The Anticancer Agent 3,3'-Diindolylmethane Inhibits Multispecies Biofilm Formation by Acne-Causing Bacteria and Candida albicans

ABSTRACT The Gram-positive anaerobic bacterium Cutibacterium acnes is a major inhabitant of human skin and has been implicated in acne vulgaris formation and in the formation of multispecies biofilms with other skin-inhabiting organisms like Staphylococcus aureus and Candida albicans. Indoles are widespread in nature (even in human skin) and function as important signaling molecules in diverse prokaryotes and eukaryotes. In the present study, we investigated the antibacterial and antibiofilm activities of 20 indoles against C. acnes. Of the indoles tested, indole-3-carbinol at 0.1 mM significantly inhibited biofilm formation by C. acnes without affecting planktonic cell growth, and the anticancer drug 3,3′-diindolylmethane (DIM) at 0.1 mM (32 μg/mL) also significantly inhibited planktonic cell growth and biofilm formation by C. acnes, whereas the other indoles and indole itself were less effective. Also, DIM at 0.1 mM successfully inhibited multispecies biofilm formation by C. acnes, S. aureus, and C. albicans. Transcriptional analyses showed that DIM inhibited the expressions of several biofilm-related genes in C. acnes, and at 0.05 mM, DIM inhibited hyphal formation and cell aggregation by C. albicans. These results suggest that DIM and other indoles inhibit biofilm formation by C. acnes and have potential use for treating C. acnes associated diseases. IMPORTANCE Since indoles are widespread in nature (even in human skin), we hypothesized that indole and its derivatives might control biofilm formation of acne-causing bacteria (Cutibacterium acnes and Staphylococcus aureus) and fungal Candida albicans. The present study reports for the first time the antibiofilm and antimicrobial activities of several indoles on C. acnes. Of the indoles tested, two anticancer agents, indole-3-carbinol and 3,3′-diindolylmethane found in cruciferous vegetables, significantly inhibited biofilm formation by C. acnes. Furthermore, the most active 3,3′-diindolylmethane successfully inhibited multispecies biofilm formation by C. acnes, S. aureus, and C. albicans. Transcriptional analyses showed that 3,3′-diindolylmethane inhibited the expressions of several biofilm-related genes including lipase, hyaluronate lyase, and virulence-related genes in C. acnes, and 3,3′-diindolylmethane inhibited hyphal formation and cell aggregation by C. albicans. Our findings show that 3,3′-diindolylmethane offers a potential means of controlling acne vulgaris and multispecies biofilm-associated infections due to its antibiofilm and antibiotic properties.

C. acnes readily forms biofilm on plastic, glass, steel, titanium, silicone, and other surfaces, and its biofilms are often resistant to conventional antibiotics and host immune systems (2). Also, C. acnes is often found in multispecies biofilms with skin-colonizing organisms such as Staphylococcus aureus and Candida albicans (3)(4)(5). C. acnes biofilms are composed of DNA, proteins, and glycosyl residues (6). Although the effects of medium, culture time, and surface type on the biofilm characteristics of several C. acnes isolates have been studied in vivo and in vitro (7), the mechanism responsible for biofilm formation remains unclear despite the attempts made to control the formation of opportunistic C. acnes biofilms (8)(9)(10).
Skin-colonizing C. acnes is almost certainly accustomed to indole containing environments, as indole constitutes nearly 30% of the volatile headspace of sweat (22). Hence, we investigated the antibacterial and antibiofilm potentials of 20 natural or synthetic indole derivatives against C. acnes. To investigate how indoles influence biofilm development, we used confocal laser scanning microscopy (CLSM), scanning electron microscopy (SEM), and quantitative realtime PCR (qRT-PCR), and extracellular polymeric substance (EPS) production, hyphal, and combinatorial antibiotic assays. In addition, we investigated the antibiofilm activities of 3,39-diindolylmethane (the more active indole) against multispecies biofilms of C. acnes, S. aureus, and C. albicans. The present study reports for the first time the antibiofilm and antimicrobial activities of several indoles on multispecies microbes including acne-causing bacteria and C. albicans.
The antibiofilm activity of DIM was compared with those of three antibiotics, namely, benzoyl peroxide, gentamicin, and ciprofloxacin, which also dose-dependently inhibited C. acnes biofilm formation (Fig. 2). While benzoyl peroxide, which is commonly used to treat acne, was less effective at inhibiting C. acnes biofilm formation than DIM, gentamicin and broad-spectrum antibiotic ciprofloxacin were more effective at inhibiting C. acnes cell growth and biofilm formation. For example, benzoyl peroxide at 200 mg/mL inhibited biofilm formation by 60% ( Fig. 2a) with MIC 400 mg/mL, which concurs with a recent report (23). Gentamycin at 20 mg/mL inhibited biofilm formation by 98% ( Fig. 2b) with MIC 100 mg/mL, and ciprofloxacin at 1 mg/mL inhibited biofilm formation by 94% ( Fig. 2c) with MIC 5 mg/mL.
Combinatory antibiofilm efficacies of DIM and antibiotics against C. acnes. Antibiofilm activities were enhanced when DIM and antibiotics were used in combination. Concentrations of antibiotics resulting in 40 ; 60% biofilm reduction were selected for this study. Gentamycin at 5 mg/mL reduced biofilm formation by 53%, whereas treatment with gentamicin (5 mg/mL) and DIM (0.05 mM) reduced biofilm formation by 93% (Fig. 2e). Also, ciprofloxacin at 0.5 mg/ mL reduced biofilm formation by 37%, whereas treatment with ciprofloxacin (0.5 mg/mL) and DIM (0.05 mM) reduced biofilm formation by 84% (Fig. 2f). However, DIM did not enhance the efficacy of benzoyl peroxide (Fig. 2d). Fractional inhibitory concentration index (FICI) was calculated to check synergistic effect of DIM and antibiotics. FICI values are 1.0 for gentamicin and 1.0 for ciprofloxacin, respectively. Hence, it appeared that the combinatory effects were additive and not synergistic.
Antibiofilm activities of DIM against C. albicans and S. aureus biofilms. C. acnes often forms biofilms on skin with other microorganisms like S. aureus or C. albicans (3,4), and these biofilms tend to be more tolerant of antimicrobials (24). Hence, we investigated the inhibitory effects of DIM on S. aureus or C. albicans biofilms under aerobic and anaerobic conditions. Biofilm formations by C. albicans and S. aureus were dose-dependently inhibited by DIM after culture for 1 day under aerobic conditions ( Fig. 3a and b). For example, DIM at 0.1 mM inhibited C. albicans biofilm formation by 94% (Fig. 3a) and inhibited S. aureus biofilm formation by 71% under aerobic conditions (Fig. 3b). Furthermore, DIM significantly inhibited biofilm formations by C. albicans and S. aureus after culture for 6 days under anaerobic conditions ( Fig. 3c and d). More specifically, DIM at 0.1 mM inhibited C. albicans biofilm formation by 83% (Fig. 3c) and inhibited S. aureus biofilm formation by 65% (Fig. 3d). The MICs of DIM against C. albicans and S. aureus were both . 1 mM, respectively, which suggest that the antimicrobial activity of DIM was partially responsible for its antibiofilm activity. Microscopic observations of C. acnes biofilm inhibition by DIM. Biofilm inhibition was analyzed using a live cell imaging system and by CLSM and SEM. 3-D color mesh plots revealed that DIM at 0.02 or 0.05 mM inhibited biofilm formation by C. acnes (Fig. 4a). CLSM and COMSTAT analysis better revealed biofilm changes ( Fig. 4b and c). Nontreated C. acnes formed dense biofilms (thickness . 40 mm and almost 100% surface coverage), and the presence of DIM at 0.02 or 0.05 mM significantly reduced biofilm densities and thicknesses (Fig. 4c). Specifically, biofilm biomass, thickness, and substrate coverage were reduced by . 88, 87, and 95% versus untreated controls, respectively, in the presence of 0.05 mM DIM (Fig. 4c).
SEM analysis showed that DIM decreased the production of EPS and reduced C. acnes cell densities in biofilms. Intriguingly, DIM treatment increased cell length, indicating inhibition of cell division (Fig. 4d). More specifically, C. acnes cell sizes increased by 240% (3.0 6 0.3 mm) and 252% (3.2 6 0.2 mm) in the presence of 0.02 or 0.05 mM DIM, respectively, as compared with non-treated controls (1.25 6 0.4 mm) (Fig. 4d). Interestingly, indole and indole-3-carbinol at 0.05 mM also decreased EPS production and increased C. acnes cell sizes (Fig. S1 in the supplemental material).
Antibiofilm activities of DIM against multispecies biofilms of C. acnes, C. albicans, and S. aureus. Since DIM effectively inhibited biofilm formation by C. acnes (Fig. 1), C. albicans (Fig. 3c), and S. aureus (Fig. 3d), we investigated the antibiofilm activities of DIM against polymicrobial biofilms containing two or three species. We optimized media and inoculum to form biofilms from two different combinations of two species and one combination of three species: C. acnes and S. aureus (Fig. 5a) in 1:1 combination of Reinforced Clostridium Media (RCM) and Luria-Bertani (LB); C. acnes and C. albicans (Fig. 5b) in a 1:1 mixture of RCM and Potato Dextrose Broth (PDB); and C. acnes, S. aureus, and C. albicans (Fig. 5c) in a 1:1:1 mixture of RCM, LB, and PDB. As was expected, DIM dose-dependently inhibited biofilm formation in all cases (Fig. 5a to c). Notably, DIM at 0.05 mM inhibited three-species biofilm formation by 69% after culture for 6 days under anaerobic conditions (Fig. 5c). To the best of our knowledge, this is the first report of three species biofilms of C. acnes, C. albicans, and S. aureus.
Multispecies biofilms were observed using a live cell imaging system and by CLSM and SEM. All three techniques showed DIM inhibited three species biofilms (Fig. 5d to g). They  formed biofilms of thickness ; 19 mm with almost 100% surface coverage, but DIM at 0.02 or 0.05 mM dose-dependently reduced biofilm densities and thicknesses (Fig. 5f). On the other hand, three species biofilms were substantially less dense and thinner than a single C. acnes biofilm (Fig. 4b). SEM visualized the presence of C. acnes, C. albicans, and S. aureus in mixed biofilms (Fig. 5g). In the nontreated control, C. albicans formed pseudohyphae and yeast cells that were much larger than the round cells of S. aureus cells or the rod-type cells of C. acnes. DIM treatment at 0.02 mM prevented the adhesion of most C. albicans cells, and few C. acnes cells were observed among S. aureus cells. Notably, DIM at 0.05 mM significantly reduced the production of EPS in three species biofilms, and only few numbers of S. aureus cells remained while C. albicans cells and C. acnes cells were not shown.
DIM reduced EPS production in C. acnes biofilms. EPS production is a hallmark of biofilm formation as it protects cells from environmental challenges. EPS production by C. acnes was measured in the presence of indole, DIM, and indole-3-carbinol. While indole at concentrations up to 2 mM did not affect EPS production (Fig. 6a), DIM and indole-3-carbinol dose-dependently reduced EPS production in C. acnes (Fig. 6b and c), and DIM at 0.05 mM reduced EPS production in C. acnes by 90%. These results suggest that the antibiofilm activity of DIM is partially due to the inhibition of EPS production.
DIM and indole-3-carbinol inhibited hyphal growth and cell aggregation by C. albicans. Dimorphic switching of yeast cells to hyphal cells and cell aggregation are prerequisites of C. albicans biofilm maturation (25). To study the effect of indoles on C. albicans morphology, we used a microscope and a cell aggregation assay. In untreated C. albicans, hyphae and large cell aggregations entangled by hyphae were observed after 24 h. Treatments with DIM or indole-3-carbinol at 0.1 mM all substantially suppressed hyphal formation; yeast and pseudohyphae cells were mostly observed (Fig. 7a). Indole at concentrations up to 0.2 mM did not affect this dimorphic switching, but DIM at 0.05 mM clearly inhibited hyphal formation and cell aggregation (Fig. 7b). These results indicate DIM potently inhibits C. albicans biofilm formation by inhibiting hyphal formation and cell aggregation.
DIM-induced changes in gene expressions in C. acnes. qRT-PCR was used to investigate the effect of DIM on the expressions of 11 biofilm-and virulence-related genes in C. acnes. Overall, DIM modulated the expression of several lipase genes, hyaluronate lyase genes, and virulence-related genes (Fig. 8), while the expression of the housekeeping gene (16s rRNA) was unchanged. Notably, the expressions of lipase genes (PPA1796 and PPA2105), the hyaluronate lyase gene (hly), and the precorrin-2 C (20)-methyltransferase (cbiL) gene were downregulated while the expression of hemolysin (tly) was upregulated 3-fold by DIM at 0.1 mM.

DISCUSSION
Acne vulgaris is a common chronic skin disease that results in psychological and social problems in adolescents, and it is generally accepted that biofilm formation by skin microorganisms like C. acnes, S. aureus, and C. albicans aggravate acne vulgaris (2,26). In this study, we aimed to identify a biofilm inhibitor against biofilms formed by C. acnes, S. aureus, and C. albicans singly or in combination. DIM was found to be an effective antibiofilm agent, and its action mechanism was partially revealed. Several indole derivatives have been shown to exhibit antibiofilm activity against various microbes, for example, 3-indolylacetonitrile against enterohemorrhagic E. coli O157:H7 (14), methylindoles (18) or 7-benzyloxyindole (27) against C. albicans, iodoindoles against A. baumannii (19), several halogenated indoles derivatives against S. marcescens (20), and chloroindoles against V. parahaemolyticus (21). Unlike other indoles in previous studies, in this study, DIM and indole-3-carbinol both efficiently inhibited biofilm formation by C. acnes. Notably, microbes respond differently to indoles, and it would appear that a methanol group at the C-3 position of the indole moiety might have been responsible for the antibiofilm activities of DIM and indole-3-carbinol.
Most plants have acquired defense mechanisms against environmental microbes, and thus, diverse secondary metabolites represent major pharmaceutical sources for discovery (28,29). For example, cruciferous vegetables (Brassica), such as broccoli, cauliflower, and cabbage, contain various indole derivatives such as indole-3-acetic acid, indole-3-acetonitrile, indole-3-carbinol, and 3,39-diindolylmethane (DIM) (Fig. 9). Indole-3-acetic acid and indole-3acetonitrile are growth hormones (auxins) (30), and indole-3-acetonitrile has been reported to inhibit E. coli O157:H7 biofilm formation and to act as an antivirulence compound against P. aeruginosa (14). Also, it was reported that indole-3-carbinol and DIM exhibit antimicrobial, antiviral, and anticancer activities (31,32). DIM is produced from the digestion of indole-3carbinol in Brassicas, and it is currently used to treat respiratory tumors (33). It has been reported that oral dose of DIM at 2 mg/kg/day was tolerated with no significant toxicity (34). Also, the cytotoxicity of DIM was relatively low as EC 50 of DIM was a range of 230 ; 440 mM against three glioblastoma cell lines (35). While its cytotoxicity is inevitable as an anticancer agent, the toxicity of DIM on normal skin cells or skin diseases would be further studied. In this study, DIM shows antimicrobial and antibiofilm activities against C. acnes, S. aureus, and C. albicans singly and in combination. Hence, it appears that Brassicas utilize several indole derivatives for defense purposes against diverse microbes. Further studies are required to determine the effects of these indoles on other microbes.
Natural occurring biofilms usually consist of multiple microbe species, which means the majority of studies on polymicrobial biofilms are limited, because focus has been placed on the susceptibilities of single microbe biofilms, which are more susceptible than composite biofilms (36). However, this study presents the antibiofilm activities of DIM in mono-, dual-, and three-component biofilm models against Gram-positive C. acnes, Gram-positive S. aureus, and fungal C. albicans. As was expected, the antibiofilm mechanisms of DIM appeared to depend on microbe type. C. acnes expresses multiple genes encoding lipases that degrade sebum lipids and release free fatty acids (1), and it has been reported that surface characteristics, biofilm-related genes, adhesive proteins, and lipase activity are involved in C. acnes biofilm development and that C. acnes contains multiple lipases that degrade sebum lipids (1). Our qRT-PCR study showed that DIM downregulated the expressions of lipase genes (PPA1796 and PPA2105), the  hyaluronate lyase (hly), and precorrin-2 C(20)-methyltransferase (cbiL) genes (Fig. 8). Previously, it was reported that extracellular lipases are associated with biofilm formation by C. acnes as biofilm cells produce more extracellular lipases than planktonic cells (37). Hence, it appears that suppressions of these lipase genes by DIM possibly underlie its inhibition of biofilm formation. Additionally, the inhibition of EPS production by DIM (Fig. 6a) supports its biofilm inhibitory effects, since EPS formation is a hallmark of biofilm formation. However, DIM did not affect the hydrophobic characteristics of C. acnes cells (Fig. S2 in the supplemental material). More detailed study is required to understand more clearly the mechanism of antibiofilm action.
In C. albicans, the morphogenetic switching of yeast cells to hyphal cells is considered to play an important role in biofilm formation and the pathogenesis of fungal infections (38), and it has been reported that 7-benzyloxyindole (27) and several methylindoles (18) inhibit C. albicans biofilm formation by inhibiting hyphal formation. In the present study, DIM inhibited hyphae and biofilm formation ( Fig. 3c and 7a), possibly by reducing the virulence of C. albicans. On the other hand, DIM did inhibit biofilm formation by S. aureus under aerobic and anaerobic conditions ( Fig. 3b and d). As was expected, the inhibitory effects of DIM on dual-and three-species biofilms appeared to be more complex (Fig. 5).
Indole-3-carbinol and DIM both induced the apoptosis of human breast cancer cells (39) and human prostate cancer cells (40) and exhibited antioxidative and anti-inflammatory properties (41). Currently, DIM is used to treat recurrent respiratory papillomatosis with tumors in the upper respiratory tracts caused by human papillomavirus (33), and in a preclinical study was well tolerated at a dose of 2 mg/kg/day (34). In the present study, DIM inhibited biofilm formation by C. acnes and exhibited potent antimicrobial ( Fig. 1f and Table 1) and antibiofilm formation activities by all three microbes (Fig. 5c). Furthermore, DIM and antibiotic treatment in combination enhanced the antibiofilm activities of antibiotics (Fig. 2). These results show DIM, a natural product found in cruciferous vegetables, has potential utility for the treatment of multispecies infections and as a potent antibiofilm agent and antibiotic adjuvant.

CONCLUSIONS
In conclusion, we report the antimicrobial and antibiofilm activities of a number of indoles, including DIM, against single or composite biofilms produced by C. acnes, S. aureus, and/or C. albicans. The antibiofilm activities of DIM appeared to have been due to suppressions of the expressions of lipase genes, antimicrobial activity against C. acnes, and inhibition of hyphae development by C. albicans. The study shows DIM offers a potential means of controlling acne vulgaris and multispecies biofilm-associated infections due to its antibiofilm and antibiotic properties.

MATERIALS AND METHODS
Strains and chemicals. C. acnes KCCM 41747 (ATCC 6919) (isolated from human facial acne), methicillin-sensitive S. aureus ATCC 6538, and fluconazole-resistant C. albicans strain DAY185 were used in this study. The C. acnes strain was cultured on Reinforced Clostridium Media (RCM)-agar plates for colony preparation and in liquid RCM at 37°C under anaerobic conditions (BD GasPak EZ Gas Generating Anaerobic Pouch Systems; Fisher Scientific, Pittsburgh, PA, USA) for all other experiments. The S. aureus strain was cultured in LB medium at 37°C, and the C. albicans strain was cultured in Potato Dextrose Broth (PDB) medium at 37°C.
Biofilm formation inhibition assay. Biofilm formation was quantified using crystal violet (Sigma-Aldrich), as previously described (42). Briefly, a 6-day culture of C. acnes with an initial OD of 2.3 (CFU and the housekeeping gene (16s rRNA) used for qRT-PCR are listed in Table S1. The expression of 16s rRNA was not affected by DIM. The qRT-PCR method used has been previously described (44) and was performed using SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) and an ABI StepOne Real-Time PCR System (Applied Biosystems). Analysis of gene expression by qRT-PCR was done by the Applied Biosystems StepOne Real-Time PCR System analysis software. Each gene expression level was normalized by housekeeping gene expressions. At least two independent cultures and four PCRs were used.
Statistical analysis. Replication numbers for each experiment are provided in each method, and results are presented as means 6 standard deviations. One-way ANOVA followed by Dunnett's test using SPSS Ver. 23 (Chicago, IL, USA) was performed for the statistical analysis. P values of , 0.05 were considered significant. Asterisks in the figures indicate significant changes between treated (none) and untreated samples.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.7 MB.