Transcriptome and network analyses in Saccharomyces cerevisiae reveal that amphotericin B and lactoferrin synergy disrupt metal homeostasis and stress response

Invasive fungal infections (IFIs) are difficult to treat. Few effective antifungal drugs are available and many have problems with toxicity, efficacy and drug-resistance. To overcome these challenges, existing therapies may be enhanced using more than one agent acting in synergy. Previously, we have found amphotericin B (AMB) and the iron chelator, lactoferrin (LF), were synergistic against Cryptococcus neoformans and Saccharomyces cerevisiae. This study investigates the mechanism of AMB+LF synergy, using RNA-seq and network analyses. Genes involved in iron homeostasis showed increased expression upon treatment with AMB alone. Unexpectedly, AMB+LF treatment did not lead to increased expression of iron or zinc homeostasis genes however we observed decreased expression of oxidative stress response genes. Addition of iron or zinc to AMB+LF treated cells did not rescue the synergy, supporting the likelihood that the mechanism of synergy involves more than iron and zinc chelation. We clustered genes based on patterns of co-expression and found by network analysis that many genes involved in iron and zinc homeostasis, which have dysregulated expression upon AMB+LF treatment, are targets of transcription factors Aft1p and Zap1p. Hypothesizing that these might play a key role in the synergistic response, knock-out mutants of Aft1 and Zap1 were tested for increased sensitivity to AMB and oxidative stress. Both mutants showed hypersensitivity towards these treatments. Our results suggest the mechanism of AMB+LF synergy involves disruption to oxidative stress response, in addition to chelation of iron and zinc. Since Zap1 is conserved in C. neoformans and contains a putative drug binding domain, we suggest novel Zap1 binding molecules could be combined with existing antifungals to serve as synergistic antifungal treatments for this species. Overall design: We extracted mRNA from S. cerevisiae treated with i) AMB only (0.03 µg/mL), ii) LF only (2 µg/mL), iii) a combination of AMB + LF (0.03 µg/mL and 2 µg/mL, respectively) or iv) corresponding matching untreated controls. ID20 (inhibition of cell growth by 20%) was chosen as the time point for RNA isolation from all treatments, as at this point cells are stressed but there should be few confounding effects of cell death, allowing a direct comparison among treatments. The ID20 for AMB of 60 minutes, and AMB-LF of 50 minutes. The same incubation times were applied to each of the matched control samples. RNA-seq data were generated using Illumina HiSeq 2000 and Illumina NextSeq 500 with biological triplicates multiplexed and randomized across two sequencing lanes. Using EdgeR, we identified differentially expressed genes by comparing AMB treatment with the matched control. We also compared AMB + LF treatment with its matched control. Please refer to the manuscript for more details regarding the experimental methods.