 was
developed for phase I clinical studies. Early in the phase 1 IV
formulation development program, a lyophilized solithromycin
tartaric acid-mannitol formulation, diluted in 0.45% sodium
chloride to a concentration of 1.6 mg/mL and infused as single
doses of 25 to 1000 mg and as multiple doses of 50 to 200 mg was
evaluated. At higher single doses and multiple doses of 200 mg,
(2) several subjects experienced infusion-siteerelated pain and
phlebitis. Precipitation of solithromycin due to the change on
infusion from a low pH range (4 to 6) to a physiologic pH of
approximately 7.4 was considered a potential contributor to the
infusion-site reactions. Following these initial studies, extensive
experiments were conducted to design the optimal infusion so-
lution for solithromycin.
Materials and Methods
Materials
Solithromycin-free base and lyophilized product (tartaric acid-
mannitol formulation) was provided by Cempra Pharmaceuticals,
Inc. In addition, all buffers used in the formulations of this study
were purchased from Ajinomoto and Sigma-Aldrich by Cempra and
were of United States Pharmacopeia or National Formulary grade.
All formulations tested were prepared according to instructions
provided by Cempra and filtered using a 0.45-mm Polytetrafluoro-
ethylene Acrodisc® syringe filter (Pall Life Sciences) before testing.
Isotonic Sorensen's phosphate buffer (ISPB pH ¼ 7.4), was prepared
 D. Evans et al. / Journal of Pharmaceutical Sciences xxx (2017) 1-7 3

Figure 3. Results of the dynamic in vitro tests for diluted solithromycin solution with tartaric acid at 2 (left) and 3 mg/mL (right).

using Documenta Geigy Scientific Tables8 and passed through a
0.45-mM Nalgene® nylon filter system before use.
Static Dilutions
All formulations were subjected to a static serial dilution test
described by Dai.9 Ten borosilicate test tubes were filled with 1 mL
of ISPB. Prepared formulation of 1 mL was dispensed into the first
test tube rapidly to facilitate mixing. The resultant solution of 1 mL
was then transferred to the second test tube in the same manner
and so forth until all 10 of the 1:1 serial dilutions were complete. A
visual inspection for drug precipitation was performed with the aid
of a laser pointer positioned below the test tube to produce a
Tyndall beam.
Dynamic In Vitro Testing
Formulations of solithromycin were tested using the validated
dynamic in vitro device used by Johnson et al.10 An illustration of
the device is shown in Figure 2, panel B. Using an intravenous Y-Site
port, IV formulation candidates were injected into ISPB flowing
through Tygon® tubing at 5 mL/min into a 0.1-cm flow cell.
Detection of opacity due to drug precipitation was done at 500 nm
(a wavelength at which the drug and the excipients do not absorb)
using a Cary 50 UV-Vis Spectrophotometer and WinUV® software
in Kinetics mode.
A newly described in vitro device,11 illustrated in Figure 2, panel
A, was also used to compare formulations for their potential to
precipitate. This model uses a charged coupled device camera
positioned near a quartz tube flow cell to detect scattered light
caused by drug precipitation at the site of injection.
Simulated Dilutions
The titration effect of blood can be modeled using equations
established by Surakitbanharn et al.6 and extended to other
buffer systems by incorporating the work of de Levie12 where the

Figure 4. Simulated dilution of the original solithromycin formulation at 2 mg/mL.

4 D. Evans et al. / Journal of Pharmaceutical Sciences xxx (2017) 1-7

Figure 5. Comparison of 2 pH-solubility plots of solithromycin using phthalate buffer (left) and HCl (right).

volume fraction of formulation (4F) is calculated as a function strength of the formulation and blood surrogate were calculated
of pH. using additional equations derived by de Levie.12

P     Results and Discussion

VF  F C  Hþ þ OH
4F ¼ ¼ PB B P (4)
VF þ VB FF CF  FB CB
The first generation of lyophilized solithromycin IV formulation
contained solithromycin (50 mg/mL), tartaric acid (5.8 mg/mL, 38.5
where CB and CF are the molar concentrations of the buffers in the
mM), and mannitol as a bulking agent at pH 3.8 after reconstitution
blood surrogate and formulation respectively; and [Hþ] and [OH]
in 5 mL of sterile water. The reconstituted solution was diluted in
are the concentrations of the hydrogen and hydroxide ions
half-normal saline (0.45%) to achieve desired drug concentrations.
respectively. ISPB is used as the blood surrogate (buffer) because it
Diluted solutions containing 1, 2, and 3 mg/mL of solithromycin in
has similar buffering capacity at pH ¼ 7.4 as blood, whereas blood
tartrate buffer (in half-normal saline) were each tested at infusion
or plasma are buffered by carbonate/biocarbonate which can be lost
rates of 1, 2, and 3 mL/min for 1 min using the dynamic in vitro
as carbon dioxide outside of the body. FB and FF are buffer functions,
model of Johnson and Yalkowsky10 which introduces IV drug for-
for the blood and formulation respectively, calculated using the
mulations into a continuous stream of ISPB. Figure 3 shows the
hydrogen ion concentration [Hþ], dissociation constants Ki, along
opacity scans of the diluted formulations at 2 and 3 mg/mL at each
with the maximum (p), and actual number (q) of protons for a weak
of the 3 infusion rates as they passed through the flow cell. It ap-
acid or base as
pears from these tests that solithromycin precipitated on mixing
with ISPB buffer and that increasing both the infusion rate and drug
Pp  pj Qj concentration increased the amount of precipitation. These results
j¼0
ðp  q  jÞ Hþ K
i¼0 i offer a likely explanation for the infusion-siteerelated pain and
FACID ¼ FBASE ¼ Pp  þ pj Qj (5)
H K phlebitis in the first IV clinical study.
j¼0 i¼0 i
To understand why this happened, a simulated dilution of the
The drug solubility and volume fraction of formulation are 2 mg/mL formulation of solithromycin was performed. Figure 4 is a
calculated using Equations 3 and 4 for 500 pH values equally spaced plot showing the effect of dilution on the pH of the mixture (left) as
in between the formulation pH and 7.4. For both in vitro tests and well as drug concentration and solubility (right) as calculated by
dilution calculations, ISPB is used as the blood surrogate. In turn, Equations 1-4. Because the diluted formulation of solithromycin
the concentration of drug is calculated for each volume fraction of contains very little buffer (0.77 to 2.3 mM), the ISPB blood surrogate
formulation. The concentration and solubility of drug versus the quickly titrates the mixture pH back to pH 7.4. As shown on the
volume fraction of formulation are plotted on the same axis to right side of Figure 4, this rapid increase in pH causes the theo-
determine areas of supersaturation. Buffer capacity and ionic retical drug solubility (solid line) to fall below the concentration of

Table 1
Results of the Static Dilution Tests of Solithromycin Formulations at 2 Mg/mL

Buffer System pH Dilution

1/2 1/4 1/8 1/16 1/32 1/64 1/128 1/256 1/512 1/1024

5-mM tartaric acid 4.2 1 1 1 0 1 1 1 0 0 0

10-mM tartaric acid 4.2 1 1 1 0 0 0 0 0 0 0
20-mM tartaric acid 4.2 1 1 1 1 0 0 0 0 0 0
38.5-mM tartaric acid 4.2 1 1 1 1 0 0 0 0 0 0
5-mM glutamic acid 4.2 1 1 0 0 0 0 0 0 0 0
10-mM glutamic acid 4.2 1 1 0 0 0 0 0 0 0 0
25-mM glutamic acid 4.2 1 1 0 0 0 0 0 0 0 0
15-mM aspartic acid, 15-mM glutamic acid, 4.5 0 0 0 0 0 0 0 0 0 0
and 15-mM histidine
 D. Evans et al. / Journal of Pharmaceutical Sciences xxx (2017) 1-7
Figure 6. Buffer capacity of the 2 formulations of solithromycin (left) and the contributions of the individual buffer species in the tri-amino acid formulation to the total buffer
capacity (right).
drug (dotted line). This produces a transient supersaturated solu-
tion (gray-shaded region bottom) for most of the dilution ratios.
To reduce the amount of supersaturation, a higher concentration
of buffer is required to provide more buffer capacity between the
formulation pH and physiologic pH of 7.4. The simplest approach
would be to increase the concentration of the tartaric acid. Using a
higher concentration of buffer is not without risk, however, as the
solubility of an ionizable drug can be dependent on the species and
concentration of the counter-ion. An example of species-dependent
interactions between an ionizable drug and solution counter-ions is
shown in the initial pH-solubility study for solithromycin (Fig. 5).
The plot on the left of Figure 5 indicates that lowering the pH of the
solithromycin solution with phthalate buffer (left) causes a
decrease in drug solubility. In comparison, using HCl to lower the
solution pH increases the solubility in a manner consistent with the
Henderson-Hasselbalch equation.
To determine if increasing the concentration of tartaric acid
would contribute enough buffer capacity to the formulation
without having a deleterious effect on the drug solubility, soli-
thromycin was formulated along with 5, 10, 20, 38, and 67-mM
tartaric acid at pH 4.2. In addition, formulations containing other
buffers such as lactic acid, citric acid, acetic acid, histidine, glutamic
acid, and aspartic acid, were made in case tartaric acid was deemed
unsuitable for buffering purposes. The static dilution method was
used as a quick-and-dirty method for determining the performance
of these buffered formulations. Table 1 shows the static dilution
results for some of the buffered formulations tested, where 1
Figure 7. Simulated dilution of both the tartaric acid and the tri-amino acid formulations of solithromycin.
5
indicates the presence of drug precipitation and 0 indicates no
precipitation.
Before static dilution testing, formulations containing 67-mM
tartaric acid and 50-mM citric acid showed visible signs of pre-
cipitation with the latter producing large particles. It was therefore
concluded that formulations of solithromycin and certain poly-
carboxylic buffers were incompatible at concentrations necessary
to provide highly buffered formulations.
In static tests, precipitation is usually observed as the formula-
tion becomes more dilute with ISPB. Interestingly, many formula-
tions of solithromycin precipitated on the first dilution as shown in
Table 1. For tartaric acid, as the buffer concentration increased,
precipitation was present up to the fourth dilution. For the lowest
concentration of tartaric acid (5 mM), the precipitation of drug at
higher dilution ratios (ISPB:formulation) were also observed. A
possible explanation for immediate precipitation is a conversion of
the salt to the free form as the formulation is mixed with a higher
pH solution (ISPB). This phenomenon, described by Anderson and
Flora,13 occurs because the equilibrium favors conversion of the salt
to the free base form as the pH increases.
Two strategies were employed to buffer the formulation to
prevent avoid drug-counter-ion solubility problems and rapid
precipitation of the free base. The first strategy was to use the
positive charge on an amino acid to reduce the interaction between
solithromycin and the counter-ion. Table 1 shows that the amount
of precipitation in the glutamic acid formulation was less than that
of the tartaric acid formulations. Moreover, increasing the amount
6 D. Evans et al. / Journal of Pharmaceutical Sciences xxx (2017) 1-7
Figure 8. Scaled intensity images taken near the needle tip (left) and downstream (right) during the injection of the tartaric acid (top) and final tri-amino acid (bottom)
formulations of solithromycin.
of glutamic acid to increase the buffer capacity did not produce an
increase in the immediate precipitation. The second formulation
strategy was to use smaller amounts of multiple buffers to achieve a
high buffer capacity without running into solubility issues such as
those observed with tartaric acid. This combined strategy led to a
novel formulation containing 3 amino acids with a different pKa
values (of the side chains), glutamic acid, histidine, and aspartic
acid at 15 mM of each amino acids at a pH of 4.5.
The left plot of Figure 6 compares the buffer capacities of the
solithromycin tartaric formulation with the solithromycin formu-
lation containing the 3 amino acids. The solithromycin tri-amino
acid formulation maintains a higher buffer capacity than the soli-
thromycin tartrate formulation as the dilution of the formulation
shifts the pH from 4.5 to 7.4. The rise in buffer capacity as both
solutions approach 7.4 is due to the buffering capacity of the
isotonic phosphate buffer used as the blood surrogate. The right
plot of Figure 6 shows the contribution of the individual buffer
species to the total buffer capacity of the final formulation of soli-
thromycin. Both the aspartic acid and glutamic acid contribute to
the buffering capacity at the starting pH of 4.5, whereas histidine
provides more buffering capacity as the pH rises to 7.4 during the
dilution on intravenous administration.
The graphs in Figure 7 compare the theoretical effect of the tri-
amino acid formulation with the diluted (1.5 mM) tartaric acid
formulation on the pH and on the solubility of solithromycin during
dilution. Each formulation contained 2 mg/mL of solithromycin.
The change in pH for the tri-amino acid formulation (dashed line) is
much more gradual than the diluted tartrate-buffered formulation
of solithromycin (solid line). This results in the tri-amino acid
formulation maintaining a higher solubility at all fractional di-
lutions and a much lower area of supersaturation. By reducing the
area of supersaturation using a tri-amino acid combination of
buffers, there is a decrease in the probability of drug precipitation
during infusion.
To confirm the results of the static dilution tests, both formu-
lations were tested using the new dynamic in vitro visualization
model of Evans et al.11 Both the tartaric acid and tri-amino acide
buffered formulations containing 2 mg/mL solithromycin were
injected at 4 mL/min for 1 min while images were taken at the
needle tip as well as 12 cm downstream. Results of this dynamic
visualization test are shown in Figure 8 as scaled color intensity
images taken near the needle tip (left) and 12 cm downstream
(right) during the injection of original diluted tartaric acid (top) and
final tri-amino acid (bottom) formulations of solithromycin. The
intensity of precipitation in these images is a function of the color
where blue < white < red.
Precipitation at the needle tip produced during the injection of
the diluted tartaric acid formulation is shown as the light blue-
white haze on the far-right side of Figure 8 (top left). This precip-
itation, which happens almost immediately on injection into the
flowing ISPB, increases as the mixing of the formulation and ISPB
becomes more complete downstream (top right of Fig. 8). In
contrast, the 2 images on the bottom of Figure 8 show that the
formulation containing the tri-amino acid buffer did not produce
any apparent precipitation on injection into the ISPB at either the
needle tip or downstream.
Conclusion
Selecting buffers for an IV formulation requires balancing buffer
capacity with drug-counter-ion solubility especially for compounds
with low solubility at physiological pH. The successful formulation
of solithromycin containing the 3 amino acids that was used in the
phase 3 IV-oral switch clinical trials demonstrates that using a
combination of buffers can eliminate problems that might occur
with a single buffer. Modeling of precipitation following dilution
with blood using static and dynamic in vitro methods led to the
development of an optimal formulation of solithromycin. Since
these tests along with simulated dilutions can be done during the
early drug development stages, they can save time and financial
resources associated with pre-clinical animal studies.
References
1. Van Bambeke F, Tulkens PM. The role of solithromycin in the management of
bacterial community-acquired pneumonia. Expert Rev Anti Infect Ther. 2016;14:
311-324.
2. Llano-Sotelo B, Dunkle J, Klepacki D, et al. Binding and action of CEM-101, a
new fluoroketolide antibiotic that inhibits protein synthesis. Antimicrob Agents
Chemother. 2011;54:4961-4970.
3. Yalkowsky SH. Solubility and Solubilization in Aqueous Media. New York:
American Chemical Society and Oxford Univeristy Press; 1999.
€m CAS, Luthman K, Artursson P. Accuracy of calculated pH-dependent
4. Bergstro
aqueous drug solubility. Eur J Pharm Sci. 2004;22(5):387-398.
5. Stahl PHS, Wermuth CG. Handbook of Pharmaceutical Salts Properties, Selection,
and Use. 2nd ed. Zurich, Switzerland: Verlag Helvetica; 2002.
6. Surakitbanharn Y, Simamora P, Ward GH, Samuel H, Yalkowsky S. Precipitation
of pH solubilized phenytoin. Int J Pharm. 1994;109:27-33.
 D. Evans et al. / Journal of Pharmaceutical Sciences xxx (2017) 1-7 7

7. Yalkowsky SH, Krzyzaniak JF, Ward GH. Formulation-related problems asso-
ciated with intravenous drug delivery. J Pharm Sci. 1998;87(7):787-796.
8. Geigy D. Scientific Tables. 7th ed. Ardsley: J.R. Geigy S.A.; 1970.
9. Dai W-G. In vitro methods to assess drug precipitation. Int J Pharm. 2010;393:
1-16.
10. Johnson JL, Yalkowsky SH. Prediction of precipitation-induced phlebitis: a
statistical validation of an in-vitro model. J Pharm Sci. 2003;92:1574-1581.
11. Evans DC, Kudenov MW, Sassenrath KC, Dereniak EL, Yalkowsky SH. Imaging of
in vitro parenteral drug precipitation. Int J Pharm. 2016;512:219-223.
12. de Levie R. A general simulator for acid-base titrations. J Chem Educ.
1999;76:987.
13. Anderson BD, Flora KI. Preparation of water soluble compounds through salt
formation. In: The practices of medicinal chemistry. London: Academic Press;
1996.