Tamoxifen Decreases Cholesterol Sevenfold and Abolishes Lipid Lesion Development in Apolipoprotein E Knockout Mice
Background Apolipoprotein E (apo E) knockout mice develop severe vascular lipid lesions resembling human atherosclerotic plaques, irrespective of the fat content of their diet.
Methods and Results Oral tamoxifen (TMX) at a dose of 1.9 mg·kg body wt−1·d−1 abolished lipid lesion development, assayed by oil red O staining, whether the mice were fed a normal diet or a diet with high fat content. The TMX-treated mice showed a sevenfold decrease in total cholesterol. However, the proportion of plasma cholesterol present in VLDL remained unchanged, whereas the proportion in LDL decreased by 37%, and that in HDL increased by 64%. Consistent with the shift from LDL to HDL cholesterol, there was a 62% decrease in total triglycerides. The concentrations of active and acid-activatable latent plus active TGF-β in the aorta were substantially elevated by TMX (87% and 24% increase, respectively).
Conclusions Although the mechanism of cardiovascular protection by TMX in apo E knockout mice is unknown, the inhibition of lipid lesion formation may be attributable to the changes in lipoprotein profile and the elevated levels of TGF-β, both of which are thought to be protective against atherosclerosis in humans and animal models.
When C57BL/6 mice are fed a high-fat diet, they develop fatty-streak lesions1 2 3 4 resembling early human atherosclerosis. However, if their diet is supplemented with TMX (1 mg·kg body wt−1·d−1), there is a substantial suppression of diet-induced lesion formation.1 Furthermore, in two separate studies, both Macdonald and Stewart5 and Rutqvist and Mattsson6 found that there was a significant decrease in death from myocardial infarction in postmenopausal women given TMX as adjuvant therapy after operable breast cancer. Taken together, these studies suggest that TMX therapy may be cardioprotective. However, the mechanism(s) by which TMX prevents vascular fatty-streak lesion development are unclear.
One possible mechanism through which TMX exerts its cardioprotective effects is modulation of the lipoprotein profile. In postmenopausal women5 6 7 8 and men with atherosclerosis (D.J. Grainger, J.C. Metcalfe, A.A. Grace, M.C. Petch, H.W. Bethell, unpublished data, 1996), TMX decreased total plasma cholesterol and also decreased the proportion of plasma cholesterol present in the most triglyceride-rich particles (VLDL and LDL). Epidemiological studies have suggested that lower levels of cholesterol in VLDL and LDL, together with higher levels in HDL, are associated with reduced risk of coronary artery disease.8 9 10 Consequently, the effects of TMX on the lipoprotein profile would be expected to reduce lesion development. Consistent with these observations, C57BL/6 mice fed a normal diet have the majority of their plasma cholesterol in HDL and do not develop fatty-streak lesions.1 10 When these mice are fed a high-fat diet, there is a substantial increase in VLDL and a decrease in HDL associated with lesion development.1 Inclusion of TMX in the high-fat diet slightly reduced total plasma cholesterol levels in C57BL/6 mice,1 but almost all of the observed reduction in cholesterol occurred in the HDL fraction.1 In contrast to the human studies, VLDL was not decreased but rather marginally elevated. We concluded that modulation of the lipoprotein profile was unlikely to be the major mechanism by which TMX prevented diet-induced lipid lesion formation in C57BL/6 mice,1 because the TMX-induced changes were likely to be atherogenic.
Since TMX is known to elevate TGF-β production11 12 13 by both smooth muscle and breast tumor cells in vitro and in vivo, we postulated that the cardioprotective effects of TMX are at least in part due to increased TGF-β activity. According to the “protective cytokine” hypothesis,14 TGF-β activity is necessary for the maintenance of normal vessel wall structure, preventing the activation of both endothelial cells and smooth muscle cells.15 Furthermore, loss of TGF-β activity has been correlated with the development of fatty-streak lesions. We postulate that this may allow smooth muscle cell migration and proliferation as well as invasion of inflammatory cells.15 Thus, TMX would be expected to reduce lesion formation by increasing TGF-β activity.
If TMX is to be clinically useful for prevention of human atherosclerosis, which presumably results from diverse genetic causes, then it must be cardioprotective irrespective of the genetic basis for susceptibility. We therefore examined the effects of TMX on lesion development in other mouse models of atherosclerosis, including the apolipoprotein(a)15A transgenic mouse and the apo E knockout mouse. The apo E knockout mice are a model of more severe human atherosclerosis than the C57BL/6 mice.16 17 Apo E knockout mice exhibit spontaneous elevation of total plasma cholesterol and triglycerides and reduced levels of HDL on a diet with normal fat content.17 18 As a result, they develop massive lipid-filled lesions throughout the arterial tree by 3 to 4 months of age.17 19 By 6 months of age, the vascular lesions in apo E knockout mice resemble atherosclerotic lesions in humans, with a similar pattern of lesion distribution, microscopic appearance, and cellular composition.17 18 For example, the lesions develop a raised intima with a fibrous cap and have dramatic infiltration by macrophages and other inflammatory cell types.17 18 When apo E mice are fed a diet with high fat content, lesion development occurs more swiftly and to a greater extent.
The aim of the present study was to determine whether TMX is likely to be cardioprotective irrespective of the genetic basis of susceptibility and irrespective of the severity of the disease. We examined the effects of TMX on lesion development, lipoprotein profile, and TGF-β production in the apo E knockout mouse.
Treatment of Apo E Knockout Mice With TMX
The mice used in this study were progeny of those initially described by Piedrahita and colleagues.20 At the start of the experiment (day 0), 45 male apo E knockout mice (≈3 months old) were weighed and randomly allocated into five groups. One group of five mice was killed at day 0, and the remainder were split into four equal groups. Each group of mice was fed ad libitum either normal mouse chow (ICN Pharmaceuticals Inc), or normal mouse chow containing 15 μg TMX (Aldrich) per gram of food, or a high-fat diet containing 2.5% cholesterol and 7.5% saturated fat as cocoa butter, 7.5% casein, and 0.5% sodium cholate (ICN Pharmaceuticals Inc), or the high-fat diet containing 15 μg TMX/g food. Water was freely available throughout the course of the experiment. Weight and food intake were analyzed at regular intervals during the treatment.
On the day the animals were killed, each mouse was weighed, and the heart and attached aorta were dissected, embedded in Cryo-M-bed (Bright Instrument Co), and immediately snap-frozen in liquid nitrogen. Frozen sections (4 μm) were prepared as described previously.21 Blood was collected at the time of death, and serum was prepared separately from each mouse by allowing the blood to stand for 2 hours at room temperature followed by centrifugation at 5000g for 10 minutes. Aliquots were frozen at −20°C until assayed.
Analysis of Lipid Lesion Formation by Oil Red O Staining
Sections from the aortic sinus region were collected according to the strategy of Paigen and coworkers.21 For each mouse, five sections, each separated by 80 μm, were fixed in 10% buffered formalin, stained with oil red O, and counterstained with light green as described.1 21 The area of oil red O staining in each section was quantified with a calibrated eyepiece, excluding droplets <50 μm2, and the mean lesion area per mouse and per each group of mice was calculated. In addition, regions of focal oil red O staining >500 μm2 were defined as lipid lesions, and the number of such lesions per section per mouse was determined.
Analysis of Lipoproteins
For each group of mice, a single sample (1 mL) of serum pooled from every mouse in the group was made up to a density of 1.215 g·mL−1 with KBr and subjected to density gradient ultracentrifugation at 4°C for 48 hours as previously described.1 A 0.2-mL sample of the lipoprotein fraction (d<1.215 g·mL−1) was gel-filtered through a Sepharose 6B column by FPLC at room temperature as described.1 A 100-μL sample of each 0.4-mL fraction from the column was assayed for total cholesterol with a kit for the enzymatic determination of cholesterol (Sigma Diagnostics) in accordance with the manufacturer’s instructions,22 except that the reactions were performed in ELISA plate wells (Maxisorp, Gibco BRL) in a total volume of 300 μL. Absorbance (492 nm) was measured every 15 minutes until no further change occurred. Consistent with previous studies,1 23 FPLC fractions 1 through 9 contain the VLDL class, fractions 10 through 19 contain LDL, and fractions ≥20 contain HDL.
Assays for Serum Triglycerides and Cholesterol
Total triglycerides in serum from each mouse were determined with the Triglyceride-UV kit (Sigma Diagnostics) based on the glycerol kinase enzymatic method.24 Total serum cholesterol was measured for each mouse by the cholesterol oxidase method (Sigma Diagnostics) in ELISA plates as described above. All values are expressed as the mean±SEM for each group.
Measurement of TGF-β in Aortic Vessel Wall Sections
Active and a+l TGF-β in the vessel wall were measured by quantitative immunofluorescence as described previously.15 Briefly, 4-μm frozen aortic sections adjacent to those quantified for lipid lesion area were stained for TGF-β. Active TGF-β was measured with fluorescein-labeled R2X.14 24 25 a+l TGF-β was measured with BDA19 as primary antibody at a concentration of 25 μg·mL−1 (AB-101-NA, R and D Systems) and fluorescein-labeled anti-chicken IgY (Jackson Immunoresearch Laboratories Inc) as the secondary antibody at 75 μg·mL−1.
Measurement of Smooth Muscle α-Actin and Osteopontin in Aortic Vessel Wall Sections
Frozen sections (4 μm) were stained for smooth muscle α-actin and osteopontin as previously described,1 except that each section was preincubated in a humidified chamber at 37°C for 24 hours with Affinipure (Fab′)2 fragment of donkey anti-mouse IgG with minimal cross-reactivity (Jackson Immunoresearch Laboratories Inc) at 160 μg·mL−1 to reduce background staining.
The nonparametric Mann-Whitney U test for unpaired samples was used throughout this study for two-group comparisons, since too few measurements have been made to determine whether the parameters are normally distributed. Since the follow-up times were different for the high-fat diet and normal diet groups, no statistical tests were applied to comparisons between mice receiving the different diets. The level of statistical significance was set at P<.05.
Effect of Tamoxifen on Lipid Lesion Development
Two groups of 10 apo E knockout mice each were fed either a normal mouse chow diet or the normal diet supplemented with 15 μg TMX/g food. The mice taking TMX received, on average, 1.9±0.4 mg·kg body wt−1·d−1. Twelve weeks later, when they were killed, the mice fed a normal diet had gained 3% in body weight (Table 1⇓). In contrast, mice fed the normal diet plus TMX showed a marked decrease (21%) in their body weight. However, the mice fed the normal diet and the mice fed the normal diet plus TMX consumed very similar amounts of food (Table 1⇓). Therefore, the loss in body weight during TMX treatment could not be attributed to reduced calorie intake.
Two more groups of 10 mice each were fed either the high-fat diet or the high-fat diet plus TMX. The mice fed the high-fat diet without TMX showed morbidity by day 53, and on veterinary advice, both groups fed the high-fat diets were killed. These mice showed a trend in body weight change similar to that of the groups fed the normal diet (1% increase on the high-fat diet alone and 19% decrease in mice fed the high-fat diet plus TMX; Table 2⇓). However, because the duration of the experiment was different in each case, it was not possible to directly compare the effects of TMX on mice fed a high-fat diet with its effects on mice fed a normal diet.
According to the sectioning strategy of Paigen and coworkers,21 five sections from each mouse were analyzed for the development of fatty vascular lesions. The total area stained with oil red O (excluding droplets <50 μm2) and the number of lesions in the blood vessel wall at the aortic sinus region were determined (Fig 1⇓). Apo E knockout mice ≈12 weeks old (day 0 of the experiment) had large areas of oil red O staining in the aortic sinus region (8145±1402 μm2 per mouse [n=5], consistent with previously published data17 ). After a further 3 months on the normal diet, there was a threefold increase in the number of lesions (13.7±2.1 compared with 4.8±1.1, where a lesion is defined as contiguous regions of lipid staining >500 μm2) in the aortic wall, resulting in a significant increase in the area of oil red O staining (Fig 1⇓). In marked contrast, in the mice fed the normal diet plus TMX, there was no increase in the number of lesions after 3 months of treatment compared with mice killed at day 0 (Table 1⇑). The increase in the area staining with oil red O over the 3 months of the experiment evident in mice fed a normal diet was abolished in mice receiving normal diet plus TMX (4±0.9 for normal diet plus TMX and 13.7±2.1 for normal diet alone, Table 1⇑).
The effect of TMX was very similar in mice receiving a high-fat diet, such that the increase in area stained with oil red O over the 53 days of the experiment was 96% lower in mice receiving TMX (9433±324 μm2) than in mice receiving high-fat diet alone (38 708±2597 μm2, Table 2⇑).
Effect of Tamoxifen on Lipoproteins
The effect of TMX on the lipoprotein profile in apo E knockout mice was examined. For each group of apo E knockout mice, we constructed a lipoprotein profile by separating the lipoprotein classes by gel filtration and measuring total cholesterol in each fraction, as previously described.1 In mice fed a normal diet for 3 months, there was a moderate increase in total serum cholesterol compared with day 0 (445±6 mg·dL−1 compared with 383±6 mg·dL−1, Table 1⇑), in accordance with published results.17 18 In contrast, if apo E knockout mice were fed the normal diet plus TMX, there was a sevenfold decrease in total cholesterol (63±3 mg·dL−1, Table 1⇑, Fig 2⇓), and levels were reduced to concentrations reported for wild-type mouse strains1 26 fed a diet with normal fat content. The proportion of total cholesterol present in VLDL was unchanged by the presence of TMX, but there was a decrease in the proportion present in LDL, with a corresponding increase in the proportion in HDL (Table 1⇑). Consistent with the shift from LDL to HDL cholesterol, there was a 62% decrease in total triglycerides in mice fed the normal diet plus TMX compared with mice fed a normal diet for the same period (150±26 mg·dL−1 compared with 392±34 mg·dL−1, Table 1⇑). This is in marked contrast to the effect of TMX treatment in fat-fed C57BL/6 mice, in which HDL cholesterol was depressed and triglyceride levels were significantly elevated.1
Similar trends in the concentrations of both total cholesterol and triglycerides (reductions of 52% and 28%, respectively) were observed in mice fed the high-fat diet plus TMX (Table 2⇑, Fig 2⇑). It is plausible that the magnitude of the effects is smaller because the duration of feeding was shorter (53 days on the high-fat diets compared with 84 days on the normal diets). No change in the relative proportions of LDL cholesterol and HDL cholesterol was noted when mice on the high-fat diet were treated with TMX (Table 2⇑).
Effect of Tamoxifen on TGF-β Activity
TMX has been reported to increase TGF-β activity in the vessel wall of C57BL/6 mice,1 and its effect on TGF-β activity in the vessel wall of apo E knockout mice was therefore examined. Both active and a+l TGF-β assayed by quantitative immunofluorescence were unchanged after 3 months on the normal diet compared with day 0 (Table 1⇑). However, there was a small but significant increase in a+l TGF-β (47±6 AU compared with 38±3 AU [+24%]) after 3 months on the normal diet plus TMX (Table 1⇑) and a larger increase in active TGF-β (112±10 AU compared with 60±3 AU [+87%]). These changes are similar to the increase in plasma a+l TGF-β (+22%) and active TGF-β (+169%) seen after TMX treatment of men with atherosclerosis (D.J. Grainger, J.C. Metcalfe, A.A. Grace, M.C. Petch, H.W. Bethell, unpublished data, 1996).
Similar trends in a+l and active TGF-β were seen in the vessel wall of mice fed a high-fat diet. There were no significant changes in levels of active or a+l TGF-β over a period of 53 days on a high-fat diet, but there was a 93% increase in active and a 40% increase in a+l TGF-β in mice fed a high-fat diet plus TMX over the same period (Table 2⇑).
Effect of Tamoxifen on Cellular Differentiation in the Vessel Wall
TGF-β has previously been suggested to control the differentiation state of smooth muscle cells.1 14 27 Smooth muscle α-actin, an essential component of the contractile apparatus of the smooth muscle cell, has been used as a marker of smooth muscle differentiation.28 Smooth muscle α-actin staining in the vessel wall did not change significantly over the period of the experiment in mice fed either the normal or high-fat diet alone. However, the addition of TMX to the diet caused a 50% increase (190±7 AU compared with 127±4 AU, Table 1⇑) and a 70% increase (189±9 AU compared with 111±6 AU, Table 2⇑) in staining for smooth muscle α-actin in the mice fed the normal and high-fat diets, respectively (n=10, P<.0001, both experiments). Thus, an increase in smooth muscle differentiation is correlated with the increase in TGF-β activity after TMX therapy.
By contrast, osteopontin, a marker of lesion development expressed by dedifferentiated smooth muscle cells29 30 and macrophages, showed an inverse staining pattern to that for smooth muscle α-actin. Large accumulations of osteopontin were detected in mice fed both the normal and high-fat diets without TMX (Fig 3⇓) at sites at which significant lesions marked by increased oil red O staining developed (Tables 1⇑ and 2⇑). In contrast, osteopontin staining was lower at the end of the experiment in both groups of mice that received TMX (66% decrease in mice fed a normal diet plus TMX [13±1 AU compared with 38±7 AU] and 37% decrease in mice fed a high-fat diet plus TMX [24±2 AU compared with 38±7 AU]) than in mice killed at day 0 (Tables 1⇑ and 2⇑, Fig 3⇓).
The most striking observation from this study is that TMX is able to abolish the development of lipid-filled vascular lesions in apo E knockout mice. The area of oil red O staining in the vessel wall in mice fed the normal diet plus TMX was slightly less than that in the apo E mice at the start of the experiment (day 0). Thus, TMX has a dramatic cardioprotective effect on the apo E knockout mice. This is consistent with studies showing that TMX is able to reduce diet-induced lesions by 88% in C57BL/6 mice1 and the incidence of fatal myocardial infarction among postmenopausal women receiving TMX as adjuvant therapy for breast cancer.5 6
TMX treatment lowered total plasma cholesterol in C57BL/6 mice as well as in women with breast cancer and men with atherosclerosis5 6 7 8 (D.J. Grainger, J.C. Metcalfe, A.A. Grace, M.C. Petch, H.W. Bethell, unpublished data, 1996). We show here that TMX also lowers total plasma cholesterol in apo E knockout mice, but the magnitude of the effect is much larger (sevenfold reduction in apo E mice compared with 10% reduction in C57BL/6 mice). This may be because apo E knockout mice, unlike C57BL/6 mice or the patient groups studied to date, have massive hypercholesterolemia and hypertriglyceridemia. This very large reduction in total plasma cholesterol in TMX-treated apo E knockout mice masks the effects of TMX on the lipoprotein profile, which are consistent with previous studies. Caleffi and coworkers31 reported decreased plasma cholesterol and an increase in HDL in premenopausal women with breast pain. Similarly, administration of TMX to male patients with severe atherosclerosis over a period of 10 days resulted in a reduction in the total cholesterol levels, with a significant decrease (18%) in triglyceride-rich lipoprotein (D.J. Grainger, J.C. Metcalfe, A.A. Grace, M.C. Petch, H.W. Bethell, unpublished data, 1996).
In apo E knockout mice treated with TMX, the fraction of cholesterol in HDL was increased (even though the total HDL cholesterol concentration was markedly reduced, presumably due to the very large decrease in total plasma cholesterol). Similarly, the decrease in plasma triglycerides reflects the decrease in total plasma cholesterol, because two thirds of the lipoprotein particles in apo E knockout mouse plasma are triglyceride rich. We reported an increase in plasma triglycerides in C57BL/6 mice treated with TMX,1 presumably again reflecting the decrease in total plasma cholesterol, because in this mouse strain, much of the cholesterol is present in the triglyceride-poor HDL particles. These observations suggest that the effects of TMX on triglycerides are secondary to the effects on total plasma cholesterol.
Mechanism of Cardioprotection by TMX: Cholesterol Lowering?
The changes in total plasma cholesterol and in the lipoprotein profile of apo E knockout mice treated with TMX are sufficient to explain the absence of lesion development. Epidemiological studies in humans9 and animal models10 32 have demonstrated that lowered HDL levels increase the risk of lesion formation and that the high HDL-to-LDL ratio in mice may contribute to their resistance to lipid lesion formation. Thus, the increased fraction of cholesterol in HDL would be expected to contribute to reduced lesion formation. It is interesting to note that statin-based lipid-lowering agents (such as simvastatin) are much less effective at lowering total plasma cholesterol in apo E knockout mice (D.J. Grainger, unpublished data, 1996). This is consistent with our findings that TMX is a more effective lipid-lowering agent in men with atherosclerosis (D.J. Grainger, J.C. Metcalfe, A.A. Grace, M.C. Petch, H.W. Bethell, unpublished data, 1996) than studies have shown for pravastatin (20% decrease in total cholesterol)33 and simvastatin (25% decrease in total cholesterol).34
Mechanism of Cardioprotection by TMX: Reduced Calorie Intake?
Calorie restriction has been shown to slow the genesis of several experimental diseases and can modify the lipoprotein profile in humans.35 Consistent with the effects of TMX on C57BL/6 mice and apolipoprotein(a) mice,15A TMX reduced the body weight of treated mice (Tables 1⇑ and 2⇑). However, there was no significant difference in the quantity of food consumed by mice fed the normal diet or normal diet plus TMX (Table 1⇑). Thus, the TMX-induced weight loss is presumably attributable to the marked effects of TMX on lipoprotein metabolism (Fig 2⇑ and Tables 1⇑ and 2⇑) and does not result from an altered dietary intake.
Mechanism of Cardioprotection by TMX: Elevation of TGF-β?
Consistent with previous studies of TMX-treated mice and humans1 13 (D.J. Grainger, J.C. Metcalfe, A.A. Grace, M.C. Petch, H.W. Bethell, unpublished data, 1996), TMX stimulated TGF-β activity in apo E knockout mice. Both active and a+l TGF-β are elevated in the vessel wall of the mice treated with TMX. According to the “protective cytokine” hypothesis,13 TGF-β activity will promote the differentiated smooth muscle cell phenotype and therefore maintain the structure of the normal vessel wall. Consistent with this hypothesis, we observed increased smooth muscle α-actin and reduced osteopontin accumulation (Tables 1⇑ and 2⇑, Fig 3⇑) in apo E mice treated with TMX. It is plausible that elevation of TGF-β activity by TMX contributes to the reduction in lipid lesion development.
Irrespective of the mechanism of action, the cardioprotective effects of TMX are now clearly established in three mouse models of atherosclerosis and in postmenopausal women administered TMX as adjuvant therapy after operable breast cancer.1 5 6 7 8 Since TMX inhibits lipid-lesion development in C57BL/6 mice,1 apo E knockout mice, and apolipoprotein(a) mice15A as well as reducing the incidence of fatal myocardial infarction in women, we conclude that TMX is cardioprotective irrespective of the genetic basis of susceptibility to atherosclerosis. It is possible that newer analogues of TMX will retain this cardioprotective property while eliminating the recently documented36 carcinogenic effects associated with TMX.
Selected Abbreviations and Acronyms
|a+l||=||acid-activatable latent plus active|
|FPLC||=||fast protein liquid chromatography|
|TGF-β||=||transforming growth factor-β|
|TMX||=||tamoxifen free base|
This work was funded by a programme grant from the British Heart Foundation and a Wellcome Trust grant to Dr Metcalfe and Dr Grainger, who is a Royal Society university research fellow. We are grateful to Glaxo Wellcome for providing the apo E knockout mice for this study.
J.R. and D.J.G. are now at Department of Medicine, University of Cambridge, UK.
- Received May 30, 1996.
- Revision received November 4, 1996.
- Accepted November 14, 1996.
- Copyright © 1997 by American Heart Association
Paigen B, Ishida B, Verstuyft J, Winters RB, Albee D. Atherosclerosis susceptibility differences among progenitors of recombinant inbred strains of mice. Arteriosclerosis. 1990;10:316-323.
MacDonald CC, Stewart HJ. Fatal myocardial infarction in the Scottish adjuvant tamoxifen trial. Br Med J. 1991;303:435-437.
Rutqvist LE, Mattsson A. Cardiac and thromboembolic morbidity among postmenopausal women with early-stage breast-cancer in a randomized trial of adjuvant tamoxifen. J Natl Cancer Inst. 1993;85:1398-1406.
Dewar JA, Harobin JM, Preece PE, Tavendale R, Tunstall-Pedoe H, Wood RAB. Long term effects of tamoxifen on blood values in breast cancer. Br Med J. 1992;305:225-226.
Love RR, Newcomb PA, Wiebe DA, Surawicz TS, Jordan VC, Carbone PP, Demets DL. Effect of tamoxifen therapy on lipid and lipoprotein levels in post-menopausal patients with node negative breast cancer. J Natl Cancer Inst. 1990;82:1327-1332.
Paszty C, Maeda N, Verstuyft J, Rubin EM. Apolipoprotein A1 transgene corrects apolipoprotein E deficiency–induced atherosclerosis in mice. J Clin Invest. 1994;94:899-903.
Grainger DJ, Weissberg P, Metcalfe JC. Tamoxifen decreases the rate of proliferation of rat vascular smooth muscle cells in culture by inducing production of transforming growth factor-β. Biochem J. 1993;294:109-112.
Butta A, MacLennan K, Flanders KC, Sacks NPM, Smith I, Mackinna A, Dowsett M, Wakefield LM, Sporn MB, Baum M, Colletta AA. Induction of transforming growth factor-β1 in human breast cancer in vivo following tamoxifen treatment. Cancer Res. 1992;52:4261-4264.
Lawn RM, Pearl AD, Konz LL, Rubin EM, Reckless J, Metcalfe JC, Grainger DJ. Feedback mechanism of focal vascular lesion formation in transgenic apolipoprotein(a) mice. J Biol Chem. 1996;271:31367-31371.
Zhang SH, Reddick RL, Piedrahita JA, Maeda N. Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science. 1992;258:468-471.
Reddick RL, Zhang SH, Maeda N. Atherosclerosis in mice lacking apo E. Arterioscler Thromb. 1994;14:141-147.
Nakashima Y, Plump AS, Raines EW, Breslow JL, Ross R. Apo E deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb. 1994;14:133-140.
Piedrahita JA, Zhang SH, Hagaman JR, Oliver P, Maeda N. Generation of mice carrying a mutant apolipoprotein E gene inactivated by gene targeting in embryonic stem cells. Proc Natl Acad Sci U S A. 1992;89:4471-4475.
Allain CA, Poon LS, Chan CSG, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chim Acta. 1974;20:470-479.
Yokode M, Hammer RE, Ishibashi S, Brown MS, Goldstein JL. Diet induced hypercholesterolemia in mice: prevention by overexpression of LDL receptors. Science. 1990;250:1273-1275.
Bucolo G, David H. Quantitative determination of serum triglycerides by the use of enzymes. Clin Chim. 1973;19:476-484.
Bjorkerud S. Effects of TGF-β1 on human arterial smooth muscle cells in vitro. Arterioscler Thromb. 1991;11:892-902.
Owens GK, Loeb A, Gordon D, Thompson MM. Expression of smooth muscle specific isoactin in cultured vascular smooth muscle cells: relationship between growth and cytodifferentiation. J Cell Biol. 1986;102:343-352.
Giachelli CM, Bae N, Almeida M, Denhard DT, Alpers CE, Schwartz SM. Osteopontin is elevated during neointima formation in rat arteries and is a novel component of human atherosclerotic plaques. J Clin Invest. 1993;92:1686-1696.
Shanahan CM, Cary NRB, Metcalfe JC, Weissberg PL. High expression of genes for calcification-regulating proteins in human atherosclerotic plaques. J Clin Invest. 1994;92:2393-2402.
Caleffi M, Fentiman IS, Clark GM, Wang DY, Needham J, Clark K, Laville A, Lewis B. Effect of tamoxifen on estrogen binding, lipid and lipoprotein concentrations and blood clotting parameters in premenopausal women with breast pain. J Endocrinol. 1988;119:335-339.
Rutqvist LE, Johansson H, Signomklao T, Johansson U, Fornander T, Wilking N. Adjuvant tamoxifen therapy for early stage breast cancer and second primary malignancies. J Natl Cancer Inst. 1995;87:645-651.