5 RESULTS AND DISCUSSION
5.5 RELEASE OF MODEL COMPOUNDS FROM ION-EXCHANGE FIBERS
5.5.1 Lipophilicity of bound compounds (I-IV)
The rate and extent of model compounds release from the Smopex DS-218v fiber decreased with growing lipophilicity of the compound, being clearly the lowest with 5-Br and 3-IPSA (Figure 11; Table 4 in paper I). The effect of lipophilicity was seen more clearly in dilute extraction solutions. Similarly, the release of dually loaded 3-IPSA was consistently hindered from all the fibers relative to SA (Figure 15). As assumed, the lipophilic nature of the compound was even more important when releasing the salicylates from the weakly basic and/or poly(ethylene) based fibers; the extent of release of 5-Cl was clearly smaller from the Smopex®-103pe, Smopex®-105pe and Smopex®-105v fibers compared to the release of more hydrophilic SA and 5-OH (Figure 1 in paper II). The compounds of higher lipophilicity were more strongly bound to the fibers as compared to the more hydrophilic compounds,i.e. Kchp,i was higher. Figure 12 illustrates the possible interactions formed between the compounds and the Smopex DS-218v fiber, including the hydrophobic interactions. Previous studies with cation-exchange fibers have indicated that lipophilic drugs interact especially strongly with a strong cation-exchange fiber, Smopex® -101, that contained a lipophilic poly(ethylene) framework (Jaskari et al., 2001).
Figure 11. The release profiles of salicylic acid (— —), 5-hydroxysalicylic acid (— —), 3-isopropylsalicylic acid (— —) and 5-hydroxyisophthalic acid (— —) from the Smopex DS-218v fiber. The molar amount of chloride-ions was equal (left) or ten-fold (right) as compared to the compound bound in the fiber.
When comparing the release efficiencies of a drug-sized compound and a small inorganic ion into the equal external solutions, it was observed, that the cumulative release of SA from the Smopex®-103pe fiber was significantly reduced compared to the release of chloride (III). This is consistent with the rather high estimated value of chemical
(log Poct 3.79) that the cation-exchange microspheres loaded by the above drugs were more hydrophobic as compared to the microspheres loaded by sodium-ions, which resulted in reduced swelling ratio of the microspheres in water and, thus, lower release rates of both the drugs from the ionic microspheres.
Figure 12. The structure and functionality of the Smopex DS-218v anion-exchange fiber.
The extents of SA, 5-F and 5-CH3 release from the Smopex DS-218v fiber were observed to be higher (> 10%) into the most dilute chloride solution (1/10 [n-]) than possibly could have been obtained by ion-exchange only (Table 4 in paper I). This supports the conclusion that incorporation of compounds to the fibers may be partly due to the pure non-specific interactions without ionic bonding, as has been discussed also previously in chapter 5.4.1 in relation to the relative occupancies over 100% (Table 7)(II, III, IV). The more the non-specific interactions are formed, the more difficult is the accurate control of drug release profiles from the fibers by ion-exchange approach.
5.5.2 Hydrogen bonding (I, II)
Numerous oxygen and hydroxyl groups in the viscose framework (Smopex®-105v and Smopex DS-218v) and/or nitrogen atoms of the unionised vinylpyridine groups (Smopex®-105v and Smopex®-105pe) are prone to function as hydrogen bond acceptors and/or donors with compounds such as SA, containing one hydroxyl group and two oxygen atoms (in the ionised carboxyl group), capable of forming H-bonds (see 5.4.4) (Figure 12). Of the used monovalent compounds, the structures of 5-NH2 and 5-OH possess strongly electronegative substituents in 5-position; an amine in 5-NH2 and an additional hydroxyl group in 5-OH, thereby surpassing the others in their hydrogen bonding potential. The fact that the release of 5-NH2 and 5-OH from the H-bond forming fibers was lower than that of SA, suggests, that their release behaviour was affected by the additional hydrogen bonding (Figure 11; Table 4 in paper I, Figure 1 in paper II).
Compound binding additionally via hydrogen bonds produce a stronger interaction
O
between the 5-NH2/5-OH and the fiber (i.e. Kchp,i is higher, corresponding to hydrophobic interactions) and, thus, restricts the compound release. In the case of 5-NH2, non-sink-conditions during the release experiments, may have slightly contributed to the extent of 5-NH2 release.
5.5.3 Valence of bound compounds (I, III, IV)
The release of di-COOH was consistently the lowest of all the compounds studied across the used fibers and external conditions (Table 11, Figures 11, 13 and 14; Table 4 and Figure 3 in paper I). The same phenomenon was consistent across subsequent equilibration stages. The electrostatic interaction between the bound divalent di-COOH and the ion-exchanger is stronger, as compared with the monovalent compounds, due to the dual site binding (Kunin and Myers, 1947; Sawaya et al., 1987; Bhandari et al., 1993;
Helfferich, 1995; Liu et al., 2001). The formed strong interaction hinders the release of di-COOH from the fibers, especially in the case of monovalent extracting electrolyte, by preferring the divalent di-COOH over the chloride ion (electroselectivity) (Table 11, Figures 12and 13). The release of dually bound di-COOH requires breakdown of both the ionic bonds before being released into the surrounding solution, i.e. requires effectively two chloride-ions in order to be released from the fiber, while the monovalent salicylates are capable of forming only one ionic bond (Liu et al., 2001). Similarly, the exchange of sulfate-ions from the ion-exchange resin by chloride-ions was unfavoured and yielded one half of the reverse reaction (Kunin and McGarvey, 1949; Kunin and Myers, 1947). When 1[n ] sulfate and oxalate solutions were used, the release of di-COOH was about equal compared to the release of monovalent SA into the 1[n ] chloride solution due to the charge equalization (Table 11).
By binding to the fiber via both the ionic groups, di-COOH may partly cross-link the system, e.g. when binding with anion-exchange groups from two separate fiber strands.
The cross-linked structures potentially formed may partly prevent the binding/release of di-COOH by hiding a portion of ion-exchange groups and, thus, hinder the diffusion of exchanging ions in the vicinity of fixed ionic groups (i.e. the effective capacity < number of ionisable groups in the fiber). The rate and extent of drug release has been observed to be much higher from the non-cross-linked ion-exchange fiber (staple fiber > fiber cloth) than from the cross-linked ion-exchange resin or gel (Vuorio et al., 2003). Ion-dipole type bonds and H-bonds may also form between the di-COOH and the fiber, which may further increase both cross-linking and attraction towards the ion-exchange fiber.
Jaskari et al. (2001) illustrated that the effective charge of the fiber was reduced due to the strong association of divalent calcium-ions with the fixed carboxylate cation-exchange groups, which led to a reversal of the sign of effective charge when the amount of calcium-ions was sufficiently high. A similar phenomenon is possible in the present study.
and, as a consequence of adequate loading of di-COOH, the fiber behaves like a cation-exchanger that does not release the divalent di-COOH-anions (Helfferich, 1995).
Table 11. The amount (%) of salicylic acid (SA) and 5-hydroxyisophthalic acid (di-COOH) released from Smopex DS-218v and Smopex®-103pe fibers at 24 hours and with extracting electrolyte solutions of different valence. The molar amount of extracting electrolyte was one third (1/3[n ]), half (1/2[n ]), equal (1[n ]), ten-fold (10[n ]) or hundred-fold (100[n ]) as compared to the molar amount of compound loaded in the fibers (mean ± sd, n = 3; N.D. = not determined).
5.5.4 Effect of ion-exchange groups and capacity of fibers (I-III)
The interaction between the bound salicylate and the ion-exchange fiber was affected by both the chemical nature and the number of the fixed cationic groups of the fiber. A clearly more efficient release of model salicylates was seen from the viscose fiber containing trimethylammonium groups as compared to vinylpyridines (Figure 1 in paper II). On the contrary, when the fiber framework was poly(ethylene), generally lower release efficiency was observed from the strongly basic fiber. The strength of specific
SA di-COOH SA di-COOH
sodium chloride
1 [n-] 45.0 ± 0.8 25.3 ± 2.1 22.2 ± 0.6 14.7 ± 0.2
10 [n-] 84.5 ± 1.3 47.7 ± 0.8 N.D. N.D.
100 [n-] 85.5 ± 0.3 87.2 ± 0.9 N.D. N.D.
calcium dichloride
1/2 [n-] 40.7 ± 0.7 N.D. N.D. N.D.
1 [n-] 50.9 ± 0.2 N.D. N.D. N.D.
disodium sulphate
1/2 [n-] 73.6 ± 0.8 33.4 ± 0.9 N.D. N.D.
1 [n-] 77.9 ± 0.4 40.9 ± 0.9 33.2 ± 0.4 21.7 ± 0.1
10 [n-] 83.3 ± 1.1 68.1 ± 0.6 N.D. N.D.
disodium oxalate
1/2 [n-] 71.6 ± 0.5 38.7 ± 1.1 N.D. N.D.
1 [n-] 88.8 ± 1.6 45.8 ± 1.0 N.D. N.D.
10 [n-] 97.9 ± 0.7 75.9 ± 1.4 N.D. N.D.
trisodium citrate
1/3 [n-] 85.2 ± 5.0 49.0 ± 0.6 N.D. N.D.
1 [n-] 83.4 ± 4.7 63.0 ± 0.8 47.8 ± 0.6 35.8 ± 1.0
10 [n-] 90.8 ± 0.2 81.4 ± 1.4 N.D. N.D.
Smopex DS-218v Smopex®-103pe
External electrolyte;
the molar amount of electrolyte compared to salicylate loaded to fiber
interactions between the ion-exchanger and the drug has previously been found to differ with strong and weak cation-exchange materials in a compound and fiber/resin specific manner (Jaskari et al., 2001; Sprockel and Price, 1989). Cation-exchangers consisting of a lipophilic poly(ethylene) framework released lipophilic drugs more easily from the weak than from the strong ion-exchangers. The binding between strong sulphonic acid groups and lipophilic drugs (propranolol and tacrine, log Poct 3.2) were stronger than the interactions with weak carboxylic acid groups, while the opposite was true with a hydrophilic drug (nadolol, log Poct 0.9) (Jaskari et al., 2001). However, the release profiles of metoprolol (log Poct 1.9) from these same fibers were fairly similar (Vuorio et al., 2003).
When the amount of chloride in the extracting solution was relative to the amount of bound compound (1[n-] or 10[n-]), similar fractions of the compound were released from the Smopex DS-218v fiber regardless of different ion-exchange capacities. At the same time, the molar amount of compound loaded into the fiber was higher and the molar amount of compound release was increased at increasing fiber capacity (I-III). Equal observations about the ion-exchange capacity have been reported by Jaskari et al. (2001) and Conaghey et al. (1998b) with ion-exchange fibers/resins.
The ion-exchange mechanism has been observed to play a more prominent role in the release of salicylates from the strong base fibers as compared to the weak base anion-exchangers, as the 100 fold increase of extracting chloride concentration (from 1/10[n ] to 10[n ]) resulted in a clearly more effective release of the salicylates from the strong base anion-exchange fibers (8-11 fold) than from the weak base fibers (2-5 fold) (II). The incomplete ionisation of the vinylpyridine groups (pKa ~ 5-6) during the experiments may increase the contribution of hydrophobic interactions in compound binding/release behaviour and, moreover, decrease of the fiber charge will decrease the Donnan potential between the phases (Figure 8A).
5.5.5 Valence of external counter- and co-ions (III)
In general, the higher the valence of extracting counter-ion (citrate > sulfate > chloride), the more efficient was the release of SA and di-COOH from the strongly basic anion-exchange fibers (Table 11, Figure 13), as suggested by the lowered Donnan potential (Figure 6A) and modelled fraction released (Figure 7). In the extracting electrolyte solutions containing at most an equimolar amount of counter-ions with regards to the bound SA-ions (i.e. 1[n ] solutions), the release of SA from the Smopex DS-218v fiber was clearly higher with higher valence of counter-ion (Table 11). Instead, at higher concentrations, the differences in the exchange potentials with ions of different valence diminished in the case of SA, as predicted by modelling (Figure 6). With di-COOH, the effect of valence of the extracting electrolyte on the compound release was seen even more clearly than with SA (Table 11,Figure 13).
Figure 13. The release of salicylic acid (A, C) and 5-hydroxyisophthalic acid (B, D) from Smopex DS-218v (A, B) and Smopex®-103pe (C, D) fibers. The extracting electrolytes were chloride ( ), sulfate ( ) and citrate ( ) at 1[n ] concentrations (mean ± sd, n = 3-5).
At higher valence, the extracting counter-ions are more strongly attracted to the ion-exchanger (electroselectivity) (Bhandari et al., 1993; Helfferich, 1995; Jaskari et al., 2001;
Kankkunen et al., 2002b). Multivalent ions can simultaneously bind to several ion-exchange groups in the ion-ion-exchange material, divalent with two and trivalent with three ionic groups, releasing, therefore, two/three molecules of monovalent compounds (Bhandari et al., 1993; Liu et al., 2001). The effect of electroselectivity can also be observed when fraction released was modelled in Figure 7. The pronounced effect of multivalent counter-ions on the drug release has been observed in previous studies with cation-exchangers (Sawaya et al., 1988; Jaskari et al., 2001; Liu et al., 2001). CaCl2 and MgCl2 solutions, at equivalent ionic concentrations to NaCl, caused steeper release profiles of doxorubicin from albumin microcapsules and increased clearly the extent of drug release compared to the NaCl solution (Sawaya et al., 1988). Liu et al. (2001) observed that the addition of divalent cations (Ca2+) in a sodium chloride solution resulted in clearly higher rates of doxorubicin release from sulfopropyl dextran cation-exchange microspheres than a pure sodium chloride solution. Similarly, Jaskari et al. (2001) observed that as little as 10% of CaCl2 in a NaCl solution increased clearly the drug release from cation-exchange fibers. The effect of the divalent calcium-ions was higher in a weak carboxylic acid ion-exchange fiber than in a strong sulfonic acid fiber.
Time (h)
Of the divalent extracting electrolytes studied, oxalate was more effective than sulfate in releasing SA and di-COOH from the Smopex DS-218v fiber (Table 11). In the case of SA, oxalate was, at high concentrations, even more effective than the trivalent citrate, in spite of the fact that 86% of citrate contained three negative charges during the experiments. Association of sulfate- and oxalate-ions with sodium-ions (co-ion) may be different, which may explain the observed variation in their extraction efficiency, as the ion-exchangers prefer extracting anions that are less strongly associated to their co-ions (Helfferich, 1995). The concentration of the extracting electrolyte solution may also influence the selectivity of the ion-exchanger. This is demonstrated inFigures 6and 7 that show diminished difference between extracting salt effects at higher salt concentration, irrespective of valence. Sawaya et al. (1988) observed that the selectivity of cation-exchanger was different with different divalent cations. The release of doxorubicin from albumin microcapsules was faster and the extent of release was higher with magnesium-ions (Mg2+) compared to calcium-ions (Ca2+).
When studying the effect of valence of co-ion, the release of SA was observed to be only 4% smaller into the 1/2[n ] CaCl2 solution than into the 1[n ] NaCl solution (Table 11), the former containing half the molar concentration of highly excluded Ca2+-cations compared to Na+, while the amount of Cl -ions was equal. In line with the co- and counter-ion effects modelled in Figure 6, the release of SA was only 6% higher into the 1[n ] CaCl2 solution than into the corresponding NaCl solution (1[n ]), despite twice the amount of extracting Cl -ions.
5.5.6 Concentration of external counter-ions (I-III)
First equilibrium stage (24 h)
An increase in the surrounding counter-ion concentration improved consistently the release (rate and extent) of compounds from the fibers by lowering the Donnan potential between the external and fiber phases (i.e. the interaction between the bound compound and the fiber is weaker) (Table 11, Figure 14; Table 4 and Figure 3 in paper I; II), consistent with the theory (Figures 6A and 7) and previous observations with other ion-exchange fibers and resins (Irwin et al., 1987; Conaghey et al., 1998a; Jaskari et al., 2001;
Kankkunen et al., 2002b). Accordingly, the exchange of chloride was increased when its concentration in the extracting solution was increased (III). At a high external salt concentration, the concentration gradient between the phases is greater and also the probability of interchange of chloride-ion and ionic compound is increased, resulting in a more efficient release of the compounds (Jaskari et al., 2000, 2001; Liu et al., 2001). The favourable effect of counter-ion concentration on compound release was observed more clearly with divalent di-COOH compared to the monovalent compounds (Table 11, Figure 14).
Figure 14. The release profiles of salicylic acid (A) and 5-hydroxyisophthalic acid (B) from the Smopex DS-218v fiber during 24 h (first equilibrium stage) (mean ± sd, n = 3). The amount of extracting chloride-ions was equal ( , 1[n ]), 10 fold ( , 10[n ]) or 100 fold ( , 100[n ]) to the amount of compounds loaded to the fibers.
In the dilute external chloride solution (1/10[n-], ~ 1 mM), the Donnan potential of the system is high and, therefore, the electrostatic interaction between the compound and the fiber is strong, hindering the release of compounds from the fibers (I). However, the release of compounds into pure deionised water was clearly smaller as compared to the 1/10[n-] chloride solution, implying that a supply of ions suitable for the ion-exchange is necessary for the release of a compound from the fiber. The results support the previous findings of good controllability of drug release from the ion-exchangers (Conaghey et al., 1998a; Sriwongjanya and Bodmeier, 1998; Jaskari et al., 2000, 2001).
Later equilibrium stages
The external release medium was removed and replaced by the fresh external solutions (e.g. chloride, sulphate, citrate) at 24, 48, 72, 96 and 168 hours after the experiments were started (equilibrium stages 1-5, respectively). Clearly higher cumulative amounts of anionic compounds were released from the fibers gradually by all the external solutions used (Figure 3 in paper I, Figure 1 in paper II, III). As earlier discussed, the Donnan potential accomplishes the interchange of external-ions and anionic model compounds between the two phases until an equilibrium stage (Donnan equilibrium) is reached. When the release medium was changed, a new equilibrium stage was reached. The amount of counter-ions was insufficient to release all the compound bound to the fiber at once. Liu et al. (2001) and Jaskari et al. (2001) have also observed that a fresh medium breaks down the old equilibrium between the ion-exchanger and external solution by removing the released drug and by introducing more counter-ions capable of interchange with the remaining drug-ions in the ion-exchanger.
The cumulative release profiles of the compounds from the Smopex DS-218v fiber were clearly different in the external solutions containing different concentrations of chloride-ions (Figure 3 in paper I). At the most dilute Cl--concentration (1/10[n-]), the cumulative equilibration profiles of the compounds were practically linear; similar amount
Time (h)
0 5 10 15 20 25
Released amount (%)
0 20 40 60 80 100
A
Time (h)
0 5 10 15 20 25
Released amount (%)
0 20 40 60 80 100
B
of compounds were released during each equilibration. Instead, as the chloride-ion concentration was equal to the molar amount of compounds in the fiber (1[n-]), the extent of compound release from the fiber started to decline as a function of the equilibrium stages, obviously because the amount of compound remaining in the fiber diminished markedly. At each equilibrium stage the percentage of the compound, that was released from the ion-exchange fiber in the 1[n-]-solution, remained constant with all the monovalent compounds studied, although the amount of compound in the fiber was decreased after each equilibrium stage. Earlier studies with fibers and resins have also shown that the amount of drug release from the ion-exchange material is diminished when the drug loading in the ion-exchanger is smaller, but the fraction (percentage) of drug release remains the same with the same drug and ion-exchanger (Conaghey et al., 1998b;
Jaskari et al., 2001). With other fibers and 1[n-] chloride solutions, the cumulative release profiles of monovalent salicylates across the five equilibrium stages were linear or curved, depending on the release efficiency (Figure 1 in paper II). When the highest amounts of chloride-ions were used (10[n-] and 100[n-]), the release of SA from the Smopex DS-218v fiber was similar during the first equilibration (24 h) (Table 11) and a complete release of monovalent compounds was achieved by the second equilibrium stage (48 h) (Figure 3 in paperI), demonstrating that even a significant excess of extracting monovalent anions (or multivalent anions) was not able to release SA completely at once.
5.5.7 Dual loading (IV)
Utilisation of the ion-exchange fibers for simultaneous delivery of two drugs in combination was estimated by assessing the effects of compound lipophilicity and valence on their relative affinity towards the fibrous ion-exchangers by using structurally similar model compounds (SA, 3-IPSA, di-COOH). As assumed, compound with higher lipophilicity or valence had stronger attraction towards the fibers (see 5.4.1 and 5.4.2).
Because of the lower Donnan potential under the dual loading conditions with either 3-IPSA or di-COOH, the SA loading content of the ion-exchange fibers was clearly smaller compared to the fibers loaded in equal molar amounts of SA only (Tables 8 and 10). Also the competitive binding of compounds presenting higher affinity towards the ion-exchanger (here 3-IPSA or di-COOH) restricted the binding of the other compound (here SA), especially in the cases of fibers having low ion-exchange capacity relative to compound content in loading solution. In fact, it was observed that during a second loading period none of the SA was bound into the Smopex DS-218v fiber, while the binding of 3-IPSA and di-COOH was increased further. If all the ionic groups of the fiber will be occupied during the loading (i.e. 100% relative occupancy is reached), the more lipophilic/multivalent compound may even operate as a releasing electrolyte with regards to the more hydrophilic/monovalent compound by replacing it in the ion-exchange groups.
Consistent with release studies after single loadings, the release of SA from the fibers
molar amount of SA release was higher from the fibers that were loaded in pure SA solution compared to the dually loaded fibers, as the SA loading content of the former was higher. Instead, the relative fraction of SA release from the fibers remained constant
molar amount of SA release was higher from the fibers that were loaded in pure SA solution compared to the dually loaded fibers, as the SA loading content of the former was higher. Instead, the relative fraction of SA release from the fibers remained constant