Tamoxifen effects on endothelial function and cardiovascular risk factors in men with advanced atherosclerosis.

BACKGROUND: Tamoxifen and its analogues act as selective estrogen receptor modulators (SERMs) in women, with estrogen-like activities on some plasma cardiovascular risk factors (eg, lipoproteins). Effects of SERMs on men with coronary artery disease (CAD) have not been reported. METHODS AND RESULTS: Thirty-one men with angiographically proven CAD were recruited; 16 were treated with tamoxifen (40 mg/d) for 56 days, and 15 were untreated. All the CAD patients were medicated with aspirin and an HMG-CoA reductase inhibitor for >/=6 weeks before entering the study. Ten men with angina-like symptoms but normal coronary arteries by angiography (NCA group) were also treated with tamoxifen. Blood samples were collected at days -7, 0, 7, 14, 21, 28, and 56 of treatment. Endothelium-dependent flow-mediated dilatation (ED-FMD) of the brachial artery was measured by high-resolution ultrasound at 5 visits. Tamoxifen caused an increase in %ED-FMD maximal at 28 days in the CAD group (2.1+/-0.3% to 7.5+/-0.7%; P<0.0001) and the NCA group (3.8+/-0.4% to 7.9+/-1.0%; P<0.0001), with no significant change in the untreated group. Tamoxifen also caused decreases in several plasma cardiovascular risk factors, including total cholesterol, triglycerides, lipoprotein(a), and fibrinogen. Except for the triglyceride response, these effects were similar to those reported for postmenopausal women treated with tamoxifen. CONCLUSIONS: Tamoxifen substantially increased ED-FMD in men with CAD who were taking conventional medication. Together with the effects on risk factors, the data strongly support clinical evaluation of SERMs for the treatment of men with CAD.

Estimation of systolic and diastolic free intracellular Ca2+ by titration of Ca2+ buffering in the ferret heart.

Spectroscopic Ca(2+)-indicators are thought to report values of free intracellular Ca(2+) concentration ([Ca(2+)](i)) that may differ from unperturbed values because they add to the buffering capacity of the tissue. To check this for the heart we have synthesized a new (19)F-labelled NMR Ca(2+) indicator, 1, 2-bis-[2-bis(carboxymethyl)amino-4,5-difluorophenoxy]ethane ('4, 5FBAPTA'), with a low affinity (K(d) 2950 nM). The new indicator and four previously described (19)F-NMR Ca(2+) indicators 1,2-bis-[2-(1 - carboxyethyl)(carboxymethyl)amino - 5 - fluorophenoxy]ethane ('DiMe-5FBAPTA'), 1, 2-bis-[2-(1-carboxyethyl)(carboxymethyl)amino-4-fluorophenoxy]ethane ('DiMe-4FBAPTA'), 1, 2-bis-[2-bis(carboxymethyl)amino-5-fluorophenoxy]ethane ('5FBAPTA') and 1, 2-bis-[2-bis(carboxymethyl)amino-5-fluoro-4-methylphenoxy]ethane ('MFBAPTA'), with dissociation constants for Ca(2+) ranging from 46 to 537 nM, have been used to measure [Ca(2+)](i), over the range from less than 100 nM to more than 3 microM, in Langendorff-perfused ferret hearts (30 degrees C, pH 7.4, paced at 1.0 Hz) by (19)F-NMR spectroscopy. Loading hearts with indicators resulted in buffering of the Ca(2+) transient. The measured end-diastolic and peak-systolic [Ca(2+)](i) were both positively correlated with indicator K(d). The positive correlations between indicator K(d) and the measured end-diastolic and peak-systolic [Ca(2+)](i) were used to estimate the unperturbed end-diastolic and peak-systolic [Ca(2+)](i) by extrapolation to K(d)=0 (diastolic) and to K(d)=infinity (systolic) respectively. The extrapolated values in the intact beating heart were 161 nM for end-diastolic [Ca(2+)](i) and 2650 nM for peak-systolic [Ca(2+)](i), which agree well with values determined from single cells and muscle strips.

Ca2+ buffering in the heart: Ca2+ binding to and activation of cardiac myofibrils.

The measurement of cardiac Ca(2+) transients using spectroscopic Ca(2+) indicators is significantly affected by the buffering properties of the indicators. The aim of the present study was to construct a model of cardiac Ca(2+) buffering that satisfied the kinetic constraints imposed by the maximum attainable rates of cardiac contraction and relaxation on the Ca(2+) dissociation rate constants and which would account for the observed effects of (19)F-NMR indicators on the cardiac Ca(2+) transient in the Langendorff-perfused ferret heart. It is generally assumed that the Ca(2+) dependency of myofibril activation in cardiac myocytes is mediated by a single Ca(2+)-binding site on troponin C. A model based on 1:1 Ca(2+) binding to the myofilaments, however, was unable to reproduce our experimental data, but a model in which we assumed ATP-dependent co-operative Ca(2+) binding to the myofilaments was able to reproduce these data. This model was used to calculate the concentration and dissociation constant of the ATP-independent myofilament Ca(2+) binding, giving 58 and 2.0 microM respectively. In addition to reproducing our experimental data on the concentration of free Ca(2+) ions in the cytoplasm ([Ca(2+)](i)), the resulting Ca(2+) and ATP affinities given by fitting of the model also provided good predictions of the Ca(2+) dependence of the myofibrillar ATPase activity measured under in vitro conditions. Solutions to the model also indicate that the Ca(2+) mobilized during each beat remains unchanged in the presence of the additional buffering load from Ca(2+) indicators. The new model was used to estimate the extent of perturbation of the Ca(2+) transient caused by different concentrations of indicators. As little as 10 microM of a Ca(2+) indicator with a dissociation constant of 200 nM will cause a 20% reduction in peak-systolic [Ca(2+)](i) and 30 microM will cause approx. 50% reduction in the peak-systolic [Ca(2+)](i) in a heart paced at 1.0 Hz.

Cultures of proliferating vascular smooth muscle cells from adult human aorta.

Abnormal proliferation of human vascular smooth muscle cells (hVSMCs) is a central event in the development of atherosclerosis (1-3) As a result, there is considerable interest in the establishment of hVSMC cultures as a mode1 of this disease process However, it has been noted in the past (4,5) that hVSMCs, especially when cultured by the enzyme-dispersal technique (hVSMC,(ED)), grow poorly in culture compared to VSMCs from other species (e.g., rat) This has limited their use for cell-culture studies. We have recently reported that the reduced proliferative capacity of hVSMC(ED) from adult aorta can be attributed to the endogenous production of active TGF-ß(6,7). We (6-8) and others (9-11) have shown that TGF-ß; is a potent inhibtior of smooth muscle cell proliferation. Moreover, recent studies in animal models of atherosclerosis (12) have suggested that TGF-P plays a pivotal role in regulation of vessel wall architecture (13). We have also shown that hVSMCs derived by the alternative method of explanting (hVSMC(EX)) have a greater proliferative capacity than the hVSMC(ED) (14) In accordance with our hypotheses, the cells grown from explanted tissue did not produce TGF-ß; (14).

Proliferation of human aortic vascular smooth muscle cells in culture is modulated by active TGF beta.

OBJECTIVE: Human vascular smooth muscle cells (VSMC) prepared by enzyme dispersal have been reported to proliferate more slowly and remain more differentiated than VSMC derived by explant culture. The aim of this study was to investigate whether these differences are attributable to the production of transforming growth factor beta (TGF beta), a major modulator of VSMC properties in culture, by enzyme dispersal, but not explant, cultures. METHODS: Human aortic VSMC cultures were prepared by enzymatic dissociation of aortic media or by outgrowth of cells from explants derived from the vessel media. Properties, including rate of proliferation and differentiation, of the two different types of culture obtained from 10 donors (age 3-54 years, of either sex) were compared. The role of TGF beta as mediator of these differences was investigated. RESULTS: VSMC from enzyme dispersal cultures proliferated with a long doubling time in 20% fetal calf serum of 68(SEM 2) h, reaching a low saturation density with no "hills and valleys", and the cells were predominantly of stellate morphology. VSMC from explant cultures proliferated with a shorter doubling time of 35(2) h, reached a higher saturation density with "hills and valleys", and showed a spindle shaped morphology. The enzyme dispersed cells, but not the explant cells, released TGF beta into the medium. Addition of exogenous TGF beta to the explant cells extended the doubling time, while addition of a neutralising antibody to the enzyme dispersed cells decreased the doubling time. Tamoxifen, an anti-oestrogenic drug, decreased the rate of proliferation of the explant cells (ED50 = 50 nM; n = 3) but not enzyme dispersed cells by stimulating production of TGF beta. The explant and enzyme dispersal cultures also differed in response to growth factors. The explant cultures stimulated DNA synthesis in response to platelet derived growth factor (PDGF) AA and BB. Enzyme dispersal cultures stimulated DNA synthesis in response to PDGF-BB but response to AA was weaker and also variable. CONCLUSIONS: Human explant derived VSMC systematically differ from enzyme dispersed VSMC obtained from the same donor tissue. Wherever the potential role of TGF beta was investigated, the differences between enzyme dispersal and explant cultures were due to autocrine production of TGF beta by the enzyme dispersal, but not explant cultures.

Adult human aortic smooth muscle cells in culture produce active TGF-beta.

Vascular smooth muscle cells (VSMC) from adult human aortas proliferated in culture in response to fetal calf serum (FCS) with a population doubling time of 70-85 h compared with 35 +/- 5 h for VSMC derived from adult rat aortas. Medium conditioned on cultures prepared from aortas from three different donors and mixed 1:1 with fresh Dulbecco's modified Eagle's medium plus 20% FCS [human conditioned medium (HCM)] reduced the rate of proliferation of rat VSMC by 46 +/- 6% (n = 3) after 48 h compared with cells in fresh medium. HCM did not reduce the proportion (> 65%) of rat VSMC that entered DNA synthesis but delayed entry into mitosis by at least 18 h. This effect was similar to previous observations of the action of transforming growth factor-beta (TGF-beta) on rat VSMC (G. K. Owens, A. A. Geisterfer, Y. W. Yang, and A. Komoriya. J. Cell Biol. 107: 771-780, 1988). A TGF-beta assay using DNA synthesis in mink lung epithelial cells confirmed that human, but not rat, VSMC in culture secrete active TGF-beta. Addition of a neutralizing antibody to TGF-beta to human VSMC in the presence of 20% FCS decreased the population doubling time from 74 +/- 3 to 46 +/- 6 h (n = 3). These observations demonstrate that the long population doubling time of human VSMC is due to the production of active TGF-beta and to an inhibitory autocrine loop.

Calcium measurements with a new high-affinity n.m.r. indicator in the isolated perfused heart.

A new n.m.r. indicator, 1,2-bis-(2-[1-(hydroxycarbony)ethyl- (hydroxycarbonylmethyl)]amino-5-fluorophenoxy)ethane (DiMe-5FBAPTA), with a higher affinity for calcium (apparent Kd 46 nM, pH 7.2, 30 degrees C) than the parent 5FBAPTA chelator (Kd 537 nM, pH 7.1, 30 degrees C) has been used to measure the cardiac intracellular free Ca2+ ([Ca2+]i). DiMe-5FBAPTA was loaded into Langendorff-perfused ferret hearts maintained at 30 degrees C using the acetoxymethyl ester (AM) derivative. The intracellular concentration required to achieve an adequate signal-to-noise (S/N) ratio (> 10:1) for the n.m.r. spectra caused a similar reduction in developed pressure to that obtained using 5FBAPTA-AM. The DiMe-5FBAPTA was used to estimate [Ca2+]i in diastole, through the calcium transient and at rest in the presence of the slow calcium channel blocker diltiazem. At a pacing frequency of 1.0 Hz, end-diastolic [Ca2+]i was 198 +/- 30 nM (n = 9), and reducing the pacing frequency to 0.2 Hz lowered [Ca2+]i to 89 +/- 13 nM (n = 5). Perfusion with diltiazem (100 microM) for 60 min lowered [Ca2+]i to 10 +/- 1 nM (n = 4) in unpaced hearts and to 94 +/- 24 nM (n = 4) in hearts paced at 1.0 Hz. The [Ca2+]i transient measured with DiMe-5FBAPTA was sharper and delayed compared with the transient measured previously with 5FBAPTA. Co-loading the two indicators provided evidence that the indicator with the higher Kd had a dominant effect on the end-diastolic [Ca2+]i. The lower values for end-diastolic [Ca2+]i obtained with DiMe-5FBAPTA are consistent with fluorescent indicator measurements. These observations suggest that perturbations of [Ca2+]i caused by the new indicator are less than those induced by 5FBAPTA. DiMe-5FBAPTA therefore represents a useful step in the development of 19F-n.m.r. calcium indicators.

Regulation of intracellular pH in the perfused heart by external HCO3- and Na(+)-H+ exchange.

Both Na(+)-dependent HCO3- influx and the Na(+)-H+ antiport have been shown to contribute to recovery from intracellular acidosis in avian and mammalian cardiac tissue. We have investigated the participation of these mechanisms in the recovery of intracellular pH (pHi) after an acid load in the Langendorff-perfused ferret heart. pHi was measured from the phosphorus-31 nuclear magnetic resonance chemical shift of 2-deoxy-D-glucose 6-phosphate. Basal pHi was higher in HCO(3-)-buffered solution (7.05 +/- 0.01; n = 8) than in nominally HCO(3-)-free N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) solution (6.98 +/- 0.02; n = 9). Addition of 5-(N-ethyl-N-isopropyl)amiloride (EIPA) caused a significant fall in pHi in HEPES solution (6.91 +/- 0.02; n = 5) but not in HCO3- solution (7.02 +/- 0.02; n = 5). Intrinsic intracellular buffering capacity in 0 Na(+)-HEPES solution was 37 +/- 2 mmol/l (n = 4), and additional buffering due to HCO(3-)-CO2 was approximately 13 mmol/l in HCO3- solution. After an intracellular acidosis induced by an NH4Cl prepulse, the proton efflux rate (JH) at pHi 6.90 was 0.5 +/- 0.2 nmol.l-1.min-1 (n = 14) in HEPES solution and 1.2 +/- 0.4 mmol.l-1.min-1 (n = 13) in HCO3- solution. The addition of 1 microM EIPA effectively blocked proton efflux in HEPES solution (JH < 0.1 mmol.l-1.min-1; n = 8), whereas it slowed pHi recovery in HCO3- solution (JH = 0.6 +/- 0.2 mmol.l-1.min-1; n = 9). There was no recovery of pHi in Na(+)-free HCO3- solution (JH < 0.1 mmol.l-1.min-1; n = 3). The Na(+)-H+ antiport and a mechanism requiring both external Na+ and HCO3- each contribute approximately 50% to proton efflux at pHi 6.90 during the recovery from intracellular acidosis in the isolated perfused mammalian heart.

Proliferation of human smooth muscle cells promoted by lipoprotein(a).

Elevated blood concentrations of lipoprotein(a) [Lp(a)] and its constituent, apolipoprotein(a) [apo(a)], constitute a major risk factor for atherosclerosis, but their physiological activities remain obscure. Lp(a) and purified apo(a) stimulated the growth of human smooth muscle cells in culture. This effect resulted from inhibition of plasminogen activation, and consequently the activation by plasmin of latent transforming growth factor-beta, which is an inhibitor of smooth muscle cell growth. Because smooth muscle proliferation is one of the hallmarks of atherosclerotic lesions, these results point to a plausible mechanism for the atherogenic activity of Lp(a).

Resting and end-diastolic [Ca2+]i measurements in the Langendorff-perfused ferret heart loaded with a 19F NMR indicator.

Intracellular free calcium concentration ([Ca2+]i) was measured in Langendorff-perfused ferret hearts (30 degrees C, pH 7.4) by loading paced hearts with the 19F NMR calcium indicator, the 5,5'-difluoro derivative of 1,2-bis(o-aminophenoxy)ethane-N, N,N',N'-tetraacetic acid (5FBAPTA), to an initial cytosolic concentration of approximately 120 microM. Increasing the pacing frequency raised the end-diastolic [Ca2+]i from 299 +/- 44 nM (mean +/- SEM) at 0.2 Hz to 522 +/- 54 nM at 1.0 Hz and 691 +/- 166 nM at 2.0 Hz. Raising [Ca]o from 1.8 to 7.0 mM at a pacing frequency of 1.0 Hz increased end-diastolic [Ca2+]i to 625 +/- 39 nM. In unpaced hearts perfused with diltiazem (100 microM), [Ca2+]i fell rapidly to a steady-state value of < 100 nM after 60 min. Raising [Ca]o from 1.8 to 7.0 mM had no detectable effect on resting [Ca2+]i. The time course of the [Ca2+]i transient was measured in hearts paced at 1.1 Hz and perfused with 1.8 mM [Ca]o. The peak [Ca2+]i was approximately 2 microM at approximately 150 msec after the pacing pulse, and peak developed LVP occurred at 550 msec compared with 280 msec in control hearts not loaded with 5FBAPTA. Comparisons with data obtained by other techniques, including fluorescent [Ca2+]i indicators, imply that although the end-diastolic [Ca2+]i values obtained with 5FBAPTA in beating hearts are elevated by the concentrations of intracellular 5FBAPTA required for signal detection, the changes in [Ca2+]i observed in response to experimental interventions are qualitatively consistent with previous data.

Ca2+ transient, Mg2+, and pH measurements in the cardiac cycle by 19F NMR.

19F NMR indicators have been used to measure the free cytosolic cation concentrations ([Mn+]i, where M is the atomic symbol and n is the value of the charge) of Ca2+, H+, and Mg2+ in perfused ferret hearts. The [Ca2+]i transient, cytosolic pH (pHi), and [Mg2+]i have also been followed at 16 phases in the cardiac cycle in hearts paced at 1.25 Hz at 30 degrees C. The initial [Ca2+]i rose rapidly after a 50-ms delay, was maximal at greater than 1.5 microM after 150 ms, and declined thereafter to the initial concentration. In contrast, no significant changes in pHi (pH 7.03 +/- 0.08) or [Mg2+]i (1.2 +/- 0.1 mM) were detected in the cycle. A decrease in developed pressure when the [Ca2+]i indicator (but not the pHi or [Mg2+]i indicator) was loaded into hearts was substantially reversed by the addition of 50 microM ZnCl2 to the perfusion medium. The Zn2+ was taken up into the myoplasm and displaced Ca2+ bound to the indicator, a symmetrically substituted difluoro derivative of 1,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (5FBAPTA), as evidenced by the appearance of the Zn-5FBAPTA resonance. The decrease in developed pressure caused by 5FBAPTA, therefore, may be due to its Ca2+ buffering effect on the myoplasm. By coloading hearts with the [Ca2+]i and pHi indicators, simultaneous measurement of several [Mn+]i was demonstrated, which should provide a useful addition to the methods available to monitor cardiac function and pharmacology.

Relative phosphate contents of phosphofructokinase and other soluble phosphoproteins in skeletal muscle.

In vivo phosphorylation of muscle proteins has been studied by incorporation of [32P]phosphate with emphasis placed upon the phosphorylation of glycolytic enzymes. Of the approximately 25 soluble proteins resolved by two-dimensional electrophoresis that contain significant 32P, phosphofructokinase was the sole glycolytic enzyme identified as a phosphoprotein. The extent of phosphorylation found for this enzyme was the same as determined previously for purified phosphofructokinase and was about the same as the extent of phosphorylation of phosphorylase in resting muscle. Subsequent partial purification of several glycolytic enzymes confirmed the absence of significant amount of phosphate. However, phosphoglycerate mutase contained small amounts of covalently bound 32P that was exchangeable with 3-phosphoglycerate and therefore, most likely was incorporated during the catalytic reaction cycle. Analogous results were obtained for phosphoglucomutase. Both mutases were also phosphorylated at the same sites by the catalytic subunit of cyclic AMP-dependent protein kinase.