Susceptibility of Aerococcus urinae and Aerococcus sanguinicola to Standard Antibiotics and to Nitroxoline

ABSTRACT Aerococcus urinae and Aerococcus sanguinicola have been increasingly recognized as causative agents of urinary tract infection (UTI) during the last decade. Nitroxoline achieves high urinary concentrations after oral administration and is recommended in uncomplicated UTI in Germany, but its activity against Aerococcus spp. is unknown. The aim of this study was to assess the in vitro susceptibility of clinical Aerococcus species isolates to standard antibiotics and to nitroxoline. Between December 2016 and June 2018, 166 A. urinae and 18 A. sanguinicola isolates were recovered from urine specimens sent to the microbiology laboratory of the University Hospital of Cologne, Germany. Susceptibility to standard antimicrobials was analyzed by disk diffusion (DD) according to EUCAST methodology, nitroxoline was tested by DD and agar dilution. Susceptibility of Aerococcus spp. to benzylpenicillin, ampicillin, meropenem, rifampicin, nitrofurantoin, and vancomycin was 100% and resistance was documented only against ciprofloxacin (20 of 184; 10.9%). MICs of nitroxoline in A. urinae isolates were low (MIC50/90 1/2 mg/L) while significantly higher MICs were observed in A. sanguinicola (MIC50/90 64/128 mg/L). If the EUCAST nitroxoline breakpoint for E. coli and uncomplicated UTI was applied (16 mg/L), 97.6% of A. urinae isolates would be interpreted as susceptible while all A. sanguinicola isolates would be considered resistant. Nitroxoline demonstrated high activity against clinical A. urinae isolates, but low activity against A. sanguinicola. Nitroxoline is an approved antimicrobial for UTI and could be an alternative oral drug to treat A. urinae urinary tract infection, yet clinical studies are needed to demonstrate this potential in vivo. IMPORTANCE A. urinae and A. sanguinicola have been increasingly recognized as causative agents in urinary tract infections. Currently, there are few data available on the activity of different antibiotics against these species and no data on nitroxoline. We demonstrate that clinical isolates in Germany are highly susceptible to ampicillin, while resistance to ciprofloxacin was common (10.9%). Additionally, we show that nitroxoline is highly active against A. urinae, but not against A. sanguinicola, which based on the presented data, should be considered intrinsically resistant. The presented data will help to improve the therapy of urinary tract infections by Aerococcus species.

In 2017, EUCAST introduced clinical breakpoints for A. sanguinicola and A. urinae, which now encompass benzylpenicillin, ampicillin, meropenem, ciprofloxacin, levofloxacin, vancomycin, rifampicin, and nitrofurantoin. However, for UTI treatment several limitations are obvious: Except for nitrofurantoin, none of the antimicrobials with defined breakpoints is currently recommended for oral empirical treatment of lower UTI. Benzylpenicillin and ampicillin are only useful for targeted therapy, similarly for fluoroquinolones, vancomycin, rifampicin, or carbapenems, which are usually avoided because of side effects and collateral damage. Susceptibility data for Aerococcus spp. is still scarce and resistance to currently available oral antimicrobials (such as fluoroquinolones and nitrofurantoin) has been reported, highlighting the need for alternative options (9).
Nitroxoline is a quinoline derivate (8-hydroxy-quinoline) that is structurally unrelated to any other antibiotic and was first described as an antimicrobial in 1954 (10). It has been rediscovered for the treatment of uncomplicated UTI (uUTI) and is recommended as a first-line drug in the German uUTI guideline (11). The mode of action is based on ion chelation with subsequent effects on enzymatic pathways and charges of cellular compartments of fungal and bacterial cells (12). Recently, activity of nitroxoline against emerging pathogens such as multidrug resistant Gram-negative pathogens, including carbapenemase producers, enterococci, staphylococci, and even drug resistant fungi and mycobacteria was demonstrated (13)(14)(15)(16)(17)(18)(19). High urine concentrations can be achieved after oral administration (standard dose 250 mg every 8 h) yet systemic concentrations are low, limiting its use to infections of the urinary tract (20).
The aim of the present study was to compare the susceptibility of A. urinae and A. sanguinicola clinical isolates from Germany according to EUCAST and to assess the potential of nitroxoline for the treatment of Aerococcus spp.

DISCUSSION
Urinary tract infections are a frequent cause for prescription of antibiotics in inpatients and outpatients, yet treatment has become more and more difficult with increasing antimicrobial resistance, especially in Gram-negative bacilli. Additionally, limited oral options are available for emerging microbes such as Aerococcus spp. (21). Moreover, the use of highly active agents, such as fluoroquinolones, has been restricted because of potential side effects (22,23).
Therefore, old drugs such as nitroxoline, nitrofurantoin, and amdinocillin (mecillinam) have regained interest recently, since they show high activity against drug resistant pathogens and achieve high urinary concentrations after oral administration (11-20, 24, 25). Additionally, they are well tolerated, and ample experience is available with these drugs in different countries. Oskooi et al. (26) investigated aerococcal urinary tract infections and reported clinical success in all patients (100%) treated with amdinocillin, though microbiological success was only 33%. In contrast, the activity of nitroxoline against A. urinae has not been investigated previously.
Our study demonstrates the excellent in vitro activity of nitroxoline against A. urinae isolated from urinary specimens in a tertiary care center in Germany. If the current EUCAST breakpoint for E. coli (#16 mg/L) was applied, 97.6% of A. urinae isolates would be considered susceptible to nitroxoline. Since oral antimicrobial options are limited for the treatment of A. urinae, nitroxoline could therefore be a useful alternative agent.
In contrast, activity of nitroxoline against A. sanguinicola was poor, indicating intrinsic resistance of this species to nitroxoline, which was demonstrated in both disk diffusion Susceptibility of Aerococcus spp. to Nitroxoline Microbiology Spectrum and agar dilution. However, the small number of A. sanguinicola isolates limits the generalizability of this finding.
In another study, Scholtz et al. (27) demonstrated high susceptibility (97.8% to 100%) of Aerococcus spp. to benzylpenicillin, meropenem, linezolid, and vancomycin, but comparative lower susceptibility (56%) to levofloxacin on isolates from the United States. In this study, MICs were interpreted according to CLSI M45 breakpoints. In total, 134 Aerococcus isolates were tested using different commercially available susceptibility testing methods (gradient tests with different agars, Vitek2, BD Phoenix) and compared to broth microdilution as the CLSI-recommended reference method. The highest agreement with broth microdilution was observed for gradient tests on Mueller-Hinton agar with 5% sheep blood. In our study, an additional method was assessed (EUCAST disk diffusion), equally demonstrating high susceptibility of aerococci to standard antimicrobials when EUCAST breakpoints were applied, but a lower rate of resistance to fluoroquinolones (10.9%).
In the study of Roy et al. (28) on isolates from Canada, ciprofloxacin MICs $ 2 mg/L were recorded in 17.4% of isolates (I or R according to CLSI), which was more similar to the results from our study.
The present study investigated nitroxoline as an alternative treatment option with two different methodologies (disk diffusion and agar dilution). Overall correlation of inhibition zones and MICs was good, since isolates with high nitroxoline MICs showed small inhibition zones. However, only few A. urinae isolates with high nitroxoline MICs were available (4/166 isolates . 16 mg/L) and no A. sanguinicola isolate demonstrated low nitroxoline MICs, limiting interpretation of the correlation of the two methods. Also, more data on different MIC testing methods (e.g., agar dilution versus broth microdilution) for Aerococcus spp. are needed since superiority of agar dilution (as used in this study) was demonstrated for fluoroquinolones but not for other drugs previously (21, 28).
Nitroxoline resistance is still poorly understood and has rarely been described in clinical isolates. Molecular follow-up studies for comparison of genetic variations of A. urinae and A. sanguinicola should consider previously described mechanisms, such as efflux pumps (29).
Our study has several limitations. Not all isolates from the study period were available for analysis (e.g., were not stored initially or did not grow from frozen stock), which could limit the representativeness of our results. Furthermore, low MICs do not necessarily correlate with clinical or microbiological success in vivo. For nitroxoline, data on clinical outcome is still scarce and therapeutic failure has been described in geriatric patients for UTI caused by other organisms (30). Nevertheless, the activity displayed in the present study against A. urinae is promising, not only when applying the current EUCAST breakpoint (162/166 A. urinae #16 mg/L [97.6%]), but also when considering the pharmacokinetics of nitroxoline. High urinary concentrations after oral therapy have been demonstrated (5.4 mg/L of the unconjugated form and 210.6 mg/L of the conjugated form, yet the role of the conjugated form remains elusive) (20,30). The fact that nitroxoline is an approved drug may facilitate in vivo studies and compassionate use, especially for uUTI.
Another limitation of the present study is the lack of clinical data, as we could not discriminate between asymptomatic bacteriuria and UTI as underlying conditions. This could lead to an overestimation of the uropathogenic potential of the isolates in our collection. Since most isolates were from inpatients from a tertiary care center, additional risk factors may be overrepresented compared to uncomplicated UTI for which nitroxoline is approved. Additionally, our patient population might not be comparable to other populations, e.g., from primary care, which could have an impact on the overall susceptibility. Although a recent study has investigated virulence factors of Aerococcus species isolated from UTI (31), more data on the correlation of clinical characteristics and outcome of aerococcal infections is needed.
To conclude, A. urinae isolates are highly susceptible against nitroxoline, in contrast to A. sanguinicola isolates, which can be considered intrinsically resistant. Nitroxoline could be a useful alternative oral antibiotic for the treatment of uncomplicated UTI caused by A. urinae.

MATERIALS AND METHODS
Ethics approval. We declare that the study was conducted in accordance with guidelines outlined by the Declaration of Helsinki. For this study on bacterial isolates obtained as part of routine laboratory diagnostics, no patient consent or ethical approval was required according to the regulations of the University Hospital of Cologne.
Isolates. All isolates were grown from urine samples, which were sent for bacteriological analysis to the Institute of Medical Microbiology of the University Hospital of Cologne, a large tertiary care center in the western part of Germany. Isolates were characterized as part of routine diagnostics, based on national microbiology quality standards (32) and stored at 280°C in glycerol stocks. In the study period (December 2016 to June 2018), 30,718 urine samples from 17,570 patients were analyzed; of these, 312 grew A. urinae/sanguinicola. After removing duplicate isolates, 278 isolates remained. A total of 166 A. urinae and 18 A. sanguinicola nonduplicate isolates were available for our study, representing 66.2% of all Aerococcus spp. during the study period. Of these isolates, 82 were from voided midstream urine, 56 from catheterization, and 46 others. Most samples were from inpatients (121/184; 65.7% of isolates); of these, 93/184 (50,5%) were female and the median age was 77 years. For identification, MALDI-TOF MS (Microflex LT, Bruker Daltonik, Biotyper database 5.0, Bremen, Germany) was used. All isolates with a score $2.0 by Biotyper were included in the study.
Additionally, MICs were determined for nitroxoline using an agar dilution protocol for fastidious organisms, as previously described, and recommended by EUCAST (21, 28). Briefly, Mueller-Hinton (MH) agar was supplemented with defibrinated horse blood and other supplements according to EUCAST recommendations (33). Nitroxoline powder (Rosen Pharma, St.Ingbert, Germany) was dissolved in DMSO (Sigma-Aldrich, Darmstadt, Germany) and diluted into the agar plates in serial 2-fold dilutions to reach final concentrations of 0.06 mg/L to 128 mg/L. A suspension with an inoculum corresponding to 10 7 CFU/mL was prepared, and 1 mL was inoculated on the agar plates using an automated multipoint inoculator. Reading was carried out after incubation at 35°C for 24 h in a CO 2 enriched atmosphere. MICs were correlated with inhibition zone diameters (IZD) of disk diffusion testing.
Statistical analysis. Correlation of MICs and IZD was calculated by linear regression analysis using Excel (Microsoft, Redmond, WA, USA). Statistical analysis of categorical data were performed using Fisher's exact test (two-sided). A P value of ,0.05 was considered significant.
Data availability. The original contributions presented in the study are included in the article/supplemental material. The corresponding author may be contacted for additional information.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.6 MB.