Selective inhibition of the cellular sodium pump by emicymarin and 14? anhydroxy bufadienolides

Article abstract of DOI:10.1016/j.steroids.2010.07.010

Partial inhibition of the sodium pump (Na/K-ATP-ase) by a circulating inhibitor is known to occur in humans. The objectives of this study were to determine the effects of novel bufadienolides lacking an oxygen at C14 on sodium pumps in human erythrocytes and leucocytes, dog kidney and pig brain and to document the importance of the stereochemistry at C17 on the ability to inhibit these sodium pumps. 14α bufadienolides were weak inhibitors of all preparations studied. 3?-OH,5?,14? bufadienolide produced near-total inhibition of dog kidney and pig brain Na/K-ATP-ase. Over the same concentration range, it maximally inhibited the sodium pump of erythrocytes by 70% and leucocytes by 47%. The inhibition profile induced in the leucocyte sodium pump deviated significantly from the simple sigmoidal relationship present in the other preparations over the 3 × 10^-5 to 1 × 10^-7 mol/l concentration range. Allo-emicymarin (17α) was confirmed to be a weak inhibitor of the sodium pump/ATP-ase compared with emicymarin (17?) but both were weaker inhibitors of the leucocyte sodium pump than that of the other preparations. Molecules with the C14 in the ? configuration are more efficacious than in the α configuration. In the case of emicymarin, the attachment of the furone at C17 in the α configuration results in substantially weaker inhibitory activity than in the beta configuration, seen in most cardenolides and bufadienolides. Unlike ouabain and bufalin that show no specificity of action in these preparations, 3?- OH,5?,14? bufadienolide selectively inhibits the activity of at least one low-prevalence subset of the leucocyte Na/K-ATP-ase.

Full text of DOI:10.1016/j.steroids.2010.07.010

Steroids 75 (2010) 1137–1145
Contents lists available at ScienceDirect
Steroids
journal homepage: www.elsevier.com/locate/steroids
Selective inhibition of the cellular sodium pump by emicymarin and 14ß
anhydroxy bufadienolides
Philip J. Hiltona,b,∗, William McKinnona, Edward C. Gravettb, Jean-Marie R. Peronb,
Christopher M. Framptonc, M. Gary Nichollsc, Gwyn Lordd
a Renal Laboratory, St Thomas’ Hospital, London, SE1 7EH, United Kingdom
b School of Pharmacy and Chemistry, Faculty of Science, Kingston-upon-Thames, Surrey, KT1 2EE, United Kingdom
c Department of Medicine, Christchurch Hospital, Christchurch, New Zealand
d School of Biological and Chemical Sciences, Birkbeck, University of London, WC1E 7HX, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Partial inhibition of the sodium pump (Na/K-ATP-ase) by a circulating inhibitor is known to occur in
humans. The objectives of this study were to determine the effects of novel bufadienolides lacking an
oxygen at C14 on sodium pumps in human erythrocytes and leucocytes, dog kidney and pig brain and
to document the importance of the stereochemistry at C17 on the ability to inhibit these sodium pumps.
14␣ bufadienolides were weak inhibitors of all preparations studied. 3ß-OH,5ß,14ß bufadienolide pro-
duced near-total inhibition of dog kidney and pig brain Na/K-ATP-ase. Over the same concentration
range, it maximally inhibited the sodium pump of erythrocytes by 70% and leucocytes by 47%. The inhi-
bition profile induced in the leucocyte sodium pump deviated significantly from the simple sigmoidal
relationship present in the other preparations over the 3 × 10−5 to 1 × 10−7 mol/l concentration range.
Allo-emicymarin (17␣) was confirmed to be a weak inhibitor of the sodium pump/ATP-ase compared
with emicymarin (17ß) but both were weaker inhibitors of the leucocyte sodium pump than that of the
other preparations. Molecules with the C14 in the ß configuration are more efficacious than in the ␣ con-
figuration. In the case of emicymarin, the attachment of the furone at C17 in the ␣ configuration results
in substantially weaker inhibitory activity than in the beta configuration, seen in most cardenolides and
bufadienolides. Unlike ouabain and bufalin that show no specificity of action in these preparations, 3ß-
OH,5ß,14ß bufadienolide selectively inhibits the activity of at least one low-prevalence subset of the
leucocyte Na/K-ATP-ase.
Received 10 December 2009
Received in revised form 30 June 2010
Accepted 22 July 2010
Available online 3 August 2010
Keywords:
Sodium pump
ATP-ase
Bufadienolide
Ouabain
© 2010 Elsevier Inc. All rights reserved.
1. Introduction
is with ␣-pyrone (2H-pyran-2-one) and whose commonest natu-
ral source is the venom of toads, though they are also found in
The membrane sodium pump (Na/K-ATP-ase) of animals is
known to be inhibited by two principal groups of compound. The
first of these, the cardenolides, are steroids with furone substitu-
tion at C17 and are principally found in a restricted group of plants.
Long before their structure or mechanism of action was under-
stood, they were used medicinally for the control of the ventricular
response to atrial fibrillation and as inotropic agents. However,
this medicinal use repeatedly demonstrated their essentially poi-
sonous nature since a significant proportion of patients so treated
developed toxic symptoms which are now recognised as represent-
ing excessive inhibition of the sodium pump. The second group of
compounds, the bufadienolides, differ in that the C17 substitution
a small range of plants. These have never achieved a therapeutic
role though both accidental and deliberate intoxication are recog-
nised and show considerable overlap with cardenolide toxicity
[1].
The ability of mammals to secrete a substance that inhibits
the membrane sodium pump is well recognised. Apart from its
physiological interest, partial inhibition of the sodium pump (Na/K-
ATP-ase) caused by a circulating substance is known to occur in
humans, having been reported in uraemia [2,3], volume-expansion
[4] and during the neonatal period [5]. However, arguably the great-
est impetus to this branch of research has been the association
of sodium pump inhibition with essential hypertension [6]. Some
attempts to characterise the circulating inhibitor of the sodium
pump in man have claimed that it may be the cardenolide, ouabain
[7]. We have suggested recently that evidence for ouabain being
synthesised and secreted into the circulation by the human adrenal
is insecure [8] though this perspective is disputed [9].
∗
Corresponding author at: Renal Laboratory, St. Thomas’ Hospital, Lambeth
Palace Road, London SE1 7EH, United Kingdom.
E-mail address: renallab@gmail.com (P.J. Hilton).
0039-128X/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2010.07.010

1138
P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
Fig. 1. 1HNMR of 3␤-OH,5␤,14␤ bufadienolide.
Various techniques for measuring sodium pump activity have
2. Experimental
been employed in different laboratories, but the most consistent
model showing sodium pump inhibition in essential hypertension
has been the in vitro preparation of human leucocytes [10]. The
average inhibition seen in leucocytes obtained from hypertensives
is by approximately 10% compared to control cells obtained from
normotensives and it may be calculated that the inhibition shown
in these studies would correspond to a ouabain concentration in the
extracellular fluid of over 10−8 mol/l. The concentration of digoxin
at which clinical toxicity is generally evident is approximately
4 nmol/l and the limited evidence on ouabain toxicity suggests
that both its therapeutic and toxic concentrations are similar to
those of digoxin [11]. It is reasonable to conclude that clinical
toxicity would be evident if a relatively unselective inhibitor of
the sodium pump such as ouabain were to be responsible for the
observed alterations of sodium transport seen in essential hyper-
tension.
One argument that might explain this paradox is the discov-
ery that different tissues contain different isoforms of the sodium
pump. It has been proposed that these isoforms confer ouabain
selectivity among the various tissues [12]. To the best of our knowl-
edge, however, this has never been demonstrated convincingly in
humans, though it is clearly present in rodents [13].
In an attempt to identify more selective inhibitors of the mam-
malian sodium pump, we synthesised bufadienolides lacking the
14–OH group and compared their effects with ouabain and bufalin
on sodium pumps in four tissues. We also had the opportu-
nity to study the effects on the sodium pump of a change in
the stereochemistry at position C17 using emicymarin and allo-
emicymarin.
2.1. Synthesis of bufadienolides
The synthesis of 14␣ bufadienolides and 3␤-OH,5␤,14␤ bufa-
dienolide is described in detail in the Appendix A. Each synthetic
step was monitored by 1HNMR and by the determination of the
accurate mass of the intermediate product (Q-ToF mass spec-
trometry). The accurate mass of the protonated end-product
(3␤-OH,5␤,14␤ bufadienolide) was 371.2585 which corresponds
to the required elemental formula C24H34O3 (C 77.80; H 9.25; O
12.95). Its 1HNMR spectrum is shown in Fig. 1 and the assignation
of peaks in Table 1. Its melting point was 196–198 ◦C. The structure
of 3␤-OH,5␤,14␤ bufadienolide is shown in Fig. 2 alongside those
of better-known Na/K-ATP-ase inhibitors the plant-derived poison
ouabain, the toad venom bufalin, emicymarin and allo-emicymarin.
It should be noted that 3␤-OH,5␤,14␤ bufadienolide is essentially
the toad venom bufalin with the C14–OH group replaced with a
hydrogen atom.
2.2. Sodium transport in intact cells
Blood samples were obtained from laboratory staff who gave
their informed consent to the procedure. Sodium efflux in human
erythrocytes was studied by the method of Walter and Distler [14].
10 ml of blood was drawn from a forearm vein into a heparinised
syringe. The blood was transferred to a test-tube and the erythro-
cytes isolated by centrifugation. The buffy coat was removed and
the cells washed in physiological buffer. The erythrocytes were
loaded with radioactive sodium by incubation in physiological

P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
1139
buffer containing 22NaCl for 30 min. The cells were then washed
three times in non-radioactive buffer and incubated at 37 ◦C with
the sample under test for 90 min. Aliquots were removed at inter-
vals of 15 min. These aliquots were immediately chilled to 0 ◦C and
spun at 1200 × g for 3 min. The supernatant was removed and the
amount of 22Na in the supernatant determined in a gamma counter
(Wallac, 1282, PerkinElmer, Beaconsfield, Bucks.) The increase in
radioactivity of the supernatant was related back to the radioactiv-
ity of a sample of cells removed at time 0.
Sodium efflux in human leucocytes was studied by the method
previously described [15] with minor modifications. Leucocytes
were obtained from 100 ml of blood drawn from a forearm vein into
a heparinised syringe. Leucocytes were isolated by mixing the blood
with Dextran T500 then allowing the erythrocytes to sediment
under gravity. The supernatant, consisting of leucocytes contam-
inated with some erythrocytes, was removed and the cells spun
from the supernatant at 250 × g for 5 min, followed by lysis of the
residual erythrocytes by exposure to water for 13 s before restoring
normal osmolality by the addition of hypertonic buffer. The leuco-
cytes were loaded in a 22NaCl-containing buffer and washed, before
incubation with the sample under test for 16 min at 37 ◦C. Aliquots
of the cell suspension were removed every 4 min. Once withdrawn,
the cell aliquot was immediately chilled to 0 ◦C before being spun
for 1 min at 3200 × g. The supernatant was removed and the resid-
ual 22Na in the cell pellet determined in a gamma counter. The cell
pellet was digested in NaOH (1 mol/l) and the counts normalised
for the protein content as determined by the Folin method.
In both cases, 22Na efflux (and hence Na/K ATP-ase activity)
was calculated as a pseudo-first order rate constant. The ouabain-
sensitive portion of the flux was derived by subtraction of the
residual sodium efflux seen in the presence of 1 × 10−4 mol/l
ouabain. The effect of the synthetic bufadienolides and of the emi-
cymarins was studied over a range of concentrations from 1 × 10−11
to 1 × 10−4 mol/l.
A separate experiment to compare the binding of ouabain and
3␤-OH,5␤,14␤ bufadienolide on the ATP-ase present in human ery-
throcytes was undertaken. Previous experiments had shown that
5 min incubation with the inhibitors utilised in this study was
Fig. 2. Structures of ouabain, bufalin, emicymarin, allo-emicymarin and 3␤-OH, 5␤,14ß bufadienolide.

1140
P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
the presence of 1 × 10−4 mol/l ouabain. Inorganic phosphate was
Table 1
measured in aliquots of the sample solution withdrawn every 7 min
by the method of Kraml [16].
All the studies were performed a minimum of three times
at each concentration studied. The data for each point are pre-
sented as the mean the standard error. Sigmaplot software was
used to fit sigmoid curves to estimate the slopes and EC50 for the
dose/inhibition relationships. These parameters were then com-
pared between compounds using the z-test computed from the
parameter estimates and the standard errors of these estimates.
3. Results
.
3.1. Effects of 14˛ bufadienolides on sodium pump activity
1H NMR data in d-chloroform for 3ß-OH,5ß,14ß bufadienolide
ı (ppm), J (Hz)
EG155
Six 14␣ bufadienolides were studied utilising the available
derivatives of androsterone which were coupled as described to ␣
pyrone (2H-pyran-2-one) by the Suzuki method [17] (Appendix A).
The results of four representative compounds, presented in Table 2,
show that 14␣ bufadienolides that lack an oxygen function at this
position are weak inhibitors of Na/K-ATP-ase (for clarity only the
values for dog kidney ATP-ase are presented).
3
17␣
18
19
21
4.11 (p, J = 2.8)
2.37 (dd, J17,16 = 9.4, J17,16 = 10.5)
0.90 (s)
0.91 (s)
7.25 (ddd, J21,22 = 2.6, J21,23 = 1.1, J21,17 = 0.5)
7.31 (dd, J22,23 = 9.7, J22,21 = 2.7)
6.25 (dd, J23,22 = 9.7, J23,21 = 1.1)
22
23
Others
1.95–0.81 (envelope+3-OH)
3.2. Effects of 3ˇ-OH,5ˇ,14ˇ bufadienolide on sodium pump
activity
13C NMR data in d-chloroform for 3ß-OH,5ß,14ß bufadienolide
ı (ppm)
Carbon**
ı (ppm)
Carbon**
The effect of 3␤-OH,5␤,14␤ bufadienolide on the activity of
the four sodium pump preparations used in this study is shown
in Fig. 3. For comparison, the typical pattern of inhibition by the
plant-derived poison, ouabain, and the toad venom, bufalin on the
sodium pump preparations utilised in this study is shown in Fig. 4.
As can be seen, ouabain and bufalin inhibit the sodium pumps with
similar sigmoidal profiles (slopes not significantly different) and
almostidentical inhibitorymaxima, although bufalin was moreeffi-
cacious than an equimolar concentration of ouabain (differences in
EC50 bufalin versus ouabain, all p < 0.001). The sigmoidal inhibitory
profile was still present when both dog kidney and pig brain ATP-
ase were individually exposed to 3␤-OH,5␤,14␤ bufadienolide,
but, as observed when comparing ouabain to bufalin, an increase
in the concentration of 3␤-OH,5␤,14␤ bufadienolide required to
achieve 50% inhibition (EC50) was required. 3␤-OH,5␤,14␤ bufa-
dienolide was thus less effective than either ouabain or bufalin
on an equimolar basis. Despite this, 3␤-OH,5␤,14␤ bufadienolide
maximally inhibited dog kidney ATP-ase by 88 3% and pig brain
ATP-ase by 87 1.6%.
162.33
148.90
145.25
118.23
115.27
67.07
52.15
51.71
43.94
36.69
35.04
34.29
24-C
21-CH
22-CH
20-C
23-CH
3-CH
17-CH
CH
C
CH
C
CH
O
33.43
31.95
29.67
28.47
28.10
26.71
25.96
25.69
23.93
23.02
22.72
21.02
CH2
CH
CH2
CH2
CH2
CH2
CH2
16-CH2
I9-CH3
I8-CH3
CH2
CH2
*Data acquired at 400 MHz in CDC13 (8H = 7.26 ppm) on a Jeol ECP400 FT-NMR
instrument.
**Data acquired at 100 MHz in CDC13 (8C = 77.16 ppm) on a Jeol ECP400 FT-NMR
instrument.
1Hand 13C assignedusingDEPT-135, 1H–1HFG DQF Cosyand CH–HetCor correlation.
more than sufficient to permit maximal binding of the inhibitor
to the sodium pumps. Human erythrocytes loaded with 22Na
were incubated in concentrations of ouabain (1 × 10−5 mol/l) or
3␤-OH,5␤,14␤ bufadienolide (1 × 10−4 mol/l) for 15 min, to give
maximal inhibition of the Na/K-ATP-ase. To measure binding a sec-
ond batch of cells similarly loaded with 22Na was exposed to the
inhibitor. These cells were washed three times in approximately
10 volumes of non-radioactive, inhibitor- free buffer. The sodium
efflux was then measured as previously described and the pre
and post wash inhibition results for ouabain compared with those
obtained from 3␤-OH,5␤,14␤ bufadienolide.
2.3. Na/K ATP-ase studies
Dog kidney and pig brain Na/K ATP-ases were obtained from
Sigma-Aldrich. The effect of various concentrations of the syn-
thetic compounds on 0.05 U of each Na/K ATP-ase was studied
by measuring the release of inorganic phosphate from ATP over
a 28 min period at 37 ◦C in a medium containing Na+ 100 mmol/l,
K+ 3 mmol/l, Mg2+ 4.5 mmol/l, ATP 0.3 mmol/l, EDTA 0.1 mmol/l,
EGTA 5 mmol/l, buffered to pH 7.40 with Tris (40 mmol/l) and the
substance under test. The ouabain-sensitive portion of the flux was
calculated by subtraction of the residual phosphate generation in
Fig. 3. The effect of 3␤-OH,5␤,14␤ bufadienolide on various sodium pumps.

P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
1141
Table 2
The maximum inhibitory effect of four synthetic 14␣ anhydroxy bufadienolides on dog kidney ATP-ase at 10−4 mol/1 (the approximate limit of solubility in water).
Compound
3␣-OH, 5␣, 14␣ bufadienolide
3␤-OH, 5␣, 14␣ bufadienolide
3␤-OH, 5␣, 14␣ bufadienolide
3␤-OH, 5:6-ene, 14␣ bufadienolide
Inhibition @ 1 × 10−4
M
0%
46%
86%
30%
The inhibitory profile obtained from the exposure of Na/K
ATP-ase present in erythrocytes to 3␤-OH,5␤,14␤ bufadienolide
appeared similar to that of ouabain and bufalin in its classical sig-
moidal form. However, exposure to the maximal concentration of
3␤-OH,5␤,14␤ bufadienolide studied inhibited the ATP-ase present
in erythrocytes by only 68 4.3% At this concentration, the inhibi-
tion of erythrocyte sodium efflux appeared to be reaching a plateau.
The effect of 3␤-OH,5␤,14␤ bufadienolide could not be tested at
concentrations in excess of 1 × 10−4 mol/l as this represented the
limit of its solubility in water.
The inhibitory profile obtained by the exposure of Na/K ATP-
ase present in human leucocytes to 3␤-OH,5␤,14␤ bufadienolide
deviated from the classical sigmoidal profile observed in pig brain
Na/K ATP-ase, dog kidney Na/K ATP-ase and erythrocyte Na/K ATP-
ase (Fig. 3). Examination of the leucocyte Na/K ATP-ase response
to 3␤-OH,5␤,14␤ bufadienolide showed that maximal inhibition
was attained by a concentration of bufadienolide of 1 × 10−5 mol/l.
At this point only 47 6.1% of the total leucocyte sodium pumps
were inhibited, i.e. approximately 53% of the sodium pumps in
human leucocytes were not inhibited by 3␤-OH,5␤,14␤ bufadieno-
lide. Examination of the dose–response curve for leucocyte Na/K
ATP-ase after exposure to 3␤-OH,5␤,14␤ bufadienolide in the con-
centration range of 3 × 10−8 to 1 × 10−6 mol/l showed there was
further significant deviation from the simple sigmoidal inhibitory
relationship seen in the other Na/K ATP-ase preparations studied.
Indeed, zero inhibition of the Na/K-ATP-ase present in leuco-
cytes was not observed until 3␤-OH,5␤,14␤ bufadienolide was
reduced to a concentration of 3 × 10−8 mol/l. Mathematical analy-
sis of the relationship between the concentration of 3␤-OH,5␤,14␤
bufadienolide and the inhibitory effect on ATP-ase present in leu-
cocytes shows that the data can be deconvoluted and resolved into
two component sigmoidal curves (harmonics) the sum of which
produces a curve that closely fits the inhibitory profile obtained
experimentally (Fig. 5). The presence of an additional point of
inflexion at a concentration of 3␤-OH,5␤,14␤ bufadienolide of
approximately 1 × 10−7 mol/l may indicate that at least two sub-
types of ATP-ase are present, with each have differing sensitivities
to 3␤-OH,5␤,14␤ bufadienolide.
Fig. 5. The component harmonics composing the inhibitory profile of 3␤-
OH,5␤,14␤ bufadienolide using human leucocytes. Curve 2 between 10−8 and
3 × 10−7 mol/l has been adjusted upwards by 0.6% to prevent its being rendered
invisible by the combined curve.
Fig. 4. The effect of ouabain and bufalin on various sodium pumps.

1142
P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
importantly C14. All the known cardenolides and bufadienolides
which function principally if not exclusively as poisons have an
oxygen function (–OH or epoxide) at C14.
The question of whether the erythrocytes of hypertensive indi-
viduals also exhibited the sodium pump inhibition consistently
found in leucocytes has never been satisfactorily answered, with a
wide divergence of results between different laboratories. A review
of the then available data in 1985 [10] suggested that in aggregate,
inhibition of erythrocyte sodium pump in human hypertension had
not been convincingly demonstrated. Studies in our own labora-
tory of leucocyte and erythrocyte sodium transport derived from
the same blood samples of a group of hypertensives failed to show
inhibition in the erythrocytes despite reproducing the previously
recognised inhibition of the pump in the leucocytes [19], a finding
confirmed subsequently by Boon et al. [20] in respect of cellular
sodium content. This suggested that differences in the behaviour of
subsets of the sodium pumps of the two cell types might underlie
this observation.
We therefore examined whether removing the –OH from C14
would allow for selective inhibition of different sodium pump
preparations particularly in human erythrocytes and leucocytes.
To begin we synthesised and characterised the effects of a range
of 14␣ anhydroxy bufadienolides before moving on to 14␤ anhy-
droxy bufadienolides. Professor I M Glynn kindly allowed us to
characterise the effect of isomerisation at the C17 position of the
cardenolides emicymarin and allo-emicymarin by the gift of these
compounds. Glynn had reported earlier the effect of these com-
pounds on human erythrocytes [21] and had himself received the
isomers as a gift from Professor Reichbach (Basel). The compounds
had been stored in sealed glass vials and when examined by LC–MS,
LC–QToF MS and 1H NMR proved to be both pure and to have the
requisite structure.
Fig. 6. Effect of emicymarin and allo-emicymarin on various ATP-ases. The solid
curves are those for emicymarin, the broken ones are for allo-emicymarin.
3.3. Effects of emicymarin and allo-emicymarin on sodium pump
activity
For all four preparations the dose–response curves relating
concentrations of the two cardenolides to sodium pump activity
were shifted markedly to the left for allo-emicymarin compared to
emicymarin (p < 0.001, Fig. 6). Thus, for example, the EC50 for emi-
cymarin and for allo-emicymarin differed significantly (p < 0.001) in
all four preparations. For both of these cardenolides, their inhibitory
action on leucocytes was substantially less than in the other three
preparations (p < 0.001, Fig. 6).
As anticipated, the 14␣ anhydroxy bufadienolides proved to be
weak inhibitors of the sodium pump/ATP-ase (Table 2) with 3␣-OH,
5␣, 14␣ bufadienolide showing no detectable inhibition in any of
the four preparations. The inhibitory activity of the three other 14␣
compounds in Table 2 was weak and demonstrable only at concen-
trations close to saturation. For this reason, no attempt to calculate
an EC50 has been made for these compounds though there is a sug-
gestion that the 5␤ configuration is somewhat more effective an
inhibitor than either the 5␣ or 5:6-ene. Arguably none showed a
sufficiently large effect to be of either physiological or pharmaco-
logical interest. They did, however, demonstrate the feasibility of
synthesising these compounds from readily available precursors
and also that the 14␣ configuration is not totally inert even when
no oxygen function is present at this site.
3.4. Dissociation of 3ˇ-OH,5ˇ,14ˇ bufadienolide and ouabain
from human erythrocytes
After exposure to 1 × 10−5 mol/l ouabain, human erythrocytes
washed three times in normal physiological buffer still remained
inhibited by 97% compared to uninhibited control values. In
contrast, after exposure to 1 × 10−4 mol/l 3␤-OH,5␤,14␤ bufa-
dienolide, human erythrocytes washed three times in normal
physiological buffer were inhibited by only 15% compared to unin-
hibited control values. These results indicate that 3␤-OH,5␤,14␤
bufadienolide morereadilydissociates fromthehumanerythrocyte
sodium pump than does ouabain.
4. Discussion
One member of the 14ß series, 3ß-OH,5ß, 14ß bufadienoide,
exhibited interesting differences both from the 14␣ bufadienolides
and from bufalin (from which it differed solely in the lack of a
14-OH – Fig. 3). The pattern of inhibition in dog kidney and pig
brain ATP-ases resembled that of ouabain and bufalin though the
concentration required to produce 50% inhibition was consider-
ably greater than either at approximately 10−5 mol/l. The pattern
of inhibition of the sodium pump of human erythrocytes showed
a probable maximum at just under 70% in contrast to the isolated
ATP-ases where almost 90% inhibition could be achieved, albeit at
high concentrations of the compound. The sodium pump of human
leucocytes also demonstrated a limit to the inhibition achievable to
just under 50%. However, whereas the pattern of inhibition of the
erythrocyte sodium pump paralleled that of the isolated ATP-ases
at lower concentrations, that of the leucocyte at first sight appeared
to be exhibiting a lower Hill coefficient. While we have no direct
proof that this is not the explanation for the appearance of the inhi-
bition, an alternative and arguably more likely explanation is that
the relationship is actually a composite of two conventional curves
as suggested in Fig. 5. A similar appearance has been described in
Two different lines of argument led us to study semi-synthetic
bufadienolides lacking the 14-OH. First was the consideration that
the better-known cardenolides, digoxin and ouabain, would exhibit
severe toxicity if present in the concentrations required to demon-
strate sodium pump inhibition in human leucocytes obtained from
hypertensives or in transfer experiments utilising plasma from
hypertensive patients. The coefficient of variation in “classical” leu-
cocyte sodium transport experiments points to a concentration of
either cardenolide of at least 10−8 mol/l being required to demon-
strate significant inhibition with the sample size used in most
studies.
Second, studies of an inhibitor of the sodium pump obtained
from human placentas suggested that it possessed only a single –OH
group [18]. Evidence for this came from the fact that the material
deuterated only once, indicating the presence of a single exchange-
able hydrogen characteristic of an alcohol. With the exception of
oestrogens, non-substitution at C3 is unusual making it likely that
the hydroxyl group was at this site rather than any other, including

P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
1143
the case of rat testis Na/K ATP-ase exposed to ouabain, that tissue’s
sodium pumps being a mixture of the relatively ouabain resistant
justify further investigation. Further, similar approaches to syn-
thesis may yield interesting compounds that could play a useful
pharmaceutical role.
␣1 and ouabain-sensitive ␣4 isoforms [22]. If we concentrate on the
suggested pattern of inhibition in the leucocyte preparation, the
relative ineffectiveness of this compound compared with ouabain
takes on a different appearance. Essentially a high proportion of the
sodium pumps under study are resistant to 3ß-OH,5ß, 14ß bufa-
dienolide. However the pattern of inhibition of one subset of the
pump (which appears to represent only some 15% of the total) is
that 50% inhibition occurs at approximately 10−7 mol/l and this is
comparable to the EC50 of ouabain for both erythrocytes and leu-
cocytes of 2 × 10−7 mol/l. Were there to be any interest in targeting
this subset of the sodium pump, then this compound’s relative inef-
fectiveness against the other subsets gives it a considerable degree
of specificity in this respect.
Appendix A.
The steps in the synthesis of 3ß-OH,5ß,14ß bufadienolide are
described below. The synthesis of the corresponding 14␣ com-
pound follows the same pathway omitting the epimerisation steps
included between the symbols **.
Step a: Bromination of 2-pyrone to 5,6-dibromo-5,6-dihydro-
2H-pyran-2-one (photodibromide).
The pattern of inhibition of the four preparations by emicymarin
and allo-emicymarin showed both expected and unexpected fea-
tures. It is clear, as has been previously reported [21] that changing
the configuration at C17 from beta (emicymarin) to alpha (allo-
emicymarin) profoundly reduces the inhibitory capability of the
compound. What was not anticipated was the markedly reduced
effectiveness of emicymarin on the leucocyte sodium pump as com-
pared with that of the erythrocyte or the isolated ATP-ases. There
was no suggestion of the selectivity seen with the 14ß anhydroxy
bufadienolides and we have been unable to produce a satisfactory
explanation for this phenomenon, which was not observed in the
range of bufadienolides and other cardenolides we tested.
Limitations of our studies should be mentioned. We studied
human erythrocytes and leucocytes along with dog kidney and pig
brain tissue whereas, ideally, we might have used only human tis-
sues and cells since species differences in the indices measured
may exist. Second, it would be of interest to have studied cardiac
and vascular tissues in addition to the four preparations we used
since, it might be argued, these may be some of the most important
target tissues of circulating Na/K ATP-ase inhibitors.
The method used was derived from that described by Pirkle
and Dines [23]. Briefly, in a dry flask under N2 was placed DCM
(200 ml) and freshly distilled 2H-pyran-2-one (␣-pyrone) (10.56 g,
110.0 mmol). The flask was cooled to −78 ◦C (dry-ice – acetone).
Bromine (17.56 g, 110.0 mmol) was placed in a dropping funnel,
diluted in DCM (50 ml) and chilled using dry-ice pellet around the
funnel. The cold solution was added to the ␣-pyrone over 1 h whilst
shinning light on the flask (500 W floodlight). The irradiation was
continued for about 5 h (TLC:SiO2, EtOAc/hexane 3:7, UV detection,
Rf(␣-pyrone) = 0.20, Rf(photodibromide) = 0.54) with stirring, maintaining
the temperature at −78 ◦C. The light source was removed and the
contents of the flask were allowed to warm to 20 ◦C. The solvent
was removed under reduced pressure to afford the title compound
as a light-orange oil (27.42 g, 97%). By 1H NMR traces of starting
material remained. The photodibromide readily loses HBr at room
temperature [23] and was used without further purification.
Step b: Dehydrodebromination of photodibromide to 5-bromo-
2H-pyran-2-one [23].
However, in the systems reported in this paper we carried out
extensive experiments to fully characterise the inhibitory pro-
files for all commercially available cardenolides and found the
responses did not differ significantly, especially when the systems
were exposed to ouabain and bufalin.
5. Summary
(1) 14␣ anhydroxy bufadienolides are weak inhibitors of Na/K ATP-
ase.
Following the prior art [23] the title compound (18.35 g)
(2) Changing the configuration at the C17 position from ß to ␣
(emicymarin to allo-emicymarin) reduces profoundly the com-
pound’s capability to inhibit sodium pumps.
was obtained as
a crude light brown solid. Two column
chromatographies (SiO2, DCM; TLC: SiO2, DCM, UV detection,
Rf(photodibromide) = 0.82, Rf(product) = 0.34) afforded an off white solid
(13.69 g, 73%) and sublimation of the residue from relevant frac-
tions from the second column chromatography allowed isolation
of colourless needles (1.45 g) mp 61–63 ◦C.
(3) Emicymarin exhibited markedly reduced inhibition of the
sodium pump of leucocytes compared with sodium pumps in
erythrocytes, dog kidney and pig brain preparations.
(4) A synthetic compound we prepared, 3ß-OH,5ß,14ß bufadieno-
lide, exhibits selectivity in its inhibitory effect on different
sodium pump preparations in contrast to ouabain and bufalin
which showed no specificity of action between our four prepa-
rations.
Step c: Preparation of 2-pyrone-5-boronate (Suzuki precursor)
(Published in Ref. [17]).
(5) It is probable that there are at least two subtypes of Na/K-ATP-
ase present on human leucocytes with differing sensitivities to
3ß-OH,5ß,14ß bufadienolide.
(6) 3ß-OH,5ß,14ß bufadienolide more readily dissociates from ery-
throcyte Na/K-ATP-ase than does ouabain.
We would thus like to propose that this model compound,
3ß-OH,5ß,14ß bufadienolide, demonstrates some interesting
inhibitory characteristics on certain sodium pump subtypes that
A mixture of 5-bromo-2-pyrone (13 g, 74 mmol), pinacolborane
(16.2 ml, 14.7 g, 111 mmol), triethylamine (31 ml, 22.1 g, 222 mmol)
and PdCl2(PPh3)2 (1.56 g, 2.2 mmol) in dry toluene (100 ml) was

1144
P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
heated to reflux under a N2 atmosphere for 6 h. After this time TLC
analysis showed no remaining 5-bromo-2-pyrone. The dark solu-
tion was filtered through a short plug of silica and this was further
eluted with dichloromethane (250 ml). The solvent was removed
under reduced pressure to give a dark crystalline solid. Flash col-
umn chromatography using dichloromethane as eluent gave the
2-pyrone-5-boronate (13.6 g, 83%) as a pale yellow crystalline solid,
mp 91–93 ◦C. This material was used in all coupling reactions. An
analytical sample with the same mp could be obtained as white
crystals by sublimation at 0.7 mm Hg pressure.
Step 4: Acid catalysed C–C double bond isomerisation of 15:16-
enone to 14:15-enone (14-position epimerisation step 3).
In a dry flask containing Nitrogen were placed the steroid start-
ing material, dry THF and p-toluene sulphonic acid monohydrate.
The reaction mixture was heated under reflux for about 4 h. Whilst
still underreflux, Et3Nwasadded. Heatingwas stopped after 10 min
and the contents of the flask allowed to cool to 20 ◦C. The sol-
vent volume was reduced to about 15 ml in vacuo and transferred
to a separating funnel. After adding some DCM the organic layer
was washed once with NaHCO3(sat). Then the aqueous layer back
extracted in DCM. The combined DCM fractions were washed once
with brine and then dried (Na2SO4). The solvent mixture was evap-
orated to dryness. The title compound was obtained as a solid from
solvent.
Step 1: Protection of 3-hydroxy function as
butyldimethysilyl ether [24].
a
tert-
Following the literature procedure [24], the TBDMS-protected
etiocholan-3ß-ol-17-one was obtained as a white solid.
**Step 2: Preparation of 17-silylenol ether of protected steroid
(14-position epimerisation step 1).
Step 5: Palladium catalysed hydrogenation of 14:15-enone to
14bH-17-keto steroid (14-position epimerisation step 4).
The title compound was obtained as a solid from solvent.**
Step 6: Preparation of vinyl triflate of steroid.
In a dry flask under a N2 atmosphere were placed the steroid
starting material and dry THF. The flask was cooled to −78 ◦C (dry-
ice – acetone) and LDA was introduced over 2 min using a syringe.
The enolate formation was allowed to proceed for two hours. In a
dropping funnel (non-pressure equalising) featuring a sinter at its
base, above the tap, were premixed TMSCl and Et3N THF in order
to trap any HCl impurity originating from TMSCl. After mixing for
a few minutes the suspension was filtered slowly into the reac-
tion flask by pressurising the funnel over a 2–3 min period. After
15 min the reaction mixture was allowed to warm to 20 ◦C and
stirred for 5 h. The contents of the flask were then evaporated
in vacuo to afford a yellow oily residue. The residue was reconsti-
tuted in DCM and transferred to a separating funnel where it was
washed with NaHCO3(sat) followed by back extraction of the aque-
ous layer in DCM. The combined DCM fractions were washed with
brine, dried over Na2SO4(anh) and evaporated to dryness to afford
the title compound.
The title compound was obtained as a solid from solvent.
Step 7: Suzuki cross coupling of vinyl triflate with 2-pyrone-5-
boronate.
Step 3: Tsuji-type oxidation of 17-silylenol ether to correspond-
ing 15:16-enone (14-position epimerisation step 2).
Vinyl triflate (50 mg, 0.22 mmol), 2-pyrone-5-boronate (55 mg,
0.25 mmol), and PdCl2 (dppf) (20 mg, 0.022 mmol) were dissolved
in dry DMF (4 ml) under a N2 atmosphere. Finely crushed K3PO4
(180 mg, 0.82 mmol) was added to the DMF solution and this slurry
was heated to 60 ◦C for 6 h. The DMF was removed under reduced
pressure to give a dark solid that was initially purified by dis-
solving in dichloromethane and filtering through a plug of silica,
followed by washing with dichloromethane (30 ml). The solvent
was removed under reduced pressure to give a solid that was fur-
ther purified by flash column chromatography (10% ethyl acetate
in dichloromethane) to the bufatrienolide obtained as a solid from
the solvent.
The method used was derived from the work by Tsuji [25]. The
TMS enol ether was dissolved in a mixture of dry DCM and dry
acetonitrile in a flask under a N2 atmosphere. The solution was
heated to 60 ◦C for a few minutes and then palladium(II) acetate
was introduced into the reaction flask as once portion and the
orange solution stirred 60 ◦C for at least 6 h. The title compound
was obtained as a solid from solvent.

P.J. Hilton et al. / Steroids 75 (2010) 1137–1145
1145
Step 8: Hydrogenation of 16:17-double bond using Wilkinson’s
catalyst.
[9] Manunta P, Mara Ferrandi M, Bianchi G, Hamlyn JM. Endogenous ouabain in
cardiovascular function and disease. J Hypertens 2009;27:9–18.
[10] Hilton PJ. Cellular sodium transport in essential hypertension. N Engl J Med
1986;314:222–9.
[11] Haft JI, Shahabadi A-E, Fano A. Clinical experience with ouabain admin-
istered in small divided doses in the monitored patient. Chest 1973;63:
868–74.
[12] Crambert G, Hasler U, Beggah AT, Yu C, Modyanov NN, Horisberger J-D, et
al. Transport and pharmacological properties of nine different human Na,K-
ATPase isoenzymes. J Biol Chem 2000;275:1976–86.
[13] Marks MJ. Seeds NW A heterogeneous ouabain-ATPase interaction in mouse
brain. Life Sci 1978;23:2735–44.
[14] Walter U, Distler A. Abnormal sodium efflux in erythrocytes of patients with
essential hypertension. Hypertension 1982;4:205–10.
[15] Hilton PJ, Johnson VE, Jones RB, Patrick J. The effect of alterations in the external
sodium concentration on human leucocyte sodium and potassium transport in
vitro. J Cell Physiol 1981;109:323–32.
[16] Kraml M. A semi-automated determination of phospholipids. Clin Chim Acta
1966;13:442–8.
[17] Gravett EC, Hilton PJ, Jones K, Peron J-MR. A new route to 5-Aryl and 5-
Heteroaryl-2-pyrones via Suzuki coupling of a 2-Pyrone-5-boronate ester.
Synlett 2003;2:253–5.
[18] Hilton PJ, White RW, Lord GA, Gordon DB, Hilton MJ. An inhibitor of the sodium
pump obtained from human placenta. Lancet 1996;248:303–5.
[19] Gray HH, Johnson VE, Poston L, Hilton PJ. Sodium transport by leucocytes
and erythrocytes in hypertensive subjects and their normotensive relatives.
J Hypertens 1984;2(suppl 3):467–9.
The title compound was obtained as a solid from solvent.
1H NMR confirmed that the hydrogenation product was exclu-
sively the 17ß epimer.
References
[1] Gowda RM, Cohen RA, Khan IA. Toad venom poisoning: resemblance to digoxin
toxicity and therapeutic implications. Heart 2003;89:e14.
[2] Welt LG, Sachs JR, McManus TJ. An ion transport defect in erythrocytes from
uremic patients. Trans Assoc Am Physicians 1964;77:169–81.
[3] Edmondson RPS, Hilton PJ, Jones NF, Patrick J, Thomas RD. Leucocyte sodium
transport in uraemia. Clin Sci 1975;49:213–6.
[4] Haddy F, Pamnani M, Clough D. The sodium-potassium pump in volume-
expanded hypertension. Clin Exp Hypertens 1978;1:295–336.
[5] Morris JF, McEachern MD, Poston L, Smith SE, Mulvany MJ, Hilton PJ. Evidence
for an inhibitor of leucocyte sodium transport in the serum of neonates. Clin
Sci 1987;73:291–7.
[6] Edmondson RPS, Thomas RD, Hilton PJ, Patrick J, Jones NF. Abnormal leu-
cocyte composition and sodium transport in essential hypertension. Lancet
1975;1:1003–5.
[20] Boon NA, Harper C, Aronson JK, Grahame-Smith DG. Cation transport functions
in vitro in patients with untreated essential hypertension: a comparison of
erythrocytes and leucocytes. Clin Sci 1985;68:511–5.
[21] Glynn IM. The action of cardiac glycosides on sodium and potassium move-
ments in human red cells. J Physiol 1957;136:148–73.
[22] Blanco G, Mercer RW. Isoenzymes of the Na-K-ATPase: heterogeneity in struc-
ture, diversity in function. Am J Physiol 1998;275:F633–50.
[23] Pirkle WH, Dines M. The bromination of 2-pyrone.
1969;34:2239–4224.
J
Org Chem
[24] Greene TW, Wuts PGM. Protective Groups in Organic Synthesis. 2nd ed. New
York: Wiley and Sons; 1991.
[25] Tsuji J. Synthetic application of the palladium-catalysed oxidation of olefins to
ketones. Synthesis 1984;5:369.
[7] Mathews WR, DuCharme DW, Hamlyn JM, Harris DW, Mandel F, Clark MA, et
al. Mass spectral characterisation of an endogenous factor from human plasma.
Hypertension 1991;17:930–5.
[8] Nicholls MG, Lewis LK, Yandle TG, Lord GA, McKinnon W, Hilton PJ. Ouabain,
a circulating hormone secreted by the adrenals, is pivotal in cardiovascular
disease. Fact or fantasy? J Hypertens 2009;27:3–8.