pH/temperature-sensitive hydrogel-based molecularly imprinted polymers (hydroMIPs) for drug delivery by frontal polymerization

Abstract

Frontal polymerization was successfully utilized, for the first time, to obtain pH/temperature-sensitive hydrogel-based molecularly imprinted polymers (hydroMIPs). The polymerization reaction can be completed in 20 min. By this approach, hydroMIPs were synthesized using a mixture of gatifloxacin (template), N-isopropylacrylamide, acrylic acid and N,N′-methylenebisacrylamide. The influence of template concentration and the crosslinker amount on the imprinting effect of the hydroMIPs was investigated. The textural and morphological parameters of the hydroMIPs were also characterized by mercury intrusion porosimetry, DSC, and scanning electron microscopy. The gatifloxacin-imprinted polymer formed in frontal polymerization displayed high affinity than conventional thermal polymerization. The hydroMIPs were further evaluated as drug delivery devices and showed higher relative bioavailability than the corresponding non-imprinted polymer and hydroMIPs based on bulk polymerization from the results of the evaluation in vivo. The results suggest that frontal polymerization provides an alternative method to prepare hydroMIPs and may open up new perspectives in the field of MIPs.

---

Introduction

Hydrogel-based molecularly imprinted polymers (hydroMIPs) with stimuli-sensitive recognition toward target (template) molecules can reversibly swell and shrink in response to environmental changes.^1,2 HydroMIPs display a recognition ability to the template reversibly in the dense state (shrunken gels) that is activated by temperature, pH, and enzymatic activities at the diseased sites.³ Thus, hydroMIPs are materials to integrate MIP and stimuli-responsive hydrogel with novel functions to be generated. For example, when used as a catalyst, the function of catalysis of these intelligent hydrogels can be switched on and off by external stimuli;⁴ target molecule-responsive hydroMIPs revealed greater shrinkage than non-imprinted hydrogels⁵ and the swelling of MIPs can lead to drug release in a controlled manner.^6,7 In addition, imprinting in hydrogels causes significant improvements in affinity, capacity, and selectivity over conventional hydrogels for many templates, such as ions, small and moderate molecular weight molecules, proteins, viruses, DNA, and cells.⁸

Molecular imprinting involves the creation of macromolecular memory for a template molecule within a polymer network.⁹ However, imprinting within hydrogels is more complex than imprinting within rigid structures. In contrast to the traditional MIP with a rigid matrix, the relative position of the functional groups of the sensitive hydroMIPs can be controlled via the stimuli-response. At a given crosslinking density in aqueous solvent, an imprinted network that contains more hydrophilic moieties along the polymer backbone will tend to swell or expand more than gels containing hydrophobic groups, which will try to minimize their exposure to aqueous solvent. Since the first synthetic polyacrylamide matrix hydroMIPs were introduced by Tanaka and coworkers in 1999,¹⁰ the optimization of imprinting protocols for a number of biological molecules of significance has been successfully achieved.^11–13 All structural considerations controlling architecture and limiting the expansion or collapse of polymer chains, including the number and diversity of functional monomeric species and the strength of monomer–template interaction have been demonstrated to influence template binding and control the size of the imprinted cavities. Recently, polymerization mode, e.g., living polymerization strategies was demonstrated to enhance binding parameters of imprinted networks significantly.¹⁴

Frontal polymerization (FP) is an alternative synthetic technique for polymers, which allows the conversion of monomer into polymer in a localized reaction zone that propagates in an unstirred medium, able to self-sustain and propagate throughout the monomeric mixture.^15,16 Compared to batch polymerization (BP), FP has a simpler reaction route, shorter time, and lower energy consumption. FPs have created great expectations because their self-propagating nature can provide polymer with high conversion rates at short reaction times. Furthermore, the layer mode provides a homogeneity enhancement of polymer chains. Importantly, the FP can be a technique exploitable to obtain materials that cannot be prepared by the classical method. For example, poly(N-isopropylacrylamide) nanocomposite hydrogels containing graphene were synthesized by FP since graphene did not reaggregate to graphite flakes due to the fast conversion from monomer into polymer.¹⁷ In addition, the hydrogel prepared by FP exhibited higher swelling rate and swelling ratio than that prepared by BP (Table 1).

Table 1 Preparation protocol for molecularly imprinted polymers (MIPs)

Polymer	GFLX (mmol)	NIPAm (mmol)	AA (mmol)	MBAA (mmol)	APS (mmol)	DMSO (mL)	Synthesis method
M1	0.48	22.09	21.86	0.16	0.04	2.5	FP
M1-B	0.48	22.09	21.86	0.16	0.04	2.5	BP
M2	0.24	22.09	21.86	0.16	0.04	2.5	FP
M3	0.72	22.09	21.86	0.16	0.04	2.5	FP
M4	0.96	22.09	21.86	0.16	0.04	2.5	FP
M5	0.48	22.09	21.86	0.044	0.04	2.5	FP
M6	0.48	22.09	21.86	0.088	0.04	2.5	FP
M7	0.48	22.09	21.86	0.44	0.04	2.5	FP
M8	0.48	22.09	21.86	0.88	0.04	2.5	FP
M9	0.48	22.09	21.86	0.16	0.02	2.5	FP
M10	0.48	22.09	21.86	0.16	0.08	2.5	FP
M11	0.48	22.09	21.86	0.16	0.12	2.5	FP

---

When molecular recognition and sensitive dependence on temperature are integrated in the hydroMIP, the receptors with a temperature responsive function can be achieved. The most widely used synthetic thermo-responsive polymer is poly(N-isopropylacrylamide) (PNIPAm) due to its lower critical solution temperature (LCST) close to the temperature of the human body, which undergoes a reversible phase transition at 32 °C in water, changing from hydrophilicity below the temperature to hydrophobicity above it. In addition, thermoresponsive methacrylate and acrylate polymers possessing short oligo(ethylene glycol) side chains were found to have properties comparable to PNIPAm.¹⁸ Thermoresponsive ketoprofen-imprinted monolith using acrylamide and 2-acrylamide-2-methyl propanesulfonic acid as functional monomers was also reported.¹⁹

In view of the facts above, it is intriguing for us to investigate whether FP can be utilized to improve affinity of pH/temperature-sensitive hydrogel-based MIPs. In this investigation, N-isopropylacrylamide was used as thermoresponsive monomer, gatifloxacin as template, acrylic acid as functional monomer, N,N′-methylenebisacrylamide as crosslinker to prepare hydroMIPs. Gatifloxacin (GFLX) was chosen as template molecule since it is a fourth-generation fluoroquinolone antibiotic and acts by inhibiting DNA gyrase and topoisomerase IV. To our knowledge, we are the first to produce hydroMIPs by FP. The drug release of the resulting MIPs hydrogel was evaluated by batch rebinding studies. The textural and morphological characteristics of the hydroMIPs were also determined by mercury intrusion–extrusion porosimetry, differential scanning calorimetry, and field emission scanning electron microscope (FE-SEM).

Results and discussion

Preliminary experiment

In order to prepare MIP hydrogels by FP method, a number of preliminary experiments were performed to determine the conditions in which pure FP occurred, while no presence of simultaneous, spontaneous bulk polymerization (BP). Since the plot life of the reagents mixture at ambient temperature before BP is an important factor for FP, we assessed the plot life by leaving the reagents mixture at ambient temperature to observe whether it gels spontaneously. It was found that the mixture does not gel at ambient temperature, while becoming gel in twenty minutes at a oil bath of 70 °C.

Fig. 1 shows a representative time series of visual imagines illustrating the constant-speed propagation of the polymerization front of the MIP hydrogel. Obviously, a clearly interface between the polymer and unreacted monomer can be observed. The lower layer may be cross-linked MIP copolymer, and the upper layer of the mixture was unreacted monomer. Front propagation occurred at a constant velocity since the experimental data of distance vs. time were well fitted by a straight line, meaning that FP propagates at a constant velocity and a self-sustaining front was obtained. The results indicate that it is possible to perform pure FP without simultaneous occurrence of BP in preparation of hydroMIP.

Fig. 1 The procedures of synthesising MIP1 in frontal polymerization mode.

Imprinting effect can be evaluated by comparing the difference in adsorption ability between MIP and non-imprinted polymer (NIP). 0.36% MBAA were chosen as crosslinker due to a good balance of effect of imprinting and hydrogel. As shown in Fig. 2, higher amount of the template molecule is absorbed on the FP-based hydroMIP than the FP-based hydroNIP, which can be attributed to imprinting effect. In contrast, higher amount of the template molecule is absorbed on the FP-based hydroNIP than the FP-based hydroMIP was observed. Thus, the variation of polymerization mode caused the reversal of adsorption amount of hydroMIP and hydroNIP. The reason maybe is higher cross-linking density of the FP-based hydrogels comparing with the BP-based hydrogel.

Fig. 2 Adsorption isotherms of GFLX on FP-MIP1, FP-NIP1, BP-MIP1 and BP-NIP1. V = 4.0 mL, C₀ = 0–8 mmol L⁻¹, time = 24 h, 10 mg of the polymers.

The effect of polymerization factors on FP

Ammonium​ persulfate (APS) was chosen as initiator in our FP experiments. The effect of APS concentration on front velocity dependence was studied with a constant concentration of AA, NIPAm and MBAA. The results indicated that stable front can be achieved in the APS weight range (5–30 mg) under the experiment condition (Fig. 3a). At lower contents of APS, no propagation front generated due to too much heat loss. In contrast, at high APS concentration, BP will compete with FP.

Fig. 3 Front position–time curves for FP-based hydrogels with different amount of initiator (a) and different amount of cross-linking monomers (b).

In our FP experiments, the dependence of front velocity on cross-linker (MBAA) concentration was also studied. As shown in Fig. 3b, the front velocity was also affected by the amount of MBAA ranging from 0.1% to 1%. All results showed that front velocity depended on the MBAA weight ratios. The greatest front velocity can be obtained at MBAA amount of 0.1%. However, MBAA amount of 0.2% exhibited rampant fingering and could not obtain a stable front.

Characterization of hydroMIPs prepared by FP

The morphologies of hydrogels are further studied by FE-SEM (Fig. 4). Obviously, it can be seen that the FP-based hydrogels were porous and most of the pores were connected to each other to form capillary channels. For BP-based hydrogel, the inner structure of the hydroNIP and hydroMIP was more impact than the hydrogels prepared by FP. This may be attributed to the different polymerization rate of hydrogels preparation. This meant that the polymerization mode had played a significant role in the porosity of hydrogels.

Fig. 4 SEM images of FP-MIP1, FP-NIP1, BP-MIP1 and BP-NIP1.

As knowledge of the porous structure may be used to predict water diffusivity, the results of the resultant hydrogels from mercury intrusion–extrusion porosimetry are shown in Fig. 5. According to IUPAC definitions, the mean pore radius indicated the presence of mesopores in the hydrogels under study. A higher contribution of larger pores can be seen for FP-based hydroMIP, which was consistent with the shifting of smaller surface areas and median pore radius to greater values than in FP-based hydroNIP (Table S1†). Furthermore, the addition of template molecule to the prepolymerization mixture in BP resulted in increased porosity (ca. two fold) and mean pore radius (ca. 1.7 times).

Fig. 5 Differential pore size distribution curves for FP-based hydrogel (a) and BP-based hydrogel (b).

Thermosensitive characters of the FP-based hydroMIP/NIP were determined by recording the enthalpy of transition by differential scanning calorimetry (DSC) using a Seiko DSC-6220 thermal analysis system at a heating rate of 2 °C min⁻¹ from 25 to 90 °C. As shown in Fig. S1A,† the FP-based hydroMIP presented an endothermal peak with a lower critical solution temperature (LCST) of 67.2 °C, while the FP-based hydroNIP showed a broad endothermal peak with a LCST of about 66.9 °C, which indicated a continuous shrinking with the increase of temperature. Whereas the BP-based hydroMIP/NIP showed an unconspicuous endothermal peak within the temperature range (Fig. S1B†). The LCST of pure PNIPAm is about 32 °C. Incorporation of hydrophilic monomer units will increase the LCST and hydrophobic monomer units will decrease the LCST.²⁰ Therefore, the increase of the LCST revealed that the acrylic acid comprised of a free carboxyl group acted as an electrostatic monomer to the GFLX during the selectivity recognition process. The negatively charged acrylic acid groups strongly influenced changes in the hydrophilic/hydrophobic nature of the hydrogels.²¹ In addition, the FP-based hydroMIP presented a higher LCST than the FP-based hydroNIP, indicating the difference of microstructures between the FP-based hydroMIP and the control NIP. It may be caused by the fixed AA assembled around the imprinted cavity, which is the foundation of the specific recognition.²²

Polymerization variables of MIP preparation based on FP

Template to functional monomer ratio. The functional monomer to template (M/T) ratio is a major variable to imprinting efficacy. The molar ratio of M/T was varied by adjusting the amount of imprinting molecule added to an otherwise constant pre-polymerization mixture. In this investigation, the amount of the cross-linking monomers was constant (0.16 mmol, 3.6% of apparent crosslinking degree). At smaller amount of the template, there is no difference between the non-imprinted and imprinted gels (MIP2, IF = 1.04) since the insufficient templates can react with monomers, leading to a lower number of recognition sites (Fig. 6a). At larger amount of the template, no recognition was observed (MIP12, IF = 1.04) due to ineffective interactions between functional monomer with the template (Table S2†). The optimum amount of the template was 0.48 mmol (MIP1, IF = 1.33).

Fig. 6 Adsorption isotherms of GFLX on the MIPs with different amount of the imprinted molecule (a) and different amount of cross-linking monomers (b). V = 4.0 mL, C₀ = 0–8 mmol L⁻¹, time = 24 h, 10 mg of the polymers.

Amount of crosslinker. To further understand how the amount of cross-linker in the hydroMIP influences the binding properties, we prepared a number of hydroMIPs using different amount of MBAA. In this study, we fixed the amount of GFLX to 0.48 mmol throughout all the hydrogels. Adjusting the amount of MBAA added to an otherwise fixed pre-polymerization mixture varied the apparent crosslinking degree of the resulting hydroMIPs. For most of the hydrogels, a porogen of DMSO was adopted. The quality of the hydrogels made with different cross-linkers was comparable and binding properties of the resultant imprinted hydrogels for GFLX in aqueous media were analyzed by batch-type rebinding assays and Langmuir analysis (Fig. 6b and Table S3†). The density of the imprinted sites in the hydrogels ranged from 1.395 to 2.359 mmol g⁻¹ of hydrogel with a general increasing trend from higher to lower cross-linking. When the apparent crosslinking degree of the resulting hydroMIPs was 1%, the best MIP was obtained in terms of IF value (IF = 1.41) in spite of smaller binding amounts.

Adsorption kinetics. In order to study the controlling mechanism of adsorption process, pseudo first-order, pseudo second-order and second-order kinetic models were used to test experimental data. The study of adsorption kinetics was performed by changing the adsorption time from 0 to 24 h (Fig. 7). Kinetics adsorption capacity of the FP-based hydroMIP (M1) was greater than that of the corresponding FP-based hydroNIP. In contrast, kinetics adsorption capacity of the BP-based hydroMIP (M1-B) was smaller than that of the corresponding BP-based hydroNIP. Such situations are similar to the adsorption capacity measured by binding experiments above.

Fig. 7 Kinetic binding of GFLX on FP-MIP1, FP-NIP1, BP-MIP1 and BP-NIP1. The dried hydrogels of 10 mg were placed in a 10 mL centrifugal tube and mixed with 5.0 mL of GFLX solution with concentration of 2.5 mmol L⁻¹.

As illustrated in Table 2, the second-order model was more suitable to describe the adsorption kinetic than other models for the kinetic of GFLX bound onto all the hydrogels. For the FP-based MIP1, the q_e and k were calculated to be 0.854 mmol g⁻¹ and 41.9 × 10⁻³ g mmol⁻¹ h⁻¹, respectively. For the BP-based MIP1B, the q_e and k were calculated to be 0.762 mmol g⁻¹ and 48.1 × 10⁻³ g mmol⁻¹ h⁻¹, respectively. The k of both FP-based and BP-based hydroMIPs was higher than that of the FP-based and BP-based hydroNIPs, indicating that the adsorption rate of the hydroMIPs was higher than that of the hydroNIPs.

Table 2 Kinetic fitting data of hydroMIP and hydroNIP

Kinetic models	Parameters	FP-MIP1	FP-NIP1	BP-MIP1	BP-NIP1
Pseudo-first-order	k1 (h^−1)	0.289	0.348	0.461	0.261
	qe, cal (mmol g^−1)	0.727	0.667	0.614	0.667
	h1 (mmol g^−1 h^−1)	0.210	0.232	0.283	0.174
	R^2	0.985	0.953	0.962	0.961
Pseudo-second-order	k2 (10^−3 g mmol^−1 h^−1)	40.8	43.4	56.5	42.4
	qe, cal (mmol g^−1)	0.852	0.785	0.744	0.776
	h2 (mmol g^−1 h^−1)	0.297	0.267	0.313	0.255
	t1/2 (h)	2.873	2.937	2.380	3.041
	R^2	0.997	0.983	0.988	0.987
Second-order	k (10^−3 g mmol^−1 h^−1)	41.9	35.8	48.1	43.5
	qe, cal (mmol g^−1)	0.854	0.805	0.762	0.782
	R^2	0.999	0.992	0.997	0.996

---

In vitro release studies

The optimum concentration of drug soaking solution. The performance of FP-based hydroMIP was further studied by investigating the ability of drug release. In order to find the best concentration for GFLX release, the FP-based hydroMIP (M1), which is the most effective matrices on GFLX recognition synthesized under optimal FP conditions, was soaked in three different GFLX concentrations of methanol solution for three days before use, and thereafter separated from the solution by centrifugation. The amount of loaded GFLX on the hydrogel was measured by a UV spectrophotometer (Table S4†). It was observed that the loading amount for FP-based hydroMIP was slightly lower than that of control hydrogNIP. For example, the loading capacity for the FP-based hydroMIP was 9.5, 10.0 and 17.6 mg g⁻¹, respectively, and the entrapment efficiency was 50.5, 40.1 and 46.8%, respectively, when loaded with 75, 100 and 150 g mL⁻¹ soaking solution, respectively. In contrast, for the corresponding NIP, the loading capacity was 13.2, 15.2 and 29.0 mg g⁻¹, respectively (Fig. 8), and the entrapment efficiency was 70.6, 61.0 and 77.4%, respectively.

Fig. 8 Release profiles of GFLX from FP-MIP1 and FP-NIP1 soaked in three different GFLX concentrations.

After drug loading, in vitro release experiments were performed under PBS solution (pH 7.4) at 37 °C, and the sustained release was observed in all cases. From Fig. 8 and Table S4,† the release from FP-based hydroMIP soaked with 75 and 100 g mL⁻¹ of GFLX solution, exhibited the almost same longest duration time (about 14 h). Moreover, almost zero order release was observed on the latter, which is desired to maintain the drug concentration within the body at an optimum level. However, in case of FP-based hydroMIP soaked with 150 g mL⁻¹ of GFLX and control hydroNIPs at three different soaked solutions, the drug released rapidly and continuously with the duration of about 7–10 h. In addition, the percent of final release from the two hydrogels was 90%, indicating almost entire drugs release, excepted hydroNIPs at 150 g mL⁻¹ of GFLX soaked solutions, only 67.9% drug released.

Influences of temperature and pH value on the drug release. The effects of temperature and pH value on the GFLX release were investigated for both FP-based hydroMIP and control hydroNIP. In order to monitor the pH-responsive release, two typical pH values i.e., 1.0 (simulated gastric fluid) and 7.4 (simulated intestinal fluid), were selected for a comparative study. From Fig. 9, the hydroMIP and hydroNIP showed an extremely lower release of GFLX at pH 1.0 with the duration of over 120 h than at pH 7.4. Furthermore, the hydroMIP system demonstrated a slower release of GFLX compared with the hydroNIP. This observation can be attributed to the hypothesis that the presence of specific binding sites in the hydroMIP device can provide superior controlled release characteristics. Thus, the result indicated that the imprinted capacities of FP-based hydroMIP was still contained and played the exact role on regulating the release behavior of GFLX at pH 1.0.

Fig. 9 Release profiles of GFLX from FP-MIP1 and FP-NIP1 under different pH (pH 1.0 HCL aqueous solution and pH 7.4 phosphate buffer) and temperature (37 °C and 43 °C) conditions.

Two typical temperatures i.e., 43 and 37 °C were selected for a comparative study by monitoring the temperature-regulated release of GFLX. As shown in Fig. 9, the hydroMIP resembled the hydroNIP and demonstrated a fast release of GFLX when temperature was increased to 43 °C. It indicated that the characteristic of imprinted cavacities within the hydroMIP was temperature-sensitive due to the disappearance of its imprinted effect at 43 °C. Such temperature-regulated characteristic is quite advantageous for the DDS system.

Mathematical analysis of the drug release kinetics from the FP-based hydroMIP. One of the objectives of this contribution was to account for the kinetics of drug release from imprinted hydrogels at different pH values and temperatures. Data obtained during the drug release experiments were fitted to the early-time approximation equations (eqn (8)) in order to determinate the apparent diffusion coefficient (D) at the early stage of controlled release (M_t/M_∞ < 0.6). As displayed in Table 3, the hydroMIP resulted in the lowest value of D at pH 1.0. For example, the value of D was equal to 2.8 × 10⁻⁹ cm² s⁻¹ and 1.8 × 10⁻⁸ cm² s⁻¹ at pH 1.0 and pH 7.4, respectively, while for the control hydroNIP system, the D value was 6.6 × 10⁻⁹ and 4.0 × 10⁻⁸ cm² s⁻¹, respectively. The hydroMIP system at pH 1.0 also demonstrated the lowest value of the swelling ratio, which corresponds to 0.5 as compared to 0.7 for the hydroNIP, and 60.5, 57.0 for the hydroMIP and hydroNIP at pH 7.4, respectively (Table 3). Therefore, both diffusional coefficients and swelling ratio increased as the pH of the release medium was more basic. These results manifest the anionic properties and pH sensibility of these P(AA-NIPAm) hydrogel systems.⁸ Furthermore, all imprinted networks had lower diffusion coefficients than non-imprinted networks no matter at swollen (pH 7.4) or shrunken state (pH 1.0). Thus, the pH-regulated release of the hydroMIP should be still attributed to the imprinted networks, even at swelling state. It indicated that the imprinted networks did play the exact role to delay transport on regulating the release behavior of GFLX at two different pH values.

Table 3 Summary of diffusion coefficients, power law exponents and swelling data obtained from the FP-based hydroMIPs and hydroNIPs at different pH and temperatures

Polymers	pH	T (°C)	GFLX diffusion coefficient (×10^9 cm^2 s^−1)	R^2 of diffusion coefficient	Power law exponent (n)	R^2 of exponent	SR
hydroMIP	1.0	37	2.8	0.992	0.46	0.990	0.5
hydroNIP	1.0	37	6.6	0.985	0.45	0.975	0.7
hydroMIP	7.4	37	17.6	0.844	0.85	0.994	60.5
hydroNIP	7.4	37	39.7	0.867	0.54	0.965	57.0
hydroMIP	7.4	43	46.2	0.847	0.76	0.983	45.1
hydroNIP	7.4	43	41.7	0.841	0.65	0.964	74.2

---

As temperature was increased to 43 °C, the diffusional coefficients from drug release experiments resulted in larger values for both the hydroMIP and the hydroNIP, which were 4.6 × 10⁻⁸ and 4.2 × 10⁻⁸ cm² s⁻¹, respectively. This indicated disappearance of imprinted networks within the hydroMIPs. The corresponding values of SR were 45.1 and 74.2 for the hydroMIP and hydroNIP, respectively. It showed the swelling of the hydrogels were temperature-dependent, for the hydroMIP a high temperature result in a decrease in swelling, whereas for the hydroNIP a high temperature facilitated swelling.

In order to gauge the appropriateness of the fit to a Fickian diffusion or zero-order release mechanism, the diffusional exponents was calculated by eqn (8) and shown Table 3. At pH 1.0, the values of n, 0.46 and 0.45 from the hydroMIPs and the hydroMIPs, respectively, indicated that they were in agreement with a Fickian diffusion mechanism, where n was approximately equal to 0.5. At pH 7.4, the almost all of the hydrogels showed n values between 0.54 and 0.76, indicating that anomalous drug transport involving both diffusion and polymer relaxation were observed. However the n value from the hyroMIP at 37 °C was obtained 0.85. Due to the zero-order release kinetics from the release profile of the hyroMIP at 37 °C, the lower n value may be caused by spheres geometry.²³

In vivo pharmacokinetic study

The in vivo pharmacokinetic profiles in Wistar rats for the loaded FP-based hydroMIP and two control loaded samples, FP-based hydroNIP and BP-based hydroMIP, were compared. Plasma concentrations of the equivalent dose (1.0 mg kg⁻¹) following intragastric administration were shown in Fig. 10. FP-based hydroMIP showed the shortest T_max of 1.5 h compared with FP-based hydroNIP and BP-based hydroMIP, which had a T_max of 3.0 and 2.0 h, respectively. However, its plasma concentration displayed a plateau region of slightly lower C_max between 2 and 10 h, indicating that the controlled release of GFLX from FP-based hydroMIP resulted in the adsorption and elimination reached equilibrium during this period. Moreover, FP-based hydroMIP obtained the significantly highest the value of AUC_0–12 (220.9 h ng mL⁻¹), than FP-based hydroNIP and BP-based hydroMIP, which had an AUC_0–12 of 56.7 and 131.1 h ng mL⁻¹, respectively (Table S5†). These results highlight the potential of the FP-based hydroMIP as a drug-controlled release carrier.

Fig. 10 Plasma concentration–time curves of the GFLX loaded FP-based hydroMIP, FP-based hydroNIP and BP-based hydroMIP.

Conclusions

In present work, we have successfully prepared hydroMIPs via frontal polymerization. Two of polymerization variables are key factor to affect imprinting effect-mole ratios of template to functional monomer (T/M) and the amount of crosslinking agent. Compared to the hydroMIPs prepared with traditionally thermal polymerization, the hydroMIPs prepared with FP mode can favor imprinting effect in terms of imprinting factor. The loaded FP-based hydroMIP was observed almost zero order release of GFLX with the long duration time (about 14 h). The pH-sensitive release of GFLX was obtained from the hydroMIP and hydroNIP with an extremely lower release of GFLX at pH 1.0 (the duration of over 120 h). The temperature-regulated characteristic of imprinted cavities within the hydroMIP was received due to the similar fast release from the hydroMIP and the hydroNIP at 43 °C. In vivo pharmacokinetic study showed a plateau region between between 2.0 and 10.0 h on the plasma concentration from FP-based hydroMIP, indicating that the controlled release of GFLX from FP-based hydroMIP resulted in the adsorption and elimination reached equilibrium during this period. Moreover, FP-based hydroMIP obtained high relative bioavailability due to the highest value of AUC_0–12 compared with the FP-based hydroNIP and BP-based hydroMIP. In summary, the preparation of MIP by frontal polymerization might be expected to a powerful approach to hydroMIPs with enhanced affinity.

Experimental

Materials and instruments

Gatifloxacin (GFLX) was purchased from Dalian Meilun Biotech. Co., Ltd. (99.0%, Dalian, China). N-isopropylacrylamide (NIPAm) was purchased from J&K (99.0%, Beijing, China). Acrylic acid (AA) was from Damao Chemical Reagent Factory (99.0%, Tianjin, China). N,N′-methylenebisacrylamide (MBAA) was obtained from Tianjin Bodi Chemical Industry Co., Ltd. (98.0%, Tianjin, China). Ammonium persulfate (APS) was purchased from J&K (98.0%, Beijing, China). Acetonitrile (HPLC grade) was purchased from Biaoshiqi Science & Technology Development Co., Ltd. (99.9%, Tianjin, China). Other analytical reagents were from Jiangtian Chemical Reagent Co. Ltd. (Tianjin, China).

Synthesis

Frontal polymerization. FP-based hydroMIP was prepared by frontal polymerization. GFLX, NIPAm and MBAA were mixed with AA and DMSO in a 10 mL glass ampoule. After ultrasonic sound of 20 min, APS were added. The mixture was homogenized and then poured into a 100 mm long test tube (i.d. 10 mm). Initiation was achieved by immersing the bottom end (1 cm) of the tube in an oil bath with the temperature kept at 90 °C. After initiation, the tube was taken from the oil immediately, then the polymerization evolved by propagation of the temperature front upwards the entire reactor. After the reaction was completed, the tube was allowed to cool to room temperature. The resultant hydrogels were cut into small pieces and washed by Sechelt extractor using methanol–acetic acid mixture (90 : 10, v/v) for three days, followed by methanol for another one day, until the template, unreacted monomers could no longer be detected by UV spectrophotometer. Then the hydrogels were vacuum dried. NIPs were made in the same manner except GFLX was not included in the formulation.

Batch polymerization. For comparison purposes, hydrogels were also obtained from conventional synthesis. In a conventional synthesis, the same amount components were mixed in a reaction vessel and allowed to react at 60 °C for 20 min. Other procedures were as mentioned above.

Adsorption experiments

The dried hydrogels (10 mg) were placed in a 5 mL centrifugal tube and immersed in 4.0 mL of a solution containing a known concentration of GFLX (0–8 mmol L⁻¹) in ethanol solution for 24 h at 37 °C. The concentration of free GFLX in the supernatant was analyzed by HPLC using a RP ODS column. The mobile phase was composed of methanol–triethylamine–20 mmol L⁻¹ KH₂PO₄ buffer (pH = 3.5) (39.4 : 0.6 : 60). Flow-rate was 1.0 mL min⁻¹. The correlation coefficient for the calibration curve in the range 0.002–0.02 mmol L⁻¹ for GFLX was greater than 0.997.

The amount of GFLX bounded to the polymers at equilibrium (Q_e, mmol g⁻¹) is calculated by subtracting the amount of free substrate at equilibrium from the initial concentration, according to the following eqn (1):²⁴

where C₀ and C_e are the initial and equilibrium concentrations of GFLX in solution (mmol L⁻¹), respectively, V (L) is the volume of the solution, and M (g) is the weight of the polymers.

The Langmuir model was used to characterize the data obtained as follows eqn (2):²⁵

where Q_max is the amount of GFLX adsorbed for a complete monolayer (mmol g⁻¹), and K is a Langmuir constant related to the energy or net enthalpy of sorption (L mmol⁻¹).

Imprinting effect of the MIPs is assessed with imprinting factor (IF), which is defined as:²⁶

IF = Q_MIP/Q_NIP (3)

where Q_MIP and Q_NIP is the amounts of bound template on the MIP and NIP, respectively.

Kinetic adsorption experiments using ethanol as adsorption medium

10 mg of the dried hydrogels was added into 5.0 mL of the ethanol solution containing 2.5 mmol L⁻¹ of GFLX in a 10 mL centrifugal tube at 37 °C for 24 h. At definite time intervals, 50 μL of a supernatant solution was taken. Each sample was measured by at 294 nm to determine the drug concentration in the supernatants at time t. The adsorption amount at time t (min), q_t (mmol g⁻¹), was calculated as follows:²⁷where C_t (mmol L⁻¹) are concentrations of the drug at equilibrium and time t.

The pseudo-first-order, pseudo-second-order and second-order kinetic models were selected to test the adsorption dynamics kinetic experimental data as shown in eqn (5)–(7), respectively:²⁸

q_t = q_e(1 − e^−k₁t) (5)

where q_e and q_t (mmol g⁻¹) represent adsorption capacity of MIP at equilibrium and time t, respectively. k₁ (h⁻¹), k₂ (g mmol⁻¹ h⁻¹) and k (g mmol⁻¹ h⁻¹) are the rate constants of pseudo-first-order, pseudo-second-order and second-order, respectively.

In vitro release studies

Drug loading through soaking procedure. 40 mg of hydrogels were immersed in 10 mL GFLX solution in methanol for three days. Subsequently, supernatants were removed and analyzed by UV spectrophotometer at 294 nm. The drug-loaded hydrogels were dried in oven at 60 °C for 48 h.

In vitro release. Release studies were conducted in 100 mL phosphate buffer (pH 7.4) and HCl aqueous solution (pH 1.0) under different temperature conditions (37 °C and 43 °C). Samples (3 mL) were taken periodically from the receptor medium at certain time-intervals and replaced with the same volume of buffer and analyzed by UV spectrophotometer at 294 nm.

Mathematical analysis of the drug release kinetics from FP-based hydroMIP. Analysis of the GFLX release kinetics from FP-based hydroMIP was performed by calculating the diffusion coefficients D using the approximation equation (eqn (8)), which is obtained when solving Fick's second law of diffusion under initial and boundary conditions equivalent to those of testing in this work:²⁹where M_t/M_∞ is the fractional drug release, t is the release time, D is the corresponding diffusional coefficients and δ is the diffusional distance.

For the hydrogels release studies, an empirical power law equation was used to determine the order of release:³⁰

where k is a constant, and n is the release exponent related to the drug transport mechanism.

The swelling behavior. The dried hydrogels were immersed in the solution as same as the release medium for 24 h to reach swelling equilibrium. Then the excess water was taken from the surface of the hydrogels with wet filter paper. The swelling ratio (SR) was calculated as the following equation:³¹

SR = (W_s − W_d)/W_d (10)

where W_d was the weight of the hydrogels in the dry state, and W_s was the weight of the hydrogels after reaching equilibrium in the different release mediums.

In vivo pharmacokinetic study

Drug administration. Male Wistar rats weighting 180–200 g (8–10 weeks old) were purchased from Experimental Animal Centre of Academy of Military Sciences PLA China (Tianjin, China). All rats were housed and food manufactured by the Department of Laboratory Animal Science of Tianjin Medical University (laboratory animal certificate: syxk2014-0004). All experiments were performed in compliance with the Laboratory Animal Management Rules of the People's Republic of China and Tianjin Municipality Implementation of the Regulations on Laboratory Animal Management Rules.

15 mg of FP-based hydroMIP, 12 mg of ​FP-based hydroNIP, 21 mg of conventional MIP were dispersed separately in 2.0 mL of normal saline solution, and administrated intragastrically (1 mg kg⁻¹ GFLX). The blood samples (ca. 100 μL) were collected from the eye socket vein of anaesthetized rats at specified time intervals of 0.5, 1, 1.5, 2, 3, 4, 5, 6 and 8 h after dosing. Each sample was immediately placed into heparinized test tubes, and immediately centrifuged for 5 min at 13 000 rpm, and then the plasma was collected and stored in vials at −20 °C until analysis.

Preparation and quantitative assay of plasma sample. To a 50 μL aliquot of plasma in a conical centrifuge tube, 150 μL acetonitrile was added and mixed for 30 s. After centrifugation at 13 000 rpm for 5 min, the upper organic layer was transferred to another set of clean tube and dried under a gentle stream of nitrogen. The residue was reconstituted in 50 μL of the mobile phase, and an aliquot of 20 μL was injected into the HPLC system for analysis.

An Agilent 1100 system (Shimadzu, Japan) and a RP ODS column (150 × 4.6 mm i.d., 5 μm; Kromasil, Sweden) were used for HPLC. The detection was set at 293 nm and the column temperature was kept at 25 °C. The mobile phase was acetonitrile–10 mmol L⁻¹ KH₂PO₄ buffer pH 3.05 (5 : 18, v/v). Flow-rate was set 1.0 mL min⁻¹. Data processing was carried out with a HPCORE workstation. The sensitivity of detection was 50 ng mL⁻¹ for GFLX. The correlation coefficient for the calibration curve in the range 50–700 ng mL⁻¹ for GFLX was greater than 0.995.

Data analysis. In the pharmacokinetic study, the maximum plasma concentration (C_max) and time to reach the maximum plasma concentration (T_max) were obtained using actual observations. The area under the plasma concentration–time curve (AUC) was calculated by a linear trapezoidal method.

Acknowledgements

This work was supported by National Natural Science Foundation of China (Grant No. 21375096 and U1303202) and Natural Science Foundation of Tianjin Medical University (Grants No. 2014KYQ14).

References

1. H. F. EL-Sharif, D. M. Hawkins, D. Stevenson and S. M. Reddy, Phys. Chem. Chem. Phys., 2014, 16, 15483.
2. D. M. Hawkins, A. Trache, E. A. Ellis, D. Stevenson, A. Holzenburg, G. A. Meininger and S. M. Reddy, Biomacromolecules, 2006, 7, 2560.
3. H. Hiratani, C. Alvarez-Lorenzo, J. Chuang, O. Guney, A. Y. Grosberg and T. Tanaka, Langmuir, 2001, 17, 4431.
4. Z. Chen, L. Xu, Y. Liang and M. Zhao, Adv. Mater., 2010, 22, 1488.
5. A. Kawamura, T. Kiguchi, T. Nishihata, T. Uragami and T. Miyata, Chem. Commun., 2014, 50, 11101.
6. M. E. Byrne, K. Park and N. A. Peppas, Adv. Drug Delivery Rev., 2002, 54, 149.
7. B. Singh and N. Chauhan, Acta Biomater., 2008, 4, 1244.
8. M. E. Byrne and V. Salian, Int. J. Pharm., 2008, 364, 188.
9. M. J. Whitcombe, Nat. Chem., 2011, 3, 657.
10. T. Oya, T. Enoki, A. Y. Grosberg, S. Masamune, T. Sakiyama, Y. Takeoka, K. Tanaka, G. Wang, Y. Yilmaz, M. S. Feld, R. Dasari and T. Tanaka, Science, 1999, 286, 1543.
11. T. Takeuchi, T. Mori, A. Kuwahara, T. Ohta, A. Oshita, H. Sunayama, Y. Kitayama and T. Ooya, Angew. Chem., Int. Ed. Engl., 2014, 53, 12765.
12. E. Saridakis, S. Khurshid, L. Govada, Q. Phan, D. Hawkins, G. V. Crichlow, E. Lolis, S. M. Reddy and N. E. Chayen, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 11081.
13. J. L. Urraca, C. S. Aureliano, S. E. Chillinger, H. Esselmann, J. Wiltfang and B. Sellergren, J. Am. Chem. Soc., 2011, 133, 9220.
14. Z. Adali-Kaya, S. B. B. Tse, A. Falcimaigne-Cordin and K. Haupt, Angew. Chem., Int. Ed. Engl., 2015, 54, 5192.
15. D. Nuvoli, V. Alzari, J. A. Pojman, V. Sanna, A. Ruiu, D. Sanna, G. Malucelli and A. Mariani, ACS Appl. Mater. Interfaces, 2015, 7, 3600.
16. I. D. Robertson, H. L. Hernandez, S. R. White and J. S. Moore, ACS Appl. Mater. Interfaces, 2014, 6, 18469.
17. R. Sanna, D. Sanna, V. Alzari, D. Nuvoli, S. Scognamillo, M. Piccinini, M. Lazzari, E. Gioffredi, G. Malucelli and A. Mariani, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 4110.
18. N. P. Truong, M. R. Whittaker, A. Anastasaki, D. M. Haddleton, J. F. Quinn and T. P. Davis, Polym. Chem., 2016, 7, 430.
19. X. Sun, C.-Y. Zhao, X.-H. Wang, Y.-P. Huang and Z.-S. Liu, Anal. Bioanal. Chem., 2014, 406, 5359.
20. N. T. D. Tran, Z. Jia, N. P. Truong, M. A. Cooper and M. J. Monteiro, Biomacromolecules, 2013, 14, 3463.
21. N. T. D. Tran, N. P. Truong, W. Gu, Z. Jia, M. A. Cooper and M. J. Monteiro, Biomacromolecules, 2013, 14, 495.
22. Z. D. Hua, Z. Y. Chen, Y. Z. Li and M. P. Zhao, Langmuir, 2008, 24, 5773.
23. J. Siepmanna and N. A. Peppas, Adv. Drug Delivery Rev., 2012, 64, 163.
24. X.-X. Li, L.-F. Hao, Y.-P. Huang, H.-Q. Duan and Z.-S. Liu, Polym. Eng. Sci., 2012, 52, 1440.
25. A. M. Rampey, R. J. Umpleby II, G. T. Rushton, J. C. Iseman, R. N. Shah and K. D. Shimizu, Anal. Chem., 2004, 76, 1123.
26. S.-J. Zhang, Y.-P. Huang, Z.-S. Liu and H.-Q. Duan, Polym. Adv. Technol., 2011, 22, 286.
27. L. Ma, L. Tang, R.-S. Li, Y.-P. Huang and Z.-S. Liu, RSC Adv., 2015, 5, 84601.
28. L. Tang, C.-Y. Zhao, X.-H. Wang, R.-S. Li, J.-R. Yang, Y.-P. Huang and Z.-S. Liu, Int. J. Pharm., 2015, 496, 822.
29. M. Ali, S. Horikaw, S. Venkatesh, J. Saha, J. W. Hong and M. E. Byrne, J. Controlled Release, 2007, 124, 154.
30. J. Siepmann and F. Siepmann, J. Controlled Release, 2012, 161, 351.
31. D. Ran, Y. Wang, X. Jia and C. Nie, Anal. Chim. Acta, 2012, 723, 45.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20626c

---