Induction of antimicrobial peptides secretion by IL-1β enhances human amniotic membrane for regenerative medicine

Due to antibacterial characteristic, amnion has been frequently used in different clinical situations. Developing an in vitro method to augment endogenous antibacterial ingredient of amniotic epithelial and mesenchymal stem cells is desirable for a higher efficacy of this promising biomaterial. In this study, epithelial or mesenchymal side dependent effect of amniotic membrane (AM) on antibacterial activity against some laboratory and clinical isolated strains was investigated by modified disk diffusion method and colony count assay. The effect of exposure to IL-1β in production and release of antibacterial ingredients was investigated by ELISA assay. The results showed that there is no significant difference between epithelial and mesenchymal sides of amnion in inhibition of bacterial growth. Although the results of disk diffusion showed that the AM inhibitory effect depends on bacterial genus and strain, colony count assay showed that the extract of AM inhibits all investigated bacterial strains. The exposure of AM to IL-1β leads to a higher level of antibacterial peptides secretion including elafin, HBD-2, HBD-3 and cathelicidic LL-37. Based on these results, amniotic cells possess antibacterial activity which can be augmented by inflammatory signal inducers; a process which make amnion and its epithelial and mesenchymal stem cells more suitable for tissue engineering and regenerative medicine.

Microbial assay. Antibacterial property of amnion was measured by modified disk diffusion method. This method was adapted from the Clinical and Laboratory Standard Institute (CLSI) protocol. Pseudomonas aeruginosa ATCC 27853, Staphylococcus aureus ATCC 25923, Escherichia coli ATCC 25922 and two clinically isolated sensitive strains of E. coli (T3 and T4) were used for antibacterial susceptibility test.
All bacterial strains were cultured on the blood agar medium (Merck, Germany). After being incubated overnight at 35 ± 2 °C, some colonies were harvested from plates and separately suspended in the sterile normal saline. The turbidity was equivalent to a 0.5 McFarland standard. Finally, these bacteria were cultured on the Muller-Hinton agar plates (Merck, Germany).
After being isolated and washed, the fresh amnion was cut into small pieces (1 × 1 cm). Each of the pieces were placed on the seeded Muller-Hinton agar plate in 20-30 minutes after separation, according to the modified disk diffusion method 13,29 . Amnion pieces were put separately in two ways, epithelial side up and mesenchymal side up as an antibacterial peptide releasing disk. Then the plates were incubated at 35 ± 2 °C overnight and the inhibitory effects were measured.
In order to avoid probable contamination within processing and preparation of the AM, a small piece (1 × 1 cm) of fresh AM was put on the Muller-Hinton agar plates and microbial growth was controlled 24 h after cultivation of tissues.
Amnion extract antibacterial effect. Antibacterial effect of amnion extract was examined with a modified method in which a piece of amnion (1 × 1 cm) was cut into smaller sections and added to equivalent volume of PBS. The extract was obtained by sonicating the amnion on ice for 10 minutes with 80 W and 0.5 s cycle (Hielscher, Ultrasound Technology, Germany). Then membrane residues were removed by centrifuging at 800 rpm for 4 min. 500 μl of amnion extract was added to the surface of each Muller-Hinton agar plates. After drying the surface of plates, the bacterial inoculum was added to them. The number of colonies was counted after overnight incubation at 35 ± 2 °C.
Evaluating the secretory state of natural antimicrobial peptides. To evaluate secretory state of antimicrobial agents from amniotic cells, the amniotic membrane was cut into 12 small pieces that were each 2 × 2 cm, and the pieces were individually cultured in a 12-well plate. To each well, 1 mL of DMEM/F12 containing FBS 10% and penicillin-streptomycin 1% was added. The plates were incubated in 37 °C with 5% CO2.
After 24 hours, medium was changed to serum-depleted (1% FBS) for 20 hours before treatments were added; 1% FBS was used because our previous study demonstrated decreased amniotic cell viability after 24 hours of culture in the absence of serum 30 . Half of the plates were treated with 10 ng/mL recombinant human interleukin (IL)−1beta (Peprotech, London, UK) to induce production and secretion of natural antimicrobial. The plates were incubated in 37 °C with 5% CO2 for 24 h, after which the supernatant was collected from each well.
Statistical analysis. Statistical comparisons were performed as mean ± standard error of the mean (SEM).
Statistical significance was determined by means of one-way analysis of variance (ANOVA) followed by Tukey's post-test. A P-value less than 0.05 was considered statistically significant. Data availability. All data generated or analyzed during this study are included in this published article.

Results
Possible bacterial contaminations of membranes were controlled by culturing the fresh tissues on Muller-Hinton agar plates and microbiologically proper samples without bacterial contaminations were used for data interpretation. Totally, 30 amnion were examined in this study and at least 6 amnion were used for each bacterial strain in epithelial or mesenchymal sides separately. In the statistical analysis, only the groups with 50% inhibition zone in cultured plates were used.
Based on the disk diffusion method, the inhibitory effect of the AM was appeared in P. aeruginosa ATCC 27853, E. coli T3 and T4 after 24 hours of incubation ( Table 1).
The growth was decreased under and in the edge of the tissues in both epithelial side and mesenchymal side up amnion. The mean of inhibition zone was about 2 mm in the epithelial side up AM and 2.8 mm in mesenchymal side up AM for P. aeruginosa ATCC 27853 cultures. The maximum inhibition (5 mm) was seen in this group ( Fig. 1a) ( Table 1).
The inhibitory effect among the E. coli strains demonstrated different results. The inhibition zone was appeared in two sensitive clinical isolated E. coli (T3 and T4) in both epithelial and mesenchymal side up. Figure 1b and Fig. 1c demonstrate the inhibition zone in E. coli T3 and E. coli T4, respectively.
Statistical analysis showed that there is no significant difference in the number of plates in which the inhibitory effect was seen against P. aeruginosa ATCC 27853, E. coli T3 and T4 (Fig. 2). However, the inhibition zone in P. aeruginosa ATCC 27853 is significantly (P < 0.01) greater than two clinical isolated E. coli T3 and T4 (Fig. 2).
To investigate the effects of amnion side on antibacterial activity, we compared the results between mesenchymal and epithelial sides of amnion. As shown in Fig. 2, there is no significant difference between the number of plates and inhibition zone of P. aeruginosa ATCC 27853 which incubated with mesenchymal side amnion in comparison to those of incubated with epithelial side amnion. The same results were achieved from incubation of E. coli T3 and T4 with both mesenchymal and epithelial sides of amnion.
No inhibitory effect was seen in E. coli ATCC 25922 and S. aureus ATCC 25923 after 24 hours of incubation under and in the edge of both epithelial side up and mesenchymal side up amnion (Fig. 1d).
In order to consider the effects of the contents of amniotic cells, the AMs were sonicated and the inhibitory effects of amnion extract were evaluated on P. aeroginosa ATCC 27853, S. aureus ATCC 25923, E. coli ATCC 25922, and two clinical sensitive strains of E. coli (T3, T4). Consistent with the results of disk diffusion method, the number of colonies decreased in P. aeroginosa ATCC 27853 and two clinical sensitive strains of E. coli (T3, T4) significantly in comparison to control groups (Fig. 3). Surprisingly, the extract of amnion significantly reduced the number of colonies of S. aureus ATCC 25923 and E. coli ATCC25922. In addition to the number of colonies, the extract of amnion decreased the size of S. aureus ATCC 25923 and E. coli ATCC25922 colonies compared to the control groups (Fig. 4).
To evaluate secretory state of amniotic cells after induction with IL-1β, elafin, HBD-2, HBD-3 and cathelicidin LL-37 protein levels were measured in amniotic membrane culture supernatant by ELISA. The rational for choosing IL-1 beta as pro-inflammatory cytokine was a pilot study in which we compared 10 ng/ml of TNF-alpha with the same concentration of IL-1 beta in production of HBD-3 at time points 6, 12, 24 and 48 hours. Although the primary results showed that both pro-inflammatory cytokines significantly up-regulated HBD-3 production compared to unstimulated control, HBD-3 production was significantly higher at each time point in response to IL-1 beta (data not shown). As evaluated by trypan blue exclusion, treatments with IL-1β had no significant effect on amniotic cell numbers or viability after 24 hours.

Discussion
In this study, the antibacterial property of amnion was examined on different bacterial strains and the side dependent effect of amniotic membrane was also investigated on antibacterial activity. Furthermore, the amnion extract was used to optimize the inhibitory effect.
Disk diffusion is a commonly used method to evaluate antibacterial property of the AM. Talmi et al. used the modified disk diffusion method for investigating the amniotic membrane antimicrobial activity on coagulase positive staphylococcus, E. coli, Kelebsiella pneumonias, P. aeruginosa, Proteus mirabilis. They demonstrated that the inhibition zone appears exactly under the tissue due to the close contact between amnion and culture plate 13      study. They reported that a narrow inhibition zone (1 mm) appeared in Group A streptococcus, Staphylococcus saprophyticus, E. coli, P. aeruginosa and S. aureus 31 . We also previously demonstrated the antibacterial properties of cryopreserved AM by this method 32 .  In the present study, we used both standard and clinical bacterial strains to investigate antibacterial property by disk diffusion method. In accordance to the results of disk diffusion method, the inhibitory effect of fresh amnion appeared in P. aeruginosa and two clinical isolated strains of E. coli (T3 and T4) opposed to S. aureus ATCC 25923 and E. coli ATCC 25922. The results of disk diffusion method showed that the amnion inhibitory effect was dependent on bacterial genus and bacterial strain. According to our results, it seems that the inhibitory effect in disk diffusion method is due to either the contact between the amnion and bacterial strains 13 or achieving a secretion threshold of antibacterial ingredients which is enough to inhibit the bacterial activity of some strains. For instance, elafin can inhibit serine peptidase that is a virulence factor in P. aeruginosa in a critical threshold concentration which is not enough to inhibit growth of the other bacterial strains 33,34 . On the other hand, using the amnion extract reduced the size and number of colonies in all strains. In the case of two strains which were not inhibited in disk diffusion method but inhibited using the amnion extract (S. aureus ATCC 25923 and E. coli ATCC 25922), we have two options. Either amniotic membrane cannot release its ingredients in disk diffusion condition or the amniotic antibacterial ingredients are non-secretory. As shown in results, exposing the AM to IL-1β resulted in a higher amount of elafin, HBD-2, and HBD-3 and cathelicidin LL-37 secretion which confirmed secretory state of amniotic antibacterial ingredients. This result is consistent with previous study in which we showed that AM releases some anti-cancer and immune-modulatory agents in condition medium 35,36 . Therefore, the non-secretory state of amniotic antibacterial ingredients is not acceptable and it seems that the disk diffusion method does not provide an appropriate condition for amnion to display its antibacterial property. To the best of our knowledge, disk diffusion method is the most common method for assessing the AM antibacterial property; and since it is not a proper method for this aim, it is necessary to modify and develop alternative methods to attain a reliable method for investigating the antibacterial property of AM.
In this study the side dependent effect of amnion on bacterial activity was examined. There are no significant difference between epithelial side up and mesenchymal side up amnion in inhibitory effect. This similarity in two sides of amnion suggests that not only amniotic epithelial cells cause inhibitory effect, but also amnion mesenchymal cells can display antibacterial property. Although all previous studies focused on antibacterial properties of amniotic epithelial cells, amniotic mesenchymal cells also could have antibacterial characteristics 32,37 . Mesenchymal stem cells have antibacterial characteristics. It has been shown that lipocalin 2 38 , beta-defensin 2 39 and LL-37 (human cathelicidin antimicrobial peptide, hCAP-18) 40 in mesenchymal stem cells from the other sources such as bone marrow and adipose demonstrate antibacterial property against bacterial strains; hence, the amniotic mesenchymal cells also may contribute to inhibitory effect of amniotic membrane. More studies would be required to determine the antibacterial effect of amniotic mesenchymal cells. Figure 5. Antibacterial peptides concentration in the AM culture supernatant before and after treatment with IL-1β for 24 hours (P < 0.05 ★, P < 0.001 ★★★). The experiment was performed in quadruplicate (n = 5 for each time).