A Novel Simple In Vitro System Mimicking Natural Environment for the Biofilm Cultivation of Cutaneous Bacteria

Abstract

Biofilms are microbial communities of cells embedded in extracellular matrix, and they are regarded as a major form of the natural and laboratory occurrence of bacteria. Cutaneous microbiota is represented by prokaryotic and eukaryotic organisms, which form biofilms in the different niches including the skin surface, glands, and hair follicles. Despite of a large number of in vitro studies dedicated to the biofilms of cutaneous bacteria, the methods used usually do not closely take into account the specific surroundings of certain skin parts. In this study, we introduce a new simple method of biofilm cultivation on the solid keratine/agarose pellets embedded in polyacrylamide gel. In such a model system, we tried to minimize the amount of liquid phase, which makes a model close to a human comedo, and provide a prominent biofilm formation of selected cutaneous bacteria.

1. Introduction

The microbiota is one of the important components of the human organism, which largely determines the correct functioning of the skin, intestine, and other organs. Additionally, normal human microbiota helps the host organism fight pathogenic microorganisms. When the balance of the commensal microbiota is disturbed, pathogens begin to proliferate intensively, which can lead to the development of infections. Microorganisms mostly colonize the surfaces of the human body by forming multispecies biofilms on them—communities of microorganisms embedded in the extracellular matrix. Biofilms are very resistant to adverse environmental factors, including antibiotics and other biocides, and pathogenic biofilms are the cause of most chronic infections [1].

In most cases, classical methods are used to study the biofilms of microorganisms; for example, the method of immunological plates [2,3,4,5]. The advantages of this method are low cost, simplicity, and the ability to conduct the large-scale screening of active compounds. To eliminate the shortcomings of this method, e.g., a wide variability of results, as well as a limited amount of substrate, our group developed the Teflon cube method, which uses a larger carrier surface area, as well as growing biofilms on the surface of glass microfiber (GMF) filters, which is based on the lack of planktonic culture [6]. In recent years, new methods for the reconstruction of biofilms based on the imitation of natural conditions have appeared. This group of novel methods is developed for better correlation with in vivo experiments. The first group of methods includes various methods of simulating the natural surface, in particular, methods of pathogen biofilms growing on the surface of collagen to simulate wound infections, both in static systems [7] and in flow systems [8]. In the work of Bessler et al. [9], disks of coagulating plasma containing collagen were used for biofilm cultivation. In addition to using natural polymers as a carrier, it is also possible to modify the surface itself with various micropatterns to better mimic the natural microrelief [10] and obtain heterogeneous microbial populations, including using 3D bioprinting [11]. Similar trends apply to the development of flow systems. For example, a specific flow chamber has been developed for culturing biofilms on the surface of dental implants [12], microfluidic chips to mimic dental canal infections [13], or polymicrobial wound infections [14].

Currently used methods for cutaneous biofilms study are summarized in Table 1. When cultivating the biofilms of skin microorganisms, the natural temperature of the skin surface (about 33.5 °C) and special rich media, such as reinforced clostridial [3], are used. However, attempts to more fully recreate the natural environment of skin bacteria have not yet been noted in the scientific literature, with rare exceptions. For instance, a commercial EpiDermFT™ model, which includes both epidermal and dermal cell layers, was used in one of the recent studies [15]. However, cutaneous bacteria colonize different econiches in skin, and apocrine sebaceous glands are one of the most inhabited by microorganisms. To date, a few methods for obtaining artificial sebum and the cultivating of Cutibacterium acnes biofilms in such a medium are described [16,17]. Due to impossibility of conducting of large-scale in vivo experiments on animal skin, the development of a new simple model system is an extremely urgent task.

In the present study, we aimed to develop a simple, cheap, and reproducible method for the cultivation of cutaneous bacteria, based on the structure of human comedo, which matches the real skin microenvironment as closely as it is possible. This method, on the one side, does not require any complex and expensive manipulations, such as artificial keratinocyte-covered surfaces, animal models, or artificial skin; on the other side, it allows to predict bacterial behavior in the microenvironment of skin. Comedo is a form of acne vulgaris—a chronic inflammatory disease of skin pilosebacious units [18], and represents itself a “result from abnormalities in the proliferation and differentiation of ductal keratinocytes” [19]. We tried to minimize the amount of unbounded liquid in comparison with our current methods used. To simulate the surrounding tissues with nutrients, we proposed to use a high-dense polyacrylamide gel (PAAG), and a solid pellet containing hydrolyzed keratin and few components of a human sebum as a model of comedo-like structure. Three selected strains of Staphylococcus aureus, S. epidermidis, and Micrococcus luteus were selected as well-studied bacteria, forming well-developed biofilms on microtiter plates, Teflon cubes, and microtiter plates.

Table 1. Current in vitro models used for cutaneous biofilms assay.

Method	Advantages	Disadvantages	Ref.
Cultivation in 96-well plates	Suitable for a large screening of active compounds and their effects on biofilms. Simple, quick and the most widespread biofilm assay.	Lack of nutrients in wells. Cells aggregates in liquid medium are not taken into account. Sometimes separation of the biofilm and the planktonic culture is almost impossible. A potential large variation of data per wells especially during crystal violet staining. Lots of plastic waste needed to be processed.	[2,3,4,5]
Cultivation on Teflon cubes submerged in liquid medium	Less data variation due to a total Teflon cubes surface squire is higher. Teflon is a reusable material hence there is less damage to environment.	A surface cannot be suitable for some bacterial strains. Cells aggregates in liquid medium are not taken into account. Takes more time in comparison with plates. The use of highly toxic compounds (K2Cr2O7 in concentrated H2SO4) is needed to clean the cubes after use.	[6,20,21,22]
GMF filters	Conditions are more close to those on human skin or to cutaneous wounds because of the lack of liquid phase. A surface can be used as abrasive material which facilitates a CFU count assay.	Filters cannot be easily cut into small (less than 5 mm) standard pieces. A large biomass amount on a large filter makes not possible sometimes to detect minor effects. Filters cannot be stained by universal dyes as crystal violet due to high level of glass fiber staining. The only staining of metabolically active cells is applicable. A higher price of commercially available glass filters.	[6,20,21,22]
Artificial 3D-skin (EpiDermFT™)	A model consists from artificial tissues and very suitable for the study of the interaction between microbes and human cells.	Expensive method which does not allow to use it in preliminary studies A model does not contain any glands and therefore nutrients are limited.	[15]
Artificial sebum model	Conditions are close to those of human comedo.	A solid pellet contains only sebum components without keratin. Medium was mixed with sebum which is not fully conform to conditions in comedo.	[16]
Current model	Conditions are close to those in human comedo. Cheap. Underlies potentially more complicated models.	Preparation of pellets and carriers needs a significant time.	Current study.

Table 1. Current in vitro models used for cutaneous biofilms assay.

Method	Advantages	Disadvantages	Ref.
Cultivation in 96-well plates	Suitable for a large screening of active compounds and their effects on biofilms. Simple, quick and the most widespread biofilm assay.	Lack of nutrients in wells. Cells aggregates in liquid medium are not taken into account. Sometimes separation of the biofilm and the planktonic culture is almost impossible. A potential large variation of data per wells especially during crystal violet staining. Lots of plastic waste needed to be processed.	[2,3,4,5]
Cultivation on Teflon cubes submerged in liquid medium	Less data variation due to a total Teflon cubes surface squire is higher. Teflon is a reusable material hence there is less damage to environment.	A surface cannot be suitable for some bacterial strains. Cells aggregates in liquid medium are not taken into account. Takes more time in comparison with plates. The use of highly toxic compounds (K2Cr2O7 in concentrated H2SO4) is needed to clean the cubes after use.	[6,20,21,22]
GMF filters	Conditions are more close to those on human skin or to cutaneous wounds because of the lack of liquid phase. A surface can be used as abrasive material which facilitates a CFU count assay.	Filters cannot be easily cut into small (less than 5 mm) standard pieces. A large biomass amount on a large filter makes not possible sometimes to detect minor effects. Filters cannot be stained by universal dyes as crystal violet due to high level of glass fiber staining. The only staining of metabolically active cells is applicable. A higher price of commercially available glass filters.	[6,20,21,22]
Artificial 3D-skin (EpiDermFT™)	A model consists from artificial tissues and very suitable for the study of the interaction between microbes and human cells.	Expensive method which does not allow to use it in preliminary studies A model does not contain any glands and therefore nutrients are limited.	[15]
Artificial sebum model	Conditions are close to those of human comedo.	A solid pellet contains only sebum components without keratin. Medium was mixed with sebum which is not fully conform to conditions in comedo.	[16]
Current model	Conditions are close to those in human comedo. Cheap. Underlies potentially more complicated models.	Preparation of pellets and carriers needs a significant time.	Current study.

2. Materials and Methods

2.1. Bacterial Strains and Culture Conditions

S. aureus 209 P (ATCC 6538P, type strain), S. epidermidis ATCC 14990 (type strain, isolated from nasal mucosa) and M. luteus C01 (isolated from a human skin of a healthy volunteer, [20]) were stored at room temperature in semisolid lysogeny broth (LB, Dia-M, Moscow, Russia) with 0.5% agar covered with sterile mineral oil. All of these selected strains are able to form well-established biofilms which was shown in our previous studies [20,21,22]. Cultures were prepared in LB and grown overnight with shaking at 150 rpm. The incubation temperature was 33.5 °C, which is close to the temperature on the skin surface. Commercial porcine brain heart infusion (BHI, BD, Franklin Lakes, NJ, USA) was used as a basic medium for all experiments. Sterilization was performed at 121 °C for 15 min according to manufactured protocol.

Cultivation in 96-well plates was set up under aerobic conditions. Cultures were adjusted to OD₅₄₀ = 0.5 with sterile physiological saline (PS, 1% NaCl in distilled water). Two hundred microliters of BHI per well was added, and 3.3 μL of a cell suspension was inoculated. The plates were incubated in a Xmark microplate spectrophotometer (Bio-Rad, Hercules, CA, USA) at 33.5 °C with shaking.

2.2. Preparation of the Comedo-like Pellets

To prepare the solid comedo-like scaffolds a blend of keratin/agarose was used as a basic material. A method described in previous work of Nayak, Gupta [23] was used as a basis for following procedure but it was substantially modified and simplified by us. In particular, we used commercial hydrolyzed keratin and essentially shortened the dehydration time.

2.2.1. Preparation of Keratin/Agarose Blend

Hydrolyzed keratin (Shaanxi Aoji Biological Technology Co., Xi’an, China) was dissolved by gentle mixing in neutralizing buffer (Tris HCL—12.1 g, NaCl—8.7 g, HCl (concentrated)—2.5 mL, 0.5 M EDTA-4Na—200 μL per 1 L of MQ water) in concentration 500 mg/mL and pH of obtained solution was adjusted to 6.5–7.0 by 10 M NaOH. Agarose was suspended in Milli-Q^® water (5% (w/v)), mixed together with equal volume of keratin solution (1:1), then, mixture was gently heated in microwave oven with stirring until agarose was completely melted and dissolved and finally solution was placed in boiling water bath.

2.2.2. Preparation of Artificial Sebum

In this study, the preparation of artificial sebum was substantially simplified. Briefly, 1 g of glycerol tripalmitate (99%, Aladdin Biochemical Technology, Shanghai, China) was mixed with 1 mL of jojoba oil (Oleos, Russia), then heated to 85 °C, mixed and autoclaved at 121 °C for 30 min.

2.2.3. Preparation of the Pellets

Artificial sebum was melted on boiling water bath and then added to keratine/agarose mixture in final concentration of 2%. A mixture was stirred at boiling water bath until the homogeneous suspension was obtained, then mixture was poured into wells of a special plastic plate (Figure S2). After a mixture was solidified at room temperature (RT), the obtained cylindrical pellets were gently removed from wells and kept at −18 °C for 30 min. After freezing pellets were immersed in chilled 99% ethanol for 20 min at RT, then ethanol was removed and pellets were dried overnight at 37 °C and stored at RT.

2.3. Preparation of the PAAG Carriers

A concentrated solution was prepared to obtain relatively dense gel with short polymerization time (2 mL of 40% acrylamide/bis-acrylamide solution (37.5:1, Dia-M, Moscow, Russia), 10 μL of TEMED (neoFroxx GmbH, Einhausen, Germany), 100 μL of 10% ammonium per-sulfate (Helicon, Moscow, Russia) per 5 mL of MQ water) and was poured into plastic lids, then special comb-like plastic templates were placed immediately to form the wells (3 mm in diameter and 3 mm in height) in gel. After complete solidification templates were removed and cylindrical carriers (8 mm in diameter and 7 mm in height) were made using the metallic bore. Carriers were rinsed once with distilled water for 1 min to remove all potentially toxic unpolymerized reagents from gel, then rinsed quickly with 96% ethanol, placed into plastic plates, and sterilized under UV-light for 1.5 h. Then, carriers were immersed with sterile BHI medium, and control samples were incubated with sterile PS, for 20 min at RT. After the solutions were discarded, carriers were placed into sterile plastic Petri dishes.

2.4. Biofilm Cultivation on Pellets

Comedo-like pellets were sterilized under UV-light for 2 h, placed into wells of prepared PAAG cylindrical carriers, then samples were additionally sterilized under UV-light for 15 min. Bacterial cells were centrifuged and rinsed twice with sterile PS, then suspensions were adjusted to OD₅₄₀ = 0.5. Then, 5 μL of each suspension was dropped into pellet, samples were left to dry for 5 min, and then incubated at 33.5 °C for 24 h. Incubation was conducted in small Petri dishes covered with parafilm to avoid gel drying.

To estimate the biofilm quantity pellets were stained with 3-(4,5-dimethyl-2-thiazolyl)-2.5-diphenyl-2H-tetrazolium bromide (MTT, Dia-M, Moscow, Russia) for 30 min as it was described in our studies previously [3,21]. Basically, the only metabolically active cells are stained with MTT, which transforms to insoluble formazan as an electron acceptor. However, biofilms were formed only 24 h in our experiments, thus, we proposed that quantity of formazan is directly related to quantity of biofilm. After staining each pellet was removed from carrier using syringe needle, carefully washed in water for 10 s to stop the reaction and placed into Eppendorf tube. Extraction was performed by adding 1 mL of DMSO and incubation at RT for 24 h, then, tubes were vigorously vortexed, subsequently centrifuged for 2 min at 10,000× g, and OD₅₉₀ of supernatant was measured.

2.5. Scanning Electron Microscopy (SEM) of Biofilms

Samples were fixed in 2.5% glutaraldehyde for 2 h at RT immediately after incubation, then washed gently 3 times in PS with subsequent incubation in ethanol gradient according to standard protocol, then left to dry under the hood for 2 h. Dry samples were mounted on specimen stub covered previously with a glue tape and were exposed in a high vacuum and at an accelerating voltage of 20 kV; a gold was used as a coating material. Studying of a pellet surface was performed with SEM microscope JEOL JSM IT 200, Tokyo, Japan.

2.6. Plastic Stuff Fabrication

All additional necessary plastic details, e.g., a special plate for making cylindrical pellets (see Section 2.2.3), comb-like templates (see Section 2.3) were manufactured using 3D-printer (Hercules G2, Imprinta, Kransoyarsk, Russia) from polyethylene terephthalate glycol filament (PETG, Filamentarno, Moscow, Russia). Previous full-precision modelling (Figure S1) was made in FreeCAD 2019 software. The following printing parameters were selected: layer height 0.2 mm, filament diameter 1.75 mm, extrusion temperature 220 °C, disabled cooling fan, nozzle diameter 0.5 mm, retraction 0.5 mm.

2.7. Statistics and Data Processing

All experiments were conducted in triplicates. The nonparametric Mann-Whitney U test was performed for statistical data evaluation. All statistical and microbiological data plots were generated in the GraphPad Prism 2007 software. Kinetic growth parameters were calculated in Microsoft Excel 2007 Software. Where appropriate, mean relative values (control without BHI in PAAG carrier was designated as 100%) and absolute values were plotted on the graphs, and the standard error of the mean was depicted as error bars. Data were considered to be statistically significant at a confidence level of 95% (p-value < 0.05).

3. Results

3.1. Sample Preparation and MTT Biofilm Assay

Following the procedure described above, we obtained a solid material which contains hydrolyzed keratin and sebum components (Figure 1) and is surrounded with polyacrylamide hydrogel. During the incubation time, a pellet gradually becomes saturated with a gel-bound solution. However, in this model system, there is a far less amount of unbounded liquid in comparison with glass microfiber filters, which makes this condition closer to the natural environment of cutaneous bacteria. Hydrogel simulates dermal tissues and pellet material containing keratin and sebum plays a role as basic substrate for selected cutaneous strains. Commercial BHI was selected as a basic medium in this model on the one hand, due to the possible similarity in chemical composition to human blood, and on the other hand, it may be used as a basic solution for more complex media and is well standardized comparing with blood plasma. 3D-printing has been shown as a cheap and powerful tool for making the necessary plastic equipment for carriers and pellet manufacturing (Figure S1).

First, we tested the kinetic parameters of the bacterial growth of selected strains in planktonic culture (

Table 2. Kinetic parameters of bacterial growth in BHI.

Kinetic Parameters	S. aureus 209 P	S. epidermidis ATCC 14990	M. luteus C01
Maximal OD 540	1.71	1.88	1.83
Specific growth rate, h−1	0.34	0.37	0.25
Doubling time, h	2.0	1.89	2.79
Linear portion of the curve, h	2.25	3.25	2.25

, Figure S2) in BHI medium. Both staphylococci have a similar specific growth rate, and S. epidermidis ATCC 14990 also shows a longer linear interval of the growth curve and a highest doubling time value. However, M. luteus has a lower growth rate, together with shortest linear part and largest doubling time, which also appears in the specific character of the growth curve (Figure S2C). Meanwhile, the maximum OD₅₄₀ values are similar for all three selected strains.

Second, we cultivated biofilms on comedo-like pellets placed in PAAG carriers containing BHI. Carriers containing PS served as control. In these experiments, we estimated the biofilm formation of selected cutaneous bacteria and whether pellet material or nutrient gel plays a more significant role in this process. The results are indicated in Figure 2. Staphylococci shows a significantly higher level of biofilm formation on pellets in the presence of BHI. S. epidermidis reveals the most significant rate of biofilm growth (330.4 ± 61.5% vs. 100 ± 6.8% in control). These results may indicate the better adhesion of cells to the material as well as higher growth rate on the BHI medium, which has been already shown. At the same time, significant biofilm growth for M. luteus C01 was shown, however, there was no significant difference between control and BHI-treated samples observed (128.4 ± 12.4% vs. 100 ± 15.5% in control). This observation may imply better preference for pellet nutrient material by this strain in general but also, conversely, a lower adhesion of the cells to the pellet. Thus, a SEM study was conducted to study the biofilm morphology.

3.2. SEM Study of Biofilms

Biofilms were grown for 24 h on pellets, which were placed into PAAG carriers, previously incubated in BHI, then samples were prepared as described in Section 2.6. First, we have revealed that all three selected strains form biofilms on the surface of the material. Despite the porous, sponge-like structure of the pellet (Figure 3A), the material does not allow cells to form biofilms homogeneously. Cell clusters are actively formed on all roughnesses of the pellet but can colonize the internal part of the pellet only via the relatively large cracks of the surface. Second, the cells were revealed to be the major component of the biofilms, despite the extracellular matrix (EM). A thick layer of EM was not observed, which is typical for these strains. However, we can proclaim that it was fully formed, according to the way the cells are attached to each other and form layer-like structures, which were definitely observed on higher magnification (Figure 3C,E,G). S. epidermidis is more prone to form biofilm as a monolayer, but cells appear more deeply embedded in the substrate (Figure 3C). Finally, M. luteus C01 forms a well-developed biofilm (Figure 3G) despite the relatively low biofilm amount shown via staining with MTT. We can propose that this fact can be explained by the lower attachment of the cells to the surface with subsequent potential biomass loss when the pellet was washed in water to remove MTT. Contrary, S. epidermidis cells look as deeply attached to the surface, and therefore total biofilm biomass is revealed to be higher when it was stained with MTT. Additionally, this strain has a shortest doubling time of growth, as was shown by kinetic study (Table 2).

4. Discussion

The biofilms of cutaneous bacteria, their interaction in multi-species communities, and the impact of human hormones on these processes have been deeply studying in the last decade. This raises the problem of developing a method, which, on the one hand, is quite simple and allows to obtain reproducible results and, on the other hand, has a potential to further development and can be adopted to the different strains and culture conditions. Becoming closer to the natural environment may have a significant impact on microbial behavior in biofilms. Spittaels and Coenye [16] previously developed the artificial sebum model for C. acnes cultivation and have shown the key importance of the model conditions for the virulence factor production of the selected strain. In the present work, we aimed to reconstruct the conditions of human comedo.

First, we aimed to minimize the amount of unbounded liquid in our model. We propose that the main disadvantage of present in vitro models in an abundance of liquid, which is not typical for the comedo-like environment. This disadvantage also concerns the model based on GMF filters, which rather resembles a wound than a skin microniche. Thus, we relied on a recently developed artificial sebum model, first developed by Spittaels and Coenye [16].

However, we suggested that mixing the artificial sebum together with medium does not fully meet the conditions of the microniche. The cutaneous bacteria tested should obtain nutrients both from the surrounding tissues and from the material itself. Additionally, a new model should include the keratine together with the lipid components. Hence, we decided to obtain the porous material, containing both keratin and artificial sebum, and surrounded by hydrogel, saturated with medium. We used a dence PAAG instead of agarose because of its transparency and elasticity, which makes samples preparation convenient. Despite the potential toxicity of PAAG, we confirmed the ability of selected strains to form microcolonies on its surface previously (data not shown).

A third step was the simplification of a new model. Commercially available cheap hydrolyzed keratin was used, to avoid the commonly used preparation from natural sources, such as feathers. We deliberately decided to use the only two components of artificial sebum (tripalmitin and jojoba oil) in a substantially lower concentration (2%) for better mixing. A significantly simplified method of Nayak and Gupta, 2015 [23], was taken for the preparation of the pellets. All necessary plastic was made by 3D-printing from a standard, commonly used PETG. Computer aided design (CAD) was undertaken using a freely-available software, the FreeCAD, 2019, and therefore we decided to avoid the direct addition of medium into the solid pellet. BHI was selected as the basic medium as a potential model of human blood, instead of reinforced clostridium medium, which was commonly used in our previous and current studies.

The results obtained demonstrates that all selected strains are able to form biofilms on pellets, even without a medium in the gel carrier and, therefore, can use the keratin and fatty acids as a carbon source. We have also shown the higher biofilm growth of staphylococci, especially S. epidermidis ATCC 14990, in comparison with M. luteus C01. These data are in accordance with the previous data considering S. epidermidis as a commensal strain in skin microbial communities [24], and it was confirmed by SEM study that S. epidermidis cells are deeply attached to the soft material. Additionally, BHI does not affect significantly the micrococcal biofilms. This fact can be explained by lower attachment, probably by the chemical composition of the selected medium, which is not optimal for biofilm formation. The first explanation is more probable because we observed the developed biofilms of micrococci via SEM. Additionally, the precise econiche of this strain on skin remains unknown, so probably a comedo-like environment is not quite suitable for it.

Nevertheless, based on the obtained data, we can claim that the method has proven to be applicable for biofilm cultivation in further studies. We made an attempt to overcome some disadvantages of current models, e.g., the unbounded liquid phase and the lack of necessary compounds such as keratin, and to make our system cheap and simple to use. This method has potential to be improved in further studies, with different pellet composition, PAAG density, and culture conditions.

5. Conclusions

Thus, we developed a cheap and simple method, which can simulate the natural surroundings of cutaneous bacteria more closely and successfully, and we tested it on three selected well-studied strains. 3D-printing is now a widespread method to develop cheap and reliable cultivation systems. The accessibility of gel and comedo-like components allows to simulate well the skin gland microniches. The combining of these features makes our method appropriate for any type of study of skin bacteria behavior, and solves different tasks such as the screening of active compounds effects, the deep investigation of monospecies, or multispecies skin microbial biofilm communities.