Non-antifungal medications administered during fungal infections drive drug tolerance and resistance in Candida albicans

Abstract

Introduction. Fungal infections are increasingly concerning, particularly in immunocompromised patients. These patients often suffer from comorbidities and receive multiple, non-antifungal medications.Gap Statement. The effects of these co-administered medications on fungal cells - and their potential to influence antifungal drug efficacy - are poorly understood.Aim. This study investigates non-antifungal medications commonly administered in parallel to antifungals and evaluates their impact on fungal susceptibility.Methodology. We systematically reviewed clinical guidelines to identify non-antifungal medications frequently co-prescribed with antifungals. Focusing on Candida albicans, the most prevalent fungal pathogen, we examined whether the presence of these drugs influences antifungal responses of C. albicans. First, we tested the selected compounds together with antifungals in combination assays. Interactions were then characterized using checkerboard assays, and the impact on antifungal resistance and tolerance was evaluated through disc diffusion assays. To further explore these effects in vivo, the influence of selected antagonistic interactions on treatment efficacy was assessed using a Galleria mellonella model of disseminated candidiasis.Results. From 119 medications used to manage 40 conditions linked to a high risk of fungal infections, we identified 34 compounds that altered the effectiveness of the antifungals fluconazole (FLC) and/or anidulafungin. Most of these compounds reduced or antagonized antifungal efficacy, often due to increased resistance or tolerance. Validation in a G. mellonella infection model confirmed that compounds antagonistic to FLC, including loperamide, estradiol and levothyroxine, interfere with antifungal treatment efficacy in this in vivo model.Conclusion. Our findings highlight that medications frequently used by patients at risk for fungal infections can inadvertently increase fungal pathogen drug tolerance or resistance. We suggest that drugs targeting non-fungal conditions yet affecting fungal pathogens might represent an underestimated factor contributing to rising antifungal resistance and tolerance.

Keywords: Candida albicans; antifungal resistance; antifungal tolerance; drug interactions; fungal infections; treatment failure.

Fig. 1.. Workflow of the medical guideline search and experimental set-up. (a) To identify pathologies in fungal infection high-risk settings, 813 medical guidelines were systematically mined for their association with fungal infections. The search resulted in 63 guidelines which indicate a relation of 40 pathologies or medical interventions to fungal infections. A total of 119 compounds directed against non-fungal diseases which are commonly used in these identified pathologies and interventions were shortlisted. Pathology classes selected in the systematic keyword search and related shortlisted drugs are shown in the pie charts. (b) While targeted drug testing, the shortlisted drugs were systematically exposed to C. albicans SC5314 cultures in the presence and absence of the antifungals FLC and ANI. Hits, defined as compounds which increased or decreased the OD₆₀₀ after 60-h culturing, were further tested in checkerboard assays to identify synergistic and antagonistic DDIs, and DDAs to determine whether the interactions altered antifungal tolerance and/or resistance. Finally, FLC antagonists identified from checkerboard assays were co-administered with FLC treatment in G. mellonella larvae infected with C. albicans. Larval survival was tested every 24 h.

Fig. 2.. Targeted drug testing in C. albicans reveals negative and positive interactions of drugs directed against non-fungal diseases commonly administered in conditions with a risk of fungal infections. (a) Shortlisted drugs were exposed to C. albicans cultures, in the absence and presence of different concentrations of FLC (blue) and ANI (red). Fungal growth, quantified by measuring the OD₆₀₀ after 60 h culturing, of the tested shortlisted drugs is displayed as Log2 of the relative growth compared with an reference (REF) control, being antifungal+DMSO of the matched condition. Grey dashed lines represent the threshold defining a ‘hit’. Ten micromolar of the antifungals amphotericin B, FLC and ANI were included in the tested compound shortlisted as a positive control for fungal growth suppression. X-axis labels are not displayed here due to space constraints but are provided in Fig. S2. Compounds were tested in n=3 technical replicates. (b) Hits identified from targeted drug testing. Interactions with FLC (blue) and ANI (red) classified as ‘hits’ altered the fungal OD by more than 0.5× and 1.5× mean of OD₆₀₀ compared with the mean of the REF control (antifungal+DMSO).

Fig. 3.. Checkerboard assays identify synergistic and antagonistic drug interactions with FLC and ANI. (a) Experimental set-up of checkerboard assays, including a concentration range of the antifungal (FLC or ANI) on the y-axis and concentration range of the interacting drug on the x-axis of a 96-well edge plate. (b) Checkerboards with FLC (blue) and ANI (red) showing synergistic interactions, identified by a reduction of at least two steps in the typical downward staircase pattern characteristic of synergy. (c) Checkerboards of ten drugs that increased the biomass of C. albicans show the typical staircase pattern of antagonistic DDIs, with at least two steps upwards. (b) and (c) display one of the experimentally conducted two biological replicates. All performed checkerboards are shown in Fig. S3. All checkerboards were tested with two biological replicates.

Fig. 4.. Drugs directed against non-fungal diseases increase antifungal resistance and tolerance in C. albicans. (a) To perform DDAs, the interacting drug was equally spread over a Petri dish, and, subsequently, a sterile filter disc containing the tested antifungal (FLC or ANI) was placed in the centre of the dish. After 2.5-h diffusion time, the disc was removed and fungal cells were equally spread over the dish. (b) Software output from evaluation of DDAs. Resistance was evaluated by calculating the inflection point of the saturation curve measured from the center of the ZOI outward at 24 h (FLC) or 30 h (ANI), reflecting the radius of the ZOI (red line). After 72 h, the inhibition effect, determined by the saturation inside the ZOI compared to the saturation outside the ZOI, served as an estimation for tolerance. The blue lines display the interval of the radius of the ZOI, while the green line highlights the edge of the petri dish. (c) Resistance (24 h) and tolerance (72 h) assessed by DDAs for positive interactions with FLC. (d) DDAs evaluating ANI resistance at 30 h for drugs that showed positive interactions with ANI during targeted drug testing. (e) DDAs for selected drugs that caused negative interactions with FLC during targeted drug testing, showing resistance at 24 h and tolerance at 72 h. (f) Representative negative interactions with ANI, evaluated for ANI resistance. (c–f) DDAs for all drugs tested, including all replicates (n=3 biological replicates), are shown in Fig. S5. (g) Many drugs alter resistance and tolerance to FLC (normalized to FLC+DMSO control). (h) Drugs that negatively interacted with ANI while targeted drug testing mainly also showed increased ANI resistance at 30 h in DDAs. (g, h) In two cases, resistance was strongly reduced (including DDAs for the interactions of atorvastatin+FLC and cyclosporine A+ANI) and caused such a dramatic loss of antifungal activity that no ZOI was detectable. Consequently, these data are not shown here. Similarly, desogestrel dramatically increased FLC tolerance, so that growth inside the ZOI could not be distinguished from the outer part of the ZOI; thus, these data are also not shown here (see Fig. S5). (g) and (h) with the drugs presented in a consistent order are shown in Fig. S6. N=3 biological replicates of all tested drug combinations, except for DMSO+FLC (n=6). ‘_w’ indicates that the drug was dissolved in water rather than DMSO.

Fig. 5.. Impact of FLC antagonizers identified on G. mellonella survival during systemic fungal infection. (a) Experimental workflow of G. mellonella infection and drug treatment. Larvae of G. mellonella were injected with 5×10⁵ C. albicans cells. Two hours after infection, larvae were treated with either no drug (PBS/DMSO, Ca+Buffer, grey), FLC alone (Ca+FLC, blue), th antagonizing drug alone (Ca+Drug 2, black) or a combination of FLC+antagonizing drug (Ca+FLC+Drug 2, red). Additionally, control groups of larvae were injected with PBS/DMSO (Buffer control, light beige) and antagonizing drugs (Control Drug 2, dark beige) in the absence of fungal infection. After treatment, larvae were kept at 30 °C, and survival was monitored every 24 h. (b) Kaplan–Meier survival analysis of infected larvae showed significantly lower survival probabilities when infected larvae were treated with FLC and ethinyl estradiol, loperamide or levothyroxine (red) compared with FLC treatment alone (blue; Ca+FLC+estradiol vs Ca+FLC, z=4.770, P=2.76×10⁻⁵; Ca+FLC+levothyroxine vs Ca+FLC, z=3.854, P=0.001746; Ca+FLC+loperamide vs Ca+FLC, z=3.244, P=0.017681). Different treatments are represented by different colours. Survival was calculated as the living proportion of larvae from every treatment group. For each condition, n=15 replicates were included, while each experiment was conducted twice. Crosses indicate the presence of right-censored data (i.e. larvae that were not dead at the end of the experiment). Treatment marked with the same letter showed no significant difference when performing a Wilcoxon rank sum test followed by a Bonferroni correction to account for multiple comparisons. All statistical comparisons between each group are included in Table S3.