Novel In Situ Modified PVC Membrane Potentiometric Sensors for Determination of Levamisole Hydrochloride in Human Plasma, Urine and Bovine Milk

M. M. Khalil. This paper is focused on construction and investigating the characteristics of novel levamisole (LVM) in situ modified polymeric membrane sensors. Two sensors were prepared based on the cation exchangers phosphotungstic acid and phosphomolybdic acid (PME1 and PME2, respectively) using dibutyl phthalate as plasticizing solvent. These sensors showed Nernstian slopes of 59.2±0.4 and 58.31±0.4 mV decade in the concentration ranges (1.0x101.0x10 and 1.0x10-1.0x10 mol L) with low detection limits (1.0x10 and 2.8x10mol L) for PME1and PME2, respectively. The sensors were found to be very selective and usable within a wide pH range (2-8). They exhibited fast response times (˂ 10 s), good stability and long life span (40 and 37 days), respectively. LVM is determined successfully in pure solutions, pharmaceutical preparations and biological fluids (human plasma, urine and bovine milk) using the standard addition and potentiometric titration methods with high accuracy and precision.


Introduction
Levamisole (LVM) (2,3,5,6-tetrahydro-6-Phenyl imidazole [2,1-b] thiazole) ( Fig. 1 ) [1] (C 11 H 13 N 2 SCl) belongs to synthetic imidazothiazole derivatives. It is a white to almost white crystalline powder, which is almost odorless and is freely soluble in water. It is quite stable in acid aqueous media but hydrolyzes in alkaline or neutral solutions. This drug is a broad spectrum anthelminthic drug widely used to control internal parasites in large livestock and occasionally in human medicine [2]. Because LVM acts as an inhibitor of lipid peroxidation, it is also a radioprotectant drug [3]. Also, levamisole is an immunomodulator in different cancer cells including colorectal, breast cancer, melanoma, and leukemia [4]. Besides, it has been shown that levamisole has anti-cancer activity in combination with fluorouracil (5-FU) as adjuvant therapy for colon carcinoma [5]. Levamisole gained forensic interest after the increase of its use as an adulterant in illicit cocaine samples; as a result, levamisole is now found in the majority of cocaine seized worldwide, linked to debilitating and eventually fatal immunologic effects in cocaine abusers [6]. Excess use of levamisole was found to have serious adverse side effects such as agranulocytosis, cutaneous vasculopathy and leukoencephalopathy [7]. The presence of levamisole determines an additional health threat due to aminorex, one of its main metabolites which was found to cause pulmonary hypertension [8][9][10]. N N S H .HCl membrane thickness, the amount of THF was kept constant, and its evaporation was fixed for 24 h. Thickness of the membrane is about 0.2 mm. A 12 mm diameter disk was cut out from the prepared membrane and glued to one end of a Pyrex glass tube using PVC-THF paste. Ratio of membrane ingredients, time of contact and concentration of conditioning solution were optimized; so that the potentials recorded were reproducible and stable within the standard deviation [31]. Membrane to membrane reproducibility was assured by carefully following the optimum condition of fabrication. The membrane that gave reproducible results and showed best performance was selected for further studies. The optimized electrode body was filled with a solution of 1.0×10 -1 mol L -1 NaCl and 1.0×10 -2 mol L -1 LVM. The electrode was preconditioned before use by soaking in a 1.0×10 -2 mol L -1 LVM solution for 15 minutes and storing in the same solution when not in use.

Construction of calibration curves
The conditioned electrodes were immersed in conjunction with the Ag/AgCl double-junction reference electrode in solutions of levamisole hydrochloride in the range of 1.0 ×10 -7 -1.0 ×10 -2 mol L -1 . They were allowed to equilibrate whilst stirring and recording the e.m.f. readings within ±1 mV. The mV concentration profiles were plotted. The regression equations for the linear part of the curves were computed and used for subsequent determination of unknown concentrations of levamisole hydrochloride.

Effect of pH
The effect of pH on the response of the investigated electrodes was studied using 10 −2 , 10 -3 and 10 −4 mol L -1 LVM solutions over the pH range of 1-11. This is done by immersing the electrodes in the drug solution. The pH was gradually increased or decreased by addition of very small volumes of dilute NaOH or HCl solutions, respectively. The potential obtained at each pH was recorded.

Selectivity coefficient determination
The separate solution method [40,41] and the matched potential method (MPM) [42,43] were employed to determine the selectivity coefficients, log K LVM, J z+ pot , of the potentiometric sensors towards different species. In the separate solution method, the potential of a cell comprising a working electrode and a reference electrode is measured in two separate solutions, where, E 1 is the potential measured in 1.0×10 −3 mol L -1 LVM, E 2 the potential measured in 1.0×10 −3 mol L -1 of the interfering compound, z 1 and z 2 are the charges of LVM and interfering species, respectively, and S is slope of the electrode calibration plot. The selectivity coefficients were determined by the separate solution method using the rearranged Nikolsky equation [41]: log K LVM, J z+ pot = ((E 1 -E 2 )/S) + (1 + (z 1 /z 2 )) log a In 1995, IUPAC recommendation [42] prescribes the MPM [43,44] as the method of choice for ions of different charge. MPM is considered [42] as a purely operational method, not relying on any theoretical or empirical model equation. The quantity used to express the extent of interference is the ratio of the primary ion concentration increment to the interfering ion concentration that gives the same potential change in a constant initial backgroundof primary ion. The selectivity coefficient was determined by measuring the change in potential upon increasing the primary ion activity from an initial value of a A to á A and a B represents the activity of interfering ion added to the reference solution of primary ion of activity a A which also brings the same potential change. It is given by expression: In the present study a A and á A were kept at 1.0×10 −4 and 1.2×10 −4 mol L -1 levamisole hydrochloride and a B was experimentally determined.

Potentiometric determination of levamisole hydrochloride
The standard addition method was applied [33,45], in which small increments of the standard solution 10 −2 mol L -1 of LVM hydrochloride were added to 50 mL aliquot samples of various concentrations from pure drug or pharmaceutical preparations. The change in millivolt reading was recorded for each increment and used to calculate the concentration of LVM hydrochloride sample solution using the following equation: WhereC x and V x are the concentration and the volume of the unknown, respectively, C s and V s the concentration and the volume of the standard solution, respectively, S the slope of the calibration graph and ΔE is the change in mV due to the addition of the standard solution.

Potentiometric titration of levamisole hydrochloride
Aliquots of 1.0×10 −2 mol L -1 drug solution (pure or tablet) were transferred into 50-mL volumetric flasks and made up to the mark with bidistilled water. Different concentrations of LVM hydrochloride were prepared, then titrated potentiometrically with a standard solution of 1.0 ×10 −2 mol L -1 PTA. The volume of the titrant at equivalence point was obtained using the conventional S-shaped curves.

Determination of levamisole hydrochloride in pharmaceutical preparations
An accurate weight of Katrex ® (40 mg/tablet) tablets ground and finely powdered in a mortar was dissolved in the bidistilled water up to 30 mL by stirring for 1 h. The solution was filtered in a 50-mL measuring flask. The residue was washed three times with bidistilled water; the volume was completed to the mark by the same solvent to form 1.0 ×10 −2 mol L -1 solution. The resulting potentials of the drug solution were directly measured using its corresponding ion-selective electrode.

Determination of levamisole hydrochloride in biological fluids
Different amounts of LVM hydrochloride and 5 mL of plasma or urine of a healthy person or milk of healthy bovine were transferred to a 50 mL measuring flask and completed to the mark using bidistilled water. The contents of the measuring flask were transferred to a 100-mL beaker, and then subjected to standard addition method.

Results and Discussion
Optimization of membrane composition Levamisole cation was found to form 3:1 stable water insoluble ion-pair complex with each of phosphotungstic acid and phosphomolybdic acid as indicated conductometricaly (Fig. 2). The electroanalytical performance and electrode potential of an ISE are dependent upon the selective extraction of the target ion which creates the electrochemical phase boundary potential due to thermodynamic equilibria at the sample/electrode interface. Using of a suitable ion pairing agent in the electrode matrix has the advantage of reducing the time required for the electrode preparation where there is no need for ion pair (IP) preparation as well as expansion of the application of ion selective electrodes (ISEs) for the determination of drugs that cannot be precipitated as suitable IPs.
In order to determine the suitable content of ion-pairing agents (PTA and PMA), several electrodes of a varying nature and ratio of ion pairing agent were prepared from the systematic investigation of each electrode composition. Experimental trials proved that a certain percentage of each ion pairing agents was optimum, indicated by the Nernstian behavior of the electrodes. However, further increase of the ion pairing agents over this percentage resulted in a diminished response slope of the electrode, most probably due to some inhomogenities and possible saturation of the membrane [46]. The results given in Table 1 show that the optimum ion pairing agent percentage is 5% for PTA and PMA that gave the highest slope values of 59.2±0.36, 58.3±0.37 mV decade -1 for PME1and PME2, respectively.  . 3. Calibration graphs for a) PME 1 and b) PME 2 at optimum membrane composition

Effect of plasticizer
The plasticizer mainly acts as a fluidizer, allowing homogeneous dissolution and diffusion mobility of the ion-pair inside the membrane. The nature of the plasticizer must be properly controlled in order to minimize the electrical asymmetry of the membrane and to limit fouling of the sensor. The nature of the plasticizer has a marked influence on the response slope, linear domain and also on the selectivity of the PVC membrane electrodes. In exploration for a suitable plasticizer for constructing these electrodes, four plasticizers with different polarities including DBP, DOP, DOA and TCP were used as shown in Table 2. The results revealed that DBP was the best plasticizer tested. Poor sensitivities for the electrodes plasticized using DOP, DOA and TCP are due to low solubilities or low distributions of PTA and PMA ion-pairing agents in these solvents [47]. The electrodes using DBP as a plasticizer provide higher Nernstian slope, wide response range, more stable potential reading and lower limit of detection due to the better extraction of the drug in the organic layer of the membrane [48][49][50].

Effect of internal solution
Varying the composition of internal reference solution could considerably improve the linear range and limit of detection [51,52], therefore, the response of the electrode in relation to the variation of internal solution was investigated using the optimum membrane composition. Three different concentrations of LVM.HCl (1.0×10 −2 , 1.0×10 −3 , and 1.0×10 −4 mol L -1 ) or with 1.0×10 −1 mol L -1 NaCl were used. It was found that the best results in terms of slope and working concentration range have been obtained with internal solution of concentration 1.0 ×10 −2 mol L -1 LVM.HCl and 1.0×10 −1 mol L -1 NaCl for PME 1 and PME 2 electrodes.

Performance characteristics of the constructed electrodes
The final performances of the constructed electrodes were investigated according to IUPAC recommendations [53].

Effect of soaking and lifetime of the electrodes
Freshly prepared electrode must be soaked to activate the surface of the membrane to form an infinitesimally thin gel layer at which ion exchange occurs. This preconditioning process requires different times depending on diffusion and equilibration at the electrode test solution interface; a fast establishment of equilibrium is certainly a condition for a fast potential response [54]. The lifetimes of the electrodes were determined for intervals till the electrode loses its Nernstian behavior. This behavior established that the loss of plasticizer, ionic site from the polymeric film due to leaching into the bathing solution is a primary reason for the limited lifetimes of the electrodes. The response of the electrodes has been measured by recording the calibration graph at 25 o C at different intervals. Lifetimes of the electrodes were found to be 40 and 37 days for PME 1 and PME 2 electrodes, respectively during which the electrodes showed a slight gradual decrease in the slope and an increase in the detection limit. Effect of paste duration on the response characteristic of the proposed electrodes is shown in Table 3. After preparation of the proposed electrodes, they were kept at 4 o C and directly used for potentiometric measurements.

Dynamic response time and repeatability of the electrode
The dynamic response time [53], is defined as the time which elapses between the instant at which an ion-selective electrode and a reference electrode (ISE cell) are brought into contact with a sample solution. The dynamic response time of the electrode was tested by measuring the time required to achieve a steady-state potential (within ±1 mV) after successive immersions of the electrode in a series of drug solutions, each having a 10-fold increase in concentration from 1.0×10 −6 to 1.0×10 −2 and 1.0×10 −5 to 1.0×10 −2 mol L −1 for PME 1 and PME 2 , respectively. The electrodes yielded a steady potential within 10 s as shown in Fig. 4. The repeatability of the potentials readings for each electrode was examined by subsequent measurement in 1.0×10 −3 mol L −1 levamisole hydrochloride solution immediately after measuring in 1.0×10 −2 mol L −1 levamisole hydrochloride solution (Fig. 5). An insignificant difference in potential readings was obtained, indicating good repeatability of the constructed electrode.

Fig. 5.
Dynamic response of a) PME 1 and b) PME 2 for several high-to-low sample cycles

Effect of pH
Since pKa of levamisole is 8.0 [55], therefore, at pH 6.20 levamisole is nearly completely ionized, i.e. levamisole will be in the cationic form. The concentration distribution diagram for levamisole hydrochloride species is constructed using SPECIES program [56] (Fig. 6). The results showed that potential response remained almost constant over the pH range 2-8 for the two investigated sensors as shown in Fig. 7. At higher pH values, the potential showed a sharp decrease; due to the formation of nonprotonated LVM, leading to a decrease in concentration of LVM + . However, at lower pH values, the decrease in potential may be attributed to interference of hydronium ion. The sensors response was checked with bidistilled water, 0.1 mol L -1 acetate buffer pH 5.0 or 0.1 mol L -1 phthalate buffer pH 5.0. The best results were achieved in bidistilled water; because it provided not only a higher Nernstian slope but also a stable potential reading. Therefore, bidistilled water was used for all the constructed sensors.

Effect of temperature
The thermal coefficients of the investigated sensors were determined as mentioned recently [31]. The values were found to be 0.00042 and 0.0034 V/ o C for PME 1 and PME 2 , respectively; indicating that the sensors had high thermal stabilities within the studied temperature range and PME 1 sensor is more thermally stable than PME 2 sensor.

Selectivity of the sensors
The selectivity behavior as one of the most important characteristics of ISEs, was studied for levamisole hydrochloride sensors with respect to a variety of ionic and nonionic species using separate solution method (SSM) and matched potential method (MPM), respectively [34]. The selectivity coefficients values of the sensors (Table 4) reflect their very high selectivity for the LVM cation. The high selectivity of sensors toward inorganic cations can be attributed to the differences in ionic size and consequently their mobilities and permeabilities as compared with LVM cation. In addition, the low interference of nonionic species may be ascribed to the difference in polarity and to the lipophilic nature of their molecules [32].

Potentiometric determination of levamisole
Analytical applicability of the investigated sensors was tested by applying potentiometric titration and standard addition methods in pure solutions and pharmaceutical preparations ( Table 5, Fig. 8). The calculated F-and t-values, shown in Table 6 did not exceed the theoretical values, reflecting the precision and accuracy of the applied method. Determination of LVM in spiked urine, plasma and bovine milk samples was carried using standard addition method; the mean recoveries obtained were in the range of 97-102 and 98.3-101.8% for PME 1 and PME 2 sensors, respectively (Table 7). For ruggedness of the method a comparison was performed between the intra-and inter-day assay results for levamisole obtained by two M. Sc. candidates. The RSD values for the intra-and inter-day assays of levamisole in the cited formulations performed in the same laboratory by the two analysts did not exceed 2.43%. On the other hand, the robustness was examined while the parameter values (pH of the medium and the laboratory temperature) were being deliberately slightly changed. Levamisole recovery percentages were good under most conditions, not showing any significant change when the critical parameters were modified. This result indicates that the proposed sensors show a good reproducibility and stability.

Comparison with reported methods
Although some performance characteristics of the reported methods [20,25,26] are better than those of the proposed sensors, our technique still shows superiority in many important ways. It is low coast, fast, more available and precise (RSD reaches 0.61%) as shown in Table 8, indicating the ability of the constructed sensors to face such automated methods.

Conclusion
The present work involves the preparation of novel PVC membrane sensors with in situ mode of modification. The described sensors are sufficiently selective for the quantitative determination of LVM in pure form, pharmaceutical dosage form, human urine, plasma and bovine milk. The sensors showed Nernstian slopes with low detection limit (1.0×10 -6 mol L -1 ) with fast response time (10 s) and long operational life time (40 days) in the concentration range 1 .0×10 -6 to 1.0×10 -2 mol L -1 . They also show high sensitivity, adequate selectivity, high thermal stability and applicability over a wide pH range (2)(3)(4)(5)(6)(7)(8) with no sample pretreatment.