Intestinal Candida parapsilosis isolates from Rett syndrome subjects bear potential virulent traits and capacity to persist within the host

The participants’ data related to the 50 RTT patients and 29 Healthy Controls (HC) included in this study are available in [26]. Stool samples from enrolled subjects [26] were collected, aliquoted as it is and stored at − 80 °C until analysis. Samples were homogenized in sterile Ringer’s solution and plated on solid YPD medium (1% Yeast extract, 2% Bacto-peptone, 2% D-glucose, 2% agar) supplemented with 25 U/ml of penicillin, 25 μg/ml of streptomycin (Sigma-Aldrich) and incubated aerobically at 27 °C for 3–5 days. All fungal isolates grown on the selective medium were further isolated to obtain single-cell pure colonies. Genomic DNA was extracted from pure cultures of the isolated colonies as previously described [27]. Fungal isolates were identified by amplification and sequencing of the ribosomal Internal Transcribed Spacer (ITS) region, using ITS1 (5’-GTTTCCGTAGGTGAACTTGC-3′) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3′) primers [28]. ITS1–4 sequences were then classified by using the BLAST algorithm in the NCBI database (minimum 97% sequence similarity and 95% coverage with a described species).

The ability of fungal strains to penetrate YPD solid medium was tested as previously described [29]. M28-4D and BY4742 S. cerevisiae strains, known to be invasive and non-invasive respectively, were used as controls. The strain invasiveness was assigned with scores from 3 (highly invasive) to 0 (non-invasive).

Fungal cells (~ 10⁵ cells/ml) were grown for 7 days in liquid YPD and YNB media (0.67% Yeast Nitrogen Base w/o aminoacids and (NH₄)₂SO₄ (Sigma-Aldrich)_, 2% glucose), both at 27 °C and 37 °C in order to evaluate hyphae or pseudohyphae formation. Formation of hyphae was inspected by optical microscope observation with a Leica DM1000 led instrument (magnification 40× and 100×) [30].

Biofilm formation was quantified according to a previous published protocol [31]. Briefly, fungal cells (~ 10⁵ cells/ml) were grown in liquid YPD at 37 °C for 48 h in flat-bottom 96-well plates. After the incubation period, cell suspensions were aspirated and each well with the adhered fungal cells was washed three times with deionized H₂O and one time with PBS 1X. Biofilm-coated wells were then incubated with 0.01% of crystal violet (Sigma) for 30 min and washed as above. Finally, each well of the dried 96-well plate was incubated with 100 μl of 100% EtOH for 10 min and biofilm formation was quantified by optical density measurement at 570 nm with a microplate reader (Synergy2, BioTek, USA).

Candida albicans and C. parapsilosis isolates were genotyped by Random Amplification of Polymorphic DNA (RAPD) using the primer Oligo 2 (5’-TCACGATGCA-3′) as described previously [34]. Amplifications were performed according to the following protocol: 5 min at 94 °C, 40 cycles of 30s at 94 °C, 30s at 36 °C and 2 min at 72 °C, followed by a final extension of 10 min at 72 °C. The PCR reaction mix contained 1X PCR buffer 2 mM MgCl₂, 200 μM of dNTPs, 0.4 μM of the primer, 2.5 U of Taq Polymerase and 10 ng of gDNA as template. PCR amplicons were separated using a 2% agarose gel in 1× TAE buffer at 90 V for 2 h and visualized with 0.5 μg/ml ethidium bromide staining. The presence or absence of an amplicon at any position of the gel was used for the construction of a binary matrix for the calculation of samples’ distance similarity according to the Jaccard index [35] by mean of the “vegdist” function within the vegan R package. The samples have been then clustered hierarchically according to the UPGMA method by using the “hclust” function within the stats R package.

Human peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Hypaque density gradient centrifugation (Biochrom, Berlin, Germany) from buffy coats provided by the Transfusion Unit of Ospedale Santa Chiara in Trento, Italy. The experimental plan was approved by the local hospital ethical committee, and informed consent was obtained from all the healthy donors (protocol No: 54896583). Candida isolates were cultured in YPD medium for 18 h at 37 °C. Fungal cells were harvested by centrifugation, washed twice with PBS, heat-killed for 3 h at 65 °C and resuspended in culture medium (RPMI1640; Sigma Aldrich). For stimulation experiments, 5 × 10⁵ PBMCs in RPMI1640 were incubated with 5 × 10⁶ heat-killed C. albicans, C. parapsilosis or RPMI1640 medium alone (negative control) [36]. After the incubation periods (24 h for IL-1β, IL-6, TNF-α, IL-10 production and 120 h for IL-17A, INFγ, IL-22, IL-10 production) cell suspensions were centrifuged and supernatants were collected and stored at − 20 °C until assayed. Each experiment was performed in triplicate.

Cytokine detection i.e. IL-17A, INFγ, IL-22, IL-1β, IL-6, TNF-α, IL-10 production, were assayed using the MAP human cytokine/chemokine kit (Merck Millipore) according to the manufacturer’s instructions (MagPix technology).

Wilcoxon rank-sum tests and Spearman’s correlations were performed using the R software [37] through the stats R package (version 3.1.2) and the psych R package, respectively. Permutational MANOVA (PERMANOVA) test was performed using the adonis()function of the vegan R package with 999 permutations. All p-values have been corrected for multiple hypothesis controlling the false discovery rate (FDR) [38].

We identified 122 fungal isolates belonging to different species (Additional file 1: Table S1). Twenty-four of such isolates were obtained from stool samples of RTT subjects (Additional file 1: Table S1). We discovered a significant reduction of fungal species richness in RTT subjects compared to HC (p = 3.9e-05, Wilcoxon rank-sum test) in agreement with the results obtained using a metagenomic approach from the same study cohort [26]. Candida was the most abundant genus present in both RTT subjects (91.7%) and HC (71.4%) with C. albicans and C. parapsilosis as the two most abundant species in both RTT subjects and HC. Interestingly, we observed an inversion in the relative abundances of C. albicans and C. parapsilosis between RTT subjects and HC. While in RTT subjects 4 out of 24 fungal isolates belonged to C. albicans (16.7%) and 14 out of 24 belonged to C. parapsilosis (58.3%), in the HC 49 out of 98 fungal isolates belonged to C. albicans (50%) and 15 out of 98 belonged to C. parapsilosis (15.3%) (Additional file 2: Figure S1). We then characterized the fungal isolates for putative virulence-associated traits and resistance to antifungals (Additional file 1 Table S1). We found that 50% and 63.8% of fungal isolates from RTT subjects and HC, respectively, were able to form hyphae or pseudohyphae (Additional file 1: Table S1). In addition, we observed that the morphotype switch to hyphae or pseudohyphae was related to the isolates’ invasiveness, with hyphae- and pseudohyphae-forming isolates being the most invasive (Additional file 3: Figure S2A). We also observed that RTT isolates produced more biofilm (p = 1.3e-05, Wilcoxon rank-sum test; Additional file 3: Figure S2B) and were significantly more resistant to fluconazole compared to HC isolates (45.8% of RTT isolates were resistant vs 18.1% of HC isolates; p = 5.1e-06, Wilcoxon rank-sum test). As previously observed, we found the co-occurrence of azole cross-resistance between fluconazole and itraconazole (Spearman’s correlation r = 0.57; p = 2.2e-10) [39]. Almost the totality of the isolates (96.7%) were susceptible to 5-flucytosine, with MIC≤0.125 μg/ml (Additional file 1: Table S1). Candida parapsilosis isolates from HC were susceptible to fluconazole (MIC₉₀ = 2 μg/ml; R = 7.1%) and itraconazole (MIC₉₀ = 0.0156 μg/ml; R = 0%) while C. parapsilosis isolates from RTT subjects exhibited a high resistance to these antifungals (fluconazole, MIC₉₀ > 64 μg/ml, R = 35.7%, p = 0.003, Wilcoxon rank-sum test, Figure 1A; itraconazole, MIC₉₀ > 8 μg/ml, R = 35.7%; Table 1). On the contrary, C. albicans isolates from HC were resistant to fluconazole (MIC₉₀ > 64 μg/ml, R = 24.4%) and itraconazole (MIC₉₀ > 8 μg/ml, R = 63.4%; p = 0.03, Wilcoxon rank-sum test; Fig. 1B) while only one RTT C. albicans isolate was resistant to fluconazole (Table 1). Noteworthy, all other Candida species isolated from RTT subjects (i.e. C. glabrata, C. pararugosa and C. tropicalis) were resistant to fluconazole (MIC90 > 64 μg/ml; R = 100%) and itraconazole (MIC90 = 8 μg/ml; R = 100%) while Candida spp. isolated from HC (i.e. C. deformans, C. intermedia and C. lusitaniae) were completely susceptible to these azoles (Table 1). Taken together these results suggest that RTT isolates may be more difficult to eradicate in case of infection than HC isolates. Finally, we have been able to isolate Trichosporon asteroides and Saccharomyces cerevisiae only from RTT subjects. These isolates were both resistant to fluconazole (with MIC>64 μg/ml and MIC = 8 μg/ml respectively) while Trichosporon asteroides was also resistant to itraconazole (MIC>8 μg/ml). Such species are recognized as potential new emerging fungal pathogens [40] thus representing a potential threat for RTT subjects.

The genetic diversity among Candida isolates was determined by UPGMA hierarchical clustering analysis of Jaccard distances calculated from RAPD genotyping. We observed that C. parapsilosis isolates from RTT samples were genetically unrelated to those from HC, with most of RTT C. parapsilosis isolates clustering in a single group (Fig. 2 and Additional file 4: Figure S3; p = 0.002, PERMANOVA). On the contrary, C. albicans isolates from RTT subjects were genetically more diverse, clustering in different clades of the tree (Additional file 5: Figure S4; p = 0.779, PERMANOVA). It is worth noting that we only obtained 4 C. albicans isolates from RTT samples.

The first step in the immunological response against Candida is the production by innate immune cells of pro-inflammatory cytokines, such as IL-1β, IL-6 and TNFα. Successively, these cytokines promote adaptive immunity mediated by Th1 or Th17 responses [2]. Stimulation of PBMCs with C. albicans and C. parapsilosis isolates from HC and RTT subjects revealed that RTT Candida parapsilosis isolates induced higher levels of IL-1β (p = 0.01, Wilcoxon rank-sum test) and, although only close to statistical significance, TNFα (p = 0.058, Wilcoxon rank-sum test) with respect to HC Candida parapsilosis isolates (Fig. 3). On the contrary, no significant differences were observed in the levels of IL-17A, IL-22 and INFγ (Fig. 3). Furthermore we observed that C. parapsilosis isolates from RTT subjects induced highly significant levels of IL-10 compared to HC C. parapsilosis isolates and RTT C. albicans isolates (p < 0.003, Wilcoxon rank-sum test; Fig. 3) suggesting an increased fungal tolerance towards these C. parapsilosis isolates, potentially favouring fungal persistence within the host. Nevertheless, we did not observe significant differences in the expression of IDO1 in PBMCs stimulated by C. parapsilosis isolates (median HC = 30%, IQR = 22.8–39.3%; median RTT = 44%, IQR = 29.1–48.3%; p = 0.24, Wilcoxon rank-sum test). Since we observed variable levels of Th-driving cytokines, we asked whether Candida isolates were able to induce a different Th1/Th17 polarization. Therefore, we measured the intracellular levels of the key transcription factors T-bet and RORγt (Additional file 6: Figure S5), involved in the differentiation of CD4⁺ naïve cells in Th1 and Th17 cell, respectively [41]. As previously observed in culture supernatants, we measured variable, but not statistically significant levels of T-bet and RORγt in response to the different C. parapsilosis and C. albicans isolates (Additional file 6: Figure S5), reflecting the potential of different strains to elicit an immune reaction at different extents. This could be due to the diverse immune reactivity shown by the different isolates of the same species, as previously reported [42, 43]. However, C. parapsilosis isolates from RTT subjects induced more CD4⁺ cells co-expressing both RORγt and T-bet compared to HC C. parapsilosis isolates (raw p-value = 0.04, FDR-corrected p-value = 0.12; Wilcoxon rank-sum test; Additional file 6: Figure S5c).