In vitro kinetic release study of ketoprofen enantiomers from alginate metal complexes

To explore the release behavior of ketoprofen enantiomers from alginate-metal-complexes. Five mathematical models of drug release kinetics were investigated. Beads of alginate-metal complexes, loaded with racemic ketoprofen, were prepared by the ionotropic method. Divalent (Ca, Ba, Zn) and trivalent (Fe, Al) metals were used in the preparation of single-metal and mixed-metal alginate complexes. In vitro release experiments were carried out in an aqueous phosphate buffer medium at pH = 7.4. The concentrations of ketoprofen released enantiomers were determined using chiral HPLC technique. The obtained data were used to simulate the release kinetic of ketoprofen enantiomers using various mathematical models. The Korsmeyer-Peppas model was the best fit for Ca, Al, and Fe beads. Moreover, alginate-iron beads tend to release the drug faster than all other cases. In contrast, the drug release for alginate-barium complex was the slowest. The presence of barium in alginate mixed-metal complexes reduced ketoprofen release in the case of Fe and Zn, while it increased the release in the case of Al complex. In all the studied cases, ketoprofen showed very slow release for both enantiomers over a period exceeded 5 h (10 days in some cases). The release rate modification is possible using different multivalent metals, and it is also feasible by using two different metals for congealing either consecutively or simultaneously.

In our previous work [35], the prepared alginate metal complexes' beads were characterized in terms of size, metal content, drug content, and drug loading efficiency. Moreover, IR spectra were used to explore possible chiral interactions between the chiral drug and the chiral matrix. Therefore, chiral HPLC was exploited to monitor the probable enantioselective release of the drug from the chiral matrix. The HPLC results showed notable release selectivity in favor of S-enantiomer in the case of alginate-calcium complex.
This work aims to explore the effect of complexing multivalent metal ion on the release kinetics of ketoprofen enantiomers from alginate-metal beads. Thus, common mathematical models of drug release were investigated. Subsequently, a better understanding could result in using these complexes in sustained-release matrixes, or may be in enantioselective release matrices.

Methods and materials
Alginic acid sodium salt was purchased from brown algae BioReagent, suitable for immobilization of microorganisms from Sigma-Aldrich. (S)-Ketoprofen was from Sigma-Aldrich; assay 99%, calcium chloride, iron chloride, aluminum chloride, barium chloride dehydrate, and zinc chloride were obtained from Merck. Racemic ketoprofen (KTP) was from Infinity Laboratories Private Limited (Infinity, India) 99.6%. All other chemicals and reagents used were of analytical grade or HPLC grade.
Beads were prepared using ionotropic congealing method in which, racemic KTP was dissolved in sodium alginate aqueous solution PBS, pH = 7.4 to obtain 2% (weight/volume) alginate solution in a ratio of KTP to sodium alginate 1 to 3.75 (w/w). The obtained solution was then congealed in a bath contained a metal chloride (3% w/v) at room temperature. The formed beads were stirred continuously in the solution for 24 h, separated and washed with PBS and distilled water for 1 min; then, the beads were dried at 40°C for 48 h, and the described details were in [33]. However, dried beads were released in phosphate buffer solution (PBS) of pH = 7.4 to simulate that of gastric medium (6.8-7.4). Aliquots, at different time intervals, were taken, extracted, and analyzed using chiral HPLC to follow the KTP enantiomers concentrations in the release medium.
Optical microscope zoom 2000 from Leica microsystem use for imaging the surface of samples, Germany. Photos of dried beads were taken with × 40 magnification.

Drug release kinetics modeling
Zero-order, first-order, Higuchi, Hixon-Crowell and Korsmeyer-Peppas models were abundantly used in the literature to simulate ketoprofen release kinetics from various matrices [9-13, 15, 16, 20, 36]. These models described in Tables 1 and 2, and they will be exploited in this work to study the ketoprofen release from the prepared alginate metal complexes beads.

Beads' characteristics
A summary of beads characteristics, congealing conditions, bead size, and shape are shown in Table 3.
In vitro ketoprofen release study Release study of KTP in PBS (pH 7.4) Concentrations of released KTP enantiomers were determined quantitatively (μg/ml) in the withdrawn aliquots using chiral HPLC. The intent of using this analysis technique is observing each KTP enantiomers as enantioselective release is probable. This probability come from chiral interactions established between chiral drug and used chiral matrices [20,25,35]. However, chiral KTP analysis results are expressed in terms of enantiomeric excess ee% as showed in Table 4 for the first 2 h of the release experiment. The ee% is given by the expression ee% = │(R − S)/(R + S)│.
Before modeling the KTP release from the prepared beads, we defined release rate R L to be used in comparing between release behaviors of various beads. R L could be described as the concentration alteration C t over a specified time T, i.e., R L = C t /T (μg ml −1 h −1 ). The specified time T was chosen to be equals to the half-life time of KTP, i.e., 2 h. Bearing in mind that the total volume of release solution is conserved as the volume loss because of withdrawn aliquots was compensated by equal volume addition of PBS solution.
The KTP released quantity percentages, for all beads kinds, were plotted against time as depicted in Figs. 2, 3, 4, and 5. Typical HPLC chromatograms for different alginate beads are also presented in Figs. 6 and 7.

Modeling KTP release kinetics
The required data to model the release kinetics of KTP from the prepared single-metal alginate complexes were obtained and used to find the best-fit models. Table 6 summarizes the resulted correlation coefficient R 2 and the release kinetic constants corresponding to each studied model.

Discussion
The shape and size of beads varied depending on complexing ion metal; the divalent ion beads are smaller in size (5-7-mm diameter for wet and 2-3 mm for dry beads) than trivalent ion beads (7-10-mm diameter for wet, 2-3.5 mm for dry beads). The beads' shape was not a regular sphere; it is tear-like for Ca and Ba beads, full of sharp uplifting for Zn and Al beads, and middle-collapsed for Fe beads. All beads also have their characteristic convolutions and wrinkles. It may be probable that these differences in shape and morphology have an impact on the release behavior of the studied drug. Table 3 summarizes the characteristics of the formed beads.

Effect of complexing ion metal on the KTP enantiomers release behavior
In our previous work [35], we gave proof of chiral interactions between ketoprofen and alginate. These interactions, believed to lead to enantioselective release under certain conditions, are mainly due to hydrogen bindings between the carboxylic group of the drug and carboxylic and hydroxyl groups of alginate; similar interactions were proved between tiaprofenic acid and alginate [38]. During, the release experiment, water molecules will participate in hydrogen bonding while penetrating the beads networks. Subsequently, the prescribed chiral interactions will be less strong resulting in decreased ee% values. This logic applies to all the studied cases in Table 4 and there is no contradiction between the Fe case and the other cases. However, the tabulated ee% values can be understood as the following: ee% values marked with asterisk* are smaller or close to 1%, and they more likely indicate racemic release and not necessarily indicate an Table 1 description of some mathematical models used to determine drug release from drug delivery systems Zero-order Q 0 − Q t = K 0 t Q t is the drug released amount at time t; Q 0 is the initial concentration of drug at time. K 0 is the release constant of zero-order kinetic First-Order logQ r = log Q 0 − K 1 t/2.303 Q r is the drug remaining amount at time t, K 1 is the release constant of first-order kinetic.
Higuchi Q = K H *t 1/2 K H is the release constant of Higuchi Kinetic Hixson-Crowell Q 0 1/3 − R t 1/3 = K HC t R t is the drug remaining amount at time t; K HC is the release constant of Hixson-Crowell Kinetic.

Korsmeyer-Peppas
Q t /Q ∞ = K kp* t n Q t /Q ∞ is a drug released fraction at time t, n is the diffusional exponent or drug release exponent, and K kp is the release constant of Korsmeyer-Peppas Kinetic In Korsmeyer-Peppas equation, n characterizes different release models for different shaped matrices as summarized in Table 2 7 showing a more stable trend rather than an increasing one. Nonetheless, understanding the ee% results, hence the related enantioselective release process needs more understanding of many factors that can affect the process such as drug load and drug distribution within the bead, bead shape, and porosity [39].
The results of ABaK beads are not shown because of the very weak concentrations of the released enantiomers. Nonetheless, the ACaK beads showed the most dramatic changes in ee% results. Thus, curves shown in this study represent the sum of both released enantiomers.
As previously described [40,41], metals bind with different strength to alginate forming alginate-metal complexes with "egg-box" structures. The related strength follows the order Sr 2+ >Pb 2+ > Ca 2+ >Cd 2+ >Mg 2+ >Fe 2+ >-Fe 3+ >Co 2+ >Al 3+ . Hence, the strongest the complex the slowest dissociation, and subsequently the faster drug release from the complex. This explain the obtained results shown in Table 5, where R L values for single metal alginate beads follow the order AFeK>AAlK>AZn-K>ACaK>ABaK. Similar results were also obtained for the drug quantity released after 30 min. This quantity, Q 30 , was considered the onset release quantity (μg), Table 4 Obtained ee% results for 120 min in the release medium according to which the studied complexes followed the same order as above AFeK>AAlK>AZnK>ACaK>ABaK. Generally, the onset release concentrations and the release rates were higher in the case of trivalent metal complexes compared to those of divalent metal complexes. Thus, this could be because the studied divalent metals form stronger complexes with alginate compared to those of the studied trivalent metals.

Effect of congealing sequence with two ion metals on the KTP enantiomers release behavior
In the case of co-congealing with Ca and Fe, the simultaneous congealed beads A 1 K have the slowest rate compared to those of consecutive congealed A 2 K (Ca then Fe) and A 3 K (Fe then Ca). The relevant release curves are depicted in Fig. 3. An interesting difference between A 1 K and A 3 K behaviors was noticed; over  the first 50 min (approx) A 3 K showed slower release than A 1 K but after that period the release rate of A 3 K became the faster. Thus, this indicates an obvious effect of altering congealing sequences on the related release behaviors. According to R L values of the mixed metals' beads prepared with Ba and other metal as described in Table 3, beads can be ordered as ABaFeK>ABaAlK>ABaZnK>A-BaCaK. However, the related curves of cumulative released drug percentage, shown in Fig. 4, indicated similar trends for ABaCaK and ABaZnK beads (Ba mixed with divalent metals) different from those of ABa-FeK and ABaAlK beads (Ba mixed with trivalent metals). This again shows the different behavior of the complexes with trivalent metals from those of divalent metals. Thus, this was due to the differences in binding strengths and the resulted network structures (egg-box structures). ABaFeK and ABaAlK beads have similar R L ; the same point at 2 h as in Fig. 4, but they have different curves over the experiment time of 6 h (Fig. 4) and also for 10 days as shown in Fig. 5.
Additionally, during the experiment time for all kinds of these beads, cumulative release curves showed much less than 1% released drug. The curves also did not show flat or decreasing parts, i.e., KTP concentrations did not reach the maximum or the equilibrium state in the release medium. Thus, this indicated sustainability over the experiment time; however, a significant release drug has not been accomplished yet.

The KTP release model
Before trying to interpret the results, it is important to recall that the studied beads have different shapes and morphologies. Enantiomers release kinetics were modeled for single-metal alginate complexes. However, Table 6 contains the obtained results for each released enantiomer and the sum of released enantiomers.  Moreover, the simulated data resemble only the release profile of a very small fraction of the loaded KTP. This fraction did not arrive at 1% even after 10 days of release time. From this point of view, the best-fit model derived from this data is tricky and may not be comparable with other models of higher released fractions.
The best-fit model for resulted data of ACaK beads, for the total released KTP, and for each enantiomer R and S, was Korsmeyer-Peppas model as it has the highest correlation coefficient values. For ACaK beads, the release showed super case-II transport indicating more than one release mechanism (n > 0.85, it is 1.55, 1.56 and 1.53 for ACaK, ACaK-R, and ACaK-S respectively). Thus, the observed enantioselective release did not result from different release kinetics of each enantiomer. Simulating KTP release from AZnK beads did not result in a clear one best-fit model; Zero-order, First-order, and Hixon-Crowell models are quasi-equally probable as the related R 2 values are comparable. From first glance, this is unreasonable; the obtained data cannot follow three different models. However, to explain this, it is fair to remember that models can be similar to a short-range of data. The studied cases herein represent less than 1% release. However, the release obeyed the Korsmeyer-Peppas model with an anomalous mechanism in the case of AAlK (n = 0.46) and AFeK (n = 0.44) beads, for each enantiomer and the sum of enantiomers.
It is noteworthy that a complete KTP release was achieved from alginate-calcium beads after 3 h and no enantioselective release was reported. However, the mentioned beads were prepared in different conditions. They were congealed in 0.1 M solution of CaCl 2 within 12 h, dried 24 h at RT and subsequent 24 h at 45°C [42]. Comparing these conditions with our conditions indicates that increasing the concentration of calcium ions will decrease drug release. On the other hand, increasing the congealing time will also decrease drug release.

Conclusion
Alginate-metal complexes, reported in this work, caused weak KTP release concentrations, for both enantiomers, over a long period of time (10 days in several cases). The Korsmeyer-Peppas model was the most convenient fit in the case of Ca (for both enantiomers), Al, and Fe beads. More efforts are needed to find a better kinetic model that take into consideration the shape deviation from spheres, especially in the case of iron complexes. In all the studied cases, KTP enantiomers showed the highest rate for Fe beads and the smallest for Ba ones. The release rate modification is possible using different multivalent metals, and it is also feasible by using two different metals for congealing either consecutively or simultaneously. These results suggest alginate-metal complexes as a possible precursor for KTP enantiomers sustained release systems.