In Vitro Antimicrobial Susceptibility Profiles of Gram-Positive Anaerobic Cocci Responsible for Human Invasive Infections

The aim of this multicentre study was to determine the in vitro susceptibility to anti-anaerobic antibiotics of Gram-positive anaerobic cocci (GPAC) isolates responsible for invasive infections in humans. A total of 133 GPAC isolates were collected in nine French hospitals from 2016 to 2020. All strains were identified to the species level (MALDI-TOF mass spectrometry, 16S rRNA sequencing). Minimum inhibitory concentrations (MICs) of amoxicillin, piperacillin, cefotaxime, imipenem, clindamycin, vancomycin, linezolid, moxifloxacin, rifampicin, and metronidazole were determined by the reference agar dilution method. Main erm-like genes were detected by PCR. The 133 GPAC isolates were identified as follows: 10 Anaerococcus spp., 49 Finegoldia magna, 33 Parvimonas micra, 30 Peptoniphilus spp., and 11 Peptostreptococcus anaerobius. All isolates were susceptible to imipenem, vancomycin (except 3 P. micra), linezolid and metronidazole. All isolates were susceptible to amoxicillin and piperacillin, except for P. anaerobius (54% and 45% susceptibility only, respectively). MICs of cefotaxime widely varied while activity of rifampicin, and moxifloxacin was also variable. Concerning clindamycin, 31 were categorized as resistant (22 erm(A) subclass erm(TR), 7 erm(B), 1 both genes and 1 negative for tested erm genes) with MICs from 8 to >32 mg/L. Although GPACs are usually susceptible to drugs commonly used for the treatment of anaerobic infections, antimicrobial susceptibility should be evaluated in vitro.


Introduction
Gram-positive anaerobic cocci (GPAC), which are a major part of the normal human microbiota of the oropharynx, respiratory tract, skin, and urogenital and gastrointestinal tract (GIT), are frequently recovered from human specimens, accounting for 25-30% of all anaerobic clinical isolates [1,2]. They are also opportunistic pathogens responsible for many human infections (often polymicrobial) especially in the elderly population and immunocompromised patients, including pleuropulmonary, intraabdominal, pelvic, skin, and soft-tissue and bone and joint infections (BJIs) as well as brain abscess and bacteremia [1][2][3][4].
GPAC has long been an heterogenous group of organisms and then has undergone considerable taxonomic changes with the creation of new genera formed from species previously belonging to the genus Peptostreptococcus, such as Finegoldia, Parvimonas, Anaerococcus, and Peptoniphilus while the only remaining representative of this genus is P. anaerobius [5]. Other changes within the GPAC group include the addition of new species in these new genera. The most commonly species found in clinical material are Finegoldia magna, Parvimonas micra, Peptoniphilus harei, and Peptostreptococcus anaerobius [2,6,7].
Only few studies have focused on in vitro antimicrobial susceptibility of GPAC isolates, many of them using unreliable conventional identification methods based on biochemical tests. In addition, the published data are often based on GPAC in general (formerly Peptostreptococcus spp.) even though there are significant differences in antibiotic susceptibility between GPAC species [2]. Indeed, it is now accepted to identify to the species level GPAC isolates in clinical specimens for susceptibility testing to adapt the correct antibacterial therapy, which is now possible with the development and application of molecular methods and MALDI-TOF mass spectrometry [2,6,8,9]. Even if it is widely accepted that these microorganisms remain consistently susceptible to antimicrobial agents generally used for the treatment of such anaerobic infections, antimicrobial resistance among anaerobes has increased in recent years worldwide and clinical failures have been reported in patients receiving inappropriate treatments [10,11].
The aim of the study was to investigate the in vitro antimicrobial susceptibility of a large collection of GPAC clinical isolates responsible for human invasive infections to 10 antimicrobial agents as well as to decipher the molecular basis of clindamycin acquired resistance.

Bacterial Isolates and Identification
A total of 133 non-redundant clinical isolates of GPAC collected from patients suffering from invasive infections in nine French hospitals between 2016 and 2020, including 109 (82%) from 2019, were studied. These isolates were recovered from patients with bacteremia (n = 37; 28%), BJIs (n = 51; 38%), and deep-seated soft-tissue infections (n = 45; 34%). For microbiological investigation, strains were grown on 5% horse blood agar plates (bioMérieux, Marcy-l'Etoile, France) incubated for at least 48 h in an anaerobic chamber at 35 • C. Phenotypic identification at the species level was performed using the MALDI-TOF mass spectrometry technology (Microflex; Bruker Daltonics, Wissembourg, France) in accordance with the manufacturer's instructions, and if necessary, by sequencing of the 16S rRNA gene, as previously described [12].

Antimicrobial Susceptibility Testing (AST)
MICs were determined using the reference agar dilution method according to 2018 CLSI guidelines using Brucella agar supplemented with 5% laked sheep blood, 5 mg/L hemin, and 1 mg/L vitamin K [13]. Plates were inoculated using a Steers replicator device (delivering a final inoculum of ca. 10 5 CFU per spot) and incubated anaerobically for 48 h at 35 • C. The 10 following antibiotics were tested: amoxicillin, piperacillin, cefotaxime, imipenem, clindamycin, vancomycin, linezolid, moxifloxacin, rifampicin, and metronidazole. MICs were interpreted according to 2020 CA-SFM/EUCAST clinical breakpoints recommended for anaerobes except for cefotaxime for which the PK-PD (non-species related) breakpoint was used (https://www.sfm-microbiologie.org/, accessed on 25 March 2021). Bacteroides thetaiotaomicron ATCC 29741 and Clostridioides difficile ATCC 700057 were used as quality control strains. The production of β-lactamase was assessed using the nitrocefin disk (Cefinase, BD BBL) as recommended by the manufacturer.

Detection of Resistance Genes
The bacterial genomic DNA of macrolide-resistant isolates was extracted using the QIAamp DNA Mini Kit (Qiagen, Courtaboeuf, France). Detection of erm(A) [including subclass erm(TR)], erm(B), erm(C), erm(F), erm(T) and erm(X) genes was performed by PCR as previously described [14]. All PCR-amplified products were sequenced in both directions by the Sanger method using the same primers.
As observed here, F. magna is the most common species of GPAC recovered from human clinical specimens, accounting for 5-12% of all anaerobic isolates and 20-38% of all GPAC isolates [1]. It is also likely the most pathogenic organism and the species most frequently isolated in pure culture [11]. Typical infections due to this species are soft-tissue abscesses, wound infections (incl. diabetic ulcers, pressure ulcers), and BJIs [5].
P. micra is part of the normal commensal microbiota of the GIT and the gingival crevice and it is mainly recognized as an oral pathogen especially isolated from polymicrobial infections such as periodontitis while it has been implicated in infections in other parts of the body [2,5]. In our study, it was mainly (82%) isolated from blood that is concordant with its frequent implication in GPAC bacteremia in a recent Swedish study, representing 42% (96/226) of GPAC episodes [7]. The authors also showed that GPAC bacteremia is much more common than previously reported and is a condition with significant 30-day mortality (11%) mainly in elderly patients with comorbidities [7].
As described here, Peptoniphilus spp. (mainly P. harei) are found in chronic wound samples (e.g., ulcer samples, diabetic wounds) and in osteoarticular samples [2]. With A. lactolyticus, A. vaginalis has been identified among the predominant species in grouped samples of diabetic foot ulcers and pressure ulcers [2]. P. anaerobius is recognized as part of the GIT and is one of the most common GPAC associated with infections of the abdominal cavity and female genitourinary tract [1,2].
Resistance to macrolides and lincosamides was described in Peptostreptococcus spp. three decades ago [36]. As observed here, clindamycin resistance rates among GPAC vary widely and this resistance is rising, especially in F. magna (3-51%) and Peptoniphilus spp.
Although levels of clindamycin resistance can be as high as 50%, little is known about the genetic basis of resistance to macrolides-lincosamides-streptogramins (MLS) in GPAC. Therefore, only two studies have investigated the molecular basis of MLS basis in these organisms. The first study reported an incidence of 80% macrolide resistance among 21 clinical isolates of Peptostreptococcus spp. due to erm(A) subclass erm(TR), suggesting that these anaerobic members of the normal oropharyngeal microbiota may be an important reservoir of this gene for transfer to pathogens such as Streptococcus pyogenes [39]. In the second study, we detected 25 F. magna isolates (out of 69; i.e., 36%) that exhibited high-level MICs of erythromycin (>256 mg/L) harboring either erm(A) subclass erm(TR) or erm(B) [40]. Altogether, our findings confirm that erm(A) subclass erm(TR) is the predominant MLS resistance gene among GPAC and that erm(B) could also be detected whereas other erm-like genes have never been identified up to now. It is noteworthy that we did not perform the detection of erm genes among clindamycin-susceptible isolates, which constitutes a limit of our study since erm genes can be harbored by strains categorized as susceptible.
Almost all GPAC isolates are susceptible to metronidazole (as in our study) even if some resistant strains have been described [15][16][17]19,20,23,25,26,[28][29][30][31][32]38,41]. Indeed, reduced susceptibility to metronidazole has so far only been rarely detected in surveillance studies among some P. micra, F. magna, and Peptostreptococcus spp. isolates [42]. Several molecular mechanisms have been associated with metronidazole resistance, mainly in Bacteroides fragilis [42]. It is mainly related to drug inactivation by nitroimidazole reductase encoded by nim genes [42]. To date, 11 nim genes (nimA to nimK) sharing between 54 and 90% amino acid identities have been described [40]. Of them, only nimB genes have been identified in the chromosome of some GPAC isolates (P. anaerobius, A. prevotii, P. micra) [42]. Note that the nimB gene was detected in 34% (21/61) of GPAC isolates with MIC of metronidazole ≥0.5 mg/L, including only two highly resistant F. magna isolates (MIC > 128 mg/L) while the other 19 strains remained susceptible [43]. Since metronidazole is often the drug used for empirical treatment of anaerobic infections and some resistant GPAC clinical isolates have been described, microbiologists should verify the activity of this antibiotic if necessary.
In our study, susceptibility to moxifloxacin greatly varied depending on the species (9-43%), which is concordant with observations reported in the literature (0-48%) [20,23,25,26,[30][31][32]37]. In agreement with these studies, we also observed that F. magna was more resistant than other species. Among newer fluoroquinolones, it was demonstrated that delafloxacin was more active than levofloxacin against GPAC (n = 20), with MIC 50 and MIC 90 values of 0.003 and 0.38 mg/L versus 2 and >32 mg/L, respec-tively [44]. In this later study, none of the GPAC isolates tested presented a MIC of delafloxacin equal or higher than 4 mg/L [44].

Conclusions
In our study, a remarkable difference in antibiotic susceptibility among the five most clinically important GPAC species (i.e., F. magna, P. micra, P. harei, A. vaginalis, and P. anaerobius) was found. These differences underline the importance to identify to the species level clinical isolates responsible for human infections. Compared to the processing time for aerobic bacteria, the cultivation and identification steps for anaerobic isolates usually take much longer. Clinical laboratories should provide any clinically relevant information in a precise and timely manner to the clinicians since the lack of detection of an infection-associated anaerobic isolate can often lead to inappropriate therapeutic choices and clinical failure [45]. This is particularly true in this period of emergence of antimicrobial resistance among clinically relevant anaerobes. However, in vitro susceptibility testing is not routinely performed because it is time consuming and has some technical issues. For individual patient management, it is important to remember that AST for anaerobes should be performed when: (1) the selection of an active agent is critical for disease management, (2) long-term therapy is being considered, (3) anaerobes are recovered from sterile body sites, (4) the infection persists despite adequate therapy with an appropriate therapeutic regimen, and/or (5) in cases of single recovered anaerobic pathogen from culture [10]. Besides the aforementioned points, susceptibility testing should also be performed whenever it is possible, especially for epidemiological surveillance purposes and to guide the choice of empirical antibiotic therapy [45]. This surveillance (locally, nationally, and internationally) needs to be continuously conducted, since recommendations on firstline agents of therapy are usually based on these data.  Institutional Review Board Statement: This study was conducted according to the guidelines of the Helsinki Declaration (Ethical Principles for Medical Research Involving Human Subjects) and was conducted in accordance with the guidelines of research boards at each study site.
Informed Consent Statement: All patient data were anonymously reported, with no possibility of connecting specimens and isolates to individual patients. Because this study only investigated clinical isolates by using in vitro experiments, this study did not need to be examined by an ethical committee and patients' informed consent was not required. Data Availability Statement: All data are available upon request to the corresponding author.