Aerococcus urinae – A potent biofilm builder in endocarditis

The diagnosis of infective endocarditis (IE) remains a challenge. One of the rare bacterial species recently associated with biofilms and negative cultures in infective endocarditis is Aerococcus urinae. Whether the low number of reported cases might be due to lack of awareness and misidentification, mainly as streptococci, is currently being discussed. To verify the relevance and biofilm potential of Aerococcus in endocarditis, we used fluorescence in situ hybridization to visualize the microorganisms within the heart valve tissue. We designed and optimized a specific FISH probe (AURI) for in situ visualization and identification of A. urinae in sections of heart valves from two IE patients whose 16S rRNA gene sequencing had deteced A. urinae. Both patients had a history of urinary tract infections. FISH visualized impressive in vivo grown biofilms in IE, thus confirming the potential of A. urinae as a biofilm pathogen. In both cases, FISH/PCR was the only method to unequivocally identify A. urinae as the only causative pathogen for IE. The specific FISH assay for A. urinae is now available for further application in research and diagnostics. A. urinae should be considered in endocarditis patients with a history of urinary tract infections. These findings support the biofilm potential of A. urinae as a virulence factor and are meant to raise the awareness of this pathogen.


Introduction
Infective endocarditis (IE) is a life-threatening infection of the endocardium and is associated with high morbidity and mortality [1]. Culture-negative endocarditis, where the causative pathogens remain elusive in the routine microbiological diagnostic work-up, represents a serious risk for the patient. Any antibiotic therapy must remain empirical and unadjusted to the microbial species or resistance patterns, which is associated with increased mortality [2]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Culture-negative results in routine microbiology diagnostics may be caused by biofilm growth of bacteria, prior antibiotic treatment of the patient, or by fastidious or slow-growing bacteria [3]. In recent years molecular biological methods such as nucleic acid amplification techniques (NAT) and fluorescence in situ hybridization (FISH) have helped detect microorganisms that might otherwise be missed by routine culture methods [4]. The direct detection of microorganisms in valve specimens by FISH enables the simultaneous visualization and identification of the causative agent in situ, also in culture-negative cases [5]. FISH provides information not only about the identity of the pathogen(s) present but also about their state of activity and spatial organization as planktonic or biofilm microorganisms. One of the rare bacterial species recently associated with culture-negative IE is Aerococcus urinae (A urinae) [6][7][8][9][10].
A. urinae is a Gram-positive coccus that produces alpha-hemolytic colonies on blood agar and is catalase-negative, and positive for leucine aminopeptidase. A. urinae may be misreported as alpha-hemolytic streptococci because of their similar growth characteristics, biochemical characteristics, and hemolysis pattern [8,9]. Gram staining may further raise this suspicion, since streptococci appear as chains or pairs, whereas A. urinae typically appears in clusters or tetrads. Differentiation between these two genera is possible by hippurate hydrolysis testing or sequencing of the 16S rRNA gene [5,21,22]. Since 2013, MALDI-TOF (matrixassisted laser desorption ionization time-of-flight mass spectrometry) has also played a major role in the identification from clinical samples of A. urinae [23,24].
Concerning the treatment of A. urinae endocarditis, a combination of penicillin plus an aminoglycoside is recommended due to a described synergistic effect of the two drugs [21,25,26].
IE may be associated with microbial biofilms on the heart valves-microorganisms in sessile communities, which are more tolerant towards antibiotic treatment than their free-living, planktonic counterparts [27]. A. urinae isolated from human blood was shown in vitro to form biofilms and to trigger human platelet activation and aggregation. This may explain the potential virulence mechanisms of this pathogen, which may contribute to the ability of A. urinae to cause IE [6].
There are no direct data thus far on whether biofilm formation of A. urinae is relevant in vivo.
To further elucidate the biofilm formation of A. urinae in vivo we designed and optimized a species-specific FISH probe (AURI) to visualize and identify the pathogen in heart valves sections from two IE patients.

Specimen processing and fluorescence in situ hybridization (FISH)
Heart valve specimens from two patients were obtained during surgery at Charité-Universitätsmedizin and sent for FISH/PCR according to our standard routine diagnostics.
After incubation for 2 h in a dark humid chamber at 50˚C, the slides were rinsed with water, air-dried, and mounted with a mounting medium to prevent fading of the fluorescence signal (Vectashield, Vector Laboratories, California, USA).
The analysis was performed in accordance with the ethical guidelines of the 1964 Declaration of Helsinki and approved by the Ethics committee of the Charité-Universitätsmedizin Berlin (E82/018/18). Since the study was performed as part of the routine diagnostic work-up, the need for informed consent was waived as per the ethics committee approval.

Oligonucleotide probes
The probe EUB338 (5´-GCTGCCTCCCGTAGGAGT-3´) [29], which is complementary to a section of the 16S rRNA gene conserved in most microorganisms of the domain bacteria, was used to screen for bacterial colonization.
The probe NONEUB338, which is the antisense probe of EUB338, was used to exclude unspecific probe binding. In addition, a panel of FISH-probes for the detection of the species most prevalent in IE was applied as described [5]. To specifically detect A. urinae, we designed and optimized a 16S rRNA-directed FISH probe (AURI) (5´-AAGGCCATCGATAAGTGA-CAGCAA-3´). We compared the sequence of AURI with those of all 16S rRNA entries in the EMBL and GenBank databases to assess its specificity, and found 100% similarity only with the A. urinae 16S rRNA gene (November 2019).

Bacterial strains
An A. urinae blood culture isolate from a patient not related to the cases presented here served as a positive control. A. sanguinicola (DSM 15633), the nearest phylogenetic neighbor with three mismatches in the probe binding site, was used as a negative control to optimize specific FISH conditions for AURI. Bacterial strains for the positive and negative controls were fixed as described elsewhere [30] and were included in each FISH experiment to control the probe specificity and sensitivity. In addition, we screened different Gram-positive cocci frequently associated with IE, all of which contained more than four mismatches at the probe-binding site of AURI (

DNA processing and sequencing
DNA was extracted from sections of the embedded heart valve tissues and was then submitted to pan-bacterial amplification of part of the 16S rRNA gene using the primers TPU1 and RTU3, as previously described [28]. Amplicons were sequenced using a commercial sequencing facility (LGC Genomics, Berlin, Germany) and compared with all currently available sequences from the public databases (EMBL and GenBank) using BLAST and FastA of the sequence analysis program Husar 4.1 (Deutsches Krebsforschungszentrum, Heidelberg, Germany) and the commercial SmartGene program (SmartGene, Lausanne, Switzerland).

Patient description
We analyzed heart valve samples from two IE patients. The first patient, a 67-year-old man, was admitted to the Charité-Universitätsmedizin Berlin with the diagnosis of acute dyspnea and suspected urosepsis. The anamnesis revealed that the patient had undergone empiric ciprofloxacin antibiotic treatment in his nursing home two months before due to acute cystitis. During this treatment, a urinary catheter was placed.
On admission, a systolic murmur compatible with mitral insufficiency was noted. Because of purulence in the urinary catheter and suspected urosepsis, empiric meropenem and ciprofloxacin antibiotic therapy was initiated immediately. A urine culture revealed growth of Escherichia coli.
The day after admission, a multiplane transesophageal echocardiogram (TEE) showed severe mitral insufficiency and a vegetation measuring 26 mm x 10 mm on the posterior leaflet, and the antibiotic therapy was escalated to vancomycin and gentamicin in addition to meropenem and ciprofloxacin. Because of his clinical deterioration, the patient underwent immediate mitral valve replacement surgery.
Two sets of blood cultures taken on the day before the surgery showed Staphylococcus epidermidis after one day of incubation. A third set of blood cultures taken on the day of the surgery revealed growth of A. urinae after two days of incubation. On the day of the actual surgery, three additional blood culture sets were taken, all of which again exhibited growth of A. urinae. The antibiogram of A. urinae from the blood cultures showed susceptibility to penicillin with an MIC of 0.004 mg/L. Antibiotic therapy was tailored to ampicillin and gentamicin.
A conventional routine diagnostic microbiology of the heart valve performed one week post-operatively confirmed A. urinae with susceptibility to ampicillin and penicillin (MIC values of 0.032 mg/L and 0.004 mg/L, respectively). After two weeks, treatment was switched to monotherapy with ampicillin. The postoperative course was uneventful.
The second patient, an 86-year-old woman, was transferred to Charité-Universitätsmedizin Berlin for aortic valve replacement after TEE in a geriatric hospital had revealed left ventricular enlargement and severe aortic insufficiency with a vegetation measuring 20 mm x 10 mm on the non-coronary cusp. In addition, she had been diagnosed with a UTI due to Enterobacter cloacae. Calculated antibiotic therapy with ampicillin, gentamicin and ceftriaxone was initiated four days before the surgery and tailored to ampicillin and gentamicin one day before the surgery after four sets of blood cultures indicated alpha-hemolytic streptococci. Aortic valve specimens were sent for conventional routine diagnostic microbiology, which showed Grampositive cocci in the microscopic analysis but no growth in the culture.
Four days after the surgery, the bacteria from the preoperative blood cultures were identified as A. urinae and therapy was switched to penicillin and gentamicin for two days and was then further tailored to monotherapy with penicillin. The strain was susceptible to penicillin and cefotaxime (MIC values of 0.004 mg/L for both antibiotics) and resistant to gentamicin (MIC of 12 mg/L). The patient had an uneventful postoperative course.

Results
Both heart valves were embedded and analyzed by FISH using the pan-bacterial probe EUB338 in order to screen for bacterial colonization including yet uncultivated or fastidious bacterial species. We found extensive areas with bacterial colonization and formation of microorganisms resembling structured biofilms, which were in part FISH-positive. None of the specific FISH probes for the species most commonly associated with IE identified the EUB338-positive bacteria. Therefore, Staphylococcus spp., Streptococcus spp., Enterococcus spp., as well as Granulicatella, Tropheryma whipplei, and Bartonella were excluded as infectious agents.
DNA was isolated from consecutive sections of embedded tissue and submitted to pan-bacterial PCR amplification. Sequencing of the PCR products resulted for the first patient in an 847-base pair fragment with the highest homology (99.6% identity) to the A. urinae 16S rRNA gene (accession number U64458). Similarly, for the second patient a 465-base pair fragment was obtained with the highest homology (100%) to the A urinae 16S rRNA gene (accession number CP002512).
For direct in situ identification of the microorganisms, an AURI FISH probe specific for A. urinae was designed. The specificity and sensitivity of the probe were evaluated in silico against all sequences available in the databases. The in vitro specificity and sensitivity were evaluated using a clinical isolate strain of A. urinae (positive control) and the nearest phylogenetic neighbor A. sanguinicola (three mismatches at the probe binding site, negative control). In addition, no false positive signal with AURI was obtained from a panel of Gram-positive cocci frequently associated with IE. The stringency of probe binding was adjusted by raising the formamide concentration in the hybridization buffer. AURI showed specific and strong hybridization signals at a formamide concentration in the range of 10-40% (v/v), clearly discriminating A. urinae from all other species tested (Fig 1).

Discussion
The chances of survival of a patient in the ICU setting are 2.5 times higher when the correct antibiotic therapy is initiated as compared to empiric, broad-spectrum antibiotic therapy [2]. This makes species identification of the infectious agent highly desirable; however, this is not always successful. Reasons for culture-negative results may include biofilm growth, previous antibiotic treatment and/or rare or slow-growing bacterial species that are missed by routine culture methods.
We analyzed heart valve samples from two IE patients in whom routine culture methods did not immediately correctly identify the key causative pathogen.
In the first patient, culture revealed the presence of other bacteria such as S. epidermidis in the blood culture and E. coli in a urine sample. In addition, later on, a blood culture yielded A. urinae. In this case, PCR immediately identified A. urinae from the valve tissue and FISH confirmed A. urinae as the only causative agent (monospecies infection), revealing extensive structures consistent with microbial biofilms.
In the second case, E. cloacae was detected in the urine and the presence of alpha-hemolytic streptococci were suspected in blood culture. A blood culture later confirmed A. urinae. Again, FISH and PCR of the heart valve tissue immediately correctly identified the causative agent of the IE as A. urinae. Culture of the heart valve remained negative.
In both cases, FISH/PCR was the only method to unequivocally identify A. urinae as the only causative pathogen of IE, since all bacteria detected by the pan-bacterial probe were also positive with the specific AURI probe.
With the newly developed fluorescently labeled probe AURI specific for A. urinae described in this paper, we now have at our disposal a test system for further research and diagnostics for culture-negative heart valves that enables rapid and specific species identification and targeted antibiotic therapy.
Shannon et al. showed in vitro that blood culture isolates of A. urinae from IE patients form biofilms [6].
Here we show, to our knowledge for the first time, that impressive communities of A. urinae also exist in ex vivo material. A biofilm is determined by the arrangement of bacterial communities and the matrix in which they are embedded. While this matrix is composed of only bacterial products in vitro, in biofilm-associated infections it is also formed by the host's own substances such as fibrin, blood components, and tissue components [27]. Although FISH is unable to analyze the matrix in which the bacteria are located, it is still able to visualize the typical biofilm property of structural organization of the bacteria: the fluorescence signal intensity of FISH correlates with the bacterial ribosome content and therefore demonstrates activity of the bacteria [31,32]. The cases presented here display areas with a higher ribosome content urinae, as well as heart valve tissue were simultaneously hybridized with the probes EUB338 (FITC), AURI (Cy3) and NONEUB (Cy5). DAPI was used to stain nucleic acids. Each row shows an identical microscopic field with filter sets for FITC, Cy3, Cy5 and DAPI, respectively. AURI only detected A. urinae in the culture and the heart valve section, thus demonstrating specific hybridization. Scale bar 10 μm. Subsequently, the novel FISH probe AURI (Cy3-labeled) was applied to new sections of the heart valve samples simultaneously with the EUB338 (FITC-labeled); thus confirming A. urinae as the infectious agent of the IE in situ and demonstrating the biofilm potential of these bacteria (Figs 2 and 3).
https://doi.org/10.1371/journal.pone.0231827.g001 correlated to more active areas, which indicates a spatial organization of the bacteria in the socalled mushroom formations (Fig 2 and to a lesser extent Fig 3).
Therefore, these images show in vivo grown A. urinae biofilms in situ in IE and confirm A. urinae as the biofilm pathogen. This is a relevant finding, as bacteria grown as a biofilm exhibit a greater recalcitrance to antibiotic treatment than free-living, planktonic microorganisms https://doi.org/10.1371/journal.pone.0231827.g002 [27]. This might be particularly important for beta-lactam antibiotics, which act on cell wall synthesis and therefore are not as active on dormant cells which might rest in the less active parts of the biofilms. This hypothesis supports the data of Skov et al. (2001) who recommend the addition of gentamicin in IE based on time-kill curves in vitro [25]. Therefore, we think that diagnostic biofilm staging (Specific clinical guidelines for the antibiotic treatment of biofilm-associated infections, however, still need to be further developed [27]. planktonic cells, microcolonies, nascent or mature biofilms) is necessary to correlate therapeutic efficacy to the biofilm state and consequently to improve antibiotic algorithms.

Limitations
Unfortunately, this study is limited to a mere two cases; however, we hope to analyze more cases now that the FISH-probe is ready available. One limitation of FISH as a diagnostic tool is its application in the post-surgical setting only. Awareness of thorough documentation of the patient history including UTI must be raised, and respective sampling must be initiated in order to diagnose a possible Aerococcus involvement more timely.

Conclusions
In conclusion, A. urinae should be considered in IE patients with a history of UTI. Clearly, the biofilm potency of A. urinae as a virulence factor must be taken into account and awareness of A. urinae as a relevant pathogen in UTI and its association with IE must be raised.