Candida albicans is an opportunistic pathogen and the most prevalent fungal species of the human microbiota (da Silva Dantas et al., 2016). In individuals with healthy immune systems, C. albicans asymptomatically colonizes mucosal surfaces. However, in some conditions this species can cause infections, such as VVC (Nobile and Johnson, 2015). The overgrowth and morphological transition (e.g., yeast-to-hyphae transition) of C. albicans are very important determinants to promote the conversion from a commensal to a pathogen (Jacobsen et al., 2012). VVC is responsible for a great morbidity among women of reproductive-age and significant burden to the health care system due to rising vaginitis-related health care costs. It is estimated that about 75% of all women at the childbearing age are afflicted by VVC at least once in their lifetime with ∼40–50% experiencing at least one additional episode of infection, ∼5–8% of women suffer from at least four recurrent VVC per year (Peters et al., 2014; Cassone, 2015).

It is well accepted that the interactions between C. albicans and other components of the resident bacteria, particularly Lactobacillus spp., play important roles in determining the commensal or pathogenic outcome of this fungus (Förster et al., 2016). Lactobacilli are predominant microorganisms in the normal vaginal microbiota, where the most frequently isolated species are L. crispatus, L. gasseri, and L. jensenii (Petrova et al., 2015). In a healthy status, C. albicans and lactobacilli naturally co-inhabit in the female genitourinary tract, where the former lives as a minority co-habitant. It is reported that lactobacilli could inhibit the overgrowth of C. albicans by producing a wide variety of secondary metabolites with antimicrobial activity, such as lactic acid, hydrogen peroxide (H₂O₂), bacteriocin-like compounds, and biosurfactants (Borges et al., 2014). For instance, bacteriocin L23 produced by L. fermentum isolate displayed inhibition effect on C. albicans growth (Pascual et al., 2008). However, the mechanisms underlying anti-Candida activity are still not clearly understood.

When the balance of vaginal flora is disturbed by the weakened immune system, antibiotic usage and other risk factors, the lactobacilli biomass would decrease and its species composition might shift, while C. albicans could overgrow and undergo a morphogenetic change from a round-ovoid typical yeast cell to a hyphal growing organism (Ma et al., 2012; Cassone, 2015). The yeast-to-hyphae transition is believed to be a determinant of C. albicans pathogenesis, and it could be triggered by various environmental cues in vitro, including neutral pH, presence of serum or N-acetyl glucosamine, elevated carbon dioxide concentration and physiological temperature (Mayer et al., 2013; Palková and Váchová, 2016). In addition, the morphological transition from yeast to hyphae is regulated by a complex network of signaling pathway in C. albicans. Several virulence-related genes that are specifically expressed in hyphae have been suggested to play essential roles in C. albicans pathogenesis (Sudbery, 2011; Jacobsen et al., 2012). For example, the hyphal wall protein 1 gene (HWP1), agglutinin-like protein gene 3 (ALS3), and extent of cell elongation gene 1 (ECE1) encode cell surface proteins, which are important for hyphal growth and adhesion to host cells (Fan et al., 2013; Vila et al., 2016). Expressions of these hyphae-specific genes are positively regulated by transcription factors Efg1 and Tec1, and are also negatively regulated by the transcriptional repressors Nrg1 (Shapiro et al., 2011).

The use of probiotic lactobacilli against VVC has emerged as an attractive therapeutic option in the era of antibiotic resistance (Matsubara et al., 2016a). Several clinical trials reported that the commercially available probiotic Lactobacillus spp., such as L. rhamnosus GR-1, L. reuteri RC-15, and L. plantarum I1001, are beneficial to women’s urogenital health by decreasing the risk of VVC (Reid et al., 2003; Martinez et al., 2009; Murina et al., 2014; Palacios et al., 2016). However, they are seldom reported as the main resident Lactobacillus species in healthy vagina. Moreover, little is known about the roles and underlying mechanisms of yeast-to-hyphae transition in C. albicans regulated by lactobacilli, especially by those resident Lactobacillus species (Orsi et al., 2014b; Matsubara et al., 2016b; Oliveira et al., 2016).

The aim of this study was to investigate the effects of the most frequently found Lactobacillus spp. in the healthy vagina, i.e., L. crispatus, L. gasseri, and L. jensenii, on growth, hyphal formation and virulence-related genes expression of C. albicans. We found that all 18 Lactobacillus strains tested could inhibit the growth and hyphal formation of C. albicans in a strain-specific manner. In addition, for the first time, we demonstrated that L. crispatus inhibited C. albicans hyphal formation by down-regulating the hyphae-specific genes ALS3, HWP1, and ECE1 possibly through up-regulating the expression of transcriptional repressor gene NRG1. Having the strongest inhibition effects on C. albicans growth and hyphal formation, and being the most commonly found resident Lactobacillus species in the healthy vagina, L. crispatus shows a great potential as a probiotic candidate.

Vulvovaginal candidiasis, primarily caused by C. albicans, is one of the most common urogenital diseases affecting women of reproductive-age and associated to the disruption of the healthy vaginal microbiota (Achkar and Fries, 2010). Lactobacillus-dominated microbiota is thought to be a valuable biomarker for vaginal health since the Lactobacillus species can create a barrier against pathogen invasion (Petrova et al., 2015). L. crispatus, L. gasseri, and L. jensenii were the most frequently isolated species in the vagina of healthy women (Pendharkar et al., 2013; Al et al., 2014; Madhivanan et al., 2015), where C. albicans is often a minority co-habitant of vaginal microbiota. In this study, we investigated how these three Lactobacillus species affected C. albicans growth and yeast-to-hyphae transition, which are implicated in its pathogenesis.

We studied the antimicrobial activities of Lactobacillus strains against C. albicans using two assays, i.e., the agar overlay method and the microplate-based liquid medium assay. The results were similar in terms of having or not having inhibition activities except for three strains of L. jensenii (A083, B021, and B161) (Table 2 and Figure 1). The same phenomenon was also found by other researches. For example, Coman et al. (2014) evaluated the growth inhibition activities of L. rhamnosus IMC 501^® and L. paracasei IMC 502^® on different pathogens by several antimicrobial assays, such as agar well diffusion method and liquid co-culture assay. They found that the inhibition results were different within the different methods, and suggested that this might be due to the various antimicrobial mechanisms. In this study, the liquid medium assay using Lactobacillus CFS seemed to be more sensitive to assess the antimicrobial activities, where 100% (18/18) of the tested Lactobacillus strains showed anti-Candida activities rather than 83.3% (15/18) in agar overlay assay. We speculated that the solid agar might block the diffusion of antimicrobial compounds to reach the target C. albicans, which might affect those compounds with hydrophobic nature more significantly. This also suggests the necessity of using different methods in assessing the antimicrobial activities and comparing results from different studies in an objective way. Anyhow, our data suggests that the tested Lactobacillus strains had moderate to strong anti-Candida activities with inter-strain variations. Our results are consistent with some previous studies showing that several strains of lactobacilli were effective at inhibiting the growth of C. albicans (Parolin et al., 2015; Hütt et al., 2016).

Lactobacilli are known for their production of various antimicrobial compounds to prevent vaginal infection (Borges et al., 2014). H₂O₂ is one of the antimicrobial compounds. We found that Lactobacillus strains had different H₂O₂ production abilities, ranging from no evidence of H₂O₂ production to strong production (Table 2). However, all tested Lactobacillus strains showed growth inhibition effects on C. albicans regardless of H₂O₂ production, suggesting that they might produce some antimicrobial compounds in addition to H₂O₂. Organic acid production and consequent pH reduction by lactobacilli growth have also been suggested as one of the mechanisms for growth inhibition on Candida in previous studies (Jiang et al., 2015). However, we found that the anti-Candida activities of the tested CFS might not be largely attributed to low pH (Figure 1). Each of CFS from the tested Lactobacillus strains showed significantly higher antimicrobial activity than MRS-HCl liquid (pH 4.0) (P < 0.01). It suggests that other antimicrobial compounds might play roles in the growth inhibition effects of CFS on C. albicans. The antimicrobial compounds, such as bacteriocin-like substances and biosurfactants, have been suggested to contribute to the antagonistic effects of probiotic Lactobacillus strains against a variety of vaginal pathogens (Madhu and Prapulla, 2014; Maldonado-Barragán et al., 2016).

In this study, for the first time, the effects of CFS from three species of Lactobacillus (L. crispatus, L. gasseri, and L. jensenii) on yeast-to-hyphae transition of C. albicans were compared. The ability of yeast-to-hyphae transition is essential for C. albicans virulence (Mayer et al., 2013). The hyphae form, which can promote tissue penetration and escape from immune cell, is more prevalent in the infection process than yeast form (Lu et al., 2014). Blocking of virulence traits of pathogens, such as yeast-to-hyphae transition in C. albicans, has been recently considered as a new antifungal paradigm (Tsang et al., 2012). In this study, L. crispatus showed good inhibition activity on hyphal formation (Figures 4, 5). Longitudinal studies have also shown that the presence of L. crispatus promoted stability of the vaginal microbiota toward a healthy status (Verstraelen et al., 2009; Santiago et al., 2012). Our findings of L. crispatus strains with generally high inhibition activities on Candida growth and yeast-to-hyphae transition is very attractive. As far as we know, the clinical approved L. crispatus strains are very limited. Our results hint that L. crispatus strains are highly worthwhile for further investigation in developing the new antifungal strategies.

Our preliminary results revealed that the antimicrobial compounds from L. crispatus B145 against Candida growth were heat stable and protease resistance. But the same treatments would greatly reduce the residual activity against yeast-to-hyphae transition. This hints that the growth and yeast-to-hyphae transition inhibition activities would be attributed to different secretory antimicrobial compounds of L. crispatus B145. Tendulkar et al. (2007) have reported to identify a heat stable and protease resistance biosurfactant compound, which had anti-fungi activity. The compound was characterized to be lipopeptide in chemical nature. To further study whether L. crispatus B145 was able to inhibit different clinical relevant C. albicans strains, we tested the growth and hyphae inhibition effect of its CFS on five additional vaginal isolates of C. albicans. We found that the growth inhibition rates were from 50.7 to 68.6%, and the hyphae formation inhibition rates were from 77.9 to 91.7% (Supplementary Figure S1). These results confirmed our findings in C. albicans ATCC 10231. Further studies to purify the bioactive compounds from L. crispatus B145 and to distinguish different compounds in growth and hyphal inhibition would be interesting.

To elucidate the potential mechanisms involved in the inhibition of hyphal formation of C. albicans by the L. crispatus, we studied the expression levels of five hyphae-specific genes, i.e., ALS3, HWP1, ECE1, EAP1, and SAP5, in response to L. crispatus CFS under hyphae-inducing conditions. These genes encode proteins that are essential for hyphal formation and also play roles in Candida pathogenesis. ALS3 encodes Als3, which is a member of the agglutinin-like sequence gene family and contributes to the invasion of cells and subsequently cell damage (Liu and Filler, 2011). HWP1 encodes a cell wall mannose protein, which functions as an adhesion, required for hyphal formation and yeast adhesion to epithelial cells (Orsi et al., 2014a). ECE1 encodes a membrane protein, which is essential for cell elongation and biofilm formation (Fan et al., 2013). EAP1 gene encodes a glycosylphosphatidylinositol-anchored, glucan-cross-linked cell wall protein, in adhesion and biofilm formation (Li and Palecek, 2003). SAP5 encodes a member of the secreted aspartic proteases family, which are important for the pathogenesis of candidiasis (Sanglard et al., 1997). The results of qRT-PCR analysis revealed that all seven L. crispatus CFS significantly down-regulate the expression of hyphal genes ALS3, HWP1, and ECE1, associated with the inhibition of hyphal formation. Moreover, C. albicans hyphal maintenance requires specific transcriptional regulation mechanisms (Fan et al., 2013). We measured the expression of positive regulator genes of EFG1 and TEC1, and the negative regulator gene of NRG1 (Braun et al., 2001; Argimón et al., 2007; Daniels et al., 2015). Expression of the negative hyphal regulator gene NRG1 was found to be significantly up-regulated in C. albicans in the presence of L. crispatus CFS, implying that L. crispatus might repress the expression of hyphae-specific genes in a NRG1-dependent manner. To the best of our knowledge, this is the first study demonstrating the role of L. crispatus in inhibition of C. albicans virulence-related gene expressions. Interestingly, L. crispatus B145 was shown to strongly up-regulate NRG1 expression (3.398-fold) and down-regulate ALS3 (0.057-fold), HWP1 (0.033-fold), and ECE1 (0.003-fold) gene expressions. If the antibodies are available, experiments such as Western blot would further confirm this finding.

The present study demonstrated the in vitro inhibitory effects of vaginal L. crispatus, L. gasseri, and L. jensenii on C. albicans growth and hyphal formation. L. crispatus had a generally better ability to inhibit C. albicans and might down-regulate the expression of hyphae-specific genes ALS3, HWP1, and ECE1 in a NRG1-dependent manner. The effects might be due to their productions of some secretory antimicrobial compounds in addition to H₂O₂ and organic acids. The growth and hyphal inhibition compounds from L. crispatus B145 seemed to be different in chemical nature. Further investigations on L. crispatus B145 and other clinical isolates would facilitate the development of probiotic agents with great potentials against C. albicans.

SW and TL designed experiments, analyzed data and wrote the paper; SW and QW carried out experiments; LY helped with assay set-up; HZ supervised the project and paper writing; EY gave critical review on the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.