Pharmacological effects in animals and normal human subjects of the diuretic amiloride hydrochloride (MK-870> Amiloride is a new diuretic drug producing moderate natriuretic effects coupled with prominent potassium retention. Though chemically distinct, this compound has some structural similaritites to triamterene, which it also resembles in its effects in normal human subjects, adrenalectomized rats, and on sodium transport across the isolated frog skin. It has no appreciable effect on the urinary concentrating mechanisms. These actions suggest a primary, direct effect upon the distal renal tubule. A dose-response relationship in man is obtained with little increase in effect above 40 mg. At lower doses the effects are manifested within the first 10 hours after administration. Doses above 40 mg. prolong these effects to 24 hours. Walten I. Baba, M.B., Ph.D.,* Ariel F. Lant, B.Se., M.B., M.R.C.P.,** Anthony J. Smith, D.M., M.R.C.P., Mary M. Townshend, Ph.D.,*** and Graham M. Wilson, M.D., D.Se., F.R.C.P.**** Sheffield, England Department of Pharmacology and Therapeutics, University of Sheffield Several reports have appeared already of the pharmacological effects and therapeutic applications of amiloride hydrochloride (MK-870, amipramizine). 5, 6, 8, 13 This compound, a guanidine diuretic, possesses modest natriuretic and diuretic potencies combined with a marked ability to Supported in part by grants from the Research Fund of the University of Sheffield and the United Sheffield Hospitals Endowment Research Fund. Amiloride supplied by Merck, Sharp, and Dohme Limited. Received for publication Nov. 10, 1967. Accepted for publication Jan. 5, 1968. OPresent address: Department of Pharmacology and Therapeutics, Medical College, Baghdad, Iraq. Formerly held the J. G. Graves Medical Research Fellowship of Sheffield University. oOPresent address: Laboratory of General Physiology, University of Leicester, University Road, Leicester, England. “o”Financial support provided by the Medical Research Council. ~~OQPresent address: Department of Medicine, Western Infirmary, Glasgow, W.l., Scotland. 318 reduce urinary potassium loss. In this respect it resembles closely both triamterene and the aldosterone antagonists: but it has been claimed that its action is not mediated by an inhibition of aldosterone secretion or a reversal of this hormone’s effects on the renal tubule.s The structural formula of amiloride, though distinct from that of triamterene, shares certain features with it (Fig. 1). Both contain an amino-substituted pyrazine ring and each molecule has three free amino radicals. An analogy can be drawn between the 5-amino-6-chloro substitution on the pyrazine ring of amiloride and the meta-sulfamoyl-halogenated substitution on the benzene ring which features in the structure of all benzothiadiazines, as well as frusemide. Our concern with this new diuretic has been to define its duration of action, its Volume 9 Number 3 , , , C1t’#”NX’. CO—’NH __ / / I : I Amiloride / H N “‘” I J..-NH2 : 2 N ,’NH2 HN ~ …. – ,,/ – -… _ … Triamterene Fig. 1. Comparison of structural formulas of amiloride and triamterene. effect on osmolal and free-water clearance, and to examine the dose-response relationship in normal human subjects. The results of these studies and the close resemblance of amiloride to triamterene led to a comparison of their effects in the normal human subject, in adrenalectomized animals, and on the active transport of sodium ions through the isolated frog skin. The results presented here were obtained as a preliminary to the performance of clinical trials in patients with fluid retention.

Materials and methods

Dose-response relationship and duration of action. Six healthy male subjects, aged 21 to 30 years, acted as volunteers for these experiments. They ate a normal diet and were allowed to drink water freely, although tea and coffee were withheld during the experiments. Urine was obtained in two collection periods-one over the first 10 hours and the second over the 14 hours from 10 to 24 hours after taking the drug. Each subject received amiloride in oral doses of 5, 10, 20, 30, 40, and 60 mg. These doses were administered according to a Latin-square design at intervals of one week. A control day preceded each drug day. In order to define the period of maximal activity of amiloride, a single oral dose of 40 mg. was given to six more healthy male subjects. The drug was taken at 8 A.M. and urine collections made at hourly intervals for the next 10 hours. Similar coIlecPharmacological effects of amiloride 319 tions without prior administration of amiloride were made by each subject on a separate day and from these were derived the control values. Venous blood was withdrawn from each subject at 2 P.M. on the control day and at 10 A.M. and 2 P.M. on the day the drug was administered. Urinary and serum electrolytes were estimated by the methods described below. Effects on osmolal and free water clearance. Two male subjects, aged 23 and 35 years, fasted overnight and at 7 A.M. initiated a water diuresis by drinking 1 L. of water. A further half liter was taken at 8 A.M. after the bladder had been emptied and thereafter a constant state of hydration was maintained. Urine collections were made at hourly intervals and water equivalent to the volume of urine passed was then taken. Control observations were carried out until 12 noon when both subjects received an oral dose of 40 mg. of amiloride. Urine collections and water replacement were continued at hourly intervals for a further 4 hour period. At the midpoint of both control and postdrug periods, venous blood samples were withdrawn without stasis. Comparison of the effects of amiloride and triamterene. In the normal human sub;ect. On two separate days a healthy male subject made urine collections at hourly intervals for 10 hours. The mean results of the 2 days served as control values. On a third day, a single dose of 200 mg. of triamterene was taken orally and the effects on hourly urine volume and composition measured over a 10 hour period. This response was compared with that following an oral dose of 40 mg. of amiloride in the same subject, urine collections being made in an identical manner. In adrenalectomized rats. Adrenalectomy was performed on 18 rats under ether anesthesia. Animals were considered to be fully adrenalectomized only if both adrenal glands were removed with their capsules intact. Following surgery the rats were given a standard diet of commercial 320 Baba et al. rat cake and were allowed free access to 1 per cent sodium chloride as drinking fluid. All experiments were performed between the second and third weeks following the adrenalectomy. The animals were divided into 6 groups of 3 and matched with 6 further groups of 3 rats each with intact adrenals. Nine adrenalectomized and 9 normal animals received amiloride by stomach tube in a dose of 1.5 mg. per kilogram. The remaining 9 adrenalectomized and 9 normal animals served as controls and received only water by tube. The pooled urine from each group of 3 rats placed in metabolic cages was collected over periods of 5 hours and the electrolyte contents measured. A week later the control and treated animals were interchanged and the experiments repeated. On sodium transport across the isolated frog skin. CURRENT MEASUREMENTS. Isolated frog skin was mounted between Perspex halfcells in a medium of frog Ringer’s solution and circuits arranged to enable the measurement of the current generated by the skin when short-circuited by an externally applied potential difference.2, 12 Following a control period to allow stabilization of the short-circuit current, amiloride (12 or 24 p.g) or triamterene (300 p.g) dissolved in frog Ringer’s solution was added to the fluid bathing either the outer or inner skin surface and changes in short-circuit current measured. Following the full development of the drug’s effect the fluid containing the drug was removed, the skin washed in frog Ringer’s solution, and further readings of the short-circuit current made. FLUX MEASUREMENTS. The unidirectional passage of sodium ions across the frog skin was measured by serial sampling from the fluid bathing the opposite side when approximately 100 flCi 24Na had been added to the outside (for influx) or inside (for effiux) bath fluid. Following a control period of one hour, amiloride was added to the bath fluid on the outside of the skin and flux measurements determined by sampling for one more hour. Simultaneous reClinical Pharmacology and Therapeutics cording of the short-circuit current enabled comparison between sodium flux expressed in milli-Coulombs per square centimeter per hour and short-circuit current expressed in the same units. From measurements of both influx and effiux of sodium, the effect of amiloride on the skin resistance to the passage of sodium ions and the electromotive force of the sodium transporting mechanism were calculated with the formulas of U ssing.ll These results were compared with those previously obtained in the frog skin preparation with triamterene. 4 Analytical methods. The volume of each urine collection was recorded and an aliquot stored under toluene for subsequent analysis. Sodium and potassium concentrations were measured with a Technicon auto-analyzer. Urinary chloride levels were measured both by a mercuric nitrate method10 and also on the auto-analyzer. Good agreement was found between these values and, accordingly, the data obtained from the auto-analyzer are presented in the tables of results and the figures. Urine bicarbonate levels were measured by a microdiffusion method. Urinary and serum creatinine were measured on the auto-analyzer by a modification of the Jaffe alkali-picrate method and both urine and serum osmolalities were measured by an Advanced Instruments osmometer. Determinations of pH were made on a Pye pH meter. Results Dose-response and duration of action. Increasing effects are produced by doses ranging from 5 to 40 mg., but above 40 mg. the responses reach a plateau (Fig. 2). This applies to the increase in urinary volume, sodium and chloride excretion, and also to the consistent reduction in potassium loss (Table I). The effects over the periods 0 to 10 hours and 0 to 24 hours after administration of amiloride have been compared. Some diuretic effect persists for more than 10 hours after the drug has been given and this is seen more clearly with doses in excess of 30 mg. (Fig. 2). Volume 9 Number 3 Pharmacological effects of amiloride 321 UCinarl Chloride (mEq) o Urinary -10 Potassium (mEq) Urinary Sodium (mEq) Urinary Volume (ml) -20 -30 -40 200 150 100 50 0 1200 1000 800 600 400″: 2001 0 0 T 5 10 20 30 40 60 Dose (mg) Fig. 2. Dose-response relationship: Changes from control values ± S.E.M. in urinary electrolytes and volume following six different doses of amiloride in six healthy subjects (e–e = o to 10 hours; 0—-0 = 0 to 24 hours).

Table I.

Increase in sodium and potassium excretion values over control after single doses of 40 mg. of amiloride in 6 healthy subiects Hours after drug administration Age Weight Nad-Nac I 10-24 Kd-Kc I 10-24 Sub;ect (years) (Kg.) 0-10 (mEq.) 0-10 (mEq.) WB 29 74.0 + 81.90 +71.98 -17.51 + 1.85 RB 23 66.5 + 83.16 +33.23 -12.36 -24.20 JGS 23 61.4 +118.75 – 1.94 -35.39 -15.42 HL 32 68.5 + 79.56 +71.93 -20.55 – 7.96 JWS 23 86.0 + 90.20 +69.07 -36.60 -13.36 MT 22 74.5 +170.97 -12.42 -29.32 -26.42 Mean +104.09 +38.64 -25.28 -14.25 ± S.E.M. ± 14.62 ±15.51 ± 4.06 ± 4.26 Nad and Kd stand for sodium and potassium excretion following drug. Nac and Kc stand for sodium and potassium excretion on control day. 322 Baba et al. Clinical Pharmacology and Therapeutics

Table II.

Effect of 40 mg. of amiloride on maximum water diuresis in two normal sub;ects Subject JS Age: 23 Weight: 86 Kg. Male AS Age: 35 Weight: 70 Kg. Male uv – urinary volume; = serum creatinine. pH Bicarbonate Chloride mEq I per Potassium hour Sodium Urinary Volume (ml/hr) Treatment Control Amiloride Control Amiloride UOsm. – urinary osmolality; Control UV U08 m. Cosmo (ml.!min.) (mOsm./Kg.) (ml./min.) 17.5 59 3.5 17.5 75 4.4 7.2 113 2.7 7.2 127 3.0 Cosmo = osmolal clearance; CH20 free-water clearance; UN.V Triamterene 200mg Amiloride 40mg Hours Fig. 3. Comparison of effects of 200 mg. of triamterene and 40 mg. of amiloride in a normal subject. Control data represent tbe mean of two separate periods of observation. The response to a single dose of 40 mg. begins in the second hour, reaching a maximum between 4 and 6 hours and waning thereafter. The reduction in potassium loss falls to control values close to 9 hours, although the natriuretic and chloruretic effects persist beyond 10 hours. Urinary bicarbonate excretion is also increased by 40 mg. of amiloride and this is most prominent between the second and sixth hours after taking the drug. Effects on osmolal and free-water clearance. The data obtained from these experiments are presented in Table II. Comparable urine flow rates were achieved by each subject during control and amiloride periods. The natriuretic effects of the drug were apparent in both subjects and were accompanied by an increase in osmolal clearance. Potassium retention also occurred and no significant change was seen in serum creatinine levels. The free-water Volume 9 Number 3 CH,o (ml.!min.) 14.0 13.1 4.5 4.2 UNaV UKV (p,Eq/min.) (p,Eq/min.) 116 159 329 42 131 89 264 36 urinary sodium output; UKV – urinary potassium output; UCIV n CONTROL • AMILORIDE URINARY :::1 [l POTASSIUM n… 200 ~Eq/5hr OL–L–.. ——~–.. URINARY SODIUM ~Eq/Shr URINARY VOLUME ml/Shr 2000 1500 1000 500 OL–L- “~1 cI [J 0 NORMAL ADRENALECTOMIZED Fig. 4. Effects of amiloride (1.5 mg. per kilogram) in rats showing potassium-sparing and natriuretic responses in both normal and adrenalectomized animals. clearance was not influenced significantly in either subject. Comparison of the effects of amiloride and triamterene. In the normal human subject. The doses compared were 200 mg. of triamterene and 40 mg. of amiloride and at these levels amiloride produced the greater effects on water diuresis and sodium, chl<;lride, and Pharmacological effects ot amiloride 323 UcN SO:;m. Sa, (p,Eq/min.) (mOsm./Kg.) (mg./1OOml.) 138 296 0.9 175 296 0.9 151 295 1.1 151 299 1.0 – urinary chloride output; SOsm. – serum osmolality; and SCr bicarbonate excretion (Fig. 3). The depression of potassium loss was comparable with each compound. Adrenalectomized rats. Both the natriuretic and potassium-sparing response to amiloride were preserved in adrenalectomized animals (Fig. 4). On sodium transport across the isolated frog skin. CURRENT MEASUREMENTS. Amiloride produced no alteration in the short-circuit current across isolated frog skin when added to the Ringer’s solution bathing the internal surface. When added to the outer solution, a dose of 12 fLg (concentration: 8 x 1O-7M) produced an immediate and profound drop in the short-circuit current. This effect was readily reversed by washing the skin surface with frog Ringer’s solution (Fig. 5). The pH of the solution was not altered by the addition of amiloride. The results of similar experiments with triamterene are set out in Fig. 6. FLUX MEASUREMENTS. Amiloride produced a consistent reduction in the inward movement ( influx) of sodium ions, although the effect was not always apparent during the first half hour after addition of the drug(Table III).No consistent effect was observed on the efflux of sodium ions. The electromotive force of the sodiumtransporting mechanism and the partial resistance of the skin to the flow of sodium ions were calculated from measurements of sodium efflux and the short-circuit cur- 324 Baba et al. MK 870 Wash with Ringer (x2) 20 0 PERCENTAGE CHANGE IN -20 SHORT CIRCuiT -40 CURRENT -60 -80 -100 12pg ! 0 10 20 30 40 TIME (min) 50 24 pg ! l 60 70 Clinical Pharmacology and Therapeutics 80 Fig. 5. Effects of 12 and 24 Ilg of amiloride on the short-circuit current of isolated frog skin and reversal by washing the skin. Triamterenc 300pg ! 1 Wash with Ringer (x2 ) 20 0 \. / PERCENTAGE CHANGE IN SHORT -20 CIRCUIT CURRENT -40 -60 -80 i i ,. –,——, 80 90 130 140 180 190 TIME(min) Fig. 6. Effect of 300 Ilg of triamterene on the short-circuit current of isolated frog skin. Current restored to normal by washing the skin with Ringer’s solution. Volume 9 Number 3 Pharmacological effects of amiloride 325

Table III.

Sodium (Z4Na) influx: Effect of 20 fhg of amiloride applied to outside skin surface (4 experiments) Mean Na influx short-circuit current Na influx (mCoul./cm.z/ Short-circuit current Period (/lAmps) (/lEq/cmN30 min.) 30 min.) (mCoul.) 1. Control 1 68.1 2 64.5 After drug 1 13.7 2 9.0 3 9.0 2. Control 1 94.5 2 113.0 After drug 1 30.0 2 23.0 3 23.0 3. Control 1 80.2 2 77.7 After drug 1 25.6 2 10.0 3 10.0 4. Control 1 50.0 2 39.0 After drug 1 11.5 2 9.0 3 10.0 rent (Table IV). Amiloride consistently lowered the electromotive force and increased the partial resistance of the skin. Discussion Amiloride is an interesting addition to the available range of diuretic drugs and possesses the important property of reducing urinary potassium loss. The results of the experiments discussed in this paper show that at many points this compound behaves in a way similar to triamterene. This resemblance is seen in the chemical formulas, as both compounds possess a 0.377 36.4 43.3 0.257 24.8 41.0 0.225 21.7 8.7 0.192 18.5 5.7 0.154 14.8 .’5.7 0.704 67.9 60.1 0.710 68.5 71.9 0.509 49.1 19.1 0.316 30.5 14.6 0.277 26.8 14.6 0.417 40.3 51.0 0.410 39.5 49.4 0.391 37.7 16.3 0.162 15.6 6.4 0.221 21.3 6.4 0.456 44.0 31.3 0.298 28.7 24.4 0.379 36.5 7.2 0.216 20.8 5.6 0.207 20.0 6.3 pyrazine ring with amino substitution in the 5 position. In the isolated frog-skin preparation, amiloride produced a reduction in shortcircuit current and active sodium transport. This is apparently mediated by an increase in resistance and diminution in the electromotive force of the sodium-transporting mechanism. These results correspond closely to those obtained with triamterene.4 In adrenalectomized rats, both the natriuretic and potassium-sparing effects of amiloride were preserved, suggesting that neither is mediated by an interference 326 Baba et al. Clinical Pharmacology and Therapeutics

Table IV

Sodium eWa) efflux: Effect of 20 f-tg of amiloride applied to outside skin surface 1. 2. 3. 4. Period Control 1 2 After drug 1 2 Control 1 2 After drug 1 2 Control 1 2 3 After drug 1 2 3 Control 1 2 3 After drug 1 2 3 Mean S.C.C. (/lAmps) 105.8 103.4 29.6 18.5 77.3 80.9 35.2 32.0 134.5 143.5 132.3 24.1 10.0 10.0 96.2 115.8 111.4 22.5 11.0 11.0 Efflux Na (pEq/cmH 30 min.) 0.0500 0.0137 0.0262 0.0391 0.113 0.165 0.135 0.156 0.0232 0.0446 0.0291 0.0483 0.0313 0.0361 0.0444 0.0639 0.0288 0.0497 0.0434 0.0872 Efflux Na (mCoul.) 4.83 1.32 2.52 3.77 10.93 15.94 13.05 15.08 2.24 4.30 2.81 4.66 3.02 3.48 4.29 6.16 2.77 4.79 4.19 8.41 S.C.C. (mCoul.) 67.3 65.8 18.8 11.8 48.3 50.6 22.0 20.0 84.1 89.7 82.7 15.1 6.3 6.3 60.1 72.4 69.6 14.1 6.9 6.9 ENa (mV.) 69.4 100.8 54.8 36.4 43.4 36.7 25.4 21.7 93.7 79.2 87.7 37.0 28.8 26.4 69.6 65.3 83.7 35.2 24.9 15.3 RNa (Ohms/cm. 2 ) 1,857 2,760 5,243 5,567 1,616 1,305 2,075 1,950 2,097 1,589 1,908 4,424 8,291 7,599 2,082 1,624 2,164 4,500 6,528 4,017 S.C.C. short-circuit current; EXa electromotive force of sodium transport mechanism; R:’\ a = partial resistance to passage of sodium ions. with the production or peripheral action of adrenocortical steroids and, in particular, of aldosterone. Similar results were obtained with triamterene. 3 The third comparison, in a normal human subject, again revealed a qualitative similarity between the two compounds. Amiloride (40 mg.) proved more potent than 200 mg. of triamterene in promoting sodium loss and water diuresis but produced closely similar effects on urinary potassium and bicarbonate excretion and on urinary pH. Many potent diuretics reduce free-water clearance, a fact which has been turned to therapeutic use in the treatment of diabetes insipidus. 1 Thus it was of interest to find that amiloride produced no change in freewater clearance despite a moderate natriuretic effect in two subjects undergoing water diuresis. This observation may be related to that of Dow and Irvine, 7 who showed that amiloride did not affect corticomedullary gradients of sodium and urea in the rat kidney. It is also in accord with results previously obtained with triamterene.i , U Wilson and colleagues13 have de- Volume 9 Number 3 scribed a slight decrease in free-water clearance with amiloride. The relatively modest natriuretic action of amiloride, the lack of effect on urinary concentrating mechanisms, and its close resemblance to triamterene suggest that the predominant site of action is the distal renal tubule. It appears to act independently of aldosterone or carbonic anhydrase inhibition and may produce its effect by a direct interference with cellular transport mechanisms. The results obtained from this assessment suggested that amiloride deserved further investigation as a diuretic in patients with fluid retention. The results of a clinical investigation will be reported separately. We wish to thank all the volunteers who took part in these studies; the technicians of the Department of Pharmacology and Therapeutics, Royal Infirmary, Sheffield; and Mrs. J. Leicester for her secretarial help.

References

1. Baba, W. I., Lant, A F., and Wilson, G. M.: The action of oral diuretics in diabetes insipidus, Proc. Roy. Soc. Med. 58:911-912, 1965. 2. Baba, W. I., and Smith, A J.: The effect of guanethidine on sodium transport across the isolated frog skin, Quart. J. Exper. Physiol. 49: 194-198, 1964. 3. Baba, W. I., Tudhope, G. R., and Wilson, G. M.: Triamterene, a new diuretic drug, Brit. M. J. 2:756-764, 1962. 4. Baba, W. I., Tudhope, G. R., and Wilson, G. M.: Site and mechanism of action of the diPharmacological effects of amiloride 327 uretic, triameterene, Clin. Sc. 27:181-193, 1964. 5. Baer, J. E., Jones, C. B., Spitzer, S. A, and Russo, H. F.: The potassium-sparing and natriuretic activity of N-amidino-3,5-diamino6-chloropyrazinecarboxamide hydrochloride dihydrate (amiloride hydrochloride), J. Pharmacol. & Exper. Therap. 157:472-485, 1967. 6. Baer, J. E., Mucha, C. M., Spitzer, S. A, and Yee, H. W.: A potassium-sparing pyrozinamide derivative, Fed. Proc. 25:197, 1966. 7. Dow, J., and Irvine, R. O. H.: The effects of diuretics on cortico-medullary gradients of sodium and urea in the rat kidney, Nephron 4: 25-31, 1967. 8. Glitzer, M. S., and Steelman, S. L.: N-amidino3,5-diamino-6-chloropyrazine-carboxamide: An active diuretic in the carboxamide series, Nature 212:191-193, 1966. 9. Lant, A F., Baba, W. I., and Wilson, G. M.: Localization of the site of action of oral diuretics in the human kidney, Clin. Sc. 33: 11- 27, 1967. 10. Schales, 0., and Schales, S. S.: A simple and accurate method for the determination of chloride in biological fluids, J. BioI. Chern. 140:879-884, 1941. 11. Ussing, H. H.: The active ion transport through the isolated frog skin in the light of tracer studies, Acta physiol. scandinav. 17:1- 37, 1949. 12. Ussing, H. H., and Zerahn, K.: Active transport of sodium as the source of electric current in the short-circuited, isolated frog skin, Acta physiol. scandinav. 23:110-127, 1951. 13. Wilson, J. D., MK-870 Richmond, D. E., Simmons, H. A, and North, J. D. K.: MK 870: A new potassium-sparing diuretic, New Zealand M. J. 65:505-511, 1966.