Azoles

We evaluated the activity of the free and loaded extracts against C. albicans ATCC 10231 and SC5314 strains, and against C. glabrata ATCC 2001 strain. In addition the free extract was also tested against some C. albicans strains that have mechanisms of resistance that have been previously characterized in the Lopez-Ribot laboratory (Perea et al., 2001 (link); Li et al., 2004 (link); Vila et al., 2015 (link)). We also used several C. albicans clinical isolates with previously characterized molecular mechanisms of azole resistance, TW2, TW3 and TW17, which represent a series of matched susceptible and resistant isolates from the same patient with oropharyngeal candidiasis (White, 1997 (link)). In addition, we used a set of genetically engineered C. albicans “gain of function” strains in key transcriptional regulators of azole resistance including TAC1, MRR1 and UPC2 (a kind gift from David Rogers) (Schubert et al., 2011 (link); Flowers et al., 2012 (link)). In the case of C. glabrata, we also used a series of clinical isolates with different levels of resistance against azoles or echinocandins, which were provided by the Fungus Testing Laboratory (FTL) at The University of Texas Health Science Center at San Antonio. Antifungal susceptibility testing was performed using a broth microdilution technique following the methodology described in document M27-A3 published by the Clinical and Laboratory Standards Institute (CLSI) (Clinical and Laboratory Standards Institute, 2008 ) with minor modifications, with determination of MIC₉₀ as the concentration that inhibits 90% of the fungal yeast growth when compared to the growth control. MIC values were initially evaluated by visual analysis and confirmed by spectrophotometry at 490 nm in a microtiter plate reader.

Representative time-kill studies were performed to investigate the activity of C₁₂ and C₁₄ in the presence or absence of POS against one azole-resistant strain, C. albicans ATCC 64124 (B). These assays were performed in 15 mL culture tubes using RPMI 1640 medium as previously described25 (link). Different sets of cell suspensions were prepared with C₁₂ (8 μg/mL), C₁₄ (4 μg/mL), and POS alone (10 μg/mL), or combinations of C₁₂ (2 μg/mL) plus POS (2.5 μg/mL) or C₁₄ (2 μg/mL) plus POS (1.25 and 2.5 μg/mL), or growth control (no drug) and sterility control (no cells and no drug). The final inoculum size of yeast cells used was 10⁵ cfu/mL as confirmed by colony count. The cell suspensions were then incubated at 35 °C with constant shaking (200 rpm). Aliquot of 100 μL from each tubes were removed at 0, 3, 6, 9, 12, and 24 h, and serially diluted in sterile ddH₂O. 50 μL of each dilution was plated onto potato dextrose agar (PDA) and then incubated at 35 °C. Colony counts were determined after 48 h of incubation. The experiments were performed in duplicate.

A total of 105 participants were enrolled (55 patients with NVAF and 50 controls). Eligible participants comprised individuals with a history of AF documented by electrocardiogram within the 12 months before collection, and for whom chronic oral anticoagulation was indicated (CHA₂DS₂-VASC ≥ 2). The patients were enrolled from the outpatient clinics of the hospitals of Lifecenter, Semper, and Ipsemg (Belo Horizonte, Minas Gerais, Brazil) during the period from October 2013 to January 2017. A group of controls composed of individuals in sinus rhythm with no previous diagnosis of AF or use of any anticoagulant therapy was recruited from the local community.
Participants were excluded if they had used any antiplatelet agent, non-steroidal anti-inflammatory drugs, heparin, hormone replacement therapy, antifibrinolytics, amiodarone, verapamil, quinidine, azole antifungals, or ritonavir in the 4 weeks prior to the study. Subjects were also excluded if they presented a current diagnosis of an alcohol use disorder, chronic kidney disease (creatinine clearance < 30 mL/min), severe dyslipidemia, bleeding disorders, liver and thyroid diseases, infectious, inflammatory, autoimmune, or malignant diseases, pregnancy, puerperium, and breastfeeding.

To verify the use of the new workflow for read-across, we used a dataset of 326 azole compounds with experimental values on human aromatase breast cancer cell line (MCF-7aro, cell-based assay). The starting dataset was initially collected from the Tox21 library considering only Tox21_Aromatase_Inhibition (activity test). This contained 20,992 compounds encoded as SMILES, name, and CAS number [29 ]. The assay was performed using aromatase breast cancer cell line (MCF-7 aro) (cell-based assay) and the concentrations of testosterone (an androgen and estradiol (an estrogen)) were measured before and after exposure to azole compounds tested. The qualitative outcome was recorded as an active agonist, active antagonist, and inactive, where quantitative agonist and antagonist activities were expressed in nanomolar (nM) units represented by AC50 in the original database [29 ]. Once the data was collected, it was subjected to a rigorous data curation procedure. The first step involved the retrieval of SMILES following the workflow developed by Gadaleta et al., 2018 [30 (link)]. The maximum purity was labeled “A” and only compounds with this label were considered. The detection of inorganic compounds, organometallic compounds, mixtures, neutralization of salts, tautomeric forms, and chemotype normalization was performed using the KNIME platform [31 (link)]. The compounds with inconclusive assay outcomes were discarded and duplicate structures were classified into two cases as follows: (i) activity range lower or equal to 1:3, and (ii) activity range higher than 1:3. In the first case, the mean of the activity was calculated, and in the second case, the structures were rejected. 3459 compounds were kept from the original dataset which had the purity “A” label. Furthermore, 67 compounds with ambiguous values, 10 compounds with trace element or inorganic compounds, 3 mixtures, 6 duplicates, and 6 ionic liquid compounds were removed. After this, the dataset was subjected to a manual inspection process and 119 compounds were found to have incorrect structures, and therefore removed. At this point, the dataset contained 3248 compounds and was filtered to extract azoles only. The total number of azoles was 351, from them 25 were tetrazoles and were discarded due to their poor representation. The quantitative outcome in nanomolar (nM) units was converted to molar (mole/liter) using the formulae (−logAC50 + 9). The qualitative activity values, active agonist and active antagonist, were recorded as “active”. The distribution of compounds in the final dataset of 326 azoles, considering the numbers of nitrogen in the azole ring and their qualitative activity value was:

82 monoazoles compounds of which 61 were inactive and 21 active.

198 diazoles of which 148 were inactive and 50 active.

46 triazoles which contained 26 inactive and 20 active.

More details regarding the data collection and data curation process are available in Caballero et al. [9 ,11 ].