Metabolically programmed polyamine analogue antidiarrheals

Article abstract of DOI:10.1021/jm950827h

The design, synthesis, and testing of a novel class of antidiarrheal drugs based on a tetraamine pharmacophore are reported. While N¹,N¹⁴- diethylhomospermine (DEHSPM) (5 mg/kg) completely prevents diarrhea in rodents, tissue distrib

Full text of DOI:10.1021/jm950827h

J . Med. Chem. 1996, 39, 2461-2471
2461
Meta bolica lly P r ogr a m m ed P olya m in e An a logu e An tid ia r r h ea ls
Raymond J . Bergeron,* Guo Wei Yao, Hua Yao, William R. Weimar, Charles A. Sninsky,† Brian Raisler,
Yang Feng, Qianhong Wu, and Fenglan Gao
Department of Medicinal Chemistry and College of Medicine, University of Florida, Gainesville, Florida 32610
Received November 9, 1995X
The design, synthesis, and testing of a novel class of antidiarrheal drugs based on a tetraamine
pharmacophore are reported. While N1,N14-diethylhomospermine (DEHSPM) (5 mg/kg)
completely prevents diarrhea in rodents, tissue distribution studies demonstrated that the
principal metabolite of DEHSPM, homospermine (HSPM), accumulates and persists in tissues
for a protracted period of time. This accumulation accounts for a large part of the chronic
toxicity of DEHSPM. Thus a major objective was to develop a metabolically labile analogue of
DEHSPM which retained the desirable biological properties of the parent drug. Hydroxyl
groups, sites vulnerable to further metabolic transformation, were introduced into the external
aminobutyl segments providing N1,N14-diethyl-(3R),(12R)-dihydroxyhomospermine [(HO)2-
DEHSPM]. The design concept was assisted by molecular modeling, which predicted that
(HO)2DEHSPM would have a Ki for polyamine transport essentially identical with that of
DEHSPM. The experimentally measured Ki and also the observed values of other biological
properties of (HO)2DEHSPM were in fact identical with those of DEHSPM, including IC50
against L1210 cells, impact on the NMDA receptor, and impact on L1210 native polyamine
pools. Most significantly, however, there was no accumulation of the dideethylated metabolite
in tissues from mice treated chronically with (HO)2DEHSPM, and (HO)2DEHSPM was 3-fold
less toxic than DEHSPM. Finally, (HO)2DEHSPM completely prevented diarrhea in the castor
oil-treated rat model at a dose of 5 mg/kg, just as did DEHSPM.
In tr od u ction
als clearly represents a critical need in the care of
patients with ARD. It has now become quite clear that
polyamine analogues represent a serious therapeutic
option for ARD, and clearly ARD is but one, albeit
dramatic, of a number of clinical indications that would
benefit from the development of potent new antidiar-
rheals.
Gastrointestinal involvement frequently occurs in the
course of human immunodeficiency virus (HIV) infec-
tion.1 The etiology of AIDS-related diarrhea (ARD) is
complicated and often difficult to assess and includes
both enteric pathogens2,3 and certain drugs.4 Statistics
show that about 50% of HIV patients in the United
States will experience serious diarrhea during the
course of their disease.5,6 While a variety of treatment
protocols have been attempted against various patho-
gens, the response is often only partial, and the relapse
rate is very high.7 Patients not only break through the
therapeutic blanket, but even if the microorganism
responds, an alternate pathogenic infection frequently
develops.
Treatment options for ARD include the application
of antimotility agents,7 luminal-active agents,7 or hor-
monal therapy.3,8 Luminal-acting agents, generally
either fibrous material or resins which absorb unbound
bile acids, rarely provide any assistance in this disease.
Antimotility agents, e.g., tincture of opium, morphine
sulfate, and codeine, can be useful, although there are
concerns with the addictive properties of these drugs.7
Again, none of these agents is exceptionally effective in
ARD. This may well be due to the fact that the problem
has a large secretory component, very similar to that
of cholera.9
In a series of elegant experiments, Tansy was able to
demonstrate that polyamines have a pronounced effect
on the motility of the gastrointestinal (GI) tract.11
His
original work focused on poly(ethyleneimine) and gastric
emptying in rodents12 and dogs.13 Branched-chain poly-
(ethyleneimine)s effected significant inhibition of gastric
emptying in rodents;12 however, they caused a severe
retch response in dogs.13 Because of the structural
relationship between the poly(ethyleneimine)s and natu-
ral polyamines, Tansy elected to evaluate the effect of
spermidine and spermine and a group of polyamine
analogues on the gastric emptying of rodents.14 It soon
became clear that polyamines had a profound influence
on gastric emptying and that “endogenous spermine and
spermidine may have some unrecognized GI secreto-
motor activity”.14 From a structure-activity perspective
it also became obvious that very small changes in the
polyamine’s structure could completely eradicate the
molecule’s ability to inhibit gastric emptying. These
studies strongly suggested that the polyamine pharma-
cophore was an excellent candidate for construction of
antitransit, antidiarrheal drugs. Indeed this has now
been demonstrated to be true.
Until recently octreotide was regarded as a potential
solution to the problem. This somatostatin analogue,
originally developed for diarrhea associated with vipo-
mas, slows gastrointestinal transit and reduces secre-
tory function. Results of a double blind study with this
drug and ARD unfortunately were very disappointing.10
Thus, identification of alternative effective antidiarrhe-
Terminally diethylated homospermine, DEHSPM, a
polyamine analogue designed and synthesized in these
laboratories, is a very potent antidiarrheal.15 This has
been demonstrated in a number of animal models and
recently in the clinic against ARD.16 However, it does
present with some troublesome metabolic properties in
† College of Medicine.
X Abstract published in Advance ACS Abstracts, May 15, 1996.
S0022-2623(95)00827-2 CCC: $12.00 © 1996 American Chemical Society

2462 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13
Bergeron et al.
F igu r e 1. Antidiarrheal activity of polyamine analogues DENSPM, DPNSPM, HSPM, DEHSPM, and (HO)2DEHSPM. All drugs
were administered as an sc injection followed, after 30 min, by gastric gavage of castor oil (5 mL/kg). Stool output (a) and body
weight loss (b) were measured at 2, 4, and 6 h. Both DEHSPM- and (HO)2DEHSPM-treated rats showed virtually no stool output
and lost little of their body weight as compared to castor oil control rats. HSPM-, DPNSPM-, and DENSPM-treated rats all
became diarrheal during the 6 h assay. Data are not corrected for slight evaporative losses that occur during the assay. Such
corrections do not significantly affect the results.
animals. DEHSPM is N-deethylated to homospermine
(HSPM), which has a very protracted half-life, 2-3
weeks in mice and even longer in the dog.17 Thus each
subsequent dose of DEHSPM results in a further
accumulation of HSPM until toxic levels of the parent
tetraamine are reached. The current study is aimed at
the design, synthesis, and testing of an antidiarrheal
DEHSPM derivative into which we incorporate meta-
bolically labile sites in the tetraamine backbone to
ensure that such toxic tetraamines cannot accumulate,
i.e., a metabolically programmed drug.
successfully treating patients with ARD with 30 mg/
m2 DEHSPM, the equivalent of 5 mg/kg in the rat.
However, a problem associated with buildup and reten-
tion of its metabolite, HSPM, in tissues needed to be
addressed.
In order to understand the buildup of the HSPM in
DEHSPM-treated animals, we will briefly summarize
our extensive metabolic studies on DEHSPM17 and
DENSPM.18 While a number of the pharmacokinetic
parameters of DEHSPM and DENSPM were similar,
the metabolic profiles are very different. Although
DENSPM and its metabolites were found in all tissues
examined in mice treated with the drug, the liver and
kidney had the highest level of metabolites. These
catabolic products included N1-ethylnorspermine
(MENSPM), N1-ethylnorspermidine (MENSPD), N1-
ethyl-1,3-diaminopropane (MEDAP), and norspermidine
(NSPD), suggesting that DENSPM is metabolized (Fig-
ure 2) by (1) N-deethylation and (2) stepwise removal
of 3-aminopropyl equivalents by spermine/spermidine
N1-acetyl transferase (SSAT)/polyamine oxidase (PAO),
a metabolic pathway unique to the polyamines.18 Five
days after the end of a 6 day DENSPM treatment
schedule, essentially all of the drug and metabolites had
cleared the tissues.
Resu lts
Design Con cep t. The purpose of the initial antidi-
arrheal studies with the polyamine analogues was to
assess which polyamine pharmacophore backbone was
suitable for the development of antidiarrheals. Four
compounds were evaluated as follows: homospermine
(HSPM), N1,N14-diethylhomospermine (DEHSPM),
N1,N11-diethylnorspermine (DENSPM), and N1,N11
-
dipropylnorspermine (DPNSPM). The antidiarrheal
activity of the polyamine analogues was measured in a
castor oil-induced diarrhea model in rats.
In the case of norspermines, neither of the compounds
was an active antidiarrheal. DPNSPM at 15 mg/kg and
DENSPM even at 40 mg/kg did not demonstrate any
significant activity. When rodents were treated with 5
or 20 mg/kg HSPM, no significant antidiarrheal activity
was observed. However, when animals were adminis-
tered 5 mg/kg DEHSPM, there was an overall 100%
reduction in stooling relative to castor oil controls
(Figure 1a). Not surprisingly animal weight loss paral-
leled stool output (Figure 1b). We are currently and
In the case of DEHSPM, again the liver and kidney
had the highest concentration of drug and its metabo-
lites, N1-ethylhomospermine (MEHSPM) and HSPM.
The data clearly showed that as with DENSPM, N-
deethylation is a key metabolic step in processing
DEHSPM (Figure 3), but, that unlike NSPM, HSPM
does not undergo further metabolism. As pointed out
above, normal polyamine catabolism involves stepwise

Metabolically Programmed Polyamine Antidiarrheals
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13 2463
as drug or metabolites. Interestingly 84% of the drug
remaining in the animal at this time is in the form of
HSPM. By 5 days after the end of DEHSPM treatment,
essentially no DEHSPM remains in the form of the
parent drug, yet about 25% of the total dose is present
as HSPM. It is quite clear that the increased chronic
toxicity of DEHSPM over DENSPM is related to the
buildup of HSPM.17 The chronic toxicity of HSPM is 5
times greater than that of NSPM. Interestingly the
chronic LD50 of DEHSPM is almost the same as that of
HSPM and 5 times greater than that of DENSPM.
The key to a less toxic DEHSPM-like antidiarrheal
is one in which the metabolite can be cleared from the
tissues. Again because of the aminobutyl fragments of
HSPM, this metabolite cannot be processed through the
polyamine biosynthetic network; thus, it remains in the
tissues for protracted periods of time. Neither the
primary nor the secondary nitrogens of HSPM offer an
opportunity for facile conversion to an easily cleared
metabolite. Certainly the methylene backbones cannot
be easily oxidized to an excretable metabolite. We thus
elected to raise the oxidation state of two of the
methylene groups of DEHSPM by introducing a single
hydroxyl in the R-configuration at each of the external
aminobutyl fragments, resulting in N1,N14-diethyl-
(3R),(12R)-dihydroxyhomospermine, [(HO)2DEHSPM]
(7), Scheme 1. This offered potential sites for either
conjugation (e.g., glucuronidation) or further oxidation
and clearance.
F igu r e 2. Metabolic breakdown (catabolism) of DENSPM.
Two enzymatic systems are operative: (a) N-deethylation (solid
arrows) and (b) de-3-aminopropylation (open arrows) by the
constituitive system for catabolism of native polyamines,
SSAT/PAO. The various permutations of these enzymatic
reactions are shown, and each of the possible metabolites
indicated in the figure was in fact observed in tissues of
DENSPM-treated mice.18
Com p u ter Mod elin g. The potential of comparative
molecular field analysis (CoMFA) as a predictive tool
in identifying analogues which would likely be success-
ful as antidiarrheals has now been demonstrated.
CoMFA is a relatively new technique used to correlate
receptor-ligand affinities with the molecular, steric,
and electrostatic fields presented by the ligands. The
output of these studies is an interactive 3D color contour
plot from which one can deduce which structural
modifications are likely to improve activity.19 Thus, the
approach allows one to predict the activity of new
structures which have not been synthesized and are not
part of the model data set. Preliminary studies suggest
this is a very useful tool in polyamine analogue design.
As part of a National Cancer Institute-sponsored pro-
gram focused on the development of polyamine ana-
logues as antineoplastics, we have already applied the
method to search for a quantitative relationship be-
tween structure and the ability of the analogue to
compete with spermidine for the polyamine transport
apparatus, Ki values. The initial quantitative structure-
activity relationship (QSAR) CoMFA included 26
polyamine analogues and provided a model with a cross-
validated r2 ) 0.697 (optimum components five) and the
conventional R2 ) 0.973, F test ) 152.726, S ) 0.143.20-22
The molecular modeling shows that the relative contri-
butions of the steric and electrostatic terms in the QSAR
equation are 0.673 and 0.327, respectively, specifying
the importance of the geometry of the groups fixed to
the nitrogen and of charge.
F igu r e 3. Catabolism of DEHSPM. Only the N-deethylation
pathway is operative; the terminal 4-aminobutyl segments of
HSPM are not recognized by SSAT/PAO, so that further
degradation does not occur.
removal of aminopropyl equivalents by SSAT/PAO. In
this pathway only primary 3-aminopropyl nitrogens are
acetylated, and since HSPM has only 4-aminobutyl
fragments, it cannot be processed. The accumulation
and persistence of HSPM in the tissues of DEHSPM-
treated animals are especially striking. One day after
a 6 day treatment schedule of DEHSPM, a regimen
identical with that of DENSPM, 35% of the drug
administered to mice remains in the liver and kidney
The library of compounds includes both linear ana-
logues and the cyclic dipiperidyl and dipyridyl com-
pounds along with N1,N12-bis(2,2,2-trifluoroethyl)sper-
mine (FDESPM). The program was presented with a
group of tetraamines in which the nitrogens were both
charged and uncharged at physiological pH; thus it was
able to assess the contribution of charge to Ki data

2464 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13
Bergeron et al.
Ta ble 1. Comparison of Experimental and Predicted Ki Values of Tetraamine Analogues Using CoMFA Models
aCompetitive inhibition of radiolabeled spermidine uptake in L1210 cells [Km(SPD) ) 2.1 ( 0.2 µM]. These values29,30 represent the
mean of at least three such experiments with a standard deviation typically about 10%.
(Table 1). Prior to embarking on the synthesis of (HO)2-
DEHSPM, we “asked” the computer to predict the Ki
value. SYBYL suggested that (HO)2DEHSPM should
be recognized by the polyamine transport apparatus as
effectively as DEHSPM with nearly identical calculated
Ki values of 2.0 and 1.9 µM, respectively. The agree-
ment between the actual Ki values was also outstanding,
1.8 vs 1.4 µM, respectively (Table 2).
In this regard we elected to assess how effective a Ki
value would be in projecting the compound’s IC50 values
against L1210 cells, its impact on polyamine pools, its
effect on the NMDA receptor, and most importantly its
antidiarrheal activity. A prori one could not anticipate
any quantitative relationship between competition for
the polyamine transport apparatus and any of these
other biological properties. However, perhaps it is not
unreasonable to expect that if two molecules are rec-
ognized by this transport apparatus as being structur-
ally similar, i.e., close in Ki values, it would not be
surprising to see they have similar pharmacological
properties in several areas.
make hypusine.23 N,N′-Dibenzylputrescine (1)24 added
regiospecifically to (S)-(+)-epichlorohydrin (2 equiv) to
form (S,S)-bis-chlorohydrin 2 (Scheme 1). Treatment
of 2 with KCN in the presence of 18-crown-6 gave
dinitrile 3, in which the alcohols are in the R-configu-
ration. The cyano groups of 3 were hydrogenated with
Raney nickel in methanolic ammonia resulting in
primary R,ω-diamine 4. Hydrogenolysis of the N-benzyl
protecting groups of 4 under mild conditions (1 atm, 10%
Pd-C, methanol, 4 equiv of HCl) furnished (HO)2HSPM
(8) as its tetrahydrochloride salt. Compound 4 was also
converted to its tetraacetyl derivative 5 with acetic
anhydride in methylene chloride. Exposure of 5 to
LiAlH4 in THF simultaneously reduced the acetamides
to ethylamines and unmasked the secondary carbinols
generating tetraalkylated polyamine 6. Catalytic re-
moval of the benzyl protecting groups of 6 over Pd-C
in ethanolic HCl afforded terminally diethylated homo-
spermine R,R-diol 7 as its tetrahydrochloride salt.
Biologica l Stu d ies
Syn th esis
IC50 for L1210 Cell Gr ow th . While (HO)2DEHSPM
is more highly substituted than DEHSPM, it neverthe-
less maintains enough structural similarity to be “read”
by L1210 cells as DEHSPM. The 48 and 96 h IC50
values of (HO)2DEHSPM vs DEHSPM are very similar
The synthesis of polyamines (HO)2DEHSPM (7) and
(3R),(12R)-dihydroxyhomospermine [(HO)2HSPM] (8),
which contain chiral alcohols on the outer methylene
chains, was accomplished by our methodology used to

Metabolically Programmed Polyamine Antidiarrheals
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13 2465
Sch em e 1. Synthesis of
N1,N14-Diethyl-(3R),(12R)-dihydroxyhomospermine
[(HO)2DEHSPM] (7) and
(3R),(12R)-Dihydroxyhomospermine [(HO)2HSPM] (8)a
F igu r e 4. Comparison of the effects of DEHSPM and
(HO)2DEHSPM on L1210 cell 96 h growth curves. The
concentration of drug that results in 50% inhibition of cell
growth, i.e., the IC50, is between 0.05 and 0.09 µM for
DEHSPM and (HO)2DEHSPM as indicated by the intersection
of the ‘50% growth line’ with the respective curves.
Ta ble 3. Impact of Polyamine Analogues on Polyamine Poolsa
compd
concn (µM) PUT SPD SPM analogueb
DEHSPM
(HO)2DEHSPM
(HO)2DEHSPM
10
0.7
3.5
0
45
0
0
57
0
61
96
68
2.94
1.32
2.73
a
Native polyamine levels in analogue-treated L1210 cells are
relative to untreated controls (control ) 100%). Typical control
values in pmol/106 L1210 cells are PUT ) 192 ( 38, SPD ) 2751
a
Reagents: (a) (S)-(+)-epichlorohydrin/MgSO4/MeOH; (b) KCN/
b
18-crown-6/CH3CN; (c) H2/Ra Ni/NH3/MeOH; (d) Ac2O/CH2Cl2; (e)
LiAlH4/THF; (f) H2/10% Pd-C/MeOH or EtOH/HCl.
( 163, and SPM ) 726 ( 34. Analogue amount is expressed as
nmol/106 cells. Untreated L1210 cells (106) correspond to about 1
µL volume; therefore, concentration can be estimated as nmol/
106 cells f mmol/L f mM. Treatment with polyamine analogues
generally causes a shrinkage of L1210 cell volume dependent on
time of exposure.30
Ta ble 2. Analogue Structures, Acronyms, L1210 Growth
Inhibition, and Effects on [3H]Spermidine Transport
Mod u la tion of [3H]MK-801 Bin d in g to th e NMDA
Recep tor by (HO)2DEHSP M. The retention of cer-
tain characteristic pharmacological properties of the
DEHSPM pharmacophore by (HO)2DEHSPM is clearly
exemplified in Figure 5. We have previously reported
the marked activity of terminally dialkylated tet-
raamines including DEHSPM on the NMDA receptor
and have shown that very small structural changes can
result in striking differences in this activity among
polyamine analogues.25 The dose-effect curves of DE-
HSPM and (HO)2DEHSPM are virtually superimpos-
able. Agonist activity occurs at 0.5-1.0 µM, but marked
antagonist activity is observed at concentrations above
5 µM with maximal antagonism at concentrations of 50
µM and above for both polyamines.
Meta bolism of (HO)2DEHSP M vs DEHSP M in
Mou se Liver a n d Kid n ey. Dansylated samples of
authentic (HO)2DEHSPM or (HO)2HSPM resulted in a
single fluorescent peak in the HPLC with a molar
response similar to that of DEHSPM or HSPM. Dan-
sylation of the HClO4 extracts of (HO)2DEHSPM-treated
liver and kidney contained the peaks corresponding to
the natural polyamines, PUT, SPD, and SPM as well
as unchanged (HO)2DEHSPM. In addition there was
a single additional fluorescent peak with a retention
time identical with that of authentic dansylated
(HO)2HSPM. We were unable to identify a peak in the
area in which we would anticipate the dansylated
monodeethylated metabolite of (HO)2DEHSPM. How-
aThe Ki values were estimated from Lineweaver-Burke plots,
which showed simple substrate-competitive inhibition of radiola-
beled spermidine transport in L1210 cells. The Vmax was typically
about 1 nmol/h/106 cells; Km(SPD) ) 2.1 ( 0.2 µM.
(Table 2). In fact the actual 48 and 96 h IC50 curves
for DEHSPM and (HO)2DEHSPM are nearly identical
(see Figure 4 for 96 h curves).
Ki Stu d ies. The ability of (HO)2DEHSPM and DE-
HSPM to compete with [3H]spermidine for uptake into
L1210 cells in vitro is very similar, with Ki values of
1.8 and 1.4 µM, respectively (Table 2).
P olya m in e P ools. When L1210 cells were grown in
10 µM DEHSPM (Ki ) 1.4 µM) for 48 h, they achieved
an intracellular concentration of 2.9 mM DEHSPM. At
0.7 µM (HO)2DEHSPM (0.4Ki) for 48 h, the concentra-
tion of intracellular drug reached 1.3 mM, while in 3.5
µM (HO)2DEHSPM (2Ki), intracellular level of analogue
reached 2.7 mM, almost identical with the intracellular
DEHSPM concentration. At this essentially equimolar
intracellular concentration, each analogue had the same
impact on polyamine pools. Putrescine and spermidine
were depleted to below detectable limits, while spermine
was reduced to about 60% of control values (Table 3).

2466 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13
Bergeron et al.
F igu r e 5. Effects of (HO)2DEHSPM and DEHSPM on NMDA
receptor (MK-801) binding. Each data point represents the
mean ( SD of three experiments. Each experiment was in turn
performed in triplicate; “100%” activity is defined as the
amount of specific MK-801 binding to washed rat cerebral
cortical membranes in the presence of saturating concentra-
tions of coagonists (100 µM L-glutamate, 100 µM glycine) and
2 nM [3H]MK-801.
F igu r e 7. Quantitation of parent drug, (HO)2DEHSPM,
dideethylated metabolite, (HO)2HSPM, and native polyamine
pools in mouse liver (a) and kidney (b). Groups (N ) 3; (SD)
were sacrificed for HPLC analysis 24 h after cessation of a
subchronic dosage regimen (20 mg/kg/day (HO)2DEHSPM as
a single ip injection daily for 5 days; total dose ) 3.0 mg or
6.46 µmol for a 30 g mouse). For comparison, similar data for
untreated control and DEHSPM-treated (received “equimolar”
subchronic dose of DEHSPM ) 2.7 mg or 6.25 µmol for a 30 g
mouse) mice are plotted. Total polyamine N+ equivalents (
SD (nequiv × 10-1/g of wet wt) for these tissues are calculated
and given in Table 4.
lar for (HO)2DEHSPM and DEHSPM. In the liver the
level of unchanged drug is similar and low: 83.6 ( 11.7
vs 43.9 ( 12.6 nmol/g of wet weight for (HO)2DEHSPM
and DEHSPM, respectively (Figure 7a). Levels in the
kidney are also similar, though substantially higher
than those in the liver: 481 ( 46 [(HO)2DEHSPM] vs
409 ( 178 [DEHSPM] (Figure 7b). However, there is
no substantial accumulation of metabolic products of
(HO)2DEHSPM in either tissue. This is in marked
contrast with DEHSPM, in which tissues from treated
animals contained the dideethylated metabolite HSPM
in substantially higher concentrations than the parent
drug. Thus HSPM is present in large quantities in the
DEHSPM-treated liver (525 ( 44 nmol/g of wet weight)
and at even higher levels in the kidneys, where it is
easily the most abundant polyamine in the tissue (815
( 61 nmol/g of wet weight), equivalent to all the other
polyamines combined. The corresponding dideethylated
metabolite (HO)2HSPM is, in contrast, present in both
liver (63 ( 2 nmol/g of wet weight) and kidney (63 ( 9
nmol/g of wet weight) but in concentrations lower than
the parent drug (liver, 84 ( 12 nmol/g of wet weight;
kidney, 481 ( 46 nmol/g of wet weight). The conse-
F igu r e 6. Proposed catabolism of (HO)2DEHSPM. The N-
deethylation pathway is operative as with DEHSPM, but in
contrast to DEHSPM, the terminal 2-hydroxy-4-aminobutyl
segments of (HO)2HSPM are further degraded by other, yet
to be characterized, enzymatic systems. Thus the metabolite
(HO)2HSPM does not accumulate.
ever without an authentic sample of this metabolite, it
is not clear that its dansylated derivative is not buried
under other signals. This remains to be established.
Nevertheless it is clear that the initial step in (HO)2-
DEHSPM metabolism involves simple deethylation
(Figure 6) as with DEHSPM.
It appears that the initial steps in the tissue distribu-
tion and metabolism of (HO)2DEHSPM are indistin-
guishable from DEHSPM. The levels of unmetabolized
drug present in liver and kidney 1 day after cessation
of a subchronic equimolar treatment regimen are simi-

Metabolically Programmed Polyamine Antidiarrheals
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13 2467
Ta ble 4. Effects of Chronic (HO)2DEHSPM/DEHSPM
Treatment on Polyamine Homeostasis
30% of their body weight (data not shown) but recovered
in a period of 20 days after the dosing regimen. This
was in contrast to DEHSPM-treated animals wherein
a 20% loss of body weight was a sign of imminent death.
Several severely affected (HO)2DEHSPM-treated ani-
mals were necropsied. It was noted that the stomachs
of these animals appeared herniated and were ex-
tremely distended with food. This suggests that this
compound potently blocks gastric emptying.
untreated
control
tissue
liver
(HO)2DEHSPMa DEHSPMb
native [N+] 6552 ( 609
6118 ( 395
588 ( 84
6706 ( 479
3251 ( 149
2176 ( 192
5427 ( 340
5506 ( 455
2300 ( 250
7806 ( 706
1560 ( 474
5034 ( 1060
6594 ( 1534
drug [N+]
0
total [N+]c 6552 ( 609
kidney native [N+] 4115 ( 296
drug [N+]
total [N+] 4115 ( 296
0
P r even tion of Ca stor Oil-In d u ced Dia r r h ea . The
dosage of DEHSPM administered to the rats in this
study (5 mg/kg) is equivalent to the clinically used dose
of 30 mg/m2 in humans. At this dosage DEHSPM
effectively prevented the onset of diarrheal stooling for
the 6 hour time period of the assay in this castor oil-
induced diarrhea rat model. At the same dose of 5 mg/
kg, (HO)2DEHSPM was as effective as DEHSPM at
preventing diarrhea. In addition, the body weight loss
for each group was very similar. Neither group exhib-
ited any diarrheal stooling during the assay. This
suggests that (HO)2DEHSPM could be an effective
antidiarrheal agent for clinical use. The dose-response
relationship shows, as expected, the higher dose of 18
mg/kg (HO)2DEHSPM effectively prevented the onset
of diarrhea but indicated that 5 mg/kg (HO)2DEHSPM
was the minimum effective dose for prevention of
diarrhea in this assay. At 2.5 mg/kg, (HO)2DEHSPM
was partially effective, resulting in reduced stool output
(Figure 8a) and maintenance of body weight (Figure 8b),
although some diarrhea did occur within the 6 h
observation period. At 1 mg/kg, (HO)2DEHSPM was not
an effective antidiarrheal.
a
Total dose: 20 mg/kg/day × 5 days f 100 mg/kg (215.3 µmol/
b
kg). Total dose:15 mg/kg/day × 6 days f 90 mg/kg (208.3 µmol/
kg). c Total [N+] equivalents ) 2[diamine] + 3[triamine] + 4[tet-
raamine]. Values expressed as nequiv of [N+]/g of wet weight.
quences of these polyamine levels on the net polyamine
[N+] equivalents present in each tissue are depicted in
Table 4.
The lack of accumulation of the dideethylated me-
tabolite from (HO)2DEHSPM is best illustrated in Table
5, which analyzes the data from the perspective of mass
balance of the cumulative dose, i.e., how much of the
original total dose remains in the liver and kidney and
what form is it in, parent drug or metabolite? A 30 g
mouse would have received a cumulative dose of 2.7 mg
(6.25 µmol) of DEHSPM or 3.0 mg (6.46 µmol) of
(HO)2DEHSPM. One day after cessation of treatment,
33.8% (2.11 µmol) of the original dose of DEHSPM can
be accounted for in liver and kidneys. However, only a
small portion, 5.0% (or 0.311 µmol), of the original dose
is present in the form of unchanged parent drug; 28.8%,
or 1.80 µmol, of the total original dose remains in these
organs in the form of the dideethylated metabolite,
HSPM. In (HO)2DEHSPM-treated animals, a similar
amount (6.96% of dose, or 0.450 µmol) of unmetabolized
parent drug is present in liver and kidneys 24 h after
cessation of treatment, but only 3.03% of the original
dose is present in the form of the dideethylated me-
tabolite, (HO)2HSPM. In each tissue the amount of
parent drug exceeds that of the metabolite, reflecting
the lack of accumulation of (HO)2HSPM.
Although the onset of diarrhea was slightly delayed
relative to control rats and the total weight of stool
output was somewhat less, the norspermine analogues,
DENSPM and DPNSPM, even at doses of 40 and 15 mg/
kg, respectively, were far less active as antidiarrheal
compounds. Data for both stool output (Figure 1a) and
weight loss (Figure 1b) clearly illustrate the similarity
between DEHSPM and (HO)2DEHSPM and their dif-
ferences from DENSPM, DPNSPM, and HSPM.
Acu te Toxicity. The acute toxicity of (HO)2DEHSPM
was assessed after ip administration of a single dose of
325 mg/kg to a group of five female CD-1 mice. Three
of the five animals died in this experiment, suggesting
that the acute toxicity of (HO)2DEHSPM is similar to
that of DEHSPM (LD50 ) 325 mg/kg)17 and DENSPM
(LD50 ) 325 mg/kg).18
Discu ssion
On the basis of animal studies in rodents and dogs,
we must anticipate the possibility for toxicity as patients
receive more chronic dosage regimens of DEHSPM in
the treatment of chronic, severe ARD. Thus our central
goal was to develop a less toxic polyamine with similar
potency to DEHSPM as an antidiarrheal. Since the
HSPM backbone appeared to be required for potent
antidiarrheal activity, our strategy was to render these
terminal aminobutyl segments susceptible to further
metabolic degradation without significantly altering the
nonmetabolic biological properties of the DEHSPM
pharmacophore. Thus while (HO)2DEHSPM is indeed
more highly substituted than DEHSPM, it nevertheless
maintains enough structural similarity to DEHSPM to
be “read” by L1210 cells as DEHSPM. First, the
abilities of (HO)2DEHSPM and DEHSPM to compete
with radiolabeled spermidine for transport into L1210
cells were found to be essentially identical, just as
computer-assisted molecular modeling had predicted.
HPLC analysis of L1210 cells grown for 48 h in the
presence of analogue showed that at an intracellular
concentration of ca. 3 mM, each analogue had a very
similar effect on intracellular native polyamine pools.
Finally, (HO)2DEHSPM and DEHSPM have the same
Ch r on ic Toxicity. In comparing the chronic toxicity
of DEHSPM with that of (HO)2DEHSPM, we elected not
to use historical controls. The chronic LD50 for DEH-
SPM was determined to be 38 mg/kg/day × 5 days in
this study. This value confirms the LD50 value we had
previously reported.17 No deaths were observed in
groups of six animals each at (HO)2DEHSPM doses of
20, 40, and 60 mg/kg/day × 5 days, so experiments were
initiated at higher doses (no. of deaths/no. of animals):
70 (0/5), 80 (0/5), 90 (0/10), 100 (1/5), and 120 (2/5) mg/
kg/day × 5 days. These data indicate an LD40 for (HO)2-
DEHSPM of 120 mg/kg/day × 5 days. We estimate the
LD50 to be 130 mg/kg based on the slope of the toxicity
curve. These data establish (HO)2DEHSPM to be at
least 3 times less toxic in this chronic treatment regimen
than its analogue DEHSPM. While the lower doses of
(HO)2DEHSPM did not cause deaths, animals did
exhibit signs of toxicity. At 90 mg/kg for 5 days, the
animals all exhibited lethargy, poor coat condition, and
decreased, ataxic movements. The animals in this
(HO)2DEHSPM-treated group lost between 20% and

2468 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13
Ta ble 5. Amounta (%) of Total Dose Present in Liver and Kidney
Bergeron et al.
unmetabolized parent drug,
DEHSPM/(HO)2DEHSPM (µmol)
dideethylated metabolite,
HSPM/(HO)2HSPM (µmol)
total drug +
metabolite (µmol)
treatment
liver
DEHSPMb
0.115 (1.84 ( 0.53%)
0.219 (3.39 ( 0.48%)
0.196 (3.14 ( 1.37%)
0.231 (3.57 ( 0.34%)
0.311 (4.98%)
1.394 (22.30 ( 2.10%)
0.166 (2.57 ( 0.38%)
0.408 (6.53 ( 0.67%)
0.030 (0.47 ( 0.01%)
1.802 (28.83%)
1.509 (24.15%)
0.385 (5.96%)
0.604 (9.67%)
0.261 (4.04%)
2.113 (33.82%)
0.646 (10.01%)
(HO)2DEHSPMc
DEHSPMb
kidneys
both tissues
(HO)2DEHSPMc
DEHSPMb
(HO)2DEHSPMc
0.450 (6.96%)
0.196 (3.03%)
a
b
Mass balance accounting for fate of the total orginal dose normalized for a 30 g mouse. Total dose ) 15 mg/kg/day × 6 days ) 90
mg/kg f 2.7 mg (6.25 µmol) for a 30 g mouse. c Total dose ) 20 mg/kg/day × 5 days ) 100 mg/kg f 3.0 mg (6.46 µmol) for a 30 g mouse.
ment to enable metabolic clearance of terminal dieth-
yltetraamine analogues is N-deethylation to provide a
primary amine terminus. It is interesting to note that
under our in vitro conditions, L1210 cells apparently
do not N-deethylate either (HO)2DEHSPM or DEHSPM
to any significant extent. In contrast, tissues examined
from animals (mouse, rat, dog) treated with DENSPM,18
DEHSPM,17 or (HO)2DEHSPM do contain N-deethy-
lated metabolites, thus meeting the first requirement
of SSAT for a primary amine terminus. However, if the
second criterion of an exposed aminopropyl segment is
not met, further degradation via the SSAT/PAO path-
way is precluded. Thus HSPM, a tetraamine comprised
exclusively of aminobutyl segments, can neither be
further degraded nor exported from the cell, resulting
in its accumulation. We postulated that this accumula-
tion and persistent retention of HSPM was responsible
for the markedly greater toxicity (5-fold) of DEHSPM
vs DENSPM when administered chronically, even though
the acute toxicities of DENSPM and DEHSPM are
essentially identical.17,18 Therefore the design concept
was to modify the polyamine backbone of DEHSPM in
such a manner as to render the aminobutyl segments
susceptible to further metabolic transformation and
clearance.
We elected to introduce hydroxy groups into the
external aminobutyl segments of DEHSPM in a position
and configuration analogous to the (R)-(-)-2-hydroxy-
putrescine segment of (+)-hypusine.23 The rat liver is
able to convert 2-hydroxyputrescine itself to γ-amino-
â-hydroxybutyric acid.28 Thus, we hoped that enzymes
involved in 2-hydroxyputrescine and/or hypusine me-
tabolism or perhaps other enzyme systems present in
normal tissues might be able to degrade or otherwise
facilitate elimination of the tetraamine. Possibilities
could include, for example: (1) direct conjugation of the
OH to make the molecule exportable and excretable,
e.g., glucuronidation, or (2) oxidation at the hydroxyl-
bearing carbon by either (a) a constitutive enzyme
system normally functioning in hypusine metabolism,
or (b) some other more ‘nonspecific’ oxidase/dehydroge-
F igu r e 8. Antidiarrheal dose-response for (HO)2DEHSPM
nase to result in cleavage of the external aminobutyl
in the rat castor oil-induced diarrheal model. Drug was
segments. While we do not yet have any evidence as to
administered as an sc injection followed, after 30 min, by
which of these or other degradative mechanisms is
gastric gavage of castor oil (5 mL/kg). Stool output (a) and body
operating, it is clear that there is no significant ac-
cumulation of (HO)2HSPM. One day after a chronic
treatment regimen, only 10% of the total dose of
(HO)2DEHSPM remains in liver and kidney compared
with 34% of the total DEHSPM dose. In both liver and
kidney, DEHSPM had significantly disrupted normal
tissue polyamine homeostasis as indicated by the effects
on native polyamine pools and total [N+] equivalents
(‘net charge balance’). In contrast equimolar treatment
with (HO)2DEHSPM had no significant impact on
polyamine homeostasis in the liver, and while there was
weight loss (b) were measured at 2, 4, and 6 h. The mininum
effective dose for (HO)2DEHSPM was 5 mg/kg. Data are not
corrected for slight evaporative losses that occur during the
assay. Such corrections do not significantly affect the results.
effect on L1210 cell growth, with nearly identical 48 and
96 h dose-effect curves.
Tetraamines are only poorly exported from cells
unless acetylated,26 and the cytosolic acetylase SSAT
only recognizes aminopropyl moieties with a primary
amino terminus as substrates.27 Thus the first require-

Metabolically Programmed Polyamine Antidiarrheals
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13 2469
program.19 The CoMFA were performed using the QSAR
option of SYBYL 6.02 program running on a Silicon Graphics
Indigo2 workstation.
some impact on the kidney, it was mild compared to that
of DEHSPM. These results support our view that the
degree of tetraamine buildup and its duration are
responsible for a substantial portion of the toxicity that
can occur with DEHSPM and explain the 3-fold decrease
in chronic toxicity of (HO)2DEHSPM compared to DE-
HSPM and thus validate the hydroxylated polyamine
analogue design concept.
The energies of steric and electrostatic interaction between
each of the analogues were calculated as parameters, and a
probe atom was placed at the various intersections of a regular
3D lattice, large enough to include all of the analogues in the
database and with a 2.0 Å lattice space. The probe atom had
the shape of sp3 carbon and a charge of +1.0. The van der
Waals values were taken from the standard Tripos force field,
and the atomic charges were calculated by the method of
Gasteiger and Marsili.20 Wherever the probe atom experiences
a steric repulsion greater than “cutoff” (30 kcal/mol in these
studies), the steric interaction is set to the value “cutoff”, and
the electrostatic interaction was set to the mean of the other
molecular electrostatic interactions at the same location. The
resulting steric and electrostatic field values, on a regularly
spaced lattice around the analogues, were then correlated with
the Ki data for the polyamine transport system by a very
efficient statistical method, partial least squares (PLS).21
Statistical techniques, bootstrapping, and cross-validation are
used to determine the quality of the correlation.21,22 The cross-
validation technique involves random elimination of one or
more analogues from the original data set with subsequent
equation development and activity prediction for the elimi-
nated analogues in an interactive manner, thus yielding a
QSAR equation that is generally of greater predictive value
than that derived from conventional regression analysis. The
implementation of PLS also rotates the PLS solution back into
the original data space, thus generating a “conventional”
QSAR equation showing r-square, F test, and the standard
error S.
Con clu sion
The data clearly support earlier findings that the
critical issue in how the cell “reads” a polyamine
analogue is very dependent on the molecule’s charge and
the separation between the charge centers.29 Thus it
is possible to operate on the methylene backbones, i.e.,
the insulators between charged centers, of these mol-
ecules, altering their metabolic properties while still
maintaining the desired pharmacological activity. As
we have demonstrated, such metabolic programming
can result in polyamine analogues with substantially
reduced toxicities. Finally, it seems clear that the
CoMFA profiles generated from our expanding library
of polyamines are very useful in drug design by assess-
ing whether a cell will recognize modified analogues,
at least at the level of polyamine transport. These
modifications in the backbone can be made while still
maintaining the desired pharmacological properties.
The initial alignment of the molecules, the positioning of a
molecular model within the fixed lattice, is by far the most
critical step in developing a successful CoMFA model, since
the relative interaction energies depend strongly on relative
molecular positions. In this series of molecules, both the end
nitrogen atoms of the analogues and the third nitrogen atom
were finally found as atom pairs between two molecules for
performing a best fit and then followed again by energy
minimization using the standard Tripos force field option, and
the resulting structures were used to generate the CoMFA
analysis. DESPM (15) was chosen as the template in the
alignment of polyamine analogues.
Cell Cu ltu r e. Murine L1210 leukemia cells were main-
tained in logarithmic growth as a suspension culture in RPMI-
1640 medium containing 10% FBS, 2% HEPES-MOPS buffer,
1 mM aminoguanidine, and 1 mM L-glutamine (Sigma) at 37
°C in a water-jacketed 5% CO2 incubator.
Exp er im en ta l Section
Chemical reagents were purchased from Aldrich Chemical
Co. Reactions using hydride reagents were run in distilled
DMF or THF (from sodium and benzophenone) under a
nitrogen atmosphere. Fisher Optima grade solvents were
routinely used, and organic extracts were dried with sodium
sulfate. Silica gel 32-63 (40 µM “flash”) from Selecto, Inc.
(Kennesaw, GA) was used for column chromatography. Melt-
ing points were determined on a Fisher-J ohns melting point
apparatus and are uncorrected. Proton NMR spectra were run
at 300 MHz in CDCl3 (not indicated) or D2O with chemical
shifts given in parts per million downfield from tetramethyl-
silane or 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid, sodium
salt, respectively. Coupling constants (J ) are in hertz. FAB
mass spectra were run in a m-nitrobenzyl alcohol matrix.
Elemental analyses were performed by Atlantic Microlabs,
Norcross, GA.
Com p u ter Mod elin g. A series of N-alkylated polyamines,
terminally dialkylated analogues and homologues of spermine
which have been synthesized by methods developed in these
laboratories,29-31 were used to generate the CoMFA analysis.
The structures and Ki values, both calculated and actual, for
the polyamine uptake apparatus of these polyamine analogues
are listed in Table 1.
On the basis of pKa values measured29 for DENSPM (11),
DESPM (15), DEHSPM (20), FDESPM (17), PIP(4,4,4) (29),
PYR(4,4,4) (30), and PIP(5,4,5) (31) (Table 1), we have made
the following assumptions about the protonation state. All of
the nonaromatic tetraamines are largely in the form of the
tetracation at physiological pH 7.2. The spermine analogues
13-16 and 34 and the homospermine analogues and homo-
logues 18-26, 29, and 31-33 should be 85% and 97%,
respectively, in the form of the tetracations. The norspermine
analogues 9-12 and 27 should be at least 74% in the form of
the tetracations. Compounds 17 and 30 should be almost
exclusively in the form of the internal dication at physiological
pH. Compound 28 should also be a dication, but because of
resonance it should be a terminal dication. Thus the calcula-
tion of atomic charges and the conformational search of these
analogues were carried out based on the anticipated proto-
nated structures.
IC50 Deter m in a tion s. Cells were grown in 25 cm2 tissue
culture flasks in a total volume of 10 mL and treated while in
logarithmic growth (0.5-1.0 × 105 cells/mL) with polyamine
derivatives dissolved in sterile water and filtered through a
0.2 µm filter. For the 96 h IC50 determinations, cells were
reseeded after 48 h and incubated for an additional 48 h. Cells
were routinely counted electronically (Model ZF Coulter counter,
Coulter Electronics, Hialeah, FL) with this result confirmed
periodically by a hemocytometer. The percentage of control
growth was calculated as follows:
% of control growth )
final treated cell no. - initial inoculum
final untreated cell no. - initial inoculum
× 100
(
)
The IC50 is defined as the concentration of compound necessary
to reduce cell growth to 50% of control growth after defined
intervals of exposure.
P olya m in e P ool An a lysis. Cells in logarithmic growth
were treated with the polyamine derivatives. At the end of
the treatment period, cell suspensions were washed twice with
cold medium (RPMI-1640) and homogenized in 0.6 N perchloric
acid.32 This homogenate was stored frozen at -80 °C until
analysis by HPLC.
Ki Det er m in a t ion s. The polyamine derivatives were
studied for their ability to compete with [3H]SPD or [14C]SPD
uptake into L1210 leukemia cells in vitro.32 Cell suspensions
were incubated in 1 mL of RPMI-1640 containing 1, 2, 4, 6, 8,
The low-energy conformations of polyamine analogues were
obtained by systematic conformational searching supported by
Tripos Associates, Inc. in SYBYL 6.02 molecular modeling

2470 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13
Bergeron et al.
and 10 µM radiolabeled SPD alone or with the additional
presence of 10, 25, and 50 µM polyamine analogue for 20 min
at 37 °C. At the end of the incubation period, tubes were
centrifuged at 900g for 5 min at 0-4 °C, and the pellet was
washed twice with 5 mL of ice-cold RPMI-1640 containing 1
mM SPD. The washed pellet was dissolved in 200 µL of 1 N
NaOH at 60 °C for 1 h, neutralized with 1 N HCl, and
transferred to a vial for scintillation counting. Lineweaver-
Burke plots indicated a simple competitive inhibition with
respect to SPD [Km(SPD) ) 2.1 ( 0.2 µM].
NMDA Assa y. The procedure for assaying specific binding
of [3H]MK-801 to the N-methyl-D-aspartate receptor complex
was a modification of the method of Ransom and Stec.33
Cerebral cortices from young male Sprague-Dawley rats
(200-300 g) were homogenized with 10 vol of ice-cold 0.32 M
sucrose using a motor-driven glass/Teflon homogenizer. The
homogenate was centrifuged at 1000g for 10 min, and the
pellet was discarded. The supernatant was then centrifuged
at 18000g for 20 min. The 18000g pellet was then resuspended
in 10 vol of buffer A (5 mM Tris-HCl, pH 7.7) at 4 °C,
homogenized using high-intensity ultrasound, and centrifuged
at 8000g for 20 min. The supernatant and upper buffy coat
of the pellet were combined and centrifuged at 50000g for 20
min. The pellet was then resuspended in 10 vol of buffer A
and homogenized using high-intensity ultrasound. These
resuspended membranes were then centrifuged at 50000g,
discarding the supernatant. The membranes were washed in
this manner an additional three times and stored as a frozen
suspension at -80 °C for at least 18 h but no longer than 2
weeks before use.
Acu te a n d Ch r on ic Toxicity of P olya m in e An a logu es
in Mice. For acute toxicities, the polyamine analogues were
administered as a single ip injection. The animals were scored
2 h after administration of drug. All survivors were observed
for a period of 10 days after the treatment. In the chronic
toxicity regimen, mice were administered the polyamine
analogue as 1 ip dose/day for 5 days. Body weight and
observations of coat health and overall physical appearance
were recorded at the time of each injection. After the treat-
ment regimen had been completed, animals were examined
daily for an additional 10 days. The chronic toxicity dosing
regimen was set up so that at least one test group presented
with a high fraction of lethalities.
Ca st or Oil-In d u ced Dia r r h ea . Male Sprague-Dawley
rats (350-400 g) were fasted overnight in hanging wire cages
and allowed free access to water. A typical experiment
involved 12 rats: six untreated controls and six pretreated
with polyamine analogues (sc injection 30 min prior to castor
oil). All animals were then challenged with castor oil35 (gastric
gavage of 1 mL/200 g of body weight) at T ) 0 and monitored
for a 6 h period during which they received no food or water.
Onset of diarrhea for the control rats was between 30 and 90
min and lasted for at least 6 h. Time of diarrheal onset was
recorded for each animal in the control and treated groups,
and the animal weight and stool weight were recorded at 2, 4,
and 6 h.
4,9-Diben zyl-1,12-d ich lor o-(2S),(11S)-d ih yd r oxy-4,9-d i-
a za d od eca n e (2). MgSO4 (5 g) was added to a solution of
124 (5.45 g, 20.3 mmol) and (S)-(+)-epichlorohydrin (4.13 g, 44.7
mmol) in distilled methanol (120 mL). The reaction mixture
was stirred at room temperature for 2 days until completion
of the reaction as monitored by TLC. The solid was filtered,
and the filtrate was concentrated in vacuo to leave an oil. Flash
chromatography with 80/15/5 hexane/EtOAc/EtOH afforded
6.23 g (68%) of 2 as an oil: NMR δ 1.40 (m, 4 H), 2.50 (m, 8
H), 3.50 (m, 4 H), 3.56 (dd, 4 H, J ) 25.6, 1.2), 3.80 (m, 2 H),
7.27 (m, 10 H). Anal. (C24H34Cl2N2O2) C, H, N.
4,9-Diben zyl-1,12-d icya n o-(2R),(11R)-d ih yd r oxy-4,9-d i-
a za d od eca n e (3). A mixture of 2 (2.91 g, 6.4 mmol), KCN
(4.18 g, 64 mmol), and 18-crown-6 (0.17 g, 0.64 mmol) in dry
acetonitrile (80 mL) was heated at 60 °C for 2 days. The solid
was filtered, and the solvent was removed by rotary evapora-
tion. The residue was purified using flash column chroma-
tography with 60/35/5 hexane/EtOAc/EtOH to afford a solid,
which was recrystallized from 30/70 CH2Cl2/hexane to furnish
1.8 g (65%) of 3 as a white solid: mp 85 °C; NMR δ 1.46 (m,
4 H), 2.48 (m, 12 H), 3.60 (dd, 4 H, J ) 24.2, 4.5), 3.73 (m, 2
H), 7.29 (m, 10 H). Anal. (C26H34N4O2) C, H, N.
For binding experiments, frozen membranes were thawed,
pelleted at 50000g for 20 min, and washed as described above,
except that 20 vol of buffer A was used for resuspension, for a
total of four times. The final pellet was resuspended in buffer
B (5 mM Tris-HCI, pH 7.5, 23 °C). The binding assay mixture
was 1.00 mL of buffer B containing 200-300 µg of membrane
protein (Lowry method34), 100 µM L-glutamate, 100 µM glycine,
2 nM [3H]MK-801, and tetraamine at the following con-
centrations: 0 (“100% + L-Glu + Gly control”) and 0.05, 0.1,
0.5, 1, 5, 10, 50, 100, 500, and 1000 µM. Nonspecific binding
was determined using 100 µM MK-801. Binding assays were
performed in triplicate at 23 °C for 1 h and terminated by
filtration through Whatman GF/B glass fiber filters followed
by three 4.0 mL rinses of ice-cold buffer B using a Brandel
M-48 Cell Harvester.
Meta bolism of (HO)2DEHSP M vs DEHSP M in Mou se
Liver a n d Kid n ey. Groups of three female CD-1 mice
received a subchronic treatment regimen of either (a) a total
dose of 100 mg/kg (216 µmol/kg) (HO)2DEHSPM or (b) a total
dose of 90 mg/kg (208 µmol/kg) DEHSPM. Animals were
sacrificed for HPLC analysis 1 day after the final dose.
N5,N10-Diben zyl-(3R),(12R)-dih ydr oxyh om osper m in e (4).
W-2 grade Raney nickel (0.7 g) was added to a solution of 3
(1.42 g, 3.04 mmol) in methanol (100 mL) in a 500 mL Parr
bottle, and a slow stream of NH3 was bubbled through the
mixture for 20 min at 0 °C. Hydrogenation was carried out
with shaking at 50 psi for 7 h. The suspension was filtered
through Celite, and the solvents were evaporated in vacuo to
afford 1.12 g (86%) of 4 as an oil: NMR δ 1.41 (m, 4 H), 2.42
(m, 8 H), 2.82 (m, 8 H), 3.55 (d, 4 H, J ) 2.6), 3.71 (m, 2 H),
7.52 (m, 10 H); HRMS calcd for C26H42N4O2 443.3386 (M + 1),
found 443.3383 (M + 1).
HP LC An a lysis of P olya m in es in Mou se Tissu es. Vari-
ous tissues including the liver and kidney were prepared for
HPLC analysis of polyamine content. In order to facilitate
their handling, the organs were frozen in liquid nitrogen,
weighed, and homogenized (Tissuemizer, Tekmar, Cincinnati,
OH) in 1.2 N perchloric acid (containing 1,7-diaminoheptane
internal standard) in a 1:20 (w/v) ratio. The tissue homoge-
nates were then freeze-thawed three times and stored in a
-70 °C freezer until HPLC analysis was performed.
Each chromatographic assay included calibration standards
which were treated in the same manner as the samples. The
calibration standards (typical retention times in minutes are
indicated) were prepared by adding known amounts of PUT
(9.63), 1,6-diaminohexane (13.04), 1,7-diaminoheptane (15.39),
SPD (20.88), (HO)2HSPM (24.28), SPM (26.75), HSPM (27.60),
(HO)2DEHSPM (28.37), MEHSPM (29.80), and DEHSPM
(31.85) to a matrix that resembled the sample matrix. The
concentration of each polyamine was calculated from the peak
area by calibration curves obtained by nonweighted least-
squares linear regression. Peak area and linear regression
calculations were performed on a Macintosh Centris 650
instrument with Rainin Dynamax HPLC Method Manager
software (Rainin Instrument Co., Ridgefield, NJ ). The method
had a detection limit e 0.1 nmol/mL and was reproducible and
linear over the range 0.5-100 nmol/mL.
(3R),(12R)-Dia cetoxy-N1,N14-d ia cetyl-N5,N10-d iben zyl-
h om osp er m in e (5). A solution of 4 (2.4 g, 5.42 mmol) in CH2-
Cl2 (100 mL) was treated with acetic anhydride (10 mL) at
room temperature for 3 h. The volatiles were evaporated in
vacuo, and the concentrate was purified by flash chromatog-
raphy with 65/25/15 hexane/EtOAc/EtOH to give 2.39 g (73%)
of 5 as an oil: NMR δ 1.44 (m, 4 H), 1.96 (m, 3 H), 1.98 (m, 4
H), 2.07 (s, 3 H), 2.46 (m, 8 H), 2.57 (m, 2 H), 3.02 (m, 2 H),
3.42 (m, 2 H), 3.59 (m, 4 H), 7.28 (m, 10 H). Anal. (C34H50N4O6)
C, H, N.
N5,N10-Dib en zyl-N1,N14-d iet h yl-(3R),(12R)-d ih yd r oxy-
h om osp er m in e (6). LiAlH4 (1 M in THF, 15 mL, 15 mmol)
was added to a solution of 5 (1.73 g, 2.83 mmol) in dry THF
(50 mL). The mixture was stirred at 65 °C for 3 h under N2.
The reaction was cautiously quenched with water (5 mL)
followed by filtration of the solids. Evaporation of the filtrate
and flash chromatography on silica gel, eluting with 5%
concentrated NH4OH in methanol, produced 0.6 g (44%) of 6

Metabolically Programmed Polyamine Antidiarrheals
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 13 2471
(14) Belair, E. J .; Carlson, G. R.; Melamed, S.; Moss, J . N.; Tansy,
M. F. Effects of Spermine and Spermidine on Gastric Emptying
in Rats. J . Pharm. Sci. 1981, 70, 347.
as an oil: NMR δ 1.09 (m, 6 H), 1.47 (m, 8 H), 2.64 (m, 16 H),
3.59 (m, 4 H), 3.75 (m, 2 H), 7.28 (m, 10 H). Anal. (C30H50N4O2)
C, H, N.
(15) Sato, T. L.; Sninsky, C. A.; Bergeron, R. J . Structural Specificity
of Synthetic Analogues of Polyamines and their Effect on
Gastrointestinal Motility. In Polyamines and the gastrointestinal
tract, Falk Symposium No. 62; Dowling, R. H., Folsch, U. R.,
Loser, C., Eds.; Kluwer Academic Publishers: Boston, MA, 1991.
(16) Sninsky, C. A.; Bergeron, R. Potent Anti-Diarrheal Activity of a
New Class of Compounds: Synthetic Analogs of the Polyamine
Pathway. Gastroenterology 1993, 104, A584.
N1,N14-Dieth yl-(3R),(12R)-dih ydr oxyh om osper m in e Tet-
r a h yd r och lor id e [(HO)2DEHSP M] (7). Pd-C (10%, 0.4 g)
was added to a solution of 6 (2.26 g, 4.53 mmol) in concentrated
HCl (3 mL) and EtOH (100 mL), and the suspension was
degassed three times with N2. After stirring under hydrogen
(1 atm, 3 h), the catalyst was filtered off and washed with
water (10 mL). The solvents were removed under reduced
pressure to give a white solid. Recrystallization from aqueous
EtOH gave 1.5 g (72%) of 7 as a crystalline solid: NMR (D2O)
δ 1.07 (t, 6 H, J ) 2.4), 1.60 (m, 4 H), 1.74 (m, 4 H), 2.91 (m,
16 H), 3.84 (m, 2 H). Anal. (C16H42Cl4N4O2) C, H, N.
(3R),(12R)-Dih yd r oxyh om osp er m in e Tetr a h yd r och lo-
r id e [(HO)2HSP M] (8). HCl (1 N, 10 mL) and 10% Pd-C
(0.08 g) were added to a solution of 4 (0.764 g, 1.73 mmol) in
methanol (100 mL), and the suspension was flushed three
times with nitrogen. The mixture was exposed to hydrogen
for 3 h at atmospheric pressure followed by filtration of catalyst
on Celite. The filtrate was evaporated in vacuo to give a white
solid. Recrystallization from aqueous ethanol produced 0.56
g (80%) of 8 as a crystalline solid: NMR (D2O) δ 1.63 (m, 8
H), 2.97 (m, 12 H), 3.92 (m, 2 H). Anal. (C12H34Cl4N4O2) C,
H, N.
(17) Bergeron, R. J .; Weimar, W. R.; Luchetta, G.; Sninsky, C. A.;
,N14-Dieth-
Wiegand, J . Metabolism and Pharmacokinetics of N1
ylhomospermine. Drug Metab. Dispos. 1995, 23, 334-343.
(18) Bergeron, R. J .; Weimar, W. R.; Luchetta, G.; Streiff, R. R.;
Wiegand, J .; Perrin, J .; Schreier, K. M.; Porter, C.; Yao, G. W.;
Dimova, H. Metabolism and Pharmacokinetics of N1,N11-Dieth-
ylnorspermine. Drug Metab. Dispos. 1995, 23, 1117-1125.
(19) SYBYL, Version 602, Tripos Associates, Inc., St. Louis, MO; J uly
21, 1993.
(20) Gasteiger, J .; Marsili, M. Interactive Partial Equalization of
Orbital Electronegativity-A Rapid Access to Atomic Charges.
Tetrahedron 1980, 36, 3219-3228.
(21) Cramer, R. D., III; Bunce, J . D.; Petterson, D. E.; Frank, I. E.
Crossvalidation, Bootstrapping, and Partial Least Squares
Compared with Multiple Regression in Conventional QSAR
Studies. Quant. Struct.-Act. Relat. Pharmacol. Chem. Biol. 1988,
7, 18-25.
(22) Wold, S. Cross-validatory Estimation of the Number of Compo-
nents in Factor and Principal Components Models. Technomet-
rics 1978, 20, 397-405.
Ack n ow led gm en t. Financial support was provided
by SunPharm Corp., J acksonville, FL, and National
Institutes of Health, Grant No. CA37606.
(23) Bergeron, R. J .; Xia, M. X. B.; Phanstiel, O., IV. Total Synthesis
of (+)-Hypusine and Its (2S,9S)-Diastereomer. J . Org. Chem.
1993, 58, 6804-6806.
(24) Samejima, K.; Takeda, Y.; Kawase, M.; Okada, M.; Kyogoku, Y.
Synthesis of 15N-Enriched Polyamines. Chem. Pharm. Bull.
1984, 32, 3428-3435.
(25) Bergeron, R. J .; Weimar, W. R.; Wu, Q.; Austin, J . K., J r.;
McManis, J . S. Impact of Polyamine Analogues on the NMDA
Receptor. J . Med. Chem. 1995, 38, 425-428.
(26) Pegg, A. E.; Wechter, R.; Pakala, R.; Bergeron, R. J . Effect of
N1,N12-Bis(ethyl)spermine and Related Compounds on Growth
and Polyamine Acetylation, Content, and Excretion in Human
Colon Tumor Cells. J . Biol. Chem. 1989, 264, 11744-11749.
(27) Seiler, N. Acetylation and Interconversion of the Polyamines.
In The Physiology of Polyamines; Bachrach, U., Heimer, Y. M.,
Eds.; CRC Press: Boca Raton, FL, 1989; Vol. I, Chapter 11, pp
159-176.
Refer en ces
(1) Gillin, J . S.; Shike, M.; Alcock, N.; Urmacher, C.; Krown, S.;
Kurtz, R. C.; Lightdale, C. J .; Winawer, S. J . Malabsorption and
Mucosal Abnormalities of the Small Intestine in the Acquired
Immunodeficiency Syndrome. Ann. Intern. Med. 1985, 102, 619-
622.
(2) Chandrasekar, P. H. “Cure” of Chronic Cryptosporidiosis During
Treatment with Azidothymidine in a Patient with the Acquired
Immune Deficiency Syndrome. Am. J . Med. 1987, 83, 187.
(3) Simon, D.; Weiss, L.; Tanowitz, H. B.; Wittner, M. Resolution
of Cryptosporidium Infection in an AIDS Patient after Improve-
ment of Nutritional and Immune Status with Octreotide. Am.
J . Gastroenterol. 1991, 86, 615-618.
(4) Abrams, D. I.; Goldman, A. I.; Launer, C.; Korvick, J . A.; Neaton,
J . D.; Crane, L. R.; Grodesky, M.; Wakefield, S.; Muth, K.;
Kornegay, S.; Cohn, D. L.; Harris, A.; Luskin-Hawk, R.; Markow-
itz, N.; Sampson, J . H.; Thompson, M.; Deyton, L. A Comparative
Trial of Didanosine or Zalcitabine after Treatment with Zidovu-
dine in Patients with Human Immunodeficiency Virus Infection.
N. Engl. J . Med. 1994, 330, 657-662.
(28) Nakao, J .; Hasegawa, T.; Hashimoto, H.; Noto, T.; Nakajima,
T. Formation of GABOB from 2-Hydroxyputrescine and Its
Anticonvulsant Effect. Pharmacol. Biochem. Behav. 1991, 40,
359-366.
(29) Bergeron, R. J .; McManis, J . S.; Weimar, W. R.; Schreier, K. M.;
Gao, F.; Wu, Q.; Ortiz-Ocasio, J .; Luchetta, G. R.; Porter, C.;
Vinson, J . R. T. The Role of Charge in Polyamine Analogue
Recognition. J . Med. Chem. 1995, 38, 2278-2285.
(30) Bergeron, R. J .; McManis, J . S.; Liu, C. Z.; Feng, Y.; Weimar,
W. R.; Luchetta, G. R.; Wu, Q.; Ortiz-Ocasio, J .; Vinson, J . R.
T.; Kramer, D.; Porter, C. Antiproliferative Properties of
Polyamine Analogues: A Structure-Activity Study. J . Med.
Chem. 1994, 37, 3464-3476.
(5) Lopez, A. P.; Gorbach, S. L. Diarrhea in AIDS. Infect. Dis. Clin.
North Am. 1988, 2, 705-718.
(6) Antony, M. A.; Brandt, L. J .; Klein, R. S.; Bernstein, L. H.
Infectious Diarrhea in Patients with AIDS. Dig. Dis. Sci. 1988,
33, 1141-1146.
(7) Simon, D.; Weiss, L. M.; Brandt, L. J . Treatment Options for
AIDS-Related Esophageal and Diarrheal Disorders. Am. J .
Gastroenterol. 1992, 87, 274-281.
(31) Bergeron, R. J .; Neims, A. H.; McManis, J . S.; Hawthorne, T.
R.; Vinson, J . R. T.; Bortell, R.; Ingeno, M. J . Synthetic
Polyamine Analogues as Antineoplastics. J . Med. Chem. 1988,
31, 1183-1190.
(8) Grosman, I.; Simon, D. Potential Gastrointestinal Uses of
Somatostatin and Its Synthetic Analogue Octreotide. Am. J .
Gastroenterol. 1990, 85, 1061-1072.
(32) Bergeron, R. J .; Hawthorne, T. R.; Vinson, J . R. T.; Beck, D. E.,
J r.; Ingeno, M. J . Role of the Methylene Backbone in the
Antiproliferative Activity of Polyamine Analogues on L1210
Cells. Cancer Res. 1989, 49, 2959-2964.
(9) Farthing, M. J . G. Octreotide in the Treatment of Refractory
Diarrhea and Intestinal Fistulae. Gut 1994, 35 Suppl. 3, S5-
10.
(10) Powell, D. W.; Szauter, K. E. Nonantibiotic Therapy and
Pharmacotherapy of Acute Infectious Diarrhea. Gastroenterol.
Clin. North Am. 1993, 22, 683-707.
(11) Tansy, M. F.; Martin, J . S.; Landin, W. E.; Kendall, F. M.;
Melamed, S. Spermine and Spermidine as Inhibitors of Gas-
trointestinal Motor Activity. Surg. Gynecol. Obstet. 1982, 154,
74-80.
(12) Melamed, S.; Carlson, G. R.; Moss, J . N.; Belair, E. J .; Tansy,
M. F. GI Pharmacology of Polyethyleneimine I: Effects on
Gastric Emptying in Rats. J . Pharm. Sci. 1977, 66, 899-901.
(13) Tansy, M. F.; Martin, J . S.; Innes, D. L.; Kendall, F. M.;
Melamed, S.; Moss, J . N. GI Pharmacology of Polyethyleneimine
II: Motor Activity in Anesthetized Dogs. J . Pharm. Sci. 1977,
66, 902-904.
(33) Ransom, R. W.; Stec, N. L. Cooperative Modulation of [3H]MK-
801 Binding to the N-Methyl-D-Aspartate Receptor-Ion Channel
Complex by L-Glutamate, Glycine, and Polyamines. J . Neuro-
chem. 1988, 51, 830-836.
(34) Lowry, O. H.; Rosebrough, N. J .; Farr, A. L.; Randall, R. J .
Protein Measurement with the Folin Phenol Reagent. J . Biol.
Chem. 1951, 193, 265-275.
(35) Eaker, E. Y.; Bixler, G. B.; Mathias, J . R. WHR 1049, a Potent
Metabolite of Lidamidine, Has Antidiarrheal and Antimotility
Effects on the Small Intestine in Rats. J . Pharmacol. Exp. Ther.
1988, 246, 786-789.
J M950827H