Use of Gelatin as Tannic Acid Carrier for Its Sustained Local Delivery

Background : Polyphenols are macromolecules that play a pivotal role in plant protection against external aggression from microorganisms or radiations. Their antimicrobial properties have raised interest in the medical community for several years now. Among them, tannic acid is one of the cheapest polyphenols extracted from plants. However, its broad use is impeded because it is unstable, with rapid hydrolysis leading to byproducts such as epigallocatechin with less activity. In this work, we developed a porous material to deliver locally tannic acid. Methods : The preparation of this matrix relied on the interaction between gelatin and tannic acid in an aqueous medium. Structure properties of the developed matrix were investigated through rheology measurement and by scanning electron microscopy (SEM). Tannic acid quantity and release over time were measured. Antimicrobial activity was tested in vitro on E. coli and S. aureus. Results : The presence of two tannic acid populations, one free to diffuse across the scaffold and the second stabilized by gelatin with a more structural activity, was demonstrated. We showed that the formulation is possible above a molar ratio of 1:15 (gelatin:tannic acid). Gelatin was used here as a carrier for the tannic acid. Tannic acid content could be modulated depending on its initial concentration during the preparation.


INTRODUCTION
Polyphenols are macromolecules that play a pivotal role in plant protection against external aggression from microorganisms or radiations [1]. The family is broad with epigallocatechin, gallic acid derivatives, tannins, flavonoids, isoflavonoids, and anthocyanins to cite only a few. However, it is quite unstable with a rapid degradation into smaller macromolecules such as gallic acid in an oxidative environment. Its oxidation tends to diminish its potential as an antimicrobial agent. It can also be cytotoxic depending on its concentration. Hence, it has to be delivered at a concentration high enough for antimicrobial activity but with antimicrobial properties [4]. TA is known to interact strongly with gelatin, mostly via the proline residues of the protein [5]. Other strategies aimed at grafting of tannic acid via their hydroxyl groups onto the carboxyl group of polymers via esterification methods to produce poly(methacrylic acid) microparticles releasing TA for more than 40 h [6].
Tannic acid-based particles were successfully produced from the macroto nanosize range. Such particles incorporated in porous poly(HEMA) cryogels helped to add antimicrobial properties to this future wound dressing material [7]. Apart from its biological properties, tannins in general and tannic acid in particular are well known for their interactions with proteins. They strongly interact with proteins via hydrogen bonds.
This explains why their use in the leather industry was so prominent. It can be used as a crosslinking agent to stabilize nanoparticles or scaffolds.
However, its use has to be modulated to insure the functionality of the materials. Indeed, by depositing tannic acid and alkaline phosphatase on substrates layer-by-layer, we previously showed that alkaline phosphatase was inactivated when the concentration of tannic acid was increased [8].
In this study, we developed a method for the local delivery of tannic acid using gelatin (G) as a carrier. It resulted in a material that could be used for example as a wound dressing. Tannic acid was used both for the shaping of the gelatin matrix and as a therapeutic agent. The preparation process relied on the interaction of tannic acid with the protein to form a precipitate when raise above a threshold in the tannic acid:gelatin molar ratio. We showed that these matrices after crosslinking with formaldehyde can be manipulated and are stable up to seven days. More interestingly, the designed matrices showed a slow release of tannic acid.
Such wound dressing materials present a good antimicrobial activity.
Their conservation was ensured by lyophilization.

G/TA Matrix Preparation
All preparations were made in a citrate/phosphate buffer at pH = 7.00 to work in close physiological conditions. Citrate/phosphate buffer was of tannic acid in water is 100 mg·mL −1 at ambient temperature. TA has an isoelectric point close to 8.5 [9] and is hence non-charged in the experimental conditions where the TA/gelatin blends were prepared. The gelatin solutions were prepared at 2 or 10 mg/mL depending on the size of the matrix prepared. Gelatin solutions were heated at 37 °C to ensure complete dissolution. Once dissolved, the gelatin solution was cooled to room temperature before use. G/TA matrices were prepared by mixing equal volumes of TA and gelatin solutions followed by 30 s of vortexing.
The obtained solution was poured in a mold or in a well of a multiwell plate and centrifuged at 1200 rpm for 15 min in a plate centrifuge (Allegra The TA content in the matrix (TAm), in mg, was calculated with Equation (1) where TA0 is the quantity of tannic acid used during the matrix preparation and TAs is the quantity of TA measured in the supernatant after completion of the centrifugation. TAm was plotted as a function of TA0 to get an adsorption isotherm.

Measurement of Tannic Acid Release from the G/TA Matrices
Tannic acid release from the matrices was measured by UV-Visible spectroscopy. After preparation of the G/TAx matrices, they were incubated at room temperature in citrate/phosphate buffer at pH = 7.00.
The release of TA was measured in the buffer supernatant at various time points. One hundred microliters were withdrawn, and the absorbance measured at λ = 277 nm with a SAFAS mc² spectrophotometer.  [10].

Growth inhibition in planktonic culture
Antimicrobial activity measurements were performed on matrices

G/TAx Matrix Characterization
G/TAx matrices were prepared based on the protocol described in for 24 h at room temperature. After three rinsing steps, the matrices are lyophilized (6).
Both the quantity of gelatin used and the shape of the mold determine the size and shape of the matrix. We crosslinked the matrix after it was centrifuged to stabilize it. Indeed, protein-tannic acid interactions are non-covalent, mainly hydrogen bond and hydrophobic interactions, and can be destabilized by several parameters such as pH or osmotic pressure.
The range of molar ratios tested was from 1:15 to 1:200.
Scanning electron microscopy micrographs of the matrix after lyophilization showed that the obtained matrix is porous. The pores seem to be interconnected. However, more interestingly, the walls of the matrix show a granular aspect at high magnification (60,000×) ( Figure 2).    The same behavior was recorded for S. aureus. For the same molar ratio, we recorded a better growth inhibition of G/TA15, G/TA30, and G/TA50 for S.
aureus than for E. coli with more than 90% inhibition.
Finally, as we are aware that tannic acid, a hydrolysable polyphenol, is not stable in oxidative conditions, we tested the effect of lyophilization on the antimicrobial properties of the G/TA matrices. G/TA70 matrices were prepared as before but followed by a lyophilization process. Before their use, they were kept in dry conditions and protected from light. Their

DISCUSSION
We succeeded in preparing gelatin/tannic acid matrices that represent a cost effective and easy way to produce a tannic acid delivery system. Our strategy was based on the natural interaction of tannic acid with proteins through hydrogen bonds and hydrophobic interactions. With such a method, we were able to produce matrices with gelatin concentration that do not allow gel formation, such as 1% w/v solutions [11]. As the interaction between gelatin and tannic acid relies on non-covalent binding, we can expect, depending on the pH and ionic strength, for example, the release of non-strongly-bound TA [12]. The process seems to be overly simple but different parameters have to be taken into account such as pH and ionic strength because they can affect the interaction between both components and therefore the structuration of the matrix and the release of tannic acid. First, not only hydrogen bonding takes place, but, according to some other studies, hydrophobic and electrostatic interactions also occur. At pH = 7.00, lower than the pKa value of tannic acid, electrostatic interactions are possible with the negatively charged gelatin of type A (pHi = 9). However, the ionic strength has to be close to physiological value.
Indeed, when ionic strength increases, gelatin takes a more compact conformation due to internal charge screening. It avoids interchain repulsion and favors the interaction with tannic acid [13]. Hydrophobic bonding occurs between L-proline in gelatin and the phenolic group in tannic acid. This interaction mode could generate some cavities in the  [13]. Overall, all these weak interactions have the potential to generate macromolecules mobility. In our case, we observed a modification of the area of our matrices over time (Supplementary Figure S2) with a diminution of 10% after four days in buffer at pH = 7.00 and 35% after 23 days. It shows that the system is not at equilibrium once produced. We observed that the shrinking of the matrix is dependent on the medium, being more pronounced in DMEM, a cell culture media, than in citrate/phosphate buffer. It suggests that the interactions undergone by gelatin are influenced by the nature and concentration of the present electrolytes, as observed in the literature [14].
It has been demonstrated that the interaction of tannic acid with protein is also dependent on its concentration. Under the critical micellar concentration (CMC), the interaction is specific, and we can find few tannic acid molecules interacting with proteins. Above the CMC, tannic acid form aggregates and interaction with proteins are less specific. The interaction of tannic acid aggregates can be done with different protein in their surrounding media. Large TA-protein aggregates can be formed [15]. In our case, the concentration of tannic acid was always above the CMC of tannic acid (3.10 × 10 −4 mol/L, i.e., 0.53 mg/mL) [15]. We thus produced large aggregates that could be centrifuged to prepare the material in the form of a pellet. This also explains the porous aspect of the matrix, as found at high magnification in SEM.
We observed a release of tannic acid over time from each formulation.
Such a release was observed during about four days with an initial burst release lasting 24 h followed by a slower release up to four days.
Considering the quantity of tannic acid left in the matrices after four days, we observed that all matrices made from a molar ratio above 50 between TA and gelatin led to the same amount remaining in the matrix. Hence, above a molar ratio of 50, there are two tannic acid populations: one can be considered as strongly bound to gelatin and the other considered as mobile. Below this threshold value, TA release was minimal. The tannic acid released over time was therefore present in the water content of the matrices or resulted from tannic acid aggregates that disassembled. In fact, TA spontaneously forms solution aggregates at high concentrations [16].
From a practical point of view, using gelatin as a binder for tannic acid implies using molar ratios above 50. Otherwise, we cannot have two populations of tannic acid; only the "structural" one is available with no diffusible molecules. The increase of the G:TA molar ratio allows an increase in polyphenol release. TA is also reported as a crosslinking agent [5,22,23]. It could help the overall structuration of the scaffold. Researchers using tannic as a crosslinker tend to use far lower TA:protein ratios than what we did in this study. At pH = 8.00, covalent binding between tannic acid and gelatin can occur between the amine group of gelatin (lysine, arginine, and histidine residues) and the oxidized phenolic group of tannic acid [24]. However, at this pH, oxidation can occur rapidly and would diminish the antimicrobial activity, which is why we decided to work at a controlled pH = 7.00. Our matrices remained stable for at least 23 days.
Use of tannic acid as an antimicrobial agent tends to be tricky because of its instability. Indeed, under oxidative environment, it degrades into smaller molecular weight products by hydrolysis [25]. Such byproducts lose their antimicrobial properties. In our matrices, oxidation of the tannic acid could be easily seen by a color change from a yellowish color to dark green. The same observation can be made when oxidation is triggered by sodium periodate [8].
For this reason, we tested the antimicrobial activity after lyophilization of the matrix. Indeed, once lyophilized, tannic acid can be conserved for a longer period of time. No loss in antimicrobial activity was recorded with such preparation.
Tannic acid could have some cytotoxicity related to its abilities to interact with membrane, inhibit enzyme activity, and chelate metal especially iron. We tested cytotoxicity on NIH3T3 mouse fibroblasts. The threshold concentration above which we recorded cytotoxicity is 125 µM (representing 200 mg/mL) (Supplementary Figure S3). Considering the concentration of tannic acid released over time, we are in a range that is not toxic for eukaryotic cells. However, further in vitro testing is mandatory before testing such material in vivo in clinical condition such as wound closure model.
The easy handling of this porous material to give it various shapes and sizes is promising for future use. Indeed, applications could be found in the preparation of wound dressings with anti-microbial properties. The material can be manipulated with tweezers and can then be put in contact with a wound and comply with the actual protocol of wound dressing.
Research tends to avoid the use of antibiotics to diminish the appearance of bacterial resistance. For example, we can cite the use of silver derivatives in various materials such as cellulose or carrageenan-based  [26,27]. Such cost-effective antimicrobial agents delivering matrices could also be used in periodontology. Indeed, adjuvant therapy to scaling and root planning therapy using a degradable material loaded with chlorhexidine inserted in periodontal pockets is already available in the clinic (Periochip ® ) [28,29]. The material designed herein could be an alternative in the case of chlorhexidine allergy.

CONCLUSIONS
Herein, we show an easy and cost-effective way to prepare a porous material that can be used to locally deliver tannic acid. Conserved as lyophilized, tannic acid keeps its activity and has been proved to be antimicrobial against S. aureus and E. coli once released into the surrounding media after matrix hydration. The content of tannic acid can be modulated easily, and thus also the concentration of tannic acid release.
It is thus suitable to fight various bacteria even if their sensibility to tannic acid varies. The method allows producing matrices with various shapes and sizes. It therefore could be used in different applications from wound dressing to periodontology.

SUPPLEMENTARY MATERIALS
The supplementary materials on Multilayer Perceptron Network are available online at https://doi.org/10.20900/pf20200002.

DATA AVAILABILITY
The dataset of the study is available from the authors upon reasonable request.

AUTHOR CONTRIBUTIONS
FM and VB designed the study. FR and JH performed the antimicrobial experiments. EB, JH and FM performed the matrix characterization. FM and VB analyzed the data. FM and VB wrote the paper with input from all authors.

CONFLICTS OF INTEREST
The authors declare that there is no conflict of interest.