Porous materials as delivery and protective agents for Vitamin A

Abstract

The suitability of porous materials to immobilize and release under control bioactive molecules prompted us to design and study delivery systems of Vitamin A (VitA). This molecule, relevant in several physiological functions, is easily oxidized. Commercial VitA was immobilized in two different clays, montmorillonite K-10 (MMT) and sepiolite (SEP), and in MCM-41, by impregnation. Characterization of the resulting hybrid materials by XRD, FTIR and ¹³C and ²⁹Si (MAS) NMR spectroscopies revealed its presence. The photo-stability tests showed decreased degradation of VitA in the clays, compared to MCM-41 and the pure VitA, while thermostability is observed until ∼100 °C. The kinetics of the release depended on the structural features of the support material and the pH. Sepiolite originated a classic profile of increasing amount of VitA with time, indicating that no oxidation was taking place. In both montmorillonite and MCM-41 the amount of released VitA dropped after ∼2 hours, reflecting oxidation. Oxidation and degradation products obtained when VitA was immobilized in MCM-41, both under nitrogen and in air, were identified by mass spectrometry experiments. Sepiolite is therefore a suitable material to use in controlled release of VitA, since it appears to prevent oxidation.

---

1 Introduction

Nutriceuticals, supplements and chemopreventive agents have emerged in biotechnology associated with the solubilization and consequent delivery of functionally active compounds, usually lipophilic, to living organisms. Much attention has been paid to a class of compounds generically labelled “Vitamin A” (VitA), which encompasses any compound, with important nutritional and pharmaceutical roles in humans, possessing the biological activity of all-trans-retinol. The term ‘retinoids’ usually refers to both the naturally occurring forms of VitA and the many synthetic analogues of retinol (Scheme 1), namely retinyl acetate, with or without biological activity.¹

Scheme 1 Structure of retinol.

Retinyl acetate is the main compound of the VitA family and one of the esterified forms of retinol present in foods.^2,3 VitA has been found to play a relevant role in a large number of physiological functions,⁴ and to be essential for human survival from embryogenesis to adulthood. The range of cellular activities in which VitA participates is still being investigated. The molecular mechanism of VitA physiological functions was first elucidated for vision⁵ but biological functions of VitA have been discovered in almost every vertebrate organ, including reproduction, embryonic growth and development, maintenance of epithelial surfaces, and, above all, proper functioning of the adult brain.^6,7 VitA is known for modulating growth, differentiation, and apoptosis of normal, premalignant, and malignant cells. Such effects have been probed under both in vitro and in vivo⁸ conditions. Moreover, it represents not only a major class of chemopreventive agents⁹ but is also able to achieve a protective role via antioxidant properties,^10,11 namely in the prevention from various cancers.⁸

VitA has been identified to be crucial in the immune system, therefore helping to protect against infections. In fish, high dietary levels of VitA have been shown to enhance certain immune functions.^12–15 Deficiency of VitA is a major cause of disease in cases of unbalanced diet and malabsorption diseases. A solution to these problems has been provided by supplementing the daily diet with large, pharmaceutically administered doses of VitA. This kind of protocol has been found to reduce successfully the incidence and severity of some infectious diseases.¹⁶ In particular, either the administration of VitA alone¹⁷ or the development/application of an alternative VitA-rich diet¹⁸ have been found to have a large impact on human survival by decreasing childhood mortality by 20–70% in developing countries. There are drawbacks, however, as excess of VitA is associated with harm and even death.^19–21

Further disadvantages are assigned to both the poor solubility in water and the high chemical instability, which lead to an easy loss of the physiological activity under UV irradiation. Therefore, the protection of VitA towards photodecomposition and oxidation and the enhancement of its solubility in aqueous systems have been a challenging task for both scientists and industrialists. The encapsulation method revealed to be very promising for improving not only VitA stability and bioavailability, but also targeting and drug delivering directly to the action site, such as the brain for the elderly or Alzheimer's patients,²² or to short-circuit the liver or bone, for example, in order to avoid any adverse effects caused by an excess of VitA.

We propose to use three different porous materials, the clays montmorillonite and sepiolite, and the mesoporous material MCM-41, to immobilize VitA and study its releasing profile, aiming at delaying its photodegradation. Montmorillonite [(1/2 Ca, Na)_0.5–1(Al, Mg, Fe)_4–6(Si, Al)₈O₂₀(OH)₄·nH₂O] is a clay mineral bio-inert of the group of smectites, whose structure consists of a layer of aluminium octahedral (O) sandwiched between two layers of silicon tetrahedra (T), known as TOT packing. These layers are separated by a space containing water molecules, and calcium, sodium and potassium ions, to balance the net negative charge of the clay lamellae. The montmorillonite lattice can be expanded by cation exchange (100–200 meq. per 100 g).²³ Sepiolite (Si₁₂O₃₀Mg₈(OH)₄(H₂O)₄·8H₂O) is a natural, fibrous clay mineral with fine microporous channels running parallel to the length of the fibres. The structure is in some aspects similar to those of other 2 : 1 trioctahedral silicates, such as montmorillonite, but it has discontinuities and inversion of the silica sheets, which give rise to channels. Some of the Si atoms at the corners of the outer blocks are bound to hydroxyls (Si–OH).²⁴ The MCM-41 (SiO₂) is a thermally, chemically and mechanically stable mesoporous silica material. Its structure is characterized by ordered hexagonal channels running in one direction, with much larger diameter then the sepiolite channels.²⁵ These three materials (see also Fig. S1 in ESI‡) are characterized by a large surface area, high absorption capacity, good biocompatibility and low or no toxicity, being thus widely used as drug delivery systems.^25–27 We investigated the encapsulation of VitA in these inorganic matrices to determine the role of the matrix on VitA photostability, as well as to elucidate the release mechanisms.

The work therefore addresses the synthesis and characterization of the inorganic/organic hybrids (IOHs), using XRD, FT-IR, solid state NMR, and TGA measurements to control the presence of VitA in the inorganic matrix, and UV-Vis spectra on solids to provide information on the VitA photostability inside the inorganic materials, followed by the study of VitA release.

2 Experimental section

2.1 General

Montmorillonite K-10 (MMT-K10) and Vitamin A (retinol acetate) were obtained from Sigma Aldrich, Sepiolite S-15 (SEP) from Tols, and were used as received. Vitamin A from a local Palermo pharmacy was used as received for comparison. Commercial grade dichloromethane was dried and deoxygenated by standard procedures (over calcium hydride), distilled under nitrogen, and kept over 4 Å molecular sieves. MCM-41 (MCM) was prepared by adopting a methodology previously described,²⁵ using [(C₁₄H₃₃)N(CH₃)₃]Br as template agent. The template agent was removed by calcination (823 K, 6 h, air atmosphere). Prior to the loading experiments, MCM was heated under vacuum at 453 K for 2 h.

FTIR spectra were obtained as KBr pellets (Vitamin A) and diffuse reflectance measurements (DRIFT) (inorganic materials and hybrids) on a Nicolet 6700 in the 400–4000 cm⁻¹ range using 4 cm⁻¹ resolution.

¹³C and ²⁹Si solid-state NMR spectra were recorded at 79.49 and 59.63 MHz, respectively, on a TecMag/Bruker 300 wide bore spectrometer. The standard magic angle spinning (MAS) cross polarization–dipolar decoupling RF pulse sequence (CP–DD) was used under about 4 kHz spinning rate.¹³C and ²⁹Si spectra were both recorded with 1 ms contact time, but 5 s and 3 s recycle delays were selected for the observation of ¹³C and ²⁹Si signals, respectively. The Hartmann–Hahn condition was optimized using glycine for ¹³C and tetrakis(trimethylsilyl)silane for ²⁹Si, also the external references to set the chemical shift scales (¹³CO at 176.03 ppm and ²⁹Si(CH₃)₃ at −9.8 ppm, respectively).

TGA studies were performed using a Perkin-Elmer TGA7 thermobalance system at a heating rate of 10 °C min⁻¹ under N₂.

Powder XRD data was collected on a Phillips PW1710 diffractometer using Cu-Kα radiation filtered by graphite. UV-Vis spectra were measured with a Shimadzu UV-2450PC spectrometer equipped with a Shimadzu TCC-240A Peltier temperature controlled cell.

Electrospray mass spectrometry (ESI-MS) was performed on a LCQ Duo ion trap mass spectrometer from Thermo Finnigan (San Jose, CA, USA). Samples were introduced, via a syringe pump (flow rate of 5 mL min⁻¹), into the stainless steel capillary of the ESI source. All mass spectrometry data were acquired in positive ion mode, the full scan spectra were recorded in the range m/z 50–600 and the applied spray voltage in the source was 4.5 kV, the capillary voltage was 10 V and the capillary temperature was 493 K using N₂ both as nebulizing and drying gas in the source.

Microanalyses for CH quantification were performed at CACTI, University of Vigo, on a Fisons EA 1108. N₂ sorption measurements were obtained on a Quantachrome Autosorb iQ porosimeter. BET specific surface areas (S_BET, p/p₀ from 0.03 to 0.30) and specific total pore volume V_p were estimated from N₂ adsorption isotherms measured at 77 K. The pore size distributions (PSD) were calculated by the BJH method from the desorption branch of the isotherm, using the modified Kelvin equation with correction for the statistical film thickness on the pore walls. The statistical film thickness was calculated using Harkins–Jura equation in the p/p₀ range from 0.1 to 0.95.

2.2 Materials

Loading of Vitamin A. The hybrid materials were obtained by mixing 1 g of inorganic material (MMT, SEP, MCM) and 0.3 g of VitA in dichloromethane (20 mL). The mixture was stirred, with a magnetic stirrer at 500 rpm, for 3 days at room temperature, under a nitrogen atmosphere, in a Schlenk covered with aluminum foil to prevent any action from light. After filtration, the solid was washed once with dichloromethane and dried under vacuum for 2 hours. UV-Vis analysis on the washing solvent did not reveal any noticeable amounts of VitA released from any of the materials.
MMT–VitA. IR (ν, cm⁻¹): 3627 (vs), 3420 (vs-br d), 2935 (w), 1630 (m), 1460 (w), 1055 (vs), 931 (s), 800 (m), 530 (s). ¹³C CP/MAS NMR (δ ppm): 12–19, 21.0, 24.7, 28.6, 41.5, 59.2, 71.8, 91.7, 102.6, 172.4. Elemental analysis found (%): C 12.5, H 2.61.
SEP–VitA. IR (ν, cm⁻¹): 3690 (w), 3563 (vs), 3390 (vs-br d), 2957 (w), 1659 (m), 1556 (w), 1436 (w), 1210 (m), 1082 (s), 1013 (s), 977 (s), 692 (m), 646 (m). ¹³C CP/MAS NMR (δ ppm): 12–17, 20.4, 24.6, 29.4, 41.5, 58.9, 71.3, 80.7, 101.6, 172.1. Elemental analysis found (%): C 12.0, H 2.57.
MCM–VitA. IR (ν, cm⁻¹): 3407 (vs br d), 2965 (w), 2856 (w), 1630 (m), 1465 (w), 1380 (w), 1230 (vs), 1086 (vs), 951 (s), 800 (m), 560 (m). ¹³C CP/MAS NMR (δ/ppm): 12–18, 20.4, 24.6, 29.2, 42.1, 59.6, 71.9, 80.7, 101.3, 171.8 (the signal identified at about 50 ppm is temptatively assigned to residual dichloromethane). Elemental analysis found (%): C 24.2, H 4.02.

3 Results and discussion

3.1 Immobilization of Vitamin A and characterization of the materials

The loading of VitA on the three inorganic materials (MMT, SEP, MCM) was immediately detected with naked eyes by the color change of the inorganic materials from white to deep blue, blue, and yellow, respectively. Elemental analysis of the resulting inorganic–organic hybrid materials, MMT–VitA, SEP–VitA, and MCM–VitA, showed that MMT–VitA contained 0.38 mmol g⁻¹ material, SEP–VitA 0.36 mmol g⁻¹ and MCM–VitA 0.74 mmol g⁻¹ of VitA. These materials were characterized by different complementary techniques, namely XRD, FTIR, solid state ¹³C NMR, and TGA.

The XRD powder patterns of all host (MMT, SEP and MCM) and composite materials (MMT–VitA, SEP–VitA and MCM–VitA) are shown in Fig. 1. Most of the characteristic peaks in the patterns of the two MMT systems are in the same position, except the characteristic peak indexed to the 001 reflection, observed at 2θ = 8.71° for MMT, and shifted to 2θ = 6.07° for MMT–VitA. According to Bragg's law, peaks shifting from higher to lower diffraction angles results from a d-spacing increase (from 10.1 Å to 14.5 Å after VitA loading). Assuming that the alumino-silicate layer is ca. 9.6 Å thick,²⁸ the drug intercalation forces the gallery to expand from 0.5 Å to 4.9 Å.

Fig. 1 XRD powder patterns of the pristine inorganic materials (black line) and the new hybrids (red line) (a) MMT and MMT–VitA; (b) SEP and SEP–VitA; (c) MCM and MCM–VitA. In (c) the inset shows the zoomed pattern for MCM–VitA.

The pristine MCM shows the characteristic four reflections in the 2θ range 2–10° (Fig. 1), indexed to a hexagonal cell as (100), (110), (200), and (210), with a d value of 38.5 Å for the 100 reflection, corresponding to a lattice constant of a = 44.5 Å. The XRD powder pattern of MCM–VitA is quite similar to that of MCM, with observed (100), (110) and (200) reflections (Fig. 1c), indicating retention of the long range hexagonal symmetry. The value of d for the (100) reflection is 37.2 Å, associated with a lattice constant of a = 43.0 Å. There is a strong attenuation of the XRD peaks intensities that must not be interpreted as a loss of crystallinity. Instead, it is most likely due to a reduction in the X-ray scattering contrast between the silica wall and the pore filling material (Fig. 1c, inset).^29,30 This observation is corroborated with elemental analysis data, which indicated a loading a 0.737 mmol g⁻¹ of VitA as discussed earlier.

The SEP based materials XRD diffraction patterns (Fig. 1) do not change after loading the VitA molecules. These results not only indicate that the structure and crystallinity of SEP were maintained in the hybrid material, but also suggest that VitA adsorption occurred on the surface rather than inside the clay or by partial replacement of inter layer water.³¹

Nitrogen sorption/desorption studies at 77 K were also performed for MCM and MCM–VitA and showed reversible type IV isotherms (Fig. 2 left), typical of mesoporous solids (pore width between 2 nm and 50 nm, according to IUPAC). The calculated textural parameters (S_BET and V_p) of these materials (Table 1) agree with literature data for this type of materials.^32–34

Fig. 2 Nitrogen adsorption studies of MCM and MCM–VitA materials at 77 K: isotherm (a) and pore size distribution curves (b). In the isotherm, both the adsorption (closed symbols) and desorption (open symbols) branches are shown.

Table 1 Textural parameters of host MCM and composite material MCM–VitA, from powder XRD data and N₂ isotherms at 77 K, for MCM materials

Material	2θ/°	d100/Å	a/Å	SBET^a (m^2 g^−1)	ΔSBET/%	Vp/cm^3 g^−1	ΔVp^b/%	dBJH/nm
a Surface area variation relatively to parent MCM.b Total pore volume variation relative to parent MCM.
MCM	2.29	38.5	44.5	1077	—	1.23	—	3.6
MCM–VitA	2.29	37.2	43.0	609	−44	0.75	−39	3.6

---

The capillary condensation/evaporation step in pristine MCM sample appears as a narrow step in the 0.25–0.35 relative pressure range (Fig. 2 right); the sharpness of this step reflects a uniform size distribution. According to Table 1, the functionalized material MCM–VitA isotherm reveals a lower N₂ uptake, accounting for the decrease in both S_BET (−44%) and V_p (−39%). These results are in agreement with the p/p₀ coordinate decrease in the isotherm inflection points after post-synthesis treatments.³² Hysteresis was also observed in the isotherms. This phenomenon in adsorption/desorption isotherm measurements is usually due to capillary condensation in conjunction with nonlinearity of equilibrium where adsorption/desorption dynamics is high.

The FTIR spectra of both the pristine inorganic materials and the inorganic/organic hybrid materials are depicted in Fig. 3, as well as the spectrum of VitA for comparison.

Fig. 3 FTIR spectra of the pristine inorganic materials (blue line) and the new hybrids (red line) (a) MMT and MMT–VitA; (b) SEP and SEP–VitA; (c) MCM and MCM–VitA; (d) VitA. Arrows indicate relevant VitA bands in the hybrid materials.

The spectra of the two MMT and SEP clays reveal the presence of characteristic bands of the octahedral layer for MMT at 931 cm⁻¹, 530 cm⁻¹ and for SEP at 3690 cm⁻¹, 692 cm⁻¹, 646 cm⁻¹ (ref. 31 and 35) corresponding to the νAl–Al–OH, νAl–O–Si modes, respectively,^36,37 as well as the O–H stretching vibrational modes around 3400 cm⁻¹ and the Si–O stretching in the range 1000–1200 cm⁻¹.

The spectrum of MCM is dominated by the broad and intense bands at 3422 cm⁻¹ assigned to νO–H modes of hydrogen bonded hydroxyl groups, and in the range 1000–1200 cm⁻¹ corresponding to νSi–O–Si modes. The spectrum of MCM also displays some unexpected bands in the 2900–3000 cm⁻¹ range, probably due to νC–H modes arising from the template agent (CTAB) that has been incompletely removed during the calcination step.

Although the spectra of the composite materials MCM–VitA, MMT–VitA and SEP–VitA are dominated by the bands of the host material, it is possible to detect the presence of VitA owing to small, but relevant, spectral features³⁸ that offer a positive identification of the guest (Fig. 3).

Therefore, in MMT–VitA, a set of three broad and very weak bands in the νC–H modes region centered at 2935 cm⁻¹ indicates that VitA is present in the interlayer spacing of the host clay, without dramatic structural changes. Also visible is a band at 1460 cm⁻¹, assigned to the C–H bending and C=C stretching modes of the VitA molecule. The corresponding wavenumbers in the experimental FTIR spectrum of pure VitA are found at 2925 cm⁻¹ (νC–H modes) and 1458 cm⁻¹ and 1380 cm⁻¹ (βC–H and νC=C modes).

The spectrum of MCM–VitA is characterized by νO–H vibrational modes, present as broad and intense bands at 3407 cm⁻¹, slightly shifted from those of MCM, and νSi–O–Si modes of the siliceous matrix of MCM at 1000–1200 cm⁻¹. The band corresponding to the νC–H mode of loaded VitA is observed at 2965 cm⁻¹, while the βC–H and νC=C modes are evidenced by the presence of bands at 1465 cm⁻¹ and 1380 cm⁻¹. The observation of this pattern indicates that the drug was successfully loaded inside the pores of the MCM mesoporous material.

The ¹³C NMR solid-state spectra of the two clays containing VitA, MMT–VitA, SEP–VitA, and MCM–VitA are displayed in Fig. 4.

Fig. 4 ¹³C NMR solid-state spectra of VitA and the hybrid materials MMT–VitA, SEP–VitA and MCM–VitA (* denotes ¹³CO spinning side bands; # denotes the presence of dichloromethane used in the loading process).

The ¹H and ¹³C NMR spectra of neat VitA were recorded (not shown) to check its condition; both spectra allowed assignment of all signals found as belonging to VitA.

The spectra of the hybrids do not show any significant changes relative to the spectrum registered for the neat drug. This suggests that VitA is loaded into the inorganic materials and the host interacts with the guest to some extent stabilizing the drug confined in the clay gallery, though the immobilized VitA does not undergo any significant chemical change. The observed line narrowing of the spectrum of MCM–VitA is consistent with VitA mobility being more favored inside the MCM than in the clay materials. Nevertheless, the presence in the spectra of additional peaks, which do not belong to the carbons of VitA, has been detected, both in the drug and in the two materials. These observed peaks could be indexed to carbon atoms, such as carbons 5, 6, 7, and 8, resulting from oxidation products of VitA,³⁹ as shown in Scheme 2. Their presence is also observed in the spectrum of another VitA sample bought in the pharmacy, revealing that the organic compound was already partially oxidized and suggesting that these oxidized species are not considered to be harmful for the health.

Scheme 2 Molecular formula of VitA acetate and its oxidation products.

²⁹Si NMR spectra of MCM and MCM–VitA are shown in Fig. 5. In the ²⁹Si CP MAS NMR spectra of the MCM and MCM–VitA materials three convoluted resonances are clearly observed at −91.3 ppm, −100.7 ppm and −110.7 ppm, assigned to Q², Q³ and Q⁴ species, respectively, for the MCM and at similar deviations for MCM–VitA [Qⁿ = SiO_4−nSi(OH)_n]. These results show that loading of Vitamin A inside the pores of MCM did not affect to a great extent the structure of the host materials.

Fig. 5 ²⁹Si NMR solid-state spectra of MCM and the hybrid material MCM–VitA.

The knowledge of the photo-stability of the VitA is mandatory for any application of the drug in different industrial fields. It was rendered even more relevant by the finding that the supposedly pure product bought in the pharmacy was already oxidized, as described above. In order to establish the VitA photostability, both in its pristine form or confined into the hybrid materials, some tests were performed. All samples were exposed to sunlight for two weeks, and UV-visible spectra on every solids were recorded after seven and fourteen days⁴⁰, respectively (Fig. 6).

Fig. 6 Solid state UV-Vis spectra (a) MMT–VitA hybrid, (b) SEP–VitA hybrid, (c) MCM–VitA hybrid, (d) VitA, at t = 0 days (black line), t = 7 days (red line), and t = 14 days (blue line).

Even though the reflectance for the pristine VitA and the IOH share almost the same profiles, comparison of the trends suggests that the confinement plays a relevant role to an extent that varies with the inorganic materials nature. In particular, after 14 days exposure, for the pristine VitA and the hybrid MCM–VitA, the reflectance shows a significant increase up to ca. 30 and 50%, respectively, while for the hybrid MMT–VitA and SEP–VitA small variations of ca. 4% and 10%, respectively, have been detected. These findings are strongly indicative of the protective effect of the clays in slowing down photodegradation/oxidation, as evidenced by a blue-shift of the UV-Vis spectrum of VitA.

This result represents a precious advantage for the application of the clay-hybrid formulations.

The thermo-gravimetric analysis (Fig. 7) of the inorganic materials containing VitA revealed their thermal stability and the amount of VitA immobilized, confirming the results of elemental analysis (discussed earlier). By examining the loading data it can be assessed that the MCM–VitA and MMT–VitA hybrids show a higher loading efficiency than the SEP–VitA counterpart, probably owing to the more accessible sites.

Fig. 7 TGA profiles of the pristine inorganic materials (blue line), the new hybrids (red line), and VitA (black line) (a) MMT, (b) SEP, and (c) MCM.

Inspection of the trends reported in Fig. 7 shows that on increasing temperature all the samples undergo mass losses when approaching the critical value of 100 °C, clearly suggesting that physisorbed water is lost. A further increase of temperature leads to an additional mass loss, whose extent depends on the nature of the materials, i.e., pristine or hybrids. The pristine inorganic materials lose only a small quantity of mass, while for the other materials a significant weight loss is detected. The VitA drug by itself shows also considerable weight loss, assigned to decomposition. Probably, dehydroxilation takes place in the materials, accompanied by decomposition of VitA whenever it is present.

Moreover, we cannot forget that the thermogravimetric analysis provides important information about the site of adsorption of the vitamin in SEP. In fact, the weight loss below 100 °C for SEP–VitA is smaller than for SEP, indicating that the latter loses a large amount of water. This trend is in agreement with the literature and indicates that the functionalization of the SEP mainly occurs on the surface of SEP by partial replacement of zeolitic water.³¹ The same trend was observed for the MCM material when compared with MCM–VitA.

3.2 The release of VitA from the inorganic hybrid materials

In vitro release kinetic studies have been performed by suspending the hybrid materials, in an aqueous solution, which mimics the physiological pH conditions (HCl pH = 1.0 and phosphate buffer pH = 7.0), in the amount needed to obtain a maximum concentration of Vitamin A equal to 3 mmol dm⁻³ (corresponding to a concentration of 1 mg mL⁻¹ of VitA in the cuvettes). The suspension was continuously stirred for 24 hours, in the dark, at 37.0 °C. These conditions reproduce the oral drug administration and the subsequent physiological release. At scheduled time intervals, 1 mL of the release medium was withdrawn, centrifuged and the supernatant was analyzed by recording UV-Vis spectra. At the end, both the centrifuged and the supernatant, vigorously shaken, were reintroduced into the releasing suspension.

The results show that the release of VitA depends strongly on the nature of the inorganic/organic hybrid material. This is most probably intrinsically correlated with the host–guest specific interactions that were established. For instance in MCM the majority of interaction will be by hydrogen bonding, while in the clays, in addition to hydrogen bonding, electrostatic interactions may arise. The pore size, much smaller in SEP, may also influence the release rate.

The MCM–VitA sample released all the VitA in 15 minutes, as can be seen in Fig. 8. Analysis of the supernatant, using UV-Vis spectroscopy, reveals only the presence of VitA^41,42 for approximately 120 minutes, when a new band appeared at lower wavelengths. Its absorbance increased during the course of the study (up to 24 h). This release behavior observed for MCM–VitA and also for MMT–VitA is unusual. In fact, the initial fast release can be explained by weak host–guest interactions that will be broken within the large void spaces of MCM and MMT (Fig. S1 in ESI‡), leading to VitA release into the medium. The following concentration decrease can be ascribed to a degradation mechanism of VitA. This is reflected by the appearance, after 2 hours, of the new blue-shifted band in the UV-Vis spectra (Fig. 8c).

Fig. 8 Kinetic profiles of VitA release at 37 °C at pH = 1.0 and pH = 7.0 from: (a) MCM–VitA and (b) SEP–VitA. The “100%” mark corresponds to the complete release of VitA into the medium (1 mg mL⁻¹). Typical UV-Vis spectra of supernatant medium containing samples VitA of released from MCM–VitA, at pH = 7.0 at 37 °C, collected at different times, are shown in (c).

The release profiles of MCM–VitA and MMT–VitA show dramatic differences in their pH dependence.

Since in the MMT clay cations and anions must be balanced, a fast release of VitA at pH = 1 can be achieved by exchanging with protons, while at pH = 7 the same is also experienced at a slight lower level and probably VitA will be released in deprotonated form (its OH protons will stay in the clay). In the case of MCM–VitA, at both pH values the release will follow a concentration gradient by diffusion into the medium. However, release into the medium at pH = 7 will incur in degradation of the guest (agreeing with the profile in Fig. 8a), whereas this fact is much less pronounced at pH = 1 where the higher concentration of protons might prevent release of the guest.

In order to identify the origin of the new absorption band, the spectrum of the VitA alone was monitored in the same releasing media for 24 hours. Since it is known⁴² that VitA undergoes oxidation in these conditions, it is possible to assign the new band to the oxidation products, identified by ¹³C CP/MAS-NMR (Fig. 4). The break of the double bonds conjugation of VitA causes the blue shift of the new band that appears in the UV-spectra below 300 nm during the release. This hypothesis has been further confirmed by performing mass spectrometry measurements (see below).

The kinetic profile shown in Fig. 9 indicates that, in the release of VitA from the hybrid SEP–VitA material, the amount of released VitA increases with time, in agreement with the intensities of the band at the selected λ_max in the spectrum of VitA, and more VitA is released at pH = 1.0. This result demonstrates that the SEP is able to prevent the oxidative degradations of the VitA, as reported in the literature.⁴³

Fig. 9 Kinetic profiles of VitA release at pH = 1.0 and pH = 7.0 from SEP–VitA at 37 °C. The “100%” mark corresponds to the complete release of VitA into the medium (1 mg mL⁻¹).

The difference between the SEP and the MMT clay will possibly arise from differences in their confined space configuration, the former having smaller void spaces (Scheme 1) thus allowing less freedom for the guest to reach a dynamic regime that would allow fast release into the medium. In this way slow kinetics were observed although release of VitA reached good to excellent levels after 24 h (1440 min).

Three different kinetic models have been applied (Table 2) to analyze quantitatively the experimental data, namely using first-order (FO), second-order (SO), and double exponential (DEM) fits. The discrimination between the different models has been performed by means of the statistical criteria shown in Table 3 and the analysis of the residues⁴⁴ (Table S1, ESI‡). The better fit was obtained with the double exponential, with R-square of 0.99665 (compared to 0.96513 for SO and 0.96592 for SO) and residual sum of squares of 6.22 × 10⁻⁴ (0.00792 for FO and 0.00774 for SO), at pH = 1.0 (similar values at pH = 7.0).

Table 2 Kinetic models applied to the release profiles of VitA at 37 °C at pH = 1.0 and pH = 7.0 (Fig. 8)

First order
Second order
DEM

---

Table 3 Kinetic constant values for the VitA releasing from SEP–VitA

pH	k1 (min^−1)	k2 (min^−1)
1.0	0.0003 ± 0.0001	0.05 ± 0.01
7.4	0.0014 ± 0.0005	28 ± 6

---

According to literature,^45,46 the DEM model describes a mechanism consisting of two parallel reactions involving species adsorbed into two different sites. Since we have found that VitA is confined in the surface and the channel of the SEP, it seems reasonable to associate the fast and the slow process to the release of the drug from the surface and the channels of SEP–VitA, respectively.

The releasing processes (Table 3) depend to a great extent on the pH of the medium. In particular, k₁ and k₂ are approximately 5 and 560 times higher at pH = 7.0 than at pH = 1.0, while the ratio between the kinetic constant values for the fast (k₂) and the slow (k₁) processes is 20 000 and 160 for pH = 7.0 and pH = 1.0, respectively, reflecting the rate enhancement at lower pH values.

3.3 Mass spectrometry

To assess the stability of immobilized VitA during the release process mass spectrometry measurements were performed on the supernatant of the release medium. Experiments were carried by shaking VitA for 24 h at pH 1 in an aqueous solution with HCl, at 37 °C, under regular and N₂ atmospheres. For comparison purposes MCM–VitA was also shaken under similar conditions and N₂ atmosphere. As results discussed earlier in this work have shown that VitA is likely to suffer oxidation in MCM, this experiment was carried out under N₂ to determine whether VitA oxidation could be prevented.

The analysis of all supernatant media revealed a few compounds that could be identified as derivatives of VitA (Fig. 10). The mass spectra obtained from the different shaking experiments show dramatic differences. The supernatant solution resulting from shaking VitA in HCl aqueous solution under regular atmosphere displayed a very complex mass spectrum. There were no traces of VitA and the major identified components were VitA-CH₃CO and its oxidized derivatives (Fig. 10), VitAperox-CH₃CO and VitAepox-CH₃CO being the most abundant species. When the experiment was run under N₂ atmosphere, a much simpler spectrum was obtained, containing only VitA, VitA-CH₃CO (most abundant derivative) and VitAepox. The supernatant arising from shaking MCM–VitA yielded a similar mass spectrum, where VitA and VitA-CH₃CO (major product) were identified, as well as a small amount of VitAperox-CH₃CO.

Fig. 10 Molecules identified by mass spectrometry after the shaking VitA and MCM–VitA in highly acidic medium (HCl aqueous solution, pH 1) mimicking the gastric environment under regular and N₂ atmospheres after 24 hours at 37 °C.

The results of these experiments confirm that VitA is always degraded after the 24 h period suffering hydrolysis of the acetate moiety, as a consequence of the highly acidic pH medium, mimicking gastric environment, and oxidation reaction. Nevertheless, the extent of VitA oxidation can be reduced under N₂ atmosphere, or even almost eliminated in the case of the MCM–VitA hybrid material. This result shows that storing conditions are relevant for this material.

4 Conclusions

VitA has been successful loaded into three different inorganic materials, namely, the mesoporous material MCM-41, and the clays MMT and SEP.

Confinement of VitA has been observed and confirmed by means of complementary techniques, i.e., vibrational spectroscopy (FTIR), solid state NMR of ¹³C and ²⁹Si and thermogravimetric analysis (TGA). Powder XRD patterns not only confirmed loading of VitA, but allowed to propose the site of adsorption of the VitA in the different inorganic materials: VitA was immobilized on the inner surface of MCM 41 channels, intercalates in the interlayer spacing of MMT, and was mainly located on the surface of SEP as replacement of zeolitic water. An additional positive effect of the encasement of the drug into the clays consists of the protection, to a large extent, of the VitA against photodegradation.

The three hybrid materials released VitA differently under physiological pH mimicking conditions simulating the oral drug administration. Drug released from the MMT–VitA and MCM–VitA hybrids undergoes a degradation process, while that coming from the SEP–VitA was stable over time suggesting the ability of SEP to prevent the oxidation process. These results indicate that SEP could be a good candidate for oral VitA delivery systems.

Acknowledgements

IC and MLT gratefully acknowledge the financial support provided by the MIUR (PRIN grant 2010EARRRZ_003) and local funds from University of Palermo (ex 60%). IC thanks the University of Palermo for founding her stay in Portugal within the Ph.D. and Erasmus placement fellowship. Fundação para a Ciência e a Tecnologia (FCT), Portugal, is acknowledged for financial support by CDN, PDV, and MJC (UID/MULTI/00612/2013) and MJF (RECI/QEQ-QIN/0189/2012 and UID/QUI/00100/2013).

References

1. C. Debier and Y. Larondelle, Br. J. Nutr., 2005, 93, 153–174.
2. O. Aydin, F. Akyuz, N. Tekin, D. Burukoglu, H. Ozden, M. C. Ustuner and I. Degirmenci, J. Food Nutr. Res., 2014, 2, 9–16.
3. L. Perusek and T. Maeda, Nutrients, 2013, 5, 2646–2666.
4. A. Sommer and K. P. West, Vitamin A Deficiency: Health, Survival, and Vision, Oxford University Press, 1996.
5. G. Wald, Nature, 1968, 219, 800–807.
6. G. Duester, Cell, 2008, 134, 921.
7. K. Niederreither and P. Dollé, Nat. Rev. Genet., 2008, 9, 541–553.
8. R. Lotan, FASEB J., 1996, 10, 1031–1039.
9. J. R. Mestre, K. Subbaramaiah, P. G. Sacks, S. P. Schantz, T. Tanabe, H. Inoue and A. J. Dannenberg, Cancer Res., 1997, 57, 1081–1085.
10. Z. Zaman, P. Fielden and P. G. Frost, Clin. Chem., 1993, 39, 2229–2234.
11. E. M. De Almeida Vasconcelos Fonseca, C. E. A. Chagas, R. P. Mazzantini, R. Heidor, T. P. Ong and F. S. Moreno, Carcinogenesis, 2005, 26, 1940–1946.
12. I. Thompson, T. C. Fletcher, D. F. Houlihan and C. J. Secombes, Fish Physiol. Biochem., 1994, 12, 513–523.
13. E. C. Amar, V. Kiron, S. Satoh, N. Okamoto and T. Watanabe, Fish. Sci., 2000, 66, 1068–1075.
14. A. Cuesta, J. Ortuño, A. Rodriguez, M. A. Esteban and J. Meseguer, Immunology, 2002, 13, 279–291.
15. L. H. H. Hernandez, S. Teshima, S. Koshio, M. Ishikawa, Y. Tanaka and M. S. Alam, Aquaculture, 2007, 262, 444–450.
16. E. Villamor and W. W. Fawzi, Clin. Microbiol. Rev., 2005, 18, 446–464.
17. A. Sommer, Calif, 1997, 13, 484–485.
18. C. E. Harjes, T. R. Rocheford, L. Bai, T. P. Brutnell, C. B. Kandianis, S. G. Sowinski, A. E. Stapleton, R. Vallabhaneni, M. Williams and E. T. Wurtzel, Science, 2008, 319, 330–333.
19. H. Bennasir, S. Sridhar, M. Tec and T. T. Abdel-Razek, Int. J. Pharm. Sci. Rev. Res., 2010, 1, 68–73.
20. O. Dary and J. O. J. Mora, Nutrition, 2002, 132, 2927S–2933S.
21. R. M. Russell, Am. J. Clin. Nutr., 2000, 71, 878–884.
22. P. Sauvant, M. Cansell, A. Hadj Sassi and C. Atgié, Food Res. Int., 2012, 46, 469–479.
23. A. López-Galindo, C. Viseras and P. Cerezo, Appl. Clay Sci., 2007, 36, 51–63.
24. S. Lazarević, I. Janković-Častvan, D. Jovanović, S. Milonjić, D. Janaćković and R. Petrović, Appl. Clay Sci., 2007, 37, 47–57.
25. C. D. Nunes, A. A. Valente, M. Pillinger, A. C. Fernandes, C. C. Romão, J. Rocha and I. S. Gonçalves, J. Mater. Chem., 2002, 12, 1735–1742.
26. I. Calabrese, G. Cavallaro, C. Scialabba, M. Licciardi, M. Merli, L. Sciascia and M. L. Turco Liveri, Int. J. Pharm., 2013, 457, 224–236.
27. B. Giménez, M. C. Gómez-Guillén, M. E. López-Caballero, J. Gómez-Estaca and P. Montero, Food Hydrocolloids, 2012, 27, 475–486.
28. K. Ohtsuka, Chem. Mater., 1997, 9, 2039–2050.
29. B. Marler, U. Oberhagemann, S. Vortmann and H. Gies, Microporous Mater., 1996, 6, 375–383.
30. W. Hammond, E. Prouzet, S. D. Mahanti and T. J. Pinnavaia, Microporous Mesoporous Mater., 1999, 27, 19–25.
31. X. Liang, Y. Xu, G. Sun, L. Wang, Y. Sun and X. Qin, Chem. Eng. J., 2011, 174, 436–444.
32. S. J. Gregg and K. S. W Sing, Adsorption, Surface Area, & Porosity, Academic Press, London, New York, 2nd edn, 1982.
33. M. D. Alba, A. I. Becerro and J. J. Klinowski, J. Chem. Soc., Faraday Trans., 1996, 92, 849–854.
34. A. A. Romero, M. D. Alba, W. Zhou and J. J. Klinowski, J. Phys. Chem. B, 1997, 101, 5294–5300.
35. T. Lóránd, P. Molnár, J. Deli and G. Tóth, J. Biochem. Biophys. Methods, 2002, 53, 251–258.
36. P. Komadel, J. Bujdák, J. Madejová, V. Šucha and F. Elsass, Clay Miner., 1996, 31, 333–345.
37. A. Nasser, M. Gal, Z. Gerstl, U. Mingelgrin and S. J. Yariv, J. Therm. Anal., 1997, 50, 257–268.
38. G. A. Englert, Helv. Chim. Acta, 1975, 58, 2367–2390.
39. F. Farzaneh, J. Taghavi, R. Malakooti and M. Ghandi, J. Mol. Catal. A: Chem., 2006, 244, 252–257.
40. F. Zonta and B. Stancher, J. Chromatogr., 1984, 301, 65–75.
41. E. Azaz and R. Segal, J. Pharm. Pharmacol., 1977, 29, 322–323.
42. I. Arsic and G. Vuleta, Int. J. Cosmet. Sci., 1999, 21, 219–225.
43. M. C. Hermosin, J. Cornejo, J. L. White and S. L. Hem, J. Pharm. Sci., 1981, 70, 189–192.
44. M. Merli, L. Sciascia and M. L. Turco Liveri, Int. J. Chem. Kinet., 2010, 42, 587–607.
45. L. Sciascia, M. L. Turco Liveri and M. Merli, Appl. Clay Sci., 2011, 53, 657–668.
46. H. Qiu, L. Lv, B. Pan, Q. Zhang, W. Zhang and Q. Zhang, J. Zhejiang Univ., Sci., A, 2009, 10, 716–724.

---

Footnotes

† In memoriam Prof. Maria Liria Turco Liveri.

‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra12026a

---