Introduction

Alpha1 adrenoreceptor (α1AR) comprises a family of the G-protein-coupled receptors that was physiologically activated by the catecholamines, epinephrine, and norepinephrine [1]. Three cDNAs encoding α1AR subtypes (α1A, α1B, and α1D) have been cloned from rat, human, and other sources, expressed in cell lines, and indentified pharmacologically in native tissues [2, 3]. Benign prostate hyperplasia (BPH) is a common disease and impacts health-related quality of life profoundly. It has been reported that the α1AAR is predominant in the human prostate [4, 5]. Tamsulosin, terazosin, alfuzosin, doxazosin, and 5-methylurapidil, namely α1AAR antagonists, have a greater selectivity in prostate smooth muscle [6], and it has been used in the treatment of many clinical BPH symptoms, such as micturition, nocturia increase, dysuresia, etc. [7, 8]. In order to fully characterize the mechanism of the drugs and screen for long-acting a1AAR antagonists, investigation of drug–receptor interactions becomes increasingly important. The interactions are usually examined by means of radioligand binding assay (RBA) [9]. Although it has many advantages such as high sensitivity and accuracy, it can result in potential radioactive pollution, and it failed to exactly simulate the fundamental processes of drug action dynamically.

Recently, receptor affinity chromatography, carried out through frontal analysis or nonlinear studies, has been a new technique for characterizing binding sites, drug discovery, as well as drug development [10–12]. Based on this thought, the cell membrane chromatography (CMC) method has been proposed and applied to investigate the affinity interaction between drug and membrane receptor [13–15], and fast screen the α1AAR antagonists from complex samples [16, 17]. We have established a CMC method using frontal analysis to investigate the K D values for calcium antagonist–L-type calcium channel interactions [18]. However, the influence of detection wavelengths and the estimation of CMC column dead volume might make the method inaccurate. Moreover, the CMC method should be developed to differentiate specific binding from nonspecific binding because only specific binding can activate the signal transduction pathway.

In this study, HEK293 α1A cell line, which expresses stably high levels of α1AAR, was used to prepare the cell membrane stationary phase (CMSP) in the CMC model. An HEK293 α1A/CMC system was established for determining the affinity of the α1AAR antagonists binding to the α1AAR. In an effort to obtain further insight into the binding sites of ligands to the α1AAR on the CMSP, a competitive binding study was modified and developed. To ensure that the location of the binding region which tamsulosin occupied varied from that of oxymetazoline, a series of two breakthrough curves were constructed. The specificity, selectivity, sensitivity, reproducibility, and robustness of the HEK293 α1A/CMC method were also investigated.

Experimental

Materials and reagents

Tamsulosin, terazosin, alfuzosin, doxazosin, 5-methylurapidil, phentolamine, oxymetazoline, nitrendipine, and nilotinib as standard drugs were provided by the National Institute for the Control of Pharmaceuticals and Biological Products (Beijing, China). The HEK293 a1A high expression cell line was a gift from Prof. Youyi Zhang at Peking University Third Hospital (Beijing, China). Dulbecco minimal essential medium (DMEM) and G418 were purchased from Invitrogen Corporation (Grand Island, NE, USA). Methanol was of LC grade (Honeywell, NJ, USA). Silica gel (ZEX-II, 200–300 mesh) was obtained from Qingdao Meigao Chemical (Qingdao, China), and C18 solid-phase extraction columns were purchased from Supelco (500 mg, 3 mL−1, Sigma-Aldrich, Bellefonte, PA, USA). All other reagents and solvents were of analytical reagent grade and used without further purification unless otherwise noted. The phosphate-buffered saline (PBS, 50 mM) was prepared by dissolving disodium hydrogen phosphate (8.954 g) in the newly double-distilled water (500 mL) and diluted into various concentrations before using (adjust pH 7.4 by phosphonic acid).

Apparatus and conditions

CMC analysis was performed on a Shimadzu LC-20A apparatus that consisted of two LC-20AD pumps, a DGU-20A3 degasser, an SIL-20A auto sampler, a CTO-20A column oven, and an SPD-M20A diode array detector (Shimadzu, Kyoto, Japan). The data were acquired by the LCsolution software (Shimadzu, Kyoto, Japan) and processed by GraphPad Prism version 5.0 (San Diego, CA, USA).

The CMC mobile phase consisted of double-distilled water or PBS was delivered at a flow rate of 0.2 mL min−1. The HPLC conditions were a Dikma C18 column (150 × 4.6 mm, 5 μm), a mobile phase of methanol–water–triethylamine (70:29.8:0.2, v/v/v, adjust pH to 6.0 by acetic acid), with 1.0 mL min−1 flow rate. The detection wavelengths for the ligands were 280 nm for tamsulosin, 246 nm for terazosin, 246 nm for alfuzosin, 248 nm for doxazosin, 283 nm for 5-methylurapidil, 277 nm for phentolamine, and 280 nm for phentolamine.

Preparation of sample solutions

The stock solutions (5 × 10−3 mol L−1) of tamsulosin, terazosin, alfuzosin, doxazosin, 5-methylurapidil, phentolamine, oxymetazoline, nitrendipine, and nilotinib were prepared by separately dissolving the standard drugs in methanol. Standard solutions at various concentrations were prepared by diluting the stock solutions with the mobile phase for each one.

Cell culture and preparation of CMSP

The pREP8/bovine a1AAR plasmid was transfected into HEK293 cell lines by the calcium phosphate precipitation method [19, 20]. HEK293 cells stably expressing α1AAR were cultured in DMEM (contained 10% fetal bovine serum, 100 U mL−1 penicillin, 100 U mL−1 streptomycin, and 0.1 mg mL−1 G418). Cells were grown at 37 °C in a humidified atmosphere with 5% CO2, and cells from exponentially growing cultures (107) were harvested using trypsin and incubated for 10 min at 4 °C. Cultured HEK293 α1AAR cells were washed three times with 5 mM PBS by centrifuging at 3,000×g for 10 min, and the pellet was re-suspended with 50 mM Tris–HCl (pH 7.4), followed by ultrasonic destruction for 30 min. The resulting homogenate was clarified by centrifugation at 1,000×g for 10 min, and the supernatant was centrifuged at 12,000×g for 10 min. The precipitate was then suspended with 5 mM PBS [16, 17]. Briefly, the HEK293 α1AAR CMSP was prepared by adsorption of the cell membrane suspension (5 mL) on the activated silica (0.05 g) under vacuum and with a gentle agitation. The CMSP was placed overnight and then washed with 5 mM PBS five times. Finally, the mixture obtained was packed into a column (10 × 2.0 mm I.D.) using a wet packing method (10 MPa, 5 min). All the procedure was performed at 4 °C. The HEK293 CMSP was also obtained by the same procedure.

The total protein was determined using BCA protein assay reagent kit (Jiankangyuan Biotech, Zhuhai, China), and the content of the α1AAR in the HEK293 and HEK293 α1AAR CMSP was determined using human α1AAR kit (Cusabio Biotech, Wuhan, China) by ELISA [21]. Experiments were performed in triplicate. The absorbance was quantitatively measured at 450 nm. Likewise, the human epidermal growth factor receptor (EGFR) and vascular endothelial growth factor receptor-2 (VEGFR-2) were also evaluated by the commercially available EGFR and VEGFR-2-ELISA kit [22, 23].

HEK293 α1A/CMC-offline-HPLC applications

A mixed standard solution containing phentolamine, tamsulosin, and doxazosin at a concentration of 5 × 10−4 mol L−1 for each one was used to validate the specificity of the HEK293 α1A/CMC-offline-HPLC system. Fractions retained by the α1AAR CMSP were separately concentrated and then injected into the HPLC system. This method was also applied to screen the mixed standard solution of tamsulosin, nitrendipine, and nilotinib acting on the α1AAR. To examine whether tamsulosin is interacting with the receptors on the HEK293 CMSP, a series of two retention profiles was constructed.

Determination of K D values

One assumption made is that the drug interacts reversibly with a single type of the binding site on the α1AAR. K D values were measured in an HEK293 α1A/CMC system, performed by continuously pumping tamsulosin ranging from 1.12 × 10−7 to 3.60 × 10−6 mol L−1 through an α1A/CMC column. The breakthrough curves of tamsulosin with different concentrations were recorded. Next, the standard solutions of tamsulosin, terazosin, alfuzosin, doxazosin, 5-methylurapidil, phentolamine, and oxymetazoline (10−4 mol L−1 for each one) were injected into the column, respectively. The capacity factor (k′) of CMC chromatographic peak in elution curve was calculated using the following equation:

$$ k{\prime}{ } = { }\left( {{t_{\text{R}}} - {t_0}} \right)/{t_0} $$
(1)

where t R is the retention time of ligand, and t 0 is the dead time of non-retained solvent.

According to Eq. 2 [18, 24],

$$ \frac{1}{{k\prime - X}} = \frac{{{K_{\text{D}}}{V_{\text{M}}}}}{{{K_{\text{M}}}{{[R]}_{\text{s}}}}}{[L]_{\text{m}}} + \frac{{{K_{\text{D}}}{V_{\text{M}}}}}{{{{[R]}_{\text{s}}}}} $$
(2)

where K D and K M are the equilibrium dissociation constants for the analyte and marker in the mobile phase, respectively. [L]m, [R]s, and V m are the molar concentration of ligands in the effluent, immobilized receptors at the surface of the stationary phase, and the dead volume of the column, respectively. The term X was introduced to the part of k′ value to eliminate the errors and obtained by iterative testing [24, 25]. When using tamsulosin both as a marker and as an analyte, Eq. 2 reduces to

$$ \frac{1}{{k\prime}} = \frac{{{V_{\text{M}}}}}{{{{[R]}_{\text{s}}}}}{[L]_{\text{m}}} + \frac{{{K_{\text{D}}}{V_{\text{M}}}}}{{{{[R]}_{\text{s}}}}} $$
(3)

[R]s/V M is obtained from the slope and substituted to Eq. 2. So, the K D values of the ligands were obtained from the linear regression by plotting 1/(k′ − X) versus [L]m. All data were expressed as means±standard errors of the means (SEM, n ≥ 3), and values of p < 0.01 were regarded as indicative of statistical significance.

Competitive binding study

For further competitive studies, the mobile phase that contained tamsulosin, terazosin, and oxymetazoline (5 × 10−7 mol L−1 for each one) was separately pumped through the α1A/CMC column to saturate the binding sites of α1AAR. Afterwards, tamsulosin was propelled towards the column. The breakthrough curves were recorded and compared to those obtained from the normal column without saturation.

Results and discussion

Biological characteristics of CMSP

In order to characterize the CMSP, the total protein and α1AAR in the cell membrane suspension before and after adsorption on the activated silica were determined, respectively. The difference values, i.e., the content of the total protein and α1AAR in the HEK293 and HEK293 α1AAR CMSP, are listed in Table 1. Other endogenous receptor protein including EGFR and VEGFR-2 was lowly detected in two types of CMSP, respectively. As a result, the HEK293 α1AAR CMSP containing high-level α1AAR was successfully prepared.

Table 1 Biological characteristics of HEK293 and HEK293 α1AAR CMSP
Full size table

Application of an HEK293 α1A/CMC-offline-HPLC system

A mixed standard solution containing phentolamine, tamsulosin, and doxazosin was used to validate the specificity of the HEK293 α1A/CMC-offline-HPLC system. As shown in Fig. 1A, there were three significant retention fractions named R1, R2, and R3 (R1, 11.7 min; R2, 15.9 min; and R3, 25.5 min). Then, the three fractions were collected and concentrated by using C18 solid-phase extraction columns, respectively, followed by injecting into HPLC system for further separation and analysis. The retention times were R1 3.7 min, R2 4.9 min, and R3 14.9 min, respectively (Fig. 1B). Compared with the HPLC chromatogram of the standard solution (Fig. 1S), the retained components were finally identified as phentolamine, tamsulosin, and doxazosin, respectively.

Fig. 1
figure 1

Chromatograms of mixed standard solution using an HEK293 α1A/CMC-offline-HPLC method. (A) HEK293 α1A/CMC chromatogram of mixed standard solution including R1, R2, and R3 fractions; (B) HPLC chromatograms of R1, R2, and R3 fractions; (S) HPLC chromatogram of mixed standard solution of phentolamine (1), tamsulosin (2), and doxazosin (3). HEK293 α1A/CMC column, 10 × 2.0 mm; flow rate, 0.2 mL min−1; column temperature, 37 °C; sample concentration, 5 × 10−4 mol L−1 for each ligand; sample volume, 5 μL; detection wavelength, 254 nm; mobile phase, 2 mM PBS, pH 7.4. RP-HPLC column, Dikma C18 (150 × 4.6 mm, 5 μm); flow rate, 1.0 mL min−1; column temperature, 37 °C; detection wavelength, 254 nm; mobile phase, methanol–water–triethylamine (70:29.8:0.2, v/v/v, adjust pH to 6.0 with acetic acid)

Full size image

A mixed standard solution containing tamsulosin, nitrendipine, and nilotinib was applied to verify the selectivity of the α1A/CMC system. Of the three drugs, nitrendipine is an antagonist acting on L-type calcium channel, and nilotinib is a tyrosine protein kinase inhibitor interacting with BCR-ABL. As shown in Fig. 2, only tamsulosin, an α1AAR antagonist, can be recognized by the system. Also, there was no obvious retention when tamsulosin was injected into an HEK293/CMC system (Fig. 3). It can be concluded that the α1AAR on the CMSP plays a key role, and the system can specifically recognize the components which interact with the α1AAR.

Fig. 2
figure 2

Chromatograms of mixed standard solution using an HEK293 α1A/CMC-offline-HPLC method. A HEK293 α1A/CMC chromatogram of mixed standard solution including R0 and R1 fractions; (B) HPLC chromatograms of R0 and R1 fractions; (S) HPLC chromatogram of mixed standard solution of tamsulosin (1), nitrendipine (2), and nilotinib (3). The chromatographic conditions were same to those of Fig. 1

Full size image
Fig. 3
figure 3

The CMC chromatograms of tamsulosin on the HEK293 (A) and HEK293 α1A (B) column. The chromatographic conditions were same to those of Fig. 1

Full size image

Affinity on the α1AAR

As shown in Fig. 4, the k′ values of ligands on the HEK293 α1A/CMC column varied from each other, as a result of the different affinities that the ligands interacted with α1AAR, and the retention times of six ligands decreased with continuously increasing the concentration of tamsulosin ranging from 1.12 × 10−7 to 3.60 × 10−6 mol L−1 in the mobile phase (Fig. 5A–F). The degree of the shift is relevant to the affinity the ligands interact with α1AAR and the efficient concentration of the binding sites that both the markers and the ligands occupy. Particularly, the injected ligands have other binding sites which do not interact with the marker in the mobile phase. So, the term X is equal to a critical k′ of the ligand at a marker concentration, above which the retention time of the ligand no longer decreased with the increase of the marker's concentration. The X values were 12.35, 9.98, 6.11, 9.83, and 10.45 for terazosin, 5-methylurapidil, phentolamine, alfuzosin, and doxazosin, respectively. In accordance with Eq. 2, the corresponding graph of the reciprocal values of (k′ − X) versus [L]m was obtained with three replicates on three different α1A/CMC columns. The calculated K D values were (1.87 ± 0.13) × 10−6 M for tamsulosin, (2.86 ± 0.20) × 10−6 M for 5-methylurapidil, (3.01 ± 0.19) × 10−6 M for doxazosin, (3.44 ± 0.19) × 10−6 M for terazosin, (3.50 ± 0.21) × 10−6 M for alfuzosin, and (7.57 ± 0.31) × 10−6 M for phentolamine, respectively, and each plot for the ligands gave a linear relationship, with the correlation coefficients varied from 0.8252 to 0.9425 over the six different concentrations tested (Fig. 5A–F). So far as we know, tamsulosin, 5-methylurapidil, terazosin, alfuzosin, and doxazosin selectively interact with α1AR. The first two have a modest selectivity for α1AAR over other subtypes, while the last three demonstrate similar selectivity for all three a1AR subtypes [26, 27]. However, little decrease in the retention time of oxymetazoline was observed (Fig. 5G). The behavior suggested that the ligand interacted with α1AAR in a manner different from tamsulosin, and in fact, it is a nonselective α1AR agonist.

Fig. 4
figure 4

The CMC chromatograms of alfuzosin (a), doxazosin (b), 5-methylurapidil (c), terazosin (d), tamsulosin (e), oxymetazoline (f), and phentolamine (g) on the HEK293 α1A/CMC column. The chromatographic conditions were same to those of Fig. 1, and the detection wavelengths for seven ligands can be found in the text

Full size image
Fig. 5
figure 5

Elution profiles of tamsulosin (A), terazosin (B), 5-methylurapidil (C), phentolamine (D), alfuzosin (E), doxazosin (F), and oxymetazoline (G) on the HEK293 α1A/CMC column with different concentrations of tamsulosin in the mobile phase and regression curves achieved by plotting 1/k′ or 1/(k′ − X) versus [L]m. The six concentrations were 1.12 × 10−7, 2.25 × 10−7, 4.50 × 10−7, 8.99 × 10−7, 1.80 × 10−6, and 3.60 × 10−6 mol L−1, respectively. Each point with a bar represents the mean±SEM (n = 3). The chromatographic conditions were same to those of Fig. 1, and the detection wavelengths for seven ligands can be found in the text

Full size image

We have established a CMC method based on frontal analysis for determining the K D values of calcium antagonist–L-type calcium channel interactions [18]. However, the results were influenced by the molar absorption coefficients of the ligands. Furthermore, the estimation of the dead volume was a complex work. Here, we described a new model combined modified competitive binding, and we only have to obtain the k′ values of the analytes and the concentrations of the marker in the mobile phase, which greatly simplified the calculation process.

Displacement study on the α1AAR

As shown in Fig. 6, tamsulosin was pumped through the column resulting in the displacement of tamsulosin, terazosin, and oxymetazoline on the α1AAR binding sites. The mean position of tamsulosin's breakthrough curve on the α1A/CMC column, which has been saturated with tamsulosin and terazosin, is shifted to the left compared to that of tamsulosin on the normal column. However, that was not the case for oxymetazoline. This means that the terazosin occupies certain binding sites of α1AAR on the column thus hindering the association of tamsulosin, while oxymetazoline interacted with α1AAR in a manner different from tamsulosin. Next, terazosin was pumped through the column saturated with oxymetazoline, which led to the few change in breakthrough times (Fig. 7). So, the two drugs occupy the different class of binding sites on α1AAR as well.

Fig. 6
figure 6

Competitive binding breakthrough curves of tamsulosin (10−6 mol L−1). The straight and dashed lines represent the breakthrough curves using the normal α1A/CMC column and the column after its saturation with tamsulosin (A), terazosin (B), and oxymetazoline (C; 5 × 10−7 mol L−1). HEK293 α1A/CMC column, 10 × 2.0 mm; flow rate, 0.2 mL min−1; column temperature, 37 °C; mobile phase, 5 mM PBS, pH 7.4; detection wavelength, 280 nm

Full size image
Fig. 7
figure 7

Competitive binding breakthrough curves of terazosin (10−6 mol L−1). The straight and dashed lines represent the breakthrough curves using the normal α1A/CMC column and the column after its saturation with oxymetazoline (5 × 10−7 mol L−1). The CMC conditions were same to those of Fig. 6

Full size image

Correlation with the RBA method

As shown in Table 2, the pKi values of tamsulosin ranked first, followed by 5-methylurapidil and doxazosin, whereas terazosin and alfuzosin ranked last [8, 26, 27]. The binding affinities of the five ligands were compared with that reported previously. Due to the use of radioactive nuclides including [3H], [125I], etc., the sensitivity of the method is greatly higher than that obtained from CMC method by two to three orders of magnitude. Even so, the K D values calculated using the CMC method correlated positively with those published pKi values obtained from the RBA technique with r 2 values of 0.9343, 0.9582, and 0.8782 (p < 0.05), respectively. Because of its easiness, simplicity, and accuracy, the CMC method is possessed of popularization and application value.

Table 2 The correlation analysis of K D values obtained from the HEK293 α1A/CMC and pKi values from RBA technique
Full size table

Methodology of the HEK293 α1A/CMC

The standard solution of tamsulosin was kept at 4 °C and was analyzed at selected time intervals after storage over the study period. After 24 h, 48 h, 72 h, and 1 week, a decrease by 3.50%, 6.29%, 11.13%, and 31.32% in retention time was observed, respectively. Thus, the activity of CMC column can keep relatively stable in 2 days.

The limits of detection for the ligands were determined at a signal-to-noise ratio of 3, and they were 5.77 × 10−7, 2.93 × 10−7, 4.12 × 10−7, 3.58 × 10−7, 5.12 × 10−7, 5.45 × 10−7, and 8.02 × 10−7 mol L−1 for tamsulosin, terazosin, alfuzosin, doxazosin, 5-methylurapidil, phentolamine, and oxymetazoline, respectively.

The reproducibility of the α1A/CMC method was performed by continually injecting standard solution (10−4 mol L−1) of phentolamine, tamsulosin, and doxazosin into five different α1A/CMC columns (n = 50), and the results showed that the retention time decreased by merely 18.95%, 11.68%, and 17.08%, respectively, which were used to eliminate the errors caused by decay of membrane protein activity. In addition, the relative standard deviations of the retention times were 6.16%, 8.86%, and 9.95%, respectively.

The robustness of the method was examined by replicate injections (n = 6) of standard solution of tamsulosin at a concentration of 10−4 mol L−1 with slight changes made to the mobile phase composition, pH value, and sample volume (Table 3). Few changes in k′ values were observed.

Table 3 Robustness evaluation of α1A/CMC method
Full size table

Conclusion

The CMC method combined competitive binding provides a powerful technique for the characterization of the α1AAR binding interactions without disturbing the receptor's quaternary structure. There was a positive correlation between K D and the published pKi values, and the binding sites on the α1AAR have been validated successfully.