new secondary metabolites of phenylbutyrate in humans

0090-9556/04/3201-10–19$20.00
DRUG METABOLISM AND DISPOSITION Vol. 32, No. 1
Copyright © 2004 by The American Society for Pharmacology and Experimental Therapeutics 1181/1111187
DMD 32:10–19, 2004 Printed in U.S.A.
NEW SECONDARY METABOLITES OF PHENYLBUTYRATE IN HUMANS AND RATS
Takhar Kasumov, Laura L. Brunengraber, Blandine Comte, Michelle A. Puchowicz,
Kathryn Jobbins, Katherine Thomas, France David, Renee Kinman, Suzanne Wehrli,
William Dahms, Douglas Kerr, Itzhak Nissim, AND Henri Brunengraber
Departments of Nutrition (T.K., L.L.B., B.C., M.A.P., K.J., K.T., F.D., R.K., H.B.) and Pediatrics (R.K., W.D., D.K.), Case Western
Reserve University, Cleveland, Ohio; and Children’s Hospital (S.W.) and Department of Pediatrics (I.N.), University of
Pennsylvania, Philadelphia, Pennsylvania
ABSTRACT:
Phenylbutyrate is used to treat inborn errors of ureagenesis, malignancies,
cystic fibrosis, and thalassemia. High-dose phenylbutyrate
therapy results in toxicity, the mechanism of which is unexplained.
The known metabolites of phenylbutyrate are phenylacetate,
phenylacetylglutamine, and phenylbutyrylglutamine. These are excreted
in urine, accounting for a variable fraction of the dose. We
identified new metabolites of phenylbutyrate in urine of normal
humans and in perfused rat livers. These metabolites result from
interference between the metabolism of phenylbutyrate and that
of carbohydrates and lipids. The new metabolites fall into two
Sodium phenylbutyrate (PB 1 ) is a highly effective drug for the
treatment of patients with hyperammonemia resulting from inborn
errors of urea synthesis (Batshaw et al., 1981, 2001; Brusilow, 1991).
These patients excrete nitrogen as phenylacetylglutamine (PAGN)
(Batshaw et al., 1981). The latter is also formed when the patients are
treated with phenylacetate (PA). However, PB is preferred as a
prodrug of PA because it does not have the foul smell of the latter. In
addition, PB shows promise for the treatment of cystic fibrosis because
it increases trans-membrane chloride conductance (Rubenstein
and Zeitlin, 2000; Zeitlin et al., 2002). Also, PB is used in clinical
trials for the treatment of sickle-cell anemia and thalassemia because
it induces the formation of fetal hemoglobin (Dover et al., 1994;
Hoppe et al., 1999). Lastly, PB is used in clinical trials as a cytostatic
antineoplastic agent, because it inhibits histone deacetylases and potentiates
the effect of cytotoxic agents on tumors (Samid et al., 1997;
Gilbert et al., 2001).
This work was supported by the National Institutes of Health (Research Grants
DK58126, CA79495, and DK53761, and Training Grant DK07319) and the Cleveland
Mt. Sinai Health Care Foundation.
1 Abbreviations used are: PB, 4-phenylbutyrate; PAGN, phenylacetylglutamine;
PA, phenylacetate; PBGN, 4-phenylbutyrylglutamine; PHB, 3-hydroxy-4phenylbutyrate;
PtC, 4-phenyl-trans-crotonate; PKB, 4-phenyl-3-ketobutyrate;
TMS, trimethylsilyl; GC-MS, gas chromatography-mass spectrometry; SW,
sweep width; TD, data points; COSY, correlation spectroscopy; HSQC, heteronuclear
single-quantum coherence; AUC, area under the curve.
Address correspondence to: Henri Brunengraber, Department of Nutrition,
Room 280, Case Western Reserve University, 11000 Cedar Rd., Cleveland OH
44106-7139. E-mail: hxb8@cwru.edu
(Received May 14, 2003; accepted September 2, 2003)
This article is available online at http://dmd.aspetjournals.org
10
categories, glucuronides and phenylbutyrate �-oxidation side
products. Two questions are raised by these data. First, is the
nitrogen-excreting potential of phenylbutyrate diminished by ingestion
of carbohydrates or lipids? Second, does competition
between the metabolism of phenylbutyrate, carbohydrates, and
lipids alter the profile of phenylbutyrate metabolites? Finally, we
synthesized glycerol esters of phenylbutyrate. These are partially
bioavailable in rats and could be used to administer large
doses of phenylbutyrate in a sodium-free, noncaustic form.
The clinical effectiveness of PB in some of these situations is
limited by occasional incidences of toxicity at high doses (Carducci et
al., 2001; Gore et al., 2002). Concerns have been raised by clinical
investigators who treat patients with large doses of PB as a sodium
salt. First, the total amount of PB and its known metabolites excreted
in urine (PA, PAGN) is less than the administered PB dose, sometimes
as low as 50%. Some of the unknown metabolites might
contribute to PB toxicity at high doses. Second, the large sodium load
of the treatment is potentially dangerous for patients with impaired
cardiac and/or renal function. Third, the causticity of sodium PB can
result in esophagal and/or gastric distress, even when it is administered
as a powder suspended in water (this has been extensively
debated on the Internet discussion group Metab-L at http://lists.franken.de/mailman/listinfo/metab-L).
Neither the biochemical mechanism(s)
of PB toxicity nor the identity of the missing metabolites of
PB is known. Also, it is not clear whether the metabolism of PB (a
modified fatty acid) interferes with, or is influenced by, the metabolism
of fatty acids and carbohydrates present in foodstuffs. Lastly, it
is not known whether stimulation of lipolysis under stress conditions
interferes with PB metabolism.
Our original interest in PB metabolism was related to it being a
precursor of PAGN, which can be used as a noninvasive probe of the
14 C- or 13 C-labeling pattern of citric acid cycle intermediates in
human liver (Magnusson et al., 1991; Yang et al., 1996). As part of
these investigations, we recently identified phenylbutyrylglutamine
(PBGN) as a new metabolite of PB (Comte et al., 2002). PBGN is
presumably formed from the reaction of PB-CoA with glutamine, by
analogy with the formation of PAGN from the reaction between
PA-CoA and glutamine (Webster et al., 1976). In normal adult sub-
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

jects who ingested a fairly low dose of PB (5 g/75 kg), we found that
the total excretion of PB � PA � PAGN � PBGN accounted for only
half the ingested PB dose (Comte et al., 2002). The missing fraction
of the dose may be disposed of either in urine as unknown metabolites,
or in feces as unabsorbed PB and/or PB metabolites.
In the present study, we report the identification of additional
metabolites of PB in humans, i.e., R- and S-3-hydroxy-4-phenylbutyrate
(PHB), phenylacetone, and 1-phenyl-2-propanol, as well as PA
and PB glucuronides. We studied the mechanism of PHB formation
from PB in perfused rat livers and identified two additional metabolites,
4-phenyl-trans-crotonate (PtC) and 4-phenyl-3-ketobutyrate
(PKB). Lastly, we prepared sodium-free esters of PB and investigated
their bioavailability in rats. Glycerol-PB esters appear promising for
the administration of large amounts of PB without the corresponding
sodium load.
Materials and Methods
Materials. All chemicals used in syntheses, general chemicals, and solvents
were obtained from Sigma-Aldrich (St. Louis, MO). All organic solvents were
dried and distilled immediately before use. [ 2 H7]Phenylacetic acid (99%) and
[ 2 H6]benzene were purchased from Isotec Inc. (Miamisburg, OH). The derivatization
agent N-methyl-N-(trimethylsilyl)trifluoroacetamide was supplied by Regis
Technologies, Inc. (Morton Grove, IL). All aqueous solutions were made with
water purified with the Milli-Q system (Millipore Corporation, Bedford, MA).
Preparation of Unlabeled and Deuterated Standards. S-2-Phenylbutyryl
chloride was prepared by reacting S-2-phenylbutyric acid with SOCl2, and was
vacuum-distilled and stored at 4°C as a 0.5 M solution in benzene. [ 2 H5]PB was prepared by aluminum chloride-catalyzed condensation of �-butyrolactone
with [ 2 H6]benzene as previously described (Comte et al., 2002). �-Phenyl-trans-crotonic
acid was prepared by a Fridel-Crafts reaction of benzene
with ethyl �-bromo-trans-crotonate, followed by acid hydrolysis of ethyl
�-phenyl-trans-crotonate (Loffler et al., 1970). The trans configuration of the
product was confirmed by 1 H NMR. [ 2 H7]Phenylacetylglycine was synthesized
by reacting [ 2 H7]phenylacetyl chloride with glycine, as described previ-
NEW PHENYLBUTYRATE METABOLITES
FIG. 1.Scheme for the synthesis of unlabeled and deuterated standards.
ously for the unlabeled analog (Ramsdell and Tanaka, 1977).
[ 2 H 7]Phenylacetyl chloride was prepared by activation of [ 2 H 7]phenylacetic
acid with freshly distilled dichloromethyl methyl ether and used immediately
after evaporation of excess dichloromethyl methyl ether and of the methyl
chloroformate byproduct. The synthetic protocol for preparation of PHB, PKB,
[ 2 H 5]PHB, phenylacetone, and 1-phenyl-2-[2- 2 H]propanol is outlined in Fig. 1.
Ethyl 4-Phenyl-3-ketobutyrate (2; Fig. 1) was synthesized by a method
adapted from Capozzi et al. (1993). The commercial isopropylidene malonate
(2,2-dimethyl-4,6-diketo-1,3-dioxane, also called Meldrum’s acid; Aldrich
Chemical Co., Milwaukee, WI) was reacted with phenylacetyl chloride in the
presence of dry pyridine in anhydrous methylene chloride. The crude phenylacetylated
Meldrum’s acid (compound 1, Fig. 1) was refluxed in absolute
ethanol until evolution of CO 2 ceased (about 3 h). After evaporation of the
solvent, ethyl 4-phenyl-3-ketobutyrate was purified on a silica gel column.
4-Phenyl-3-ketobutyric acid (PKB). A 10% molar excess of 1 N NaOH
solution was added slowly to ice-cooled ethyl 4-phenyl-3-ketobutyrate and
stirred at room temperature until the organic phase disappeared (approximately
12 h). The solution was acidified with 1 N HCl to pH 2 and extracted three
times with 3 volumes of ethyl ether. After drying over Na 2SO 4, the solvent was
evaporated to give the white solid product (yield 94%, m.p. 70°C). 1 H NMR
(300 MHz, �, CDCl 3): keto, 3.33 (s, 2H, CH 2COO), 3.78 (s, 2H, CH 2Ph),
7.17–7.33 (m, 5H, Ph); enol, 4.82 (s, 1H, CH), 12.13 (s, 1H, OH); keto/enol �
7.8:1. 13 C NMR (75 MHz, �, CDCl 3): 48.36 (CH 2COO), 49.87 (CH 2Ph),
127.41 (C-4 Ph), 128.73 (C-3, C-5 Ph), 129.50 (C-2, C-6 Ph), 133.14 (C-ipso,
Ph), 169.26 (COO), 201.57 (CO).
The identity of the product was also confirmed by 1) reduction of 4-phenyl-
3-ketobutyrate with NaBH 4, 2) extraction of R,S-PHB with ethyl acetate, and
3) TMS derivatization and NH 3-positive chemical ionization GC-MS. The
mass spectrum of the derivative was identical to that of PHB synthesized as
described below. The 4-phenyl-3-ketobutyric acid was stored at �80°C to
prevent decomposition. Just before use in liver perfusion experiments, the acid
was dissolved in water and titrated to pH � 7.40.
R,S-3-Hydroxy-4-phenylbutyric acid (PHB). Ethyl 4-phenyl-3-ketobutyrate
(2.06 g, 10 mmol) was mixed with 10 ml of H 2O and a calculated amount of
NaOH/H 2O was added dropwise with cooling to give 0.5M solution of [OH - ].
11
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

12 KASUMOV ET AL.
NaBH 4 (0.37g, 10 mmol) was added and the mixture stirred for 24 h. After the
reaction mixture was cooled to 0°C, the pH was brought to 7.0 with HCl (6N)
and more than half of water removed by lyophilizer. The pH was brought to 2
by HCl (6 N) and the solution extracted 3 times with 3 volumes of ethyl ether.
After drying the combined ether extract, the ether was evaporated giving a
white solid product. Yield 79%. Purity of product was assayed by GC-MS after
derivatization with TMS.
R-3-Hydroxy-4-phenylbutyrate (R-PHB) was prepared by reducing the corresponding
4-phenyl-3-ketobutyric acid with B-chlorodiisopinocampheylbo
rane [(�)-DIP-Cl] (Wang et al., 1999). The yield was 89% from PKB;
enantiomeric excess was 97% [GC-MS of the methyl S-2-phenylbutyryl derivative
(see below), and purity was confirmed by NMR].
R,S-3-Hydroxy-4-phenyl[2,2,3,4,4- 2 H 5]butyrate (R,S-[ 2 H 5]PHB). 4-Phenyl-3-ketobutyric
acid (0.267 g, 1.5 mmol) was suspended in 3 ml of 2 H 2O
(99.9%) to which 0.36 ml of 40% of NaO 2 H (3.6 mmol) in 2 H 2O was slowly
added at 0–5°C. This procedure exchanges 1 H for 2 H atoms on the methylene
groups adjacent to the carbonyl. The solution was stirred at room temperature
overnight and then lyophilized. The residue was dissolved in 3 ml of 2 H 2O and
stirred for another 5hatroom temperature. The solution was cooled on ice,
treated with NaB 2 H 4 (63 mg, 1.5 mmol), and stirred overnight at room
temperature (Des Rosiers et al., 1988). After acidification to pH 1 to 2 with
HCl (6 N), the solution was saturated with NaCl and extracted three times with
3 volumes of diethyl ether. Solvent evaporation yielded R,S-[ 2 H 5]PHB (yield
74%, purity 99% by NMR, M5 isotopic enrichment 95% by GC-MS of the
TMS derivative). The free acid was titrated to pH 8 with NaOH and stored
frozen as a 0.5 M solution until use.
Phenylacetone was prepared by decarboxylating phenylketobutyrate in acid
at 100°C. It was reduced to 1-phenyl-2-[2- 2 H]propanol with NaB 2 H 4.
Esters of Phenylbutyrate. Dihydroxyacetone-di-PB, glycerol-tri-PB, ribose-tetra-PB,
glucose-penta-PB, and sorbitol-hexa-PB were prepared by reacting
the polyol/sugar with excess phenylbutyryl chloride in the presence of
pyridine and catalytic amounts of N,N-dimethylaminopyridine. Products were
purified by flash column chromatography on silica. To prepare glycerol-mono-
PB, isopropylidene glycerol was reacted with phenylbutyryl chloride as above,
and the isopropylidene group was removed by mild acidic hydrolysis in water.
The structure and purity of all products were confirmed by 1 H and 13 C NMR.
The structure of glycerol-mono-PB was confirmed by acetylation with acetic
anhydride, followed by GC-MS analysis of the derivative.
Sample Preparation. For the determination of the free acids concentration
(PA, PHB, PKB, PtC, and PB) in perfusate, samples (0.1 ml) were spiked with
0.17 �mol of [ 2 H 7]PA, [ 2 H 5]PB, and R,S-[ 2 H 5]PHB before deproteinization
with 20 �l of saturated sulfosalicylic acid. The slurries were saturated with
NaCl, acidified with one drop of 6 M HCl, and extracted three times with 5 ml
of diethyl ether. For the assay of conjugates of PB and PA, 0.1-ml aliquots of
final liver perfusate or of human urine were spiked with internal standards and
treated with 1.0 ml HCl (6 N) at 90°C overnight to hydrolyze the conjugates.
Glucuronides of PB and PA were identified by the amount of these compounds
released after incubation of perfusate and urine samples with �-glucuronidase
in 0.2 M ammonium acetate buffer, pH 5.0, overnight at 37°C.
Bile samples were analyzed for free and total (conjugated � free) PB and
its metabolites. In the first series of assays, 0.05-ml samples of bile were spiked
with internal standards, acidified to pH 2 to 2.5, and extracted three times with
diethyl ether. In the second series of assays, samples spiked with internal
standards were hydrolyzed with 0.3 ml of NaOH (6 N) at 90°C for 3 h before
acidification and extraction.
Phenylacetylglycine was extracted in acid and derivatized with methanol/
HCl. For the assay of phenylacetone and 1-phenyl-2-propanol, urine and
perfusate samples were spiked with the structural analog 1-phenyl-
[ 2 H 5]ethanol and then treated with NaB 2 H 4 to reduce phenylacetone to monodeuterated
1-pheny-2-propanol. The labeled and unlabeled 1-phenyl-2propanol
were assayed as TMS derivatives.
For the assays in urine, 0.1-ml samples were spiked with 0.15 �mol
[ 2 H 5]PHB, acidified to pH 1 to 2 with HCl, saturated with NaCl, and extracted
three times with 3 ml of diethyl ether. The combined extracts were dried with
Na 2SO 4 and evaporated before reacting the residues with 70 �l of TMS at
60°C for 20 min.
For the chiral assay of PHB enantiomers (Powers et al., 1994), 0.1-ml
samples were spiked with R,S-[ 2 H 5]PHB, and either deproteinized with 50 �l
of saturated sulfosalicylic acid (if containing proteins) or acidified to pH 1 to
2 with HCl (for urine). Then, the slurries or solutions were saturated with NaCl
and extracted three times with 3 ml of diethyl ether. The combined extracts
were dried with Na 2SO 4 and evaporated before reacting the residues with 0.15
ml of methanol/HCl for 1hat65°C, to derivatize the carboxyl groups of the
PHB enantiomers. After cooling, 1 ml of water was added to the mixture and
the hydroxyacid methyl ester was extracted with diethyl ether (three times in
3 ml). After complete evaporation of the combined ether extract, S-2phenylbutyryl
chloride benzene solution (0.1 ml, 0.5 M) and 0.05 ml of
aqueous 12 N NaOH were added. After vortexing, the mixture was incubated
for1honaslow shaker at room temperature. The derivatives were extracted
with ether (three times in 3 ml) and 1 ml of water. The combined ether phase
was dried with Na 2SO 4 and evaporated completely. The residue was dissolved
in 0.1 ml of ethyl acetate, and 1 �l was injected into the GC-MS.
GC-MS Methods. All of the metabolites, except phenylacetylglycine, were
analyzed as their TMS derivatives on a Hewlett-Packard 5890 gas chromatograph
equipped with a ZB-5 capillary column (60 m � 0.25 mm i.d., 0.5 mm
film thickness; Hewlett Packard, Palo Alto, CA) and coupled to a 5989A mass
selective detector. Samples (0.2–1 �l) were injected with a split ratio 20 to
50:1. The carrier gas was helium (1 ml/min) and nominal initial pressure was
20.61 psi. The injector port temperature was at 270°C, the transfer line at
305°C, the source temperature at 200°C and quadrupole at 150°C. The column
temperature program was: start at 100°C, hold for 1 min, increase by 8°C/min
to 236°C, increase by 35°C/min to 310°C, 6 min at 310°C. After automatic
calibration, the mass spectrometer was operated under ammonia-positive ionization
mode. Appropriate ion sets were monitored with a dwell time of 25 to
35 ms/ion, at m/z 226/233 (PA/[ 2 H 7]PA), 254/259 (PB/[ 2 H 5]PB), 325/330
(PHB/[ 2 H 5]PHB), and 252 (PtC). Note that PKB and phenylacetone had been
reduced with NaB 2 H 4 to monodeuterated R,S-3-hydroxy-4-phenylbutyrate
(monitored at m/z 326/330) and 1-phenyl-2-propanol (monitored at m/z 200/
210 and/or 217/227). Also, since 1-phenyl-2-propanol was assayed with a
standard of 1-phenyl[ 2 H 5]ethanol, ions monitored for this assay were 200/209
or 217/226.
For the analysis of chiral PHB derivatives, the GC injector temperature was
set at 280°C. The column (DB-5, 60 m � 0.25 mm i.d., 0.5 mm film thickness;
Hewlett Packard) program was modified to the initial 150°C for 2 min,
increased by 15°C/min to 230°C, 25 min at 230°C, increased by 35°C to
290°C, and held 10 min. Ions monitored were 1) 358 (M � 18, i.e., M �
NH 4 � ) for analytes and 2) 363 (M � 18 � 5, i.e., M � 5�NH4 � ) for
[ 2 H 5]PHB. The mass spectrometer was operated under ammonia- positive
chemical ionization and was tuned automatically.
Phenylacetylglycine was analyzed as its methyl ester derivative using an
OV-225 column (29 m � 0.32 mm i.d., 1 �m film thickness; Quadrex
Corporation, Woodbridge, CT). This column yielded better resolution of
N-phenylacetylglycine methyl ester with no peak tailing. Samples (0.2–1 �l)
were injected with a split ratio 20:1. The carrier gas was helium (constant flow:
1.2 ml/min). The injector port temperature was at 220°C, the transfer line at
240°C, the source temperature at 200°C, and quadrupole at 106°C. The column
temperature program was: start at 90°C, hold for 1 min, increase by 10°C/min
to 240°C, 15 min at 240°C. After automatic calibration, the mass spectrometer
was operated under ammonia-positive ionization mode (pressure adjusted to
optimize peak areas). Ions monitored were 1) 208 (M � 1, i.e., M � H �) and
225 (M � 18, i.e., M � NH 4 � ) for the analyte and 2) 215 (M � 7 � 1, M �
7 � H � ) and 232 (M � 7 � 18, M � 7 � NH 4 � ) for N-[ 2 H7]PA-glycine with
a dwell time of 25 ms/ion.
Areas under each chromatogram were determined by interactive computer
integration, and corrected for naturally occurring heavy isotopes and light
isotopic impurities in the synthesized labeled internal standards.
NMR Spectroscopy. Proton NMR spectroscopy was performed at 400
MHz on a Bruker Avance (Bruker, Newark, DE) DMX 400 wide-bore spectrometer
using a 5-mm inverse probe. Full-strength urine samples were obtained
by lyophilizing 5 ml of urine to dryness. The residue was dissolved in
0.5 ml of 2 H 2O, and the solution was introduced into a 5-mm NMR tube. An
external standard made of a sealed capillary containing a solution of trimethylsilylpropionic
acid in 2 H 2O was introduced into the NMR tube and used as
chemical shift reference. Standard acquisition conditions were as follows for
one-dimensional spectra: 45° pulse, 8-s repetition time, water saturation during
the relaxation delay, sweep width (SW) 6775 Hz, 64K data points (TD), and
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

32 scans of data collection. Two-dimensional correlation spectroscopy
(COSY) spectra were obtained with the following conditions for the second
dimension: SW 3500 Hz, TD 2K, 16 scans, and for the first dimension, 512
increments of 278 �s zerofilled to 1K. A nonshifted sinebell window was
applied in both dimensions, and magnitude spectra were calculated. Twodimensional
1 H/ 13 C correlations via double insensitive nuclei enhanced by
polarization transfer (HSQC) were performed in the phase-sensitive mode
(TPPI; time-proportional receiver phase incrementation) using gradients for
coherence selection and carbon decoupling during acquisition. The following
conditions were used in the second dimension: SW 3200 Hz, TD 2K, 128
scans, and in the first dimension: SW 12 kHz, 256 increments of 20.7 �s
zerofilled to 512. A shifted sinebell window was applied in both dimensions.
Proton-decoupled carbon spectra of the concentrated urine samples were
obtained at 100.62 MHz in a 5-mm dual probe. Acquisition conditions were as
follows: 20° pulse, repetition time 1.3 s, SW 25 kHz, TD 64K, 40,000 scans.
The free induction decays were zerofilled to 128K, and a Lorentz to Gauss
transformation (LB ��1 Hz, GB � 0.1) was applied before Fourier transformation.
Clinical Investigation. The protocol was reviewed and approved by the
Institutional Review Board of University Hospitals of Cleveland. All subjects
were free of any chronic or acute illness. Women had a negative pregnancy test
and were not breastfeeding. Seven subjects (three men, four women; 31.7 �
5.0 years; 171.3 � 3.4 cm; 79.5 � 5.9 kg) received detailed information on the
purpose of the investigation and signed an informed consent form. After an
overnight fast, the subjects were admitted to the Clinical Research Center at
7:30 AM. They remained fasting until the completion of the study. An
intravenous line was installed in the forearm with a saline infusion (20 ml/h)
and a short blood sampling catheter was inserted into a superficial vein of the
contralateral hand. The hand was placed in a heating box at 60°C for sampling
of arterialized venous blood. At 8:00 AM, after baseline blood and urine
sampling, each subject ingested 0.36 mmol/kg (5 g/75 kg) Na-PB. This dose
corresponds to 11 to 17% of the doses commonly used in the treatment of
patients with inborn errors of urea synthesis (0.4–0.6 g � kg �1 � day �1 ). Water
intake was adjusted to induce a diuresis of at least 100 ml/30 min. Urine
samples were collected at 30-min intervals for the first 3 h after PB ingestion,
and then every hour until 8 h. Urine samples were quickly frozen and stored
at �80°C until analysis.
Organ Perfusion Experiments. Livers from fed male Sprague-Dawley rats
kept on standard rat chow (200–230 g) were perfused (Brunengraber et al.,
1975) with recirculating Krebs-Ringer-bicarbonate buffer containing 4% bovine
serum albumin (fraction V, fatty acid poor; Intergen, Purchase, NY) and
10 mM glucose. The bile duct was cannulated with PE 10 tubing (BD
Biosciences, San Jose, CA) for bile collection. Throughout the 2-h experiment,
sodium taurocholate (38 �mol/h) was infused into the perfusion reservoir to
stimulate bile flow (Robins and Brunengraber, 1982). After 30 min of equilibration,
a calculated amount of either PB, R,S-PHB, or PKB was added to the
perfusate to set an initial concentration of 5 mM. The perfusion continued until
120 min. The pH of the perfusate was monitored and kept at 7.3 to 7.4 by
adding 0.3 M NaOH. Samples of bile and perfusate were collected at regular
intervals. For the assay of PKB, perfusate samples (2 ml) were treated immediately
with 0.2 ml of 0.1 M NaB 2 H 4 in 0.1 mM NaOH to convert unstable
PKB to stable monodeuterated R,S-PHB. Bile samples were collected every 30
min. At the end of the experiment, the livers were quick-frozen with aluminum
tongs precooled in liquid nitrogen.
Rat in Vivo Experiments. We tested the bioavailability of two PB esters as
a means to deliver large amounts of PB without the corresponding sodium
load. Overnight-fasted rats (330–400 g) were divided into six groups (5–7 rats
per group) for the testing of three different PB preparations: Na-PB, glycerolmono-PB,
glycerol-tri-PB, ribose-tetra-PB, glucose-penta-PB, and sorbitolhexa-PB.
Each rat received one stomach gavage of the sodium salt or ester in
an amount that delivered 2.15 mmol PB/kg. The weighed dose for each rat was
mixed with 3 ml of Tween and administered to the rats through a stomach
gavage needle.
Whole blood samples (100–200 �l) were taken at �5, 15, 30, 60, 150, 240,
330, 420, and 480 min from a small incision in a tail vein. Blood was collected
in heparinized microcapillary tubes and centrifuged. The plasma (50–100 �l)
was transferred to an Eppendorf tube and quick frozen.
NEW PHENYLBUTYRATE METABOLITES
Results
Human Study. In our previous study, we had identified PBGN in
the plasma and urine of normal adults who had ingested a small dose
of PB. We now report the data of additional analyses conducted on the
same samples of human urine. First, we subjected to NMR analysis
two samples of urine produced by each subject before and 2 h after
ingestion of PB. The NMR spectrum of the second sample, but not of
the first, was highly suggestive of the presence of a product of
hydroxylation of the side chain of PB. In the COSY spectrum of the
lyophilized urine dissolved in D2O (after PB ingestion), we identified
a proton at 4.25 ppm coupled with two CH2’s. The first CH2 has
protons at 2.87 and 2.70 ppm; the second has nearly identical protons
at 2.4 ppm. A chemical shift at 4.25 ppm is likely corresponding to a
proton coupled to the OH group. Therefore, the COSY spectrum
revealed the presence of a metabolite having –CH2–CHOH–CH2– moiety. In the HSQC spectrum these proton signals correlated with
the following carbons: CH at 70 ppm overlapping with other CH
carbohydrate carbons, CH2 at 44.6 ppm and CH2 at 42.5 ppm. Lastly,
in the aromatic region, signals at 129.6, 128.7, and 126.6 ppm corresponded
to a monosubstituted phenyl group having the same intensity
per carbon as the signals of the –CH2–CHOH–CH2– group. We
therefore concluded that �-hydroxy-PB (PHB) was present in the
urine of patients treated with PB. PHB would be a very likely
metabolite since it would be formed via partial �-oxidation of PB to
PHB-CoA (presumably the S-enantiomer), which would be hydrolyzed
to free PHB.
To further confirm the identity of the urinary metabolite detected by
NMR, we synthesized unlabeled and R,S[ 2 H5]PHB. The di-TMS
derivative of synthetic R,S-PHB was analyzed by GC-MS in parallel
with an extract of human urine (after PB ingestion) that had been
reacted with TMS. In the sample derived from urine, we found a peak
at the same retention time and with the same mass spectra (electron
ionization and NH3-positive chemical ionization) as the standard of
R,S-PHB. In addition, the NMR spectrum of synthetic PHB had the
same chemical shifts as the material identified in human urine. This
confirmed the identity of PHB in human urine but did not yield
information about its chirality. The chirality of excreted PHB yields
information on the mechanism of its formation (see below).
For the chromatographic separation of PHB enantiomers from the
synthetic racemate, we tried various chiral hydroxyl derivatization
reagents before selecting the combination of 1) methylation of the
carboxyl group, and 2) reaction of the hydroxyl group with S-2phenylbutyryl
chloride (Powers et al.,1994). The expected derivatives
of R- and S-PHB were well separated, and their order of elution was
confirmed using a sample of R-PHB that we had synthesized. R-PHB
elutes ahead of the S-derivative (Fig. 2).
Chiral GC-MS analysis of the human urine samples (after PB
ingestion) revealed that PHB is present as an enantiomeric mixture
with 10% R-PHB and 90% S-PHB (Fig. 2). Figure 3 shows the time
profile of (R�S)-PHB excretion in urine after an oral bolus of Na-PB.
The excretion of (R�S)-PHB peaked at 120 to 240 min. Eight hours
after ingestion of PB, the cumulative excretion of (R�S)-PHB
(1.35 � 0.13 mmol) amounted to 4.4 � 0.56% of the PB dose.
Small amounts of phenylacetone and 1-phenyl-2-propanol were
identified in the urine samples (Table 1). Treatment of urine samples
with �-glucuronidase increased their PB � PA content. The total
amount of PB � PA released by �-glucuronidase amounted to 2.4 �
0.3% of the PB dose.
Perfused Rat Liver Study. The metabolism of PB was studied in
perfused rat livers by the addition of 5 mM PB to the recirculating
perfusate. The time profile of the PB concentration was curvilinear
13
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

14 KASUMOV ET AL.
FIG. 2.Chiral GC-MS assay of R- and S-PHB excreted in one sample of human
urine after oral ingestion of 0.36 mmol/kg Na-PB (bold trace).
The sample was spiked with R,S-[ 2 H5]PHB, the enantiomer profile of which is
shown by the thin trace.
FIG. 3.Time course of (R�S)-PHB urinary excretion in normal humans after
oral ingestion of 0.36 mmol/kg Na-PB (mean � S.E.M.; n � 7).
(Fig. 4A). Plotting of the data under semilogarithmic coordinates
yielded a linear relationship compatible with a first order kinetic
process for PB uptake. The total uptake of PB amounted to 310 � 16
�mol � (90 min �1 ) � liver �1 .
Figure 4B shows the time accumulation of metabolites derived
from PB and released into the perfusate. Phenylacetylglycine is a
TABLE 1
Recovery of PB and its metabolites in human urine (n � 7) after the oral
ingestion of 0.36 mmol/kg Na-PB and production of PB metabolites by rat liver
(n � 5) perfused with 5 mM PB
Metabolite
% of PB Uptake
Humans Rats
Free PB 0.97 � 0.23
PB-�-glucuronide 1.29 � 0.33 21.52 � 4.32
Free PA 0.26 � 0.06 7.64 � 0.87
PA-glycine 7.68 � 0.84
PA-�-glucuronide 1.11 � 0.19 5.25 � 0.88
PAGN a 32.6 � 1.9
PBGN a 21.5 � 2.4
PHB 4.4 � 0.56 15.71 � 1.63
PKB 4.54 � 0.29
PtC 5.15 � 0.59
Phenylacetone 0.11 � 0.007 3.67 � 0.23
1-Phenyl-2-propanol 0.01 � 0.001 Trace amounts
Total bile met 2.97 � 0.61
Total 62.4 � 2.1 74.12 � 5.58
a Previously identified metabolites (Comte et al., 2002).
FIG. 4.Metabolism of PB in perfused rat livers.
A, uptake of PB from the recirculating perfusate. B, accumulation of PB
metabolites in the perfusate (n � 5 for all compounds except for PKB, where n �
4).
known metabolite of PA in rats and dogs (Knoop, 1904; Ambrose and
Sherwin, 1933; James et al., 1972). About 16% of the uptake of PB
was accounted for by the production of R- and S-PHB, of which about
90% is the S-enantiomer. We also identified PtC, PKB, phenylacetone,
and 1-phenyl-2-propanol, which, to our best review of the
literature, have not been previously described as metabolites of PB.
The identity of these metabolites was confirmed by GC-MS using
standards we synthesized. Incubation of liver perfusate samples with
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

FIG. 5.Biliary excretion of PB and its metabolites (R�S)-PHB and PA by
perfused rat livers.
The white bars show the biliary concentrations of the free metabolites. The dark
bars show the total concentrations of metabolites after alkaline hydrolysis.
�-glucuronidase revealed the formation of PB and PA glucuronides,
which accounted for 22 and 5% of the PB uptake, respectively (Fig.
4B; Table 1).
Figure 5 shows the time profile of biliary excretion of PB and its
metabolites. At the end of the experiments, the concentrations of free
PB, PA, and PHB in bile (2.14, 0.08, and 0.18 mM, respectively, Fig.
5) were 50 to 70% of the corresponding concentrations in the perfusate
(3.11, 0.16, and 0.28 mM, respectively, Fig. 4). Alkaline treatment
of the bile significantly increased the amounts of PB, PA, and PHB,
presumably via hydrolysis of glucuronides. The total amount of free
� alkali-released PB, PA, and PHB (Fig. 5, dark bars) amounted to
3.0 � 0.6% of the amount of PB taken up by the liver from the
perfusate.
To compare the metabolism of the PHB enantiomers, one rat liver
was perfused with 5 mM R,S-PHB added at 30 min. Figure 6 shows
the GC-MS chromatograms of the assays of R- and S-PHB in the
perfusate at 32 and 120 min. At 32 min, the perfusate contained equal
concentrations of R- and S-PHB, which was expected since a racemic
mixture of PHB had been used (Fig. 6A). However, at 120 min, the
concentration of S-PHB had decreased much more than that of R-PHB
(Fig. 6B). Using the internal standards of R,S-[ 2 H 5]PHB, we found
that the uptake of R-PHB was immeasurable. In contrast, the uptake of
S-PHB was 134 �mol. A very small amount of PKB (3 �mol) was
found, derived presumably from the oxidation of R-PHB by R-�hydroxybutyrate
dehydrogenase. At 120 min, the perfusate contained
55 �mol of PA resulting from the �-oxidation of S-PHB.
To test for the interconversion of R-PHB and PKB, another rat liver
was perfused with 5 mM PKB added at 30 min. At 120 min, 358 �mol
of PKB had disappeared from the perfusate, and 90% of this had
NEW PHENYLBUTYRATE METABOLITES
FIG. 6.Metabolism of R,S-PHB in a perfused rat liver.
A, chiral assay of the PHB enantiomers in a sample of perfusate taken 2 min after
adding 5 mM R,S-PHB to the recirculating perfusate. The thin trace corresponds to
the R,S-[ 2 H5]PHB internal standard added during sample analysis. B, similar analysis
on a sample of perfusate taken after 90 min of recirculating perfusion.
appeared as phenylacetone, the product of PKB decarboxylation. Only
traces of 1-phenyl-2-propanol were detected. Figure 7 shows the
GC-MS chromatogram of the assays of R- and S-PHB in the perfusate
at 120 min. Small amounts of R-PHB (2.3 �mol) and S-PHB (about
0.2 �mol) were detected. Table 1 presents a balance of the conversion
of PB to metabolites in humans (urinary metabolites) and in perfused
rat livers (metabolites released in the perfusate and bile).
Bioavailability of PB Esters. Rats were gavaged with equivalent
doses of Na-PB, glycerol-mono-PB, or glycerol-tri-PB suspended in
Tween. The plasma concentrations of PB and PA were then followed
for 8 h (Fig. 8). Table 2 shows the comparison of the areas under the
curves (AUC; determined using the linear trapezoidal rule) for each of
the compounds tested. Based on the AUCs for PB, the bioavailability
of the glycerol monoester or triester are about 63 and 42%, respectively,
of that of the sodium salt. However, the bioavailability of the
glycerol-mono-PB ester is not significantly different from that of the
sodium salt. The bioavailability of the glycerol-tri-PB ester is significantly
lower than that of the sodium salt. In all cases, the concentrations
of PB and PA had decreased to very low levels after 8hof
gavage. In fact, Na-PB and glycerol-mono-PB esters have the same
T max at 30 min, with a similar C max for PB concentration in plasma,
namely, 0.32 mM and 0.22 mM. However, for the glycerol-tri-PB, the
C max is much lower, reaching a concentration of only 0.16 mM with
its T max at 15 min. For all three preparations, the T max values for the
PA metabolite are identical at 150 min, whereas their respective C max
values are different, namely, 0.25 mM, 0.23 mM, and 0.14 mM for
Na-PB, glycerol-mono-PB, and glycerol-tri-PB, respectively. Other
rats were gavaged with equivalent doses of ribose tetra-PB, glucose
penta-PB, or sorbitol-hexa-PB. Over the next 6 h, no PB or PA was
detected in their plasma (not shown), suggesting that these compounds
were not bioavailable.
Discussion
Phenylbutyrate (PB) was one of the �-phenyl fatty acids used
originally by Knoop as model compounds to elucidate the mechanism
of fatty acid �-oxidation (Knoop, 1904). He showed that after oral
administration of PB to dogs, phenylacetylglycine (phenaceturic acid)
is excreted in urine. This suggested that PB underwent one cycle of
chain shortening, forming PA, which was then coupled with glycine.
The metabolism of PB is different in humans, in whom PA conjugates
with glutamine to form PAGN, rather than phenylacetylglycine (Ambrose
and Sherwin, 1933). The mechanism of this conjugation was
studied by Webster et al. (1976), who showed that it proceeds via
phenylacetyl-CoA and is catalyzed by an acyl-CoA: L-glutamine
N-acyltransferase.
15
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

16 KASUMOV ET AL.
FIG. 7.Production of the R- and S- enantiomers of PHB from PKB in a perfused
rat liver.
The liver was perfused for 90 min with recirculating perfusate initially containing
5 mM PKB. The final perfusate sample was analyzed for PHB enantiomers (thick
trace). The sample was spiked with R,S-[ 2 H5]PHB, the enantiomer profile of which
is shown by the thin trace.
In the 1980s, Brusilow and collaborators started using PA for the
treatment of patients with inborn errors of the urea cycle (Batshaw et
al., 1981). Later, PB was used as a precursor of PA because it did not
have the foul smell of PA, and thus was more acceptable to patients
(Brusilow, 1991; Batshaw et al., 2001). Most recently, PB has been
used in various clinical trials for the treatment of malignancies, cystic
fibrosis, sickle cell anemia, and thalassemia (Dover et al., 1994;
Hoppe et al., 1999; Rubenstein and Zeitlin, 2000; Gilbert et al., 2001).
The use of large doses of Na-PB for therapeutic purposes has raised
several issues. First, the recovery of known urinary metabolites of PB
(PAGN � traces of PB and PA) was very variable, ranging from 50
to 90% of the ingested dose (Piscitelli et al., 1995, Comte et al., 2002).
Since PB is known to have toxic effects at high doses, some of the
unknown metabolites might contribute to its toxicity. It is also possible
that part of the ingested PB was not absorbed in the gut and was
excreted as such (or as secondary metabolites formed in the colon) in
feces. Second, in cancer patients (and presumably other patients as
well), the high sodium load associated with the treatment (0.4–0.6
g/kg) results in an expansion of the extracellular fluid, which could be
dangerous in patients with cardiac or renal dysfunction.
When we learned of the under-recovery of the PB dose, we hypothesized
that part of PB could be excreted in the form of PBGN, an
analog of PAGN. Indeed, we previously identified PBGN in the urine
of normal adult humans who had ingested small doses of PB (5 g)
(Comte et al., 2002). Still, the total recovery of the PB dose as urinary
PBGN � PAGN � PB � PA accounted for only about half of the
dose. We thus looked for other metabolites of PB in the urine of the
original subjects, concentrating on 1) compounds related to intermediates
of fatty acid �-oxidation, and 2) glucuronide conjugates. The
compounds identified are shown in the scheme of PB metabolism
presented in Fig. 9.
Logical intermediates in the �-oxidation of PB are 4-phenyl-transcrotonyl-CoA,
S-3-hydroxy-4-phenylbutyryl-CoA, 4-phenyl-3-keto-
FIG. 8.Bioavailability of phenylbutyrate administered as sodium salt (A),
monoglycerol ester (B), and triglycerol ester (C).
The bioavailability is reflected by the plasma concentrations of PB and PA
following gavage of the rats with 2.15 mmol PB equivalents/kg.
TABLE 2
Comparison of AUCs for PA and PB concentrations
Rats were gavaged with the indicated compounds at a dose of 2.15 mmol PB
equivalents/kg. The AUCs (mM � min) were calculated from the plasma PA and PB
concentrations. Data are presented as mean � S.E.M., n � 7 for each group.
Compound
Administered
AUC PB AUC PA
Ratio of the Means
[(AUC PB)/AUC PA]
Na-PB 83.1 � 14.7 74.9 � 21 1.1
Glycerol mono-PB 51.7 � 7.1 71.5 � 24 1.4
Glycerol tri-PB 34.6 � 5.0* 26.6 � 8.0 1.3
* Significantly different from Na-PB controls (p � 0.01).
butyryl-CoA, and phenylacetyl-CoA. The latter is known to undergo
partial hydrolysis to free PA, which accumulates transiently in body
fluids after the ingestion of PB. We hypothesized that the other CoA
intermediates could also undergo partial hydrolysis to free acids,
which would then be excreted in urine. Although we identified S-3hydroxy-4-phenylbutyrate
(S-PHB), we were unable to find PKB or
4-phenyl-trans-crotonate in the urine of the human subjects who had
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

FIG. 9.Scheme of PB metabolism in humans and rats.
The newly identified metabolites are identified by asterisks. Enzymes: (1), acyl-CoA synthetase; (2), acyl-CoA dehydrogenase; (3), enoyl-CoA hydratase; (4),
S-3-hydroxyacyl-CoA dehydrogenase; (5), 3-ketoacyl-CoA thiolase; (6), acyl-CoA-L-glutamine N-acyltransferase; (7), UDP-glucuronyl transferase; (8), acyl-CoA
hydrolase; (9), S-3-hydroxyacyl-CoA synthetase; (10), R-�-hydroxybutyrate dehydrogenase; (11), glycine N-acylase; (12), spontaneous phenylketobutyrate decarboxylation;
(13), alcohol dehydrogenase.
ingested a small dose of PB (5 g). It is possible that these compounds
are excreted after the ingestion of therapeutic doses of PB (0.4–0.6
g/kg).
The chiral GC-MS assay we used to identify S-PHB (Powers et al.,
1994) revealed that the R-enantiomer accounted for only about 10% of
the total PHB excreted by humans. The probable origin of this
compound would involve hydrolysis of PKB-CoA to PKB, which
would then be reduced to R-PHB by R-�-hydroxybutyrate dehydrogenase
(Fig. 9). We did not identify PKB in the human urine, possibly
because it undergoes spontaneous decarboxylation similar to acetoacetate
(Pedersen, 1929). Indeed, in the subjects’ urine, we found
phenylacetone, which derives from the presumably spontaneous decarboxylation
of PKB. In addition, we found measurable amounts of
1-phenyl-2-propanol, which is also formed from the reduction of
phenylacetone in microsomes (Coutts et al., 1981). The absence of
1-phenyl-2-propanol in the urine before PB ingestion confirms its
origin from PB. Note that phenylacetone and 1-phenyl-2-propanol are
also products of amphetamine catabolism (Shiiyama et al., 1997; Lee
et al., 1999). It is not known whether phenylacetone and 1-phenyl-2propanol
contribute to the toxic side effects of amphetamine.
In searching for metabolites of PB, we looked for glucuronides
since it was known that benzoate, which is also used for the treatment
of urea cycle disorders, is eliminated in part as a glucuronide (Riddick,
1998). Small amounts of PB and PA glucuronides were indirectly
identified in human urine by the increases in PB and PA
concentrations induced by incubating the urine with �-glucuronidase.
The total recovery of PB and its metabolites in human urine
accounted for only 62% of the ingested dose. Part of the missing
fraction could still be disposed of as unknown urinary metabolites.
Also, one cannot exclude the possibility that some PB and/or PB
metabolites are excreted in feces. We do not know whether PB
NEW PHENYLBUTYRATE METABOLITES
glucuronide excreted in bile is entirely hydrolyzed in the gut with
complete reabsorption of the PB moiety of the glucuronide.
To gain more insight into the metabolism of PB that does not
involve glutamine conjugation (which is specific to humans and
primates), we perfused rat livers with recirculating perfusate containing
an initial PB concentration of 5 mM. To ensure that PKB (which
could possibly be formed in these experiments) would not undergo
spontaneous decarboxylation, we treated some of the samples with
NaB 2 H 4. This converts PKB to R,S-[ 2 H]PHB, which can then be
differentiated from the original PHB by GC-MS. This technique has
been used previously to protect acetoacetate and �-ketopentanoate
prior to analysis (Des Rosiers et al., 1988; Leclerc et al., 1995).
The uptake of PB by rat liver [310 � 16 �mol � (90 min �1 ) � liver �1
or roughly 0.34 �mol � min �1 � g �1 ] is of the same magnitude as the
uptake of long-chain fatty acids such as oleate (Brunengraber et al.,
1978). The apparent first order character of PB uptake shows that it
was not saturated at 5 mM substrate (Fig. 4A). Perfusion of rat livers
with PB resulted in the production of PKB and PtC. Although these
compounds have not been found in human urine after ingestion of a
small dose of PB, the presence of phenylacetone and R-PHB is
indirect evidence for the formation of PKB (Fig. 9). It is quite possible
that, after ingestion of large therapeutic doses of PB, humans would
excrete measurable amounts of PKB and PtC.
Sixteen percent of the uptake of PB by the perfused rat liver was
released in the form of PHB, 90% of which was the S-enantiomer,
similar to what we found in humans. To probe the origin of the R-PHB
enantiomer, one liver was perfused with 5 mM PKB, and the perfusate
was analyzed for the presence of PHB enantiomers. As expected, most
of the PHB generated from PKB was in the R form (Fig. 7), which
most likely derived from the action of R-�-hydroxybutyrate dehydrogenase
on PKB (Fig. 9). Presumably, the production of small amounts
of R-PHB in humans who ingested PB can also be attributed to the
17
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

18 KASUMOV ET AL.
reduction of PKB. Note that the bacterial R-�-hydroxybutyrate dehydrogenase
used in the enzymatic assay of acetoacetate and R-�hydroxybutyrate
does not use PKB or R,S-PHB as a substrate.
Rat livers perfused with 5 mM PB produced large amounts of PB
and PA glucuronides, corresponding to 22% and 5% of the PB uptake,
respectively. These glucuronides were formed presumably from free
PB and PA, since glucuronidation of carboxylates does not require
activation to a CoA ester. The difference in the fractions of the PB
uptake recovered as PB versus PA glucuronide results probably from
the low concentration of PA derived from PB (Fig. 4).
Figure 9 summarizes the proposed metabolism of PB, including the
new compounds identified in this report. The proportion of PB and PA
glucuronides is much smaller in humans compared with perfused rat
livers (Table 1). Note that the humans were overnight-fasted and,
therefore, had presumably low portal vein glucose concentrations, and
fairly low glycogen concentrations in the liver. In contrast, the livers
from rats fed ad libitum were perfused with a high glucose concentration
(10 mM), a concentration likely to occur in the portal vein
during the absorption of digested food. Liver glucokinase has a high
K m for glucose (10 mM); thus, the uptake of glucose is practically
proportional to the portal glucose concentration in the physiological
range. Since the production of glucuronides uses glucose 1-phosphate
derived from glucose phosphorylation and from glycogenolysis, the
rate of glucuronidation may be partially dependent on the glucose
concentration in the portal vein. We could not find in the literature
reports on studies on the influence of portal vein glucose concentration
on rates of xenobiotic glucuronidation in intact livers. We found
one report on the stimulation of acetaminophen glucuronidation in
humans fed a high-carbohydrate diet compared with a low-carbohydrate
diet (Pantuck et al., 1991). Although the stimulation of glucuronidation
can result from a high glucokinase flux or a high rate of
glycogen cycling, both mechanisms would operate after the ingestion
of a high-carbohydrate diet.
Our data raise several clinically relevant questions. First, if patients
with urea cycle defects ingest their PB dose with a high-carbohydrate
meal, or if decompensated patients are infused with intravenous
glucose and PA, would a large fraction of the PB/PA dose be excreted
as glucuronides, i.e., compounds that do not carry nitrogen out of the
organism? Second, if patients ingest their PB dose with a high-fat
meal, would competition between PB and dietary long-chain fatty
acids for �-oxidation result in excretion of side products (PHB, PKB,
PtC), which are also non-nitrogenous? Either of these scenarios would
decrease the clinical effectiveness of PB therapy. Third, independent
of the composition of the diet, could saturation of �-oxidation by PB
itself lead to the excretion of these same side products? Fourth, could
the accumulation of �-oxidation side products such as PtC explain
episodes of toxicity encountered occasionally during treatment with
PB (Hoppe et al., 1999; Carducci et al., 2001; Gilbert et al., 2001;
Gore et al., 2002)? Although we were unable to find reports of
phenylcrotonate toxicity, this compound is present in toxic pesticides
(Dinocap). Other side products of PB metabolism, phenylacetone and
1-phenyl-2-propanol, are also formed during amphetamine intoxication
(Shiiyama et al., 1997; Lee et al., 1999), but it is not clear whether
they contribute to the toxicity of amphetamine or phenylbutyrate. The
production of phenylacetone and 1-phenyl-2-propanol may be increased
under conditions associated with stress and stimulation of
lipolysis. It is well known that stimulation of lipolysis raises the
concentration of plasma free fatty acids, thus increasing the competition
for �-oxidation.
In conclusion, by presenting our preliminary data, we hope to draw
the attention of clinicians to the potential interference between the
metabolism of PB and the metabolism of dietary carbohydrates and
fatty acids. It might be advisable to administer PB in multiple doses
between meals, rather than with meals. Also, the potential toxicity of
side products of PB metabolism needs to be thoroughly investigated.
From the metabolic balances of Table 1, it is clear that a substantial
fraction of the metabolites of PB has not yet been identified (38% in
humans and 26% in our perfused rat liver experiments). Carnitine
derivatives of PB and PA, as well as compounds resulting from
hydroxylation of the benzene ring of PB, might be formed, and we are
presently searching for them.
Lastly, the sodium-free glycerol-mono-PB ester (Fig. 8; Table 2)
could be considered for patients who might be harmed by a large
sodium load and/or by the causticity of sodium PB. Since this ester is
a tasteless solid compound, it could be incorporated in processed
foods and ingested as snacks in between meals. The above questions
need to be investigated in detailed clinical and animal studies.
References
Ambrose AM and Sherwin CP (1933) Further studies on the detoxication of phenylacetic acid.
J Biol Chem 101:669–675.
Batshaw ML, MacArthur RB, and Tuchman M (2001) Alternative pathway therapy for urea cycle
disorders: twenty years later. J Pediatr 138:S46–S54.
Batshaw ML, Thomas GH, and Brusilow SW (1981) New approaches to the diagnosis and
treatment of inborn errors of urea synthesis. Pediatrics 68:290–297.
Brunengraber H, Boutry M, Daikuhara Y, Kopelovich L, and Lowenstein JM (1975) Use of the
perfused liver for the study of lipogenesis. Methods Enzymol 35:597–607.
Brunengraber H, Boutry M, and Lowenstein JM (1978) Fatty acid, 3-�-hydroxysterol and ketone
synthesis in the perfused rat liver. Effects of (�)-hydroxycitrate and oleate. Eur J Biochem
82:373–384.
Brusilow SW (1991) Phenylacetylglutamine may replace urea as a vehicle for waste nitrogen
excretion. Pediatr Res 29:147–150.
Capozzi G, Roelens S, and Talami S (1993) A protocol for the efficient synthesis of enantiopure
beta-substituted beta-lactones. J Org Chem 58:7932–7936.
Carducci MA, Gilbert J, Bowling MK, Noe D, Eisenberger MA, Sinibaldi V, Zabelina Y, Chen
TL, Grochow LB, and Donehower RC (2001) A Phase I clinical and pharmacological
evaluation of sodium phenylbutyrate on a 120-h infusion schedule. Clin Cancer Res 7:3047–
3055.
Comte B, Kasumov T, Pierce BA, Puchowicz MA, Scott ME, Dahms W, Kerr D, Nissim I, and
Brunengraber H (2002) Identification of phenylbutyrylglutamine, a new metabolite of phenylbutyrate
metabolism in humans. J Mass Spectrom 37:581–590.
Coutts RT, Prelusky DB, and Jones GR (1981) The effects of cofactor and species differences on
the in vitro metabolism of propiophenone and phenylacetone. Can J Physiol Pharmacol
59:195–201.
Des Rosiers C, Montgomery JA, Desrochers S, Garneau M, David F, Mamer OA, and Brunengraber
H (1988) Interference of 3-hydroxyisobutyrate with measurements of ketone body
concentration and isotopic enrichment by gas chromatography-mass spectrometry. Anal Biochem
173:96–105.
Dover GJ, Brusilow S, and Charache S (1994) Induction of fetal hemoglobin production in
subjects with sickle cell anemia by oral sodium phenylbutyrate. Blood 84:339–343.
Gilbert J, Baker SD, Bowling MK, Grochow L, Figg WD, Zabelina Y, Donehower RC, and
Carducci MA (2001) A phase I dose escalation and bioavailability study of oral sodium
phenylbutyrate in patients with refractory solid tumor malignancies. Clin Cancer Res 7:2292–
2300.
Gore SD, Weng LJ, Figg WD, Zhai S, Donehower RC, Dover G, Grever MR, Griffin C,
Grochow LB, Hawkins A, et al. (2002) Impact of prolonged infusions of the putative
differentiating agent sodium phenylbutyrate on myelodysplastic syndromes and acute myeloid
leukemia. Clin Cancer Res 8:963–970.
James MO, Smith RL, Williams RT, and Reidenberg M (1972) The conjugation of phenylacetic
acid in man, sub-human primates and some non-primate species. Proc R Soc Lond B Biol Sci
182:25–35.
Knoop F (1904) Der Abbau aromatischer Fettsauren im Tierkörper. Ernst Kuttruff, Freiburg,
Germany.
Leclerc J, Des Rosiers C, Montgomery JA, Brunet J, Ste-Marie L, Reider MW, Fernandez CA,
Powers L, David F, and Brunengraber H (1995) Metabolism of R-beta-hydroxypentanoate and
of beta-ketopentanoate in conscious dogs. Am J Physiol 268:E446–E452.
Lee MR, Jeng J, Hsiang WS, and Hwang BH (1999) Determination of pyrolysis products of
smoked methamphetamine mixed with tobacco by tandem mass spectrometry. J Anal Toxicol
23:41–45.
Loffler A, Norris F, Taub W, Swanholt KL, and Dreiding AS (1970) Thermische Lactonisierung
von Estern gamma-Brom-alpha, beta-ungesattigter Carbonsauren zy Delta alpha Butenolieden.
Helv Chim Acta 53:403–417.
Magnusson I, Schumann WC, Bartsch GE, Chandramouli V, Kumaran K, Wahren J, and Landau
BR (1991) Noninvasive tracing of Krebs cycle metabolism in liver. J Biol Chem 266:6975–
6984.
Pantuck EJ, Pantuck CB, Kappas A, Conney AH, and Anderson KE (1991) Effects of protein and
carbohydrate content of diet on drug conjugation. Clin Pharmacol Ther 50:254–258.
Pedersen KJ (1929) The ketonic decomposition of beta-keto carboxylic acids. J Am Chem Soc
51:2098–2107.
Piscitelli SC, Thibault A, Figg WD, Tompkins A, Headlee D, Lieberman R, Samid D, and Myers
CE (1995) Disposition of phenylbutyrate and its metabolites, phenylacetate and phenylacetylglutamine.
J Clin Pharmacol 35:368–373.
Powers L, Ciraolo ST, Agarwal KC, Kumar A, Bomont C, Soloviev MV, David F, Desrochers
S, and Brunengraber H (1994) Assay of the enantiomers of 1,2-propanediol, 1,3-butanediol,
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012

1,3-pentanediol and the corresponding hydroxyacids by gas chromatography-mass spectrometry.
Anal Biochem 221:323–328.
Ramsdell HS and Tanaka K (1977) Gas chromatographic studies of twenty metabolically
important acylglycines. Clin Chim Acta 74:109–114.
Riddick DS (1998) Drug biotransformation, in Principles of Medical Pharmacology (Kalant H
and Roschlau WHE eds) pp 38–54, Oxford University Press, New York.
Robins SJ and Brunengraber H (1982) Origin of biliary cholesterol and lecithin in the rat:
contribution of new synthesis and preformed hepatic stores. J Lipid Res 23:604–608.
Rubenstein RC and Zeitlin PL (2000) Sodium 4-phenylbutyrate downregulates Hsc70: implications
for intracellular trafficking of DeltaF508-CFTR. Am J Physiol 278:C259–C267.
Samid D, Hudgins WR, Shack S, Liu L, Prasanna P, and Myers CE (1997) Phenylacetate and
phenylbutyrate as novel, nontoxic differentiation inducers. Adv Exp Med Biol 400A:501–505.
Shiiyama S, Soejima-Ohkuma T, Honda S, Kumagai Y, Cho AK, Yamada H, Oguri K, and
Yoshimura H (1997) Major role of the CYP2C isozymes in deamination of amphetamine and
NEW PHENYLBUTYRATE METABOLITES
benzphetamine: evidence for the quinidine-specific inhibition of the reactions catalysed by
rabbit enzyme. Xenobiotica 27:379–387.
Wang Z, Zhao C, Pierce ME, and Fortunak JM (1999) Enantioselective synthesis of beta-hydroxy
carboxylic acids: direct conversion of beta-oxocarboxylic acids to enantiomerically enriched betahydroxy
carboxylic acids via neighboring control group. Tetrahedron Asymmetry 10:225–228.
Webster LT, Siddiqui UA, Lucas SV, Strong JM, and Mieyal JJ (1976) Identification of separate
acyl-CoA:glycine and acyl-CoA:L-glutamine N-acyltransferase activities in mitochondrial
fractions from liver of rhesus monkey and man. J Biol Chem 251:3352–3358.
Yang D, Previs SF, Fernandez CA, Dugelay S, Soloviev MV, Hazey JW, Agarwal KC, Levine
WC, David F, Rinaldo P, et al. (1996) Noninvasive probing of citric acid cycle intermediates
in primate liver with phenylacetylglutamine. Am J Physiol 270:E882–E889.
Zeitlin PL, Diener-West M, Rubenstein RC, Boyle MP, Lee CK, and Brass-Ernst L (2002)
Evidence of CFTR function in cystic fibrosis after systemic administration of 4-phenylbutyrate.
Mol Ther 6:119–126.
19
Downloaded from
dmd.aspetjournals.org by guest on December 6, 2012