Reduction in glucose levels in STZ diabetic rats by 4-(2,2-Dimethyl-1-oxopropyl)benzoic acid: A prodrug approach for targeting the liver

Article abstract of DOI:10.1021/jm000264w

The overproduction of glucose by the liver in NIDDM patients markedly contributes to their fasting hyperglycemia and is a direct consequence of the increased oxidation of excess free fatty acids (FFA) being released from the adipocyte. 2-(1,1-Dimethylethyl)-2-(4-methylphenyl)-[1,3]dioxolane (SAH51-641, 1)has previously been demonstrated to reduce glucose levels in animal models of diabetes by reducing fatty acid oxidation and hence depriving the system of the energy and cofactors necessary for gluconeogenesis. However, attempts at lowering glucose levels in vivo with 1 have been associated with toxicity in other organs such as the testes. An approach was developed utilizing the natural processing of triglyceride-like intermediates as a basis for selectively targeting the absorption, processing, and delivery of a prodrug to the liver. Compounds were identified by this method which lowered glucose levels in vivo without releasing toxic amounts of the active metabolites of 1 into circulation.

Full text of DOI:10.1021/jm000264w

512
J . Med. Chem. 2001, 44, 512-523
Articles
Red u ction in Glu cose Levels in STZ Dia betic Ra ts by
4-(2,2-Dim eth yl-1-oxop r op yl)ben zoic Acid : A P r od r u g Ap p r oa ch for
Ta r getin g th e Liver
Gregory R. Bebernitz,*,† J eremy G. Dain,‡ Rhonda O. Deems,§ Dario A. Otero,† W. Ronald Simpson,† and
Robert J . Strohschein†
Metabolic and Cardiovascular Diseases, Novartis Institute for Biomedical Research, Novartis Pharmaceuticals Corporation,
556 Morris Avenue, Summit, New J ersey 07901
Received J une 20, 2000
The overproduction of glucose by the liver in NIDDM patients markedly contributes to their
fasting hyperglycemia and is a direct consequence of the increased oxidation of excess free
fatty acids (FFA) being released from the adipocyte. 2-(1,1-Dimethylethyl)-2-(4-methylphenyl)-
[1,3]dioxolane (SAH51-641, 1) has previously been demonstrated to reduce glucose levels in
animal models of diabetes by reducing fatty acid oxidation and hence depriving the system of
the energy and cofactors necessary for gluconeogenesis. However, attempts at lowering glucose
levels in vivo with 1 have been associated with toxicity in other organs such as the testes. An
approach was developed utilizing the natural processing of triglyceride-like intermediates as
a basis for selectively targeting the absorption, processing, and delivery of a prodrug to the
liver. Compounds were identified by this method which lowered glucose levels in vivo without
releasing toxic amounts of the active metabolites of 1 into circulation.
In tr od u ction
of normoglycemia has been established for both IDDM
and NIDDM patients. The initial therapy for newly
diagnosed NIDDM patients has conventionally been diet
and exercise. Although this leads to a marginal im-
provement in insulin sensitivity and a corresponding
reduction in hyperglycemia, this treatment usually fails
within a few months due to both a further progression
of the disease as well as a problem with compliance and
an oral antidiabetic agent is prescribed. The standard
regimen has consisted of treatment with a member of
the class of sulfonylurea drugs, which reduce glycemia
by inducing the â-cell to release more insulin. However,
undesired consequences of prolonged use of sulfonyl-
ureas include hypoglycemic episodes and ultimate ex-
haustion of the â-cell, as well as the long-term angio-
genic side effects which are a result of chronic 24-h
exposure to increased insulin levels.3-5 Agents which
reduce the glycemia without these above side effects are
highly desirable and have been the aim of many
research groups over the past 20 years.
Diabetes mellitus may be categorized into two sub-
classes: type I, known as insulin-dependent diabetes
mellitus (IDDM); and type II, non-insulin-dependent
diabetes mellitus (NIDDM). IDDM accounts for about
10% of all diabetes and results from autoimmune-
mediated destruction of insulin-secreting â-cells of the
pancreas. In contrast, NIDDM is a chronic and progres-
sive metabolic disorder of carbohydrate and lipid me-
tabolism and accounts for the remaining 90% of diabetes
mellitus.1 It is the fourth leading cause of death in
developed countries and affects more than 5% of the
world’s population (and one in four people over the age
of 60). Fewer than half of all diabetics receive treatment,
and of these only a very small proportion achieve a level
of glucose control which is sufficient to avoid the
morbidity associated with the disease, namely macro-
vascular (coronary artery disease, stroke) and microvas-
cular (retinopathy, neuropathy, nephropathy, other
microangiopathies) complications. In the 10-year Dia-
betes Control and Complication Trial (DCCT)2 study it
was found that tight control of blood glucose levels
reduced the incidence and progression of neuropathy,
retinopathy, and nephropathy, each by more than 50%
in IDDM patients. Because therapeutic improvements
were continuously related to incremental reductions in
glycemia (toward normal levels), a new treatment goal
Ap p r oa ch
NIDDM is characterized by excessive hepatic glucose
production, impaired glucose disposal in the periphery,
impaired lipid metabolism, and an inability of the â-cell
to secrete sufficient insulin in a timely manner in order
to achieve normalization of glucose and lipid homeosta-
sis. The overproduction of glucose by the liver signifi-
cantly contributes to fasting hyperglycemia in these
individuals and is a direct result of the increased
oxidation of the excess circulating free fatty acids (FFAs)
being released from the adipocyte.6-10 Clinical studies
* To whom correspondence should be addressed. Phone: 908-277-
7089. Fax: 908-277-2405. E-mail: greg.bebernitz@pharma.novartis.com.
† Department of Medicinal Chemistry.
‡ Department of Pharmacokinetics and Drug Metabolism.
§ Department of Pharmacology.
10.1021/jm000264w CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/29/2000

Prodrug Approach for Targeting the Liver
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 513
Sch em e 1. Metabolic Activation of 1
F igu r e 1. Mechanism of CoA sequestration.
have shown that decreasing the supply of fatty acids or
inhibiting fatty acid oxidation (FAO) per se results in a
reduction in fatty-acid-driven gluconeogenesis.11,12 One
method which has been found useful in limiting FAO
has been through control of the transport of fatty acids
into the mitochondria where FAO takes place. Whereas
octanoate passively diffuses across the inner mitochon-
drial membrane, long-chain fatty acids (LCFA), which
are by far the major source of energy stored as fat,
require active transport via the carnitine palmitoyl
transferase (CPT) system, which includes CPT1, a
translocase, and CPT2. Administration of 2-(1,1-di-
methylethyl)-2-(4-methylphenyl)[1,3]dioxolane (1) has
been previously shown to reduce â-oxidation (FAO) in
normal fasted animals by actively sequestering the
intramitochondrial CoA required for the transport of
fatty acids into the mitochondria. 1 has also resulted
in reduced glucose levels when administered to diabetic
animal models.13 It has been shown that 1 is converted
by means of metabolic oxidation to 2 followed by
subsequent reduction into its active form, 3 (Scheme 1).
3 then acts as a substrate for the medium-chain fatty
acyl CoA ligase with the result of interfering in the FAO
process by sequestering the available CoA. However,
sequestration of CoA, which results in the inhibition of
FAO and consequently glucose production in the liver,
has proven detrimental to tissues other than the liver,
which also depend on the processing of acyl CoA
intermediates, and has produced undesired toxicity. If
the active CoA-sequestering agent 3, derived from 1 by
metabolic transformation, could be localized to the liver
and prevented from reaching general circulation, this
should help to eliminate the observed peripheral tox-
icities (Figure 1). A prodrug approach was employed to
provide a compound which was absorbed, sequestered
by the liver, converted by the liver into the active agent
where its effects could be manifested, and then ulti-
mately conjugated and cleared by the liver thereby
avoiding undesired concentrations of the active agent
in general circulation. This strategy involves protection
of the acid functionality in such a way that hepatic
cleavage by either nonspecific esterases or P-450 en-
zymes is necessary for the biotransformation of the
prodrug to 3. Both methods require that the agent: (1)
not be cleaved in the intestine, the enterocyte, or the
blood stream; (2) be taken up by the liver preferentially;
(3) be cleaved in the liver to the active form at a rate
sufficient to inhibit FAO; and (4) not escape the liver
into the blood stream once activated but be eliminated
preferentially through the bile. Glycerol ester prodrugs
have been previously utilized to enhance absorption and
stability of various biological agents as well as provide
selective delivery of these agents to the liver14-19 (see
Discussion section for further explanation). This tech-
nique was applied in this setting using various polyol
esters and ethers of ketone 2 to modulate the targeting
and release of the active pharmacophore 3.
Ch em istr y
The key carboxylic acid 2 was prepared as previously
described and converted to its acid chloride 4 using
thionyl chloride.20 Bromide 5 was prepared by treating
2,2,4′-trimethylpropiophenone with bromine.21 Solketal
(6), 1,3-benzylidene glycerol (7), pentaerythritol (13),
pivaloyl chloride, and oleic acid were available com-
mercially as were the alcohols for the preparation of
compounds 18-24. Triethyl 1,1,2-ethanetricarboxylate
was reduced with LAH using standard conditions to
provide the triol used in the preparation of 12a .
The general strategy used for the preparation of
differentially substituted glycerol adducts 9a -g utilized
protected glycerols as exemplified by 6 and 7. Acylation
under standard conditions followed by deprotection of
the glycerol fragment provided cleanly 1- or 2-acylated
glycerols. If desired, the subsequent diol may be further
derivatized with a fatty acid (i.e. palmitic acid) to
generate pre-formed triglyceride analogues. As glycerols
acylate preferentially on the 1-position, a more lengthy
procedure was required to generate the 2,3-differentially
disubstituted glyceride 9c involving silylation of the
primary position of 9a followed by acylation of the
2-position and subsequent desilylation. In addition as
2-acylated glycerols have a tendency to migrate to the
primary 1- or 3-position of glycerol, 9c was unstable and
converted slowly to the 1,3-disubstituted adduct. 9g was
prepared in a manner similar to 9f except with substi-
tution of 4-(2,2-dimethyl-1-oxopropyl)benzenepropanoyl
chloride in the initial acylation of 7. Complete acylation
of glycerol (9) or other polyols (12 or 13) using additional
equivalents of 4 provides the corresponding polyacylated

514 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Sch em e 2. Preparation of Acylated Analoguesa
Bebernitz et al.
a
Conditions: (a) 4, Et3N, THF; (b) 0.5 N HCl, THF, water, CH3(CH2)14COCl, Et3N, THF; (c) TBDPSCl, imidazole, DMF, CH3(CH2)14COCl,
Et3N, THF, n-Bu4NF, THF, AcOH; (d) Na, methanol (octanol); (e) CH3(CH2)14COCl, Et3N, THF.
Sch em e 3. Preparation of Ether Analoguesa
a
Conditions: (a) 5, NaH, THF; (b) 0.5 N HCl, THF, water, CH3(CH2)14COCl, Et3N, THF, CH3(CH2)7CHdCH(CH2)7COCl, Et3N, THF;
(c) 0.5 N HCl, THF, water, CH3(CH2)14COCl, Et3N, THF.
products. The three ketones of 9h were reduced to the
corresponding alcohols with NaBH4 to provide 9i. Reac-
tion of epichlorohydrin with sodium alcoholates resulted
in glycerol analogues 11a ,b with the 1- and 3-positions
of glycerol converted to methyl or octyl ether adducts,
respectively. Subsequent acylation under normal condi-
tions (method A) provides 9j,k (Scheme 2 and Table 1).
Alkylation of 6, 7, or 15a with bromide 5 using NaH
provided the substituted glycerol adducts. If required,
removal of the protecting groups followed by acylation
was performed in a similar manner as for the acylated
adducts (vide supra) to provide 15a -e. Alkylation of
other alcohols and diols using NaH provided 18-24.
(See Scheme 3 and Table 2.)
â-oxidation of their CoA esters (Figure 1). Whereas
octanoate passively diffuses across the inner mitochon-
drial membrane, LCFAs, which are by far the major
source of energy stored as fat, require active transport
via the CPT system, which includes CPT1, a translocase,
and CPT2. Once inside the mitochondria, FFAs are then
oxidized to produce ATP, acetyl-CoA, and NADH which
are necessary cofactors for the process of gluconeogen-
esis. 1 had previously been demonstrated a potent
hypoglycemic agent13 whose effects are manifested
through inhibition of the oxidation of LCFAs, thereby
reducing gluconeogenesis. This requires activation of 1
by a suspected P-450 process to provide 2 followed
subsequently by the reaction of a reductase to generate
3 as the major CoA-sequestering agent (Scheme 1). Both
2 and 3 are postulated to enter the mitochondria where
they are converted into their respective CoA esters, via
the medium-chain CoA ligase. After CoA ester forma-
Resu lts a n d Discu ssion
The process of FAO requires the movement of FFAs
across the mitochondrial membrane with subsequent

Prodrug Approach for Targeting the Liver
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 515
Ta ble 1. Biological Data for Ester Prodrugs
a
b
Acute screen: animals dosed at 100 µmol/kg, data reported as percent of control at 3 h postdose. Chronic screen: animals dosed at
70 µmol/kg in STZ-treated diabetic rats, data reported as percent efficacy (where 100% efficacy is lowering of blood glucose levels to
normal levels - 60 mg/dL). c Area under the curve of active metabolite 3 in blood plasma after dosing orally at 300 µmol/kg (see Experimental
Section for description of in vivo models). *p < 0.01, **p < 0.001. #ED50 values are reported in mg/kg/day.
tion, the second step in the liver, under normal condi-
tions, would be the xenobiotic elimination of these
foreign aromatic acids through formation of their cor-
responding glucuronidate or hippurate from the corre-
sponding CoA ester. As this second step is rate-limiting
for aromatic acids such as 3, formation of their CoA
esters effectively sequesters the intramitochondrial
supply of CoA. Lower levels of intramitochondrial CoA
would then have a direct effect on CPT2 thereby limiting
the respective transport of LCFA into the mitochondria.
In addition, lower levels of intramitochondrial CoA
would also have a direct effect on the long-chain acyl
dehydrogenase (LCAD) thereby limiting the oxidation
of LCFA within the mitochondria. Diminished FAO
results in lower acetyl CoA levels and a reduced supply
of ATP and NADH required for enzymes in the pathway
of gluconeogenesis, such as pyruvate carboxylase, the
first step of gluconeogenesis. The immediate outcome
would be reduced liver glucose synthesis and thus
reduced serum glucose levels.
FAO is an increase in circulating ketone levels (â-
hydroxybutyrate, â-HBA) as a consequence of the oxida-
tion of LCFAs. CoA-sequestering agents, such as 3, have
been shown to acutely reduce the levels of â-HBA in
these animals and consequently have also reduced
glucose levels. It has been observed in normal fasted
rats, by either reducing the dose of certain analogues
or by selecting analogues with reduced ability to be
converted to the active entity 3, that acute reductions
in â-HBA levels (i.e. reduction in FAO) could be ac-
complished with only marginal reductions in glucose
levels. Although the pharmacological effect on glucose
levels would be expected to lag behind the primary effect
on FAO through CoA sequestration, this delay was
considered constant for our purposes based on achieved
concentrations of active agent. Consequently, any fur-
ther separation between the effects on â-HBA levels and
glucose levels may be due to other pharmacokinetic
effects, such as absorption rate and route (portal versus
lymphatic) or metabolic activation rate of the prodrug
to 3. This separation of effects on â-HBA versus glucose
levels was examined as a potential marker for those
agents most likely to be activated at a rate sufficient
for their effects to be manifested in the liver without
being activated too rapidly so as to elicit subsequent
effects in other tissues. It should be kept in mind,
In normal 18-h fasted rats, glycogen levels are ex-
hausted and gluconeogenesis maintains blood glucose
levels in the normal range. Fatty acids are released from
the adipocyte, and their subsequent metabolism pro-
vides the energy and cofactors necessary for glucose
production in the liver. Concurrent with an increase in

516 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Ta ble 2. Biological Data for Ether Prodrugs
Bebernitz et al.
a
b
Acute screen: animals dosed at 100 µmol/kg, data reported as percent of control at 3 h postdose. Chronic screen: animals dosed at
70 µmol/kg in STZ-treated diabetic rats, data reported as percent efficacy (where 100% efficacy is lowering of blood glucose levels to
normal levels - 60 mg/dL). c Area under the curve of active metabolite 3 in blood plasma after dosing orally at 300 µmol/kg (see Experimental
Section for description of in vivo models). *p < 0.01, **p < 0.001. #ED50 values are reported in mg/kg/day.
Ta ble 3. Comparison of Organ and Body Weights to AUC
however, that in normal animals, counter-regulatory
Values
measures would also prevent hypoglycemia thus limit-
ing the maximum glucose lowering that one could
achieve.
c
compd body wta heart wtb testes wtb liver wtb AUC0-48h
1
76%**
95%
97%
92%
46%*
102%
100%
215%*
113%*
101%
1513
128
0
9h
12a
The testes have proven to be the primary organ of
101%
100%
toxicity for this class of CoA-sequestering agents. Evi-
dence from the literature has demonstrated that agents
such as p-tert-butylbenzoic acid,22-25 a known CoA-
sequestering agent, reduce aspermatogenesis in the
testes. Whereas 1 is effective at inhibiting FAO in the
liver, it was also found to escape the liver after conver-
sion into 3 and thereby also result in undesired effects
on other tissues including the testes. It was hypoth-
esized that localizing the active CoA-sequestering agent
to the liver, the desired organ of pharmacological action,
should help reduce the observed peripheral toxicities.
The goal of this research was therefore to generate the
active CoA-sequestrating agent 3 only in the liver where
its effects would be desired and at such a rate that its
effects in the liver would be complete without release
of the agent into general circulation.
a
Body weights are presented as percent of control values.
b Organ weights were calculated as follows: [organ weight (treated)/
body weight (treated)]/[organ weight (control)/body weight (con-
trol)]. c Area under the curve of active metabolite 3 in blood plasma
after dosing orally at 300 µmol/kg (see Experimental Section for
description of in vivo models). *p < 0.05, **p < 0.01. Compounds
9h , 12a , and 1 were evaluated in a 28-day safety study in normal
Sprague-Dawley rats at 380, 550, and 700 µmol/kg/day, respec-
tively (10× ED50 for glucose lowering in STZ-treated diabetic
animals).
sired. An interesting observation was noted between the
incidence and severity of testicular toxicity in these 28-
day safety studies and the corresponding blood levels
of the active agent 3, following a single oral dose. This
is illustrated for compounds 1, 9h , and 12a in Table 3.
It is clear that 1, which releases active pharmacophore
3 into circulation as confirmed by its high AUC0-48h
value, has altered body, testes, and liver weights,
whereas 9h and 12a had values which were much closer
to control. This finding allowed the AUC0-48h (area
under the curve) of 3 in blood (collected at seven time
points following a single oral dose) to serve as a
preliminary assessment of toxicity prior to a full 28-
Early on in the development process, 28-day safety
studies were employed to demonstrate the testicular
toxicity associated with these agents. As it was impos-
sible to perform full 28-day safety studies on each
candidate, an alternative method for the rapid identi-
fication of potentially undesirable compounds was de-

Prodrug Approach for Targeting the Liver
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 517
day safety study. More importantly, the AUC0-48h was
used as a guide for directing the focus of the chemistry
efforts before extensive pharmacology was practical.
AUC0-48h values of less than 500 µg‚h/mL have not
shown toxicity to the testes in comparable 28-day safety
studies.26
effect in the chronic diabetic model without eliciting
peripheral toxicities and were further evaluated for
AUC0-48h. Acylation of the remaining sites on glycerol
with palmitate or another LCFA should allow the
compounds to more closely resemble true triglycerides
and so be absorbed and processed as triglycerides. Each
of the acylated triglyceride analogues (9a -h ) was active
in the chronic STZ diabetic rat model at 6 h after oral
dosing with values ranging from 42% to 100% efficacy
(100% efficacy being indicative of a return of glucose
levels to that of a normal nondiabetic rat). Based on the
normal processing of triglycerides by pancreatic-gut
lipases, substitution at the 2-position of glycerol by a
specific acyl group should allow passage through the
intestinal wall without hydrolysis from its glycerol
carrier. Therefore, placement of the masked pharma-
cophore at the 2-position as in compounds 9e,f (palmi-
toylated 9e) should provide a better selectivity profile
without exhibiting higher AUC0-48h values than 1- and/
or 3-acylated glycerol analogues. However, as can be
inferred from the AUC0-48h data, both proforms (9a ,e)
produced 3 at levels that would be predictive of toxicity.
It was noteworthy that 9h , a triacylated glycerol
analogue, would have been predicted to be toxic, based
on normal processing of triglycerides resulting in cleav-
age of the acyl group from the 1- or 3- position of glycerol
during processing and absorption from the gut. However
9h produced a very low AUC0-48h of 128 µg‚h/mL
predictive of no toxicity. As the abilities of the gut
lipases to cleave benzoyl esters are poor, 9h likely passes
through the gut intact and proceeds to the liver before
being unmasked to the active agents. Indeed it has been
shown that medium-chain triglycerides, which contain
C6-C12 fatty acids and are similar in size to 9h , can
be absorbed from the gut intact and are not partitioned
into the lipoprotein fraction but proceed directly to the
liver via the portal vein.28 If 9h had been cleaved in the
gut or enterocyte to any significant rate, then 3 would
have entered general circulation and resulted in toxicity.
The search for a development candidate required a
very specific selection profile in terms of rates of
absorption, delivery, metabolism, and clearance. The
desired absorption should preferably be over an ex-
tended period of time so that the peak concentration of
the agent in any given tissue or fluid, such as intestine,
blood stream, and even the liver would not be more than
the liver’s ability to sequester, unmask, and be inhibited
by the active agent. An additional aid in controlling the
rate of delivery would be a design which promoted
lymphatic transport, a mechanism which would slow the
rate of delivery to the blood stream thereby avoiding
exceeding the capacity of the liver to dispose of the agent
as well as providing a longer ultimate duration of action.
However the delivery of drugs through the lymphatic
system may be neither timely nor sufficient to produce
the desired biological effect. As the primary site of action
will be the liver, portal delivery and first-pass uptake
by the liver is in this approach a desirable event. The
physical characteristics of the prodrug should be de-
signed such that its selective uptake by the liver would
be nearly complete. The unmasking of the prodrug to
the active agent in the liver should be at such a rate as
not to overload the liver’s capacity to dispose of the
agent and allow metabolites to leak into circulation, but
also not so slow that the agent is eliminated before the
active portion of the prodrug can be unmasked and find
its target. The ultimate elimination of the agent should
preferably be through the bile so as to avoid a second
pass through the blood stream, and the method of
elimination should be irreversible so that enterohepatic
recirculation does not occur. Whereas one can design
certain of these characteristics into potential drug
candidates, other factors are more tenuous as the
following results, the ultimate effect in animal models,
clearly support.
Ethers 9j,k were effective in the 18-h fasted normal
rat model at lowering both â-HBA as well as glucose
levels. However, profound differences were apparent in
the chronic STZ diabetic rat model where the dimethyl
ether 9j was efficacious while the dioctyl ether 9k
exhibited no activity. From cursory examination, one
might argue that some physicochemical parameter is
responsible for the difference as 9k (cLog P ) 9.7) is
much more lipophilic than 9j (cLog P ) 2.3) and may
require metabolic cleavage or activation in order to be
absorbed. This, however does not explain why both 9j,k
were active in the 18-h fasted normal rat model,
suggesting similar bioavailability and processing, while
9k was not active in the chronic STZ diabetic rat model.
A more reasonable explanation would be that the
chronic STZ diabetic model is different perhaps in one
or more pharmacokinetic parameters such as absorption
or metabolic activation, and this leads to the lack of
activity with 9k in this model. This will be discussed
further below.
The compounds described herein were all evaluated
first for their ability to inhibit oleate-dependent gluco-
neogenesis in freshly prepared hepatocytes from 18-h
fasted rats.13,27 In general, most compounds significantly
decreased glucose production with IC50 values of less
than 100 µM (data not shown). However it was found
that due to the particular nature of each prodrug’s
environment, even compounds which exhibited no activ-
ity in hepatocytes could provide significant activity
when placed in an in vivo setting.
Acyla t ed An a logu es: Each of the mono-, di-, and
triglyceride equivalents of 3 (9a -h ) significantly de-
creased â-HBA levels to less than 25% of control levels
in a 18-h fasted normal rat model indicating pronounced
inhibition of â-oxidation (FAO) (Table 1). Glucose levels
were also reduced to only 56-82% of control in this
model as expected and consistent with the effects on
â-HBA preceding and being more pronounced than the
corresponding effects on glucose levels. Agents with a
profile of significantly reducing â-HBA while having
only modest effects on glucose levels were selected as
the most promising to elicit the desired glucose-lowering
Whereas the activity of the various intestinal lipases
on glycerol analogues is well-documented and quite
efficient, even modest changes to the glycerol backbone
produce structures which are remarkably resistant to
the activity of these lipases. Thus, whereas esters of

518 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Bebernitz et al.
methanol, ethylene glycol, and glycerol are cleaved by
pancreatic lipase at 63, 200, and 1900 µmol FFA
released/min/mg protein, respectively, hydrolysis of
erythritol, adonitol, and other higher sugars does not
occur at measurable rates.29 In addition, as it had
become apparent that even highly lipophilic structures
such as 9h were able to be absorbed intact without
cleavage in the gut and be sequestered by the liver as
evident from a significant effect on â-HBA in comparison
to a low AUC0-48h value, other backbones were inves-
tigated for their potential as prodrug carriers which
would be resistant to hydrolysis. For two such com-
pounds, 12a and 13a , sequential derivatizable alcohol
units were added. Interestingly, 12a was an effective
FAO inhibitor in both normal (lowering â-HBA) and
diabetic animals and exhibited an undetectable AUC0-48h
for the active species 3. A 28-day safety study performed
in normal rats with 12a administered at 10× the ED50
(for glucose lowering in STZ diabetic rats) produced no
detectable histological effects in the testes thereby
supporting the predictability of the AUC0-48h data.24
This would imply that prodrug 12a was absorbed intact,
delivered to the liver, unmasked to 3, subsequently
inhibited FAO in the liver, and ultimately cleared.
Structure 12a therefore fulfilled the objectives of the
program. A tetraacyl isomer, 13a , on the other hand,
showed no activity in vitro (data not shown) or in vivo.
Since 13b (the diacyl isomer of 13a ) produced significant
activity in all assays, this suggests that 13a apparently
is neither absorbed intact nor activated to any signifi-
cant extent by pancreatic-gut lipases which would allow
for its absorption.
suggests the liver and intestinal esterases and lipases
responsible for generating active pharmacophore 3 are
functioning in similar fashions in each model. However,
the lack of glucose-lowering activity of the ether glycerol
analogues in the chronic model, while at the same time
showing significant lowering of â-HBA in the acute
model, suggests the enzymes or factors available to
unmask the active pharmacophore in the normal animal
are evidently not available in the STZ diabetic animal.
It has already been shown that 1 is readily converted
to 2 and 3 in both normal as well as STZ diabetic
animals indicating that the oxidases (P-450) and reduc-
tases necessary for these conversions are available in
both models. There have been some reports with respect
to differences in the amounts of various P-450 isozymes31
which are available for metabolic transformation in
diseased animals such as the STZ rat. This may account
for lower levels of 3 or possibly different metabolic
products altogether. Whereas it was satisfying that the
ether glycerol prodrugs provided significant amounts of
active agent 3 and subsequently pronounced activity in
the 18-h fasted normal rat model, this series has thus
far proven disappointing in terms of their activity in
the chronic STZ diabetic rat model.
The ether linkage was also utilized in a number of
other alcohol and diol units to determine whether they
could also be prodrug carriers (18-24). Each of these
compounds reduced â-HBA levels in normal animals
with minor to modest reductions in glucose levels. When
utilizing the sulfide linker as in 21 and 22, one which
is generally considered metabolically unstable, it was
interesting that no reduction of â-HBA was observed.
It is apparent that oxidation of sulfur must generate
intermediates (perhaps sulfoxides) which are not as
readily converted to 3.
Eth er An a logu es: As ether analogues are unlikely
to be cleaved by lipases or esterases in the gut, entero-
cyte, or serum, they possess the advantage of being
delivered to the liver prior to any unfortuitous metabolic
activation, given that the enterocyte possesses at best
minor P-450-metabolizing activity. The position on the
glycerol backbone should not be critical in this case for
ultimate delivery to the liver. The use of the glycerol/
triglyceride unit could, in addition, capitalize on slower
delivery via the lymphatic system and minimize the rate
of release of 3 into circulation. Many of the compounds
(see Table 2) produced marked reductions in â-HBA
levels in the 18-h fasted normal rat model as had been
observed before with the acylated analogues. There were
only modest changes in glucose levels in this model,
which was initially perceived to be in line with the
desire for agents with slower onset of action. The lack
of any significant effect of the ether analogues in the
chronic STZ diabetic rat model was, however, puzzling.
As most compounds provided a significant effect on
â-HBA levels in the 18-h fasted normal rat model, this
would indicate they are absorbed and unmasked to the
active agent 3. Bioavailability should be comparable to
or even better in the chronic versus the normal rat
models as lipophilic agents have routinely shown en-
hanced bioavailability in models where ad libitum food
is present30 (as in the chronic STZ diabetic rat model).
This would suggest that in the chronic STZ diabetic rat
model something is quite different. For the acylated
prodrugs, the reduction in glucose in the chronic model
correlates well with the reduction in â-HBA and glucose
in the normal 18-h fasted normal rat model. This
Con clu sion
It has been shown that CoA sequestration by these
agents provides reductions in â-HBA levels as a marker
for the reduction in FAO. Reduction in FAO results in
a corresponding reduction in fatty acid-driven glucone-
genesis and ultimately glucose levels. In the 18-h fasted
normal rat model many compounds produced a signifi-
cant effect on â-HBA levels without showing dramatic
effects on glucose production. This profile was desired
in our attempt to maximize delivery and activity to the
liver target while minimizing peripheral circulation of
3 and its corresponding toxicity. In the case of the STZ
diabetic rat model, the portion of glucose production
driven by the high levels of circulating FFAs is a
significant contributor to the undesired hyperglycemia.
This overproduction of glucose is reduced upon treat-
ment with these agents, while the basal gluconeogenesis
necessary to sustain basal glucose levels should be
unaffected. The prodrug stategies employed herein
provided two agents, 9g and 12a , which inhibited FAO
and reduced glucose levels without inducing toxicity via
release of active CoA-sequestering agents into circula-
tion. Agents employing the ether linkage however
provided attachment of 3 which was apparently too
resistant to activation by the liver of the STZ diabetic
animals to provide meaningful amounts of active CoA-
sequestering agent 3. However, it should be recognized
that the ability of the liver to effectively mobilize a

Prodrug Approach for Targeting the Liver
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 519
Plasma samples from each group were pooled to provide
sufficient sample for analysis by HPLC. Animals were allowed
ad libitum access to food and water during the study. Results
are presented in µg‚h/mL. Compounds with values <500 µg‚
h/mL are considered nontoxic in terms of effects on the testes
as verified with historical 28-day safety studies.
Gen er a l. Melting points were determined on a Thomas-
Hoover capillary melting point apparatus and are uncorrected.
1H and 13C NMR were recorded on a Bruker AC 300 MHz NMR
or a J EOL 200 MHz NMR using tetramethylsilane (TMS) as
an internal standard and are reported in ppm (δ). Mass spectra
were performed on a Finnigan Mat 4600 spectrometer. El-
emental analysis was performed with a Carlo Erba CHNS-O
EA 1108 elemental analyzer. Data are within 0.4% of theoreti-
cal values unless otherwise indicated. All compounds were
routinely checked by TLC with Macherey-Nagel Polygram SIL
G/UV254 plates. Yields are of purified products and were not
optimized.
Meth od A: Gen er a l P r oced u r e for Acyla tion of Alco-
h ols Usin g Acid Ch lor id e 4. To the starting alcohol (0.10
mol) and DMAP or Et3N (0.34 mol) in anhydrous THF or
methylene chloride (800 mL) at 0 °C was added acid chloride
4 (0.11 mol equiv for each alcohol unit) portionwise over 30
min. The reaction was allowed to warm to room temperature
and stirred for 15 h. The crude reaction mixture was evapo-
rated and then partitioned between CH2Cl2(100 mL) and water
(50 mL). Subsequent extraction with 2 portions of CH2Cl2 (100
mL) followed by drying the combined organic layer over
MgSO4, filtering and evaporating resulted in crude product.
Compounds were purified by chromatography.
F or Com p ou n d s su ch a s 8 or 10 Th a t Ha ve a n Acid -
Sen sitive P r otectin g Gr ou p . The crude material was depro-
tected using 0.5 N HCl in THF at room temperature for 3 h.
Ether (100 mL) was added and the mixture partitioned
between ether and water. The aqueous layer was extracted
with ether (2 × 50 mL) and the combined organic layer was
washed with aqueous NaHCO3 (satd), brine, dried over MgSO4
and concentrated to afford crude product. Chromatography
over silica gel afforded pure product.
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2,3-d ih y-
d r oxyp r op yl ester (9a ): acylation of 6 by method A -
chromatographed on silica gel (CHCl3 to 6% MeOH/CHCl3) to
afford 9a in 72% yield; mp 74-75.5 °C; 1H NMR (CDCl3) δ
1.32 (s, 9H), 2.62 (bs, 1H), 3.07 (bs, 1H), 3.75 (bm, 2H), 4.09
(bs, 1H), 4.43 (m, 2H), 7.64 (d, J ) 8.5 Hz, 2H), 8.06 (d, J )
8.5 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 281 (100), 298
(36); 13C NMR (CDCl3) δ 27.7, 44.4, 63.5, 66.0, 70.3, 127.4,
129.5, 131.2, 143.4, 166.2, 209.6. Anal. (C15H20O5) C, H.
Ben zoic acid, 4-(2,2-dim eth yl-1-oxopr opyl)-, 2-h ydr oxy-
1-(h yd r oxym eth yl)eth yl ester (9e): acylation of 7 by method
A - chromatography on alumina (neutral CHCl3 followed by
EtOAc) provided the title compound which was crystallized
from ether to give white crystals (77%); mp 65-67 °C; 1H NMR
(CDCl3) δ 1.31 (s, 9H), 2.74 (bs, 2H), 3.95 (d, J ) 4.7 Hz, 4H),
5.18 (m, 1H), 7.63 (d, J ) 8.2 Hz, 2H), 8.07 (d, J ) 8.2 Hz,
2H); MS (DCI, NH3) m/z (rel intensity) 298 (100); 13C NMR
(CDCl3) δ 27.6, 44.4, 62.3, 76.0, 127.3, 129.6, 131.3, 143.4,
165.9, 209.8. Anal. (C15H20O5) C, H.
Meth od B: Gen er a l P r oced u r e for Acyla tion of Glyc-
er ol An a logu es w ith Lon g-Ch a in Acid Ch lor id es. The
glycerol analogue (7.72 mmol) was dissolved in anhydrous CH2-
Cl2 (75 mL) and DMAP (1.1 g, 9.0 mmol) was added followed
by palmitoyl chloride (2.6 mL, 8.5 mmol per alcohol unit). After
stirring for 1 h the reaction was partitioned between water
(30 mL) and CH2Cl2 (75 mL). Subsequent extraction with 2
portions of CH2Cl2 (75 mL) followed by drying the combined
organic layer over MgSO4, filtration and evaporation resulted
in crude product. Chromatography on silica gel (CHCl3) and
recrystallization from ethanol if appropriate afforded the title
compound in 70-95% yield.
prodrug via cleavage and oxidation of an ethereal
linkage to a carboxylic acid was clearly demonstrated
with 10 out of 12 of these analogues significantly
reducing â-HBA levels in 18-h fasted normal rats. On
the other hand, the ester approach provided 13 out of
14 compounds which were delivered to and activated
by the liver at a rate sufficient to produce significant
reductions in â-HBA levels in 18-h fasted normal rats.
Of these analogues, 11 exhibited significant reductions
in glucose levels in the chronic STZ diabetic rat model.
Compound 12a provided significant glucose lowering in
this model (ED50 ) 39 mg/kg/day) and in conjunction
with its undetectable AUC0-48h was selected as a
development candidate. It is envisioned that the ap-
plication of ester prodrugs should be species-indepen-
dent and correlate well to humans as the complement
of intestinal lipases and their respective activities
appear similar across species.32 The application, how-
ever, of ether prodrugs and compounds which require
more extensive modification may experience some spe-
cies or model differences as observed above which would
affect the tuned prodrug activation rate demonstrated
here for the rat and so need to be further evaluated in
the clinic. The goal of defining an agent which was
activated in the liver without releasing the active
pharmacophore 3 into general circulation, reduced FAO-
driven gluconeogenesis and most importantly achieved
this without the subsequent toxic effects to the testes,
was thereby achieved.
Exp er im en ta l Section
Acu te In h ibition of F AO: 18-h F a sted Nor m a l Ra t
Mod el. The ability to acutely inhibit FAO was assessed in
normal male 18-h fasted Sprague-Dawley rats. Animals were
dosed orally (po) with vehicle (0.5% carboxymethylcellulose and
0.2% Tween-80 in water) or compound in vehicle (N ) 5
animals/group). Serum glucose and â-hydroxybutrate (â-HBA)
levels were determined from blood obtained via cardiac
puncture at 3 h postdose. Comparisons between metabolite
levels of control and treated animals were made using a
Student’s t-test and expressed as percent of control values.
Standard values for levels of metabolites in control animals
are ketones, 11.7 mg/dL; glucose, 115 mg/dL.
Ch r on ic Glu cose Low er in g in Dia betic Ra ts: Ch r on ic
STZ Dia betic Ra t Mod el. Sprague-Dawley rats fed a high
fat diet were rendered diabetic with a single intravenous
injection of streptozotocin (40 mg/kg) into the tail vein. After
5 days, animals with blood glucose levels >200 mg/dL were
fasted for 18 h and challenged with an oral glucose load
(dextrose 1.35 g/kg). Animals with blood glucose between 40
and 80 mg/dL at 3 h were used in the study. These low-dose
STZ rats on a high-fat diet have been shown to approximate
NIDDM with fed glucose levels between 175 and 260 mg/dL.
Animals were dosed orally once per day for 11 days with
vehicle (0.5% carboxymethylcellulose and 0.2% Tween-80 in
water) or compound in vehicle (N ) 9 animals/group). Blood
glucose levels were determined from blood obtained from the
tip of the tail at 0 and 6 h postdose on days 1, 4, 8, and 11.
Food was removed at 0 h and returned after the 6 h measure-
ment on each day. The results presented in Tables 1 and 2
are from 6 h postdose on days 8 and 11 and are presented as
percent efficacy where 100% efficacy is defined as normaliza-
tion of glucose to nondiabetic control levels (60 mg/dL).
Ar ea Un d er th e Cu r ve Exp osu r e Over Tim e (AUC0-48h
)
of 3. Normal Sprague-Dawley rats were dosed orally with
compound (300 µmol/kg) in vehicle (N ) 5 animals/group;
vehicle same as above). Blood samples were obtained from the
tip of the tail at 2, 4, 7, 8, 24, 36, and 48 h postdose. Blood
samples were centrifuged after collection at each time point.
Ben zen ep r op a n oic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-,
2-[(1-oxoh exa d ecyl)oxy]-1-[[(1-oxoh exa d ecyl)oxy]m eth -
yl]eth yl ester (9g): acylation of 7 using method A substitut-
ing 4 with 4-(2,2-dimethyl-1-oxopropyl)benzenepropanoyl chlo-

520 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Bebernitz et al.
ride provided a 2-acylated diol; subsequent acylation according
4.57 (m, 1H), 5.33 (m, 1H), 7.66 (d, J ) 8.6 Hz, 2H), 8.07 (d, J
) 8.6 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 536 (24), 242
(100), 291 (44); 13C NMR (CDCl3) δ 14.1, 22.7, 25.0, 27.7, 29.1,
29.2, 29.3, 29.4, 29.7, 29.9, 31.9, 34.1, 34.3, 61.4, 61.5, 62.1,
63.1, 72.1, 73.5, 127.4, 129.5, 209.3. Anal. (C31H50O6) C, H. This
material is thermally unstable and rearranges to the 1,3-diacyl
material if stored at room temperature for prolonged periods.
Ben zoic acid, 4-(2,2-dim eth yl-1-oxopr opyl)-, 2-h ydr oxy-
1,3-p r op a n ed iyl ester (9d ). To 9a (2.0 g, 7.14 mmol) in
anhydrous THF (100 mL) at 0 °C was added Et3N (1.12 mL,
8.0 mmol) followed by the acid chloride 4 (1.68 g, 7.50 mmol)
portionwise over 45 min. After stirring for an additional 2 h,
the reaction was evaporated and then partitioned between
CHCl3 (50 mL) and water (25 mL). Subsequent extraction with
2 portions of CHCl3 (50 mL) followed by drying the combined
organic layer over MgSO4, filtration and evaporation resulted
in crude 9d . Chromatography on silica gel (CHCl3) provided
pure product which was crystallized from ether/heptane to give
the title compound (12.95 g, 82%) as white crystals: mp 87-
88 °C; 1H NMR (CDCl3) δ 1.33 (s, 18H), 2.88 (bs, 1H), 4.40
(bs, 1H), 4.55 (m, 4H), 7.65 (d, J ) 8.5 Hz, 4H), 8.08 (d, J )
8.5 Hz, 4H); MS (DCI, NH3) m/z (rel intensity) 469 (32), 486
(100); 13C NMR (CDCl3) δ 27.7, 44.4, 66.0, 68.4, 127.4, 129.5,
131.1, 143.4, 165.9, 209.5. Anal. (C27H32O7) C, H.
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 1,2,3-p r o-
p a n etr iyl ester (9h ): acylation of glycerol with 3.3 equiv of
acid chloride 4 according to method A - chromatography on
silica gel (1% MeOH/CHCl3 to 2% MeOH/CHCl3) provided the
title compound which was crystallized from methanol to give
9h as white crystals (87%); mp 96-97 °C; 1H NMR (CDCl3) δ
1.32 (s, 9H), 1.33 (s, 18H), 4.76 (bm, 4H), 5.86 (bm, 1H), 7.65
(d, J ) 8.2 Hz, 6H), 8.05 (m, 6H); MS (DCI, NH3) m/z (rel
intensity) 469 (72) 486 (25); 13C NMR (CDCl3) δ 27.7, 44.4,
63.1, 70.1, 127.5, 129.5, 129.6, 130.9, 143.5, 143.6, 165.0, 165.3,
209.4. Anal. (C39H44O9) C, H.
to method B - chromatography on silica gel (CHCl3) afforded
1
9g in 92% yield; mp 44 °C; H NMR (CDCl3) δ 0.88 (m, 6H),
1.25 (bs, 48H), 1.35 (s, 9H), 1.60 (bm, 4H), 2.30 (t, J ) 7.5 Hz,
4H), 2.67 (t, J ) 7.8 Hz, 2H), 2.98 (t, J ) 7.8 Hz, 2H), 4.21
(bm, 4H), 5.26 (bm, 1H), 7.25 (d, J ) 8.2 Hz, 2H), 7.68 (d, J )
8.2 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 802 (22) 376
(100); 13C NMR (CDCl3) δ 14.1, 22.7, 24.9, 28.1, 29.1, 29.3, 29.4,
29.5, 29.6, 29.66, 29.70, 30.5, 31.9, 34.0, 35.2, 44.1, 62.0, 69.4,
127.9, 128.5, 136.4, 143.5, 171.7, 173.3, 208.3. Anal. (C49H84O8)
C, H.
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2,3-bis[(1-
oxoh exa d ecyl)oxy]p r op yl ester (9b): acylation of 9a using
method B - chromatography on silica gel (CHCl3) and recrys-
tallization from ethanol afforded the title compound (91%); mp
1
51-52 °C; H NMR (CDCl3) δ 0.88, (m, 6H), 1.25 (bs, 46H),
1.33 (s, 12H), 1.63 (m, 3H), 2.33 (m, 4H), 4.40 (bm, 4H), 5.44
(m, 1H), 7.66 (d, J ) 8.5 Hz, 2H), 8.04 (d, J ) 8.5 Hz, 2H); MS
(DCI, NH3) m/z (rel intensity) 774 (100); 13C NMR (CDCl3) δ
14.1, 22.7, 24.88, 24.93, 27.7, 29.08, 29.13, 29.3, 29.4, 29.5, 29.6,
29.67, 29.7, 31.9, 34.1, 34.2, 44.4, 62.1, 63.2, 68.8, 127.4, 129.5,
131.0, 143.3, 165.3, 172.9, 173.3, 209.4. Anal. (C47H80O7) C,
H.
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2-[(1-oxo-
h exa decyl)oxy]-1-[[(1-oxoh exa d ecyl)oxy]m eth yl]eth yl es-
ter (9f): acylation of 9e using method B - crystallization from
ethanol provided pure product as white crystals (4.13 g, 96.9%);
mp 42-43 °C; 1H NMR (CDCl3) δ 0.88 (m, 6H), 1.25 (bs, 46H),
1.33 (s, 9H), 1.60 (bm, 6H), 2.32 (t, J ) 7.5 Hz, 4H), 4.38 (m,
4H), 5.52 (bm, 1H), 7.67 (d, J ) 8.5 Hz, 2H), 8.05 (d, J ) 8.5
Hz, 2H); MS (DCI, ammonia) m/z (rel intensity) 774 (100); 13
C
NMR (CDCl3) δ 14.0, 22.6, 24.8, 27.6, 29.0, 29.1, 29.25, 29.33,
29.5, 29.55, 29.58, 31.8, 33.9, 44.3, 61.9, 70.1, 127.3, 129.4,
131.0, 143.2, 164.8, 173.2, 209.2. Anal. (C47H80O7) C, H.
Ben zoic acid, 4-(2,2-dim eth yl-1-oxopr opyl)-, 3-h ydr oxy-
2-[(1-oxoh exa d ecyl)oxy]p r op yl ester (9c). To 9a (7.0 g, 25
mmol) in 125 mL of anhydrous DMF was added imidazole (4.26
g, 62.5 mmol) followed by tert-butyldiphenylsilyl chloride (7.23
g, 26.3 mmol) and the mixture stirred for 16 h at room
temperature. The crude reaction was evaporated and then
partitioned between CHCl3 (100 mL) and water (50 mL).
Subsequent extraction with 2 portions of CHCl3 (100 mL)
followed by drying the combined organic layer over MgSO4,
filtering and evaporating resulted in crude product. Chroma-
tography on silica gel (5% EtOAc/hexane to 15% EtOAc/hex-
ane) provided ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-,
3-[(1,1-d im eth yleth yl)d ip h en ylsila n yl]oxy-2-h yd r oxyp r o-
p yl ester (14.7 g, 79%) as a clear oil: 1H NMR (CDCl3) δ 1.1
(s, 9H), 1.37 (s, 9H), 2.6 (m, 1H), 3.8 (m, 2H, 4.13 (m, 1H)m
4.45 (d, J ) 5.5 Hz, 2H), 7.4 (m, 6H), 7.7 (m, 6H), 8.05 (d, J )
8.2 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 536 (24), 242
(100), 291 (44).
The above material was acylated according to method B -
chromatography on silica gel (5% EtOAc/hexane to 10% EtOAc/
hexane) provided ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r o-
p yl)-, 3-[(1,1-d im eth yleth yl)d ip h en ylsilyl]oxy-2-[(1-oxo-
h exa d ecyl)oxy]p r op yl ester as a clear oil (94.6%): 1H NMR
(CDCl3) δ 0.88 (t, J ) 5.5 Hz, 3H), 1.07 (s, 9H), 1.25 (s, 9H),
1.3 (bs, 24H), 1.60 (m, 2H), 2.28 (m, 2H), 3.85 (m, 2H), 4.50
(dd, J ) 12.5, 6.2 Hz, 1H), 4.62 (dd, J ) 12.5, 3.1 Hz, 1H),
5.33 (m, 1H), 7.39 (m, 6H), 7.61 (m, 6H), 7.97 (d, J ) 8.1 Hz,
2H).
To the above material (18.1 mmol, 13.7 g) in anhydrous THF
(175 mL) at 0 °C was added AcOH (1.81 mL, 31.6 mmol) and
1 M tetrabutylammonium fluoride (22.6 mmol, 22.6 mL in
THF). After stirring for 2 h, the reaction was evaporated and
then partitioned between CHCl3 (150 mL) and pH 7 buffer (100
mL). Subsequent extraction with 2 portions of CHCl3 (100 mL)
followed by drying the combined organic layer over MgSO4,
filtering and evaporating resulted in crude product. Chroma-
tography on silica gel (5% EtOAc/hexane to 20% EtOAc/
hexane) provided 9c as a clear oil (8.5 g, 90.5%): 1H NMR
(CDCl3) δ 0.90 (m, 3H), 1.25 (bs, 24H), 1.32 (s, 9H), 1.63 (bm,
3H), 2.10 (bm, 1H), 2.36 (t, J ) 7.5 Hz, 2H), 3.86 (bm, 2H),
Ben zoic a cid , 4-(1-h yd r oxy-2,2-d im eth ylp r op yl)-, 1,2,3-
p r op a n etr iyl ester (9i). 9h (6.0 g, 9.14 mmol) was dissolved
in 2-propanol (100 mL) and acetic acid (100 mL) and cooled to
0 °C before the addition of NaBH4 (1.0 g, 26 mmol) portionwise.
The reaction was stirred for 2 h and evaporated to a crude
residue which was partitioned between CH2Cl2 and water,
dried and concentrated to a crude oil. Chromatography on
silica gel (CH2Cl2 to 1% MeOH/CH2Cl2) afforded 6.05 g of 9i
as an amorphous solid (71%): 1H NMR (CDCl3) δ 0.91 (s, 27H),
2.14 (bs, 3H), 4.43 (s, 3H), 4.73 (bm, 4H), 5.81 (m, 1H), 7.37
(m, 6H), 7.97 (bm, 6H); MS (DCI, NH3) m/z (rel intensity); 13
C
NMR (CDCl3) δ 25.7, 35.6, 62.9, 69.6, 81.7, 127.6, 128.26,
128.31, 128.9, 147.6, 147.8, 165.5, 165.9 Anal. (C39H50O9) C,
H.
1,3-Dim eth oxy-2-p r op a n ol (11a ) a n d 1,3-Bis(octyloxy)-
2-p r op a n ol (11b). Sodium metal (8.8 g, 0.38 mol) was added
to anhydrous methanol (200 mL) over 30 min and then stirred
for an additional 30 min. Epichlorohydrin (15.0 mL, 0.19 mol)
was added over 45 min and the reaction mixture refluxed for
2 h. The reaction was evaporated and then partitioned between
ether (100 mL) and water (50 mL). Subsequent extraction with
2 portions of ether (100 mL) followed by drying the combined
organic layer over MgSO4, filtering and evaporating resulted
in crude 1,3-dimethoxypropan-2-ol in 53% yield. The material
was used without further purification in the subsequent
reaction. Similar conditions were used to prepare 1,3-bis-
(octyloxy)propan-2-ol (2.1 equiv of octanol and sodium to
epichlorohydrin-toluene as solvent).
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2-m eth -
oxy-1-m eth oxym eth yleth yl ester (9j): acylation of 11a by
method A - chromatography on silica gel (CHCl3) afforded 9j
in 33% yield; 1H NMR (CDCl3) δ 1.33 (s, 9H), 3.40 (s, 6H), 3.69
(d, J ) 5.0 Hz, 4H), 5.41 (m, 1H), 7.65 (d, J ) 8.4 Hz, 2H),
8.09 (d, J ) 8.4 Hz, 2H); MS (DCI, NH3) m/z (rel intensity)
760 (100); 13C NMR (CDCl3) δ 27.7, 44.4, 59.3, 71.3, 72.3, 127.4,
129.6, 131.8, 143.1, 165.4, 209.5. Anal. (C17H24O5) C, H.
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2-octyl-
oxy-1-(octyloxy)m eth yleth yl ester (9k ): acylation of 11b
by method A - chromatography on silica gel (CHCl3) afforded

Prodrug Approach for Targeting the Liver
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 521
9k in 69% yield; 1H NMR (CDCl3) δ 0.88 (m, 6H), 1.25 (bs,
18H), 1.33 (s, 9H), 1.56 (bm, 6H), 3.49 (m, 4H), 3.70 (d, J )
5.1 Hz, 4H), 5.38 (bm, 1H), 7.65 (d, J ) 8.5 Hz, 2H), 8.08 (d,
J ) 8.5 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 309 (100);
13C NMR (CDCl3) δ 14.0, 22.6, 26.0, 27.6, 29.2, 29.3, 29.5, 31.7,
44.3, 69.2, 71.6, 72.7, 127.2, 129.3, 129.4, 131.9, 142.9, 165.3,
209.3. Anal. (C31H52O5) C, H.
) 8.2 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 284 (100);
13C NMR (CDCl3) δ 28.1, 28.2, 44.3, 64.2, 70.8, 72.2, 73.1, 127.2,
127.9, 128.3, 138.1, 140.8, 209.0. Anal. (C15H22O4) C, H.
1-P r op a n on e, 1-[4-[(2-h yd r oxy-1-(h yd r oxym eth yl)eth -
oxy)m eth yl]p h en yl]-2,2-d im eth yl- (17): alkylation of 7
using method C - purified on silica gel (20% MeOH/CH2Cl2
to 80% MeOH/CH2Cl2) to provide the product (82%); 1H NMR
(CD3OD) δ 1.33 (s, 9H), 3.5 (m, 1H), 3.7 (m, 4H), 4.75 (s, 2H),
7.4 (d, J ) 8.1 Hz, 2H), 7.70 (d, J ) 8.1 Hz, 2H); MS (DCI,
NH3) m/z (rel intensity) 284 (100).
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 1,2,4-bu -
ta n etr iyl ester (12a ): acylation of 2-hydroxymethyl-1,4-
butanediol33 according to method A - chromatography on silica
gel (1% MeOH/CHCl3 to 2% MeOH/CHCl3) afforded 12a in 92%
Hexa d eca n oic a cid 2-[[4-(2,2-d im eth yl-1-oxop r op yl)-
p h en yl]m eth oxy]-, 1,3-p r op a n ed iyl ester (15d ): acylation
of 17 according to method B - chromatography (silica gel/
gradient hexane to 66% hexane/ether) then provided a 74%
1
yield; mp 99-100 °C; H NMR (CDCl3) δ 1.32 (s, 27H), 2.10
(m, 2H), 2.59 (m, 1H), 4.53 (m, 6H), 7.63 (d, J ) 8.3 Hz, 2H),
7.64 (d, J ) 8.2 Hz, 4H), 8.03 (d, J ) 8.3 Hz, 2H), 8.04 (d, J )
8.2 Hz, 4H); MS (DCI, ammonia) m/z (rel intensity) 702 (100);
13C NMR (CDCl3) δ 27.7, 27.9, 35.3, 44.3, 62.8, 64.8, 127.4,
129.3, 131.3, 143.3, 165.6, 209.3. Anal. (C41H48O9) C, H.
1
yield of 15d ; mp 31-33 °C; H NMR (C6D6) δ 0.83 (t, J ) 6.4
Hz, 6H), 1.13 (s, 9H), 1.24 (s, 48H), 1.51 (m, 4H), 2.07 (t, J )
7.4 Hz, 4H), 3.55 (m, 1H), 4.09 (dd, J ) 11.7, 5.3 Hz, 2H), 4.22
(dd, J ) 11.7, 4.8 Hz, 2H), 4.32 (s, 2H), 7.15 (d, J ) 8.0 Hz,
2H), 7.63 (d, J ) 8.1 Hz, 2H); MS (DCI, NH3) m/z (rel intensity)
760 (100); 13C NMR (C6D6) δ 14.2, 23.0, 25.2, 28.0, 29.3, 29.6,
29.7, 29.8, 29.9, 30.0, 30.04, 32.3, 34.1, 43.9, 62.7, 71.2, 75.5,
127.0, 128.5, 138.0, 141.4, 172.7, 206.5. Anal. (C47H82O6) C,
H.
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2,2-bis[[4-
(2,2-d im et h yl-1-oxop r op yl)ben zoyloxy]m et h yl]-1,3-p r o-
p a n ed iyl ester (13a ): acylation of pentaerythritol according
to method A - chromatography on silica gel (1% MeOH/CHCl3
to 2% MeOH/CHCl3) afforded 13a in 49% yield; mp 168-170
°C; 1H NMR (CDCl3) δ 1.32 (s, 36H), 4.72 (s, 8H), 7.62 (d, J )
8.2 Hz, 8H), 8.03 (d, J ) 8.2 Hz, 8H); MS (DCI, ammonia) m/z
(rel intensity) 906 (100); 13C NMR (CDCl3) δ 27.7, 43.2, 44.4,
Hexa d eca n oic a cid , 1-[[[4-(2,2-d im eth yl-1-oxop r op yl)-
p h en yl]m eth oxy]m eth yl]-1,2-eth a n ed iyl ester (15b): acy-
lation of 15a according to method B - chromatography on
silica gel (hexane to 66% ether/hexane) then afforded a 71%
63.6, 127.5, 129.4, 130.7, 143.6, 165.3, 209.2. Anal. (C53H60O12
C, H.
)
1
Ben zoic a cid , 4-(2,2-d im eth yl-1-oxop r op yl)-, 2,2-bis-
(h yd r oxym eth yl)-1,3-p r op a n ed iyl ester (13b): acylation of
2,2-dimethyl-1,3-dioxane-5,5-dimethanol34 according to method
A - acetonide in 98% yield; 1H NMR (CDCl3) δ 1.35 (s, 18H),
1.47 (s, 6H), 3.92 (s, 4H), 4.51 (s, 4H), 7.63 (d, J ) 8.1 Hz,
4H), 8.05 (d, J ) 8.1 Hz, 4H); MS (DCI, NH3) m/z (rel intensity)
570 (100); 13C NMR (CDCl3) δ 23.6, 27.7, 38.0, 44.4, 62.5, 64.2,
98.9, 127.5, 129.3, 131.2, 143.3, 165.5, 209.3. The crude
material from above was deprotected using 0.5 N HCl in THF
at room temperature for 3 h. Ether (100 mL) was added and
the mixture partitioned between ether and water. The aqueous
layer was extracted with ether (2 × 50 mL) and the combined
organic layer was washed with aqueous NaHCO3 (satd), brine,
dried over MgSO4 and concentrated to afford 13b in 98%
yield: mp 104-105 °C; 1H NMR (CDCl3) δ 1.33 (s, 18H), 2.76
(bs, 2H), 3.79 (s, 4H), 4.52 (s, 4H), 7.65 (d, J ) 8.5 Hz, 4H),
8.04 (d, J ) 8.5 Hz, 4H); MS (DCI, NH3) m/z (rel intensity)
513 (7), 530 (100); 13C NMR (CDCl3) δ 27.7, 44.4, 45.6, 62.8,
63.4, 127.5, 129.5, 130.9, 143.6, 166.3, 209.4. Anal. (C29H36O8)
C, H.
Meth od C: Alk yla tion of Glycer ols To P r ep a r e Eth er
An a logu es of 5. Sodium hydride (26 mmol 1.15 equiv per
alcohol equiv) was washed with pentane (2 × 10 mL) and then
suspended in anhydrous THF (35 mL) and cooled to 0 °C. An
alcohol (22.5 mmol) was added and stirred for 15 min. To this
solution 5 (22.5 mmol) was added and the resulting mixture
stirred for 4 h while warming to ambient temperature. The
reaction was concentrated and the solids were diluted with
ether (50 mL), acidified to pH 1 with 2 N HCl, extracted with
4 × 30 mL of ether, dried over MgSO4 and concentrated to a
crude solid. The crude material was either triturated with
hexane or chromatographed to afford pure product.
yield of 15b; mp 32-34 °C; H NMR (C6D6) δ 0.71 (t, J ) 6.9
Hz, 6H), 1.01 (s, 9H), 1.13 (s, 48H), 1.40 (m,4H), 2.03 (m, 4H),
3.22 (m, 2H), 4.00 (d, J ) 3 Hz, 2H), 4.05 (dd, J ) 11.8,6.5
Hz,1H), 4.29 (dd, J ) 11.9,3.5 Hz, 1H), 5.27 (m, 1H), 6.94 (d,
J ) 8.2 Hz, 2H), 7.49 (d, J ) 8.2 Hz, 2H); MS (DCI, NH3) m/z
(rel intensity) 760 (100); 13C NMR (C6D6) δ 14.4, 23.2, 25.3,
25.4, 28.2, 29.5, 29.6, 29.8, 29.9, 30.17, 30.21, 30.22, 30.26, 32.4,
34.3, 34.6, 44.1, 63.1, 69.2, 70.5, 72.8, 127.1, 128.7, 138.2, 141.3,
172.7, 172.9, 206.7. Anal. (C47H82O6) C, H.
9-Octa d ecen oic a cid , 1-[[[4-(2,2-d im eth yl-1-oxop r op yl)-
p h en yl]m eth oxy]m eth yl]-1,2-eth a n ed iyl ester ((9Z)-15c).
To 15a (2.0 g, 7.5 mmol) in CH2Cl2 (30 mL) at 0 °C was added
DMAP (0.3 g, 3 mmol) and oleic acid (4.3 g, 15.2 mmol) followed
by DCC (3.29 g, 16.0 mmol) and the mixture stirred for 36 h
at ambient temperature. The reaction mixture was diluted
with ether, washed with 10% HCl, satd sodium bicarbonate,
dried over MgSO4 and evaporated to an oil. Chromatography
on silica gel (hexane to 30% ether/hexane) afforded 5.73 g (96%
yield) of 15c as a clear oil: 1H NMR (C6D6) δ 0.72 (t, J ) 6.4
Hz, 6H), 1.01 (s, 9H), 1.02 (bs, 40H), 1.39 (m, 4H), 1.90 (t, J )
6.8 Hz, 8H), 2.01 (m, 4H), 3.22 (m,2H), 4.01 (d, J ) 2.8 Hz,
2H), 4.05 (dd, J ) 11.9,6.5 Hz, 1H), 4.29 (dd, J ) 11.9, 3.4 Hz,
1H), 5.30 (m, 5H), 6.95 (d, J ) 8.1 Hz, 2H), 7.50 (d, J ) 8.1
Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 812 (100); exact
mass M + Na 817.6336 (817.6322 theoretical); 13C NMR (C6D6)
δ 14.2, 23.0, 25.2, 27.5, 27.6, 28.0, 29.3, 29.4, 29.5, 29.6, 29.9,
30.0, 30.1, 32.2, 34.1, 34.4, 43.9, 62.9, 68.9, 70.3, 72.5, 126.9,
128.6, 130.0, 130.1, 138.0, 141.1, 172.5, 172.7, 206.5. Anal.
(C51H86O6) C, H.
1-P r op a n on e], 1,1′,1′′-[1,2,3-p r op a n etr iyltr is(oxym eth -
ylen e)-4,1-p h en ylen e]tr is[2,2-d im eth yl- (15e). Sodium hy-
dride (0.52 g, 10.8 mmol) was washed with pentane (2 × 10
mL) and then suspended in anhydrous THF (35 mL) and cooled
to 0 °C. 15a (2.5 g, 9.4 mmol) was added in THF (5 mL) and
the mixture stirred for 30 min before adding 1-[4-(bromom-
ethyl)phenyl]-2,2-dimethyl-1-propanone (5; 2.5 g, 9.5 mmol).
The reaction was allowed to warm to room temperature over
15 h and then cooled to 0 °C before addition of an additional
portion of sodium hydride (0.56 g, 11.6 mmol). After 50 min
an additional amount of 5 (2.6 g, 9.9 mmol) was added and
the reaction was warmed to room temperature over 4 h before
being heated to 60 °C for 24 h. The reaction was cooled and
partitioned between methyl tert-butyl ether and pH 4 buffer.
Subsequent extraction with methyl tert-butyl ether (2 × 50
mL) was followed by drying the combined organic layer over
MgSO4. Chromatography on silica gel (10% EtOAc/hexane to
30% EtOAc/hexane) afforded 15e as a gummy solid (3.4 g,
F or Com p ou n d s su ch a s 14 or 16 Th a t Ha ve a n Acid -
Sen sitive P r otectin g Gr ou p . The crude material was depro-
tected using 0.5 N HCl in THF at room temperature for 3 h.
Ether (100 mL) was added and the mixture partitioned
between ether and water. The aqueous layer was extracted
with ether (2 × 50 mL) and the combined organic layer was
washed with aqueous NaHCO3 (satd), brine, dried over MgSO4
and concentrated to afford crude product. Chromatography
over silica gel afforded pure product.
1-P r op a n on e, 1-[4-[(2,3-d ih yd r oxyp r op oxy)m eth yl]-
p h en yl]-2,2-d im eth yl- (15a ): alkylation of 6 using method
C - purified on silica gel (20% MeOH/CH2Cl2 to 80% MeOH/
1
CH2Cl2) to provide 15a in 92% yield; H NMR (CDCl3) δ 1.35
(s, 9H), 2.33 (bs, 1H), 2.77 (bs, 1H), 3.58 (m, 2H), 3.66 (m, 2H),
3.92 (m, 1H), 4.58 (s, 2H), 7.35 (d, J ) 8.2 Hz, 2H), 7.69 (d, J

522 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4
Bebernitz et al.
69%): 1H NMR (CDCl3) δ 1.34 (s, 27H), 3.67 (d, J ) 6 Hz,
4H), 3.85 (m, 1H), 4.58 (s, 4H), 4.75 (s, 2H), 7.35 (d, J ) 8.0
Hz, 4H), 7.38 (d, J ) 8.0 Hz, 2H), 7.68 (d, J ) 8.0 Hz, 2H),
7.70 (d, J ) 8.0 Hz, 4H); MS (DCI, NH3) m/z (rel intensity)
632 (100); 13C NMR (CDCl3) δ 28.0, 44.2, 70.7, 71.8, 72.9, 77.7,
126.9, 127.0, 128.1, 128.13, 137.7, 137.8, 141.2, 141.7, 208.7.
Anal. (C39H50O6) C, H.
Ack n ow led gm en t. The authors thank Dr. Michael
J . Shapiro and Bertha Owens for NMR and MS analy-
ses. We also thank Dr. Thomas Aicher for helpful
comments while reviewing this manuscript.
Refer en ces
The following were all prepared using method C and the
appropriate alcohol.
(1) Cowie, C. C.; Eberhardt, M. S. Diabetes 1996 Vital Statistics;
American Diabetes Association: Alexandria, VA.
(2) Diabetes Control and Complications Trial Research Group. The
effect of Intensive Treatment of Diabetes on the Development
and Progression of Long-term Complications in Insulin-Depend-
ent Diabetes Mellitus. N. Engl. J . Med. 1993, 329, 977-86.
(3) Sowers, J . R.; Standley, P. R.; Ram, J . L.; J acober, S.; Simpson,
L.; Rose, K. Hyperinsulinemia, Insulin Resistance, and Hyper-
glycemia: Contributing Factors in the Pathogenesis of Hyper-
tension and Atherosclerosis. Am. J . Hypertens. 1993, 6, 260S-
70S.
(4) Depres, J . P.; Lamarche, B.; Mauriege, P.; Cantin, B.; Dagenais,
G. R.; Moorjani, S.; Lupien, P. J . Hyperinsulinemia as an
Independent Risk Factor for Ischemic Heart Disease. N. Engl.
J . Med. 1996, 334, 952-7.
(5) Nestler, J . E.; Clore, J . N.; Blackard, W. G. Dehydroepiandros-
terone: The “Missing Link” Between Hyperinsulinemia and
Atherosclerosis? FASEB J . 1992, 6, 3073-5.
(6) Struck, E.; Ashmore, J .; Wieland, O. Effects of Glucagon and
Long Chain Fatty Acids on Glucose Production by Isolated
Perfused Rat Liver. Adv. Enzyme Regul. 1966, 4, 219-24.
(7) Herrera, M. G.; Kamm, D.; Ruderman, N.; Cahill, G. F. Non-
hormonal Factors in the Control of Gluconeogenesis. Adv.
Enzyme Regul. 1966, 4, 225-35.
(8) Friedmann, B.; Goodman, E. H. J r.; Weinhouse, S. Effects of
Insulin and Fatty Acids on Gluconeogenesis in the Rat. J . Biol.
Chem. 1967, 242, 3620-7.
(9) Golay, A.; Swislocki, A. L. M.; Chem, Y.-DI.; Reaven, G. M.
Relationships Between Plasma Free Fatty Acid Concentration,
Endogenous Glucose Production, and Fasting Hyperglycemia in
Normal and Non-Insulin-Dependent Diabetic Individuals. Me-
tabolism 1987, 36, 692-6.
(10) Ferrannini, E.; Barrett, E. J .; Bevilacqua, S.; DeFronzo, R. A.
Effect of Fatty Acids on Glucose Production and Utilization in
Man. J . Clin. Invest. 1983, 72, 1737-47.
(11) (a) Dulbecco, A.; Albenga, C.; Vacca, G.; Milanesi, G.; Lavezzari,
M. Effect of Acipimox on Plasma Glucose Levels in Patients with
NIDDM. Curr. Ther. Res. 1989, 46, 478-83. (b) Blessath, H.;
Haupt, E.; Luhmann, R.; Hoppe, F.; Radtke, H. W. First
Administration of Etomoxir to type II (NIDDM) patients. Dia-
betologia 1987, 30, 501A.
1-P r op a n on e], 1,1′-[(tetr a h yd r ofu r a n -2-ylid en e)bis-
(m eth ylen eoxym eth ylen e)-4,1-p h en ylen e]bis[2,2-d im eth -
yl-(18): chromatography on silica gel (hexane to 30% diethyl
ether/hexane) afforded 18 in 40% yield; 1H NMR (CDCl3) δ
1.37 (s, 18H), 1.9 (bm, 4H), 3.52 (q, J ) 9 Hz, 4H), 3.90 (t, J )
6.8 Hz, 2H), 4.63 (abq, LS ) 15.1, 5.2 Hz, 4H), 7.37 (d, J ) 8.1
Hz, 4H), 7.71 (d, J ) 8.1 Hz, 4H); MS (DCI, NH3) m/z (rel
intensity) 481 (14), 498 (100); exact mass M + Na 503.2761
(503.2773 theoretical); 13C NMR (CDCl3) δ 26.0, 28.1, 30.8,
44.2, 68.7, 72.9, 73.0, 84.0, 126.8, 128.1, 137.5, 141.6, 208.7.
Anal. (C30H40O5) C, H.
1-P r op a n on e], 1,1′-[(2-bu ten e-1,4-d iyl)bis(oxym eth yl-
en e)-4,1-p h en ylen e]bis[2,2-d im eth yl- ((2Z)-19): chroma-
tography on silica gel (hexane to 30% diethyl ether/hexane)
afforded 19 in 37% yield; 1H NMR (CDCl3) δ 1.38 (s, 18H),
4.11 (d, J ) 5.6 Hz, 4H), 4.72 (s, 4H), 5.82 (m, 2H), 7.34 (d, J
) 8.2 Hz, 4H), 7.69 (d, J ) 8.2 Hz, 4H); MS (DCI, NH3) m/z
(rel intensity) 437 (15), 454 (100); 13C NMR (CDCl3) δ 28.0,
44.2, 66.1, 71.7, 127.1, 128.2, 129.4, 137.8, 141.1, 208.8. Anal.
(C28H36O4) C, H.
1-P r op a n on e], 1,1′-[(2,5-fu r a n d iyl)bis(m eth ylen eoxy-
m eth ylen e)-4,1-p h en ylen e]bis[2,2-d im eth yl- (20): chro-
matography on silica gel (hexane to 30% diethyl ether/hexane)
afforded 20 in 64% yield; 1H NMR (CDCl3) δ 1.34 (s, 18H),
4.50 (s, 4H), 4.58 (s, 4H), 6.30 (s, 2H), 7.37 (d, J ) 8.2 Hz,
4H), 7.68 (d, J ) 8.2 Hz, 4H); MS (DCI, NH3) m/z (rel intensity)
494 (69), 302 (100); 13C NMR (CDCl3) δ 28.0, 44.2, 64.3, 71.4,
110.3, 127.2, 128.1, 137.8, 140.9, 151.9, 208.8. Anal. (C30H36O5)
C, H.
1-P r opan on e], 1,1′-[(1,2-eth an ediyl)bis(th iom eth ylen e)-
4,1-p h en ylen e]bis[2,2-d im eth yl- (21): crude product was
crystallized from hexane to afford 21 in 90% yield; mp 81-82
°C; 1H NMR (CDCl3) δ 1.36 (s, 18H), 2.58 (s, 4H), 3.72 (s, 4H),
7.30 (d, J ) 8 Hz, 4H), 7.68 (d, J ) 8 Hz, 4H); MS (DCI, NH3)
m/z (rel intensity) 443 (17), 460 (100); 13C NMR (CDCl3) δ 28.1,
31.1, 36.0, 44.2, 96.1, 128.4, 128.5, 137.2, 141.3, 208.4. Anal.
(C26H34O2S2) C, H, S.
(12) Young, D. A.; Deems, R. O.; Foley, J . E. Pharmacological
Evaluation of Fatty Acid Oxidation Inhibitors In Vivo. In Lessons
From Animal Diabetes V; Shafrir, E., Ed.; Smith-Gordan Pub-
lishing Co.: Great Britian, 1995; pp 197-204.
(13) Young, D. A.; Ho, R. S.; Bell, P. A.; Cohen, D. K.; McIntosh, R.
H.; Nadelson, J .; Foley, J . E. Inhibition of Hepatic Glucose
Production by SDZ 51641. Diabetes 1990, 39, 1408-13.
(14) Delie, F.; Couvreur, P.; Nisato, D.; Michel, J . P.; Puisieux, F.;
Letourneux, Y. Synthesis and in vitro Study of a Diglyceride
Prodrug of a Peptide. Pharm. Res. 1994, 11, 1082-7.
(15) Deverre, J . R.; Loiseau, P.; Puisieux, F.; Gayral, P.; Letourneux,
Y.; Couvreur, P.; Benoit, J . P. Synthesis of the Orally Macro-
filaricidal and Stable Glycerolipidic Prodrug of Melphalan,
1,3-dipalmitoyl-2-(4′(bis(2′′-chloroethyl)amino)phenylalan-
inoyl)glycerol. Arzneim. Forsch/ Drug Res. 1992, 42, 1153-6.
(16) Counsell, R. E.; Pohland, R. C. Lipoproteins as Potential Site-
specific Delivery systems for Diagnostic and Therapeutic Agents.
J . Med. Chem. 1982, 25, 1115-20.
(17) Wiechert, J . P.; Longino, M. A.; Schwendner, S. W.; Counsell,
R. E. Potential Tumor- or Organ-Imaging Agents. 26. Polyiodi-
nated 2-Substituted Triacylglycerols as Hepatographic Agents.
J . Med Chem. 1986, 29, 1674-82.
(18) Wiechert, J . P.; Groziak, M. P.; Longino, M. A.; Schwendner, S.
W.; Counsell, R. E. Potential Tumor- or Organ-Imaging Agents.
27. Polyiodinated 1,3-Disubstituted and 1,2,3-Trisubstituted
Triacylglycerols. J . Med. Chem. 1986, 29, 2457-64.
(19) Wiechert, J . P.; Longino, M. A.; Bakan, D. A.; Spigarelli, M. G.;
Chou, T. S.; Schwendner, S. W.; Counsell, R. E. Polyiodinated
Triglyceride Analogues as Potential Computer Tomography
Imaging Agents for the Liver. J . Med. Chem. 1995, 38, 636-46.
(20) Aicher, T. D.; Bebernitz, G. R.; Bell, P. A.; Brand, L. J .; Dain, J .
G.; Deems, R.; Fillers, W. S.; Foley, J . E.; Knorr, D. C.; Nadelson,
J .; Otero, D. A.; Simpson, R.; Strohschein, R. J .; Young, D. A.
Hypoglycemic Prodrugs of 4-(2,2-Dimethyl-1-oxopropyl)benzoic
Acid. J . Med. Chem. 1999, 42, 153-63.
1-P r op a n on e, 2,2-d im eth yl-1-[4-[[(p h en ylm eth yl)th io]-
m eth yl]p h en yl]- (22): chromatography on silica gel (hexane
to 30% diethyl ether/hexane) afforded 22 in 83% yield; 1H NMR
(CDCl3) δ 1.36 (s, 9H), 3.61(s, 4H), 7.30 (m, 7H), 7.68 (d, J )
8 Hz, 2H); MS (DCI, NH3) m/z (rel intensity) 299 (11), 316
(100); 13C NMR (CDCl3) δ 28.1, 31.1, 36.0, 44.2, 96.1, 128.4,
128.5, 137.2, 141.3, 208.4 Anal. (C19H22OS) C, H, S; C: calcd,
76.47; found, 76.05.
1-P r op a n on e], 1,1′-[2,6-p yr id in ed iylbis(m eth ylen eoxy-
m eth ylen e)-4,1-p h en ylen e]bis[2,2-d im eth yl- (23): crude
product was crystallized from hexane to afford 23 in 70% yield;
1
mp 49-50 °C; H NMR (CDCl3) δ 1.35 (s, 18H), 4.68 (s, 4H),
4.70 (s, 4H), 7.42 (d, J ) 8.2 Hz, 6H), 7.70, (d, J ) 8.2 Hz,
4H), 7.74, (m, 1H); MS (DCI,NH3) m/z (rel intensity) 488 (100);
13C NMR (CDCl3) δ 28.0, 44.2, 72.4, 73.3, 120.2, 127.1, 128.2,
137.5, 137.9, 141.0, 157.7, 208.7. Anal. (C31H37NO4) C, H, N.
1-P r op a n on e, 2,2-d im et h yl-1-[4-[[2-(4-m or p h olin yl)-
eth oxy]m eth yl]p h en yl]- h yd r och lor id e (24): chromatog-
raphy on silica gel (hexane to 30% EtOAc/hexane) followed by
conversion to the HCl salt afforded 24 in 70% yield; mp 151-
152 °C; 1H NMR (CDCl3) δ 1.35 (s, 9H), 3.00 (m, 2H), 3.27 (bs,
2H), 3.57 (bd, J ) 12.1 Hz, 2H), 3.95 (dd, J ) 12.9, 2.8 Hz,
2H), 4.11 (s, 2H), 4.32 (bt, J ) 12.2 Hz, 2H), 4.59 (s, 2H), 7.34
(d, J ) 8.2 Hz, 2H), 7.69 (d, J ) 8.2 Hz, 2H), 13.24 (bs, 1H);
MS (DCI, NH3) m/z (rel intensity) 306 (100); 13C NMR (CDCl3)
δ 28.0, 44.2, 53.0, 57.5, 63.7, 65.1, 72.9, 127.3, 128.2, 138.4,
139.7, 208.8. Anal. (C18H28NO3Cl) C, H, N, Cl.
(21) Linder, J .; Houlihan, W. J . U.S. patent, 1976, 4 pp; CODEN:
USXXAM US 3962342 19760608 CAN 85:77912 AN 1976:
477912.

Prodrug Approach for Targeting the Liver
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 4 523
(22) McCune, S. A.; Durant, P. J .; Flanders, L. E.; Harris, R. A.
Inhibition of Hepatic Gluconeogenesis and Lipogenesis by Ben-
zoic Acid, p-tert-Butylbenzoic Acid, and a Structurally Related
Hypolipidemic Agent SC-33459. Arch. Biochem Biophys. 1982,
214, 124.
(23) Swartzentruber, M. S.; Harris, R. A. Inhibition of Metabolic
Processes By Coenzyme-A-Sequestering Aromatic Acids. Bio-
chem. Pharmacol. 1987, 36, 3147.
(24) Lu, C. C.; Cagen, S. Z.; Darmer, K. I.; Patterson, D. R. Testicular
Effects Induced by Dermal or Inhalation Exposure to para-
tertiary butyl Benzoic Acid (ptBBA) in Fischer 344 Rats. J . Am.
Coll. Toxicol. 1987, 6, 233.
(25) Hunter, C. G.; Chambers, P. L.; Stevenson, D. E. Studies on the
Oral Toxicity of p-tert-butyl Benzoic Acid in Rats. Food Cosmet.
Toxicol. 1965, 3, 289-98.
(26) For further examples supporting the use of AUC0-48h in evaluat-
ing the release of 3 into circulation as a means of estimating
potential toxicity, see ref 20.
(27) Berry, M. N.; Friend, D. S. high-yield Preparation of Isolated
Rat Liver Parenchymal Cells: A Biochemical and Fine Structure
Study. J . Cell. Biol. 1969, 42, 506-20.
(28) Eckel, R. H.; Hanson, A. S.; Chen, A. Y.; Berman, J . N.; Yost, T.
J .; Brass, E. P. Dietary Substitution of Medium-Chain Triglyc-
erides Improves Insulin-Mediated Glucose Metabolism in NID-
DM Subjects. Diabetes 1992, 41, 641-7.
(29) Mattson, F. H.; Volpenheim, R. A. Hydrolysis of Fully Esterified
Alcohols Containing From One to Eight Hydroxyl Groups By
the Lipolytic Enzymes of Rat Pancreatic J uice. J . Lipid Res.
1972, 13, 325-8.
(30) Lau, D. T.-W.; Kalafsky, G.; Aun, R. L.; Tse, F. L. S. The effect
of the fat content of food on the pharmacokinetics and pharma-
codynamics of SDZ FOX 988, an antidiabetic agent, in the dog.
Biopharmacol. Drug Dispos. 1995, 16, 137-50.
(31) Shimojo, N. Cytochrome P450 Changes In Rats With Strepto-
zocin-Induced Diabetes Int. J . Biochem. 1994, 26, 1261-8.
(32) Wells, M. A.; Direnzo, N. A. The Enzymes; 1983; Vol. XVI,
Chapter 4 Glyceride Digestion.
(33) Complete reduction of triethyl 1,1,2-ethanetricarboxylate with
LAH using standard conditions provided 2-hydroxymethyl-1,4-
butanediol.
(34) Orthner, L. Uber Aceton-Verbindungen des Pentaerythrits.
Chem. Ber. 1928, 61, 116-8.
J M000264W