Lignocellulosic membrane grafted with 4-vinylpiridine using radiation chemistry: antimicrobial activity of loaded vancomycin

Lignocellulosic membranes were obtained from the skin of Agave salmiana (MS), a digital Vernier was employed to perform measures, and found thickness was 104 ± 5.4 μm and density 0.05 mg cm⁻³. The membranes were rinsed in a basic solution (0.1 M NaOH) and subsequently soaked in a mixture of water/methanol (1/1 vol%) to remove biomass and other residues (arabinoses, hemicelluloses, etc.). Vancomycin was acquired from Fagron (Colombia); 4VP, NaOH, and methanol were purchased from Sigma Aldrich (St. Louis, MO, USA). The monomer 4VP was distilled at reduced pressure before its use.

The lignocellulosic membrane (MS) was cut into 1.2 × 5 cm pieces and dried at room temperature under reduced pressure. The MS membranes were placed in an ampoule containing 7 mL of 4VP solution, varying the monomer concentration (0–13 vol%) in a mixture of water/methanol (67/33 vol%). Air was removed by argon bubbling (20 min) and sealed. Subsequently, the ampoules were exposed to a ⁶⁰Co gamma-rays source at different doses, up to 60 kGy. After gamma irradiation, the residual monomer and homopolymer were removed by stirring in water/methanol (70/30 vol%), changing the solvent twice each 24 h, followed by vacuum drying at 40 °C until constant weight.

The grafting yield (G%) was calculated using Eq. 1:where W_f and W_i are the weights after and before graft, respectively.

Kruss DSA 100 drop shape analyzer was used to measure the contact angle on the surface of different samples.

Flat and dry films were placed inside the chamber. External angle of a drop of water was measured at 1 min using the native software; each experiment was performed in triplicate.

MS and MS-g-4VP membranes were immersed in 20 mL of a phosphate buffer at different pH values (2–11) at 25 °C for 24 h. The membranes were extracted, removed excess buffer with filter paper, and weighed. The critical pH of the membranes was calculated by recording the degree of swelling (DS) at each pH value using the following equation:where W_d and W_w are the weights of the films before and after swelling, respectively.

Mechanical properties of the samples were studied by applying uniaxial tension test, firstly the samples were soaked in distilled water before cutting the dumb-bell shape, and subsequently the test specimens were dried at vacuum before carrying out the testing following the method described in ASTM D1708. All tests were conducted on an INSTRON 1125 (Instron Inc., MA, USA) universal tensile testing machine at a crosshead speed of 10 mm/min. All experiments were carried out in triplicate.

Pieces of grafted membranes (MS-g-4VP) (1 × 1.2 cm) were placed in vials with 5 mL of a vancomycin aqueous solution (50 µg/mL) at buffer [0.01 M] pH = 5, 7, and 9, and the load was monitored at room temperature and at pre-established time intervals. Samples were taken out from the medium, and measured the UV absorbance of the solution at 280 nm; the amount of loaded drug was calculated using Eq. 2, which was obtained from experimental values using a calibration curve and the membrane area. The experiment was carried out in triplicate.where A₁ and A₂ represent the initial and final absorbance of the loading medium.

Drug release was evaluated by transferring the loaded membranes to vials with 5 mL of buffer [0.01 M] at different pH = 5, 7, and 9, and kept under stirring at 180 rpm and 25 °C. The released drug was monitored by UV, measuring the absorbance at 280 nm at different times, and calculated the drug concentration using a calibration curve. The tests were carried out in triplicate.

In addition, the drug release from MS-g-4VP membranes was assessed using representative mathematical models (Dash et al. 2010) as follows:where M_t is the fraction of drug released at each time (t), and K₀, K₁, and K_h are the zero-order, first-order, and Higuchi kinetic constants, respectively. Furthermore, the drug release mechanism was studied using the semi-empirical Korsmeyer–Peppas model (Korsmeyer et al. 1983).K is a kinetic constant that considers the structural and geometric characteristics of the device; n is the release exponent, which is indicative of the release mechanism; if the value of n is 0.45 ≤ n corresponds to a Fickian diffusion, 0.45 < n < 0.89 to a non-Fickian transport, n = 0.89 to Case II transport, and n > 0.89 to super Case II transport mechanism.

The capability of the drug-loaded membranes to inhibit bacterial growth was tested against Staphylococcus aureus (S. aureus) and Escherichia coli (E. coli). Petri plates containing Müller–Hinton agar were seeded in concentration for S. aureus 1.2·10⁹ CFU/mL and E. coli 3.4·10⁹ CFU/mL (where CFU is a colony-forming unit). Vancomycin-loaded MS-g-4VP and blanks of MS-g-4VP pieces (⬂1 × 1.2 cm and Ø  0.5 cm diameter) were tested. Plates containing the sample were incubated at 37 °C for 24 h. The inhibition zone was measured using a scale. Each experiment was performed in triplicate.

4VP grafting was successfully carried out on MS membranes at a low monomer concentration (2.4% 4VP) with approximately ≈ 13% grafting, with similar results at the lowest irradiation dose of 10 kGy. However, the effect of concentration did not show a maximum of grafting, indicating that the degree of modification was increased by increasing the monomer concentration without requiring to increase in the radiation dose above 20 kGy, as shown in Fig. 1a. In the case of the irradiation dose kinetics with a fixed concentration of 5% (v/v) of monomer, it was not possible to increase the grafting degree because monomer homopolymerization is favored at an irradiation dose of 20 kGy, reaching an average maximum graft degree of ≈ 35% and remaining constant at higher doses. Therefore, irradiation doses above this value are an excess that can degrade the matrix and promote other reactions such as crosslinking of the grafted polymeric chains (Fig. 1b). In general, the grafting degree on MS membranes depends on these two variables, yield increasing as monomer concentration and irradiation dose increased.

The hydrophilicity and pH responsiveness of the MS-g-4VP membranes were measured by contact angle and swelling in water at different pH values. The contact angles were obtained on both sides of the membranes, showing a side more hydrophobic than the other one, as was observed in the non-grafted MS membranes (Fig. 2a). The behavior of hydrophobicity was explained by the surface structuring and composition of each side of the MS membrane, because the internal part has a higher content of nitrogenous and oxygenated groups from the possible biomass residues, which increased the hydrophilicity (Silva et al. 2016), in addition to a more complex surface structuring (“C-CPMAS NMR and XPS analysis” and “Scanning electron microscopy (SEM) study” sections). However, the contact angle on both sides tended to be similar as the graft yield was increased, indicating that despite the chemical composition and surface structure, the graft was uniformly made on both sides, reassigning an average contact angle for both sides of 91.2°. Despite the contact angle, the samples had an excellent swelling, and the pH responsiveness was also evaluated using different pH buffers at room temperature (Lina et al. 2009). The MS-g-4VP membranes exhibited a pH response reaching a swelling degree of up to 111% in an acid medium (pH = 3) for the membranes with a graft of 25.7%, which decreased as the grafting degree decreased. In the case of the response to pH, the MS-g-4VP membranes exhibited a response to the change in pH, decreasing their degree of swelling compared to the MS membranes, which showed a degree of swelling without significant changes despite the content of nitrogenous groups present in the membrane, which could present an acid–base behavior. In both cases, the MS-g-4VP (13.9 and 25.7%) membranes showed a critical pH of 6.5 (Fig. 2b), considerably reducing the degree of swelling, even below the MS membrane.

The infrared study of pristine MS membrane indicated the presence of several chemical functional groups such as –OH (1461 and 3316 cm⁻¹), –NH– (3372 cm⁻¹), –CH– (2995 cm⁻¹), and –COC– (1161 cm⁻¹) (Silverstein et al. 2005). However, each side of the MS membrane showed differences, such as less hydrophilic functional groups (–OH and –NH–) on the outer part (Fig. 3a, b); this finding agreed with contact angle and later studies. On the other hand, the grafted MS membranes showed the bands corresponding to the -g-4VP, which were assigned to –C=CH (3098 cm⁻¹), –C=C– (1639 cm⁻¹), and harmonic bands from 1800 to 1982 cm⁻¹ of the -g-4VP (Fig. 3c, d), confirming the unequivocal presence of the P4VP chains grafted on the MS membranes since similar bands are observed in the P4VP homopolymer spectrum as can be seen in Fig. 3e.

The results obtained from the thermogravimetric analysis (Table 1) showed that the MS membranes had a slight weight loss below a temperature of 100 °C, probably caused by volatile residues (i.e., water) and not by the decomposition of the sample, being stable at a temperature below 250 °C, and showing several decompositions corresponding to the sample above this (Fig. 4d). P4VP grafted chains provided the MS membrane with better thermal stability, which was according to the grafting degree (Fig. 4a–c), there was an improved thermal stability as the grafting increased and maximum decomposition temperatures at 354 and 434.1 °C, being this last decomposition attributed to -g-4VP (Fig. 4e).

¹³C-NMR spectra of the MS membrane indicated not only the signals of the cellulose chains but also the presence of probable aliphatic chains linked to amino groups from the residues of biomass, as showed in the signals located at 32.2–35.0 and 59.1 ppm, respectively, being these findings in accordance with other studies such as infrared, contact angle, and XPS analysis. The signals at 65.0 and 67.1 ppm correspond to –CH₂OH, while the other signals were assigned to cellulose endocyclic carbons (from 59.1 to 107.4 ppm) (Flores-Rojas et al. 2020b), which are shown in Fig. 5. Meanwhile, the spectra of MS-g-4VP confirm the modification by means of grafting exhibiting the corresponding signal of P4VP, from these signals, the most characteristics are the correspond to pyridine moiety located at 102.2, 124.8–127.7, and 156.6–156.4 ppm, as well as the aliphatic chain from the polymerization of vinyl moieties at 42.4 ppm.

XPS analysis determined the chemical compositions of the MS and MS-g-4VP membranes, showing the corresponding peaks of carbon (C1s at 285.0 eV), oxygen (O1s at 531.0 eV), and nitrogen (N1s at 399.4 eV) in MS and MS-g-4VP membranes and all the peaks were detected in the survey scan spectrum shown in Fig. 6. The atomic ratio of membranes was calculated from the areas of C1s, O1s, and N1s peaks in the XPS survey scan, multiplied by the appropriate sensitivity factors. The measured atomic composition of C, O, and N for each side are listed in Table 2.The results indicated an increase in the ratio of N1s on both sides of the membrane, which is consistent with the grafting of 4VP onto the MS membrane, and a decrease of the C1s and O1s on the surface of membranes, caused by the coating of the grafted P4VP chains. In addition, the results showed an increase of approximately 1.5 times the amount of nitrogen in the internal than in the outer part, suggesting that possibly there was a greater degree of grafting in the internal part of the membrane compared to the outer part, a behavior perhaps caused by due to the possible biomass residues present, which could form a greater number of graft sites.

Mechanical properties exhibited by the pristine membrane were excellent, and this was in accordance with their easy handling, but these suffered some considerable changes after the modification with the grafting polymerization using gamma-rays (Fig. 7 and Table 3). The most important modifications in the mechanical properties after the grafting was the increase of Young's modulus until 76.58 ± 3.82 MPa. Even though, Young’s modulus of both grafted membranes was not considerably different. Similarly, tensile, tenacity, and elongation exhibited minor differences with pristine MS, where the tensile break increased, and elongation showed a decrease as graft degree increased, while tenacity had not an appreciable tendency. Finally, the maximum load did not differ considerably with a load of 3 N regardless of grafting degree (0, 13.2, and 31.0%).

SEM images showed a complex morphology on the surface of the pristine MS membranes (Fig. 8a). The lignocellulosic material showed two completely different types of morphology depending on the outer and inner parts. The outer part had knitting fabric-like patterns with small channels, while in the inner part, a series of channels with a more complex structure was observed, which could be related to the plant's water absorption system. The surface morphology after graft polymerization indicated a homogeneous modification of -g-4VP without generating cracks in the membrane and other damage caused by gamma radiation or graft polymerization. High magnification SEM images showed the formation of small aggregates on the grafted membranes, undoubtedly caused by the P4VP chains (Fig. 8b, c).

Vancomycin loading was performed using MS membranes with different percentages of 4VP grafted. Drug loading in all cases was prolonged and pH dependent, as indicated by the kinetic loading, with the maximum loading being obtained around 60 h at pH = 7. Grafted P4VP chains modified these changes in the loading rate; being loading increased with grafting percentage, obtaining the highest vancomycin loading at pH = 5 for all membranes, and decreasing loading at higher pH values showing the lowest vancomycin loading at pH = 7.

In general terms, results indicated loading on all membranes, including MS membranes, even without modification. Nevertheless, -g-4VP membranes enhanced vancomycin loading at any pH value compared to pristine MS membranes, showing decreased vancomycin loading for 31.0% grafted membranes at pH values = 7 and 9, contrasting these values with those obtained at pH = 5, where the higher the degree of grafting, the greater the drug load, with values from 28.76 to 8.62 and 12.8 µg mg⁻¹ at pH = 7 and 9, respectively. These results indicated there is a greater affinity of the drug for the graft at pH = 5, a behavior corroborated by the release of vancomycin because it presented the lowest vancomycin released with 28.95% compared to a higher release of 55.74% at higher pH (Fig. 9a, Table 4). Indicating weaker interactions between the drug and -g-4VP-modified MS membranes, thus a graft above pH = 5 did not benefit from vancomycin loading via N– –H and O– –H interactions between the P4VP chains and the drug, but through ionic and dipole–dipole forces between the drug and the -g-P4VP chains (Dancy et al. 2016). Therefore, the study suggests that the optimal graft found is around 31.0% and a pH = 5.

Vancomycin release from MS-g-4VP membranes was tested at 25 °C in phosphate buffer at different pH values (Fig. 9b), showing a sustained release of vancomycin for 30 h at pH = 5 and considerably decreasing at pH = 7 and 9, the shortest release time being for pH = 9 with a release time of approximately 10 h. The prolonged release of vancomycin is consistent with the marked increase in affinity conferred by -g-4VP. This finding became even more evident when the percentages of drug released after 7 h were compared. As shown in Table 4, where the release of vancomycin was almost complete in 7 h in the case of the MS membrane at pH = 7 and 9. However, a delay was noted for the membranes at pH = 5, observing a release of less than 30%.

The vancomycin release profiles at 25 °C at the sampling points were sufficient to investigate the fit of the different mathematical models to the first 60% release interval. As shown in Table 5, the pH value modified the release mechanism showing a good release fit for the zero and first-order models as well as Korsmeyer–Peppas with a super Case II transport mechanism at pH = 5, regardless of the degree of engraftment, revealing a possible drug release mechanism mix. However, at pH = 7 and 9, the best fit was recorded for the Korsmeyer–Peppas model and indicated that vancomycin release followed the quasi-Fickian diffusion mechanism at pH = 7 and non-Fickian at pH = 9 indicated by the values of n (quasi-Fickian transport ≤ 0.45 and non-Fickian transport 0.45 ≤ n ≤ 0.89). This behavior is because, in these membranes, the release of vancomycin occurs through the channels of the membranes, as well as the relaxation of the polymer graft chains, which occurs once the material is in contact with the medium. Therefore, these properties are important for the controlled loading and release of vancomycin (Xu et al. 2020).

Finally, the antimicrobial performance of vancomycin-loaded membranes was evaluated against S. aureus (Gram-positive) and E. coli (Gram-negative). The antimicrobial activity of vancomycin against S. aureus is well known (Chakraborty et al. 2012), so the aim of study MS and MS-g-4VP membranes loaded with vancomycin is to determine possible differences in their antibiotic capability. Non-loaded MS membranes did not inhibit the growth of the bacteria (Fig. 10a). Differently, the same samples loaded with vancomycin showed inhibition zones in the S. aureus plate (Fig. 10b). Interestingly, the inhibition zone caused by vancomycin-loaded MS-g-4VP membrane was significantly greater (2.0 vs. 1.5 cm) than that observed for vancomycin-loaded MS membrane, which can be related to the fact that MS-g-4VP membrane can load more vancomycin. Overall, these results highlight that the amounts loaded are sufficient to inhibit the growth of S. aureus (Moses et al. 2020), a common pathogen present in chronic wounds.

Subsequent antimicrobial activity tests at different pH values showed a greater zone of inhibition at pH = 5 for S. aureus, but release kinetics studies showed less vancomycin release, indicating the drug has a better activity at acidic pH values. While E. coli strain basically showed resistance at all pH values (Fig. 11a, b).