Monocyclic β-lactams loaded on hydroxyapatite: new biomaterials with enhanced antibacterial activity against resistant strains

The development of biomaterials able to act against a wide range of bacteria, including antibiotic resistant bacteria, is of great importance since bacterial colonization is one of the main causes of implant failure. In this work, we explored the possibility to functionalize hydroxyapatite (HA) nanocrystals with some monocyclic N-thio-substituted β-lactams. To this aim, a series of non-polar azetidinones have been synthesized and characterized. The amount of azetidinones loaded on HA could be properly controlled on changing the polarity of the loading solution and it can reach values up to 17 wt%. Data on cumulative release in aqueous solution show different trends which can be related to the lipophilicity of the molecules and can be modulated by suitable groups on the azetidinone. The examined β-lactams-HA composites display good antibacterial activity against reference Gram-positive and Gram-negative bacteria. However, the results of citotoxicity and antibacterial tests indicate that HA loaded with 4-acetoxy-1-(methylthio)-azetidin-2-one displays the best performance. In fact, this material strongly inhibited the bacterial growth of both methicillin resistant and methicillin susceptible clinical isolates of S. aureus from surgical bone biopsies, showing to be a very good candidate as a new functional biomaterial with enhanced antibacterial activity.


General Methods.
All chemicals and solvents were of analytical grade; anhydrous solvents were obtained commercially and used without further drying. Azetidinones 1 and 2 are commercially available (Sigma-Aldrich) and used as such. Deionized water was obtained from a Millipore analytical deionization system (MilliQ). ATR-FTIR spectra were recorded on an Alpha FT IR Bruker spectrometer with platinum ATR single reflection diamond module. As reference, the background spectrum of air was collected before the acquisition of each sample spectrum. Spectra were recorded with a resolution of 4 cm 1 , and 32 scans were averaged for each spectrum (scan range 4000-450 cm 1 ). TLC: Merck 60 F254 plates. Column chromatography: Merck silica gel 200-300 mesh. HPLC-MS: Agilent Technologies HP1100 instrument, equipped with a ZOBRAX-Eclipse XDΒ-C8 Agilent Technologies column, mobile phase: H 2 O/CH 3 CN, 0.4 mL/min, gradient from 30 to 80% of CH 3 CN in 8min, 80% of CH 3 CN until 25 min, coupled with an Agilent Technologies MSD1100 single-quadrupole mass spectrometer: full scan mode from m/z = 50 to 2600, scan time 0.1 s in positive ion mode, ESI spray voltage 4500 V, nitrogen gas 35psi, drying gas flow 11.5mL/min, fragmentor voltage 20 V. 1 H and 13 C spectra were recorded with an INOVA 400 or Gemini 300 instruments with a 5mm probe. All chemical shifts are quoted relative to deuterated solvent signals (δ in ppm and J in Hz).The solid state NMR spectra were obtained on a 500MHz Agilent DD 2 spectrometer by using a 22 L rotor. The 13 C and 1 H spectra were recorded at a MAS frequency of 6kHz. TGA analysis was carried out heating under air stream from 40.0°C at 10.0°C/min in a TGA7 Perkin Elmer instrument. XRD patterns were collected by using a PANalytical X'PertPro diffractometer equipped with a fast solid state X'Celerator detector and a copper target ( = 0.15418 nm). Data were acquired in the 9 -60° 2 interval, by collecting data for 50 s at each 0.05° step. For TEM investigations, a small amount of powder was dispersed in ethanol and submitted to ultrasonication. A drop of the calcium phosphate suspension was transferred onto holey carbon foils supported on conventional copper microgrids. A Philips CM 100 transmission electron microscope operating at 80 kV was used.

Synthesis of hydroxyapatite.
It was carried out using CO 2 -free distilled water in N 2 atmosphere [Bigi et al., 2004]. 50 ml of 1.08 M Ca(NO 3 ) 2 4H 2 O solution at pH adjusted to 10 with NH 4 OH was heated at 90°C and 50 ml of 0.65 M (NH 4 ) 2 HPO 4 solution, at pH 10 adjusted with NH 4 OH, was added dropwise under stirring. The precipitate was maintained in contact with the reaction solution for 5 hours at 90°C under stirring, then centrifuged at 10,000 rpm for 10 minutes and repeatedly washed with CO 2 -free distilled water. The product was dried at 37°C overnight.

Loading of Azetidinones
The loading of azetidinones on HA was conducted in H 2 O (method A) or H 2 O/organic solvent mixtures (method B), see below. Loading processes were set up in a parallel-synthesis fashion with a Carousel 6 reaction station using two necks round bottom flasks (50 mL) with a water-cooled aluminum head. This apparatus is well suited to keep constant some experimental conditions, such as stirring and warming, important parameters in heterogeneous phase processes, and, moreover, it improved and speeded up optimization steps. For loading on HA in water alone was used method A, in the presence of a co-solvent method B, instead. Method A: 200 mg of HA nanoparticles were suspended in 2 mL of H 2 O and warmed at 40°C under magnetic stirring. Azetidinone (50 mg) was added in one portion to the suspension which was then warmed up to 70°C. Reaction mixtures were controlled via TLC on the supernatant solution to monitoring the starting azetidinone disappearance. After 4 h the mixture was quantitatively transferred with 1 mL of H 2 O /MeCN (1:1) in an open test tube and centrifugated for 1 min at 700 rpm. The solid phase was perfectly separated and the supernatant aqueous phase was collected and extracted with dichloromethane (1x3 mL). The aqueous and dichloromethane phases were separately evaporated and analyzed to quantify the unloaded azetidinone and its distribution in the two phases. Data were expressed as loading efficiency % back-calculated from the added up residues obtained in DCM and H 2 O in comparison with the amount of azetidinones in the loading solution by the equation: where: LE = loading efficiency %; A = amount (g) of azetidinone in the loading solution; rw = residue (g) of azetidinone in water extract; rDCM = residue (g) of azetidinone in dichloromethane(DCM) extract.
The solid functionalized HA material was oven dried at 35°C for 24 h, and kept in dessicator (CaCl 2 ) for 24 h before the analyses. Method B: 200 mg of HA nanoparticles were suspended in 1 mL of H 2 O and warmed at 40°C under magnetic stirring, then azetidinone (50 mg) was solubilized in 1 mL of organic solvent and added to the suspended HA, then the mixture was warmed at 70°C under stirring for 4 h. Reaction mixtures were controlled via TLC on the supernatant solution to monitoring the starting azetidinone disappearance. The work-up procedure was as for Method A.
Loading amount of the azetidinone molecules on HA was evaluated through thermogravimetric analysis as difference between the total weight loss measured between 38 and 800°C for each loaded sample and that measured for pristine HA. Moreover, the determination was also performed through the evaluation of the intensity of the adsorption band of C=O at 1790 cm -1

Statistical analysis.
The statistical evaluation of data was performed using the software package SPSS/PC+StatisticsTM 23.0 (SPSS Inc., Chicago, IL USA). The study is the results of three independent experiments and data are reported as mean  standard deviations (SD) at a significance level of p<0.05. After having verified normal distribution and homogeneity of variance, a one-way ANOVA was done for comparison between groups. Finally, post hoc multiple comparison test and Pearson correlation test were performed to detect significant differences among groups and controls.

Antibacterial susceptibility testing 2.7.1 Bacterial strains.
The in vitro effect of the HA nanocrystals loaded with the monocyclic azetidinones was evaluated against Gram-positive and Gram-negative reference bacterial strains: Staphylococcus aureus (ATCC 25923), Escherichia coli (ATCC 25922). In addition, clinical isolates obtained from surgical bone biopsies were included in the study and they were categorized based on their antimicrobial susceptibility to methicillin. The tested strains were isolated on BD Columbia Agar with 5% sheep blood (Becton Dickinson, Germany) and confirmed by MALDI-TOF MS (Bruker Daltonik, Germany) (Croxatto et al. 2012). Their susceptibility was analyzed by the Vitek2 semiautomated system (bioMerieux, France) and interpreted following EUCAST guidelines. MRSA strains were confirmed by growth on BD oxacillin screen agar (Becton Dickinson, Germany), as in