Fabrication and characterization of LA/TCS@PLGA-NPs
The L. acidophilus envelope fragments and TCS@PLGA-NPs were fused to prepare LA/TCS@PLGA nanoparticles (LA/TCS@PLGA-NPs) by an extrusion method [35]. The measurement of LA/TCS@PLGA-NPs with dynamic light scattering showed a hydrodynamic diameter of 132.8 ± 9.1 nm. The diameter increased by approximately 30 nm compared to that of the bare TCS@PLGA-NPs (107.60 ± 2.04 nm), consistent with the addition of a cell envelope onto the nanoparticle surface [36, 37]. Despite the fact that the diameter of envelope fragments was 269.35 ± 28.61 nm, the diameter of LA/TCS@PLGA was close to that of TCS@PLGA-NPs, suggesting that the envelope was wrapped around the TCS@PLGA-NPs (Fig. 1A). In addition, the surface zeta potential of the nanoparticles increased from − 21.3 ± 0.46 mV for TCS@PLGA-NPs to − 12.40 ± 0.75 mV for LA/TCS@PLGA-NPs, which was similar to that of envelope fragments (− 10.68 ± 1.65 mV) (Fig. 1B). The increase in the zeta potential indicated the charge screening effect conferred by the coated envelope [38]. We further analyzed the morphologies of LA/TCS@PLGA-NPs and TCS@PLGA-NPs using TEM. Both LA/TCS@PLGA-NPs (Fig. 1C) and TCS@PLGA-NPs (Fig. 1D) showed a uniform spherical morphology. LA/TCS@PLGA-NPs displayed a clear envelope surrounding the sphere of TCS@PLGA-NPs. Overall, these results demonstrated that the L. acidophilus envelope was successfully coated onto TCS@PLGA-NPs.
Moreover, LA/TCS@PLGA-NPs showed a relatively constant hydrodynamic diameter after 48 h of storage at 4 °C, indicating their favorable stability properties (Fig. 1E, F). This result is consistent with a previous study on bacterial membrane-coated nanoparticles [39]. However, compared with PLGA-NPs, the stability of LA/TCS@PLGA-NPs decreased after coating. Although the mechanism for this phenomenon is unknown, we speculated that it is related to the surface properties of the LA envelope, which was inherited from L. acidophilus. It has been reported in the literature that L. acidophilus exhibits autoaggregation, which is the bacterium–bacterium adhesion of identical strains [40, 41]. Autoaggregation is generally mediated by self-recognizing surface structures of bacteria, such as proteins and exopolysaccharides [42]. Therefore, we proposed that by inheriting the surface properties of L. acidophilus, LA/TCS@PLGA-NPs may have an autoaggregating ability, which make the stability of the nanoparticles less than that of PLGA-NPs in this experiment.
Characterization of proteins
Accumulating evidence in the literature suggests that cell surface proteins are critical for the adhesion ability of L. acidophilus[22, 23]. Thus, we examined whether LA/TCS@PLGA-NPs maintained the cell surface proteins of L. acidophilus. The surface protein content on LA/TCS@PLGA-NPs was quantified with a protein bicinchoninic acid (BCA) assay. While no protein content was detected from bare TCS@PLGA-NPs, LA/TCS@PLGA-NPs showed a clear presence of protein, as indicated by absorption at 562 nm. Further quantification indicated a protein loading yield, defined as the weight ratio between the immobilized proteins and LA/TCS@PLGA-NPs, of approximately 0.33 ± 0.04 wt% (Fig. 2A).
Furthermore, the surface protein of LA/TCS@PLGA-NPs was characterized by sodium dodecyl sulfonate-polyacrylamide gel electrophoresis (SDS–PAGE). The results suggested that L. acidophilus envelope fragments and LA/TCS@PLGA-NPs were highly consistent in protein bands, revealing that almost all envelope proteins were retained throughout the LA/TCS@PLGA fabrication (Fig. 2B).
Identification of the envelope orientation of LA/TCS@PLGA-NPs
Surface layer (S-layer) proteins are an array of single proteins that are noncovalently bound to the outermost cell envelope of L. acidophilus and can be extracted by lithium chloride (LiCl) [43]. Owing to the asymmetric distribution of the S-layer protein on the extracellular side of the envelope, the protein can be used as an indicator to quantitatively analyze the envelope sidedness on LA/TCS@PLGA-NPs. Moreover, since the L. acidophilus envelope is impermeable to LiCl, LiCl extraction was applied to examine the S-layer protein content on the outer surface of LA/TCS@PLGA [37]. As shown in Fig. 2B, C, S-layer protein (≈ 46 kDa) on the surface of LA/TCS@PLGA-NPs was clearly visible (Fig. 2B), and the average content of S-layer protein on LA/TCS@PLGA-NPs was approximately 93.82% of the equivalent amount in free envelope fragments (Fig. 2C). This quantification suggests that the S-layer protein is strongly retained on the outside surface of the LA/TCS@PLGA-NPs, confirming the “right-side-out” orientation of the L. acidophilus envelope on the nanoparticles.
Drug loading and in vitro drug release of LA/TCS@PLGA-NPs
TCS has been widely used in oral hygiene products. Nevertheless, the extremely low water solubility (10 µg/mL) has significantly restricted its application in the treatment of bacterial infections [44]. Here, we improved its solubility (approximately 5 times) by successfully loading 49.7 µg of TCS into 1 mL of PLGA aqueous solution (1 mg/mL). As shown in Additional file 1: Fig. S1A, the loading efficiency (LE) and encapsulation efficiency (EE) of TCS@PLGA-NPs were 19.1% and 49.7%, respectively. In addition, envelope cloaking TCS@PLGA-NPs did not lead to a sharp drop in the LE and EE of LA/TCS@PLGA-NPs, which were 14.8% and 47.2%, respectively.
The release kinetics of TCS from TCS@PLGA-NPs and LA/TCS@PLGA-NPs were investigated in PBS (pH 7.4) solution to simulate the physiological environment. In the first 6 h, 30.25% and 25.36% of TCS was released quickly from TCS@PLGA-NPs and LA/TCS@PLGA-NPs, respectively. The early sudden release phenomenon may be due to the free TCS that was not wrapped in nanoparticles or attached to the surface of nanoparticles. Then, the speed of release gradually slowed down. After 77 h of incubation in PBS, 80.82% of TCS was released from TCS@PLGA-NPs, and 64.97% was released from LA/TCS@PLGA-NPs (Additional file 1: Fig. S1B). Except at 0 h and 4 h, the cumulative TCS release of LA/TCS@PLGA-NPs was significantly lower than that of TCS@PLGA-NPs at other time points (P < 0.05). This result is consistent with previous studies reporting that the PLGA-NPs coated by cell membrane showed a slower drug release profile than uncoated PLGA-NPs [35, 45]. This result indicated that LA/TCS@PLGA-NPs had better sustained release ability than that of TCS@PLGA-NPs, which may be ascribed to the additional cell envelope bilayer acting as a diffusion barrier.
Cytotoxicity assay
The cytotoxicity of TCS, PLGA-NPs, TCS@PLGA-NPs and LA/TCS@PLGA-NPs at a series of concentrations was evaluated by CCK-8 assay using HOK cells. As Additional file 1: Fig. S2 shows that TCS exhibited low cytotoxicity at concentrations lower than or equal to 7 µg/mL but high cytotoxicity at concentrations higher than 15 µg/mL. This result is consistent with a previous study on the cytotoxicity of TCS [46]. In addition, PLGA-NPs, TCS@ PLGA-NPs and LA/TCS@ PLGA-NPs had high safety performances. The highest concentrations of PLGA, TCS@PLGA and LA/TCS@PLGA in the experiment were all 120 µg/mL. The cell viability of HOK cells treated with PLGA, TCS@PLGA and LA/TCS@PLGA at the highest concentration for 72 h was more than 75%, which were all in the safe range.
Adhesion of LA/PLGA-NPs to planktonic S. mutans
The coaggregation between planktonic L. acidophilus and S. mutans is shown in Additional file 1: Fig. S3. Based on this information, we hypothesized that the inheritance of native properties of the L. acidophilus cell surface confers adhesion ability to LA/PLGA-NPs. To test this hypothesis, we labeled LA/PLGA-NPs with fluorescence dye and mixed them with planktonic S. mutans. Following incubation and washing, CLSM showed that the LA/PLGA-NPs were retained (green, representing the PLGA-NPs; red, representing the LA envelope) by S. mutans (blue), which indicated nanoparticle-cell adhesion (Fig. 3A). In the control group, the fluorescence of PLGA-NPs (green) nearly disappeared after washing (Fig. 3B). The results demonstrated that the L. acidophilus envelope could indeed endow PLGA-NPs with the ability adhere to S. mutans. Next, we utilized fluorescence spectrophotometry to quantitatively analyze the adhesion ability of LA/PLGA-NPs to S. mutans according to the aggregation assay with slight modifications. As shown in Additional file 1: Fig. S4, the adhesion rate of LA/PLGA-NPs was 66.24 ± 4.27%, which was much higher than that of PLGA-NPs (9.83 ± 2.23%) (P < 0.05).
To test the potential adhesin that mediates the adhesion of LA/TCS@PLGA-NPs to S. mutans, we employed sodium periodate, bovine serum albumin, LiCl and pepsin to block the adhesion mediated by polysaccharides, teichoic acids and S-layer proteins, respectively [37, 47, 48]. The fluorescence of LA/PLGA-NPs after LiCl and pepsin treatment almost disappeared around S. mutans, but the nanoparticles were still retained by S. mutans after treatment by sodium periodate and bovine serum albumin (Fig. 3A). The results indicated that the adhesion ability of LA/TCS@PLGA-NPs was polysaccharide- and teichoic acid-independent, and the S-layer protein was probably the major adhesin mediating the adhesion of LA/TCS@PLGA-NPs to S. mutans. This result is consistent with previous studies reporting that the S-layer protein plays a key role in the adhesion of lactobacilli [22, 49].
Integration of LA/PLGA-NPs into the S. mutans biofilm structure
According to a report by Nobbs [50], the formation period of S. mutans biofilms includes the following key stages: 0 h, initial bacterial adherence; 6 h, initial bacterial colonization; 12 h: initial early biofilm formation; 24 h: maturation of early-stage biofilm and 48 h: maturation of the later-stage biofilm. In this study, we added fluorescence-labeled LA/PLGA-NPs at these five stages to interfere with the formation of S. mutans biofilms and observed their spatial distribution in S. mutans biofilms by CLSM.
Based on 3D CLSM images, we calculated the biofilm thickness using a Leica Application Suite X (Leica Microsystems, Wetzlar, Germany). As shown in Fig. 4A–E, for the biofilms mediated at 0 h, 6 h, and 12 h, the biofilm thicknesses were all approximately 17 µm, while for the biofilms mediated at 24 h and 48 h, the biofilm thicknesses were approximately 20 µm and 30 µm, respectively. Then, the biofilms were evenly divided into three layers based on these calculated results, representing the inner, medium and outer layers.
We discovered that regardless of which biofilm stage the LA/PLGA-NPs and PLGA-NPs were added at, they mostly existed in the medium layer of biofilms (5–15 µm) without being washed out easily. Such a spatial distribution may be related to the unique structure of biofilms, in which the inner and outer layers were loose and dispersed but the medium layer was dense and cross-linked. Additionally, as shown in Fig. 4F and G, the quantity of LA/PLGA-NPs increased with the progress of biofilm formation, while the change trend of PLGA-NPs remained relatively consistent during biofilm formation. Thus, we speculated that the quantity of LA/PLGA-NPs existing in biofilms was closely related to that of S. mutans. One possible reason is that LA/PLGA-NPs possessed an ability to adhere with S. mutans. Therefore, we proposed that the integration of LA/PLGA-NPs into S. mutans biofilms relies not only on their particle size but also on their adhesion to S. mutans.
Interfering Effect of LA/PLGA-NPs on S. mutans biofilm formation
To explore whether by inheriting the surface properties of L. acidophilus, LA@PLGA-NPs without TCS could interfere with the biofilm formation of S. mutans, we further quantitatively analyzed biofilm inhibition by the crystal violet method. Because it was previously reported in literatures that inactivated probiotic strains, such as inactivated L. acidophilus, have no antibacterial activity, but can bind with planktonic pathogens through surface properties, and thus block planktonic pathogens from colonization and biofilm formation.
The results showed that at different time points, the biofilm inhibition rates were 64.1%, 40.3%, 21.9%, 6.7% and 2.6%, respectively (Fig. 4H), suggesting that LA/PLGA-NPs mainly inhibited the early stage of biofilm formation. In addition, PLGA-NPs had no significant inhibitory effect on the biofilm of S. mutans. Similar results were observed in previous studies on the inhibitory effect of live or heat-killed Lactobacillus strains against S. mutans biofilms [18, 51,52,53]. Although the specific mechanism is not clear, we proposed that this effect could be attributed to the surface properties of LA/PLGA-NPs, which were able to coaggregate with planktonic S. mutans in the medium solution and thus inhibit biofilm formation at the early stage. However, once the biofilm matures, it is difficult for LA/PLGA-NPs to exclude and displace S. mutans in the biofilm.
Effect of LA/TCS@PLGA-NPs on S. mutans biofilm activity
In this experiment, the effect of LA/TCS@PLGA-NPs on the biofilm activity was evaluated by a live and dead bacterial staining kit and was observed using CLSM. As shown in Fig. 5, live bacteria appeared as fluorescent green, while dead bacteria appeared as fluorescent red. The biofilms treated with TCS and TCS@PLGA-NPs showed few live bacteria. In contrast, the biofilms treated with LA/TCS@PLGA-NPs had many live bacteria. The intensities of red fluorescence (dead bacteria) and green fluorescence (live bacteria) were quantitatively measured by ImageJ software. Then, the biofilm activity was calculated as the percentage of live bacteria over both live and dead bacteria. The results revealed that compared with the control groups, in which the biofilm activity remained constant at approximately 89%, the activity of biofilms treated with TCS, TCS@PLGA-NPs and LA/TCS@PLGA-NPs was significantly decreased (P < 0.05). However, as shown in Fig. 6, the activity of biofilms treated with LA/TCS@PLGA-NPs was significantly higher than that treated with TCS and TCS@PLGA-NPs when nanoparticles were added at any stage of biofilm development (P < 0.05). These results indicated that during the short period of time (24 h), LA/TCS@PLGA exhibited less of an inhibitory effect on the biofilm activity of S. mutans biofilms than that of TCS and TCS@PLGA-NPs. We proposed that this phenomenon was probably due to the dual barrier effect of LA/TCS@PLGA-NPs. In the biofilms, the cumulative release of TCS from LA/TCS@PLGA-NPs in the short term was less than that of TCS@PLGA-NPs.
Moreover, when TCS and TCS@PLGA-NPs were added at the maturation stage, the biofilm activity was higher than that when they were added at the early stage (P < 0.05). This was probably because at the maturation stage, S. mutans was embedded in an extracellular matrix of polymeric substances, which had a negative impact on the antibacterial effects of both TCS and TCS@PLGA. However, for LA/TCS@PLGA-NPs, the biofilm activity remained constant at approximately 75% regardless of when the nanoparticles were added (P > 0.05). This result indicated that the TCS delivered by LA/TCS@PLGA-NPs was less affected by the biofilm matrix at the maturation stage. We proposed that on the one hand, the presence of a cell envelope and polymeric nanoparticles could protect TCS from being deactivated by matrix components or enzymatic modifications. On the other hand, due to the ability to adhere to with S. mutans, LA/TCS@PLGA-NPs could reach and release TCS more closely to S. mutans in the biofilm, while TCS released from TCS@PLGA-NPs may be blocked and inactivated by the matrix on the way to S. mutans. Thus, even when added at the mature stage, LA/TCS@PLGA-NPs could maintain a consistent inhibitory effect.
Effect of LA/TCS@PLGA-NPs on the biomass of S. mutans biofilms
The biomass growth of the S. mutans biofilm was measured from the dry weight and the total soluble protein. As shown in Fig. 7, 8 days after the treatment, the dry weight and total soluble protein of biofilms in the LA/TCS@PLGA-NPs group were 0.93 ± 0.06 mg and 0.46 ± 0.03 mg, respectively, which were the lowest among the different groups, although there was no statistically significant difference in the dry weight of biofilms between the LA/TCS@PLGA-NPs group and TCS@PLGA-NPs group. This result indicated that LA/TCS@PLGA-NPs could effectively inhibit the growth of S. mutans biofilms and had a longer-lasting antibiofilm effect than that of TCS and TCS@PLGA-NPs. We speculated that this may be related to the integration of LA/TCS@PLGA-NPs in S. mutans biofilms and the cumulative release of TCS from LA/TCS@PLGA-NPs.
Effect of LA/TCS@PLGA-NPs on virulence gene expression in S. mutans biofilms
Bacterial cells in biofilms tend to exhibit biological and phenotypic traits that are extraordinarily distinct from those of their planktonic counterparts, and these traits are accompanied by significant changes in the bacterial gene expression profile [54]. Studies have shown that the cariogenic characteristics of S. mutans biofilms are regulated by a variety of genes. In this study, real-time fluorescence quantitative PCR was used to detect and analyze the gene expression regulating the caries-associated virulence factors of S. mutans biofilms.
As revealed in Fig. 8, we found that after a period of 8 days, the expression of caries-associated virulence factors, including gtfB, gtfC, fif, spaP, gbpB, ldh, atpF, comD and vicR, was significantly downregulated in the biofilms treated with LA/TCS@PLGA-NPs compared with that in the control group (P < 0.05). In addition, the downregulation of gtfB, gtfC, and comD in response to LA/TCS@PLGA-NPs was significantly greater than that in response to TCS and TCS@PLGA-NPs (P < 0.05). Among these genes, gtfB, gtfC, and ftf are related to polysaccharide synthesis and sucrose-dependent adhesion; gbpB is related to sucrose-dependent adhesion and biofilm formation; spaP is related to sucrose-independent adhesion; ldh and atpF are related to acid production and acid resistance; and comD and vicR are related to the stress regulation system [55]. All of these genes are important factors leading to the formation of cariogenic S. mutans biofilms. Therefore, these results indicated that the antibiofilm effect of LA/TCS@PLGA-NPs relied not only on the inheritance of native properties of the L. acidophilus cell surface but also on the sustained release of antimicrobial agents that inhibit the expression of virulence genes in S. mutans biofilms. Based on these results, we proposed that, in terms of virulence gene expression, LA/TCS@PLGA-NPs also exhibited a long-lasting antibiofilm effect, which was better than that of TCS and TCS@PLGA-NPs.
LA/TCS@PLGA-NPs for dental caries in vivo
To further verify the effectiveness of LA/TCS@PLGA-NPs in vivo, we established a rat model of caries that was induced by the combination of S. mutans infection and a cariogenic diet. LA/TCS@PLGA-NPs and other treatments were performed twice daily for 7 consecutive days. After 45 days, the rats were sacrificed. Carious lesions on the smooth surface (Smo) and the sulcal surface (Sul) were quantitatively assessed via Keyes’ scoring [56]. According to Keyes’ scoring, the depth of carious lesions was divided into the following levels: enamel only (E), slightly dentinal (Ds, < 1/4 of the dentin region), moderate dentinal (Dm, 1/4–3/4 of the dentin region), and extensive dentinal (Dx, > 3/4 of the dentin region).
For the smooth surface, the carious lesions were evaluated based on the scoring of E. As shown in Fig. 9A, compared with the control group, the carious lesions were significantly decreased in the LA/TCS@PLGA-NPs group (P < 0.05), while there were no significant differences between the control group and other treatment groups (P > 0.05). For the sulcal surface, the carious lesions were evaluated based on the following levels [56]: total lesions (E + Ds + Dm + Dx), initial lesions (Ds + Dm + Dx), moderate lesions (Dm + Dx), and extensive lesions (Dx). As shown in Fig. 9B, there were no statistically significant differences among the different groups in terms of the total lesions and initial lesions (P > 0.05). However, compared with those in the control group, the moderate lesions and extensive lesions in the LA/TCS@PLGA-NP group were significantly decreased (P < 0.05), while no significant differences were observed between the control group and other treatment groups. These results indicated that compared with TCS and TCS@PLGA-NPs, LA/TCS@PLGA-NPs exhibited a more lasting inhibitory effect on the progression of caries in vivo. We proposed that this effect could be attributed to the ability of LA/TCS@PLGA-NPs to adhere to S. mutans biofilms, which facilitated the retention and sustained drug release of LA/TCS@PLGA-NPs in the oral cavity.
The in vivo biosafety of LA/TCS@PLGA-NPs was assessed via body weight monitoring, blood biochemical assays, routine blood examination and histological analysis of the main organs. During the treatment period, there was no significant difference in the body weight of the rats among the different groups (Fig. 10). The blood biochemical assays of total cholesterol (TC), triglyceride (TG), high-density lipoprotein (HDL), low density lipoprotein (LDL), immunoglobulin M (IgM), immunoglobulin G (IgG), and glucose (Glu) were at normal levels (Additional file 1: Table S1). Moreover, routine blood examination showed that the levels of leukocytes, erythrocytes, hemoglobin and platelets did not vary significantly between the different groups. In addition, based on the results of hematoxylin–eosin (H&E) staining, no distinguishable change were found in the main organs (heart, liver, spleen, lung, and kidney). Accordingly, these results demonstrated that LA/TCS@PLGA-NPs did not induce significant adverse effects in vivo, indicating that these nanoparticles have potential as a safe treatment for dental caries.