Risk assessment for the spread of Candida sp. in dental chair unit waterlines using molecular techniques

Abstract

Aims: The aim of this study was to evaluate the presence of yeasts in dental chair unit waterlines (DCUWLs) and to test their ability to form biofilms. Materials and methods: Eighteen dental waterlines were analysed by culture in liquid Sabouraud in order to allow the quantification and the purification of isolated yeasts from their internal surfaces. All isolates were identified by standard laboratory procedures, including CHROMagar Candida medium for orientation, commercial yeast identification system Api Candida, MALDI-TOF MS and DNA sequencing. To evaluate their kinetics of antifungal susceptibility during different phases of biofilm formation, these yeasts were subjected to three antifungal agents. Results: From the 18 DCUWLs studied, 10 were altered (55.56%). Eleven strains of Candida sp. [Candida albicans (2), Candida guilliermondii (5) and Candida glabrata (4)] and two species of non-Candida; Rhodotorula spp. (1) and Trichosporon spp. (2) were identified. The majority of yeasts in planktonic form were susceptible to amphotericin B, caspofungin and voriconazole, except C. albicans was resistant to voriconazole. In the biofilm form, caspofungin was the most effective antifungal agent for all isolated strains. For the other antifungal agents, sessile cells were resistant. Conclusion: Several types of yeasts were identified; the most frequently isolated genus was Candida. The majority of these yeasts had the ability to form biofilms and resisted antifungal agents used in this study

Introduction

Dental practice involves invasive surgical procedures in a naturally septic environment1., 2.. In a dental clinic, care is delivered in a dental unit (chair).

Dental chair units consist of a network of interconnected narrow-bore tubes called dental chair unit waterlines (DCUWLs). These DCUWLs constitute an ideal growth medium for microorganisms due to many factors, including: ambient temperatures, diameter of dental waterlines and stagnating water in these canalisations³.

However, once inside the DCUWLs, microorganisms settle on the internal face and trigger a series of events leading to: colonisation, formation of microscopic colonies and, eventually, to biofilm⁴.

In addition, few documented reports have evaluated the presence of yeasts in DCUWLs. Most scientific studies published in this area focus on bacterial contamination5., 6.. In this work, we were interested in: (i) the research on the presence of yeasts inside DCUWLs; (ii) the identification of isolated yeasts by conventional and molecular techniques; (iii) the study of their susceptibility to three antifungal agents and their ability to form biofilms; and (iv) the study of biofilms’ kinetics of susceptibility during different phases of development.

Materials and methods

Samples of DCUWLs (18) were taken from the dental clinic (13) and stomatology unit (5) at the university hospital of Tlemcen (Algeria). After disinfecting the external surface with alcohol, these samples were cut using a sterile scalpel. The parts collected from the DCUWLs were taken at random positions along the line. A sterile swab was inserted in the aperture of the samples and immersed in sterile Sabouraud liquid.

Strain identification

CHROMagar Candida medium (BBL-CHROMagar Candida, Becton Dickinson) was used for isolation and rapid purification of yeasts. Api Candida and MALDI-TOF MS were used to identify the isolated strains. For MALDI-TOF MS identification, the procedure recommended by the manufacturer (Bruker Daltonics GmbH, Germany) was followed.

DNA sequence-based identification of clinical strains

The sequencing of fungal ribosomal DNA was based on the amplification of internal transcribed spacer regions ITS1–ITS2.

Primers used for the amplification of ITS1-5.8s-ITS2 region were ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′)⁷.

The obtained sequences were subjected to a sequence homology search using the program of alignment Blast (Basic Alignment Search Tool).

Biofilm growth kinetics on the surface of microtitre plate wells

All yeasts recovered from the DCUWLs were selected to study their ability to form biofilms on the surfaces of microtitre plates. The biofilm-forming species were retained to study their capacity to form biofilms against antifungal agents over time interval of 4, 24, 48 and 72 hours⁴.

Antifungal susceptibility

Planktonic cells [minimal inhibitory concentrations (MICs)]

The determination of MICs of planktonic cells was performed by the Clinical Laboratory Standards Institute (CLSI) M27-A3 broth microdilution method with a reading of endpoints at 24 hours (caspofungin) and 48 hours (amphotericin B and voriconazole)⁸.

Sessile cells [sessile minimal inhibitory concentrations (SMIC)]

The action of amphotericin B, voriconazole and caspofungin was tested on formed biofilm. The ability of Candida sp. strains to form biofilms was considered positive if the SMICs were above the MIC. MICs and SMICs were determined using the XTT-reduction assay⁹.

XTT-reduction assay

For the quantification of living cells, the XTT-reduction assay was used. The principle of XTT-reduction assay was to measure metabolic activities of cells within the biofilm⁹.

Results

A total number of 14 yeasts were isolated and identified.

In terms of frequency, among the 11 isolates of Candida sp., Candida albicans accounted for 14.3% (n = 2) of the isolates, while species not C. albicans accounted for 64.3% (n = 9) of isolates, mainly represented by Candida guilliermondii 35.7% (n = 5) and Candida glabrata 28.6% (n = 4).

Other species not members of the Candida genus (21.4%), such as Rhodotorula spp. 7.1% (n = 1) and Trichosporon spp. 14.3% (n = 2) were also isolated.

The conventional identification of yeasts was based on the use of morphological (CHROMagar Candida medium) and biochemical tests (API Candida, bioMérieux).

MALDI-TOF MS was used to confirm or to correct the identification of Candida sp. previously identified by Api Candida. The log score (LS) of each strain was used for the identification.

MALDI-TOF MS technique allowed the confirmation of the identification of two strains; C. guilliermondii/Candida famata to C. guilliermondii and to correct Candida parapsilosis to C. glabrata.

The results of our study showed that all the values of best scores LS were above 2. The strains presenting scores < 2 were subjected to further tests.

The results of the identification at the species level by API Candida, CHROMagar Candida and MALDI-TOF MS are presented in Table 1.

Table 1.

Identification by API Candida, CHROMagar Candida and the values of the best-score and the second-score (LS) of identification by MALDI-TOF MS

Api Candida	CHROMagar Candida	MS-based identification
		Best score LS	Species	Second score LS
C. albicans	C. albicans	2.373	C. albicans	2.337
C. albicans	C. albicans	2.37	C. albicans	2.369
C. guilliermondii/C. famata	C. guilliermondii	2.238	C. guilliermondii	2.017
C. guilliermondii/C. famata	C. guilliermondii	2.103	C. guilliermondii	1.875
C. guilliermondii/C. famata	C. guilliermondii	2.195	C. guilliermondii	1.981
C. guilliermondii/C. famata	C. guilliermondii	2.231	C. guilliermondii	2.066
C. guilliermondii/C. famata	C. tropicalis	2.094	C. guilliermondii	1.916
C. glabrata/C. famata/Geotrichum spp	C. glabrata	2.543	C. glabrata	2.515
C. glabrata/C. famata/Geotrichum spp	C. glabrata	2.528	C. glabrata	2.495
C. parapsilosis	C. glabrata	2.565	C. glabrata	2.496
C. glabrata	C. glabrata	2.577	C. glabrata	2.49

LS, log score.

Comparing those results, there was a discrepancy in the identification between the techniques; 63.6% of strains had a correct identification by the three techniques. In effect, all isolates have been correctly identified using the MALDI-TOF MS.

DNA sequencing

Only yeasts with discordant identification results were selected for this technique.

The obtained sequences were then compared using an alignment tool sequences fungi (Table 2).

Table 2.

Characteristics of yeast included in the reference DNA sequence database

DNA sequence-based identification
Species	n	Strains no.	BLAST ref. with ≥ 99% identity
C. guilliermondii	2	5	KC119207.1
		17	JQ678690.1
C. glabrata	3	13	CR380958.2
		8	CR380958.2
		T3B2G	CR380958.2

The sequencing of inter-genic regions ITS1 and ITS2 confirmed the identification of strains by MALDI-TOF MS.

The final result of identification of isolated yeasts from DCUWLs revealed the following: two (2) strains of C. albicans; five (5) strains of C. guilliermondii; four (4) strains of C. glabrata; one strain (1) of Rhodotorula spp.; and two (2) strains of Trichosporon spp.

Determination of MIC

The MIC of caspofungin was calculated after 24 hours, whereas the MIC of voriconazole and the MIC of amphotericin B were calculated after 48 hours of incubation. The same conditions were applied to determine the SMIC.

According to the Clinical Laboratory Standard Institute (CLSI; M27-S3)¹⁰, the yeast was considered susceptible to amphotericin B if the MIC was ≤ 1 μg/mL and was susceptible to casposfungin and voriconazole if the MIC was ≤ 2 μg/mL.

The results indicated that all isolated yeasts were susceptible to amphotericin B (MIC ≤ 1 μg/mL), caspofungin (MIC ≤ 0.5 μg/mL) and voriconazole (MIC ≤ 1 μg/mL), with the exception of two C. albicans strains that were resistant to voriconazole (MIC = 16 μg/mL).

Determination of SMIC

Table 3 reported the results of SMICs of amphotericin B, caspofungin and voriconazole against isolated and reference strains.

Table 3.

In vitro antifungal susceptibility of isolated yeasts recovered from 18 DCUWLs under planktonic (MIC) and biofilm (SMIC) growing conditions with amphotericin B, caspofungin and voriconazole (μg/mL)

Isolates	Amphotericin B			Caspofungin			Voriconazole
	Planktonic	Sessile		Planktonic	Sessile		Planktonic	Sessile
	MIC	MIC50	MIC80	MIC50	MIC50	MIC80	MIC50	MIC50	MIC80
C. albicans 1	0.25	2	4	0.0313	0.25	1	16	32	˃ 32
C. albicans 2	0.5	4	8	0.0313	0.5	2	16	˃ 32	˃ 32
C. glabrata 1	0.125	0.5	1	0.0313	0.0625	0.125	0.5	1	1
C. glabrata 2	0.125	0.5	2	0.125	0.25	0.5	0.25	2	4
C. glabrata 3	0.125	0.5	2	0.125	0.125	0. 25	0.25	1	2
C. glabrata 4	0.25	1	2	0.0625	0.125	0.5	0.5	2	4
Rhodotorula sp.	1	2	4	0.0625	0.25	0.5	0.0625	2	4
C. albicans ATCC 10231	0.25–0.5	2	4–8	0.0625	0.5	1	1	1	2

MIC, minimal inhibitory concentration.

Caspofungin was more effective against all isolated species (0.0625 ≤ SMIC₅₀ ≤ 2 μg/mL), with SMIC₅₀ ≤ 2 μg/mL for C. guilliermondii and SMIC₅₀ ≤ 0.5 μg/mL for C. albicans, C. glabrata, Rhodotorula spp. and C. albicans ATCC 10231.

Furthermore, for amphotericin B, a single strain of C. glabrata was susceptible (MIC ≤ 1 μg/mL) under planktonic and sessile forms. For the other C. glabrata and isolated strains, all were resistant (SMIC₈₀ > 1 μg/mL).

The results indicated that biofilms formed were, intrinsically, resistant to amphotericin B as indicated by SMIC₈₀. Also, activity against biofilms was increased up to 4–512 times compared with that against planktonic cells.

However, SMIC₅₀ of voriconazole showed that 50% of C. glabrata, 40% of C. guilliermondii and Rhodotorula spp. were susceptible, at a dependent dose, to voriconazole (SMIC = 2 μg/mL). For C. albicans all strains were resistant to this antifungal agent with SMIC = 32 μg/mL.

For C. albicans ATCC 10231, the obtained MICs of amphotericin B, caspofungin and voriconazole were 4, 0.5 and 1 μg/mL, respectively. These MICs were superior to those obtained by CLSI method, except for voriconazole, where the SMIC was the same as that of planktonic cells.

Determination of SMICs during different phases of formation of biofilms

The SMICs were measured according to the protocol of Pierce et al.⁹ after 4, 24, 48 and 72 hours of incubation at 37 °C.

The result of kinetic susceptibility of selected sessile cells in the presence of amphotericin B, caspofungin and voriconazole are presented in Figure 1.

Figure 1.

Kinetics of biofilm formation of isolated and reference strains Candida sp. and Candida albicans ATCC 10231 against (a) amphotericin B, (b) voriconazole and (c) caspofungine during the different phases of biofilm development.

The XTT reduction test was used to assess the metabolic activities of biofilms and deduct SMICs during different phases of their development.

Four phases of biofilm development were reported: (i) the initial adherence phase (adhesion); (ii) the growth phase; (iii) the maturation phase; and (iv) the final phase.

During adherence and growth phases, we observed that the SMICs were nearest to the values of MIC given previously.

During the maturation phase (24–48 hours), resistance to antifungal agents was clearly increased: from 2 to 16 μg/mL for amphotericin B, from 0.5 to 4 μg/mL for caspofungin, and from 2 to ˃ 32 μg/mL for voriconazole.

Candida glabrata and C. albicans showed significant susceptibility to caspofugin under the different environmental conditions (planktonic or sessile), but C. guilliermondii was less susceptible.

In the last phase of biofilm formation (72 hours), the yeasts were extremely resistant. All strains react in the same way, with increased resistance, with different SMIC (4–16 μg/mL for amphotericin B, 1–4 μg/mL for caspofungin, and 2 to ˃ 32 μg/mL for voriconazole). Moreover, sessile cells were more resistant compared with planktonic cells when the biofilm reached maturity.

Discussion

Yeast are part of the contamination of DCUWLs. From 14 isolated yeast strains, 11 were identified as Candida sp. Indeed, the majority of Candida sp. strains were capable of forming biofilms with a decrease in their susceptibility to the antifungal agents tested.

The presence of Candida sp. and non-Candida sp. in water of dental units was reported by several authors11., 12., 13.. The results of predominant isolated strains from DCUWLs obtained from this study corroborated those of Szymańska¹³ and Kadaifciler et al.¹⁴, which demonstrated that the most isolated yeasts were C. albicans, C. guilliermondii, C. famata, C. curvata and Geotrichum candidum.

Szymańska¹³ reported that there were several origins of fungal contamination of DCUWLs; dental unit feed water, return and re-aspiration of biological fluids as a result of the negative pressure produced when the equipment is turned off.

Other factors may also explain the presence of yeasts in DCUWLs, such as: stagnant water in the tubing (nights, weekends and holidays); the size and nature of tubing surfaces; as well as water temperature¹². These factors promoted not only fungal proliferation, but also biofilm formation¹⁶.

The isolated yeasts were identified by MALDI TOF MS. According to Cassagne et al.⁷ and Normand et al.¹⁷, if the score (LS) was ≥ 2, the identification at the species level was verified.

The strains with a score ˂ 2 had undergone sequencing of their DNA. For this, C. guilliermondii/C. famata, with a score ˂ 2, the sequencing of the inter-genic regions ITS1 and ITS2 confirmed their identification as C. guilliermondii.

Desnos-Ollivier et al.¹⁸ and Savini et al.¹⁹ reported that ITS regions were good tools for biochemical differentiating of C. guilliermondii and C. famata species. This was confirmed in our study.

The results for the determination of MIC were corroborated with those of Seghir et al.²⁰, which showed that all yeasts isolated at Tlemcen University Hospital were susceptible to amphotericin B and caspofungin.

Indeed, according to Ramage et al.²¹ and Pfaller et al.²², amphotericin B and caspofungin were the most effective antifungal agents with very few cases of resistance, and offer an interesting alternative in the management of invasive fungal infections in hospitals²³.

The results were similar to those of Kovacicova et al.²⁴ that showed that the MICs of amphotericin B against the sessile cells of Candida sp. were 30–2000 times higher than those for planktonic forms.

The resistance of these yeasts to amphotericin B were reported by several researchers24., 25.. Generally, the adhesion and biofilm production capacity depends on the specific characteristics of each isolate26., 27., 28..

The results of SMIC during the different phases of biofilm formation were in agreement with other studies29., 30., which showed that the exopolymeric matrix (extracellular), which prevented the penetration and the diffusion of the drug in the biofilm, were missed during the initial phase of biofilm development. This explained the low SMIC in the adhesion phase (4 hours). It was reported that the exopolymeric matrix did not influence the action of antifungal agent on yeasts. This fact suggested that the matrix was not the major determining fact in resistance31., 32..

In the maturation phase, resistance to antifungal agents increased; the biofilm completes its formation, its architecture became more complex (24–48 hours)25., 33.. The increase of SMIC from one stage to another explained the amplification of resistance of sessile cells against antifungal agents.

These results are in agreement with those reported in the literature33., 34.. The biofilms formed by C. albicans, C. glabrata and C. guilliermondii showed more resistance to antifungal agents than their counterparts (planktonic cells)35., 36.. Amphotericin B had an inhibiting effect on these sessile cells with higher concentrations of the drug³⁷.

The azoles (including voriconazole) showed no activity against biofilms formed by Candida sp38., 39.. In our study, the majority of C. glabrata were less susceptible to voriconazole. The prolongation of the incubation time caused the increase of MIC. This fact was confirmed in other studies36., 40..

It was demonstrated that biofilm’s fluconazole resistance could be explained by the expression of the efflux pumps in the first phase of biofilm formation or they could be due to the attachment of azole to glucan in the exopolymeric matrix41., 42., 43., 44..

In addition, factors responsible for the resistance of fungal biofilms included: limited penetration of antifungal agents due to the extracellular matrix; physiological status of sessile cells (a slow growth rate); differential expression of genes (expression of efflux pumps); disturbances of sterols; high cell density in the biofilm; and the presence of persistent cells27., 32..

Conclusion

This study showed that DCUWLs were contaminated with yeasts, the majority of which had the capacity to form biofilms. Mycological contamination of DCUWLs remains a serious problem. Effective cleaning and disinfection practices should be applied to reduce or eliminate pathogenic microorganisms. It is also necessary to check the maintenance protocols of dental equipment to improve the quality of the water, and to regularly check for contamination to minimise the risks of exposure to potentially pathogenic agents.

Acknowledgements

The authors thank Dr Stéphane Ranque and Pr Renaud Piarroux for having allowed them to perform a part of this work in their laboratory (Laboratoire de Parasitologie-Mycologie, APHM Timone, Marseille, France). This research was supported by Abou Bekr Belkaid University Tlemcen, Higher Education and Scientific Research Ministry of Algeria.

Conflict of interest

None declared.