Electrophysiological effects of ginseng and ginsenoside Re in guinea pig ventricular myocytes

Panax ginseng is a folk medicine with various cardiovascular actions; however, its underlying mechanisms of action are not well known. In the present study, we examined the effects of ginseng and its main component, ginsenoside Re, on action potentials and membrane currents recorded from isolated guinea pig ventricular myocytes with the whole-cell patch clamp technique. Ginseng (1 mg/ml) shortened the action potential duration in a rate-dependent manner. Ginseng depressed the L-type Ca2+ current (I(Ca-L)) in a mode of both tonic block and use-dependent block, and enhanced the slowly activating component of the delayed rectifier K+ current (I(Ks)). Ginsenoside Re 3 microM exhibited similar electrophysiological effects to those of 1 mg/ml ginseng, but of slightly smaller magnitude. Inhibition of I(Ca,L) and enhancement of I(Ks) by ginsenoside Re appear to be one of the main electrophysiological actions of ginseng in the heart, although contributions from other ingredients should be considered.     

Electrophysiological effects of bethanidine sulfate on canine cardiac Purkinje fibers and ventricular muscle cells

Bethanidine sulfate is a congener of bretylium tosylate, which has been reported to have antiarrhythmic and antifibrillatory effects. We studied the effects of bethanidine on transmembrane potentials recorded from canine Purkinje fibers and ventricular muscle cells, using standard microelectrode techniques. Normal Purkinje fibers were exposed to bethanidine (10-80 mg/L) for 30-40 min. Bethanidine produced dose-dependent decreases in the maximal rate of depolarization (MRD) and over-shoot of phase 0, but did not affect maximum diastolic potential (MDP). Action potential plateau duration (APD, to -60 mV) was decreased by bethanidine at all cycle lengths of stimulation between 1,000 and 300 ms. Bethanidine depressed the membrane responsiveness curve, and its effects on MRD showed marked use dependence. In ventricular muscle cells, bethanidine 20 mg/L decreased MRD but did not affect MDP or APD. The rate of normal automaticity in Purkinje fibers with MDPs greater than -85 mV was increased to 21.5 +/- 5.6 beats/min after exposure to bethanidine (10 mg/L for 30 min) from control values of 10.2 +/- 5.3 beats/min. Abnormal automaticity (MDPs = -40 to -60 mV) was induced in Purkinje fibers by superfusion with Tyrode solution containing 0.25 mM BaCl2; this activity also was accelerated after exposure to bethanidine 10 mg/L. The effects of bethanidine on automaticity are similar to those of bretylium, and may be caused by release of endogenous catecholamines. In contrast, the effects of bethanidine on action potentials of normal (driven) Purkinje fibers are markedly different from those of bretylium, and resemble those of lidocaine after 30-60 min of exposure.     

Electrophysiological effects of labetalol on canine atrial, cardiac Purkinje fibres and ventricular muscle.

1. Using conventional microelectrode techniques for the intracellular recordings of the membrane potential, the effects of labetalol were studied on cardiac Purkinje, atrial and ventricular muscle fibres of the dog. 2. Labetalol (1-10 microM) reduced, in a concentration-dependent manner, the action potential amplitude (APA) and the maximum rate of rise of the action potential (Vmax) in Purkinje fibres. 3. The action potential duration (APD) was decreased in Purkinje fibres but significantly increased in ventricular fibres after small concentrations of labetalol (1-3 microM). The atrial fibres were not very sensitive to labetalol. 4. Depolarization of the cardiac Purkinje fibres by increasing the external potassium concentration (8-12 mM), potentiated the labetalol effects on APA and Vmax but blocked its effects on the APD. 5. The effects of labetalol on Vmax of Purkinje fibres were dependent on the frequency of stimulation. 6. The ratio of the effective refractory period to the APD was increased both in normally polarized and depolarized Purkinje fibres after treatment with labetalol (10 microM). 7. Labetalol (10 microM) shifted the membrane responsiveness curve of Purkinje fibres by about 10 mV in the hyperpolarizing direction. 8. The slow response obtained in K-depolarized, Ba-treated Purkinje fibres was not significantly affected by labetalol (10-100 microM). 9. It is suggested that labetalol can exert Class I and Class III antiarrhythmic actions in cardiac muscle of the dog in vitro.     

Validation of an in vitro contractility assay using canine ventricular myocytes

Measurement of cardiac contractility is a logical part of pre-clinical safety assessment in a drug discovery project, particularly if a risk has been identified or is suspected based on the primary- or non-target pharmacology. However, there are limited validated assays available that can be used to screen several compounds in order to identify and eliminate inotropic liability from a chemical series. We have therefore sought to develop an in vitro model with sufficient throughput for this purpose.Dog ventricular myocytes were isolated using a collagenase perfusion technique and placed in a perfused recording chamber on the stage of a microscope at ~ 36 °C. Myocytes were stimulated to contract at a pacing frequency of 1 Hz and a digital, cell geometry measurement system (IonOptix™) was used to measure sarcomere shortening in single myocytes. After perfusion with vehicle (0.1% DMSO), concentration-effect curves were constructed for each compound in 4-30 myocytes taken from 1 or 2 dog hearts.The validation test-set was 22 negative and 8 positive inotropes, and 21 inactive compounds, as defined by their effect in dog, cynolomolgous monkey or humans. By comparing the outcome of the assay to the known in vivo contractility effects, the assay sensitivity was 81%, specificity was 75%, and accuracy was 78%.With a throughput of 6-8 compounds/week from 1 cell isolation, this assay may be of value to drug discovery projects to screen for direct contractility effects and, if a hazard is identified, help identify inactive compounds.     

Electrophysiological effects of flecainide enantiomers in canine Purkinje fibres

Summary Since flecainide is a chiral class I antiarrhythmic agent, we examined the basic electrophysiological effects of its enantiomers (10^–6–3 × 10^–5 mol/l) in isolated canine Purkinje fibres using standard microelectrode techniques. Frequency-dependent block was studied by pacing at cycle lengths of 300 and 2000 ms. Both enantiomers produced a marked, concentration-dependent decrease in maximum upstroke velocity (V _max), with greater depression occurring at a cycle length of 300 ms. The concentrations causing a 50% decrease in V _max (EC₅₀) at this pacing rate were 5.0 ± 0.6 × 10^-6 mol/l for (+)-flecainide and 6.2 ± 0.8 × 10^-6 mol/l for (-)-flecainide, and were not significantly different. Both enantiomers also produced a concentration and frequency-dependent decrease in action potential amplitude and an increase in conduction time with no significant differences between the enantiomers. However, at the highest concentration studied (3 × 10^-5 mol/l), none of the nine tissues exposed to (+)-flecainide could be paced at 300 ms cycle length while five of eight tissues exposed to (-)-flecainide could be paced, suggesting the possibility of a small difference between the effects of the enantiomers on membrane responsiveness. The effect of the enantiomers on action potential duration (APD) also depended on pacing rate. At the longer cycle length, ADP₅₀ was shortened to a maximum of 61 ± 7% and 67 ± 7% from baseline by (+)- and (-)-flecainide, respectively, whereas APD₉₅ was shortened by 30 ± 6% and 32 ± 3% from baseline by (+)-and (-)-flecainide, respectively. At the 300 ms cycle length, there was a modest decrease in both ADP₅₀ and APD₉₅ at low concentrations (up to 3 × 10^-6 mol/l) followed by a return toward baseline at the higher concentrations 10^-5 and 3 × 10^-5 mol/l. No significant differences between the enantiomers were noted at either pacing frequency. The results suggest there is no significant difference between the effects of flecainide enantiomers on basic electrophysiological parameters of canine Purkinje fibres. Although the (+)-enantiomer may produce a slightly greater decrease in cellular responsiveness, the results generally agree with prior in vivo data showing little or no difference in the antiarrhythmic activities of the enantiomers. Send offprint requests to J. K. Smallwood at the above address     

Electrophysiological effects of a novel antiarrhythmic drug, EO-122, on guinea pig ventricular muscle and isolated myocytes

EO-122, a newly developed structural analog of lidocaine, has recently been shown to suppress ventricular arrhythmias in a few clinical studies in patients and in experimental animals. In the present study, we investigated the effects of EO-122 on the electrophysiological properties of guinea pig papillary muscle and ventricular myocytes by means of standard microelectrode and whole-cell recording techniques, respectively, At the concentration range of 10(-7)-10(-4) M (cycle length, 2000 ms), resting potential and action potential duration (APD90) were not altered by the drug. Action potential amplitude and APD50 were reduced (p less than 0.01) by 10(-4) M, and Vmax was reduced (p less than 0.01) by EO-122 greater than or equal to 10(-5) M. The effect of EO-122 on Vmax was use-dependent. At 10(-6) and 10(-5) M (cycle length, 2000 ms), the time constant for onset of block (tau on) was 37.0 +/- 13.2 and 26.0 +/- 3.4 s, respectively. The recovery kinetics from use-dependent block was not monoexponential, and the estimated "time constant" for recovery was 76.5 s. We examined the effects of EO-122, 10(-5) M, on the membrane currents in ventricular myocytes and found that the drug attenuated the slow inward current (Isi). EO-122 reduced peak Isi by 68.6 +/- 5.2% (p less than 0.005), whereas the outward current was unchanged. The present study demonstrates that EO-122 blocks both the fast inward (Na+) and the slow inward (Ca2+) channels, and these effects are probably responsible for the antiarrhythmic effects of the drug.     

Asynchronous activation of calcium and potassium currents by isoproterenol in canine ventricular myocytes

Adrenergic activation of L-type Ca²⁺ and various K⁺ currents is a crucial mechanism of cardiac adaptation; however, it may carry a substantial proarrhythmic risk as well. The aim of the present work was to study the timing of activation of Ca²⁺ and K⁺ currents in isolated canine ventricular cells in response to exposure to isoproterenol (ISO). Whole cell configuration of the patch-clamp technique in either conventional voltage clamp or action potential voltage clamp modes were used to monitor I _Ca, I _Ks, and I _Kr, while action potentials were recorded using sharp microelectrodes. ISO (10 nM) elevated the plateau potential and shortened action potential duration (APD) in subepicardial and mid-myocardial cells, which effects were associated with multifold enhancement of I _Ca and I _Ks and moderate stimulation of I _Kr. The ISO-induced plateau shift and I _Ca increase developed faster than the shortening of APD and stimulation of I _Ks and I _Kr. Blockade of β₁-adrenoceptors (using 300 nM CGP-20712A) converted the ISO-induced shortening of APD to lengthening, decreased its latency, and reduced the plateau shift. In contrast, blockade of β₂-adrenoceptors (by 50 nM ICI 118,551) augmented the APD-shortening effect and increased the latency of plateau shift without altering its magnitude. All effects of ISO were prevented by simultaneous blockade of both receptor types. Inhibition of phosphodiesterases decreased the differences observed in the turn on of the ISO-induced plateau shift and APD shortening. ISO-induced activation of I _Ca is turned on faster than the stimulation of I _Ks and I _Kr in canine ventricular cells due to the involvement of different adrenergic pathways and compartmentalization.     

Biphasic effect of bimoclomol on calcium handling in mammalian ventricular myocardium

1. Concentration-dependent effects of bimoclomol, the novel heat shock protein coinducer, on intracellular calcium transients and contractility were studied in Langendorff-perfused guinea-pig hearts loaded with the fluorescent calcium indicator dye Fura-2. Bimoclomol had a biphasic effect on contractility: both peak left ventricular pressure and the rate of force development significantly increased at a concentration of 10 nM or higher. The maximal effect was observed between 0.1 and 1 microM, and the positive inotropic action disappeared by further increasing the concentration of bimoclomol. The drug increased systolic calcium concentration with a similar concentration-dependence. In contrast, diastolic calcium concentration increased monotonically in the presence of bimoclomol. Thus low concentrations of the drug (10 - 100 nM) increased, whereas high concentrations (10 microM) decreased the amplitude of intracellular calcium transients. 2. Effects of bimoclomol on action potential configuration was studied in isolated canine ventricular myocytes. Action potential duration was increased at low (10 nM), unaffected at intermediate (0.1 - 1 microM) and decreased at high (10 - 100 microM) concentrations of the drug. 3. In single canine sarcoplasmic calcium release channels (ryanodine receptor), incorporated into artificial lipid bilayer, bimoclomol significantly increased the open probability of the channel in the concentration range of 1 - 10 microM. The increased open probability was associated with increased mean open time. The effect of bimoclomol was again biphasic: the open probability decreased below the control level in the presence of 1 mM bimoclomol. 4. Bimoclomol (10 microM - 1 mM) had no significant effect on the rate of calcium uptake into sarcoplasmic reticulum vesicles of the dog, indicating that in vivo calcium reuptake might not substantially be affected by the drug. 5. In conclusion, the positive inotropic action of bimoclomol is likely due to the activation of the sarcoplasmic reticulum calcium release channel in mammalian ventricular myocardium.     

Effect of yohimbine on action potentials recorded from isolated canine ventricular myocytes

Yohimbine is used extensively as an alpha 2-adrenoceptor antagonist. This drug also inhibits sodium channels in the squid axon. In this study we investigated the electrophysiological effects of yohimbine in isolated canine ventricular myocytes. Yohimbine produced a reversible and concentration-dependent decrease in the maximum upstroke velocity and a slight increase in the action potential duration. The maximum upstroke velocity was reduced by over 90% by 10(-4) M yohimbine. It did not effect the resting or plateau potentials in control or isoproterenol-treated cells. Delayed after-depolarizations were also inhibited. These data suggest that yohimbine can produce a local anesthetic effect which is primarily due to inhibition of sodium channels.     

Electrophysiological effects of CRE-1087 in guinea-pig ventricular muscles.

1. The electrophysiological effects of CRE-1087, a new antiarrhythmic drug, were studied in guinea-pig papillary muscles. 2. CRE-1087 (10(-7) M-10(-4) M) produced a concentration-dependent decrease in the action potential amplitude and Vmax of the upstroke without altering the action potential duration or the resting membrane potential. 3. At 5 x 10(-6) M, CRE-1087 produced a 5.0 +/- 1.0% tonic Vmax block and this value was not modified at the different rates of stimulation tested. In the presence of CRE-1087, trains of stimuli at rates between 0.5 and 3 Hz led to an exponential decline in Vmax (frequency-dependent Vmax block) which augmented at higher rates of stimulation. At 2 Hz the onset kinetics of the frequency-dependent Vmax block was fitted by a monoexponential function being the K value 0.099 +/- 0.012 AP-1. The time constant for the recovery of Vmax from the frequency-dependent block was prolonged to 18.3 +/- 1.5 s. 4. CRE-1087 (5 x 10(-6) M) did not shift the membrane responsiveness curve. 5. CRE-1087 had no effect on the characteristics of the slow action potentials elicited by isoprenaline in ventricular fibres depolarized by 27 mM KCl, which suggested that CRE-1087 did not exhibit class IV (Ca antagonist) antiarrhythmic actions. 6. The slow onset and the slow offset kinetics of frequency-dependent Vmax block during repetitive activity suggested that in guinea-pig ventricular muscle fibres CRE-1087 exhibited class Ic antiarrhythmic actions.     

Electrophysiological effects of cyclic GMP on canine cardiac Purkinje fibers

The effect of cyclic 3'5'-guanosine monophosphate (8-bromo-cGMP) on action potential characteristics was investigated. Standard microelectrode techniques were used to study the effects of 8-bromo-cGMP on canine cardiac Purkinje fibers in vitro. Canine Purkinje fiber tissue preparations were exposed to increasing concentrations of 8-bromo-cGMP (10(-6), 10(-5), 10(-4) M). The action potential duration at 50% (APD50) and 90% (APD90) repolarization, resting membrane potential (RMP), action potential amplitude (APA), rate of rise of phase 0 (Vmax), spontaneous rate (SR), escape time (ET), and effective refractory period (ERP) did not change at these concentrations of 8-bromo-cGMP. The effect of 8-bromo-cGMP on isoproterenol (10(-7) M) treated Purkinje fibers was tested. Predictably, isoproterenol shortened APD and ERP and increased SR. APD or ERP shortening was not affected by 8-bromo-cGMP, but the increase in SR produced by isoproterenol was prevented. Eleven of sixteen Purkinje fiber preparations treated with isoproterenol alone became spontaneously arrhythmic, whereas none of six treated with 8-bromo-cGMP and isoproterenol became arrhythmic (p less than 0.05). Slow-response action potentials elicited by potassium depolarization and catecholamines were abbreviated and eventually abolished by 8-bromo-cGMP. In conclusion, 8-bromo-cGMP has no effect on action potential characteristics in normally polarized canine Purkinje fibers but depressed slow response action potentials. The effects of isoproterenol on SR are antagonized and the production of arrhythmias in this model are prevented by 8-bromo-cGMP.     

Electrophysiological effects of ketamine on human atrial myocytes at therapeutically relevant concentrations

1. Ketamine is widely used for the induction of anaesthesia in high-risk patients with cardiovascular instability or severe hypovolaemia. However, the ionic mechanisms involved in the effects of ketamine at therapeutically relevant concentrations in human cardiac myocytes are unclear. The present study was designed to investigate the effects of ketamine on L-type Ca2+ (I(Ca)), transient outward K+ (I(to)), ultra-rapid delayed rectifier K+ (I(Kur)) and inward rectifier potassium (I(K1)) currents, as well as on action potentials, in human isolated atrial myocytes. 2. Atrial myocytes were isolated enzymatically from specimens of human atrial appendage obtained from patients undergoing coronary artery bypass grafting. The action potential and membrane currents were recorded in both current- and voltage-clamp modes using the patch-clamp technique. 3. Ketamine inhibited I(Ca) with an IC(50) of 1.8 micromol/L. In addition, 10 micromol/L ketamine decreased the I(Ca) peak current at +10 mV from 5.1 +/- 0.3 to 2.1 +/- 0.4 pA/pF (P < 0.01), but did not change the threshold potential, peak current potential and reverse potential. 4. Ketamine had no effect on I(to), I(Kur) or I(K1), but it reversibly shortened the duration of the action potential in human atrial myocytes. 5. In conclusion, ketamine, at a clinically relevant concentration, shortens the action potential duration of the human atrial myocytes, probably by inhibiting I(Ca).     

In vivo and in vitro acute cardiovascular effects of bimoclomol

Effects of bimoclomol, the novel heat shock protein (HSP) coinducer, was studied in various mammalian cardiac and rabbit aortic preparations. Bimoclomol decreased the ST-segment elevation induced by coronary occlusion in anesthetized dogs (56% and 80% reduction with 1 and 5 mg/kg, respectively). In isolated working rat hearts, bimoclomol increased coronary flow (CF), decreased the reduction of cardiac output (CO) and left ventricular developed pressure (LVDP) developing after coronary occlusion, and prevented ventricular fibrillation (VF) during reperfusion. In rabbit aortic preparations, precontracted with phenylephrine, bimoclomol induced relaxation (EC(50)=214 microM). Bimoclomol produced partial relaxation against 20 mM KCl, however, bimoclomol failed to relax preparations precontracted with serotonin, PGF(2) or angiotensin II. All these effects were evident within a few minutes after application of bimoclomol. A rapid bimoclomol-induced compartmental translocation of the already preformed HSPs may explain the protective action of the compound.     

Electrophysiological effects of protopine in cardiac myocytes: inhibition of multiple cation channel currents

Protopine (Pro) from Corydalis tubers has been shown to have multiple actions on cardiovascular system, including anti-arrhythmic, anti-hypertensive and negative inotropic effects. Although it was thought that Pro exerts its actions through blocking Ca(2+) currents, the electrophysiological profile of Pro is unclear. The aim of this study is to elucidate the ionic mechanisms of Pro effects in the heart. In single isolated ventricular myocytes from guinea-pig, extracellular application of Pro markedly and reversibly abbreviates action potential duration, and decreases the rate of upstroke (dV/dt)(max), amplitude and overshoot of action potential in a dose-dependent manner. Additionally, it produces a slight, but significant hyperpolarization of the resting membrane potential. Pro at 25, 50 and 100 microM reduces L-type Ca(2+) current (I(Ca,L)) amplitude to 89.1, 61.9 and 45.8% of control, respectively, and significantly slows the decay kinetics of I(Ca,L) at higher concentration. The steady state inactivation of I(Ca,L) is shifted negatively by 5.9 - 7.0 mV (at 50 - 100 microM Pro), whereas the voltage-dependent activation of I(Ca,L) remains unchanged. In contrast, Pro at 100 microM has no evident effects on T-type Ca(2+) current (I(Ca,T)). In the presence of Pro, both the inward rectifier (I(K1)) and delayed rectifier (I(K)) potassium currents are variably inhibited, depending on Pro concentrations. Sodium current (I(Na)), recorded in low [Na(+)](o) (40 mM) solution, is more potently suppressed by Pro. At 25 microM, Pro significantly attenuated I(Na) at most of the test voltages (-60 approximately +40 mV, with a 53% reduction at -30 mV. Thus, Pro is not a selective Ca(2+) channel antagonist. Rather, it acts as a promiscuous inhibitor of cation channel currents including I(Ca,L), I(K), I(K1) as well as I(Na). These findings may provide some mechanistic explanations for the therapeutic actions of Pro in the heart.     

Electrophysiological effects of tocainide on canine subendocardial Purkinje fibers surviving infarction

The effects of 10 and 20 mg/l of tocainide on transmembrane action potential characteristics were examined in Purkinje fibers surviving infarction. Infarcted tissue was obtained from canine hearts 24 h after coronary artery ligation. The preparation was stimulated at basic cycle lengths (BCL) of 1000-300 ms. Tocainide reduced the overshoot and amplitude of Purkinje fibers surviving infarction. The maximum upstroke velocity (Vmax) was decreased by tocainide in a dose dependent manner. This effect was more prominent at the shorter BCL. Statistical analysis revealed a significant interaction of the BCL with the drug effect on overshoot, amplitude, Vmax and action potential durations (APD50% and APD90%). Tocainide reduced the effective refractory period (ERP) at the BCL of 1000 ms, but had no significant effect at the BCL of 300 ms. Membrane responsiveness and steady state characteristics of Vmax were shifted significantly to more negative membrane potentials by tocainide. Investigation of the recovery kinetics of Vmax in the presence of tocainide showed an exponential recovery of Vmax with a time constant of 514 ms. These results support the finding that the effect of tocainide on Vmax and conductions is enhanced at faster rates of stimulation. Thus tocainide may be able to depress conduction to produce bidirectional block in the termination of ventricular tachycardia caused by reentry, while having minimal effect on conduction at normal heart rates.     

Electrophysiological effects of the neuroleptanalgesic drugs on the canine cardiac tissue

Summary Parameters of action and resting potentials of dog Purkinje and ventricular muscle fibers were studied after the administration of droperidol and fentanyl given separately and in combination (in a weight proportion 50:1 as in neuroleptanalgesia). Droperidol in all concentrations studied (10^–6 M, 5×10^–6 M, 10^–5 M) moderately shortened action potential duration in electrically driven Purkinje fibers. Effective refractory period was also shortened but to a lesser degree than action potential duration. The drug was also able to diminish the automaticity in Purkinje fibers. It exerted nearly no effect on resting membrane potential, action potential amplitude and on maximum rate of depolarization. The latter was only slightly decreased by the highest concentration of the drug. Also conduction velocity in Purkinje fibers was not changed by the drug. Ventricular action potentials were not significantly influenced by droperidol. Fentanyl administered in concentrations 2.2×10^–7 M and 1.1×10^–6 M was devoid of any action on cardiac transmembrane potentials. The effects exerted by both drugs administered simultaneously were similar to the effects exerted by droperidol administered as a sole agent. The electrophysiological effects of droperidol on the isolated cardiac tissue are very similar to the action of lidocaine and diphenylhydantoin.This work was presented in the preliminary form at the Eighth Annual Meeting of the European Society for Clinical Investigation, April 25–27, 1974, Rotterdam     

Electrophysiological effects of desmethylimipramine and desmethyldoxepin on canine cardiac Purkinje fibers

Two studies were conducted to assess the toxicity of rosaramicin (a macrolide antibiotic) when given intravenously for 30 consecutive days to beagle dogs with and without a tapetum lucidum (a light reflecting structure within the choroid of the eye). In the initial study, groups of three tapetal dogs/sex were given 20, 40, or 80 mg of rosaramicin/kg, twice daily. Ophthalmoscopic examination during Week 4 revealed dose-related, bilateral ocular changes characterized by a brown-tan discoloration and general pallor or loss of reflectivity of the normally blue-purple or yellow-green, highly reflective tapetum lucidum. These findings were restricted to the tapetal fundus; recovery occurred between Weeks 4 and 10 of the postdose period. To further investigate these changes, a second study was conducted in which groups of three tapetal dogs were given rosaramicin or erythromycin lactobionate (comparative macrolide antibiotic) at 80 mg/kg, twice daily. A third group of atapetal dogs was given 80 mg of rosaramicin/kg, twice daily. A similar change was observed in tapetal dogs given 80 mg of rosaramicin/kg, twice daily, in the follow-up study, but not in the other two groups. No other compound-related changes were observed in either study. The ocular changes observed in dogs given rosaramicin were reversible and structure-specific, occurring only in animals possessing a tapetum lucidum.