The FDA-approved anti-cancer drugs, streptozotocin and floxuridine, reduce the virulence of Staphylococcus aureus

In Staphylococcus aureus, an important Gram-positive human pathogen, the SaeRS two-component system is essential for the virulence and a good target for the development of anti-virulence drugs. In this study, we screened 12,200 small molecules for Sae inhibitors and identified two anti-cancer drugs, streptozotocin (STZ) and floxuridine (FU), as lead candidates for anti-virulence drug development against staphylococcal infections. As compared with STZ, FU was more efficient in repressing Sae-regulated promoters and protecting human neutrophils from S. aureus-mediated killing. FU inhibited S. aureus growth effectively whereas STZ did not. Intriguingly, RNA-seq analysis suggests that both compounds inhibit other virulence-regulatory systems such as Agr, ArlRS, and SarA more efficiently than they inhibit the Sae system. Both compounds induced prophages from S. aureus, indicating that they cause DNA damages. Surprisingly, a single administration of the drugs was sufficient to protect mice from staphylococcal intraperitoneal infection. Both compounds showed in vivo efficacy in a murine model of blood infection too. Finally, at the experimental dosage, neither compound showed any noticeable side effects on blood glucose level or blood cell counts. Based on these results, we concluded that STZ and FU are promising candidates for anti-virulence drug development against S. aureus infection.

Protection of neutrophils from S. aureus-mediated killing. Neutrophils are the most abundant white blood cells in human blood and critical for the defense against staphylococcal infection 24 . However, it is also well established that S. aureus can kill human neutrophils 25 . To understand the protective effect of STZ and FU on the host, we assessed whether the compounds could protect human neutrophils from killing by S. aureus. As with the Sae inhibition assay, FU protected human neutrophils from S. aureus-mediated killing much more efficiently than STZ did (IC 50 , 0.2 μM vs. 92.4 μM) (Fig. 3). Since STZ and FU showed similar in vivo efficacy (Fig. 1), these results might indicate that the neutrophil protection activity of a compound is not a good indicator for its in vivo efficacy. To examine this notion further, we measured IC 50 for doxorubicin, which showed the least in vivo efficacy among them (Fig. 1). Again, doxorubicin protected human neutrophil more efficiently than STZ did (IC 50 , 4.2 μM vs. 92.4 μM) (Fig. 3), showing that the neutrophil-protection activity of a compound does not correlate well with its in vivo efficacy in a murine model of intraperitoneal infection.
Bacterial growth inhibition by STZ and FU. STZ and FU are known to have not only anti-cancer activity but also antibacterial activity 26,27 . Therefore, it is possible that the excellent in vivo efficacy of the compounds is due to their antibacterial activity. To examine this possibility, we determined the MIC (minimum inhibitory concentration) of the compounds using S. aureus USA300 in two different growth media: Mueller Hinton broth and tryptic soy broth. In the condition employed, STZ showed almost negligible antibacterial activity whereas Repression of the SaeRS system by the anti-cancer agents. S. aureus USA300 carrying either pYJ-P1gfp or pCL-P hlamin -gfp was grown to exponential growth phase in TSB; then a varying concentration of the anticancer agents was added. At 3 h post-incubation, GFP expression was measured and normalized by OD 600 . Figure 3. Protection of human neutrophils by the anti-cancer drugs. S. aureus USA300 (10 6 CFU) and human neutrophils (10 5 cells) were mixed, and the test compounds were added to the concentration indicated for 4 h. The viability of human neutrophils was measured by CellTiter assay (Promega). In the graph, the OD 490 in the absence of compound was set to 100%. FU inhibited staphylococcal growth efficiently ( Table 1). As compared with STZ, even doxorubicin showed more than 10 times higher activity of growth inhibition. Therefore, it is possible that the in vivo efficacy of FU might be, in part, due to its growth inhibitory activity, whereas the in vivo efficacy of STZ cannot be explained by its antibacterial activity.
Transcriptional alteration in S. aureus by STZ and FU. To understand the molecular basis of the in vivo efficacy of STZ and FU, we analyzed genome-wide transcriptional changes caused by the compounds. S. aureus USA300 at exponential growth phase was treated with 1 μg/mL (~4 μM) of either compound at 37 °C for 3 h (Supplementary Fig. 2); then total RNA was purified and subjected to RNA-seq. Both STZ and FU caused significant transcriptional changes in S. aureus (STZ, 818 genes; FU 1180 genes, Fig. 4a, and Supplementary Table  1 -4). The majority of STZ-affected genes (76% up-regulated and 85% of down-regulated) were also similarly affected by FU, indicating that most anti-virulence mechanisms of STZ are shared by FU. At 4 μM, STZ is not expected to inhibit the SaeRS TCS (Fig. 2). Indeed, in the qRT-PCR analysis, STZ did not reduce the transcription of saeQ and saeS (Fig. 4b), although it did reduce the transcription of saeP. On the other hand, FU reduced the transcription of saeP and saeQ, but it increased the transcription of saeS (Fig. 4b). The molecular mechanism of the differential sae-repression is not clear. Intriguingly, both compounds appeared to inhibit the transcription of other virulence-regulatory systems such as Agr, ArlRS, and SarA more effectively than they inhibit the Sae system (Fig. 4b).
The repression also appears to be specific because the transcription of the control gene gyrB and most other staphylococcal two-component systems including the SrrAB TCS was not significantly affected (gyrB and srrB in Fig. 4b and Supplementary Tables 1-4). Taken altogether, these data suggest that the highly effective in vivo efficacy of STZ and FU could be due to their repression of multiple regulatory systems.

Prophage induction by STZ and FU.
Most clinical isolates of S. aureus carry prophages. Since induction of prophages kills the S. aureus host, it can be an effective antibacterial mechanism. The genome of the strain USA300 contains two prophages (ΦSa2usa and ΦSa3usa), of which ΦSa2usa is replication-defective 28 . The RNA-seq analysis showed that the treatment with the anti-cancer drugs increased the transcription of ΦSa3usa genes (Supplementary Tables S1 and S3), indicating induction of the prophage. To confirm the results, we grew S. aureus USA300 in TSB containing either STZ or FU (0.2 μg/mL) for 18 h and analyzed the culture supernatant by the soft-agar method. As shown, the culture supernatants produced a large number of plaques whereas the non-treated culture supernatant did not (Fig. 5), confirming the prophage induction by the compounds. These results also indicate that both compounds probably cause damages in the staphylococcal chromosome and that the prophage induction likely contributes to in vivo efficacy of the compounds.
Dose-dependent in vivo efficacy of STZ and FU. Next, we tested the dose-dependency of STZ and FU in their protective effect on the mice. A daily dose of 0.25 mg/kg (STZ) or 0.5 mg/kg (FU) was sufficient to show statistically significant protection against S. aureus infection (Fig. 6).

Effect of administration frequency on in vivo efficacy of STZ and FU. The biological half-life of
STZ is 35-40 min 29 . The half-life of FU is also rather short (t 1/2 β = 2 h in human) 30 . Although in the previous experiments, the compounds were administered every day, due to their short half-lives, it is likely that S. aureus cells were exposed to the compounds for a brief period. Therefore, we hypothesized that a brief exposure to the compounds is sufficient to give significant in vivo efficacy. To examine the hypothesis, at 1 h post-infection with S. aureus, we administered the compounds only once (1.25 mg/kg body weight for STZ and FU; 2.5 mg/kg body weight for doxorubicin) and watched the infected mice for 7 days. As shown in Fig. 7, the single administration of the compounds showed in vivo efficacy similar to that of daily administration.

In vivo efficacy of STZ and FU in a murine model of blood infection. In our drug screening and
subsequent animal studies, we used a murine model of intraperitoneal infection. However, the peritoneum is not a typical site of S. aureus infection. Therefore, STZ and FU were further tested in a murine model of blood infection, a common type of staphylococcal infections. S. aureus USA300 was administered into mice via retro-orbital injection; then, at 1 h post-infection, the compounds (2.5 mg/kg body weight) were injected into the mice via the intraperitoneal route. The drugs were administered once every day, and the mice were watched for 2 weeks. As shown in Fig. 8, both STZ and FU showed a statistically significant protective effect on murine survival whereas doxorubicin did not, confirming in vivo efficacy of STZ and FU in a clinically relevant animal model.

Effects of STZ and FU on murine blood glucose and bone marrow. At high doses, STZ is known to
induce diabetes in rodents by killing β cells in the pancreas 31,32 . Besides, a high dose of 5-fluorouracil (5-FU), a metabolite of FU, is known to cause bone-marrow depression 33 . Therefore, we examined whether the dosage used in our animal experiment can cause those side effects. As expected, the 5-FU-treated mice showed a lower level of blood cell counts along with a sign of hypoglycemia, whereas the mice treated with either STZ or FU showed no such changes (Fig. 9). Even at 100 mg/kg, 20 times higher dose than the highest dose we used, STZ did not cause any significant changes in either blood glucose level or blood cell counts. These results indicate that both STZ and FU can reduce staphylococcal virulence at a safe dosage.

Discussion
In this study, through a small molecule screening with a reporter for the SaeRS TCS, we identified two anti-cancer drugs, STZ and FU, as attractive lead candidates for novel anti-virulence drugs against staphylococcal infections. For lead candidates, these two compounds have the following promising properties: (1) simple molecular structure with low molecular weight (Fig. 1); (2) excellent in vivo efficacy even at low dosage and with single administration (Figs 1 and 6); (3) repression of multiple virulence regulators (Fig. 4); and (4) no overt adverse effect at the administration dosage (Fig. 9). In particular, with its excellent anti-growth and anti-virulence activities against    S. aureus, FU has a potential to be developed into a dual agent, which inhibits not only the growth but also the virulence of S. aureus.
Although three compounds identified in the screening are all anti-cancer drugs, not all anti-cancer drugs have anti-staphylococcal activity. In the 12,200 small compounds screened in this study, 89 anti-cancer drugs were included. However, only three drugs, STZ, FU, and doxorubicin, repressed the Sae system and showed in vivo efficacy against S. aureus infection (Fig. 1c), suggesting that the anti-cancer activity itself is not sufficient to inhibit the SaeRS system and to protect the host from S. aureus-mediated killing. Intriguingly, of the 85 compounds showing anti-Sae activity in the initial screening (Fig. 1a), only three compounds showed significant in vivo activity, demonstrating that the in vitro Sae-inhibition activity does not guarantee in vivo efficacy of the compound. Since the sae-deletion mutant completely lost its virulence at the bacterial dosage used in the experiment 16 , these results imply that a partial repression of the Sae system is not sufficient to protect the host from staphylococcal infections. The repression of multiple regulatory systems (Fig. 4) and phage induction (Fig. 5) likely contribute to the in vivo efficacy of STZ and FU. Also, other pharmacological characteristics including in vivo half-life, plasma protein binding, and toxicity will also affect the overall in vivo efficacy of a compound. Therefore, in a drug development process, it might be prudent to employ an in vivo efficacy assay at an early stage as possible.
Both STZ and FU are known to have antibacterial activity. In fact, STZ was initially identified as an antibiotic that inhibits the growth of both Gram-negative and Gram-positive bacteria including S. aureus 26 . In the original study, STZ was reported to inhibit S. aureus growth by 50% at 0.75 μg/mL 26 . FU was also reported to be a very potent inhibitor of staphylococcal growth (MIC, 0.025-0.00313 μM) 34 . However, in our study, STZ showed a negligible anti-growth effect on S. aureus USA300 (Table 1). Although much more potent than STZ, the anti-growth effect of FU was about 10 times lower than that reported previously (Table 1). These discrepancies might be due to the differences in the test strain and the assay conditions employed. Interestingly, despite its low growth-inhibition activity, STZ showed a robust efficacy in vivo similar to that of FU (Figs 1, 6-8), signifying the importance of the anti-virulence activity of the compounds. Since 76-85% of the transcriptional effect of STZ are also observed in S. aureus treated with FU (Fig. 4), it is likely that most of the anti-virulence effects of STZ are also shared by FU.
Quinolone antibiotics such as ciprofloxacin are known to induce prophages via SOS-response [35][36][37][38] . Indeed, the 2 h-treatment of S. aureus 8325 with ciprofloxacin induced 16 prophage genes 39 . Analysis of non-prophage genes affected by STZ, FU, or ciprofloxacin showed that only a total of 22 genes (11 up-regulated, and 11 down-regulated) were commonly regulated by the three compounds ( Supplementary Fig. 3 and Supplementary Table 5), suggesting that the effect of ciprofloxacin on S. aureus is largely different from that of STZ or FU. Of the 11 up-regulated genes, eight are involved in DNA repair, showing the induction of SOS-response is the primary effect shared by the three compounds. Intriguingly, along with alpha-hemolysin (hla), the virulence regulatory systems, Agr and SarA, were also commonly repressed by all three compounds. Although the molecular mechanism of the repression is not known, it is tempting to hypothesize that the induction of SOS-response gives negative impact on the virulence gene expression. Also, since STZ and FU appear to inhibit Agr more effectively than they inhibit the Sae system (Fig. 4b), it is possible that the repression of hla is through the repression of Agr, not the Sae system 40 .
As with cancer cells, bacteria can replicate indefinitely as long as a permissible growth condition is provided. Indeed, similarities between cancer cells and bacterial pathogens have been noted: high replication rates, damages to the host, spreading and dissemination within the host, and rapid development of resistance against therapeutic agents 41 . Even the immune response against bacterial pathogens has been used to treat certain types of tumor [42][43][44] . Although those similarities might be purely superficial without any biological connections, it is surprising to see that, of the 12,200 compounds screened, three compounds with in vivo efficacy are all anti-cancer agents. It is likely that the rapidly replicating nature of both types of cells makes them vulnerable to the anti-replication activity of the anti-cancer agents. Nonetheless, it remains to be seen whether there are other similarities between the pathogenic bacteria and cancer cells that can be targeted by the same therapeutic agent.

Methods
Ethics Statement. The human subject experiment (i.e., purification of human neutrophils) was approved by the Indiana University Institutional Review Board (Study number: 1010002390). Before taking blood, informed written consent was obtained from each human subject. All experiments were performed in accordance with relevant guidelines and regulations. The animal protocol was approved by the Committee on the Ethics of Animal Experiments of the Indiana University School of Medicine-Northwest (Protocol Number: NW-43). The animal experiment was performed by following the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Every effort was made to minimize the suffering of the animals.
Bacterial strains and growth conditions. The S. aureus strains and plasmids used in this study are described in Supplementary Tables 6. S. aureus was grown in tryptic soy broth (TSB). When necessary, antibiotics were added to the growth media at the following concentrations: erythromycin, 10 μg/mL; and chloramphenicol, 5 μg/mL.

Construction of a GFP-promoter fusion.
The reporter plasmids, pYJ-P1-g fp, pCL-P hlamin -g fp, pCL-P mecA -gfp, and pCL-P hrtB -gfp were generated by a ligation independent cloning method 45 . First, vectors were PCR-amplified from pYJ-gfp 16 with the primers P1969/P1747 or from pCL55-gfp 14 with the primers P1990/P1991 (for pCL-P hlamin -gfp), P3016/P3017 (for pCL-P mecA -gfp), or P3286/P3287 (for pCL-P hrtB -gfp) (Supplementary Table 7). The insert DNA fragment containing the promoter sequence was amplified with primer pairs P1971/ P1972 for P1, P1992/P1993 (for P hlamin ), P3014/P3015 (for P mecA ), or P3284/P3285 (for P hrtB ) (Supplementary Table 7). The PCR products were treated with T4 DNA polymerase in the presence of dCTP (vector) or dGTP (insert DNA) and mixed. DNA mixture was used to transform E. coli DH5α. Once verified, all plasmids were electroporated into S. aureus strain RN4220 and subsequently transduced into S. aureus strain USA300 with φ85. Animal test. An overnight culture of S. aureus was 100-fold diluted into fresh TSB and further incubated at 37 °C for 4 h. Cells were collected by centrifugation, washed with sterile PBS, and suspended in sterile PBS to OD 600 = 4 (i.e., 1 × 10 9 CFU mL −1 ). Sex-matched 8-week-old C57BL/6 mice were i.p. injected with S. aureus (2 × 10 8 CFU); then, at 1 h post-infection, a varying amount of compound was administered through i.p. injection. DMSO was used as a negative control. For daily dose experiment, the compound was i.p. injected every 24 h for seven days.

High
For retro-orbital studies for the compounds, the bacterial suspension (10 7 CFU in 100 μL) was administered into 10 sex-matched 8-week-old C57BL/6 mice via retro-orbital injection. At 1 h post-infection, test compounds (50 μg, 2.5 mg/kg) was administered by i.p. injection. The compounds were i.p. injected every 24 h for 14 days.
Statistical significance in murine survival was measured by Log-rank (Mantel-Cox) test with Prism 6 (GraphPad). RNA-seq analysis. An overnight culture of S. aureus in TSB was diluted 100 times in a fresh TSB (2 mL in a 15 mL test tube) and incubated in a shaking incubator at 37 °C for 2 h. Then 1 μg/mL of streptozotocin or floxuridine were added to the culture and further incubated for 3 h. After immediate stabilization of RNA in all samples by RNAprotect Bacteria Reagent (Qiagen), total RNA was isolated with RNeasy Mini Kit (Qiagen) according to the manufacturer's recommendations. For each condition, three RNA samples were isolated from three independent bacterial cultures. The isolated RNAs were sent to the Center for Genomics and Bioinformatics at Indiana University. Sequencing libraries were constructed using the ScriptSeq Complete Kit for Bacteria (Epicentre). The statistical analysis of the RNA-seq results was done with DeSeq. 2 as described before 47 . qRT-PCR analysis. The total RNA used for RNA-seq was also used for qRT-PCR analysis. cDNA was synthesized using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantification of transcripts was carried out by real-time PCR using SYBR Green PCR Master Mix (Applied Biosystems) in a QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems). Primers used to detect specific transcripts are shown in Supplementary Table 6. The relative amount of cDNA was determined using a standard curve obtained from PCR with serially diluted genomic DNA, and results were normalized to the levels of gyrB RNA. Statistical significance was measured by unpaired, two-tailed Student's t-test with Prism 6 (GraphPad).

Determination of MIC.
SCIeNTIfIC RepoRTS | (2018) 8:2521 | DOI:10.1038/s41598-018-20617-5 GFP assay. An overnight culture of S. aureus harboring pYJ-P1-gfp, pCL-Phla min -gfp, pCL-P mecA -gfp, or pCL-P hrtB -gfp was diluted 100 times in a fresh TSB and incubated in a shaking incubator at 37 °C for 2 h. Then various concentrations of streptozotocin and floxuridine were added to cells, and the cells were incubated for 3 h. The GFP signal (excitation 485 nm, emission 515 nm) and OD 600 were measured with Enspire Plate Reader (Perkin Elmer). All GFP expressions were normalized by OD 600 . The half maximal inhibitory concentration (IC 50 ) was calculated from at least three independent experiments via nonlinear regression analysis (sigmoidal dose-response with variable slope) using Prism 6 (GraphPad).

Neutrophil killing assay. Human neutrophils were purified from healthy adult blood donors by
Ficoll-Paque gradient method 48 . The purified neutrophils were added to a 96-well tissue culture plate (2 × 10 5 / well) in 100 μL RPMI (Gibco) supplemented with 10 mM HEPES and 10% human serum. S. aureus cells (1 × 10 6 CFU) and various concentrations of streptozotocin, floxuridine, and doxorubicin were added to each well. Then, 10 μL of CellTiter 96 ® AQ ueous One Solution (Promega) was added to each well and incubated at 37 °C in 5% CO 2 for 3-5 h. Neutrophil viability was assessed with the Enspire plate reader (Perkin Elmer).
Prophage induction by streptozotocin and floxuridine. S. aureus USA300 was grown in TSB containing 0.2 μg/mL of the compounds (STZ or FU) or no compound (a negative control) at 37 °C for 18 h; then the supernatant was collected by filtration (0.22 μm). The supernatant (100 μL) was mixed with the overnight culture (100 μL) of S. aureus RN4220 suspended in HIB (heart infusion broth) containing 5 mM of CaCl 2 . The mixture was incubated at 37 °C for 10 min, mixed with soft TSA (0.8%), spread on TSA, and incubated at 37 °C for 18 h.
Blood glucose and bone marrow suppression tests in mice. Streptozotocin