Impact of probiotics on pathogen survival in an innovative human plasma biofilm model (hpBIOM)

All strains were obtained from the Leibniz-Institute DSMZ-German Collection of Microorganisms and Cell Cultures. Details are given in Table 1.

Plasma preserves and buffy coats from anonymous donors were obtained from the DRK-Blutspendedienst West (Hagen, Germany) and biofilm model were constructed as described previously [24]. In brief, residual erythrocytes in the buffy coat were removed by a centrifugation for 30 min at 3000 rpm at room temperature (RT). The plasma preserve and the buffy coat were fused and collected in a sterile glass bottle. The content of the bottle was gently mixed and continuously shaked at 22 °C.

A single hpBIOM was comprised of 1.5 ml plasma solution. 1*10⁶ cfu/1.5 ml pathogens were supplemented. 18.26 µl CaCl₂ (500 mM) per ml plasma were applied, gently mixed and quickly transferred into wells of 12-well culture plates (Sarstedt AG & Co., Nürnbrecht, Germany). The plates were incubated for 1 h on a rotation shaker at 50 rpm and 37 °C. During this time, the plasma polymerized and a stable biofilm disc/clot with integrated pathogens was generated and could be used for further analyzes.

L. plantarum, B. lactis and S. cerevisiae were grown in MRS, CSB or MEB medium for 2 days at 37 °C with shaking at 50–100 rpm. The probiotics were diluted to provide 1*10⁹ cfu in a maximum volume of 100 µl medium. This concentration was applied on top of the biofilms, followed by an additional dose of 1*10⁹ cfu after 2 h of incubation. Each pathogen was challenged to one single probiotic strain.

2 h or 24 h after the second application of probiotics, the biofilm models were dissolved by the incubation with 1.5 ml (1:1 v/v) 10% (w/v) bromelain solution (Bromelain-POS^®, RSAPHARM Arzneimittel GmbH, Saarbrücken, Germany) in 100 ml phosphate buffered saline (PBS). By using a pipette tip, the discs were detached off the well margins and subsequently punctured to make the models more permeable for the enzymatic digestion. After 2 h, the biofilm models were completely dissolved. For the quantification of the potentially survived pathogenic bacteria, 100 µl aliquots from different dilution preparations were streaked out on CSA, MEA or MRS agar plates. The bacterial burden (cfu/ml) was determined by counting colonies with a Colony Counter Pen (eCount™, VWR Leicestershire, UK) after incubation over night at 37 °C.

Scanning electron microscopy (SEM) was used to analyse the bacterial morphology. The coagula were fixed with 0.1 M cacodylate buffer containing 2.5% glutaraldehyde, 2% polyvinylpyrrolidone and 75 mM NaNO₂ for 1 h at 4 °C. Samples were washed in 0.1 M cacodylate buffer without glutaraldehyde and subsequently incubated in a solution containing 2% arginine-HCl, glycine, sucrose and sodium glutamate for 18 h at RT. The specimens were rinsed in distilled water followed by immersion in a mixture of each 2% tannic acid and guanidine-HCl for 5.5 h at RT. The samples were rinsed again in distilled water and incubated in a 1% OsO₄ solution for 30 min at RT. After three rinsing steps with distilled water the specimens were dehydrated, dried in liquid CO₂, sputtered with gold palladium and finally examined with a Zeiss Sigma SEM (Zeiss, Oberkochen, Germany) using 2 kV acceleration voltage and an inlens detector.

The hpBIOM was produced by fusion of human plasma and the corresponding buffy coat from the same donor. After the addition of the bacteria and activation of the coagulation cascade, stable coagula-like biofilm discs with integrated pathogens were generated (Fig. 1a). By means of scanning electron microscopy, bacterial colonies were detected on the fibrin scaffold (Fig. 1b). Staining of the glycokalyx revealed the development of the EPS after 1 h.

This study involved a systematic analysis of the antimicrobial activity of three probiotics L. plantarum, B. lactis or S. cerevisiae against five clinically relevant pathogens P. aeruginosa, S. aureus, S. epidermidis, E. faecium and C. albicans. Lactobacillus plantarum eliminated the Pseudomonas infection after 4 h of incubation, except for biofilms from donor 1 and 2 (Fig. 2a). Finally, after 24 h P. aeruginosa was successfully eradicated by L. plantarum in hpBIOMs from all donors. No recurrence of the pathogen was detected after 24 h in all plasma probes. The growth of S. aureus was also significantly affected in all hpBIOMs by L. plantarum, especially after 4 h (Fig. 2b). A log₁₀ reduction rate between 0.9–2.1 cfu/ml was detected. In biofilms of plasma from donor 1 and 4, the effect was negated after 24 h. The influence of L. plantarum on the growth of S. epidermidis displayed variances between the individual donors (Fig. 2c). On the one hand no alteration was observed in hpBIOMs from donor 1 and 3, but, on the other hand, a slight reduction of pathogens was quantified in biofilms from donor 2. The application of L. plantarum on biofilms of E. faecium resulted in significant inhibition of bacterial growth with a reduction of > 1.8 log₁₀ phases. In contrast to the antibacterial effect of L. plantarum, no relevant antifungal response was detected towards C. albicans (Fig. 2e). B. lactis exerted a pathogen-reducing capacity towards P. aeruginosa as well as E. faecium, while the influence on E. faecium growth was strongly donor-specific (Fig. 3a, d). The growth rates of S. aureus, S. epidermidis and C. albicans showed no differences between B. lactis treated and non-treated conditions after 4 h of incubation (Fig. 3b, c, e). The application of the yeast S. cerevisiae resulted into moderate but significant reduction of the pathogens S. aureus and S. epidermidis (Fig. 4b, c). The antimicrobial efficiency towards Pseudomonas varied in the biofilms. Inhibitory as well as slightly growth-promoting effects were detected (Fig. 4a).

SEM analysis should give more insight into the organization of L. plantarum while eliminating Pseudomonas (Fig. 5). During the experiments, L. plantarum was applied on top of the biofilm. The eradication process was documented after 1 h and 4 h of incubation. The SEM micrographs illustrated, that L. plantarum moved into the hpBIOM and arrived at the Pseudomonas colony after 1 h (Fig. 5a arrow, straight lines). The number of Lactobacilli increased with time. Scattered probiotic–pathogen interactions were visible (Fig. 5a). L. plantarum produced a complex glycokalyx, more rapidly compared to Pseudomonas (Fig. 5a, b). This matrix seemed to coat the pathogen, finally, leading to the death of the bacteria (Fig. 5c).

Health-promoting effects of “good “ lactic acid-producing bacteria were already described centuries ago, especially those belonging to the species Bifidobacterium and Lactobacillus, by inhibiting the growth of pathogenic bacteria within the colon. Different probiotics are already in use to treat dysbiosis and infections of the gastrointestinal- and urinary tract and dental diseases, e.g. pouchitis [7, 16, 30, 31].

Many studies propose better outcomes after bacteriotherapy by using L. plantarum, e.g. in animal models of P. aeruginosa infected burn wounds or chronic wounds in diabetic mice. Even a topically applied prophylactic administration of L. plantarum induced a health benefit [17, 30, 32]. Some in-vitro studies using surface-attached biofilms, challenged the pathogens to different types of living lactic acid-producing bacteria as well as supernatants or isolated proteins, and confirmed the antimicrobial activity and healing-promoting effects [33‐39]. The success was dependent on the applied pathogens and probiotics and their concentrations. However, there is a great need for research addressing the potential of bacteriotherapy and the understanding of the mechanisms in more detail. This study transferred the investigation to the newly established human plasma biofilm model. The selection of pathogenic bacteria was based on the WHO list of priority pathogens for R&D of new antibiotics published in February 2017 [15]. Additionally, a fungal contamination with C. albicans was examined.

Plasma preserves from different donors were used for the investigation. The results were not pooled, due to the different immune competences of the donors and the potential influence on the antimicrobial efficiency. In the hpBIOM, it was possible to demonstrate and to confirm the enormous antimicrobial efficiency of L. plantarum towards Pseudomonas infections (Fig. 2a). By means of SEM, it was possible to visualize the migration into the biofilm and direct pathogen-probiotic interaction (Fig. 5a, b). Furthermore, L. plantarum extensively produced a glycokalyx, which seemed to cover and destroy Pseudomonas (Fig. 5c). Supplementation of L. plantarum to S. aureus, S. epidermidis and E. faecium also induced slight but significant growth reductions (Fig. 2b–d), which was not shown before. The exact mechanisms resulting in the reduction or elimination of these bacteria is currently under investigation in this system. Different possibilities are postulated in other publications. For instance, different lactobacilli species have anti-elastase activity against P. aeruginosa [33]. Additionally, the effects of L. plantarum were assigned to the secretion of antimicrobial substances, like 4,5-dihydroxy-2,3-pentanedione and 2-methyl-2,3,3,4-tetrahydroxytertahydrofurane, which inhibits quorum sensing [38]. Other antimicrobial substances like hydrogen peroxide, benzoic acid or lactic acid are also secreted by L. plantarum [36]. The effect was donor- and time-specific, and thereby considered to be dependent on the immune system of the donor. This thesis was already proved in the gut, where different Bifidobacteria as well as Lactobacilli exerted stimulatory effect on the immune system [16]. This has to be evaluated in progressive studies. Additionally, the constitution of the bacterial cell membrane seems to be a limiting factor, because the highest growth-reducing effects were detected against gram-negative bacteria. The growth rate of C. albicans was not affected (Fig. 2e). This species is also surrounded by a strong cell wall. Interestingly, B. lactis also exerted a reducing activity towards Pseudomonas and E. faecium (Fig. 3a, d) and even the yeast S. cerevisiae showed slight but significant inhibitory effects on S. aureus, S. epidermidis and E. faecium (Fig. 4b–d). These capacities were not yet determined in human biofilms. Although the reduction of the bacterial burden seemed not to be tremendous in some combinations, it can have major relevance for the wound therapy, because it enhances the chance of reducing bacterial load by the individual immune system. Further tests with a higher number of probiotics or their combinations will be performed, to examine, whether this will improve antimicrobial outcome.

Summarized, this study successfully reproduced a novel human biofilm model. This system still represents an in-vitro model and bares limitations like a time-limited stability or the lack of skin cells. Nevertheless, several improvements were developed compared to current biofilm models. It involves essential factors for analysing biofilms in a translational research approach, namely the individual immune competence and a human wound environment. By means of the hpBIOM, it was possible to systematically screen growth-reducing activity of three probiotics towards five clinically relevant pathogens. It was possible to visualize the elimination process of L. plantarum against P. aeruginosa. Finally, additional insights into the influence of the probiotic microorganisms B. lactis and S. cerevisiae could be efficiently obtained. These effects are described for this study design and could be differ after using other concentrations of probiotics or pathogens, respectively. In future studies, the investigation of bacteriotherapy by means of the hpBIOM should be expanded with regard to subcellular and molecular insights. Additionally, the portfolio of probiotics should be increased and in particular, combined therapies of L. plantarum and other effective probiotics should be investigated using the hpBIOM.