FormalPara Key Points

Brimonidine and oxymetazoline are both topical α-adrenergic agonists that have been developed to treat the persistent erythema of rosacea patients.

Pharmacological profiling showed brimonidine to be highly selective for the α2 adrenergic receptors, whereas oxymetazoline was much less selective with a significant number of off-target activities, including the 5-hydroxytryptamine 2B (5-HT2B) serotonin receptor.

1 Introduction

Rosacea is a chronic inflammatory skin disease characterized by transient or persistent central facial erythema and a variety of other signs and symptoms [1, 2]. Topical α-adrenergic agonist therapy has been developed to treat the persistent erythema of rosacea patients. Oxymetazoline is a direct-acting sympathomimetic amine, an α1 and α2 adrenergic receptor agonist used as a nasal decongestant [3, 4]. Brimonidine tartrate is a highly selective α2 adrenergic receptor agonist with potent vasoconstrictive activity used as an ophthalmic solution for the initial and long-term management of ocular hypertension and glaucoma [5]. Brimonidine and oxymetazoline are both now approved as topical α-adrenergic agonists approved for treating erythema of rosacea.

The selectivity (target-based assays and profiling), vasoconstrictive (wire myography assays) and anti-inflammatory (ex vivo human skin model and in vivo inflammation model) properties of brimonidine have previously been demonstrated [6].

The objective of this in vitro safety pharmacology study was to compare the potential safety profiles of brimonidine and oxymetazoline during the drug development process before obtaining long-term safety data.

2 Materials and Methods

This in vitro pharmacology research did not involve any human subjects and so did not require IRB review.

2.1 Receptor Profiling

Pharmacological profiling was performed with a standard panel (BioPrint® Full Profile, Cerep, Celle l’Evescault, France) of 151 assays, including α-adrenergic receptors and 5-hydroxytryptamine (5-HT) receptors [6]. Both compounds were dissolved in dimethyl sulfoxide (DMSO) then diluted in HEPES sodium salt buffer and tested. The amount of a particular substance needed to inhibit by 50% (IC50), or to stimulate by 50% (half maximal effective concentration [EC50]), the assay signal was computed from a dose–response assay. A compound was considered active against the target if a 10-µM concentration produced > 30% effect (stimulation or inhibition) on the assay signal.

2.2 5-Hydroxytryptamine 2B (5-HT2B) Agonist Assay

HEK-293 cells stably expressing human recombinant 5-hydroxytryptamine 2B (5-HT2B) receptors were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and 250 μg·mL−1 G-418. CHO-K1 cells stably expressing the human 5-HT2A, 5-HT2B or 5-HT2C receptors were maintained in Ham’s F-12 supplemented with 10% newborn calf serum and 300 μg·mL−1 G-418.

HEK-293 cells expressing human 5-HT2B receptors were seeded in 98-well poly-d-lysine coated plates at a density of 50,000 cells·well−1 in DMEM containing 10% (v.v−1) dialyzed fetal bovine serum and 250 μg·mL−1 G-418. After overnight incubation, the cells were washed three times with Hank’s buffered saline solution containing 2 mM CaCl2, 10 mM HEPES and 2.5 mM probenecid. The cells were incubated with 2 μM Fluo-3 AM in Hank’s buffered saline solution for 60 min at 37 °C. The Fluo-3 loaded cells were then washed four times with Hank’s buffered saline solution. Agonist-induced changes in intracellular calcium concentrations were determined at room temperature using a Fluorometric Imaging Plate Reader (Molecular Devices Corporation, Sunnyvale, CA, USA).

2.3 Valvular Interstitial Cell Proliferation Assay with Oxymetazoline Hydrochloride

A valvular interstitial cell (VIC) proliferation assay was performed, as described by Setola et al. [7]. VICs were isolated from apparently non-diseased human aortic and mitral valves, obtained from one donor heart deemed unsuitable for transplantation and obtained with appropriate ethical approval and informed consent for research use (Asterand, Royston, Herts, UK). Immediately upon arrival at Asterand, leaflets from the apparently healthy aortic and mitral valves were dissected free of the surrounding heart tissue then digested in a solution of collagenase II in serum-free medium for 20 min. After the 20-min incubation, the surface of the tissue was rubbed with a sterile cotton swab to remove endothelial cells. The valve tissue was then finely minced and digested in a solution of collagenase III for 4 h. The resulting cell suspension was then filtered, washed and then resuspended in DMEM/Ham’s F12 medium supplemented with 10% fetal bovine serum and 1% antibiotic-antimycotic solution.

Isolated cells were seeded into T25 flasks and after 4 days the cells were trypsinized and transferred to T75 flasks. When the cells reached approximately 80% confluency they were trypsinized and seeded into 48-well plates for the proliferation assay.

Cells were incubated in medium containing reduced serum overnight and then exposed to oxymetazoline hydrochloride or DMSO (0.1%) for 48 h. One assay was performed corresponding to a single donor and all treatment conditions were in triplicate wells.

Proliferation assay cells were pulsed with [3H] thymidine (2 µCi·mL−1) and assayed for tritium using liquid scintillation counting.

3 Results

3.1 Receptor Profiling

Brimonidine was highly selective for the α2 adrenergic receptors, specifically α2A. Conversely, oxymetazoline was highly active against a wide range of targets, including the α2A, α2B, α2C, α1A, and α1B adrenergic receptors, as well as the 5-HT1A, 5-HT1B, 5-HT1D, 5-HT2A, 5-HT2B, 5-HT2C and 5-HT7 serotonin receptors (Fig. 1).

Fig. 1
figure 1

Receptor profiling of brimonidine and oxymetazoline at 10 µM concentration for human α-adrenergic receptors and other ligand binding sites. All active targets (> 30% effect) for brimonidine and/or oxymetazoline are shown and the brimonidine response was normalized on the maximum signal detected in each assay (assay was carried out in duplicate). 5-HT 5-hydroxytryptamine, D dopamine, CYP2D6 cytochrome P450 2D6, M muscarinic, NAC Na+ channel (site2), H histamine

Full size image

3.2 5-HT2B Agonist Assay

Negligible 5-HT2B agonist activity was observed with brimonidine (Fig. 2). Conversely, oxymetazoline demonstrated a potent 5-HT2B agonist effect as the concentration needed to elicit half the maximum biological response (pEC50: − log EC50) was 7.82, corresponding to an EC50 of 15 nM (Fig. 2).

Fig. 2
figure 2

Hydrochloride 5-hydroxytryptamine 2B (5-HT2B) agonist-induced changes in intracellular calcium concentrations, measured by fluorescence and expressed as a mean percentage of the control agonist response (assay was carried out in duplicate). EC50 half maximal effective concentration, pEC50  – log EC50

Full size image

3.3 Valvular Interstitial Cell Proliferation Assay

Oxymetazoline caused proliferation of passage 2 mitral VICs in vitro; oxymetazoline (10 µM) clearly increased [3H] thymidine incorporation and oxymetazoline (100 µM) approximately doubled the response (Fig. 3).

Fig. 3
figure 3

Effect of oxymetazoline hydrochloride on the proliferation response of valvular interstitial cells (assay in triplicate from a single donor). DMSO dimethyl sulfoxide

Full size image

4 Discussion

Although brimonidine and oxymetazoline are both vasoconstrictors, they have very different receptor selectivity. Pharmacological profiling showed brimonidine to be highly selective for the α2 adrenergic receptors, specifically α2A, whereas oxymetazoline was much less selective with a significant number of off-target activities, including the 5-HT2B serotonin receptor. Since 5-HT2B agonism has been associated with drug-induced valvulopathy, human recombinant 5-HT2B receptor agonist-induced changes in intracellular calcium concentrations were determined using fluorometric imaging. Negligible 5-HT2B agonist activity was observed with brimonidine. Conversely, oxymetazoline demonstrated a potent 5-HT2B agonist effect and the EC50 of 15 nM corroborates previously published results [8]. Calcium-based high-throughput screening of 2200 drugs and drug-like compounds found only 27 were 5-HT2B receptor agonists and the authors recommended that five currently prescribed drugs, including oxymetazoline, should be studied further in terms of their pharmacodynamics and potential for valvulopathy [8].

An ‘off-target’ activation of the 5-HT2B receptor expressed on human heart interstitial cells is a possible mechanism for the pathogenesis of drug-induced valvulopathy [9]. In a VIC proliferation assay, oxymetazoline was shown to cause proliferation of passage 2 mitral VICs in vitro; the VIC proliferation assay was not performed with brimonidine as it is inactive on the 5-HT2B receptor. The use of valvular tissue from a single patient is a limitation of the present study. However, it has been suggested that candidate drug molecules should be screened at 5-HT2B receptors and potent agonists should be avoided, especially for treating non-life-threatening disease and for long-term use since duration of therapy is an important determinant of valvulopathic risk [8, 10]. Rosacea is a chronic disease so long-term drug use may be desirable. It remains unclear what concentrations are safe and admittedly systemic concentrations from topical treatment may be very low since mean maximum observed systemic concentrations reported for oxymetazoline cream were subnanomolar (0.4 nM) [11]. In that study, patients applied oxymetazoline cream (0.5, 1.0, or 1.5%) to the facial area once daily or twice daily for 28 days, and the authors concluded that systemic exposure was minimal [11]. However, even though there may be minimal immediate risk of adverse effects following topical application to the facial area, further studies and post-marketing data are required to determine the relevance of any potential systemic absorption during long-term chronic use.

In conclusion, on the basis of our findings and previous publications, it appears prudent to carefully analyze the benefit/risk ratio when using oxymetazoline in the chronic management of rosacea, especially in patients with cardiovascular disease or other comorbidities.