Potential conflict of interest: Nothing to report.
Tamoxifen is an anti-estrogen drug widely used for the treatment of hormone-sensitive breast cancer. Approximately 43% of breast cancer patients treated with tamoxifen develop hepatic steatosis. The mechanism or mechanisms by which tamoxifen may induce lipid accumulation in the liver are unclear. Mice were injected with tamoxifen or vehicle (sesame oil containing 1% benzyl alcohol) for 5 consecutive days. In comparison with the vehicle, tamoxifen increased hepatic triacylglycerol levels by 72%. The levels of plasma triacylglycerol were similar between the tamoxifen-treated and control groups. We found increased radiolabeling of triacylglycerol and phospholipids from [3H]acetate (∼50%) but not [14C]oleate in hepatocytes from tamoxifen-treated mice versus control mice. Fatty acid uptake, triacylglycerol secretion, and fatty acid oxidation remained unchanged in isolated hepatocytes after tamoxifen treatment. The apparent increase in fatty acid synthesis was explained by a marked decrease in the phosphorylation of acetyl coenzyme A carboxylase, which resulted in its activation. Conclusion: Our data suggest that increased de novo fatty acid synthesis is the primary event leading to tamoxifen-induced steatosis in the mouse liver. Inhibition of fatty acid synthesis might, therefore, ameliorate steatosis/steatohepatitis in breast cancer patients treated with tamoxifen. (HEPATOLOGY 2010 )
Tamoxifen (TMX) was developed in 1973 as an inhibitor of estrogens in the pathogenesis of breast cancer.1 In the United States, approximately 190,000 women were diagnosed with breast cancer in 2009, and 40,000 died.2 TMX reduces the rates of reappearance of breast cancer (47%) and death due to breast cancer (26%).3 TMX is widely used for the treatment of hormone-responsive breast cancer because of two significant advantage: it is inexpensive and is well tolerated.4 TMX is a selective estrogen receptor modulator that can act as either an estrogen agonist or an estrogen antagonist according to the tissue.4, 5 TMX acts as an estrogen agonist in bone3 and in the female lower genital tract.6 The benefits associated with the pro-estrogen activity of TMX are underscored by recent clinical trials designed to treat anovulatory infertility7 and prevent bone loss among postmenopausal women.8
One frequent side effect of TMX is the development of nonalcoholic fatty liver disease (NAFLD).9-12 Forty-three percent of women with breast cancer who are treated with TMX develop steatosis within the first 2 years of treatment9, 10; steatohepatitis12 and cirrhosis11 can also develop, particularly in overweight women. To date, the mechanism of TMX-induced steatosis remains unclear.
The effect of TMX on hepatic lipid metabolism has been investigated. Both the inhibition of fatty acid oxidation13, 14 and increased triacylglycerol (TG) biosynthesis15 have been implicated in the process. One possible explanation for the apparently different conclusions in these studies is that TMX is an appetite suppressor,16 yet the animals have not always been pair-fed.13, 15 Another possibility is that the dose (0.5-200 mg/kg/day) and duration (5-28 days) of TMX treatment varied among the studies, so lipid metabolism was investigated at different stages of TMX-mediated steatosis.13-15 Our aim was to elucidate the initial mechanism underlying the accumulation of hepatic TG after TMX treatment. Furthermore, it has been established that mice lacking the phosphatidylcholine (PC) biosynthetic enzyme phosphatidylethanolamine N-methyltransferase (PEMT) develop hepatic steatosis.17-19 PC synthesis is an important regulator of the secretion of TG into very low density lipoproteins from the liver.19-21 Therefore, we investigated whether TMX-mediated inhibition of PEMT22 contributes to the accumulation of hepatic TG. The data indicate that increased de novo fatty acid synthesis is the primary event leading to TMX-induced steatosis in the mouse liver.
ACC, acetyl coenzyme A carboxylase; AMPK, adenosine monophosphate–activated protein kinase; C, unesterified cholesterol; CE, cholesteryl ester; CPT1, carnitine palmitoyltransferase 1; CTα, cytidine triphosphate: phosphocholine cytidylyltransferase alpha; DGAT, diacylglycerol acyl coenzyme A acyltransferase; DMEM, Dulbecco's modified Eagle's medium; FA, fatty acid; FAS, fatty acid synthase; LCAD, long-chain acyl coenzyme A dehydrogenase; LXRα, liver X receptor alpha; MCD, malonyl coenzyme A decarboxylase; mRNA, messenger RNA; NAFLD, nonalcoholic fatty liver disease; PC, phosphatidylcholine; PDI, protein disulfide isomerase; PE, phosphatidylethanolamine; PEMT, phosphatidylethanolamine N-methyltransferase; PL, phospholipid; PPAR, peroxisome proliferator-activated receptor; PS/PI, phosphatidylserine/phosphatidylinositol; qPCR, quantitative polymerase chain reaction; SCD1, stearoyl coenzyme A desaturase 1; SEM, standard error of the mean; SRB1, scavenger receptor B1; SREBP-1C, sterol regulatory element binding protein 1C; TBP, TATA box binding protein; TG, triacylglycerol; TGH, triacylglycerol hydrolase; TMX, tamoxifen; UCP2, uncoupling protein 2; Veh, vehicle.
Materials and Methods
All procedures were performed in accordance with the University of Alberta Animal Policy and Welfare Committee, which adheres to the principles for biomedical research involving animals developed by the Council for International Organizations of Medical Sciences. Male C57BL/6J mice (Jackson Laboratory, Bar Harbour, ME) and male C57BL/6 Pemt−/− mice were 8 to 12 weeks old and were maintained on a standard chow diet containing 6% (wt/wt) fat and 0.02% (wt/wt) cholesterol (LabDiet). TMX (Sigma) was dissolved at a concentration of 0.2 mg/mL in sesame oil containing 1% benzyl alcohol. The mice were injected subcutaneously with either TMX (0.5 mg/kg of mouse) or vehicle (2.5 μL/g of mouse). TMX is a known appetite suppressor causing reduced food intake and weight gain in rodents.16 Therefore, the animals were pair-fed. The vehicle-injected group received an amount of the diet (g/g of mouse) corresponding to the average amount that the TMX-treated animals had eaten the day before. The mice were injected at the same time of day on 5 consecutive days and were anesthetized 4 hours after the final injection.
Tissue Preparation and Histology.
C57BL/6J mice were euthanized, and the liver was excised, fixed in 10% buffered formalin, cryosectioned (10 μm), and stained with Oil Red O to reveal neutral lipid accumulation. Photographs were taken with an AxioCam (Zeiss) mounted on a stereomicroscope (Axioskop) with a 40× objective lens.
Isolation of Hepatocytes.
Primary cultures of adult mouse C57BL/6J hepatocytes were isolated with the collagenase perfusion technique as previously described.23
Incubation of Primary Cultures of Mouse Hepatocytes with Radioisotopes.
Primary cultures of hepatocytes were isolated from vehicle-treated and TMX-treated C57BL/6J mice. After 2 hours of incubation at 37°C, the cells were rinsed with serum-free Dulbecco's modified Eagle's medium (DMEM) and incubated with 5 μCi of [3H]ethanolamine, 2 μCi of [14C]acetate, or 5 μCi of [3H]glucose per dish. The cells were harvested, washed twice with ice-cold phosphate-buffered saline (pH 7.4), scraped, and sonicated in 1 mL of phosphate-buffered saline. Hepatocytes were also incubated with 5 μCi of [3H]oleate per dish in the presence of 0.4 mM oleate and 0.4% bovine serum albumin. At the indicated times, the medium was removed and centrifuged for 5 minutes at 2500g to remove dead cells. For pulse-chase experiments, cells were incubated with [3H]oleate, washed twice with serum-free DMEM, and incubated for 4 hours in fresh serum-free DMEM in the absence of oleate. Total lipids were extracted from hepatocytes (1 mg of protein) or media (2 mL),24 and the incorporation of radioisotopes was determined.
Data are expressed as means and standard errors of the mean (SEMs) unless otherwise stated. Comparisons between treatment groups were performed with the unpaired Student two-tailed t test for most analyses. One-way analysis of variance with Tukey post hoc analysis was used to compare differences in experiments with primary cultures of hepatocytes. A probability value <0.05 was considered significant. Three to eight samples were examined for each measurement.
Body Weight and Food Intake During TMX Treatment.
TMX is an appetite suppressor causing reduced food intake and weight gain in rodents.14, 16 Therefore, the mice were pair-fed. Consequently, there was no significant difference in the food intake or body weight between the two experimental groups (Supporting Information Fig. 1).
After TMX treatment, the hepatic TG level was 72% higher than that in control mice (Fig. 1A), whereas the amount of hepatic cholesterol or cholesteryl ester (CE) between the experimental groups was not different (Fig. 1B). The accumulation of hepatic TG in TMX-treated mice was confirmed by Oil Red O staining of liver sections (Fig. 1C). Thus, our model recapitulates the hepatic steatosis observed in breast cancer patients treated with TMX.
TMX Does Not Decrease Plasma TG.
TMX did not significantly change the total plasma TG level (Supporting Information Fig. 2A,B). However, the level of plasma CE (Supporting Information Fig. 2C) was 20% lower in TMX-treated mice versus vehicle-treated mice, and this correlated with a slight reduction in total cholesterol from the high-density lipoprotein fraction of TMX-treated mice (Supporting Information Fig. 3D).
Messenger RNAs (mRNAs) in Livers of TMX-Treated Mice.
Next, we determined whether hepatic lipogenic genes were elevated in response to TMX (Fig. 2). The levels of mRNAs encoding fatty acid synthase (FAS) and stearoyl coenzyme A desaturase 1 (SCD1) were increased by 200% and reduced by 60%, respectively, by TMX.
Genes associated with fatty acid β-oxidation were not different between the treatment groups (Fig. 2). Similarly, the two animal groups had comparable mRNA levels of cytidine triphosphate:phosphocholine cytidylyltransferase alpha (CTα), peroxisome proliferator-activated receptor delta (PPARδ), and scavenger receptor B1 (SRB1). The expression of uncoupling protein 2 (UCP2) was significantly up-regulated in response to TMX. Furthermore, TMX increased expression of the mRNA encoding fatty acid translocase CD36 and significantly reduced the mRNA level of the PC biosynthetic enzyme PEMT (Fig. 2).
Inhibition of PEMT Does Not Contribute to TMX-Mediated Steatosis.
The quantitative polymerase chain reaction (qPCR) analysis of hepatic mRNAs revealed that the level of PEMT mRNA was lower in the TMX-treated animals compared to the controls (Fig. 2). Figure 3A,B shows that the amounts of PEMT protein and activity in the livers of mice treated with TMX were significantly lower (40%-60%) versus the control mice. However, the amount of hepatic PC or the PC/phosphatidylethanolamine (PE) ratio was not different (Fig. 3C).
To determine whether or not the TMX-mediated inhibition of PEMT reduced hepatic TG secretion, we used primary hepatocytes isolated from TMX-treated and vehicle-treated mice. Despite a significant reduction of the formation of [3H]PC from [3H]ethanolamine in hepatocytes isolated from TMX-treated mice (Fig. 3D), TG secreted into the culture medium was not altered by TMX treatment (Fig. 3E). In contrast, when poloxamer 407 was used to block lipoprotein catabolism and assess hepatic TG secretion in vivo, TMX-treated mice accumulated 31% less plasma TG (Supporting Information Fig. 3). To clarify whether inhibition of PEMT contributes to TMX-mediated hepatic steatosis, we used Pemt−/− mice. After TMX treatment for 5 consecutive days, Pemt−/− mice had accumulated 73% more hepatic TG than the vehicle-treated Pemt−/− mice with no change in cholesterol or CE (Fig. 3F). Thus, the TMX-induced steatosis is not due to the decreased activity of PEMT.
Hepatic Fatty Acid Synthesis Is Increased by TMX.
The real-time qPCR analysis identified several candidate lipid pathways that may be targets of TMX action in the liver. We examined the involvement of each pathway with primary cultures of hepatocytes isolated from mice treated for 5 days with either vehicle or TMX. The mRNA level of the fatty acid translocase CD36 in the livers of TMX-treated mice was higher than that in control mice (Fig. 2). Therefore, primary cultures of hepatocytes isolated from vehicle-treated and TMX-treated mice were incubated with [3H]oleate. The rate of transport of [3H]oleate into hepatocytes was similar in the two treatment groups (Supporting Information Fig. 4), and this suggests that increased fatty acid uptake is not responsible for TMX-mediated lipid accumulation in the liver.
To ascertain whether TMX promotes hepatic TG accumulation by increasing de novo fatty acid synthesis, we incubated primary cultures of hepatocytes with [14C]acetate. The incorporation of [14C]acetate into TG was 60% higher in hepatocytes from TMX-treated mice versus control hepatocytes (Fig. 4A). Similarly, the incorporation of [14C]acetate into PC, PE, and phosphatidylinositol/phosphatidylserine was increased (50%-75%) by TMX treatment (Fig. 4B-D). In contrast, TMX did not alter the incorporation of [14C]acetate into cholesterol or CE (Fig. 4E,F). The increased incorporation of acetate into TG and phospholipids indicates that fatty acid synthesis is elevated in response to TMX.
To confirm that the TMX-induced increase in the radiolabeling of glycerolipids from [14C]acetate reflected an increase in fatty acid synthesis rather than an increase in glycerolipid synthesis, we incubated hepatocytes isolated from vehicle-treated and TMX-treated mice with [3H]oleate. The formation of [3H]TG from oleate in the TMX-incubated hepatocytes was not significantly different from that in control hepatocytes (Supporting Information Fig. 5A). In addition, the incorporation of [3H]oleate into PC, PE, and phosphatidylinositol/phosphatidylserine was similar in hepatocytes with or without TMX (Supporting Information Fig. 5B-D). Thus, radiolabeling experiments with hepatocytes indicate that TMX enhances fatty acid synthesis, which subsequently promotes hepatic steatosis.
TMX Decreases the Phosphorylation of Acetyl Coenzyme A Carboxylase (ACC) in the Liver.
The rate-limiting enzyme in de novo fatty acid synthesis, ACC, is activated upon dephosphorylation at Ser-79.25, 26 With an antibody specifically recognizing Ser-79 (Fig. 5A), ACC phosphorylation (Ser-79) was reduced to undetectable levels after TMX treatment, whereas there was no change in the total amount of ACC protein or FAS protein upon TMX treatment (Fig. 5A). In an attempt to explain the decreased phosphorylation of ACC, we monitored the activity of the upstream kinase, adenosine monophosphate–activated protein kinase (AMPK).25 Phosphorylation of the AMPK catalytic subunit (AMPK-α) at Thr-172 is essential for AMPK activation25; therefore, we used an antibody that specifically recognized AMPK phosphorylation at Thr-172. The level of AMPK phosphorylation was significantly decreased by TMX (Fig. 5B). Together, these results suggest that TMX-induced lipogenesis is due to attenuated inhibition of ACC by AMPK.
Reduced AMPK activity is associated with impaired fatty acid oxidation.25 However, when we measured the rate of oxidation of [3H]oleate in hepatocytes from vehicle-treated and TMX-treated mice, the same amount of acid-soluble metabolites was released into the media (Fig. 5C). Thus, TMX does not appear to inhibit fatty acid oxidation.
This is the first study addressing the antecedent mechanisms by which TMX promotes hepatic lipid accumulation in mice that is not confounded by differences in food intake. The results demonstrate that a primary effect of TMX is to increase hepatic fatty acid synthesis via activation of AMPK and ACC.
We have found that TMX promotes hepatic steatosis by increasing lipogenesis at least in part through the reduction of the phosphorylation (Ser-79) of the rate-limiting enzyme of fatty acid synthesis (ACC). Reduced phosphorylation of ACC in the liver activates fatty acid synthesis.26 In the current study, AMPK phosphorylation (at Thr-172) decreased in response to TMX; this indicated that AMPK kinase activity was reduced,25 and this reduction accounted for the decreased phosphorylation of ACC.26 Interestingly, estrogen has been shown to reduce adiposity in pair-fed ovariectomized mice through the activation of AMPK (Thr-172) in myocytes.27 TMX might inhibit AMPK activity directly through the classical nuclear estrogen receptors (estrogen receptor alpha and estrogen receptor beta)28 or the more recently described G-protein receptor 30.29 Estrogen activates multiple cellular kinase pathways, including those involving the epidermal growth factor receptor and protein kinase A, via estrogen receptors and GRP30.29 Both of these cell signaling pathways have been previously implicated in the regulation of AMPK phosphorylation and activity.30, 31
In previous reports, TMX reduced the hepatic activity of fatty acid biosynthetic enzymes ACC and FAS while simultaneously increasing the activity of the TG biosynthetic enzyme glycerol-3-phosphate acyltransferase.14, 15 The apparent contradiction with the current study likely reflects the difference between in vitro enzymatic assays using homogenates and measurements obtained from intact hepatocytes. Increases in both the activity of SCD1 and the ratio of its product (18:1n-9) to the substrate (18:0) have been implicated in NAFLD.32 Although the level of SCD1 mRNA has been consistently reduced with TMX treatment,14, 15 the hepatic 18:1n-9/18:0 ratio has remained unchanged.15 Interestingly, sterol regulatory element binding protein 1C (SREBP-1C), which is a determinant of the biosynthesis of fatty acids because it promotes the expression of lipogenic genes such as FAS and ACC, was not changed in the current study in response to TMX. This result is consistent with several other reports on TMX action14, 33 and might reflect the complex effects of TMX and its metabolites on gene expression and cell signalling.29
We found that TMX did not alter the oxidation of [3H]oleate in hepatocytes. We also found no difference in the levels of hepatic mRNAs encoding enzymes involved in fatty acid oxidation. These results indicate that the accumulation of TG in the liver after TMX treatment is not due to reduced fatty acid oxidation. Our results are consistent with those showing no change in palmitoyl coenzyme A oxidation or carnitine palmitoyltransferase 1 (CPT1) activity in the livers of rats treated orally with TMX (40 mg/kg/day) versus vehicle.15 The previously reported TMX-mediated inhibition of fatty acid oxidation may reflect the much larger oral dose of TMX (200 mg/kg/day) used.13 It is likely that inhibition of fatty acid oxidation does occur at a later stage of TMX-mediated steatosis because increased fatty acid synthesis typically results in the inhibition of fatty acid oxidation.25 Furthermore, there is evidence that estrogen might regulate fatty acid oxidation. Mice deficient in estrogen34 have impaired hepatic fatty acid oxidation, which can be prevented with 17β-estradiol replacement.
In our study, the increase in hepatic TG after TMX treatment for 5 days was not accompanied by a decrease in plasma TG levels; this was similar to the findings in rats treated with TMX (0.5 mg/kg/day) for 5 days.14 Several other publications have shown decreased serum TG levels after a longer duration (14-84 days) of TMX treatment15, 35 or with a large oral dose of the drug (∼370 mg/kg/day).13 Administering TMX more aggressively dramatically reduced hepatic TG secretion (∼70%) by an unknown mechanism.13 In our study, the modest decrease in the secretion of TG (31%) likely contributed to the development of hepatic steatosis. However, we were unable to detect any difference in TG secretion between hepatocytes isolated from vehicle-treated mice and hepatocytes isolated from TMX-treated mice; this suggests that the 60% increase in TG synthesis is the predominant effect.
It has been established that mice lacking PEMT develop hepatic steatosis.17-19 On the basis of the TMX-mediated decrease in the mRNA, protein, and activity of PEMT, we anticipated a major role for the enzyme in TMX-mediated steatosis. However, the magnitude of the TMX-mediated increase in hepatic TG levels was similar after TMX treatment of Pemt−/− mice (73%; Fig. 3F) and Pemt+/+ mice (72%; Fig. 1A). These results suggest that the decrease in PEMT expression is not a major initial contributor to TMX-mediated hepatic steatosis. This result might be explained by the absence of any significant decrease in hepatic PC levels after TMX treatment. Previously, it has been shown that mice heterozygous for the Pemt gene are not susceptible to hepatic steatosis.36
Several studies have indicated that TMX treatment is associated with the development of NAFLD.9-12 This is supported by the current study, which showed that TMX induced hepatic steatosis in mice. In overweight and obese women, TMX treatment has been found to be a significant independent predictor for the development of NAFLD that increases the risk of developing steatohepatitis 2-fold versus placebo.12 On the basis of the prevalence of obesity, a known risk factor for breast cancer,37 and the association of NAFLD with both insulin resistance and cardiovascular disease,38 establishing the mechanism by which TMX induces NAFLD will aid in risk factor management and in achieving maximal pharmaceutical benefit for patients with estrogen receptor–positive breast cancer.
In conclusion, we have shown that TMX causes hepatic steatosis in mice primarily by increasing fatty acid synthesis. We have identified activation of ACC as a primary target of TMX. We suggest that the inhibition of fatty acid synthesis may prove to be beneficial in reducing hepatic steatosis induced by TMX.
The authors thank Susanne Lingrell, Russ Watts, Audric Moses, and Elviche Lenou for technical assistance and Professors Andrew Mason and Jean Vance for helpful comments on the article.