Prevention of Molecular Self-Association by Sodium Salicylate: Effect on Insulin and 6-Carboxyfluorescein

E. TOUITOU'~, F. ALHAIQUE~, P. FISHER*, A. MEMOLI*, F. M. RICCIERI*, AND E. SANTUCCI~ Received April 13, 1987, from the ,'School of Pharmacy, Hebrew University of Jerusalem, Jerusalem 91 120, Israel, and the 'Istituto di Chimica Farmaceufica e Tossicologica, Unwersita di Roma, Rome, Italy. Accepted for publication July 21, 1987.

- ~ ~ _ _ -~ - ~ ~ ~ _ _ - ~-

Abstract 0 The effect of sodium salicylate on the concentration- dependent self-association of insulin and 6-carboxyfluorescein (CF), as expressed by metachromasy, fluorescence, and changes in aqueous solubility, was learned. By decreasing the CF concentration from 12 to 0.48 pqmL-', A,, peaks shift from the shorter wavelengths (451,474 nm), indicating the presence of oligomers, toward the monomer wave- length region (484 nm). Sodium salicylate shifts the peaks of a 12 pgmL- ' CF solution towards the monomer region, eliminating the peak at the lower wavelengths and generating a spectrum with one peak at 490 nm, the effect being concentration dependent. The fluorescence of insulin and CF solutions increases with their concentration. Quenching of these solutions was observed, up to complete elimination of fluores- cence, when various concentrations of salicylate were added. The water solubility of both molecules, CF and insulin, was considerably increased with the addition of increasing concentrations of salicylate to the solutions: at 37 "C, 2.5 M sodium salicylate solution increases the CF solubility 532 times from 12.2 to 6.5 mgmL-', and 1.5 M salicylate increases the solubility of insulin 7875 times, thus an aqueous solution containing 630 mg-mL-' of insulin may be prepared. The results obtained here, together with our previously reported data, indicate that the interference between sodium salicylate and drug self-association behavior, by increasing drug solubility, may substantially contribute to the improved drug bioavailability mediated by salicylate.

~ ~.

In a number of interesting publication~1-~ it was reported that sodium salicylate improves the rectal absorption of a considerable number of drugs. This effect of sodium salicylate was explained by enhanced membrane permeability as a result of membrane alteration at the protein It is noteworthy that although the drugs tested in the papers cited above are known to exhibit molecular self-association in water, the possibility that sodium salicylate may interfere with the drug association-having a deaggregative effect- was not considered. We have previously reported that sodium salicylate interferes with the self-association in water of molecules with different physicochemical properties such as hydroxyethyl methylcellulose," a nonionic polymer, and methylene blue,g a cationic dye. The present investigation focuses on the effects of sodium salicylate on the self-associa- tion properties of two other molecules, insulin, a polypeptide, and 6-carboxyfluorescein (CF), an anionic dye. These effects were measured by changes in water solubilities as well as in UV and fluorescent spectra.

Experimental Section M a t e r i a l d o d i u m salicylate (Merck), urea (Merck), and 6-car-

boxyfluorescein (CF; Kodak) were analytically pure. Porcine insulin, a gift of Wellcome, had the following characteristics: potency 24.2 U1 mg, Zn 0.34% wlw, Ash 0.7%, and loss on drying of 6.29.

Solubility Measurements-The solubility values of CF or insulin a t 37 "C in water and in solutions with differing molarities of sodium salicylate have been determined. The test was assessed by addition of 5 mg of drug to 10 mL of solution a t intervals of up to 20 h, until a "visual" saturation point was achieved. The solubility was the cumulative amount of drug dissolved before this experimental end- point. The experimental error was <0.5 mglmL.

The "solubility method"'" was not used for CF due to the metachro- maay of CF spectra in the presence of various concentrations of salicylate. It was assumed that CF remained stable during the course of the measurement of solubility a t 37 "C." To test whether insulin underwent any deterioration, insulin solutions were passed through a Sephadex GS0 column a t the end of the experiment. The elution time was found to be the same as a t the beginning of the experiment.

Spectrophotometric Determinations-A bulk solution of CF (so- lution A) was prepared by suspending 5 mg of CF in 100 mL of distilled water and mixing for 20 h. Following this, the suspension was allowed to stand for 2 h and then was filtered. The resulting filtrate was centrifuged a t 2 000 rpm. The clear supernatant was used for direct UV and fluorescent spectrophotometry, for dilutions, or for the preparation of CF solutions containing various additives (e.g., sodium salicylate, urea, sodium sulfate, or sodium hydroxide).

Ultraviolet Spectra Measurements-The UV spectrophotometric determinations were performed in the visible range of 380-550 nm where no interference with the salicylate spectra occurs. The A,,, and absorbance values were measured at room temperature using a Uvikon 800 spectrophotometer (Kontron).

Fluorescence Measurement-Fluorescence measurements were carried out using an mpF 44-A fluorescence spectrophotometer (Perkin-Elmer). The fluorescence of solutions of CF or insulin was tested a t excitation and emission wavelengths where no overlap with sodium salicylate absorption spectra occurred. Solutions of CF in water, in various molarities of salicylate, and in urea or sodium sulfate were tested at excitatiodemission wavelengths of 3231 518 nm. The fluorescence of insulin at two concentrations, 0.06 and 2.1 mgmL-', in the presence of various concentrations of salicylate, was tested either at the excitatiodemission wavelengths of 2781301 nm or at 280/440 nm.

Results and Discussion Effect of Sodium Salicylate on the Visible and Fluores-

cence Spectra-To eliminate the deaggregative effect of salicylate, changes in the CF visible spectrum were investi- gated. Figure 1 presents the s ectra of CF a t three concentra-

solution A diluted 1:25; the darker line marks the spectrum of the initial concentrated solution (solution A). The data clearly indicate that a concentration dependent A,,, shift exists. The spectrum of the concentrated solution shows two peaks, one at 451 nm and a smaller one a t 474 nm. A 10-fold decrease in concentration, obtained by dilution of solution A, shifts the peaks towards the higher wavelengths ke. , 454 and 477 nm). Following a 25-fold dilution of solution A, a spectrum with only one peak at 484 nm was obtained.

This concentration-dependent metacromasy indicates that the visible spectrum of CF in aqueous solution is affected by the molecular self-association of the drug. Based on these results, the CF visible spectrum was used to indicate the effect of sodium salicylate on the self-association of CF. The spectra of a 12 pgmL-' aqueous solution of CF containing increasing concentrations of sodium salicylate are shown in Figure 2, where the darker line represents the spectrum of CF with no salicylate added to the solution (solution A). It can be observed that in 0.1 M sodium salicylate, the CF

tions: solution A (12 pgmL- P ), solution A diluted l:lO, and

0022-3549/67/1000-0791$01 .oO/O 0 1987, American Pharmaceutical Association

Journal of Pharmaceutical Sciences / 791 Vol. 76, No. 70, October 1987

spectrum has a A,,, (484 nm) identical to that obtained by diluting solution A 25 times, and in 1 M sodium salicylate, the peak was shifted to 490 nm. This clearly indicates that the addition of salicylate moved the CF spectrum towards the monomer region (higher wavelengths) until a complete elim- ination of the peaks at the lower wavelengths occurred.

In order to test that the metachromasy in the presence of salicylate was not due to the pH increase in the tested solutions, the spectra were compared with those of CF solutions at similar pH values adjusted by addition of sodium hydroxide. The results summarized in Table I clearly indi- cate that for concentrations >0.1 M salicylate, at the same pH, the two solutions show significant discrepancies between their max values (e.g., at pH 6.64, the A,, in the presence of 1.5 M salicylate was 491 nm, while that of the solution without salicylate was only 483 nm). Although the system is complex, the effect of salicylate becomes dominant at higher concentrations. This was also shown by the solubility data presented below.

Since visible spectrophotometry cannot be used for insulin testing, fluorescence measurements were undertaken. The experiments were carried out for insulin at excitation and

emission wavelengths of 278 and 301 nm, respectively. The fluorescence in the presence of sodium salicylate and in various other media is presented in Figure 3. It can be seen that the fluorescence of insulin solutions increases with increasing insulin concentrations. By addition of sodium salicylate a t any concentration, the fluorescence is complete- ly su pressed. No effect was observed when 3 M urea was addel to the system.

To test whether quenching was possibly related to the absorption of salicylate at the wavelengths used, an addition- al experiment was carried out using excitatiodemission wavelengths of 2801404 nm. Although a t 404 nm sodium salicylate has no absorbance properties, it was found that the fluorescence decreases as salicylate concentration increases. In this region, sodium salicylate has no absorbance proper- ties. These results point towards a strong interaction be- tween insulin and salicylate molecules.

Fluorescence measurements were also assessed for CF solutions and the spectra are presented in Figure 4. The

Table CThe Amax Values of 6-Carboxyfluoresceln In Aqueous Solutlons at Varlous pH Values, wlth and without the Addltlon of Sodlum Salicylate'

I

X. nm

Flgure 1-The UV absorpfion specfra of solutions of various concenfra- fions of CF in wafer. Cells of various lengths have been used: Key: (1 ) solution A (see Experimental Section), 7-cm cell; (2) solution A diluted 1 : 10, 1 -cm cell; (3 ) solufion A diluted 1 :25, 2-cm cell.

380 550 380 550 380 550 380 550

X . nm

Flgure 2-Concenfration-dependent effect of sodium salicylafe on the UV specfrum of CF. Key: (1 ) in wafer; (2) in 0.1 M sodium salicylafe; (3) in 1 M sodium salicylafe.

A,,, nm

Salicylate Salicylateb

6-Carboxyfluoroscein, PH M With Sodium Without Sodium

4.16 - 474.2 - 5.78 10.2 476.0 476.4 5.92 5 x 10-2 480.4 - 6.01 lo-' 483.7 483.5 6.55 1 .o 490.0 484.2 6.64 1.5 491 .O 483.1

a For the preparation of solutions, see Experimental Section (UV spectra measurements). bThe pH was adjusted using sodium hydroxide.

INSULIN SOLUTIONS

Flgure 3-€ffect of sodium salicylate and urea on the fluorescence of insulin in aqueous solutions. Key: (1) 10-'M HCl; (2) 0.06 mg/mL of insulin in M HCl; (3 ) 2.1 mg/mL of insulin in 10- M HCI; ( 4 ) 0.06 mg/mL of insulin (pH 6.8); (5) 0.06 mg/mL of insulin in 7 F 2 M HCl:10- ' M NaCl; (6) 2.7 mg/mL of insulin m 70 M HCl:10 ' M NaCl; (7 ) 0.06 mg/mL of insulin in 3 M urea: 10- ' M NaCl; (8) 0.06 mg/mL of insulin in 1 M NaCl; (9) 7 M sodium salicylafe; (1 0) 2 M sodium salicylare; (1 1 ) 0.06 mg/mL of insulin in 7 M sodium salicylafe; (1 2) 2.1 mg/mL of insulin in 7 M sodium salicylafe; ( 1 3 ) 0.06 mg/mL of insulin in 2 M sodium salicylate; (14) 2.1 mg/mL of insulin in 2 M sodium salicylare.

792 / Journal of Pharmaceufical Sciences Vol. 76, No. 70, October 7987

100-

75

9 50- ? €9 N €9

LL

25 -

-

9 I 0-

7

3

4 - 3

I

0 0.5 1 .o 1.5 2.0 2.5

E -

SODIUM SALICYLATE CONCN., M

Flgure !&-Effect of increasing concentrations of sodium salicylate on the solubility of CF and insulin. Key: (0) CF; (0) insulin.

CF SOLUTIONS

Flgure 4-€ffect of sodium salicylate and other additives on the fluorescence of CF in aqueous solutions. Key: (1 ) CF aqueous solution, no additive; (2 ) in 0.01 M sodium salicylate; (3) in 0.05 M sodium salicylafe; (4) in 0.1 M sodium salicylate; (5) in 1 M sodium salicylate; (6) in 3 M urea; (7) in 0.5 M Na,SO4.

excitatiodemission wavelengths were 323 and 518 nm, re- spectively. In this region, salicylate has no absorption prop- erties. However, it can be clearly seen that the intense fluorescence of CF was quenched, the quenching effect in- creasing with salicylate concentration le.g., at 1 M sodium salicylate (pH 6.51, the fluorescence was completely quenched as a result of a salicylate-CF interactionl. On the other hand, the addition of 3 M urea and 0.5 M sodium sulfate greatly enhanced the fluorescence of solutions. This last observation may be explained by the formation of charged CF anions in the urea and sodium sulfate solutions (pH 7.6 and 6.7, respectively).

Effect of Sodium Salicylate on the Solubility of 6-Car- boxyfluorescein and Insulin-Figure 5 presents the effect of salicylate on the solubilities of CF and insulin at 37 "C. 6- Carboxyfluorescein is a hydrophobic molecule that is practi- cally insoluble in water. At room temperature, we found a solubility of 12.2 pgmL-'. By the addition of increasing concentrations of salicylate, the CF aqueous solubility was significantly increased. In the concentration range tested, a maximum increase of 532-fold in CF solubility was measured in a 2.5 M salicylate solution (i.e., 6.5 mgmL-'1.

The porcine insulin molecule also has a relatively low solubility in water; a t 37 "C its water solubility was mea- sured to be 80 WmL-' . By the addition of sodium salicylate,

the insulin solubility was markedly increased; thus, 630 mgmL-' insulin dissolved in 1.5 m salicylate, representing a 7 875-fold increase in solubility.

For both molecules, the slope of the solubility plot abruptly increased above a salicylate concentration of 0.5 M for the insulin plot and 0.2 M for the CF plot. This was interesting to note since a similar effect was observed in our previous work with methylene blue system^,^ where the self-association of the dye decreased in solutions containing sodium salicylate concentrations >0.45 M.

The above results, in addition to previous findings,elo indicate that salicylate increases the solubilities of different molecules, such as insulin, fluorescein, methylene blue, and hydroxyethyl methylcellulose, which have the common prop- erty of undergoing self-association in aqueous solutions.

References and Notes 1. Nishihata, T.; Rytting, J. H.; Higuchi, T. J . Pharm. Sci. 1982,

2. Nishihata, T.; Rytting, J. H.; Higuchi, T. J . Pharm. Sci. 1982,

3. Nishihata, T.; Rytting, J. H.; Kamada, A.; Higuchi, T.; Routh, M.; Caldwell, L. J. Pharm. Pharmacol. 1983,35, 148-151.

4. Caldwell, L.; Nishihata, T.; Fix, J.; Selk, S.; CarTll, R.; Gardner C. R.; Hi chi, T. Rectal Therapy, Proceedings o the Sym osium on the Axantages and Problems Encountered in Rectal Tierapy; Glas, B.; de Blaey, C. J., Eds.; J. R. Prous: Barcelona, 1983; pp

5. Nishihata, T.; Lee, C.; Yamamoto, M.; Rytting, J. H.; Higuchi,T. J . Pharm. Sci. 1984, 73, 1326-1328.

6. Nishihata, T.; Higuchi, T. Biochim. Biophys. Acta 1984, 775, 269-271.

7. Ka'ii, H.; Horie, T.; Hayashi, M.; Awazu, S. Life Sci. 1985, 37, 52d-530

8. Touitou,'E.; Donbrow, M. Int. J. Pharm. 1982, 11, 131-148. 9. Touitou, E.; Fisher, P. J. Pharm. Sci. 1986, 75, 384-386.

71, 865-868.

71, 869-872.

57-61.

10. Hi chi, T , Zuck, D. A. J . Am. Pharm. Assoc. Sci. Ed. 1953,42, 13c145. .'

11. Martindale: The Extra Pharmacopoeia, 28th Ed.; Reynolds, J. E. F., Ed.; Pharmaceutical: London, 1982; p 517.

Journal of Pharmaceutical Sciences / 793 VOl. 76, NO. 10, October 1987