The affinity and selectivity of α‐adrenoceptor antagonists, antidepressants, and antipsychotics for the human α1A, α1B, and α1D‐adrenoceptors

Abstract α1‐adrenoceptor antagonists are widely used for hypertension (eg, doxazosin) and benign prostatic hypertrophy (BPH, eg, tamsulosin). Some antidepressants and antipsychotics have been reported to have α1 affinity. This study examined 101 clinical drugs and laboratory compounds to build a comprehensive understanding of α1‐adrenoceptor subtype affinity and selectivity. [3H]prazosin whole‐cell binding was conducted in CHO cells stably expressing either the full‐length human α1A, α1B, or α1D‐adrenoceptor. As expected, doxazosin was a high‐affinity nonselective α1‐antagonist although other compounds (eg, cyclazosin, 3‐MPPI, and ARC239) had higher affinities. Several highly α1A‐selective antagonists were confirmed (SNAP5089 had over 1700‐fold α1A selectivity). Despite all compounds demonstrating α1 affinity, only BMY7378 had α1D selectivity and no α1B‐selective compounds were identified. Phenoxybenzamine (used in pheochromocytoma) and dibenamine had two‐component‐binding inhibition curves at all three receptors. Incubation with sodium thiosulfate abolished the high‐affinity component suggesting this part is receptor mediated. Drugs used for hypertension and BPH had very similar α1A/α1B/α1D‐adrenoceptor pharmacological profiles. Selective serotonin reuptake inhibitors (antidepressants) had poor α1‐adrenoceptor affinity. Several tricyclic antidepressants (eg, amitriptyline) and antipsychotics (eg, chlorpromazine and risperidone) had high α1‐adrenoceptor affinities, similar to, or higher than, α blockers prescribed for hypertension and BPH, whereas others had poor α1 affinity (eg, protriptyline, sulpiride, amisulpiride, and olanzapine). The addition of α blockers for the management of hypertension or BPH in people already taking tricyclic antidepressants and certain antipsychotics may not be beneficial. Awareness of the α‐blocking potential of different antipsychotics may affect the choice of drug for those with delirium where additional hypotension (eg, in sepsis) may be detrimental.


| INTRODUC TI ON
The α1-adrenoceptors are expressed in a wide range of tissues including blood vessels, kidney, spleen, liver, brain, and lower urinary tract. [1][2][3] There are three subtypes: α1A, α1B, and α1D-adrenoceptors. [1][2][3][4] All are present in blood vessels, and whilst α1A and α1D and are both important in smooth muscle contraction (and control of blood pressure), the role of the α1B-adrenoceptors is less certain. 2,3,5,6 α-adrenoceptor antagonists (α blockers) were first used to reduce systemic blood pressure with dibenamine, phentolamine, and phenoxybenzamine used in the diagnosis and management of pheochromocytoma, an adrenal catecholamine-secreting tumor. 7,8 While phenoxybenzamine is still important for pheochromocytoma, longer acting, nonselective α1-antagonists were developed (doxazosin, terazosin, indoramin, and prazosin) and remain important in the management of resistant hypertension.
α blockers are also used in benign prostatic hypertrophy (BPH) where α1A blockade induces prostate and lower urinary tract smooth muscle relaxation, improving urinary flow. 9 Phenoxybenzamine was the first α blocker to be used in BPH 10 although its α2 effects limited its use. 11 The nonselective α1-antagonists doxazosin, terazosin, indoramin, and prazosin were used effectively for BPH, but caused hypotension, particularly postural hypotension, and required dose titration to manage this problematic side effect. 9,12 Selective α1A-antagonists were developed, hoping to minimize hypotension by reducing α1B-antagonsim. 11,13 Tamsulosin, alfuzosin, and silodosin were developed as prostate-specific (α1A selective) drugs and are used without dose titration. 9 Despite reports of "better tolerability," 11,14 alfuzosin is reported to be a nonselective α1-antagonist and tamsulosin to have equal α1A-and α1D-adrenoceptor affinity, 15,16 suggesting they may be pharmacologically indistinguishable from drugs used for hypertension. Indeed, tamsulosin (the most commonly prescribed α blocker for BPH) is associated with increased hypotension, falls, and fractures. 12,13,17 Although effective for BPH, silodosin appears to have more sexual side effects, whereas its cardiovascular effects remain uncertain. 18 α1-adrenoceptors are the most abundant adrenoceptors in the brain and modulate neurotransmitter release. 3 Many antidepressants prevent the reuptake of neurotransmitters (serotonin and noradrenaline), and therefore increase synaptic neurotransmitter concentration. However, several antidepressants have significant α1-adrenoceptor affinity. [19][20][21] This high affinity is seen in brain homogenates. 22 In theory these two effects (increased neurotransmitter presence, but receptor blockade) could cancel each other out. 20 However, antidepressants cause hypotension, particularly postural hypotension (up to 58% users 23,24 ). Not surprisingly therefore, antidepressant use is associated with twice the risk of falls. 25 Several antipsychotics (neuroleptics) bind to α1-adrenoceptors in blood vessels and brain homogenates. 6,26,27 Many antipsychotics cause postural hypotension, 28,29 and again, rates are high (eg, 48% taking risperidone 24 ) including postural hypotension in those taking long-term antipsychotics (77% 30 ). Interestingly, the degree of postural hypotension seen with several antipsychotics correlates well with the α1A-adrenoceptor affinity. 29 Antipsychotic drug use is also associated with falls and hip fractures and regular use is associated with twice the risk of falls (even after controlling for other risks 31,32 ).
There are many studies examining the affinity of α1-adrenoceptor ligands. Many are older studies before the identification of the three subtypes and many are in whole tissue where multiple subtypes will be present. Most studies only report the two or three ligands under investigation. Here we aimed to investigate the subtype selectivity of a wide range of α-antagonists including those used in hypertension, BPH, antidepressants, antipsychotics as well as laboratory compounds. Human α1A, α1B, and α1D-adrenoceptors were expressed in intact mammalian cells, in order to build a comprehensive and directly comparable picture of α1-subtype selectivity in living cells.

| Materials
A list of all of the compounds studied, together with the source and supplier code from which it was purchased, is given in Table S1.

| Cell lines
CHO-K1 (RIDD: CVCL_0214) were stably transfected with the DNA of the human α1A-adrenoceptor, human α1B-adrenoceptor (DNAs from Guthrie DNA Resource Centre), or human α1D-adrenoceptor (full-length DNA from Andre Pupo 33 ; using Lipofectaime and Optimem according to the manufacturers' instructions. Transfected cells were selected for 3 weeks using resistance to neomycin (at 1mg/ml). Single clones from each transfection were then isolated by dilution cloning giving rise to the stable cell lines CHO-α1A, CHO-α1B, and CHO-α1D.

| Cell culture
CHO cells were grown in Dulbecco's modified Eagle's medium nutrient mix F12 (DMEM/F12) containing 10% fetal calf serum and 2 mmol/L L-glutamine in a 37°C humidified 5% CO 2 : 95% air atmosphere. Cells were seeded into white-sided, clear bottomed 96well view plates and grown to confluence.

| [3H]prazosin binding-saturation binding
The K D value for [3H]prazosin was determined in each cell line by saturation binding. [3H]prazosin was diluted in serum-free media.
[3H]prazosin was then added to the wells (quadruplicates per condition, 1 in 2 dilution in well), and the plates incubated for 2 hour at 37°C in a humidified 5% CO 2 : 95% air atmosphere.
After 2 hours, the cells were washed twice by the addition and removal of 2 × 200 µL cold (4°C) phosphate-buffered saline. 100 µL Microscint 20 was added to each well and a white base applied to the plate to convert the wells into white-sided/white-bottomed wells.
Plates were left at room temperature for at least 6 hours before being counted on a Topcount (PerkinElmer), with a counting time of 2 minutes per well.

| [3H]prazosin whole-cell bindingcompetition binding
Ligands were serially diluted in serum-free media (DMEM/F12 containing 2 mmol/L L-glutamine only) to twice their final required concentration. Media was removed from each well of the 96-well view plate and 100 µL ligand added to triplicate wells. This was immediately followed by the addition of 100 µL [3H]prazosin (diluted in serum-free media) and the cells incubated for 2 hours at 37°C (5% CO 2 , humidified atmosphere). After 2 hours the plates were washed as above. Cells were inspected under a light microscope to ensure cells were still present after the wash and before the addition of Microscint 20. In a few cases, high concentrations of competing ligand caused the cells to round up and be washed off the plates. These concentrations were excluded from the analysis. Total binding (6 wells/plate) and nonspecific binding (6 wells/ plate) determined by the presence of 10 µmol/L tamsulosin (α1A and α1B) or 100 µmol/L tamsulosin (α1D) was defined in every plate.
Sodium thiosulfate reacts with 2-chloroethylamines in a 1:1 stoichiometry to inactivate the ethyleniminium ions generated in solution (see Discussion). Sodium thiosulfate had no effect on [3H] prazosin binding up to concentrations of 10 mmol/L. Therefore, to ensure that all ethyleniminium ions were inactivated, sodium thiosulfate was used in excess, with a final well concentration of 1 mmol/L. When used, competing ligands were serially diluted in serum-free media (just as above) in the absence and presence of thiosulfate and both dilution series were then incubated for 30 minutes at 37°C (5% should not make much difference for antagonists, the receptors (and therefore measurements taken) in this living system are therefore more akin to how drugs bind in people, than studies conducted in membrane preparations.

| Data analysis
In all cases where a K D value is stated, increasing concentrations of the competing ligand fully inhibited the specific binding of [3H]prazosin (unless otherwise annotated in the tables).
The following equation was then fitted to the data using Graphpad Prism 7 and the IC 50 was then determined as the concentration required to inhibit 50% of the specific binding. In some cases, the maximum concentration of competing ligand was not able to inhibit all of the specific binding. Where no inhibition of [3H]prazosin binding was seen, even with maximum concentration of competing ligand possible, "no binding" is given in the tables. Where the inhibition produced by the maximum concentration of the competing ligand was 50% or less, an IC 50 could not be determined and thus a K D value not calculated. This is shown in the tables as IC 50 > top concentration used (ie, IC 50 > 100 µmol/L means that 100 µmol/L inhibited some but less than 50% of the specific binding). In cases where the competing ligand caused a substantial (greater than 60%, but not 100%) inhibition of specific binding, an IC 50 value was determined by extrapolating the curve to nonspecific levels and assuming that a Selectivities are given as a ratio of the K D values for the different receptors.  The maximum concentration of competing ligand inhibited most but not all of specific binding (as in Figure 1E). An respectively, Figure 1, Table 1). Of all the compounds studied, SNAP 5089 had the highest receptor selectivity, being over 1700-fold selective for the α1A-adrenoceptor ( Figure 1,   Table 1). Several compounds had affinities of less than 0.25nM, including ligands with α1A selectivity (silodosin, RS100329, and tamsulosin), cyclazosin with slight α1D selectivity), and nonselective 3-MPPI (Table 1).

| RE SULTS
Two compounds were best described by a two-component-bind- ing inhibition curve at all three receptors-phenoxybenzamine and dibenamine ( Figure 2, Table 1). Both of these are N,N-disubstituted-

2-chloroethylamines. Preincubation of phenoxybenzamine and
dibenamine with sodium thiosulfate before addition to the cells yielded a single-component-binding inhibition (Figure 2, Table 2), whereby the high-affinity-binding component of the parent curve had been abolished. Sodium thiosulfate had no effect on the binding of tamsulosin (Figure 2, Table 2). At the α1D-adrenoceptor, several   The affinity of several antidepressants and antipsychotics was then examined. Several of these were found to have high α1-adrenoceptor affinity (Figures 3 and 4, Table 3 and 4).

F I G U R E 3 Inhibition of [3H]prazosin binding to whole cells by two commonly prescribed antidepressants amtriptyline (
Risperidone (previously suggested to have α1B selectivity, 4,36 had slight α1A selectivity, in keeping with the findings of 37 . There have also been discrepancies in the affinity of olanzapine: Richelson and Souder 27 found it to have high affinity (44nM for α1-adrenoceptor) and Nourain et al, 29 had conflicting data with low rat α1-adreoceptor affinity, but significant hypotension in rats. However, here, olanzepine had low affinity, in keeping with 38 and the findings of 39 where olanzapine was described as having low postural hypotension potential. WB4104, was also initially thought to have α1B selectivity, 40 however, it had higher and equal affinity for α1A and α1D-adrenoceptors (in keeping with 41,42 ).
Tables combing all of these ligands are presented in Supplementary Data. Table S1 has the ligands arranged in alphabetical order (together with their suppliers and individual ligand codes). Table S2 has the ligands organized in order of α1A affinity.  Table 1). In aqueous solution at physiological pH, the nitrogen mustard group cyclizes to form ethyleniminium ions. 44 These highly reactive, unstable ions are pharmacologically active and covalently bind to a cysteine in transmembrane 3 of the α adrenoceptors, giving these compounds their "irreversible" properties. 43 Phenoxybenzamine has a longer duration of action in clinical studies than phentolamine 7 and hence its continued use in pheochromocytoma (although similar outcomes have been reported with doxazosin, terazosin, and prazosin,. [45][46][47] Sodium thiosulfate also rapidly reacts with the ethyleniminium ions thus prevents them from interacting with α adrenoceptors. 44 Pretreatment with intravenous sodium thiosulfate prevented dibenamine binding to α adrenoceptors (in cats, 48 and pretreatment with sodium thiosulfate prevented the The maximum concentration of competing ligand inhibited most but not all of specific binding (as in Figure 1E). An IC 50 was determined by extrapolating the curve assuming that all specific binding would be inhibited if a higher concentration of competing ligand were possible. Thus an app K D was calculated.

| D ISCUSS I ON
app, apparent.  The maximum concentration of competing ligand inhibited most but not all of specific binding (as in Figure 1E). An IC 50 was determined by extrapolating the curve assuming that all specific binding would be inhibited if a higher concentration of competing ligand were possible. Thus an app K D was calculated.

TA B L E 4 Log K
app, apparent.
harmful interactions of the chemical weapon mustard gas in humans.
Here, preincubation of phenoxybenzamine or dibenamine with sodium thiosulfate yielded single-component-binding inhibition curves ( Figure 2, Table 2). Abolishment of the high-affinity-binding component suggests that it was due to specific α1-adrenoceptor interaction. The K D values of the low-affinity components were very similar to those obtained in the presence of thiosulfate, suggesting that this component is a non-orthosteric site or non-receptor-mediated effect.
Several other ligands were found to have a [3H]prazosin inhibition best described a two-component curve at the α1D-adrenoceptor, including tamsulosin (and hence why 100 μM was used to define nonspecific binding in CHO-α1D cells, rather than 10 μM used in α1A and α1B cells), and the only α1D-selective ligand, BMY7378. As expected for these nonmustard compounds, preincubation with sodium thiosulfate had no effect on binding. The reason for the second component is therefore unknown. Affinity (K D value) obtained for the high-affinity component the α1D-adrenoceptor has been used to determine receptor selectivity. was still 10-fold less than that for the β2 adrenoceptor. 52 Labetolol (used in hypertension particularly in pregnancy, and intravenously in hypertensive emergencies), is also considered a dual α/β blocker.
Labetolol has lower affinity than carvedilol for all β 52 and α1-adrenoceptors (log K D −7.33 at α1A), but very poor affinity for the α1B and α1D-adrenoceptors. Labetolol should be considered a β1/β2/α1A blocker rather than dual pan α1 and β blocker. Given these dual α/β ligands, the affinity of a few β blockers with very high β-adrenoceptor affinity were examined (Table 1). With the exception of bucindolol, the affinity was poor at all three α1-adrenoceptors, confirming their β selectivity. Although the affinity of bucindolol was reasonably high (log K D at α1A −7.57), this is 54-fold and 263-fold lower than that for the human β1 and β2 adrenoceptor, respectively. 53 The most selective ligand detected here was SNAP5089, with 1700-fold selectivity for the α1A over the α1B or α1D-adrenoceptors. Other α1A-selective ligands were silodosin, RS100329, and niguldipine (in keeping with 50,54 ). As well as tamsulosin, several ligands had higher affinity for the α1A and α1D receptors than the α1B-for example, 2-MPMDQ, MK-912, 2-PMDQ, and ifenprodil. BMY7378 was the only compound with substantial α1D selectivity (the 100-to 200-fold selectivity is similar to 6,20,41,50,55 ). Several antipsychotics (including first-generation chlorpromazine and flupenthixol and second-generation sertindole, risperidone, and clozapine) had high α1-adrenoceptor affinity. The very high affinity of sertindole (and 300-fold selectivity for α1A over α1D-adrenoceptors) was similar to previous reports. 6,29 The degree of α1A-adrenoceptor affinity observed here correlates well with the rankings for observations in rats. 29 The high α1A affinity of sertindole, risperidone, and ziprasidone (log K D −9.3 to −8.7) is similar to studies, 38 including in brain tissue, 27 and similar to (or even higher than) that for many drugs used to treat hypertension. The high rate of postural hypotension observed with these drugs 31,32 is therefore not surprising. A similar hypotensive effect would be expected with other antipsychotics α1 affinities equal or greater than that for α1 blockers used for hypertension, for example, chlorpromazine, flupenthixol, perphenazine, paliperidone, quetiapine, and lurasidone.
Aripiprazole had lower α1 affinity (in keeping with 56 and indeed has a relative lack of reported postural hypotension in clinical studies 57 ).
However, sulpiride and amisulpiride would be expected to have even less hypotensive effect.
Thus, the high α1 affinity and selectivity profile of many antipsychotics is comparable to the α1 blockers intentionally prescribed for hypertension. Equivalent reductions in blood pressure are a likely very common side effect. These drugs are used to manage schizophrenia where their effect on blood pressure in agitated patients is less likely to be an issue. However, antipsychotics are also widely used to manage delirium in sick patients including the intensive care unit 58,59 and in palliative case 60,61 even though recent studies have questioned their effectiveness. 62 Delirium is common in older unwell patients who are more likely to be suffering from conditions with lower blood pressure such as sepsis. In these cases, the choice of antipsychotic may well be important in order not to exacerbate already low or labile blood pressures. This study suggests that sulpiride, amisulpiride, ariprazole, and olazepine should have the least effect on blood pressure.
The drugs used for hypertension and BPH have a very similar pharmacological profile in terms of α1-adrenoceptor subtype affinity and selectivity. Several antidepressants and antipsychotics have high α1-adrenoceptor affinities, similar to, or even greater than, those seen for α blockers prescribed for hypertension and BPH. The addition of further α blockers for the management of hypertension or BPH in these patients may not be beneficial.
The excellent correlation between the affinity values determined from this cell studies with the affinities measured in blood vessels, brain tissue, and whole animals (including humans) means that many, but not all antipsychotics and antidepressant may cause significant peripheral α-adrenoceptor blockade and associated hypotension. Finally, awareness of the α-blocking potential of certain, but not all antipsychotics may affect the choice of drug used for the management of delirium in the intensive care unit where additional α blockade and blood pressure lowering in a sick patient may be detrimental.

ACK N OWLED G EM ENTS
We

D I SCLOS U R E
JGB has been on the Scientific Advisory Board for CuraSen Therapeutics since 2019.

AUTH O R S CO NTR I B UTI O N S
JGB designed the research study. ASP contributed the α1D DNA and aided discussions. RGWP and JGB performed the research. JGB analyzed the data. JGB and ASP wrote the paper.

E TH I C A L S TATEM ENT
No animals, human tissue, human volunteers, or patients were used in this study.

DATA S H A R I N G
Further information and requests for data and reagents should be directed to and will be fulfilled by the corresponding author, Jillian Baker. Please contact jillian.baker@nottingham.ac.uk.

O PE N R E S E A RCH BA D G E S
This article has earned an Open Data badge for making publicly available the digitally-shareable data necessary to reproduce the reported results.