- Email: [email protected]
Effect of Nicotinamide and Urea on the Solubility of Riboflavin in Various Solvents
+
+
Effect of Nicotinamide and Urea on the Solubility of Riboflavin in Various Solvents RENEÄ E E. COFFMAN*
AND
DANE O. KILDSIGX
Received January 11, 1996, from the Department of Industrial and Physical Pharmacy, School of Pharmacy, Purdue University, West Lafayette, IN 47907. Final revised manuscript received June 26, 1996. Accepted for publication June 27, 1996X. * Present address: Western University of Health Sciences, College of Pharmacy, Pomona, CA 91766. Abstract 0 Hydrotropy is a solubilization process whereby addition of large amounts of a second solute results in an increase in the aqueous solubility of another solute. Past investigations have focused on the potential interaction of the hydrotropic agent with the solubilized solute. Conversely, this study proposes that at least some hydrotropic agents exert their solubilizing effect predominately by interacting with the solvent. To that end, the effect of two hydrotropic agents, nicotinamide and urea, on riboflavin solubility in aqueous and nonaqueous systems was examined. The term “solutropy” is introduced to describe solubilization by addition of large amounts of a second solute in any solvent. The nonaqueous solvents used included methanol, N-methylformamide, dimethyl sulfoxide, and acetone. In water, methanol, and N-methylformamide, riboflavin solubility was found to increase with increasing nicotinamide concentration; however, riboflavin solubility decreased with increasing nicotinamide concentration in dimethyl sulfoxide and acetone, thus establishing the solvent-dependent nature of solutropy. An examination of solvent properties revealed that the solvent’s ability to be both a proton donor and acceptor is important mechanistically, while dielectric constant and polarity are not. The same solvent-dependency was observed with urea, although urea is a poorer solutrope than nicotinamide. This study proposes that some solutropic agents act by changing the nature of the solvent, specifically by altering the solvent’s ability to participate in structure formation via intermolecular hydrogen bonding.
Introduction Nicotinamide, vitamin B3, is well-known as a hydrotropic agent and has demonstrated the ability to solubilize a wide variety of therapeutic entities including riboflavin.1-9 The term “hydrotropy” refers to a solubilization process whereby addition of large amounts of a second solute results in an increase in the aqueous solubility of a sparingly soluble solute. This is in direct opposition to “normal” solution behavior in that, normally, addition of a second solute to a saturated system will cause precipitation of the less soluble solute. Urea, widely known as a protein denaturant, is also known to have some hydrotropic properties. Although its hydrotropic ability has not been studied as thoroughly as that of nicotinamide, it has demonstrated an ability to solubilize a wide array of compounds including aminophenazone, thiazide diuretics, acetazolamide, and the structural isomers of hydroxybenzoic acid.1,10,11 The term hydrotropy does not imply a specific solubilization mechanism, and in fact, few definitive mechanistic studies have been undertaken. The majority of studies on the mechanism of hydrotropic solubilization focus on potential interactions between the hydrotropic agent and the solute and conclude that the solubilization is due to complex formation between the two.12-16 However, because the hydrotropic agent is present in such large concentrations (usually 1-2 M), it is X
Abstract published in Advance ACS Abstracts, August 1, 1996.
© 1996, American Chemical Society and American Pharmaceutical Association
conceivable that the hydrotropic agent exerts its effect not via interaction with the solute but by altering the nature of the solvent. In that case, one would expect the solvent to play a significant role in the hydrotropic effect. This concept has been heretofore ignored by researchers. Therefore, in order to understand the mechanism of hydrotropic solubilization, the objective of this study is to determine the role that the solvent plays in this type of solubilization by looking at the effects of two different hydrotropic agents, nicotinamide and urea, in various solvents.
Materials and Methods ChemicalssRiboflavin USP (RFN) and nicotinamide USP (NA) were used as supplied by Pharmavite (Los Angeles, CA). Methanol (MeOH) and dimethyl sulfoxide (DMSO) were purchased from Fisher Scientific (Fairlawn, NJ), and acetone was obtained from EM Science (Gibbstown, NJ). N-Methylformamide (NMF) was purchased from Sigma Chemicals (St. Louis, MO). Urea (Microselect for Microbiology) was obtained from Fluka Chemicals (Ronkonkema, NY). All solvents were either analytical or HPLC grade. Water was double distilled prior to each solubility experiment. Effect of Nicotinamide on Riboflavin SolubilitysThe solubility of RFN in water, MeOH, NMF, DMSO, and acetone was determined using the phase solubility method. A quantity of RFN far in excess of the intrinsic solubility was placed in a 25 mL scintillation vial. The vial was covered with aluminum foil in order to prevent degradation of the RFN by light. To each vial was added 20 mL of either pure solvent or NA solution. NA concentration varied depending upon NA solubility in the solvent in question. The pH of the aqueous NA solutions was determined using a Fisher Accumet pH meter (Fairlawn, NJ). The pH ranged from 6.45 for 0.2 M NA solutions to 6.40 for 2.0 M NA solutions. A small magnetic stirbar was added and the vial sealed first with parafilm, then with the vial screw-top cap, and finally with an additional exterior layer of parafilm. This was done in order to prevent leakage from the water bath into the vials. The vials were then placed in the magnetic-stirring, circulating water bath at 30 ( 0.1 °C and allowed to equilibrate. The equilibrated mixture was filtered, diluted, and then analyzed for concentration using the Beckman DU-7 spectrophotometer. The pH of the filtrate did not differ significantly from that of the NA solution used to dissolve the RFN ((0.02 pH units). pH was determined again after dilution and was found to be 6.60 ( 0.02 pH units. Any pH changes observed during these experiments were determined to be insignificant and did not result in the precipitation of either RFN or NA. Prior to these studies, a scan (λ ) 200-500 nm) of a 1 µg/mL aqueous solution of RFN was run on the spectrophotometer to ascertain the wavelength of maximum absorption in each solvent. Concentration of the diluted filtrates was determined by comparison to a standard curve for RFN in each of the solvents. Additionally, a standard curve was generated using RFN in aqueous solutions of NA (10 and 20 µg/mL) to ensure that NA would not interfere with the RFN spectrophotometric assay. Determination of the time to achieve equilibrium is described below. Particulars for each solvent are noted in Table 1. Crystals from the supernatant of RFN/NA 1.0 M and RFN/ NA 2.0 M were recovered and analyzed by DSC and X-ray powder diffraction. No differences between recovered crystals and physical mixtures of RFN and NA in the same mole ratio were noted. Determination of Equilibration TimesThe time required for RFN solutions to reach equilibrium was determined by preparing RFN as described in the solubilization section above and placing the vials
S0022-3549(96)00012-3 CCC: $12.00
Journal of Pharmaceutical Sciences / 951 Vol. 85, No. 9, September 1996
+
+
Table 1sExperimental Conditions for the Effect of Nicotinamide on Riboflavin Solubility in Various Solvents Solvent
RFN λmax (nm)
NA Concn Range Used
Equilibration Time (h)
Water MeOH NMF DMSO Acetone
444.5 445.0 445.0 447.5 440.0
1.0 × 10 to 2.0 0.1−2.0 0.1−2.0 0.1−2.0 0.025−0.10
16 18 18 30 16
-4
Figure 2sStructures of selected solvents. Table 2sProperties of Selected Solvents H-Bond Abilityb
Figure 1sHydrotropic solubilization of RFN flavin by NA. Data points represent the average of three experiments; error bars represent standard deviations. in the 30 °C water bath. One vial was removed and its contents filtered, diluted, and analyzed for concentration after each of the following time intervals: 8, 16, 18, 24, and 48 h. Time to equilibrium was defined as the time at which RFN concentration in the filtrate did not change more than 4%. Effect of Urea on Riboflavin SolubilitysThe effect of urea on riboflavin solubility was studied in the following solvents: water, MeOH, DMSO, and acetone. The solubility studies were performed exactly as those using nicotinamide as the hydrotropic agent; however, the concentration of urea solutions prepared differed. In water, urea concentration ranged from 0.2 to 6.0 M. For MeOH, 0.5-1.5 M urea solutions were studied. Urea concentration ranged from 1.0 to 4.0 M in DMSO, while in acetone, the concentration range was 0.010.05 M. The maximum concentration employed for each solvent was a function of urea solubility in that solvent.
Results and Discussion Hydrotropic Solubilization of Riboflavin by NicotinamidesThe degree to which NA was able to solubilize RFN was studied by preparing solutions of NA in concentrations ranging from 0.1 mM to 2.0 M. Over the NA concentration range studied, a 36-fold increase in RFN solubility was seen. From the intrinsic solubility of 8.25 × 10-5 g/mL (0.219 mM), the solubility of RFN reached 2.99 × 10-3 g/mL (7.94 mM) when dissolved in 2.0 M solutions of NA. This dramatic increase in solubility is depicted graphically in Figure 1. The response is not linear, having a definite upward curvature as NA concentration increases. It should also be noted that, even at millimolar concentrations of NA, the solubility of RFN increased, so it was impossible to find a “critical hydrotrope concentration” at which hydrotropic action began. On the contrary, hydrotropic solubilization seems to be a continuous phenomenon. Although pH and ionic strength of these solutions were not controlled, the pH of NA solutions only changes 0.02 pH units (from 7.20 to 7.22) over the concentrations studied and the ionic strength ranges from 7.0 × 10-6 to 13.0 × 10-6 M. With a 0.02 pH unit change, the extent to which RFN is ionized only increases from 0.10 to 0.105%, so increases in solubility 952 / Journal of Pharmaceutical Sciences Vol. 85, No. 9, September 1996
Solvent
Classification
Dielectric Constant
Donor
Acceptor
Water MeOH NMF Acetone DMSO
Protic Protic Protic Aprotic Aprotic
78.5 32.6 182 21.4 45
1.17 0.98 0.62 0.08 0.00
0.47 0.66 0.80 0.43 0.76
a
Taken from ref 18. b Taken from ref 19.
due to a greater percentage of ionized RFN would be insignificant. Ionic strength changes are even less significant, inasmuch as ionic strength is at micromolar concentrations. The initial and most obvious reaction to these results would be to consider complexation between NA and RFN. This hypothesis was rejected following fluorescence quenching and UV/vis spectrophotometric analysis of solutions of NA and RFN. No effect on RFN fluorescence was observed upon its dissolution in NA solutions. The NA solutions ranged in concentration from 7 × 10-8 to 2.0 M, while RFN concentrations ranged from 7 × 10-8 to 1 × 10-4 M. At any given RFN concentration, addition of NA to the system caused no significant deviation from intrinsic RFN fluorescence. These data strongly suggest that complexation between RFN and NA does not occur, even at very low concentrations of NA. Similarly, changes in the UV/vis spectrum of RFN upon NA addition were not indicative of complexation. The spectrum of RFN shows two characteristic peaks at 374 and 444.5 nm. NA does not absorb in this range. No new peaks appeared, although a shift in the parent peaks of RFN was observed. A gradual bathochromic shift with increasing NA concentration accompanied by a slight decrease in absorbance occurred. At the highest NA concentration (2.0 M), the peaks at 374 and 444.5 nm had shifted to 379 and 449.5 nm, respectively.18 Shifts of this type are often associated with π-π complexation; however, π-π complexation requires a π-electron donor and acceptor. π-π complexation would not be expected to occur between NA and RFN because both are π-electron acceptors. Small changes in peak position and absorbance like those seen here are also observed when the solvent medium in which a compound is analyzed changes. This prompted us to consider the possibility that NA was affecting changes in aqueous RFN solubility not by interacting with RFN but by changing the nature of water as a solvent. Consequently, we decided to look at the role of the solvent in hydrotropic solubilization. Solubilization of Riboflavin by NicotinamidesSolvent EffectssThe role of the solvent in hydrotropic solubilization was investigated by studying the effect of NA on the solubility of RFN in nonaqueous solvents. The nonaqueous solvents chosen were MeOH, NMF, DMSO, and acetone. The struc-
+
+
Table 3sSolubility of RFN (mg/mL) in the Presence of Various Concentrations of NA in Water, MeOH, NMF, DMSO, and Acetone at 30 °Ca NA Concn (M)
Water
MeOH
NMF
DMSO
Acetoneb
0 0.1 0.2 1.0 2.0
0.0825 (0.0064) 0.151 (0.0052) 0.277 (0.0036) 1.27 (0.0073) 2.99 (0.097)
0.0328 (0.0018) 0.0433 (0.00088) 0.0529 (0.0034) 0.108 (0.0034) 0.299 (0.036)
2.47 (0.088) 2.51 (0.085) 2.55 (0.095) 3.02 (0.014) 3.66 (0.14)
39.7 (2.2) 27.8 (2.5) 25.1 (1.9) 23.7 (2.2) 15.4 (0.84)
0.0356 (0.0092) 0.0256 (0.0015)
a Solubilities represent the average of three experiments. Standard deviations are in parentheses. b For acetone, NA concentration was limited by its solubility. At 0.025 M NA, RFN solubility was 0.0274 mg/mL, and at 0.05 M NA, RFN solubility was 0.0258 mg/mL.
Table 4sSolubility Factor for Riboflavin in Water, MeOH, and NMF Solutions of Nicotinamidea
a
Figure 3sEffect of NA on the solubility of riboflRFN in water (circles), MeOH (squares), NMF (triangles), and DMSO (diamonds). Data points represent the average of three experiments; error bars represent standard deviations.
tures of these solvents are shown in Figure 2. These solvents were chosen with regard to specific properties purported to have an effect on solubility. These properties include dielectric constant, solvent classification, hydrogen-bond donor ability, and hydrogen-bond acceptor ability and are listed for each solvent in Table 2. Of the four solvents chosen, “hydrotropic” behavior was observed only in the case of MeOH and NMF. This finding is significant, since solubilization of this type heretofore has been demonstrated only in water. However, because the term “hydrotropy” refers only to aqueous systems, a new term describing this process is necessary. We define “solutropy” as a phenomenon whereby addition of a large amount of a solid solute to any solvent causes an increase in the solubility of another slightly soluble solute. Using this definition, hydrotropy would be considered a specific case of solutropy in aqueous systems. We restrict the definition of the solutropic agent to include only solids in order to distinguish solutropic solubilization from solubilization by cosolvents. Table 3 shows the effect of NA on RFN solubility in all five solvents tested. Solutropic behavior is seen only in water, MeOH, and NMF. Solubility data in water, MeOH, NMF, and DMSO are depicted graphically in Figure 3 (RFN solubility in acetone is not great enough to be seen on this graph). It is evident that the addition of NA to DMSO in increasing concentrations causes a decrease in RFN solubility; that is, typical solution behavior is noted, while in the other three solvents, solutropic behavior is seen. From Table 2, one can easily see that there is no correlation between dielectric constant or solvent polarity and solutropic solvents. In fact, NA exhibits the greatest solutropic effect in water, which has a dielectric constant of 78.5, and a moderate effect in MeOH and NMF with dielectric constants of 32.6 and 182, respectively. However, typical solution behavior occurs in DMSO (dielectric constant ) 45) and acetone (dielectric constant ) 21.4). Moreover, the intrinsic solubility of the solubilized solvent is not related to solutropic effects. In DMSO and acetone (“typical” solvents) the intrinsic solubility of RFN is the highest and lowest, respectively. One can, however, find
Nicotinamide (M)
SFwater
SFMeOH
SFNMF
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
3.35 5.39 8.59 11.56 15.39 18.7 23.0 26.9 31.0 36.2
1.61 2.10 2.39 2.68 3.28 4.00 5.01 5.93 7.15 9.08
1.03 1.08 1.15 1.20 1.23 1.26 1.28 1.32 1.43 1.48
Solubility factor (SF) ) solubilitytotal/intrinsic solubility.
a correlation between hydrogen-bonding capabilities and the solubilizing effect. Although all of the solvents listed are capable of accepting hydrogens on the lone pairs of oxygen and/or nitrogen atoms, only DMSO and acetone are incapable of donating hydrogens for hydrogen-bond formation. Therefore, the ability of a solvent to be a hydrogen donor must be a key factor in the “hydrotropic” solubilization phenomenon. Additionally, one must consider the solvent-solvent interactions which may or may not occur. All of the solvents except DMSO and acetone have the dual ability to donate and accept hydrogens, which means that intermolecular hydrogen bonds between like solvent molecules can form. This type of intermolecular hydrogen bonding in solvents has been welldocumented for water and is known as the “iceberg” or “flickering cluster” model of water structure.20 The same type of solvent-solvent interactions would be expected to occur in MeOH, and in NMF, although probably not to the extent to which it occurs in water. MeOH can only donate one hydrogen, and according to Table 2, it does so less readily than water. Similarly, NMF has only one hydrogen to donate, which it does less freely than either MeOH or water. This trend is reflected in Table 4. Table 4 represents the solubility data in terms of the solubility factor. The solubility factor is the solubility of RFN in a NA solution of a given concentration divided by the intrinsic solubility of RFN in that particular solvent. The solubility factor normalizes solubility data with respect to intrinsic solubility. Table 4 shows that, at all concentrations of NA, water is a better solutropic solvent than is MeOH, which is better than NMF. This parallels the hydrogen-donating ability as shown in Table 2. Solubilization of Riboflavin by UreasIn order to test some of the theories proposed in the previous section, the ability of urea to solubilize RFN was studied. As mentioned in the Introduction, urea has demonstrated the ability to hydrotropically solubilize a few different compounds; however, RFN solubilization has not been documented. Solvent effects were also examined by looking at the effect of urea on RFN solubility in the same nonaqueous solvents used with NA. Figure 4 is a graph of RFN solubility as a function of increasing urea concentration in aqueous solution. From that graph, one can conclude that urea does indeed hydrotropically solubilize RFN. It is interesting to note that, unlike NA, which exhibits hydrotropic effects even at low concentrations,
Journal of Pharmaceutical Sciences / 953 Vol. 85, No. 9, September 1996
+
+
in water and MeOH, while in DMSO and acetone, typical solution behavior is seen.
Summary and Conclusions
Figure 4sEffect of urea on the solubility of RFN flavin in water. Data points represent the average of three experiments; error bars represent standard deviations.
Figure 5sComparison of the hydrotropic abilities of NA (squares) and urea (circles) for RFN. Data points represent the average of three experiments; error bars represent standard deviations.
Figure 6sEffect of urea on the solubility of RFN in water (circles), MeOH (squares), and DMSO (diamonds). Because solubility in DMSO is much higher, solubility is expressed as mg/mL.
significant increases in RFN solubility are not evident until the concentration of urea reaches approximately 0.8-1.0 M. A comparison of urea and NA as hydrotropes is seen in Figure 5. Clearly, NA is a much more effective hydrotrope, since only about one-third of the maximum solubility of RFN in 2.0 M solutions of NA is attained by urea at 6.0 M concentrations. Solubility data for urea in water, MeOH, and DMSO are shown in Figure 6. Acetone displayed typical solution behavior as increasing urea concentration resulted in a decrease in RFN solubility. The data presented in this figure confirm the theories set forth earlier with NA. Solutropic effects are seen 954 / Journal of Pharmaceutical Sciences Vol. 85, No. 9, September 1996
In order to ascertain the role of the solvent in solutropic solubilization, solvent-specific effects were explored by looking at solubilization of RFN by NA and urea in different solvents. The results of these experiments lead to the novel discovery that increases in solubility due to addition of another solid solute in high concentration occur in solvents other than water, namely, those solvents which are capable of both hydrogen donation and acceptance. It is proposed here that NA, urea, and other solutropic agents exert their solubilizing effect by changing the nature of the solvent, specifically by altering the solvent’s ability to participate in structure formation via intermolecular hydrogen bonding. The results of the experiments presented in this work may be summarized as follows: 1. NA, in concentrations of up to 2.0 M, is able to increase aqueous RFN solubility approximately 36-fold. There is no discrete concentration of NA which produces a marked increase in RFN solubility; rather, it is a continuous phenomenon seen even at very low NA concentrations. 2. Solubilization effects are seen only in solvents which can both donate and accept hydrogens in hydrogen-bond formation. Therefore, large concentrations of NA can increase RFN solubility in water, MeOH, and NMF, but normal solution behavior is seen in DMSO and acetone. 3. The same solvent-dependency was seen using urea, another solutropic agent that, until this time, had not been shown to solubilize RFN. 4. Comparing the solubilizing capabilities of NA and urea, NA was found to be the more powerful solutrope.
References and Notes 1. Ibrahim, S. A.; Ammar, H. O.; Kasem, A. A.; Abu-Zaid, S. S. Pharmazie 1979, 34, 809-812. 2. Frost, D. V. J. Am. Chem. Soc. 1947, 69, 1064-1065. 3. Hussain, M. A.; DiLuccio, R. L.; Maurin, M. B. J. Pharm. Sci. 1993, 82, 77-79. 4. Truelove, J.; Bawarshi-Nassar, R.; Chen, N. R.; Hussain, A. Int. J. Pharm. 1984, 19, 17-25. 5. Rasool, A. A.; Hussain, A. A.; Dittert, L. W. J. Pharm. Sci. 1991, 80, 387-393. 6. Chen, A. X.; Zito, S. W.; Nash, R. A. Pharm. Res. 1994, 11, 398401. 7. Kenley, R. A.; Jackson, S. E.; Winterle, J. S.; Shunko, Y.; Visor, G. C. J. Pharm. Sci. 1986, 75, 648-653. 8. Elsamaligy, M. S.; Hazma, Y. E.; Abd-Elgawad, N. A. Pharm. Ind. 1992, 54, 474-477. 9. Badwan, A. A.; El-Khordagui, L. K.; Khalil, S. A. Int. J. Pharm. 1983, 13, 67-74. 10. Ammar, H. O.; Ibrahim, S. A.; El-Faham, T. H. Pharm. Ind. 1981, 43, 292-295. 11. Altwein, D. M.; Delgado, J. W.; Cosgrove, F. P. J. Pharm. Sci. 1965, 54, 603-606. 12. Hussain, M. A.; DiLuccio, R. L.; Mauring, M. B. J. Pharm. Sci. 1993, 82, 77-79. 13. Hazma, Y. E.; Kata, M. Pharm. Ind. 1990, 52, 363-368. 14. Elsamaligy, M. S.; Hazma, Y. E.; Abd-Elgawad, N. A. Pharm. Ind. 1992, 54, 474-477. 15. Ammar, H. O.; Ibrahim, S. A.; El-Faham, T. H. Pharm. Ind. 1981, 43, 292-295. 16. Fawzi, M. B.; Davison, E.; Tute, M. S. J. Pharm. Sci. 1980, 69, 104-106. 17. The Merck Index, 8th ed.; Stecher, P. G., Ed.; Merck and Co., Inc.: Rahway, NJ, 1968. 18. Marcus, Y. Chem. Soc. Rev. 1993, 93, 409-416. 19. Nemethy, G.; Scheraga, H. A. J. Chem. Phys. 1962, 36, 33823400. 20. Coffman, R. E.; Kildsig, D. O. Results presented in part at the 9th Annual AAPS Research Meeting, San Diego, CA, 1994; PDD 7405.
JS960012B
+
Effect of Nicotinamide and Urea on the Solubility of Riboflavin in Various Solvents RENEÄ E E. COFFMAN*
AND
DANE O. KILDSIGX
Received January 11, 1996, from the Department of Industrial and Physical Pharmacy, School of Pharmacy, Purdue University, West Lafayette, IN 47907. Final revised manuscript received June 26, 1996. Accepted for publication June 27, 1996X. * Present address: Western University of Health Sciences, College of Pharmacy, Pomona, CA 91766. Abstract 0 Hydrotropy is a solubilization process whereby addition of large amounts of a second solute results in an increase in the aqueous solubility of another solute. Past investigations have focused on the potential interaction of the hydrotropic agent with the solubilized solute. Conversely, this study proposes that at least some hydrotropic agents exert their solubilizing effect predominately by interacting with the solvent. To that end, the effect of two hydrotropic agents, nicotinamide and urea, on riboflavin solubility in aqueous and nonaqueous systems was examined. The term “solutropy” is introduced to describe solubilization by addition of large amounts of a second solute in any solvent. The nonaqueous solvents used included methanol, N-methylformamide, dimethyl sulfoxide, and acetone. In water, methanol, and N-methylformamide, riboflavin solubility was found to increase with increasing nicotinamide concentration; however, riboflavin solubility decreased with increasing nicotinamide concentration in dimethyl sulfoxide and acetone, thus establishing the solvent-dependent nature of solutropy. An examination of solvent properties revealed that the solvent’s ability to be both a proton donor and acceptor is important mechanistically, while dielectric constant and polarity are not. The same solvent-dependency was observed with urea, although urea is a poorer solutrope than nicotinamide. This study proposes that some solutropic agents act by changing the nature of the solvent, specifically by altering the solvent’s ability to participate in structure formation via intermolecular hydrogen bonding.
Introduction Nicotinamide, vitamin B3, is well-known as a hydrotropic agent and has demonstrated the ability to solubilize a wide variety of therapeutic entities including riboflavin.1-9 The term “hydrotropy” refers to a solubilization process whereby addition of large amounts of a second solute results in an increase in the aqueous solubility of a sparingly soluble solute. This is in direct opposition to “normal” solution behavior in that, normally, addition of a second solute to a saturated system will cause precipitation of the less soluble solute. Urea, widely known as a protein denaturant, is also known to have some hydrotropic properties. Although its hydrotropic ability has not been studied as thoroughly as that of nicotinamide, it has demonstrated an ability to solubilize a wide array of compounds including aminophenazone, thiazide diuretics, acetazolamide, and the structural isomers of hydroxybenzoic acid.1,10,11 The term hydrotropy does not imply a specific solubilization mechanism, and in fact, few definitive mechanistic studies have been undertaken. The majority of studies on the mechanism of hydrotropic solubilization focus on potential interactions between the hydrotropic agent and the solute and conclude that the solubilization is due to complex formation between the two.12-16 However, because the hydrotropic agent is present in such large concentrations (usually 1-2 M), it is X
Abstract published in Advance ACS Abstracts, August 1, 1996.
© 1996, American Chemical Society and American Pharmaceutical Association
conceivable that the hydrotropic agent exerts its effect not via interaction with the solute but by altering the nature of the solvent. In that case, one would expect the solvent to play a significant role in the hydrotropic effect. This concept has been heretofore ignored by researchers. Therefore, in order to understand the mechanism of hydrotropic solubilization, the objective of this study is to determine the role that the solvent plays in this type of solubilization by looking at the effects of two different hydrotropic agents, nicotinamide and urea, in various solvents.
Materials and Methods ChemicalssRiboflavin USP (RFN) and nicotinamide USP (NA) were used as supplied by Pharmavite (Los Angeles, CA). Methanol (MeOH) and dimethyl sulfoxide (DMSO) were purchased from Fisher Scientific (Fairlawn, NJ), and acetone was obtained from EM Science (Gibbstown, NJ). N-Methylformamide (NMF) was purchased from Sigma Chemicals (St. Louis, MO). Urea (Microselect for Microbiology) was obtained from Fluka Chemicals (Ronkonkema, NY). All solvents were either analytical or HPLC grade. Water was double distilled prior to each solubility experiment. Effect of Nicotinamide on Riboflavin SolubilitysThe solubility of RFN in water, MeOH, NMF, DMSO, and acetone was determined using the phase solubility method. A quantity of RFN far in excess of the intrinsic solubility was placed in a 25 mL scintillation vial. The vial was covered with aluminum foil in order to prevent degradation of the RFN by light. To each vial was added 20 mL of either pure solvent or NA solution. NA concentration varied depending upon NA solubility in the solvent in question. The pH of the aqueous NA solutions was determined using a Fisher Accumet pH meter (Fairlawn, NJ). The pH ranged from 6.45 for 0.2 M NA solutions to 6.40 for 2.0 M NA solutions. A small magnetic stirbar was added and the vial sealed first with parafilm, then with the vial screw-top cap, and finally with an additional exterior layer of parafilm. This was done in order to prevent leakage from the water bath into the vials. The vials were then placed in the magnetic-stirring, circulating water bath at 30 ( 0.1 °C and allowed to equilibrate. The equilibrated mixture was filtered, diluted, and then analyzed for concentration using the Beckman DU-7 spectrophotometer. The pH of the filtrate did not differ significantly from that of the NA solution used to dissolve the RFN ((0.02 pH units). pH was determined again after dilution and was found to be 6.60 ( 0.02 pH units. Any pH changes observed during these experiments were determined to be insignificant and did not result in the precipitation of either RFN or NA. Prior to these studies, a scan (λ ) 200-500 nm) of a 1 µg/mL aqueous solution of RFN was run on the spectrophotometer to ascertain the wavelength of maximum absorption in each solvent. Concentration of the diluted filtrates was determined by comparison to a standard curve for RFN in each of the solvents. Additionally, a standard curve was generated using RFN in aqueous solutions of NA (10 and 20 µg/mL) to ensure that NA would not interfere with the RFN spectrophotometric assay. Determination of the time to achieve equilibrium is described below. Particulars for each solvent are noted in Table 1. Crystals from the supernatant of RFN/NA 1.0 M and RFN/ NA 2.0 M were recovered and analyzed by DSC and X-ray powder diffraction. No differences between recovered crystals and physical mixtures of RFN and NA in the same mole ratio were noted. Determination of Equilibration TimesThe time required for RFN solutions to reach equilibrium was determined by preparing RFN as described in the solubilization section above and placing the vials
S0022-3549(96)00012-3 CCC: $12.00
Journal of Pharmaceutical Sciences / 951 Vol. 85, No. 9, September 1996
+
+
Table 1sExperimental Conditions for the Effect of Nicotinamide on Riboflavin Solubility in Various Solvents Solvent
RFN λmax (nm)
NA Concn Range Used
Equilibration Time (h)
Water MeOH NMF DMSO Acetone
444.5 445.0 445.0 447.5 440.0
1.0 × 10 to 2.0 0.1−2.0 0.1−2.0 0.1−2.0 0.025−0.10
16 18 18 30 16
-4
Figure 2sStructures of selected solvents. Table 2sProperties of Selected Solvents H-Bond Abilityb
Figure 1sHydrotropic solubilization of RFN flavin by NA. Data points represent the average of three experiments; error bars represent standard deviations. in the 30 °C water bath. One vial was removed and its contents filtered, diluted, and analyzed for concentration after each of the following time intervals: 8, 16, 18, 24, and 48 h. Time to equilibrium was defined as the time at which RFN concentration in the filtrate did not change more than 4%. Effect of Urea on Riboflavin SolubilitysThe effect of urea on riboflavin solubility was studied in the following solvents: water, MeOH, DMSO, and acetone. The solubility studies were performed exactly as those using nicotinamide as the hydrotropic agent; however, the concentration of urea solutions prepared differed. In water, urea concentration ranged from 0.2 to 6.0 M. For MeOH, 0.5-1.5 M urea solutions were studied. Urea concentration ranged from 1.0 to 4.0 M in DMSO, while in acetone, the concentration range was 0.010.05 M. The maximum concentration employed for each solvent was a function of urea solubility in that solvent.
Results and Discussion Hydrotropic Solubilization of Riboflavin by NicotinamidesThe degree to which NA was able to solubilize RFN was studied by preparing solutions of NA in concentrations ranging from 0.1 mM to 2.0 M. Over the NA concentration range studied, a 36-fold increase in RFN solubility was seen. From the intrinsic solubility of 8.25 × 10-5 g/mL (0.219 mM), the solubility of RFN reached 2.99 × 10-3 g/mL (7.94 mM) when dissolved in 2.0 M solutions of NA. This dramatic increase in solubility is depicted graphically in Figure 1. The response is not linear, having a definite upward curvature as NA concentration increases. It should also be noted that, even at millimolar concentrations of NA, the solubility of RFN increased, so it was impossible to find a “critical hydrotrope concentration” at which hydrotropic action began. On the contrary, hydrotropic solubilization seems to be a continuous phenomenon. Although pH and ionic strength of these solutions were not controlled, the pH of NA solutions only changes 0.02 pH units (from 7.20 to 7.22) over the concentrations studied and the ionic strength ranges from 7.0 × 10-6 to 13.0 × 10-6 M. With a 0.02 pH unit change, the extent to which RFN is ionized only increases from 0.10 to 0.105%, so increases in solubility 952 / Journal of Pharmaceutical Sciences Vol. 85, No. 9, September 1996
Solvent
Classification
Dielectric Constant
Donor
Acceptor
Water MeOH NMF Acetone DMSO
Protic Protic Protic Aprotic Aprotic
78.5 32.6 182 21.4 45
1.17 0.98 0.62 0.08 0.00
0.47 0.66 0.80 0.43 0.76
a
Taken from ref 18. b Taken from ref 19.
due to a greater percentage of ionized RFN would be insignificant. Ionic strength changes are even less significant, inasmuch as ionic strength is at micromolar concentrations. The initial and most obvious reaction to these results would be to consider complexation between NA and RFN. This hypothesis was rejected following fluorescence quenching and UV/vis spectrophotometric analysis of solutions of NA and RFN. No effect on RFN fluorescence was observed upon its dissolution in NA solutions. The NA solutions ranged in concentration from 7 × 10-8 to 2.0 M, while RFN concentrations ranged from 7 × 10-8 to 1 × 10-4 M. At any given RFN concentration, addition of NA to the system caused no significant deviation from intrinsic RFN fluorescence. These data strongly suggest that complexation between RFN and NA does not occur, even at very low concentrations of NA. Similarly, changes in the UV/vis spectrum of RFN upon NA addition were not indicative of complexation. The spectrum of RFN shows two characteristic peaks at 374 and 444.5 nm. NA does not absorb in this range. No new peaks appeared, although a shift in the parent peaks of RFN was observed. A gradual bathochromic shift with increasing NA concentration accompanied by a slight decrease in absorbance occurred. At the highest NA concentration (2.0 M), the peaks at 374 and 444.5 nm had shifted to 379 and 449.5 nm, respectively.18 Shifts of this type are often associated with π-π complexation; however, π-π complexation requires a π-electron donor and acceptor. π-π complexation would not be expected to occur between NA and RFN because both are π-electron acceptors. Small changes in peak position and absorbance like those seen here are also observed when the solvent medium in which a compound is analyzed changes. This prompted us to consider the possibility that NA was affecting changes in aqueous RFN solubility not by interacting with RFN but by changing the nature of water as a solvent. Consequently, we decided to look at the role of the solvent in hydrotropic solubilization. Solubilization of Riboflavin by NicotinamidesSolvent EffectssThe role of the solvent in hydrotropic solubilization was investigated by studying the effect of NA on the solubility of RFN in nonaqueous solvents. The nonaqueous solvents chosen were MeOH, NMF, DMSO, and acetone. The struc-
+
+
Table 3sSolubility of RFN (mg/mL) in the Presence of Various Concentrations of NA in Water, MeOH, NMF, DMSO, and Acetone at 30 °Ca NA Concn (M)
Water
MeOH
NMF
DMSO
Acetoneb
0 0.1 0.2 1.0 2.0
0.0825 (0.0064) 0.151 (0.0052) 0.277 (0.0036) 1.27 (0.0073) 2.99 (0.097)
0.0328 (0.0018) 0.0433 (0.00088) 0.0529 (0.0034) 0.108 (0.0034) 0.299 (0.036)
2.47 (0.088) 2.51 (0.085) 2.55 (0.095) 3.02 (0.014) 3.66 (0.14)
39.7 (2.2) 27.8 (2.5) 25.1 (1.9) 23.7 (2.2) 15.4 (0.84)
0.0356 (0.0092) 0.0256 (0.0015)
a Solubilities represent the average of three experiments. Standard deviations are in parentheses. b For acetone, NA concentration was limited by its solubility. At 0.025 M NA, RFN solubility was 0.0274 mg/mL, and at 0.05 M NA, RFN solubility was 0.0258 mg/mL.
Table 4sSolubility Factor for Riboflavin in Water, MeOH, and NMF Solutions of Nicotinamidea
a
Figure 3sEffect of NA on the solubility of riboflRFN in water (circles), MeOH (squares), NMF (triangles), and DMSO (diamonds). Data points represent the average of three experiments; error bars represent standard deviations.
tures of these solvents are shown in Figure 2. These solvents were chosen with regard to specific properties purported to have an effect on solubility. These properties include dielectric constant, solvent classification, hydrogen-bond donor ability, and hydrogen-bond acceptor ability and are listed for each solvent in Table 2. Of the four solvents chosen, “hydrotropic” behavior was observed only in the case of MeOH and NMF. This finding is significant, since solubilization of this type heretofore has been demonstrated only in water. However, because the term “hydrotropy” refers only to aqueous systems, a new term describing this process is necessary. We define “solutropy” as a phenomenon whereby addition of a large amount of a solid solute to any solvent causes an increase in the solubility of another slightly soluble solute. Using this definition, hydrotropy would be considered a specific case of solutropy in aqueous systems. We restrict the definition of the solutropic agent to include only solids in order to distinguish solutropic solubilization from solubilization by cosolvents. Table 3 shows the effect of NA on RFN solubility in all five solvents tested. Solutropic behavior is seen only in water, MeOH, and NMF. Solubility data in water, MeOH, NMF, and DMSO are depicted graphically in Figure 3 (RFN solubility in acetone is not great enough to be seen on this graph). It is evident that the addition of NA to DMSO in increasing concentrations causes a decrease in RFN solubility; that is, typical solution behavior is noted, while in the other three solvents, solutropic behavior is seen. From Table 2, one can easily see that there is no correlation between dielectric constant or solvent polarity and solutropic solvents. In fact, NA exhibits the greatest solutropic effect in water, which has a dielectric constant of 78.5, and a moderate effect in MeOH and NMF with dielectric constants of 32.6 and 182, respectively. However, typical solution behavior occurs in DMSO (dielectric constant ) 45) and acetone (dielectric constant ) 21.4). Moreover, the intrinsic solubility of the solubilized solvent is not related to solutropic effects. In DMSO and acetone (“typical” solvents) the intrinsic solubility of RFN is the highest and lowest, respectively. One can, however, find
Nicotinamide (M)
SFwater
SFMeOH
SFNMF
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
3.35 5.39 8.59 11.56 15.39 18.7 23.0 26.9 31.0 36.2
1.61 2.10 2.39 2.68 3.28 4.00 5.01 5.93 7.15 9.08
1.03 1.08 1.15 1.20 1.23 1.26 1.28 1.32 1.43 1.48
Solubility factor (SF) ) solubilitytotal/intrinsic solubility.
a correlation between hydrogen-bonding capabilities and the solubilizing effect. Although all of the solvents listed are capable of accepting hydrogens on the lone pairs of oxygen and/or nitrogen atoms, only DMSO and acetone are incapable of donating hydrogens for hydrogen-bond formation. Therefore, the ability of a solvent to be a hydrogen donor must be a key factor in the “hydrotropic” solubilization phenomenon. Additionally, one must consider the solvent-solvent interactions which may or may not occur. All of the solvents except DMSO and acetone have the dual ability to donate and accept hydrogens, which means that intermolecular hydrogen bonds between like solvent molecules can form. This type of intermolecular hydrogen bonding in solvents has been welldocumented for water and is known as the “iceberg” or “flickering cluster” model of water structure.20 The same type of solvent-solvent interactions would be expected to occur in MeOH, and in NMF, although probably not to the extent to which it occurs in water. MeOH can only donate one hydrogen, and according to Table 2, it does so less readily than water. Similarly, NMF has only one hydrogen to donate, which it does less freely than either MeOH or water. This trend is reflected in Table 4. Table 4 represents the solubility data in terms of the solubility factor. The solubility factor is the solubility of RFN in a NA solution of a given concentration divided by the intrinsic solubility of RFN in that particular solvent. The solubility factor normalizes solubility data with respect to intrinsic solubility. Table 4 shows that, at all concentrations of NA, water is a better solutropic solvent than is MeOH, which is better than NMF. This parallels the hydrogen-donating ability as shown in Table 2. Solubilization of Riboflavin by UreasIn order to test some of the theories proposed in the previous section, the ability of urea to solubilize RFN was studied. As mentioned in the Introduction, urea has demonstrated the ability to hydrotropically solubilize a few different compounds; however, RFN solubilization has not been documented. Solvent effects were also examined by looking at the effect of urea on RFN solubility in the same nonaqueous solvents used with NA. Figure 4 is a graph of RFN solubility as a function of increasing urea concentration in aqueous solution. From that graph, one can conclude that urea does indeed hydrotropically solubilize RFN. It is interesting to note that, unlike NA, which exhibits hydrotropic effects even at low concentrations,
Journal of Pharmaceutical Sciences / 953 Vol. 85, No. 9, September 1996
+
+
in water and MeOH, while in DMSO and acetone, typical solution behavior is seen.
Summary and Conclusions
Figure 4sEffect of urea on the solubility of RFN flavin in water. Data points represent the average of three experiments; error bars represent standard deviations.
Figure 5sComparison of the hydrotropic abilities of NA (squares) and urea (circles) for RFN. Data points represent the average of three experiments; error bars represent standard deviations.
Figure 6sEffect of urea on the solubility of RFN in water (circles), MeOH (squares), and DMSO (diamonds). Because solubility in DMSO is much higher, solubility is expressed as mg/mL.
significant increases in RFN solubility are not evident until the concentration of urea reaches approximately 0.8-1.0 M. A comparison of urea and NA as hydrotropes is seen in Figure 5. Clearly, NA is a much more effective hydrotrope, since only about one-third of the maximum solubility of RFN in 2.0 M solutions of NA is attained by urea at 6.0 M concentrations. Solubility data for urea in water, MeOH, and DMSO are shown in Figure 6. Acetone displayed typical solution behavior as increasing urea concentration resulted in a decrease in RFN solubility. The data presented in this figure confirm the theories set forth earlier with NA. Solutropic effects are seen 954 / Journal of Pharmaceutical Sciences Vol. 85, No. 9, September 1996
In order to ascertain the role of the solvent in solutropic solubilization, solvent-specific effects were explored by looking at solubilization of RFN by NA and urea in different solvents. The results of these experiments lead to the novel discovery that increases in solubility due to addition of another solid solute in high concentration occur in solvents other than water, namely, those solvents which are capable of both hydrogen donation and acceptance. It is proposed here that NA, urea, and other solutropic agents exert their solubilizing effect by changing the nature of the solvent, specifically by altering the solvent’s ability to participate in structure formation via intermolecular hydrogen bonding. The results of the experiments presented in this work may be summarized as follows: 1. NA, in concentrations of up to 2.0 M, is able to increase aqueous RFN solubility approximately 36-fold. There is no discrete concentration of NA which produces a marked increase in RFN solubility; rather, it is a continuous phenomenon seen even at very low NA concentrations. 2. Solubilization effects are seen only in solvents which can both donate and accept hydrogens in hydrogen-bond formation. Therefore, large concentrations of NA can increase RFN solubility in water, MeOH, and NMF, but normal solution behavior is seen in DMSO and acetone. 3. The same solvent-dependency was seen using urea, another solutropic agent that, until this time, had not been shown to solubilize RFN. 4. Comparing the solubilizing capabilities of NA and urea, NA was found to be the more powerful solutrope.
References and Notes 1. Ibrahim, S. A.; Ammar, H. O.; Kasem, A. A.; Abu-Zaid, S. S. Pharmazie 1979, 34, 809-812. 2. Frost, D. V. J. Am. Chem. Soc. 1947, 69, 1064-1065. 3. Hussain, M. A.; DiLuccio, R. L.; Maurin, M. B. J. Pharm. Sci. 1993, 82, 77-79. 4. Truelove, J.; Bawarshi-Nassar, R.; Chen, N. R.; Hussain, A. Int. J. Pharm. 1984, 19, 17-25. 5. Rasool, A. A.; Hussain, A. A.; Dittert, L. W. J. Pharm. Sci. 1991, 80, 387-393. 6. Chen, A. X.; Zito, S. W.; Nash, R. A. Pharm. Res. 1994, 11, 398401. 7. Kenley, R. A.; Jackson, S. E.; Winterle, J. S.; Shunko, Y.; Visor, G. C. J. Pharm. Sci. 1986, 75, 648-653. 8. Elsamaligy, M. S.; Hazma, Y. E.; Abd-Elgawad, N. A. Pharm. Ind. 1992, 54, 474-477. 9. Badwan, A. A.; El-Khordagui, L. K.; Khalil, S. A. Int. J. Pharm. 1983, 13, 67-74. 10. Ammar, H. O.; Ibrahim, S. A.; El-Faham, T. H. Pharm. Ind. 1981, 43, 292-295. 11. Altwein, D. M.; Delgado, J. W.; Cosgrove, F. P. J. Pharm. Sci. 1965, 54, 603-606. 12. Hussain, M. A.; DiLuccio, R. L.; Mauring, M. B. J. Pharm. Sci. 1993, 82, 77-79. 13. Hazma, Y. E.; Kata, M. Pharm. Ind. 1990, 52, 363-368. 14. Elsamaligy, M. S.; Hazma, Y. E.; Abd-Elgawad, N. A. Pharm. Ind. 1992, 54, 474-477. 15. Ammar, H. O.; Ibrahim, S. A.; El-Faham, T. H. Pharm. Ind. 1981, 43, 292-295. 16. Fawzi, M. B.; Davison, E.; Tute, M. S. J. Pharm. Sci. 1980, 69, 104-106. 17. The Merck Index, 8th ed.; Stecher, P. G., Ed.; Merck and Co., Inc.: Rahway, NJ, 1968. 18. Marcus, Y. Chem. Soc. Rev. 1993, 93, 409-416. 19. Nemethy, G.; Scheraga, H. A. J. Chem. Phys. 1962, 36, 33823400. 20. Coffman, R. E.; Kildsig, D. O. Results presented in part at the 9th Annual AAPS Research Meeting, San Diego, CA, 1994; PDD 7405.
JS960012B