**1. Introduction**

The research field dealing with the development of controlled drug delivery systems has been of relevant scientific interest since the 1970s and has grown and diversified rapidly in recent years, in

particular thanks to the benefits it brings to healthcare; furthermore, it covers a large market segmen<sup>t</sup> [1]. In general, the effectiveness of drug therapy is the main objective of controlled release systems [1,2], with a corresponding (i) reduction of the number of drug administrations; (ii) improvement in therapeutic activity [3,4]; (iii) consequent reduction of the intensity of side effects; and (iv) elimination of specialized drug administration [5]. This pharmaceutical technology, especially in recent years, has seen application in other fields ranging from cosmetics [6] to agriculture [7], including textiles [8,9], as an interesting and innovative application. Indeed, textile fabrics, thanks to their biocompatibility, breathing structure and absorptive capacity, are of grea<sup>t</sup> interest as a medium (for ex vivo applications) for controlled release of drugs, active principles or aroma substances of particular comfortability [10]. Several examples are nowadays in common use, such as the well-known transdermal patches or textile costumes, generally characterized by different layers in which the release of a specific drug substance, deposited on the textile surface, is activated by stimuli such as temperature, humidity, enzyme, perspiration types or friction [11].

In general, in controlled release systems, biocompatibility and controllability are important features; furthermore, in terms of biocompatibility, the carcinogenicity, toxicity, teratogenicity and mutagenicity are important elements to be controlled [10]. The use of textile fabrics for the realization of controlled release systems presents several advantages but on the other hand, some disadvantages with respect to the oral administration of substances. Indeed, this administration represents an attractive and easier therapy for the patient due to its efficacy since the drug avoids both the digestive apparatus and the hepatic metabolism that reduce the concentration of the drug. Furthermore, it requires lower dosages due to the higher diffusion through tissues, which correspond to lower social costs of therapies [11]. On the other hand, the disadvantages are related to the diffusion rate of the drug as function of its molecular structure and body surface administration [11]. Different methods have been employed in order to develop better textile-based delivery systems (i.e., bandages, patches), with good controllability, biocompatibility and active species entrapment/release by use of host-guest molecules (cyclodextrins [12], aza-crown ethers, fullerenes) or doping functional molecules (ion-exchange; drug-loaded hollow; nanoparticles; bioactive) [10]. Several C–C polymer heteroatom-containing (i.e., N, P, Si) backbones for controlled release applications have been tested and considered in order to improve the drug therapy effectiveness [5]. All the developed release polymeric systems have been shown to act through mechanisms of temporal controlled release, such as drug-delayed dissolution, diffusion-controlled, and drug solution flow control after interaction with environmental water or by reacting to specific skin stimuli.

The sol–gel method has been shown to be a useful method in the preparation of functional nanostructured coatings for textiles, thus combining the entrapment/encapsulation of bioactive compounds, biomolecules and their controlled release [13].

In our previous studies, we have already shown that nano-hybrid sol–gel-based coatings feature abrasion resistance, tensile strength and elongation properties of the treated fabrics [13,14]. These peculiar characteristics may be combined with a proper doping molecule, such as a dye [15–21], an antimicrobial [22], a hydrophobic [23–25] or a flame-resistant molecule [26,27], with the aim of improving the textile surface properties and making a functional nano-hybrid coating.

With the aim of developing a functional sol–gel-based coating suitable for medical application, we thought worthwhile to make a silica sol containing the *3*-glycidoxypropyltriethoxysilane (GPTES, hereafter "G"), as silica cross-linker precursor, and a PEA derivative, the *N*-Palmitoyl-(*4*-nitro-phenyl)-amine (hereafter PNPA), whose anti-inflammatory and antioxidant properties have already been tested and compared with other analogue molecules [28]. The sol was successfully applied on cotton surfaces and, after drying and curing, a stable and uniform PNPA-containing silica-based coating was obtained, as confirmed by morphological studies (SEM and AFM microscopy). As already shown in previous studies for halochromic dyestuff [20], the PEA derivative results firmly encapsulated into the 3D hybrid silica layer in absence of external stimuli (i.e., variable pH conditions), thanks to non-covalent and weak interactions (i.e., hydrogen bonds

and van der Waals interactions) acting between the non-polar active molecule and the alkoxysilane hosting network.

Finally, the functionally coated cotton samples were employed for in vitro di ffusion studies with the aim of testing their ability to release the synthesized PEA derivative in a controlled manner compared to a standard solution of the molecule. This e ffectively prepared functional hybrid system, based on the non-covalently immobilized PNPA, showed good results so that it can be considered a suitable sca ffold for fabrics in drug release applications, thus providing useful insights in the design and the development of medical textiles.

### **2. Materials and Methods**

*N*-Palmitoyl-(*4*-nitro-phenyl)-amine (PNPA) was synthetized according to a synthetic strategy described in previous researches [28]. *3*-glycidyloxypropyltriethoxysilane (GPTES) and methanol were purchased from Wacker and Aldrich, respectively, and used without further purification. Two cotton scoured and bleached 100% plain-weave textile fabrics (coded COL and CO H) kindly supplied by Albini S.p.A. (Albino, Italy) and Mascioni S.p.A. (Cuvio, Italy), respectively, were used for this research. The fabrics showed di fferent mass per unit area (COL = 119 g/m<sup>2</sup> and CO H = 331 g/m2, respectively). Cotton fabrics were washed before treatment at pH 7 and at 40 ◦C for 20 min in a non-ionic detergent, rinsed several times with de-ionized water and then dried. The cleaned samples were conditioned at 20 (± 1) ◦C and under standard atmospheric pressure at 65 (± 2)% relative humidity for at least 24 h prior to all experiments.

PNPA (25 mg) was dissolved in 40 mL of methanol through ultrasonication and left under continuous stirring. Then 2 mL of a 1 M aqueous sol–gel solution of GPTES were added drop by drop to the clear methanol solution of antioxidant molecule, thus resulting in a final GPTES sol concentration of 0.05 M (1:0.034 molar ratio with respect to PNPA). The obtained solution (G-PNPA sol) was ultrasonicated and left at room temperature under stirring for at least 90 min. The same reaction was also carried out in absence of the antioxidant molecule in order to obtain a reference GPTES sol sample. Both solutions were applied separately onto cotton textile (10 cm × 10 cm) through a two-roll laboratory padder (Werner Mathis, Zurich, Switzerland) at a nip pressure of 2 bar, then dried (80 ◦C for 5 min) and cured (100 ◦C for 1 min) in an electric laboratory oven.

PNPA, G-PNPA and G sols were fully investigated through FT-IR spectroscopy and Nuclear magnetic resonance (NMR). Untreated and treated textiles were characterized by FT-IR spectroscopy, Scanning Electron Microscopy (SEM) coupled to energy dispersive X-ray (EDS) and Atomic Force Microscopy (AFM).

The PNPA controlled release from the two prepared textiles, COL\_G-PNPA and CO H\_G-PNPA, was investigated by performing in vitro di ffusion studies using Franz di ffusion cells and according to the experimental protocol reported in a previous work [29]. For this purpose, Strat-M ® membranes (25 mm discs, Cat. No. SKBM02560, Merck Millipore, Darmstadt, Germany) were positioned between the donor and the receptor compartments of each Franz cell and the experiments were carried out at 37 ± 0.5 ◦C. The two tested items, COL\_G-PNPA and CO H\_G-PNPA, were placed on the Strat-M ® membrane with the GPTES layer facing towards the acceptor chamber. Then, the Franz cell compartments were fixed together and filled with 0.5 and 5.5 mL of phosphate bu ffer at pH 7.4 (10−<sup>3</sup> M), respectively. The content of the receptor chamber was withdrawn at di fferent times, such as 1, 2, 4, 6 and 24 h, for UV-Vis analysis and replaced with phosphate bu ffer. The same experimental conditions were applied to a control sample consisting of a standard PNPA solution. The in vitro di ffusion studies were carried out in triplicate and the obtained results were expressed as di ffused amount (%).

FT-IR spectra were performed by a Thermo Avatar 370 equipped with an attenuated total reflection (ATR) accessory and using a diamond crystal as internal reflectance element. Spectra were acquired with 32 scans and in the range from 4000 to 550 cm<sup>−</sup><sup>1</sup> with a resolution of 4 cm<sup>−</sup>1.

One- and two-dimensional NMR experiment were recorded in methanol-*d*4 at 298.2 (±0.1) K on Bruker ARX-300, equipped with a 5 mm gradient probe and operating at 300.1 MHz for 1H nucleus. All chemical shifts are shown in parts per million (δ/ppm), downfield to tetramethylsilane (Me4Si) as an internal standard (δ = 0.0 ppm), or referenced to the residual protiated solvent signal such as in methanol-*d*4 ( 1H NMR: 3.30 ppm). 1H NMR signals were assigned by means of two-dimensional homonuclear NMR gradient experiments (gCOSY, gNOESY), acquired using standard Bruker pulse sequences.

SEM morphology and SEM-EDS of the investigated samples were obtained using a FEI Quanta FEG 450 microscope. An operating voltage of 5 kV in low vacuum was used for SEM images. EDS analysis was conducted with an operating voltage of 20 kV, always in low vacuum. Samples were fixed on aluminum sample holders by means of a graphitic adhesive.

AFM characterization was performed using a stand-alone SMENA head by NTMDT, equipped with a Bruker silicon probe model NCHV working in semi-contact mode. The samples were fixed onto metallic stubs using a small piece of double-sided scotch tape and studied at RT.
