Introduction

Echinocandins play a vital role in the treatment of invasive fungal infections [1, 2]. Candida spp. isolates are commonly susceptible to echinocandins in vitro, with the majority of isolates inhibited at MICs of 0.03–0.25 µg ml−1. Intrinsically decreased susceptibility to the echinocandins is noted in Candida orthopsilosis, C. parapsilosis, Candida lusitaniae, and Candida guilliermondii [3]. Despite the growth-inhibitory activity of echinocandins at low drug concentrations, paradoxical growth (PG), defined as a re-growth at drug concentrations above the MIC, has been observed for Candida spp. isolates. To date no uniformly consistent incidence of PG phenomenon was observed. PG has been demonstrated in 14–90% of Candida spp. isolates, depending on echinocandin, Candida species, and group of isolates examined. The published data depicts that PG is more commonly observed with caspofungin than with other echinocandins [4,5,6,7,8,9].

To date, studies describing PG in Candida spp. isolates have focused mainly on planktonic cells and demonstrated the crucial role of high echinocandin concentrations in this phenomenon. There is few data on the PG incidence in Candida spp. biofilms [10,11,12]. We report here the ability of three Candida spp. growing as sessile cells to display PG, when they are exposed to caspofungin in vitro, even at relatively low drug concentrations. Clinical isolates from three Candida spp. were tested, C. albicans, C. parapsilosis, and C. glabrata. The in vitro susceptibilities of Candida spp. sessile cells to caspofungin and the frequency of PG were studied.

Materials and methods

Strains

A total of 70 Candida spp. isolates were examined: 29 of C. albicans, 26 of C. parapsilosis, and 15 of C. glabrata. All the strains were isolated in The Department of Microbiology of Dr. A. Jurasz University Hospital No.1 in Bydgoszcz, Nicolaus Copernicus University in Torun, in 2007–2010. One isolate per patient was studied. Isolates were derived from various specimens: urine (n = 23), blood cultures (n = 9), bronchoalveolar lavage fluid (n = 4), middle ear discharge (n = 7), pharyngeal swab (n = 6), wound swab (n = 4), peritoneal fluid (n = 5), tip of vascular catheter (n = 3), gastrostomy swab (n = 2), and one strain per tip of drain, piece of surgical mesh, respiratory tract swab, insertion-site skin swab, peritoneal cavity swab, tracheotomy swab, and groin swab. Biofilm-forming strain C. albicans GDH 2346 served as a positive control in the whole experiment. All examined strains were stored at −70 °C. Species identification was confirmed before research by means of germ-tube tests [13], and MALDI-TOF mass spectrometry carried out according to the manufacturer’s protocol (Bruker).

Antifungal susceptibility testing

Caspofungin acetate (Merck & Co., Inc.) concentrations ranging from 0.015 to 256 µg ml−1 were tested. MICs for planktonic cells were determined by EUCAST (European Committee on Antimicrobial Susceptibility Testing) method for antifungal susceptibility testing of yeasts [14]. C. parapsilosis ATCC 22019 served as quality control strain. Sessile MICs (SMICs) were tested by means of methods published by Ramage et al.[15], and Krom et al. [16]. These methods had been previously used in our study on micafungin and described in detail [17]. The colorimetric assay based on MTT (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H tetrazolium bromide) (AlfaAesar) reduction was engaged to assess cell viability and by this means drug susceptibility of Candida spp. sessile cells [16, 17]. The metabolic activity of Candida spp. sessile cells were determined quantitatively by the absorbance of obtained solutions measurement in a microtiter plate reader (spectrophotometer using KC4™ v3.4 and KC4 ™ Signature programs) (BIO-TEK) at 550 nm, after transferring the solutions to a new 96-well plates. SMICs were established at ≥50% biofilm inhibition comparing to untreated biofilms.

Paradoxical growth

The PG frequency in Candida spp. isolates exposed to caspofungin was studied. PG was considered as a resurgence of growth, >50% of that in the drug-free growth control well, at drug concentrations above the MIC.

Results

MICs of caspofungin

A total of 29C. albicans (MICs ≤ 0.015 µg ml−1) and 15C. glabrata (MICs ranged from ≤ 0.015 to 0.06 µg ml−1, MIC90 = 0.06 µg ml−1) isolates were susceptible, and 26C. parapsilosis isolates (MICs ranged from ≤ 0.015 to 2 µg ml−1, MIC90 = 1 µg ml−1) were intermediate to caspofungin according to EUCAST interpretative criteria [18].

SMICs of caspofungin

Caspofungin appeared to have excellent activity against C. glabrata sessile cells; SMICs ranged from 0.015 to 0.5 µg ml−1, and the median SMIC was 0.125 µg ml−1, and SMIC90 was 0.25 µg ml−1, for 24-hours old sessile cells. The median SMIC for C. albicans 24-hours old sessile cells was 0.25 µg ml−1, however, SMICs > 2 µg ml−1 were observed for 6 of 29 (20.7%) isolates (SMICs ranged from 32 to 128 µg ml−1).

The SMICs ≤ 2 µg ml−1 (0.25–2 µg ml−1) were observed for 5 of 26 (19.2%) C. parapsilosis 24-hours old sessile cells; the rest SMICs ranged from 4 to >256 µg ml−1. Detailed data are presented in Tables 1–3.

Table 1 Caspofungin SMICs and PG for C. albicans isolates (n = 29)
Full size table
Table 2 Caspofungin SMICs and PG for C. parapsilosis isolates (n = 26)
Full size table
Table 3 Caspofungin SMICs and PG for C. glabrata isolates (n = 15)
Full size table

Paradoxical growth

PG was observed for three tested Candida species, at each sessile cells age tested. The frequency of PG differed by species, as well as by sessile cells age.

PG was the most frequent among C. albicans isolates, and among 2-hours old sessile cells in case of each Candida species tested. PG was noted as follows: for 2-hours old sessile cells: in 93.1% of C. albicans (27 of 29 isolates), in 69.2% of C. parapsilosis (18 of 26 isolates), in 66.7% of C. glabrata (10 of 15 isolates); for 6-hours old sessile cells: in 86.2% of C. albicans (25 of 29 isolates), in 60.0% of C. glabrata (9 of 15 isolates), in 34.6% of C. parapsilosis (9 of 26 isolates); for 24-hours old sessile cells: in 79.3% of C. albicans (23 of 29 isolates), in 46.7% of C. glabrata (7 of 15 isolates), in 26.9% of C. parapsilosis (7 of 26 isolates).

Interestingly, PG effect of caspofungin on Candida spp. sessile cells was observed not only at high caspofungin concentrations (>2 µg ml−1), but also at concentrations ≤2 µg ml−1 (0.25–2 µg ml−1). PG at ≤2 µg ml-1 was noted in 34.5% of C. albicans for 6- and 24-hours sessile cells, in 26.7% of C. glabrata 2-hours old sessile cells, and in 19.2%, 11.5%, and 7.7% of C. parapsilosis, growing as 2-, 6-, and 24-hours old sessile cells, respectively. PG of Candida spp. was observed over one to nine dilutions (Tables 1–3). The range of caspofungin concentrations promoting PG differed by the stage of biofilm maturation. The widest range of caspofungin concentrations triggering PG of C. albicans was noted for 6-hours old sessile cells (0.25–64 µg ml−1), of C. glabrata was noted for 2-hours old sessile cells (0.5–128 µg ml−1).

Discussion

Echinocandins are antifungals, that inhibit the glucan synthesis in the fungal cell wall, by 1,3-β glucan synthase inhibition.We have described the effect of caspofungin on sessile Candida spp. cells. We focused on their susceptibility to caspofungin and the PG effect of the drug.

Our studies revealed the highest antibiofilm activity against C. glabrata isolates. All the examined strains of this species displayed SMICs ≤ 0.5 µg ml−1, regardless of sessile cells age, with SMIC90 value of 0.25 µg ml−1 for 24-hours old sessile cells. Similarly, Choi et al. [19] described caspofungin SMIC90 of 0.5 µg ml−1 for mature biofilms of 9 C. glabrata strains. Besides, Choi et al. [19] revealed also very high antibiofilm activity of caspofungin against 12 C. albicans strains, with SMICs ≤ 0.5 µg ml−1. In contrast, in our investigation, 23.3% of the C. albicans strains displayed SMICs > 2 µg ml−1 for 24-hours old sessile cells, associated with SMIC90 128 µg ml−1. Melo et al. [10] noted caspofungin SMIC of 2 µg ml-1 for 3 out of 4 C. albicans biofilms, and 4 µg ml−1 for one tested biofilm. These inter-study variations highlight the differences in susceptibilities of C. albicans biofilms, which can be due to differences in the biofilm-forming abilities and biofilm properties of the particular tested isolates.

Our results indicate the lowest antibiofilm activity of caspofungin against C. parapsilosis isolates. Caspofungin SMICs ≤ 2 µg ml−1 for 24-hours old sessile cells were noted for 21.4% of the tested isolates. The SMIC50 value for 24-hours old sessile cells was established at 128 µg ml−1. Our results are consistent with those published by Choi et al.[19], in which caspofungin SMICs ≤ 2 µg ml−1 for mature biofilms were noticed for 3 out of 12 (25%) tested strains, and SMIC50 and SMIC90 values were 8 and >16 µg ml−1, respectively. Melo et al. [10] noted caspofungin SMICs of 2 µg ml−1 for 37.5%, 4 µg ml−1 for 25%, and 256 µg ml−1 for 37.5% of 8 C. parapsilosis isolates tested. Similarly, Simitsopoulou et al. [20] noted SMIC range from 2 to 128 µg ml−1, with mean value 64 µg ml−1, for 6 C. parapsilosis biofilms examined.

Our studies, in comparison with studies published before, revealed that antibiofilm activity of caspofungin is species-specific, and strongly strain-depending among C. albicans and C. parapsilosis isolates.

We observed PG in 26.9–93.1% of tested isolates, depending on the Candida species and age of sessile cells. The lower frequency of PG among C. parapsilosis than C. albicans and C. glabrata isolates seems to be mainly due to the fact that 6- and 24-hours old sessile cells of C. parapsilosis were less susceptible to caspofungin than the biofilms of the rest two Candida species tested. Similarly, the lower frequency of PG among C. parapsilosis and C. albicans in 6- and 24-hours old sessile cells than 2-hours old sessile cells was probably due to the lower susceptibility of these forms of living to caspofungin. In contrast, among C. glabrata sessile cells the frequency of PG was lower in older sessile cells, as follows: 2-hours old sessile cells (66.7%) > 6-hours old sessile cells (60.0%) > 24-hours old sessile cells (46.7%), despite the fact that the SMICs of caspofungin persisted low, ranged from 0.015 to 0.5 µg ml−1.

The correlation between origins of the isolates and their susceptibility to caspofungin, as well as PG were analyzed. Among C. parapsilosis strains there were six isolates (23.1%) without PG effect. Their origins were different, including: wound swab, blood, peritoneal fluid, middle ear discharge, and the tip of vascular catheters – two isolates. Otherwise, C. albicans originating from the catheter displayed PG effect. The C. parapsilosis strain without PG effect, isolated from the wound expressed the lowest MIC, i.e. ≤0.015 µg ml−1, and also the lowest SMICs –0.25 µg ml−1. However, MICs of the rest five C. parapsilosis isolates without PG effect, were 0.5 µg ml−1, and their SMICs ranged 1–128 µg ml−1, thus they did not outstand.

All C. albicans isolates displayed PG in at least one biofilm-maturation step. High SMICs of caspofungin for 24-hour C. albicans biofilms, ranging 32–128 µg ml−1, were noted for six isolates, originating from different specimens: urine samples—two isolates, middle ear discharge—two isolates, BAL, and peritoneal fluid. When analyzing the most numerous C. albicans strains isolated from urine samples, it was noticed that SMICs of caspofungin widely ranged 0.06–128 µg ml−1, SMIC50 was 0.25 µg ml−1, and SMIC90 –64 µg ml−1. Likewise, no correlation between the origins of C. glabrata isolates and PG effect of caspofungin was observed. Only one isolate originating from the insertion-site skin swab did not display PG effect at all.

In our previous study PG effect of other echinocandin, i.e. micafungin was less frequent. PG effect of micafungin was detected in 3.3–31% C. albicans, and 7.1–14.2% C. parapsilosis. None C. glabrata displayed PG although SMICs values were low, ranged 0.015-0.25 µg ml−1. For six C. parapsilosis isolates without PG effect of caspofungin, the lack of PG effect of micafungin was also noted [17].

MICs and SMICs of caspofungin seem to be unrelated to the origins of the Candida spp. isolates. Although the highest MICs of caspofungin, i.e. 2 µg ml−1 were observed only in two C. parapsilosis strains isolated from blood, MICs for the rest isolates from blood ranged 0.25–1 µg ml−1.

Further analyzes showed that higher MICs of caspofungin in C. parapsilosis isolates did not result in higher SMICs of the drug. Susceptibility of the C. parapsilosis biofilm to caspofungin seems to depend on other mechanisms than the susceptibility of planktonic cells.

Although PG was more frequent at high caspofungin concentratrions, this was noted also at realtively low caspofungin concentrations, i.e. ≤2 µg ml−1. Interestigly, PG was noted in 7.7–34.5% of Candida spp. isolates at the drug concentrations of 0.25–2 µg ml−1, depending on Candida species, and sessile cells age.

PG of Candida spp. biofilms exposed to caspofungin was described before by Melo et al. [10]. The SMICs of caspofungin were determined for 30 clinical Candida spp. isolates (4 C. albicans, 6 Candida tropicalis, 7 C. parapsilosis, 8 C. orthopsilosis, and 5 Candida metapsilosis isolates), when they were grown as biofilms. The SMICs ranged from 2 to 512 µg ml−1. PG was observed for 80.0% of Candida spp. isolates and the range of caspofungin concentrations promoting PG ranged from 4 to 512 µg ml−1. Melo et al.[10] did not observe PG at caspofungin concentrations lower than 4 µg ml−1, as well as they did not note SMICs lower than 2 µg ml−1.

Next, Walraven et al. [12] analyzed biofilms formed by 12 C. albicans clinical strains, resistant to caspofungin due to FKS1 hot-spot mutations. The sessile antifungal activities of anidulafungin, caspofungin, and micafungin were compared, and PG effects were assessed. The majority of strains displayed PG to particular echinocandins, in either planktonic or sessile forms. Caspofungin at the concentrations of 64 to ≥128 µg ml−1 promoted PG in C. albicans biofilms in 58.3% of isolates tested, but did not in planktonic forms.

The clinical significance of PG effect during echinocandin therapy of candidiasis remains unclear. Clinical data on paradoxical effect of echinocandins are limited. An international, multicenter, randomized, double-blind clinical trial included 204 patients with proven invasive candidiasis; 104 patients receiving a standard (70 mg followed by 50 mg/day), and 100 patients receiving high-dose (150 mg/day) caspofungin treatment regimen [21]. Favorable responses were similar between the two groups, 71.6% of patients who received the standard regimen, and 77.9% of patients who received the high-dose regimen. Obtained results suggest no paradoxical effect associated with high caspofungin concentrations in regards to efficacy. In another study, Safdar et al. [22] demonstrated that high-dose caspofungin regimen may have favorably influenced outcomes. In the retrospective analysis 34 patients receiving high-dose caspofungin (150 mg/day) were compared with 63 patients receiving standard-dose caspofungin (70 mg followed by 50 mg/day). After 12 weeks of treatment 44% of patients revealed complete or partial response compared with 29% of patients receiving standard-dose caspofungin regimen. Further clinical studies are needed to evaluate efficacy of high-dose echinocandins in patients with invasive fungal infections.

Clemons et al. [23] observed paradoxical effect of caspofungin in animal model, in only one instance of murine candidiasis, however, the effect was not reproducible in a subsequent experiment. Published clinical, and animal data indicate that the PG is rather an in vitro phenomenon, or is somehow balanced by in vivo conditions. Shields et al. [6] demonstrated that PG is eliminated in human serum in vitro. Furthermore, brief exposures to caspofungin resulted in killing of C. albicans cells in vitro, but did not lead to PG, suggesting compensatory mechanisms occur with prolonged exposure to drug [6]. There are some potential mechanisms responsible for PG hypothesized; [5, 7, 9] they include (i) an increased cell wall chitin content, (ii) upregulation of the protein kinase C cell wall integrity pathway, and (iii) involvement of the calcineurin pathway.

In conclusion, caspofungin presents high anti-Candida spp. biofilm activity, however, this activity is species-specific, and strongly strain-depending among C. albicans and C. parapsilosis isolates. Candida spp. sessile cells can display PG in the presence of high, but also relatively low concentrations of caspofungin. This finding makes the significance of the PG in vivo even more complex and difficult to examine. In fact, the PG phenomenon might be one of the reasons for the lack of success in therapy, even when standard-dose echinocandin treatment is applied.