Studying the Drug Delivery Kinetics of a Nanoporous Matrix Using a MIP-Based Thermal Sensing Platform

The implementation of Molecularly Imprinted Polymers (MIPs) into sensing systems has been demonstrated abundantly over the past few decades. In this article, a novel application for an MIP-based thermal sensing platform is introduced by using the sensor to characterize the drug release kinetics of a nanoporous silver-organic framework. This Ag nanoporous matrix was loaded with acetylsalicylic acid (aspirin) which was used as a model drug compound in this study. The drug elution properties were studied by placing the nanoporous matrix in phosphate buffered saline solution for two days and measuring the drug concentration at regular time intervals. To this extent, an acrylamide-based MIP was synthesized that was able to detect aspirin in a specific and selective manner. Rebinding of the template to the MIP was analyzed using a thermal sensor platform. The results illustrate that the addition of aspirin into the sensing chamber leads to a concentration-dependent increase in the phase shift of a thermal wave that propagates through the MIP-coated sensor chip. After constructing a dose-response curve, this system was used to study the drug release kinetics of the nanoporous matrix, clearly demonstrating that the metalorganic framework releases the drug steadily over the course of the first hour, after which the concentration reaches a plateau. These findings were further confirmed by UV–Visible spectroscopy, illustrating a similar time-dependent release in the same concentration range, which demonstrates that the MIP-based platform can indeed be used as a low-cost straightforward tool to assess the efficacy of drug delivery systems in a lab environment.


Introduction
Molecular imprinting technology originally focused on the development of imprinted particles that could be packed into columns for affinity separation, exploiting the affinity and selectivity to extract a molecule of interest from complex matrices [1][2][3]. The concept was soon extended by using molecularly imprinted polymers (MIPs) as antibody or enzyme mimics [4,5]. One of the most interesting applications for MIPs is their incorporation into biomimetic sensing devices as they mimic the affinity a natural receptor has for its target but are superior in terms of their

Instrumentation
Scanning electron microscopy (SEM) analysis of nanoporous matrices was performed with an SEM (Philips XL 30, M4I, Maastricht University, Maastricht, The Netherlands), at an operating voltage of 20 kV. Benchmarking with UV-Vis spectroscopy was done on a Shimadzu UV-3600 UV-VIS-NIR spectrophotometer (Shimadzu Europe, Duisburg Germany). Grinding and sieving of bulk MIPs were done using a Fritsch Planetary Micro Mill Pulverisette 7 premium line (VWR International, Amsterdam, The Netherlands) and a Fritsch Analysette (VWR International, Amsterdam, The Netherlands) with a 20 µm mesh, respectively.
The thermal detection platform used for studying elution has been described thoroughly in previous work [18][19][20][21][22]. Functionalized chips (MIP or NIP) were pressed mechanically with their backside onto a copper block serving as a heat provider. The temperature of the copper underneath the sample, T 1 , was monitored by a K-type thermocouple (TC Direct, Nederweert, The Netherlands). This information was fed into a temperature control unit that stringently controlled T 1 by modifying the voltage over the power resistor (Farnell, Utrecht, The Netherlands) that heated the copper, using a software-based (Labview, National Instruments, Austin, TX, USA) proportional-integral-derivative (PID) controller (P = 10, I = 8, D = 0). The functionalized side of the chip faced a polyether ether ketone (PEEK) flow cell which was sealed with an O-ring to avoid leakage, defining a contact area of 28 mm 2 and an inner volume of 110 µL. The flow cell was connected to a tubing system, allowing the administration of liquids in a controlled and automated fashion by means of a syringe pump. The temperature of the liquid inside the flow cell, T 2 , was measured by a second thermocouple, placed 1 mm above the chip. For each rebinding measurement the signal was stabilized in PBS at pH 7.4 which was used as to mimic physiological conditions.

Synthesis of the Silver Nanoporous Matrix
Nanoporous matrixes were synthesized by mixing aqueous AgNO 3 (5 mL, 0.4 M, 2 mmol) with aqueous NaOH (150 mL, 15 M, 2.25 mol). Crosslinking was initiated by the addition of aqueous EDA (1.5 mL, 99% w/v, 0.25 mmol) and aqueous hydrazine hydrate (0.2 mL, 80% w/v, 5.0 mmol) to the solution. The mixture was then purged with N 2 and refluxed at 80 • C for 90 minutes with continuous stirring at 1500 rpm. After cooling the reaction flask to room temperature, silver nanosponges were isolated by vacuum filtration and air dried before being freeze dried for 6 hours. The resulting sponges had a loading efficiency of 73.4% ± 5.6% and a corresponding loading capacity of 65.6% ± 2.9%.
It was then heated to 60 • C during 24 hours while shielding the mixture from light to prevent the template from degrading. The resulting monolith was mechanically ground (700 rpm, 5 min, 10 mm balls) and sieved for 4 with a 20 µm mesh. The resulting powder was extracted for 96 hours at 105 • C, using a Soxhlet apparatus (VWR International, Amsterdam, The Netherlands) filled with a mixture of methanol and ethanoic acid (7:3 v/v) in order to remove the template, aspiring, from the MIP. Finally, the MIP particles were dried for three hours in an oven at 50 • C. Non-imprinted polymers (NIPs), serving as a reference, were synthesized in the same manner without the presence of a template.

Chip Preparation
Polished aluminium plates were cut to obtain chips with the desired dimensions (10 × 10 mm 2 ). To immobilize MIP particles onto the surface of the measurement chip, a 100 nm PVC adhesive layer (0.35 wt % PVC dissolved in tetrahydrofuran) was applied onto the chip by spin coating at 3000 rpm for 60 s with an acceleration of 1100 rpm/s. MIP and NIP particles were stamped into this layer using a PDMS substrate that was covered with a monolayer of polymer particles. The PVC layer was heated for 2 h at a temperature of 100 • C-chosen specifically to be significantly above its glass transition temperature (80 • C)-which allows beads to sink into the polymer layer. The samples were cooled down prior to thermal measurements and any unbound particles were washed off with distilled water.

Loading of Nanoporous Matrixes with Aspirin
Aspirin was absorbed into the nanoporous matrixes by solvent evaporation. To this extent the nanosponge (0.100 g, 1.85 mmol), was incubated with aspirin (0.26 g, 1.85 mmol) in ethanol (6.2 mL). The mixture was shaken for 48 hours at 750 rpm. The solvent was removed under vacuum (30 • C, 300 mbar, 90 rpm) and dried at 65 • C for 3 h.

Drug Elution Analysis
Loaded nanoporous matrices (0.1 g) were incubated in 300 mL PBS (pH 7.4 at 37 • C) while gently stirring at 100 rpm to mimic physiological conditions. Over the course of two days, 3 mL aliquots of the PBS solution were taken at regular time intervals and the aspirin concentration was analyzed by both thermal wave transfer analysis (TWTA) and UV-Visible spectroscopy to create an elution profile.

Surface Characterization of Ag Nanosponges
The Ag nanosponges were analyzed using scanning electron microscopy (SEM). This analysis, shown in Figure 1, confirms that the metal-organic framework has a nanosponge structure containing a large set of nano-sized cavities, providing a large a surface area that can be loaded with drug molecules.

Quantification of Aspirin in PBS
To assess whether it was possible to accurately determine the concentration of drug eluted from the nanoporous matrix, a dose-response curve was constructed by exposing a MIP-coated chip to an increasing concentration of aspirin in PBS. The thermal analysis clearly indicates that exposing the MIP to aspirin in increasing concentrations results in a decrease of the liquid temperature inside the

Quantification of Aspirin in PBS
To assess whether it was possible to accurately determine the concentration of drug eluted from the nanoporous matrix, a dose-response curve was constructed by exposing a MIP-coated chip to an increasing concentration of aspirin in PBS. The thermal analysis clearly indicates that exposing the MIP to aspirin in increasing concentrations results in a decrease of the liquid temperature inside the flow chamber ( Figure 2a) and an increase in the phase shift observed in the transmitted wave ( Figure 2b). The results in Figure 2b were used to construct a thermal Bode plot which shows the phase shift for each transmitted frequency in function of the cummulative concentration of aspirin present in the flow cell ( Figure 2c). The error bars represent the maximum error observed on the measurement signal and was calculated by comparing a programmed temperature ramp with the actual temperature profile. Although the phase shift at every concentration is most pronounced at the highest input frequency, the resolution appears to be optimal at 0.03 Hz. These findings are in line with previously obtained results with dopamine MIPs in a similar setup [21]. The time-dependent TWTA data at 0.03 Hz were used to construct a dose-response curve (Figure 2d), which will be used to assess the concentration of aspirin that has eluted from the Ag-nanoporous matrix.

Quantification of Aspirin in PBS
To assess whether it was possible to accurately determine the concentration of drug eluted from the nanoporous matrix, a dose-response curve was constructed by exposing a MIP-coated chip to an increasing concentration of aspirin in PBS. The thermal analysis clearly indicates that exposing the MIP to aspirin in increasing concentrations results in a decrease of the liquid temperature inside the flow chamber ( Figure 2a) and an increase in the phase shift observed in the transmitted wave ( Figure  2b). The results in Figure 2b were used to construct a thermal Bode plot which shows the phase shift for each transmitted frequency in function of the cummulative concentration of aspirin present in the flow cell (Figure 2c). The error bars represent the maximum error observed on the measurement signal and was calculated by comparing a programmed temperature ramp with the actual temperature profile. Although the phase shift at every concentration is most pronounced at the highest input frequency, the resolution appears to be optimal at 0.03 Hz. These findings are in line with previously obtained results with dopamine MIPs in a similar setup [21]. The time-dependent TWTA data at 0.03 Hz were used to construct a dose-response curve (Figure 2d), which will be used to assess the concentration of aspirin that has eluted from the Ag-nanoporous matrix.

Selectivity Test
In order to assess whether the aspirin recognition was selective and specific, the experiment summarized in the previous section was repeated for a NIP-coated electrode. In addition, both MIP and NIP were exposed to an increasing concentration of acetaminophen (paracetamol). The resulting dose-response curves and the corresponding fits are shown in Figure 3.

Selectivity Test
In order to assess whether the aspirin recognition was selective and specific, the experiment summarized in the previous section was repeated for a NIP-coated electrode. In addition, both MIP Polymers 2017, 9, 560 6 of 10 and NIP were exposed to an increasing concentration of acetaminophen (paracetamol). The resulting dose-response curves and the corresponding fits are shown in Figure 3.
(d) are constructed from these data and are plotted in function of the cummulative concetration present in the flow chamber.

Selectivity Test
In order to assess whether the aspirin recognition was selective and specific, the experiment summarized in the previous section was repeated for a NIP-coated electrode. In addition, both MIP and NIP were exposed to an increasing concentration of acetaminophen (paracetamol). The resulting dose-response curves and the corresponding fits are shown in Figure 3. The results in Figure 3, illustrate that although the MIP is surprisingly selective in discriminating between paracetamol and aspirin, the imprinting factor is small. This conclusion is in line with previous findings with similar AA-based MIPs and can be explained by the fact that at a neutral pH, not all binding sites and functional groups on the target will be protonated (pKa of aspirin is 3.49) [36]. Although previous work has indicated that the MIP would be more specific at acidic pH, the authors decided to continue measuring at a pH of 7.4 to simulate drug elution in physiological The results in Figure 3, illustrate that although the MIP is surprisingly selective in discriminating between paracetamol and aspirin, the imprinting factor is small. This conclusion is in line with previous findings with similar AA-based MIPs and can be explained by the fact that at a neutral pH, not all binding sites and functional groups on the target will be protonated (pK a of aspirin is 3.49) [36]. Although previous work has indicated that the MIP would be more specific at acidic pH, the authors decided to continue measuring at a pH of 7.4 to simulate drug elution in physiological conditions. The results in Figure 2 indicate that the dose-response curve is highly usable and as the release pattern will be studied in PBS no interference from other molecules is to be expected. However, if the concept was extended to complex matrices in the future, the MIP synthesis route should be revised.

Thermal Analysis of Drug Elution
The elution of aspirin from the nanoporous matrixes was studied by incubating them in PBS and retrieving a sample from the surrounding medium after 1, 10, 30, 120 and 360 minutes and after 48 hours. The elutions were diluted 5000 times with PBS to fit the linear range of the sensor. MIP-coated electrodes were exposed to these diluted elutions and their response was summarized in a temperature Bode plot (Figure 4a). The resulting phase shifts at 0.03 Hz were used to construct an elution profile that was compared to the previously obtained dose response curve (Figure 4b) to determine the aspirin concentration in each of the eluted samples.
The results obtained in Figure 4 were corrected for the dilution factor to create an elution profile. To benchmark the results obtained with our platform, the results were validated using a gold standard, relatively low-cost reference technique, i.e., UV-Visible spectroscopy/The elution profiles of both techniques demonstrate a similar behavior (Figure 5a): a characteristic exponential release pattern (fitted in Origin 8.0, Originlab, Northampton, MA, USA) which contains a sharp increase within the first two hours after which a stable plateau is reached that does not significantly change over the next two days. This indicates that the nanoporous matrix releases the drugs in a relatively quick burst which is suitable for some applications requiring immediate effect. However, to actually achieve sustained, prolonged delivery of drugs the nanoporous matrix should be functionalized with molecules that bind the drug and actually release it slowly over time.

Thermal Analysis of Drug Elution
The elution of aspirin from the nanoporous matrixes was studied by incubating them in PBS and retrieving a sample from the surrounding medium after 1, 10, 30, 120 and 360 minutes and after 48 hours. The elutions were diluted 5000 times with PBS to fit the linear range of the sensor. MIP-coated electrodes were exposed to these diluted elutions and their response was summarized in a temperature Bode plot (Figure 4a). The resulting phase shifts at 0.03 Hz were used to construct an elution profile that was compared to the previously obtained dose response curve (Figure 4b) to determine the aspirin concentration in each of the eluted samples. The results obtained in Figure 4 were corrected for the dilution factor to create an elution profile. To benchmark the results obtained with our platform, the results were validated using a gold standard, relatively low-cost reference technique, i.e. UV-Visible spectroscopy/The elution profiles of both techniques demonstrate a similar behavior (Figure 5a): a characteristic exponential release pattern (fitted in Origin 8.0, Originlab, Northampton, MA, USA) which contains a sharp increase within the first two hours after which a stable plateau is reached that does not significantly change over the next two days. This indicates that the nanoporous matrix releases the drugs in a relatively quick burst which is suitable for some applications requiring immediate effect. However, to actually achieve sustained, prolonged delivery of drugs the nanoporous matrix should be functionalized with molecules that bind the drug and actually release it slowly over time.
When analyzing the elution profile obtained by UV-Visible spectroscopy, a small decrease in the concentration of aspirin can be observed over the course of two days, which leads to a distortion of the exponential fit. This can be explained by the fact that some of the acetylsalicylic acid will be converted into salicylic acid in PBS. This is confirmed by analyzing UV absorbance at 295.5 nm, which shows that salicylic acid is indeed present in the elution and its concentration will increase slightly over the course of two days (Figure 5b). The fact that this is not shown in the TWTA data is due to the fact that both compounds will bind to the MIP in a similar manner [36]. These data suggest that although the proposed sensor platform is not able to monitor the conversion of the drug to its metabolite, and is therefore less specific in this case, it is still a valuable alternative for the gold standard. Moreover, the application of UV-Visible spectroscopy for determining drug concentration in complex media can be troublesome due to the presence of other compounds that overlap the  absorption spectrum of the drug under study [37]. The current platform on the other hand, has already proven to be capable of identifying small molecules in complex media such as blood plasma [38]. In addition, recent research has shown that the MIP-based thermal setup could be integrated into catheter-based commercial sensing devices which would enable to study drug release in vivo which is not possible with the current state-of-the-art techniques [39].  Figure 4 were compared to the drug elution profile obtained with UV-Visible and both show a initial burst of aspirin release in the first two hours in the milimolar regime (a). The decrease in aspirin concentration for the UV-Visible data shown in Figure 5a can be explained by conversion of aspirin into salicylic acid which is confirmed by analyzing absorbance at 295.5 nm (b).

Conclusions
The data shown in this article illustrate the potential use of a MIP-based thermal detection platform, which has previously been used for diagnostic purposes, for analyzing the drug release kinetics of drug delivery matrices. A proof-of-principle was demonstrated by validating the results obtained with the biomimetic sensor using UV-Visible spectroscopy, demonstrating a similar profile in the same concentration range. The metalorganic framework synthesized in this work appears to release the model drug, aspirin, within the first two hours limiting its pharmaceutical use at this point. Therefore, future research should be aimed at functionalizing the framework to get to a more gradual release of the drug. In addition, its release profile will be studied in a more realistic setting to mimic sink conditions and potential degradation of the matrix to fully understand the pharmaceutical potential of diverse drug delivery systems. The proposed platform could be a very versatile tool for these release studies by providing valuable information on the performance of the drug release system. This information can in turn be used as a feedback loop to fine-tune the drug release framework's synthesis procedure to acquire drug release matrices with a high potential for pharmaceutical application. In addition, loading and elution of other, potentially more relevant drugs  Figure 4 were compared to the drug elution profile obtained with UV-Visible and both show a initial burst of aspirin release in the first two hours in the milimolar regime (a). The decrease in aspirin concentration for the UV-Visible data shown in Figure 5a can be explained by conversion of aspirin into salicylic acid which is confirmed by analyzing absorbance at 295.5 nm (b).
When analyzing the elution profile obtained by UV-Visible spectroscopy, a small decrease in the concentration of aspirin can be observed over the course of two days, which leads to a distortion of the exponential fit. This can be explained by the fact that some of the acetylsalicylic acid will be converted into salicylic acid in PBS. This is confirmed by analyzing UV absorbance at 295.5 nm, which shows that salicylic acid is indeed present in the elution and its concentration will increase slightly over the course of two days (Figure 5b). The fact that this is not shown in the TWTA data is due to the fact that both compounds will bind to the MIP in a similar manner [36]. These data suggest that although the proposed sensor platform is not able to monitor the conversion of the drug to its metabolite, and is therefore less specific in this case, it is still a valuable alternative for the gold standard. Moreover, the application of UV-Visible spectroscopy for determining drug concentration in complex media can Polymers 2017, 9, 560 8 of 10 be troublesome due to the presence of other compounds that overlap the absorption spectrum of the drug under study [37]. The current platform on the other hand, has already proven to be capable of identifying small molecules in complex media such as blood plasma [38]. In addition, recent research has shown that the MIP-based thermal setup could be integrated into catheter-based commercial sensing devices which would enable to study drug release in vivo which is not possible with the current state-of-the-art techniques [39].

Conclusions
The data shown in this article illustrate the potential use of a MIP-based thermal detection platform, which has previously been used for diagnostic purposes, for analyzing the drug release kinetics of drug delivery matrices. A proof-of-principle was demonstrated by validating the results obtained with the biomimetic sensor using UV-Visible spectroscopy, demonstrating a similar profile in the same concentration range. The metalorganic framework synthesized in this work appears to release the model drug, aspirin, within the first two hours limiting its pharmaceutical use at this point. Therefore, future research should be aimed at functionalizing the framework to get to a more gradual release of the drug. In addition, its release profile will be studied in a more realistic setting to mimic sink conditions and potential degradation of the matrix to fully understand the pharmaceutical potential of diverse drug delivery systems. The proposed platform could be a very versatile tool for these release studies by providing valuable information on the performance of the drug release system. This information can in turn be used as a feedback loop to fine-tune the drug release framework's synthesis procedure to acquire drug release matrices with a high potential for pharmaceutical application. In addition, loading and elution of other, potentially more relevant drugs should also be studied in more challenging, biological media. As the MIP platform is generic, it can be used to study a wide variety of targets in a wide variety of matrices, by changing the MIP receptors or optimizing their selectivity or performance in more challenging media and chemical conditions (pH, temperature, etc.).