Dexamethasone and ursodeoxycholic acid protect against the arrhythmogenic effect of taurocholate in an in vitro study of rat cardiomyocytes


*Correspondence: Dr C. Williamson, Maternal and Fetal Disease Group, Institute of Reproductive and Developmental Biology, Hammersmith Campus, Du Cane Road, London W12 0NN, UK.


Objective To establish whether the therapeutic agents ursodeoxycholic acid and dexamethasone protect cardiomyocytes from taurocholate-induced arrhythmias in an in vitro model.

Design Laboratory study.

Setting Imperial College London, Hammersmith Campus.

Sample Neonatal rat cardiomyocytes.

Methods Using scanning ion conductance microscopy, we measured the rate, rhythm, amplitude of contraction and calcium dynamics of ventricular myocytes from one to two day old rats. Cells were pre-incubated for 16 hours in dexamethasone (80 or 800 nM) or 0.1 mM ursodeoxycholic acid before adding taurocholate at different concentrations (0.3–4.5 mM).

Main outcome measures Changes in rate and amplitude of contraction, calcium dynamics and rhythm.

Results Taurocholate at concentrations of up to 3 mM induces abnormal changes including reductions in rate, amplitude of contraction, abnormal calcium dynamics and dysrhythmias. Although dexamethasone had no immediate protective effect on these changes, pre-incubation with dexamethasone was protective. Ursodeoxycholic acid pre-incubation was protective at taurocholate concentrations up to 1 mM.

Conclusion The therapeutic agents dexamethasone and ursodeoxycholic acid appear protective against the arrhythmogenic effect of taurocholate on cardiomyocytes.


Obstetric cholestasis, also called intrahepatic cholestasis of pregnancy, is a liver disease of pregnancy associated with third trimester intrauterine death, fetal distress and spontaneous preterm labour1–4. It affects approximately 0.6% of pregnancies in UK white Caucasians, and double this number of pregnancies in Indian and Pakistani Asians5,6. The perinatal mortality rate has fallen from 9–11% in older studies2,3 to 2–3.5% in more recent studies where the fetus was delivered by 38 weeks of gestation1,3,4.

The aetiology of the fetal complications of obstetric cholestasis pregnancies is poorly understood. Intrauterine death is thought to occur suddenly, as there is no evidence of preceding intrauterine growth restriction or uteroplacental insufficiency and fetal autopsy is normal1. Placental histology shows non-specific changes consistent with hypoxia7. An abnormal fetal heart rate (≤100 or ≥180 beats/minute) has been observed in previous studies2,8,9.

Obstetric cholestasis causes maternal pruritus and abnormal liver function tests, including raised serum primary bile acids9–12. The increase in cholic acid is much more marked than that of chenodeoxycholic acid or deoxycholic acid12–14.

It has not been established whether the absolute levels of maternal total serum bile acids correlate with adverse pregnancy outcomes in obstetric cholestasis. In one study of 86 pregnant women with obstetric cholestasis in which the severity of the disease was classified according to the level of the serum bile acids, the incidence of meconium-stained amniotic fluid was significantly increased in the subgroup of women with the most severe cholestasis11. In another series of 117 cases that were classified in the same way, reduced fetal heart rate variability and/ or late decelerations on cardiotocography were seen more commonly in the subgroup with most severe cholestasis compared with the group with mild cholestasis9. There has also been one case report of primary sclerosing cholangitis in pregnancy in which there was marked fetal bile acidaemia associated with a variety of fetal complications including fetal bradycardia15.

We hypothesised that raised fetal serum bile acids in obstetric cholestasis may result in the development of a fetal dysrhythmia and sudden intrauterine death. Previously, using an in vitro model of rat one to two day neonatal cardiomyocytes, we have demonstrated that the primary bile acid taurocholate can alter the rate and rhythm of cardiomyocyte contraction and cause abnormal Ca2+ dynamics16. We have also shown that superfusion of a culture with taurocholate has different arrhythmogenic effects in individual cells, and that taurocholate causes reduced amplitude of contraction as well as dysrhythmias17. These data support our hypothesis that raised fetal serum bile acids in obstetric cholestasis may result in the development of a fatal dysrhythmia in the fetus and sudden intrauterine death.

Taurocholate was used because cholic acid is the main bile acid that is raised in obstetric cholestasis. In humans, the majority of bile acids are conjugated with either glycine or taurine, and glycoconjugates predominate. However, in obstetric cholestasis, the normal glycine/taurine ratio of 1.4 is reversed to 0.818.

Two drugs, ursodeoxycholic acid and dexamethasone, are clinically useful in obstetric cholestasis. There is more experience of the use of ursodeoxycholic acid than dexamethasone in the treatment of obstetric cholestasis. When the results of 10 studies with a total of 85 affected women who were treated with ursodeoxycholic acid are combined, 74 (87%) showed clinical or biochemical improvement, or both18–26. Ursodeoxycholic acid treatment has been shown to improve the serum bile acid levels measured in cord blood and amniotic fluid at the time of delivery23. In addition, there is very little accumulation of ursodeoxycholic acid in the amniotic fluid or in cord blood23. However, ursodeoxycholic acid treatment does not decrease total bile acid concentrations in the meconium27. There have been no reports of an adverse fetal outcome following maternal treatment with ursodeoxycholic acid. There has been one report of a series of 10 women treated with dexamethasone. In this series, all had a clinical and biochemical response28.

In vitro experiments using placentas from women with obstetric cholestasis have demonstrated an impaired rate and affinity of bile acid transport in the basal membrane and reduced efficiency of transport at the apical membrane. All of the observed changes were reversed in placentas from women treated with ursodeoxycholic acid29. However, there have been no studies of the effect of either ursodeoxycholic acid or dexamethasone treatment on features of fetal dysrhythmia. As it has not been possible to perform controlled clinical trials of the effects of ursodeoxycholic acid and dexamethasone on fetal outcome in obstetric cholestasis, in vitro animal models such as the one presented in this paper are an important source of knowledge about the effects of these therapeutic agents. Therefore, in order to further investigate the association between elevated bile acid and fetal heart abnormalities in this study, we aimed to investigate whether ursodeoxycholic acid and dexamethasone have a protective effect on the arrhythmogenic effect of the bile acid taurocholate on rat neonatal cardiomyocytes.


Ventricular myocytes were isolated from the hearts of one to two day old rats30, cultured as previously described16,17, and used following three to four days culture on glass coverslips (small network of cardiomyocytes). Taurocholate (Sigma-Aldrich, Dorset, UK) was added to give a final concentration of 0.3, 1.0 and 3.0 mM to a network of synchronously beating cells. The rate of contraction was calculated as described previously16. In some experiments cells were pre-incubated for 16 hours (overnight) with either 0.1 mM ursodeoxycholic acid (Sigma-Aldrich) or with either 80 or 800 nM dexamethasone (David Bull Laboratories, Warwick, UK) prior to the addition of taurocholate. In experiments where the therapeutic agents protected the cardiomyocytes from the effect of 0.3, 1.0 and 3.0 mM taurocholate, a higher dose of 4.5 mM was used. Taurocholate, ursodeoxycholic acid and dexamethasone were diluted in phosphate buffered saline to make stock solutions prior to addition to culture medium; the concentrations of the stock solutions were 150 mM, 10 mM and 80 μM, respectively. To investigate cardiomyocyte recovery, cells were transferred to taurocholate-free medium for 1 hour.

To investigate the changes of the rhythmicity of cardiomyocyte contraction caused by the addition of taurocholate, we recorded the vertical cell displacement of individual cells using a scanning ion conductance microscope as described previously31.

The cardiomyocytes were loaded with the visible wavelength fluo-4 Ca2+ indicator by cell incubation with the esterified derivative of fluo-4 (fluo-4 acetoxymethyl) (Molecular Probes, Leiden, Netherlands) in a medium containing equal volumes of Leibovitz's L-15 (Gibco, Paisley, UK) and Hank's balanced salt solution buffer (Gibco) at room temperature for 15 minutes. Cells were then rewashed five times with the medium, followed by a post-incubation period of 20 minutes to allow for complete intracellular dye cleavage32. Scanning laser confocal microscopy was used to study the Ca2+ dynamics in the cardiac myocyte network as described previously16,31.

Cells were fixed with 4% formalin solution in phosphate buffered saline, permeabilised with 0.1% Triton X-100 in phosphate buffered saline for 20 minutes. Blocking solution of 10% normal goat serum was then applied, followed by monoclonal antibodies to sarcomeric α-actinin, clone EA-53 (Sigma-Aldrich, A7811) in 1:100 dilution. The antigen was detected using fluorescein-conjugated secondary antibody. Slides were observed under a Nikon-Eclipse TE300 microscope. To localise fibrillar actin, cells were counterstained with TRITC-phalloidin (Sigma-Aldrich) in 1:500 dilution.

To establish the proportion of cardiomyoctes in the geneticin-treated cultures, cells were stained with antibodies to detect sarcomeric α-actinin and phalloidin to reveal the distribution of filamentous actin. Immunodetection showed that almost 90% of the cells possessed the myocyte marker, α-actinin (data not shown). Cells incubated normally display synchronous contractile behaviour, and taurocholate induces arrhythmia16.


We present here the results of several independent experiments for each experimental condition analysed. Figures 1–3 illustrate recordings that we typically obtained, and which were consistent between experiments. In the first set of experiments, cells were pre-incubated for 16 hours with either ursodeoxycholic acid or dexamethasone. Taurocholate was then added to the culture medium. With dexamethasone or ursodeoxycholic acid alone under standard bright-field microscopy cells displayed synchronous beating. After cell loading with the calcium indicator fluo-4 synchronous calcium waves were observed that spanned the entire network of cells (Figs 1A and 1E), that is, the cells were coupled. When taurocholate was added to the culture medium of cells that had been pre-incubated with dexamethasone to give a final taurocholate concentration of up to 3 mM, the concentration that has been previously reported to cause severe disruption of cardiomyocyte contraction17, there was no change in the rate or synchronisation of contraction (Fig. 1B) (four experiments). Only at the higher dose of 4.5 mM taurocholate did the cells show contractile irregularities (Fig. 1C). Neighbouring cells within the network had different patterns of calcium wave propagation, consistent with abnormal cell–cell communication. As soon as the medium was changed to taurocholate-free medium, the cells displayed signs of recovery. Although the rate of contraction did not return to that seen prior to taurocholate addition, it became regular and calcium waves showed intercellular coupling (Fig. 1D).

Figure 1.

Effect of taurocholate on calcium dynamics in cardiomyocytes pretreated with dexamethasone and ursodeoxycholic acid. (A) The cardiomyocyte network was incubated for 16 hours with 800 nM dexamethasone (A–D). Horizontal lines represent calcium waves which extend across adjacent cells, consistent with synchronous beating. The spacing between the lines provides an indication of frequency. (B) The addition of taurocholate up to a concentration of 3 mM does not alter calcium dynamics. (C) Addition of taurocholate to a concentration of 4.5 mM alters calcium dynamics. (D) Restoration of rhythmic contraction after transfer to taurocholate-free medium. (E) Network of cardiomyocytes incubated for 16 hours with 0.1 mM ursodeoxycholic acid. (F) The addition of 0.3 mM taurocholate does not alter calcium waves. (G) The addition of 3 mM taurocholate alters calcium wave dynamics, consistent with tachycardia—closely spaced lines (t), (H) this did not return to normal following transfer to taurocholate-free medium, and calcium overload (o) was seen with cessation of beating.

Figure 2.

Effect of various concentrations of taurocholate on the contraction of a newborn rat cardiomyocyte pre-incubated with ursodeoxycholic acid (0.1 mM, 16 hours). (A) Control; The addition of 0.3 mM taurocholate (B) and 1.0 mM taurocholate (C) does not decrease the rate of contraction; (D) Addition of 3 mM taurocholate causes an appearance consistent with fibrillation of the cardiomyocyte.

Figure 3.

(A) Effect of 3 mM taurocholate on the contraction of cardiomyocytes pre-incubated with dexamethasone (800 nM, 16 hours). The addition of 3 mM taurocholate slightly decreases the amplitude of contraction but not the rate (3 minutes after addition of taurocholate to the culture). (B) Effect of simultaneous addition of 1.0 mM taurocholate and 800 nM dexamethasone to the cardiomyocyte culture for 2 hours. The addition of taurocholate decreases the rate of contraction.

Pre-incubation with ursodeoxycholic acid protected cells from the effects of 0.3 mM taurocholate (Fig. 1F) but 3.0 mM taurocholate produced fibrillation/tachycardia and loss of synchronous contraction (Fig. 1G) and subsequent irreversible cellular overload with calcium (Fig. 1H) (four experiments).

The protective effect of dexamethasone was dependent on both the dose of taurocholate added and the concentration of dexamethasone in the pre-incubation medium (Table 1). Without pre-incubation with dexamethasone, the integrity of the network of cardiomyocytes was lost in the presence of both 3.0 and 4.5 mM taurocholate, and the effects of the latter concentration were irreversible. Pre-incubation with 80 nM dexamethasone protected cells from the effect of 0.3 and 3.0 mM taurocholate, but not from the higher dose of 4.5 mM. After transfer to taurocholate-free medium, the cells recovered but had a slower rate of contraction than before taurocholate treatment. Addition of 800 nM dexamethasone to the pre-incubation medium protected cells from the effect of 0.3 and 3.0 mM taurocholate concentrations, but cells exposed to 4.5 mM taurocholate had a slower rate of contraction than before treatment. However, they made a full recovery after taurocholate was removed.

Table 1.  Effect of 16 hours pre-incubation with dexamethasone (80 or 800 nM) prior to addition of 3.0 or 4.5 mM taurocholate on the rate of contraction of cardiomyocytes (beats per minute). Values are mean (SD).
 No dexamethasone80 nM dexamethasone800 nM dexamethasone
  1. LI = loss of integrity of the network.

  2. *P < 0.001 when compared with control data.

10 minute treatment
Control105.7 (7.9)99.6 (9.2)104.1 (13.0)
3.0 mM taurocholateLI88.2 (15)96.5 (11.3)
4.5 mM taurocholateLILI62.1 (8.7)*
1 hour treatment
Control105.5 (7.6)96.6 (8.4)100.2 (8.0)
3.0 mM taurocholateLI92.2 (12.8)96.8 (9.7)
4.5 mM taurocholateLILI61.0 (10.0)*
1 hour recovery
Control105.9 (8.7)101.9 (7.2)99.0 (11.0)
3.0 mM taurocholate46.1 (8.0)*93.7 (7.3)97.7 (9.1)
4.5 mM taurocholateLI67.7 (13.0)*87.7 (10.0)

Table 2shows that ursodeoxycholic acid could only protect cells from 1 mM taurocholate. When the cells were exposed to 3 mM taurocholate, there was loss of integrity and synchronous beating within the network of cells. When the cardiomyocytes were subsequently transferred to taurocholate-free medium, they recovered but had a slower rate of contraction than before treatment.

Table 2.  Effect of 16 hours pre-incubation with ursodeoxycholic acid (0.1 mM) prior to addition of 1.0 or 3.0 mM taurocholate on the rate of contraction of cardiomyocytes (beats per minute). Values are mean (SD).
 No ursodeoxycholic acid0.1 mM ursodeoxycholic acid
  1. LI = loss of integrity of the network.

  2. *P < 0.001 when compared with control data.

10 minute treatment
Control88.0 (6.4)88.0 (7.4)
1.0 mM taurocholate47.2 (4.5)*84.0 (6.7)
3.0 mM taurocholateLILI
1 hour treatment
Control89.2 (4.6)85.5 (6.7)
1.0 mM taurocholate43.2 (3.2)*85.5 (5.2)
3.0 mM taurocholateLILI
1 hour recovery
Control90.5 (5.2)88 (7.4)
1.0 mM taurocholate57.6 (6.0)*86.5 (9.5)
3.0 mM taurocholate53.3 (4.9)*52.8 (3.9)*

In our previous paper, we demonstrated a reduced rate of contraction and proportion of beating cells when cardiomyocytes were exposed to increasing concentrations of taurocholate (0.1–3.0 mM), more marked at higher concentrations (P < 0.001). Using scanning ion conductance microscopy, we also demonstrated reduced amplitude of contraction and calcium transients with taurocholate17.

Figure 2 shows a representative single-point scanning ion conductance microscope measurement of contraction of a small aggregate. The aggregate was pre-incubated with ursodeoxycholic acid for 16 hours then increasing doses of taurocholate were added (eight experiments). The rate of contraction did not change after addition of 0.3 or 1 mM taurocholate, but a small reduction of amplitude from 3 to 2.2 μM was seen (Figs 2B and 2C). Addition of 3 mM taurocholate caused movement oscillations which look like fibrillation (Fig. 2D).

In Fig. 3A, a typical example of a single-point scanning ion conductance microscope measurement is shown where a contracting cardiomyocyte pre-incubated with 800 nM dexamethasone for 16 hours was exposed to increasing concentrations of taurocholate (four experiments). There was no change in contraction when up to 1 mM taurocholate was added (data not shown). After 3 minutes, incubation with 3 mM taurocholate caused a slight decrease in amplitude of contraction from 3 to 2.2 μM, but the rate remained the same.

Pre-incubation with dexamethasone curtails the effects of taurocholate (Figs 1A–1D), and the question arises whether pre-incubation is a requirement for dexamethasone's action. When the cells were not pre-incubated with dexamethasone, and 800 nM dexamethasone and 1 mM taurocholate were added to the culture medium together, this lower dose of taurocholate caused a similar effect to that seen in cells that were not incubated with dexamethasone (four experiments) (Fig. 3B).


This study shows that the therapeutic agents dexamethasone and ursodeoxycholic acid protect neonatal rat cardiomyocytes from the arrhythmogenic effects of the bile acid taurocholate. We have previously demonstrated that when single cells or networks of cardiomyocytes are cultured in medium containing taurocholate, this results in abnormalities in the rate and amplitude of contraction, altered calcium dynamics and in the loss of synchronous contraction. We have proposed that the results in this in vitro model may offer an explanation for the aetiology of intrauterine death in obstetric cholestasis16. If this hypothesis is correct, the protective effect of ursodeoxycholic acid and dexamethasone in the current study is a clinically important finding. Dexamethasone was more protective than ursodeoxycholic acid. This may reflect different mechanisms of action. We have shown that dexamethasone protects against taurocholate-induced loss of physiological integrity of the cardiomyocyte network. Dexamethasone has previously been shown to maintain the structural integrity of myocytes for periods of at least 45 days in the absence of any damaging agents33.

Dexamethasone also protects against the taurocholate-induced reduction in rate of contraction. These data are in agreement with a report that demonstrated that dexamethasone protected against a decreased contraction rate and multiform arrhythmias that were induced in cultured rat myocardial cells following infection with Coxsackie B-2 virus34–36. In addition, dexamethasone inhibits endotoxin-induced changes in calcium and contractility in rat isolated papillary muscle. Pretreatment of rats with dexamethasone prevented the endotoxin-induced decrease in peak tension and inhibited the elevation in resting [Ca2+], with a resultant maintenance of Ca2+ transient magnitude37. The effect of dexamethasone is often attributed to its ability to induce gene expression or stabilise mRNA in target cells38.

The cardioprotective effect of ursodeoxycholic acid may be a consequence of protection from the apoptosis that is induced by more hydrophobic bile acids or may be due to membrane stabilisation. Most of the studies of the protective effect of ursodeoxycholic acid have compared it with more hydrophobic bile acids (e.g. glycochenodeoxycholic acid). In rat and human hepatocytes, the mechanism of cytotoxicity of glycochenodeoxycholic acid varies at different concentrations. At the higher concentrations that are seen in severe cholestasis (i.e. 0.5 mM), cytolytic cell destruction has been demonstrated39. At lower pathological concentrations of 0.05–0.1 mM, apoptosis is the predominant mechanism of bile acid toxicity40,41. Studies of human and rat hepatocytes have shown that ursodeoxycholic acid protects against these cytotoxic effects of hydrophobic bile acids39,40 and ursodeoxycholic acid at millimolar concentrations protects against membrane damage caused by more hydrophobic bile acids in in vitro studies of rat and human hepatocytes, and in artificial membranes42–44.

Ursodeoxycholic acid may also exert its cardioprotective effect by increasing ATP levels. Studies of the mechanism by which ursodeoxycholic acid promotes bile flow have demonstrated that it stimulates secretion of ATP in isolated rat hepatocytes45. In cardiac myocytes, ATP is the endogenous ligand for the P2X receptor, and separate studies of rat and mouse hearts have shown that both ATP and another P2X receptor agonist (2-meSATP) stimulated large increases in the myocyte contractile amplitude46. Thus, it is possible that ursodeoxycholic acid-induced rise of ATP in cardiomyocytes has a cardioprotective effect.

Ursodeoxycholic acid increases the density of the hepatic apical conjugate export pump, Mrp2, fourfold in rat canalicular membranes47. The authors are not aware of other studies of the effect of ursodeoxycholic acid on the insertion of other potential bile acid transporters into the canalicular membrane.

Both ursodeoxycholic acid and dexamethasone could have immediate effects on cell membrane transporters, or compete with bile acid. But it seems unlikely that this plays a part in the effect of both agents on cardiomyocyte contraction, as both agents act only after being applied to cells well before the addition of taurocholate.

It is possible that the cardioprotective effects of ursodeoxycholic acid and dexamethasone are the result of altered expression of genes that influence bile acid transport or metabolism in cardiomyocytes. We are currently not aware of studies of the expression of these bile acid transporter genes in cardiomyocytes. If the transporters are expressed in cardiomyocytes, it is also possible that ursodeoxycholic acid and dexamethasone could act by a direct effect on the function of the transporters.

This current study suggests that dexamethasone may have a more potent cardioprotective effect on the fetus, although this study is relatively small. However, there is evidence from animal and human studies that fetal exposure to glucocorticoids may have an adverse effect on subsequent blood pressure48 and brain maturation49. Our data do not suggest routine use of dexamethasone as a first line treatment for obstetric cholestasis, but we have provided evidence that it may have a beneficial effect in the prevention of intrauterine death in women who have not responded to ursodeoxycholic acid.

The results of the experiments reported in this paper imply that the clinical indications for the use ursodeoxycholic acid and dexamethasone in obstetric cholestasis include protection from the fetal complications of obstetric cholestasis in addition to treatment of the maternal symptoms and biochemical abnormalities. Some patients will be treated with a combination of ursodeoxycholic acid and dexamethasone and therefore it will be of interest to perform subsequent studies of pretreatment of cardiomyocytes with these agents in combination. In addition, it will be of interest to vary the concentration of ursodeoxycholic acid used in subsequent experiments. These experiments report the effect of 0.1 mM ursodeoxycholic acid, a similar but slightly higher concentration than has been reported in maternal serum in ursodeoxycholic acid-treated women (i.e. approximately 0.03 mM)18,49. The concentration of dexamethasone used was similar to the serum levels of dexamethasone in patients following intravenous or intramuscular boluses of the drug50.

The authors are not aware of any studies of the effect of high serum bile acids on the human heart. It would be of interest to study the effect of cholestasis in adults on cardiac function, and to investigate the effect of taurocholate on an in vitro model of the adult human heart.


We have demonstrated that dexamethasone, and to a lesser extent, ursodeoxycholic acid, protects against the arrhythmogenic effect of taurocholate in an in vitro model of the cardiac effects of bile acids.


Ventricular neonatal rat myocytes were kindly provided by Peter H. Sugden, National Heart and Lung Institute Division, Imperial College. This work was supported by grants from the Wellcome Trust, the Graham Dixon Cardiovascular Fund, the Biotechnology and Biological Science Research Council and the British Heart Foundation.