Binedaline Binding to Plasma Proteins and Red Blood Cells in Humans

DIDIER MORIN*, ROLAND ZINI*, SYLVlE LEDEWYN*, JEAN-PIERRE cOLONNA*, MONIQUE cZAJKA*, AND JEAN-PAUL TILLEMENT*' Received May 2, 1984, from the 'Dhpartement de Pharmacologie, Facult6 de M6decine de Paris XI/, F-94010 Creteil and the *Laboratoires Cassenne, F-75015 Paris, France. Accepted for publication April 2, 1985.

Abstract 0 Serum binding of binedaline, a new antidepressant drug, was studied in vitro by equilibrium dialysis. The percent of binding in serum is high, 99.2%, and remains constant within the range of therapeutic concentrations: no saturation to the binding sites was seen. Investigations performed on isolated proteins with a wide range of concentrations showed one site with a high affinity constant (K, = 2 x lo6 M-') for m,-acid glycoprotein and two sites with a low affinity constant (K, = 3 x lo4 M-') for human serum albumin. Binding to lipoproteins was nonsaturable, with a total affinity constant of 1.25 x 1 O5 < nK, < 2.79 x 1 O6 M-'. Over the range of therapeutic concentrations, the ratio of binedaline concentrations in serum and red blood cells remained constant (1%) and was shown to be dependent on the free fraction of binedaline in serum.

Previous investigations reported that lipid-soluble basic drugs, such as antidepressants, are bound to a large extent in p 1 a sm a 1 - [ [ 2 - (dime t h y 1 a mi no) e t h - yl]methylamino]-3-phenylindole, was recently introduced in therapeutics as an antidepressant. The chemical structure of this drug can be related to tryptophan due to the presence of an indolyl nucleus. This indolyl moiety differentiates bineda- line from the other classical tricyclic antidepressants. Bine- daline is a weakly basic drug, almost completely ionized at plasma pH and highly lipid soluble.

The purpose of this study was to determine to what extent binedaline is bound in serum and to identify the proteins responsible for this binding. In a first step, binding studies were performed in human serum containing proteins a t physiological concentrations. In a second step, binedaline binding was studied in solutions containing the main isolat- ed plasma proteins, i.e., human serum albumin (HSA), al- acid glycoprotein (AAG), and lipoproteins, in order to deter- mine the association constants and the number of binding sites. In addition, a possible binding to isolated red blood cells (RBC) was investigated.

These experiments allowed the determination of the distri- bution of binedaline to the different constituents of blood. Finally, binding modifications which may occur in different pathological states were simulated using the corresponding plasma protein alterations usually observed in these situa- tions.

B i ne d a1 i ne ,7

Experimental Section Binedali~~e-Binedaline,~ as the 14C-labeled and unlabeled hydro-

chloride, was supplied by Cassenne Laboratories. The specific activi- ty of [14C]binedaline was 58 Ci/mol. The radiochemical purity (>95%) was determined by TLC (silica gel F-254; Merck) in chloro- form:methanol:water (75:24:1). The drug was dissolved in a phos-

phate buffer (0.067 M) over a range of concentrations from 0.5 to 62.5 pM, including the therapeutic range of 0.2-1 pM.

Human Serum-A pool of human serum was obtained a h r collection of serum samples in healthy volunteers. The HSA concen- tration of this serum pool was adjusted to 600 pM with phosphate buffer (0.067 MI. Plasma-free fatty acids were determined by GC in a modification of the method of Sampson and Hensley:* myristic, 12.3 pM; oleic, 195.4 pM; palmitic, 154.3 pM; palmitoleic, 33.1 pM; and stearic, 42.6 pM acids. The AAG was measured by radial immunodif- fusion (M-partigen; Behringwerke). The serum pool was stored at

Human Serum Albumin and nl-Acid Glycoprotein-HSA (Sig ma A 1887) and AAG (electrophoresis purity: 99%; Behringwerke) were solubilized in phosphate buffer, pH 7.4, at a concentration of 10 pM. The molar ratio of fatty acids to HSA was 0.04.9

Lipoproteins-Lipoproteins were isolated from human serum using the method of Nelsonlo modified by Glasson et 81." The concentrations were 1 and 10 pM for low-density (LDL) and high- density lipoproteins (HDL), respectively. These concentrations were determined using average molecular masses of 3 x lo6 and 3 x 10' Da for LDL and HDL, respectively.

Red Blood Cells-The RBCs were washed in 0.9% NaCl and adjusted to a hematocrit of 0.47 in phosphate buffer (0.067 M) containing 9 g/L of NaC1. In addition, the same hematocrit (0.47) was reconstituted in the serum pool described above with the washed RBCs. [14C]Binedaline and RBCs were incubated for 1 h a t 3TC, then centrifuged at 1500xg at 37°C for 15 min. [14ClBinedaline concentrations were determined in RBC suspensions in buffer and/or serum. The [14C]binedaline RBC concentrations (CRBc) were ob- tained by:

-30°C.

where H is the hematocrit value. Equilibrium Dialysi~-['~C]Binedaline binding to various isolat-

ed proteins and to plasma were measured by equilibrium dialysis. Experiments were carried out at 37°C and pH 7.4 for 2 h, under constant stirring at 20 rpm (Dianorm apparatus). No significant binding was observed to the dialysis tubing (VisKing).

Calculations-At equilibrium, the concentrations in each com- partment were measured by liquid scintillation counting (Packard tricarb 460 CD). Free (F) and bound (B) molar concentrations of [14C]binedaline were calculated. The binding percentage was deter- mined from the concentrations obtained a t equilibrium:

drug binding (%) = - x 100 B + F

The data obtained a t equilibrium (B and F) were fitted according to:

where N,, nj, K . denote, respectively, the molar binding site concen- tration of the j & class, the number of binding sites, and the affinity constant to this specific protein; R denotes the concentration of the protein.

The parameters n and K, were calculated by a nonlinear least- squares method using a Gauss-Newton algorithm.12 The binding percentages and estimated parameters were expressed as the mean +- SD of 3-5 determinations.

Oa22-3549/85/0700-0727$0 1 . OO/O 0 1985, American Pharmaceutical Association

Journal of Pharmaceutical Sciences / 727 Vol. 74, No. 7, July 1985

Simulation of Binedaline Plasma Binding-At nonsaturating concentrations, the percentage of bound binedaline to one class of binding sites fj) for one type of protein (i) is calculated from the following:13

where T stands for the total [14C]binedaline concentration, and the sum of N,K,, products denotes the binding coefficient of all plasma proteins. Each plasma protein concentration being determined, the total percentage of [14C]binedaline binding to human serum, can be then calculated according to:

T = T T B Bj

Results [14C]Binedaline Binding to Human Serum-Over a range

varying from 0.5 to 4.5 pM [l4C1binedaline, the percentage of binding remained constant. The mean value is 99.2 * 0.6% (Ttjqle I).

[ CIBinedaline Binding to Human Serum Albumin- Over the same range (0.5-4.5 pM) of [14C]binedaline levels, the binding percentages to 600 pM HSA also remained constant, 96.1 IT 0.4% (Table I), but were significantly lower than those found in human serum. These results suggest that other plasma proteins are also responsible for the binding of [14C]binedaline. On the other hand, when [l4C1binedaline concentrations increased from 0.5 to 62.5 pM, the binding percentages to 10 pM HSA decreased from 34 to 16%. These results indicate that [l4C]bineda1ine binding to HSA was following a saturable process. The Scatchard plot is linear, and the B versus F plot, according to eq. 2, also shows one class of binding sites, with n = 1.7 * 0.2 and K, = 29,900 * 6499 M-' (Fig. 1).

[ CIBinedaline Binding to al-Acid Glycoprotein- [14C]Binedaline binding to 10 pM AAG was saturable over the range of concentrations of 0.5-62.5 pM. Subsequently, the binding percentages decreased from 96 to 20%. The Scatchard plot is linear, and the binding process involves one class of binding sites with n = 0.92 * 0.02 and a high affinity cor$ant (K,) = (2.21 t 0.35) x lo6 M-' (Fig. 2).

[ CIBinedaline Binding to Lipoproteins-When concen- trations of [14C]binedaline varied from 0.5 to 62.5 pM, the binding percentages remain constant and the average values were 72.5 f 0.7 and 57.8 & 0.8% to LDL and HDL, respec-

l2 t

F, ILM

Figure l--['4C]Binedaline binding to human serum albumin (10 pM). The number of binding sites was 1.7 t 0.2, and the affinity constant (Ka) was 29,900 t 6400 M - ' . Each point represents the mean f SD of three determinations. The inset is the Scatchard plot.

I 0 10 20 30 40 50 60

F, LLM

Figure 2-[14C]Binedaline binding to a1 -acid glycoprotein (10 pM). The number of binding sites was 0.92 f 0.02, and the affinity constant (K,) was (2.21 2 0.35) x lo6 M ~~ '. Each point represents the mean t SD of three determinations. The inset is the Scatchard plot.

tively (Table I). These data showed that 1l4C1binedaline binding to lipoproteins was nonsaturable, with nK, = (2.79 +- 0.06) x lo6 M-' for LDL and nK, = (1.25 -t 0.01) x lo5 Mpl for$DL (Fig. 3).

[ CIBinedaline Binding to Red Blood Cells-Over a range of 0.5-4.5 pM of [14C]binedaline, [binedaline]RBc/ [binedaline],,,,, and [binedaline]~~c/[binedaline]b,ff,, ratios remain constant and were 1.0 * 0.2 and 89.1 +- 0.8%, respectively. The [binedalinel~~c/[binedalinel,,,,, ratio can

Table I-['4C]Binedaline Binding Percentage to Human Serum and Plasma Proteins-'

Binedaline Concentration, pM

0.5 1 3 4.5 62.5 Protein

Serum 99.1 f 0.5 99.1 t 0.5 99.4 t 0.9 99.2 f 0.7 -

HSA 96.3 f 0.4 96.2 f 0.6 96.2 f 0.3 95.8 t 0.5 - (HSA = 600 pM)

600 pM 10 pM 34.5 2 2.1 33.6 t 1.3 35.0 t 3.7 36.0 t 3.6 16.0 t 3.0'

AAG 96.0 t 0.5 95.5 t 0.3 93.2 t 0.2' 88.8 2 0.3d 20.0 t 1.5'

LDL 71.7 f 1.0 73.1 f 1.3 73.4 t 0.7 72.1 t 1.6 72.5 f 0.8

HDL 57.3 t 0.5 56.7 f 0.6 58.8 f 2.3 58.5 I 1.5 57.8 t 1.2

(10 PM)

(1 PM)

(10 PM)

,The plasma protein concentration was 65 g/L with 600 pM of human serum albumin and 16pM al-acid glycoprotein. Each percentage represents the mean 5 SD of five determinations. The percentage changes of bound ['4C]binedaline were compared by the Mann-Whitney test. Key: (HSA) human serum albumin; (AAG) al-acid glycoprotein; (LDL) low-density lipoprotein; (HDL) high-density lipoprotein. =p < 0.05. d p < 0.01.

728 /Journal of Pharmaceutical Sciences Vol. 74, No. 7, July 1985

be viewed as the [14C]binedaline free fraction in serum (0.8%).

Simulation of Binedaline Plasma Binding-The calculat- ed binding percentage of binedaline (1 pM), from eqs. 3 and 4, in normal serum was 98.7%, i.e., close to the measured percentage of 99.2% (Table I). This percentage (98.7%) was the sum of the respective percentages of HSA (50%), AAG (41.7%), and lipoproteins (7%) binding (Fig. 4). Some varia- tions of the plasma protein concentrations induced changes in the [ 14C]binedaline binding percentage to each protein. Thus, in serum containing 5 pM AAG, the simulated bineda- line binding percentage yielded 98.1%. The same result was obtained with a concentration of 300 pM HSA in serum. When HSA and AAG concentrations were 300 and 7.5 pM, respectively, the overail serum binding of binedaline fell to 97.4% (Fig, 4). RBCs were not taken into account in this simulation because the [ l4C1binedaline concentration in the

Flgure 3--f34CfSinedaline binding to lipoproteins; the low-density (A) and high-density (B) lipoprotein concentrations were 1 and 10 pM, respectively. The nK, values were (2.79 ? 0.06) x lo6 and (1.25 +- 0.01) x lo5 M - for low- and high-density lipoproteins, respectively. Each point represents the mean t SD of three determinations.

31

A B C D E

Flgure 4-Simulation of binedaline binding in serum. The simulation was calculated from e9. 4, and the binedaline concentration was 1 pM. Key: (A) normal serum [600 pM HSA (U); 15 pM AAG (8); 15 pM lipoproteins (O)]; (B) serum (600 pM HSA; 7.5 pM AAG; 15 pM lipoproteins); (C) serum (600 pM HSA; 5.0 pM AAG; 15 pM lipopro- teins); (D) serum (300 pM HSA; 15 pM AAG; 15 pM lipoproteins); (E) serum (300 pM HSA; 7.5 pM AAG; 15 pM lipoproteins).

RBCs was viewed as the free [14C]binedaline concentration in serum.

As expected, binedaline is highly bound to human serum (-99%), and this binding was constant over the range of the studied concentrations. However, binding is not a cause of plasma drug retention because the apparent volume of distri- bution of binedaline is relatively large (3.5 L/kg). Binedaline is bound to all tested proteins, i.e., HSA, AAG, and lipopro- teins. This heterogeneous binding involves two kinds of processes: saturable and nonsaturable. This fact accounts for the binding percentage remaining constant over the range of therapeutic levels (0.5-1 pM) and even up to 4.5 pM (Table I). Binedaline binding to HSA (10 pM) is saturable, with a low affinity constant (29,900 M-'). This result is surprising because binedaline is a weakly basic molecule and, as a rule, the basic compounds, like imipramine,14 show a nonsatura- ble binding process to HSA. The pharmacological effects of binedaline and imipramine are similar, but their structural conformations are different. Binedaline shows a benzyl moi- ety linked to an indolyl residue. This structure is different from tricyclic antidepressants, such as imipramine, but the side chains are similar. The different binding processes to HSA of the two drugs can be explained by their different cyclic nuclei. At therapeutic levels, binedaline occupies 0.04% of the HSA binding sites.

Binedaline binding to AAG is also saturable but with a high affinity constant (2 x lo6 M-'), and at 1 p M , binedaline occupies -3% of the AAG binding sites. The occupation of 50% of isolated AAG binding sites would require a total concentration of 6 pM binedaline. This concentration far exceeds the therapeutic levels. Both binedaline and imipra- mine show saturable binding to AAG, but binedaline exhib- its an higher affinity constant than that of imipramine (K, = 64,000 M-l, data not shown).

The nonsaturable binding of binedaline to lipoproteins is 7% of the overall serum binding and can be regarded as an apparent partition coefficient between the saturable proteins (HSA and AAG) and lipoproteins. This percentage of bind- ing-site occupation on lipoproteins can increase in hyperlipo- proteinemia. The [binedalinel~c/[binedalinelb,,~er ratio can also be viewed as an apparent partition coefficient between the RBCs and buffer. This result can be related to the hydrophobic character of binedaline. The [binedalinel~~cl [binedaline],,, ratio is low (1%) and can be accounted for by the binedaline plasma free fraction which diffuses into the RBCs. In normal serum, and over the therapeutic range of binedaline concentrations (0.5-1 pM), the binding percent- age remains constant and the free fraction does not vary (Table I). On the other hand, at 1 pM binedaline, a decrease in plasma protein concentrations, mainly AAG and HSA, involves an increase in binedaline free fraction. The simula- tion (Fig. 4) allows the determination of the relative binding percentages to each plasma protein. When AAG decreases from 15 to 5 pM, the free fraction increases from 1.4 to 1.9%. This change shows a relative increase of -36%. The same relative increase in free fraction is observed when the HSA plasma concentration falls to 300 pM.

If AAG and HSA concentrations are reduced by one-half, the free fraction increases from 1.4 to 2.6%, leading to a relative increase of -86% (Fig. 4). The simulation shows that only a decrease in HSA and AAG concentrations leads to a large increase in the binedaline free fraction. This phenome- non should be considered in pregnancy, in the elderly, and in nephrotic syndrome and hepatic failure.

Binedaline could be displaced from AAG by other basic drugs such as disopyramide,15 pr~pranolo l ,~~ and erythromy- cinl6 if these drugs share the same binding site to AAG.

Journal of Pharmaceutical Sciences / 729 Vol. 74, No. 7, July 7985

However, this s i tuat ion seems unlikely because binedaline has a higher afinity constant than these drugs. Additionally, the effect of displacement from a specific protein m a y be minimized by the “buffering” effect of the o ther binding proteins.

1.

2.

3. 4. 5.

6.

7. 8. 9.

References and Notes Pike, E.; Skuterud, B.; Kieruf, P.; Fremstad, D.; Abdel Sayed, S. M.; Lunde, P. K. M. Clin. Pharmncokinet. 1981, 6, 367374. Burch, J. E.; Roberts, S. G.; Raddats, M. A. Psychopharmacology 1981. 75. 262-272. Bickel, M. H. J . Pharm. Pharmncol. 1975,27,733-738. Danon, A.; Chen, 2. Clin. Pharmacol. Ther. 1979,25, 316321. Reidenbera, M. M.; Odar-Cederlof, I.; Bahr, V. C.; Borga, 0.;

id . J . Med. 1971.285. 264-267. Sioavist. FT N . Er Xlekanderson. B.YBorea. 0. Eur. ‘ J . Clin. Phnrmncol. 1972. 4 .

I , I ,

195-200. Cassenne Laboratories, French Patent 2 265 365, 1975. Sampson, D.; Hensley, W. J. Clin. Chim. Acta 1975, 61, 1-8. Birkett, D. J.; Myers, S. P.; Sudlow, G. Clin. Chim. Acta 1978, 85, 253-258.

10. Nelson, R. A. “Rapid, Single-Spin Fractionation of Serum Lipo- proteins by Density Gradient Ultracentrifugation in Sorvall Vertical Rotors”; Biochemical Application Laboratory, Du Pont Go.: Wilmington, DE, 1980; pp 1-4.

11. Glasson, S.; Zini, R.; Tillement, J. P. Biochem. Phurmncol. 1982,

12. Zini, R.; #Athis, Ph.; Barre, J.; Tillement, J. P. Biochem. Phar- 31, 831-835.

rnc01. 1979.28. 2661-2665. 13. Glasson, S.; kin;, R.; d’Ath&, Ph.; Tillement, J. P.; Boissier, J. R.

14. Tillement, J. P.; Zini, R.; d’Athis, Ph.; Boissier, J. R. J. Phnrmn-

15. Lima, J. J.; Boudoulas, H.; Blandford, M. J . Pharmucol. Erp.

16. Prandota, J.; Tillement, J. P.; d’Athis, Ph.; Campos, H.; Barr6, J.

Mol. Pharmacol. 1980, 17, 187-191.

col. Clin. 1974, I , 227-232.

Ther. 1981,219, 741-747.

J . Int. Med. Res. 1980, 8 Suppl. 2, 1-8.

Acknowledgments This study was made possible with special grants from the Univer-

sity of Paris XII, from the “Direction de la Recherche au Ministhre de 1’Education Nationale,” and from Cassenne Laboratories.

730 / Journal of Pharmaceutical Sciences Vol. 74, No. 7, July 1985