Antagonist Activity of Formoterol and Salmeterol Thorax: First Published As 10.1136/Thx.51.1.54 on 1 January 1996

54 Thorax 1996;51:54-58

Evaluation of the 132 adrenoceptor agonist/

antagonist activity of formoterol and salmeterol Thorax: first published as 10.1136/thx.51.1.54 on 1 January 1996. Downloaded from

A Grove, B J Lipworth

Abstract ist), salmeterol and formoterol, like Background - Salmeterol and formoterol propranolol, exhibited P2 receptor ant- have a lower intrinsic activity at P2 re- agonism as evidenced by their attenuation ceptors than isoprenaline in human bron- of P2 receptor mediated responses. The chus in vitro. The aim ofthe present study degree of P2 blockade with formoterol and was to evaluate in vivo the P2 agonistlant- salmeterol was comparable but less than agonist activity of salmeterol and for- with propranolol. The relevance of these moterol at rest with low endogenous findings at extrapulmonary P2 receptors adrenergic tone, on exercise with raised with regard to airway P2 responses remains endogenous adrenergic tone, and in the unclear and warrants further in- presence of fenoterol, an exogenous full P2 vestigation. receptor agonist. (Thorax 1996;51:54-58) Methods - Eight normal subjects were randomised to receive single doses ofplacebo, Keywords: salmeterol, formoterol, 02 agonist/ant- salmeterol 300 sg, formoterol 72 pg, or pro- agonist, intrinsic activity. pranolol 80 mg at weekly intervals. P2 adrenoceptor responses were evaluated at rest, at peak exercise, and after treatment Salmeterol and formoterol are potent long act- with fenoterol 2 4 mg. ing 2 receptor agonists. The potency of these Results - At rest salmeterol and formoterol drugs will determine the dose required to pro- exhibited equivalent P2 agonist activity duce a given effect, but it is important to with regard to decrease in serum potas- appreciate that factors other than potency may sium levels and increase in finger tremor, influence the expression of their 12 agonist with propranolol having no effect. Sal- activity. Both salmeterol and formoterol have

meterol and formoterol, like propranolol, a lower intrinsic activity than isoprenaline in http://thorax.bmj.com/ potentiated the hyperkalaemic delta re- human bronchus in vitro, and therefore at sat- sponse to exercise compared with placebo, urating concentrations will produce a lower consistent with P2 antagonism: (mean maximal response, with salmeterol having a difference and 95% confidence interval lower efficacy than formoterol.' In other words, (CI) compared with placebo) salmeterol they are less efficient than a full agonist at 0-20 (0.02 to 0.38) mmol/l, formoterol 0-17 activating the receptor transduction mech- (0.00 to 0.34) mmol/l, propranolol 0 45 anism required to produce a maximal cellular to blunted (0.08 0.82) mmol/l. Propranolol response. Thus, as partial agonists the ex- on September 30, 2021 by guest. Protected copyright. the heart rate delta response to exercise, pression oftheir P2 agonist or antagonist activity consistent with P, blockade, whilst sal- may be modulated by the presence of a full 12 meterol and formoterol had no effect. Sal- agonist such as isoprenaline or adrenaline. It meterol and formoterol, like propranolol, could be predicted from first principles that, attenuated the hypokalaemic, tremor, and in the presence of a low concentration of a full heart rate delta responses to fenoterol agonist, salmeterol and formoterol would be compared with placebo, in keeping with P2 expected to behave as 12 agonists, occupying blockade: potassium, salmeterol 0-18 (0.0 additional receptors and therefore augmenting to 0.36) mmol/l, formoterol 0-17 (-0.03 the overall response. However, in the presence to 0.37) mmolll, propranolol 0-80 (0.54 to ofhigh concentrations ofa full agonist they may 1-06) mmolll; tremor, salmeterol -0-69 behave as 12 antagonists because, by occupying (-1.26 to -0.12) log units, formoterol receptors that would otherwise have been oc- Department of -0-71 (-1.53 to 0.11) log units, pro- cupied by the full agonist, they will reduce the Clinical pranolol -0 85 (-166 to -0.04) log units; Pharmacology, overall response.2 Ninewells Hospital and heart rate, salmeterol -6 (-13 to 1) beats/ By evaluating the 12 receptor mediated Medical School, min, formoterol -10 (-19 to -1) beats/ modulation of exercise-induced hyperkalaemia University of Dundee, propranolol -18 (-29 to -7) beats/ to raised levels of Dundee DD1 9SY, UK min, due endogenous adrenaline, A Grove min. we have previously shown that oral salbutamol B J Lipworth Conclusions - At rest with low endogenous exhibits P2 antagonist effects at extrapulmonary Reprint requests to: adrenergic tone salmeterol and formoterol 12 receptors in vivo.' We were therefore in- Dr B J Lipworth. showed equivalent P2 mediated agonist terested to investigate whether similar effects Received 30 January 1995 activity in terms of serum potassium and occurred with inhaled long acting 12 agonists Retumed to authors 20 July 1995 finger tremor responses. In the presence given that they, like salbutamol, are not full 12 Revised version received of raised endogenous adrenergic tone at receptor agonists. 14 August 1995 Accepted for publication peak exercise and in the presence of feno- The aim of the present study was to evaluate 4 September 1995 terol (an exogenous full P2 receptor agon- in vivo P2 agonist/antagonist activity of sal- /2 agonistlantagonist activity offormoterol and salmeterol 55

meterol and formoterol at extrapulmonary 02 MEASUREMENTS receptors at rest when there is low endogenous Heart rate was monitored and recorded using adrenergic tone, and during exercise when a 12-lead electrocardiograph with stress test

endogenous adrenergic tone is high. adaptor (HP4700A, Hewlett Packard, Palo Thorax: first published as 10.1136/thx.51.1.54 on 1 January 1996. Downloaded from Furthermore, recent in vitro data have shown Alto, California, USA). Heart rate was cal- that salmeterol inhibits fenoterol and sal- culated from the mean of five consecutive R-R butamol induced relaxation in precontracted intervals. Postural finger tremor was measured guinea pig airways.4 Thus, 12 receptor mediated by a previously validated method9 with an responses were also evaluated in the presence accelerometer transducer (Entran Ltd, Ealing, offenoterol, an exogenous full P2 receptor agon- UK) attached to the distal phalanx of the ist.5-7 Propranolol was included as a positive middle finger. Each measurement consisted of control for the effects of 1,3 and 12 receptor four recordings, the results ofwhich were stored antagonism. on computer disc for subsequent spectral analysis of total tremor power >2 Hz (units mg2/s) using computer assisted autocovariance. The mean of three consistent recordings was used Methods in subsequent analysis. Serum potassium was SUBJECTS analysed by flame photometry (IL943 analyser, Eight healthy volunteers (two women) with a Instrumentation Lab Ltd, Warrington, UK), mean (SE) age of 27 (1 9) years completed the samples being analysed in batches at the end study. Each volunteer had a normal physical of the study with each sample analysed in examination including 12-lead ECG, haem- duplicate. The normal reference range for our atological and biochemical screen prior to in- laboratory is 3 5-5 5 mmol/l, and the co- clusion in the study. None of the subjects was efficients ofvariability for analytical imprecision taking regular medication. The study was within and between assays are 041 % and approved by the local ethics committee, and 1 04%, respectively. subjects gave informed written consent.

STATISTICAL ANALYSIS Finger tremor data were transformed using PROTOCOL logarithm to base 10 to achieve conformation A single blind randomised (Latin square) cross- to a normal distribution prior to analysis. Data over, placebo controlled design was used. Sub- were then analysed using the Statgraphics soft- jects attended the laboratory on four occasions ware package (STSC Software, Rockville, at least one week apart. USA). Effects at rest and in response to exercise An intravenous cannula was inserted into a and fenoterol were all analysed as delta re- http://thorax.bmj.com/ forearm vein to allow blood sampling. After 30 sponses - that is, the difference between the minutes of supine rest baseline measurements response before and after drug administration (to) of heart rate, postural finger tremor, and (tl-t0), before and after exercise (t2-tl), and serum potassium were made. Subjects then before and after fenoterol (t4-t3). Comparisons received inhaled salmeterol 300 jig (Serevent were made by multifactorial analysis ofvariance metered dose inhaler, 25 gg per actuation, Allen (MANOVA) using subjects, treatments, and and Hanburys, Uxbridge, Middlesex, UK), or time as within factors for the analysis. Where

inhaled formoterol 72 gg (Foradil metered dose the overall MANOVA was significant, Dun- on September 30, 2021 by guest. Protected copyright. inhaler, 12 gig per actuation, Ciba-Geigy, Basel, can's multiple range testing was used to estab- Switzerland), or inhaled placebo or oral pro- lish where differences between treatments were pranolol, 80 mg, with double dummies as ap- significant. Differences from placebo, where propriate. One hour after administration of the significant, were calculated as means and 95% drugs measurements of potassium, heart rate, confidence intervals. A probability value of and tremor were repeated (t,) and subjects then p<005 (two tailed) was considered significant underwent a standardised three minute exercise for all tests. step test in order to produce a maximal heart rate response.8 Peak exercise heart rate was recorded, and a blood sample was taken im- Results mediately on completion of the exercise for SUPINE REST estimation of serum potassium (t2) (tremor was There were no significant differences in base- not recorded on completion of the exercise). line (to) measurements between each ofthe four Subjects then rested supine for 30 minutes. study days for any of the measured parameters. Further measurements ofpotassium, heart rate One hour after dosing (t,) salmeterol and and tremor were made (t3) before subjects formoterol significantly (p<0 05) lowered received inhaled fenoterol 2-4 mg (Berotec 200 serum potassium levels (as delta response, ti-to) metered dose inhaler, 200 jig per actuation, compared with placebo, an effect consistent Boehringer Ingelheim, Bracknell, UK). This with 12 agonism: (mean differences and 95% was administered as 12 sequential puffs over a CI versus placebo) salmeterol -038 (-0-65 period of four minutes. Final measurements to -0 11) mmol/l, formoterol -0-36 (-0-63 (t4) were made 30 minutes after the ad- to-0- 09) mmol/l. Likewise, the finger tremor ministration of the fenoterol. At the end of delta response was significantly (p<005) each of the four study days subjects received greater with salmeterol and formoterol than 32 mmol effervescent potassium (Sando K, with placebo: salmeterol 0-56 (-0-16 to 1.28) Sandoz Pharmaceuticals, Camberley, UK). log units, formoterol 0 61 (-0 19 to 141) log 56 Grove, Lipworth

Mean (SE) absolute values Treatment Pre-drug Post-druglpre-exercise Post-exercise Pre-fenoterol Post-fenoterol (ta) (td (td) ft-) (td

Potassium Placebo 3 74 (0-08) 3-84 (0 07) 4 07 (0-11) 3-86 (0 07) 2-93 (0 07) Thorax: first published as 10.1136/thx.51.1.54 on 1 January 1996. Downloaded from (mmolIl) Salmeterol 3-74 (0-08) 3-48 (0 07)* 3-91 (0-11) 3 50 (0 07)* 2 75 (0 07) Formoterol 3 81 (0 08) 3-56 (007)* 3 91 (0 11) 3-58 (0 07)* 2 80 (0 07) Propranolol 3-81 (0 08) 3 95 (0 07) 4-60 (0 10)* 4-08 (0 07) 3 95 (0O07)* Heart rate Placebo 72 (3) 73 (5) 162 (4) 84 (2) 99 (3) (beats/min) Salmeterol 70 (3) 81 (5) 162 (4) 95 (2)* 105 (3) Formoterol 73 (3) 74 (5) 165 (4) 90 (2) 97 (3) Propranolol 73 (3) 71 (5) 129 (4)* 73 (2)* 71 (3)* Finger tremor Placebo 1-73 (0 12) 1-58 (0 27) 1 70 (0-18) 2 68 (0 11) (log units) Salmeterol 1-88 (0-11) 2-20 (0-18) 2 27 (0 12)* 2-60 (0-11) Formoterol 1-93 (0-10) 2-28 (0 35) 2 28 (0 18)* 2-38 (0-16) Propranolol 1 78 (0 10) 1-58 (0-24) 1 67 (0 18) 1-69 (0 20)* t =before treatment; t1 =one hour after treatment; t2 = immediately after exercise; t3 = before treatment with 2-4 mg fenoterol; t4 30 minutes after treatment with fenoterol. * Significant difference from placebo at a given time point.

units. Absolute values for potassium levels but and formoterol had no effect on this response. not finger tremor or heart rate were significantly Propranolol significantly (p<0 05) potentiated different with salmeterol and formoterol com- the rise in potassium on exercise (t2-tI) con- pared with placebo (table). Propranolol had no sistent with 2 antagonism. Salmeterol and for- significant effect on resting serum potassium moterol, like propranolol, also potentiated the level, finger tremor, or heart rate compared potassium response to exercise compared with with placebo. placebo: salmeterol 02 (0-02 to 0-38) mmol/ 1, formoterol 0-17 (0 0 to 0 34) mmol/l, propranolol 0 45 (0-08 to 0 82) mmol/l. EXERCISE RESPONSES Propranolol produced significantly greater po- Propranolol significantly (p<005) attenuated tentiation of the delta potassium response than the P3 receptor mediated heart rate response to either salmeterol or formoterol. The absolute exercise (t2-tl) compared with placebo, con- post exercise potassium level (t2) was sig- sistent with 0l antagonism: (mean difference nificantly higher (p<0 05) with propranolol and 95% CI compared with placebo) -31 than with placebo, but salmeterol and for- (-54 to -8) beats/min (fig 1). Salmeterol moterol were not significantly different from placebo (table). http://thorax.bmj.com/

RESPONSES TO FENOTEROL Propranolol significantly (p<005) attenuated the potassium, heart rate, and tremor response to fenoterol (t4-t3) compared with placebo,

.0 consistent with P2 antagonism: potassium 0-80

(0 54 to 1-06) mmol/l, heart rate -18 (- 29 on September 30, 2021 by guest. Protected copyright. IVr to -7) beats/min, tremor -0-85 (-1-66 to -0 04) log units (fig 2). Salmeterol and formoterol also significantly (p<005) attenuated the hypokalaemic and tremor response to fenoterol compared with placebo: potassium: salmeterol 0-18 (0-0 to 0536) mmol/l, formoterol 0.8 FB 0-17 (-0 03 to 0537) mmol/l; tremor: salmeterol -0-69 (-1-26 to -0 12) log units, formoterol -0-71 (-0 53 to 0 11) log units. In 0.6 _ addition, formoterol also significantly (p<005) attenuated the heart rate response to fenoterol * 0E compared with placebo, with salmeterol show- 0.4 E _ ing a similar trend (p=0= 06): formoterol - 10 (-19 to -1) beats/min, salmeterol -6 (-13 Ve to -1) beats/min. Propranolol produced sig- ii~~~~ 0.2 _ nificantly greater blunting of fenoterol induced delta heart rate and delta potassium responses (but not tremor responses) than either sal- u.uII meterol or formoterol. There were no sig- PL SMT FORM PR nificant differences in the absolute values for serum potassium, finger tremor, or heart rate Figure 1 i,/1132 mediated effects on peak exercise. A comparison of the effect of treatment with single doses of after fenoterol following treatment with sal- placebo (PL), salmeterol 300 /ig (SMT), formoterol 72 ig meterol or formoterol compared with placebo. (FORM), and propranolol 80 mg (PR) on (A) delta The absolute values for heart rate and finger heart rate (/3) and (B) delta serum potassium (fl3). Values are given as means and pooled SE. *p<0.05 versus tremor after fenoterol were significantly lower placebo. (p<005) after treatment with propranolol than 12 agonistlantagonist activity offormoterol and salmeterol 57

0 - sodium/potassium ATPase driving potassium * back into the cells.12 receptor antagonists _-** such as propranolol have therefore been shown the rise in potassium levels ~-8 A 74to Thorax: first published as 10.1136/thx.51.1.54 on 1 January 1996. Downloaded from exercise.'' Furthermore, digitalis which directly .6 inhibits the sodium/potassium ATPase has also -0 - E been found to augment exercise-induced hyperkalaemia.11 Thus, in the present study E .-4 the potentiation of the potassium response to Je0. exercise by salmeterol and formoterol, like pro- -2 * pranolol, was consistent with 12 receptor ant- ffiHbT agonism. The heart rate response to exercise = is mediated viaP1 adrenoceptors, as indicated O. PL SMT FORM PR by the lack of effect of the selective 12 blocker ICI 118 551 on this response.'2 In this study B propranolol significantly attenuated the heart with its 11.2 rate response to exercise, consistent P1 - blocking properties. Salmeterol and formoterol 1n .0- had no effect, thus showing that the 13 receptor antagonism of these drugs, as one might pre- C O1 - .8 dict, is 12 receptor selective. In contrast, the - fenoterol 01 -* heart rate response to drugs such as 7 1314 0 ).6 Tis predominantly 12 receptor mediated.6 'a00 _.4 - 1;7, * Thus, as salmeterol and formoterol, like propranolol, attenuated this response, this is again

0 in keeping with 12 blockade.'5 -.2 It is important to consider the possible con- 0 founding influence of pre-exercise (tl) and pre- PL SMT FORM PR fenoterol(t3) absolute values on the delta re- 1,I20-C sponses to both exercise and fenoterol. Sal- meterol and formoterol significantly lowered I15 - the serum potassium level compared with placebo before exercise (at tl). It is therefore 10 - possible that changes in the delta potassium response to exercise could have been due to case, http://thorax.bmj.com/ .-0 5 _ - functional antagonism. If this were the - - also be ~0 the values after exercise (t2) would expected to be lowered to the same degree. o However, this was clearly not the case, there -5 lbeing, in fact, no significant difference in serum potassium levels after exercise following treat- 10 PL SMT FORM PR ment with salmeterol, formoterol, or placebo. _ Furthermore, propranolol, which had no effect Figun e 2 12 mediated responses to fenoterol2-4mg. A on pre-exercise potassium levels, produced a on September 30, 2021 by guest. Protected copyright. compd arison of the effect oftreatment with single doses of marked increase in the response to exercise, placed bo (PL), salmeterol 300 pg (SMT), formoterol 72,ug (FO) RM), and propranolol 80mg (PR) on (A) delta thus suggestigthat the pre-exercise potassium senumz potassium, (B) delta finger tremor, and (C) delta concentration did not confound the delta potas- heart rate. Values are given as means and pooled SE. sium response to exercise. *p

1 Anderson GP. Formoterol: pharmacology, molecular basis than a full agonist such as isoprenaline.' For- of agonism and mechanism of long duration of a highly moterol was said to have an efficacy of 0-96 potent and selective 02 adrenoceptor agonist broncho- with a dilator. Life Sci 1993;52:2145-60. compared isoprenaline (given nominal 2 Waller DG. ,B-adrenoceptor partial agonists: a renaissance efficacy value of 1 00) in terms of their ability in cardiovascular therapy? Br J' Clin Pharmacol 1990;30: Thorax: first published as 10.1136/thx.51.1.54 on 1 January 1996. Downloaded from to maximally relax human bronchi, while sal- 157-71. 3 Grove A, McFarlane LC, Lipworth BJ. Expression of meterol has an efficacy value of 0-71. Similar the 02 adrenoceptor partial agonist/antagonist activity of studies have shown that formoterol does, in- salbutamol in states of low and high adrenergic tone. Thorax 1995;50:134-8. deed, exhibit greater intrisic activity than 4 Kallstrom BL, Sjoberg J, Waldeck B. The interaction be- salmeterol.'71 In vitro studies with guinea pig tween salmeterol and P2 adrenoceptor agonists with higher efficacy on guinea-pig trachea and human bronchus in airway have also shown that salmeterol can vitro. Br_JPharmacoll994;113:687-92. inhibit fenoterol and, to a lesser degree, sal- 5 O'Donnell SR, Wanstall JC. Evidence that the efficacy (intrinsic activity) of fenoterol is higher than that of salbutamol-induced relaxation.4 In the present butamol on ,B-adrenoceptors on guinea-pig trachea. EurJ7 study we were unable to distinguish between Pharmacol 1978;47:333-40. 6 Wilson C, Lincoln C. P-adrenoceptor subtypes in human, the relative partial agonist activities of for- rat, guinea-pig, and rabbit atria. J Cardiovasc Pharmacol moterol and salmeterol as they both produced 1984;6:1216-21. 7 Goldie RG, Spina D, Henry PJ, Lulich KM, Paterson JW. comparable degrees of 132 agonism at rest and In vitro responsiveness of human asthmatic bronchus to comparable 12 blockade with exercise and feno- carbachol, histamine, P-adrenoceptor agonists and theo- phylline. Br Jf Clin Pharmacol 1986;22:669-76. terol responses. However, it was clearly evident 8 Carruthers S, Shanks RG, McDevitt DG. Intrinsic heart that the P2 blockade of formoterol and sal- rate on exercise and the measurement of ,-adrenoceptor meterol was less than that of propranolol. It blockade. Br J' Clin Pharmnacol 1976;3: 991-9. 9 Lipworth BJ, McDevitt DG. ,B-adrenoceptor responses to may, however, be possible to make such a inhaled salbutamol in normal subjects. Eur J Clin Phar- distinction between salmeterol and formoterol macol 1989;36:239-45. 10 Williams ME, Gervino EV, Rossa RM, Landsbergh L, Young if dose ranging studies were performed using JB, Silva P, et al Catecholamine modulation of rapid lower doses of partial agonist. potassium shifts during exercise. N EnglJMed 1985; 312: 823-7. This study was confined to evaluating 11 Norgaard A, Botker HE, Klitgaard NA, Toft P. Digitalis extrapulmonary 12 responses in normal in- enhances exercise-induced hyperkalaemia. EurJ'Clin Phar- macol 1991;41:609-11. dividuals, and the clinical relevance in terms 12 Arnold JMO, O'Connor PC, Riddell JG, Harron DWG, of modulation of bronchodilator response Shanks RG, McDevitt DG. Effects of the P2 adrenoceptor antagonist ICI 118,551 on exercise tachycardia and iso- therefore remains unclear. Clearly salmeterol, prenaline induced ,-adrenoceptor responses in man. BrJ a drug with relatively low efficacy, normally Clin Pharmacol 1985;19:619-30. 13 Newnham DM, Wheeldon NM, Lipworth BJ, McDevitt produces clinically significant degrees of bron- DG. Single dosing comparison of the relative cardiac 1,/ chodilation. It is, however, worth noting that, 32 activity of inhaled fenoterol and salbutamol in normal subjects. Thorax 1993;48:656-8. in vitro, differences in intrinsic activity between 14 Lipworth BJ, Newnham DM, Clark RA, Dhillon DP, Winter agonists in terms of bronchodilator activity be- JH, McDevitt DG. Comparison of the relative airways and systemic potencies of fenoterol and salbutamol in come more pronounced when bronchial tone asthmatic subjects. Thorax 1995;50:54-61. is increased' as might occur in the setting of 15 Iipworth BJ, Brown RA, McDevitt DG. Assessment of http://thorax.bmj.com/ acute asthma. This may be compounded by airways, tremor and chronotropic responses to inhaled salbutamol in the quantification of beta-2 adrenoceptor the presence of raised endogenous adrenergic blockade. Br J Clin Pharmacol 1989;28:95-102. tone which might also result in the expression 16 Tattersfield AE. Clinical pharmacology of long-acting ,- receptor agonists. Life Sci 1993;52:2161-9. of antagonist effects by partial agonists such as 17 Iinden A, Bergendal A, Ullman A, Skoogh BE, Lofdahl salmeterol. Furthermore, it is possible that a CG. Salmeterol, formoterol, and salbutamol in the isolated guinea-pig trachea: differences in maximum relaxant effect P2 agonist with lower intrinsic activity such as and potency but not in functional antagonism. Thorax salmeterol might result in a reduced broncho- 1993;48:547-53. 18 Advenier C, Zhang Y, Naline E, Grandordy BM. Effects of dilator response to an exogenous agonist with formoterol on isolated human bronchus. Am Rev Respir on September 30, 2021 by guest. Protected copyright. higher intrinsic activity such as salbutamol. Dis 1991;143:A651