Auranofin is an effective agent against clinical isolates of Staphylococcus aureus

Abstract

Aim:

The orphan drug auranofin was recently found to exhibit antimicrobial properties.

Materials & methods:

We explored the efficacy of auranofin by evaluating the minimal inhibitory concentration against a collection of over 500 clinical isolates derived from multiple institutions, inclusive of drug resistant strains. Our evaluation also included continuous exposure of bacteria to auranofin.

Results & conclusion:

We found that minimal inhibitory concentrations ranged between 0.125 and 1 mg/l, exerting robust antimicrobial activity against a sizeable clinical collection of the bacteria. Further, we evaluated the propensity of the methicillin-resistant Staphylococcus aureus strain MW2 to develop resistance through extended exposure to auranofin. After 25 days, the bacteria remained susceptible. Our data suggest that resistance mechanisms do not currently exist to block auranofin antimicrobial activity.

Although great strides have been made in curbing bacterial infections through the development and employment of antibiotics, the continued emergence of drug resistance begs for expansion of the drug arsenal. The CDC notes that 2 million people a year in the USA experience bacterial infections from drug-resistant microbes, and at least 23,000 die [1]. Although progress has provided means of treating disease, it does not stand that the treatment methods are indefinitely efficacious.

Perhaps the most prominent drug-resistant strain of bacteria is methicillin-resistant Staphylococcus aureus (MRSA). In 2009, there were an estimated 633,338 hospitalizations as a result of S. aureus infections (17.68 per 1000 hospitalizations) – and among these an estimated 463,017 were MRSA-related hospitalizations (11.74 MRSA infections per 1000 hospitalizations) [2].

Our search, and that of others, for new antibacterial compounds identified auranofin for its antimicrobial effects [3–6]. Auranofin, 2,3,4,6-tetra-o-acetyl-1-thio-D-glucanpyranosato-S-(triethyl-phosphine), is an orphan US FDA-approved drug for the treatment of arthritis [7], found to inhibit S. aureus infections through both systemic and local delivery methods in murine infection models [4,5]. Thus far, it has been effective at inhibiting reference strains and a small collection of clinical isolates of S. aureus in vitro [3].

Due to the number of drug resistant strains that already exist within the hospital community, we endeavored to investigate S. aureus susceptibility to auranofin, examining isolates derived from a clinical setting. This is the first report, to the best of our knowledge, exploring the resistance potential against auranofin using a large number of clinically derived samples, particularly those that already express drug resistance. Our findings suggest drug resistant S. aureus isolates are indeed susceptible to auranofin.

Methods

Staphylococcus aureus clinical isolates

A total of 503 isolates were obtained from the Rhode Island Hospital Clinical Microbiology Laboratory between April 2015 and January 2017. Organisms were identified as S. aureus based on expected colony morphology and a positive Staphaurex^® (Remel, Inc., KS, USA) test. A subculture of the S. aureus isolate was provided to the research team on a blood agar plate. Single colonies were picked from plates and were then grown in tryptic soy broth (TSB) overnight at 37°C with agitation. The bacteria were stored as 25% glycerol stocks at -80°C until needed. Additionally, 14 vancomycin intermediate S. aureus (VISA) strains were procured from a panel made available by the CDC. The collection is part of the antibiotic/antimicrobial resistance bank.

A separate set of 50 clinical isolates were gathered and stored as glycerol stocks at -80°C that were included in the study. The S. aureus clinical strains were part of a collection obtained from nasal swabs. Colony cultures were grown overnight and the culture was used for minimal inhibitory concentration (MIC) assays described below.

Minimal inhibitory concentration

MICs were determined against the collection of S. aureus isolates for oxacillin and auranofin from 10 mg/ml stock solutions in DMSO, and a 10 mg/ml stock solution of vancomycin dissolved in water. The compounds were evaluated in 96-well plates utilizing a VIAFLO Assist (Integra Biosciences, NH, USA). S. aureus strains were grown overnight in TSB. Before the inoculum was added to the assay plate, it was adjusted to an initial optical density (OD₆₀₀₎ of 0.03 [5 × 10⁵ CFU/ml]; colony-forming units [CFU]) in Mueller–Hinton broth (MHB). To test the bacterial MICs for auranofin, 50 μl of MHB was added to columns 2–11. A total of 100 μl of MHB was added to column 12 (as a sterility control). Next, 50 μl of 64 mg/l of each of the drugs were dispensed into columns 1 and 2. Twofold serial dilutions were made for columns 2 through 10. Finally, 50 μl of inoculum was dispended to columns 1–11. The assay plate was then incubated at 35°C for 18 h and the absorbance was measured at 595 nm. Inhibition was gaged as no growth visually and confirmed using a plate reader. MICs were designated in wells that both showed no visual growth and provided a 8× or better reduced absorbance reading compared with the absorbance reading in a twofold lower drug concentration well.

Low dose drug exposure

Serial-passage resistance development was conducted following the methodology described in a previous study [8]. Briefly, an extended range of concentrations of auranofin was generated by twofold serial dilution from three different starting concentrations (10, 12 and 16 mg/l) covering 0.0781–16.0 mg/l, which extended over three rows of a 96-well plate. Arranged in triplicate, the serial gradients of auranofin wells were seeded in a 96-well plate. The same extended range of concentrations of ciprofloxacin was used as a control. Four 15 ml test tubes containing 5 ml of TSB were inoculated with a single colony of S. aureus MW2, respectively, and were incubated at 37°C with agitation at 200 r.p.m. Each overnight culture was adjusted to 1 × 10⁶ cells/ml, and 50 μl of the diluted cultures were dispensed into the 96-well plates containing 50 μl of the serial gradient of antibiotics. After incubation for 24 h at 35°C, bacterial growth was determined by measuring optical density at 600 nm on a SpectraMax plate reader (Molecular Devices, CA, USA), and the growth was defined as an optical density reading of ≥0.1. Bacterial cells growing at the highest concentration of the antimicrobial (just below the MIC) were diluted 1:1000 in MHB (Difco, MI, USA) and then used to inoculate a subsequent serial-passage plate. The remaining bacterial cells were stored at -80°C in MHB with 16% glycerol. This process was repeated for 25 days.

Enzymatic target of auranofin

We used a thioredoxin reductase colorimetric assay kit (Cayman Chemical Company, MI, USA) following the manufacturer’s instructions with small modifications. Bacterial TrxR (Thioredoxin Reductase from Escherichia coli, T7915, Sigma-Aldrich, MO, USA) was employed instead of mammalian TrxR provided in the kit. In brief, the kit assay buffer (1×) was mixed with 20 μl of bacterial TrxR (25 U/mg) and 20 μl of serial concentrations of auranofin (Santa Cruz Biotechnology, TX, USA). The reaction was initiated by adding 20 μl of NADPH and 5,5-dithio-bis-(2-nitrobenzoic acid) (DTNB). Positive and negative controls were also included in this assay. Instead of test compound, DMSO or sodium aurothiomalate (ATM) treatment was included separately. DMSO treatment did not inhibit TrxR action and was considered as a negative control, whereas ATM inhibited the TrxR action and was considered as a positive control. The experiment was repeated in triplicate. Absorbance was measured using a Spectramax M2 plate reader (Molecular Devices, CA, USA) at 410 nm after 5 min.

Cloning of TrxB & sequence analysis

Bacterial genomic DNA was isolated using genomic DNA isolation kit (Sigma-Aldrich) and the trxB gene sequence was amplified using the primers TrxB-F-ATGACTGAAATAGATTTTGATATAGCAATTATCGGTGC and TrxB-R-TTAAGCTTGATCGTTTAAATGTTCAATATATTCCGC. Cycling conditions were as follows: initial denaturation at 95°C for 5 min followed by 40 cycles of denaturation at 95°C for 1 min, annealing at 60°C for 1 min 30 s and extension at 72°C for 2 min, with a final extension at 72°C for 10 min. The amplified PCR product was purified according to manufacturer’s instructions (QIAquick^® PCR purification kit, Qiagen, Inc., Hilden, Germany) and cloned into pGEM T-Easy vector (Promega, WI, USA) and plated on LB agar with ampicillin (100 mg/l), X-gal and IPTG and incubated overnight. Plasmid DNA was isolated and sequenced at Genewiz sequencing facility (Genewiz, Inc., NJ, USA) to ensure the identity and specific nucleotide sequence of the trxB insert.

Results & discussion

Clinical isolate susceptibility

As reported previously, auranofin inhibits S. aureus growth inclusive of both laboratory reference strains and a few clinical isolates [3–6,9]. We set forth to expand on this knowledge and evaluate the clinical efficiency of auranofin, testing a larger collection of S. aureus isolates in an effort to determine if natural resistance mechanisms to this compound exist.

One source of our S. aureus isolates was the clinical microbiology laboratory at Rhode Island Hospital (between April 2015 and January 2017). We found that 40% of the clinical isolates were MRSA, compared with 60% Methicillin-susceptible S. aureus (MSSA) strains. Isolates were designated as MRSA based on an observed MIC to oxacillin >4 mg/l. Among all isolates, the MIC for auranofin ranged from 0.125 to 1 mg/l. The distribution chart shows the majority of the Rhode Island Hospital-derived clinical isolates exhibited an MIC of 0.5 mg/l (278/503 strains). Among the 303 MSSA isolates, 0.5 mg/l was the most frequently observed MIC (159 strains; Figure 1A & Supplementary Table 1). In the group of MRSA isolates, 0.5 mg/l was again the most frequent MIC observed (119/200 strains) (Figure 1A).

Figure 1. . Minimal inhibitory concentration distribution among Staphylococcus aureus strains.

(A) MICs were determined for the MSSA and MRSA strains within the tested group. (B) The range of the auranofin MIC was plotted for the S. aureus isolates derived from 23 facilities. (C) An additional collection of Vancomycin intermediate S. aureus strains was evaluated and the distribution and frequency of the auranofin MICs were evaluated among this separate collection. (D) For totality, a graph is presented that provides auranofin MIC for all of the 567 isolates tested.

MIC: Minimal inhibitory concentration; MRSA: Methicillin-resistant S. aureus; MSSA: Methicillin-susceptible S. aureus.

To add diversity, we tested a separate set of clinically derived isolates that were collected outside of Rhode Island Hospital. The collection was gathered from patient nasal swabs >65 years of age between 2009 and 2013 from 23 long-term care facilities located within 60 miles of Boston [10,11]. The selected collection was weighted toward isolates that exhibited some form or degree of drug resistance. Thus, many of the strains in the tested collection (42/50) are MRSA isolates. The auranofin MIC range within this group was 0.125–0.25 mg/l, within the range identified in the population from Rhode Island Hospital (Figure 1B & Supplementary Table 2). Testing of this group showed the vulnerability of the colonizing strains to auranofin. It is significant that this group was inhibited by auranofin since they are potentially the colonizing isolates that can infect a patient should an entry site and conditions be favorable to the bacteria.

A separate collection of vancomycin intermediate S. aureus strains, derived from a panel available through the CDC, was also evaluated for susceptibility to auranofin. Among the strains in the panel, MICs of auranofin ranged from 0.125 to 0.5 mg/l. The majority of the strains (11/14) exhibited an MIC of 0.5 mg/l (Figure 1C).

Testing of isolates in this work was performed using planktonic bacteria; however, previous literature has shown that auranofin is also effective against biofilms. Torres et al. observed a 1-log reduction in preformed biofilm upon the addition of auranofin where antibiofilm activity was observed at 11.7 μM [6]. The inhibition found against S. aureus biofilm by Torres et al. was higher than the concentration needed to inhibit planktonic cells and higher than the plasma concentration achieved via oral delivery [6]. Thus, while effective against planktonic cells, additional measures may be needed for inhibition of preformed biofilms.

The highest MICs among all of the evaluated clinical strains in the current study exhibited susceptibility at 1 mg/l and included three isolates (two from the MSSA group and one from the MRSA group among the clinical isolates collected at Rhode Island Hospital). Overall, the MIC range suggests that the S. aureus isolates are susceptible to auranofin and do not yet harbor endogenous resistance mechanisms to this compound. Importantly, even among drug-resistant strains to two different antibiotics, auranofin susceptibility is retained.

Extended exposure to drugs

When naturally occurring resistance was not identified among over 500 clinical isolates of S. aureus we imparted to see if we could develop a resistant strain through low dose exposure over an extended period of time. An MW2 drug resistant strain was used in the course of this assay since MRSA strains are noted for developing or acquiring additional drug resistant mechanisms more quickly than drug susceptible S. aureus strains [12].

S. aureus cells that were continuously exposed to auranofin over the course of 25 days did not develop resistance to auranofin based on measuring the MIC and growth of the cells (Figure 2). S. aureus cells did however develop resistance to the drug ciprofloxacin, a compound currently used as a broad spectrum antibiotic, during the same time period. Indeed, cells began to demonstrate growth in the presence of the drug and thus resistance was developed within 5 days of exposure to ciprofloxacin. Although auranofin resistance was not observed within the time frame of ciprofloxacin-developed resistance or in the test period, it is not to say that resistance could not develop over a more extensive dosing period.

Figure 2. . Low dose drug exposure to Staphylococcus aureus cells.

The minimal inhibitory concentration was tracked over 25-days during daily, renewed exposure of S. aureus cells to auranofin or ciprofloxacin.

MIC: Minimal inhibitory concentration.

Thus, auranofin exhibits further potential to be repositioned as the new antimicrobial. Auranofin is most commonly provided to patients at an oral dose of 3 mg twice daily for extended periods of time for the purpose of treating chronic rheumatoid arthritis. Doses of 6 mg given orally twice daily are also provided to patients with the most common adverse reaction being diarrhea, a condition that could be tolerated with the much shorter term of infection treatment compared with chronic care. A study by Sharma et al. evaluated blood and serum concentration of auranofin via detection of the gold component of the compound. Within 8 weeks, they detected a mean gold concentration of 0.63 mg/l in the blood and 0.43 mg/l in the serum [13].

Although the drug concentration detected within the blood is within a range that can be inhibitory to S. aureus, the concentration in the serum is slightly low. However, it should be noted that rheumatory arthritis patients dosing is not necessarily the same dose that would be used for patients with a bacterial infection. Indeed, treatment for sepsis or other S. aureus infection manifestations would ideally be treated via injection and likely at a higher dose.

Auranofin target

The identified susceptibility of bacteria to auranofin may, in part, be attributed to the focus on a new microbial drug target. It is the prevailing hypothesis that auranofin exhibits antimicrobial activity through release of monovalent gold, Au(I) that binds catalytic cysteine or selanocysteine residue in TrxR or thioredoxin-glutathione reductase enzymes [14]. Inhibition of TrxR leads to reactive oxygen species accumulation and apoptosis. We sought to confirm that the TrxR enzyme is inhibited by auranofin.

Since the MIC value of auranofin against S. aureus is 0.25 mg/l [3], and we observed a common range between 0.25 and 1 mg/l, we encompassed these concentrations, using a concentration range of 0.25–4 mg/l in our enzymatic assay to confirm the drug target. In the TrxR colorimetric assay, TrxR reduces DTNB into 5thio-2-nitrobenzoic acid, which produces a yellow color in the absence of an inhibitor. DMSO does not inhibit the enzyme reaction and is considered a negative control. ATM is a known inhibitor and is included in the assay kit as a positive control. Vancomycin was included as a control since it is readily known to inhibit the bacteria, although via alternate mechanisms.

The assay showed all tested auranofin concentrations inhibited the TrxR reaction as measured at 410 nm compared with the DMSO control or vancomycin (Figure 3). Further, auranofin was able to inhibit bacterial TrxR in a dose-dependent manner and was specific compared with another S. aureus antimicrobial agent. Our results suggest that auranofin does indeed diminish the enzymatic activity of thioredoxin (Trx).

Figure 3. . Dose response of auranofin on thioredoxin reductase system.

Bacterial TrxR was mixed with NADPH and 5,5-Dithio-bis-(2-nitrobenzoic acid). Auranofin inhibited the enzyme reaction in a dose-dependent manner in a cell free assay. DMSO was included as a negative control and aurothiomalate served as a positive control for the ability to inhibit the thioredoxin reductase enzymatic activity.

The thioredoxin (Trx) system is comprised of NADPH, TrxR and Trx. In some systems the Trx system is backed up by the glutathione-glutaredoxin system. However, for many Gram-positive bacteria, and for some Gram-negative bacteria, the systems are not redundant, leaving the Trx system to serve the essential task of defending against oxidative stresses through disulfide reductase activity [15]. In this context, auranofin holds the potential to inhibit the Trx system from adequately responding to reactive oxygen species.

The Trx system employs selenium in the C- terminal redox center for reduction of Trx and other substrates. There are a limited number of articles that depict auranofin binding to TrxR of parasites Entamoeba histolytica and Schistosoma mansoni though x-ray crystallography [14,16]. In the case of all of the parasitic inhibition, auranofin was found to disrupt the Trx-glutathione system where Trx glutathione reductase acts as a key enzyme [17–19]. The gold from auranofin appears to be the inhibiting portion of the molecule [16]. There appear to be other compounds that show activity against the Trx system that have not yet reached the level of clinical use, but have the potential for further development: allicin (derived from garlic) [20] and ebselen [21].

While there is evidence that auranofin targets TrxR, there is also emerging literature that suggests this may not be the sole target in microbes. Indeed, auranofin exhibits antimicrobial inhibition against some Gram-positive bacteria and some yeast [3,4,22–24]. For example, Thangamani et al. studies more global effects of auranofin to microbes and found that it inhibits multiple biosynthetic pathways resulting in disruptions to S. aureus cell wall, DNA and protein synthesis in bacteria [23]. Further, the authors suggested that the lack of inhibition to Gram-negative bacteria is not due to the redundancy provided by glutathione-glutaredoxin, but rather auranofin is excluded from Gram-negative bacteria by means of a more substantial cell membrane and bacteria may become sensitive to auranofin after membrane permeabilization with agent polymixin B nanopeptides [23].

There is also evidence to suggest that auranofin has an impact on mitochondrial respiration in fungi. Gamberi et al. identify Pos5 as a Saccharomyces cerevisiae auranofin target [22]. Pos5 is a kinase confined to the mitochondria capable of phosphorylating NADH and NADP+ [25]. Using chemogenomic profiling, Thangamani et al., identified mia40 as a potential auranofin target [24]. Mia40 is involved in the import and assembly of small intermembrane space proteins that enter the mitochondria from the cytoplasm.

Single nucleotide polymorphism

With evidence that auranofin does indeed affect TrxR in S. aureus, we endeavored to further interrogate the three clinical isolates with the highest MICs of 1 mg/l (BFSA46, BFSA52 and BFSA341) in order to determine if there was any genetic variance to TrxR that would yield higher MIC values. The reference strains MRSA-MW2 trxB gene sequence was used to compare to the clinical strains BFSA46, BFSA52 and BFSA341. In comparison to the reference strain MW2, nucleotide variance occurred at position 285 (T>C) and 301 (A>G) of clinical strain BFSA341, and at position 297 (T>C), 312 (G>T), 585 (C>A), 600 (T>C), 735 (G>A), 806 (T>C), 837 (T>C) and 891 (G>A) of clinical stain BFSA46. The polymorphisms in clinical strain BFSA52 was similar with clinical strain BFSA46, but with one additional change at position 524 (G>A). MW2 and BFSA341 were closely related clones based on the single nucleotide polymorphism and changes were observed in only two positions in the entire gene sequences. However, BFSA46 and BFSA52 were closely related clones based on the single nucleotide polymorphism except for one other position. Among the listed polymorphisms, a missense change in amino acid sequence was observed at position K111E in all three strains. R175H was observed in the clinical strain BFSA52 and V269A was observed in clinical strains BFSA46 as well as BFSA52 (Figure 4A).

Figure 4. . Polymorphisms in amino acid sequences.

(A) The trxB protein sequences of three clinical strains were translated from nucleotide sequences and compared with MRSA-MW2 strain for single nucleotide polymorphism and missense changes were observed in amino acid sequences. (B) Crystal structures of trxB protein from E. histolytica and S. aureus are shown in blue and brown colored ribbons with FAD and NADPH co-factors represented with sticks. The mutations identified in this study are marked with red color and are located away from the enzyme active site cysteines or the decoy site seen in E. entamoeba, where the Au(I) from auranofin putatively binds.

High-resolution crystal structures of Trx-glutathione reductase from Schistosoma mansoni and TrxR from Entamoeba histolytica in complex with auranofin-derived Au(I) are available [14] and served as a reference for our modeling of S. aureus. Similarly, crystal structure of TrxR from S. aureus has been solved by the Joint Center for Structural Genomics. The TrxR enzymes from E. histolytica and S. aureus share 39% sequence identity and are comprised of two domains that bind FAD and NADPH moieties as shown in Figure 4B. It is proposed that Au(I) binds catalytic cysteine pair of these enzymes and inhibits enzyme activity, although in both structures these cysteines were seen in disulfide bond configuration. It is proposed that auranofin inhibits parasites such as S. mansoni by means of the gold from auranofin attraction and subsequent transfer to selenocysteine in Trx-glutathione reductase of the parasite, as opposed to inhibition by the whole compound; thus, serving as a prodrug [16]. Of note, selenocysteine is also present in the S. aureus TrxR homolog. Using x-ray chrystallography, the authors show Cys-gold-Cys adducts in SmTGR incubated with auranofin [16]. The E. histolytica homologues enzyme is also shown to bind a second Au(I) from a ‘decoy’ Cys286 site of unknown significance [14], but this residue is not conserved in the S. aureus homolog.

Based on this modeling, it does not appear that the amino acid changes observed in the clinical isolates occur directly at known auranofin binding sites. Thus, the natural sequence variations that occur do not greatly affect auranofin susceptibility. The lack of significant alterations to TrxB may be influenced by the fact that others have found that TrxB is essential for bacterial growth in S. aureus [26]. Juhas et al. noted that essential genes are the most promising ‘drugable’ targets [27]. Within the tested time period, our results show that the drug resistance strain MW2 developed resistance to ciprofloxacin, but did not bring an advantage for developing resistance to auranofin. Perhaps the importance of the trx gene is why S. aureus did not exhibit growth during extended exposure to auranofin, suggesting that the resistance mechanisms were not developed in the provided time period.

Conclusion

In summary, the concept of repurposing or repositioning drugs is thought to be a valuable avenue in drug development. We find that a large collection of S. aureus clinical isolates is indeed susceptible to auranofin and S. aureus does not readily develop resistance through continuous exposure to the compound. Thus, providing encouragement for further clinical investigation due to the reluctance to develop auranofin resistance.

Future perspective

Auranofin presents an interesting and exciting new antimicrobial compound, primarily by the fact that it engages a new target. Since auranofin is already an FDA-approved compound, it holds the potential for re-positioning and following an abbreviated pathway to the clinic. In an effort to gage the usefulness of auranofin as a means to inhibit S. aureus, we tested the efficacy of the compound against an array of clinical isolates. Our findings suggest that there is not an existing resistance mechanism against auranofin, nor does one readily develop after exposure, evidence that can be useful prior to initiating a clinical study. We expect that auranofin and the TrxR target will be valuable in treating S. aureus infections, especially among drug resistant strains.

Summary points.

• Auranofin is able to inhibit an array of clinical isolates.

• The minimal inhibitory concentration is at or below 1 mg/ml.

• Auranofin is efficacious against even drug resistant S. aureus strains.

• Auranofin resistance did not develop after extended exposure.