Clinical Laboratory Perspective on Streptococcus halichoeri, an Unusual Nonhemolytic, Lancefield Group B Streptococcus Causing Human Infections

Streptococcus halichoeri is a relatively newly identified species of pyogenic streptococci that causes zoonotic infection in humans. S. halichoeri was first described in 2004 as indigenous to seals, and only 8 reports of human S. halichoeri infection have been published. S. halichoeri grows as small, white, nonhemolytic colonies and may be strongly catalase-positive on routine blood agar media, which can lead to isolates being misidentified as coagulase-negative staphylococci. S. halichoeri tests positive for Lancefield group B antigen, like S. agalactiae, but can be identified with matrix-assisted laser desorption/ionization time of flight mass spectrometry or partial 16S rRNA sequencing. We describe 3 cases of S. halichoeri bone and joint infections in patients in the United States with underlying health conditions. In addition, we examine the microbiologic characteristics of S. halichoeri and discuss the importance of fully identifying this organism that might otherwise be disregarded as a skin commensal.

isolates available in the strain repository for further characterization with traditional biochemical tests including pyrrolidonyl arylamidase (PYR), hippurate hydrolysis, and bile-esculin growth/hydrolysis (4). We determined Lancefield type using the Hardy StrepPro Grouping kit (https://hardydiagnostics.com). This study was approved by the University of Utah Institutional Review Board (no. 24431).

MALDI-TOF Mass Spectrometry
We performed MALDI-TOF mass spectrometry as described elsewhere (6). Isolated colonies from Hardy Columbia sheep blood agar incubated at 35°C and 5% CO 2 were spread evenly on a polished steel target and overlaid with 1 µL of matrix (α-cyano-4hydroxycinnamic acid in 50% acetonitrile, 47.5 % water, and 2.5% trifluoroacetic acid) and air dried. We collected mass spectra as described elsewhere (6); each spectrum was a sum of 240 shots collected in increments of 40. We analyzed spectra using the Bruker Biotyper commercial database (https://www.bruker. com) and used scores of ≥1.9 for identification to the species level (6). Custom mass spectral profiles (main spectra, or MSPs) were created according to manufacturer's recommendations. In brief, we prepared extracts from 5-10 mg of cells using the standard formic acid-acetonitrile method; 10 replicate 1 µL aliquots were air dried on a polished steel target, then overlaid with 1 µL α-cyano-4-hydroxycinnamic acid and air dried. We collected spectra from each spot in triplicate and reviewed them in Bruker FlexAnalysis, then processed 24 spectra for MSP generation in Bruker CompassExplorer using default settings.

S. halichoeri Clinical Context
We describe 3 representative clinical cases of S. halichoeri bone and joint infection to highlight the clinical disease progression and microbiological characteristics of this unusual organism. The first case-patient was a 43-yearold man with poorly controlled type 2 diabetes mellitus, hypertension, and stage 3 chronic kidney disease who sought treatment for a 7-week history of low back pain radiating to his legs. Spinal magnetic resonance imaging (MRI) revealed diffuse edema of the fourth lumbar vertebral body with a possible fracture. Results from fluoroscopy-guided vertebral biopsy were unremarkable. He was discharged with plans for a followup MRI in 1 month but was readmitted 2 weeks later with worsening low back pain. Leukocyte count was within reference range; erythrocyte sedimentation rate (110 mm/h; reference <20 mm/h) and C-reactive protein level (CRP; 5.79 mg/dL; reference <0.5 mg/dL) were elevated. MRI results suggested L3 and L4 vertebral osteomyelitis and diskitis. Fluoroscopy-guided L4 vertebral biopsy had negative stains and cultures for bacteria, mycobacteria, and fungi. Additional biopsies from L4 bone, L3-4 disk, and both posterior paraspinal areas were negative by Gram stain, but cultures from all 4 yielded gram-positive cocci in clusters that grew as tiny, whitish-gray, nonhemolytic colonies on SBA. The isolate was initially reported as coagulasenegative Staphylococci on the basis of phenotypic testing but later identified as S. halichoeri by MALDI-TOF mass spectrometry (Bruker Biotyper 5627 database) at ARUP Laboratories. The isolate (229 in this study) was catalase-positive when grown in SBA, PYR-positive, weakly positive for Lancefield B antigen, and hippurate hydrolysis negative. Gram stain from SBA showed cocci, mostly in clusters with a few small chains, but revealed substantial chaining when grown in broth, and catalase testing from Mueller-Hinton agar was negative, consistent with the genus Streptococcus ( Table 1). The isolate was sensitive to all antimicrobials tested, including penicillin, vancomycin, levofloxacin, and linezolid. Because of a penicillin allergy, the patient was treated with renally adjusted vancomycin (1,500 mg/18 h), but he continued to have back pain, and new weakness developed in his legs after 5 weeks of therapy. Subsequent MRI revealed disease progression with diffuse marrow edema and end-plate disruption of the L3-4 vertebrae along with phlegmon formation and myositis of both psoas muscles, prompting addition of meropenem to his treatment. Needle biopsy of L4 vertebrae revealed reactive bone with degenerative changes and chronic inflammation, but routine bacterial cultures were negative. The patient's symptoms improved after 7 additional weeks of intravenous vancomycin and meropenem. This treatment was followed by 6 weeks of oral minocycline and cefuroxime, during which time his back pain was alleviated and inflammatory markers normalized.
The second case-patient was a 68-year-old man admitted for planned removal of his infected left knee arthroplasty and left knee fusion. The patient had a history of hypertension and giant cell tumor of the left knee 3 years earlier; at that time, left total knee arthroplasty, curettage, and polymethylmethacrylate packing of the left tibia were performed without complications. Three months before his admission for the removal, the patient noted purulent draining lesions around the surgical site for which he was prescribed amoxicillin/ clavulanate. Initial radiographs showed no evidence of tumor recurrence and stable-appearing changes of the left proximal tibia and left total knee arthroplasty. Chronic prosthetic joint infection was suspected, as well as left knee arthrofibrosis. Because of a rash attributed to amoxicillin/clavulanate, the patient was switched to trimethoprim/sulfamethoxazole, which was stopped ≈2 weeks before the planned second left knee arthroplasty. During the prosthetic joint removal procedure, knee fluid and resected tissue sent for culture grew 3 organisms: methicillin-susceptible Staphylococcus aureus and S. epidermidis, both identified by conventional phenotypic methods, and a third organism from thioglycolate broth that stained as gram-positive cocci in chains. This organism, which grew as white, nonhemolytic colonies on SBA and chocolate agar at 48 hours, was positive for catalase, PYR, leucine aminopeptidase, and bile-esculin and negative for growth in media with 6.5% salt. The organism (isolate 853 in this study) was sent to ARUP laboratories, where it was identified as S. halichoeri by 16S rRNA sequencing. Eight separate cultures of the knee tissue grew S. halichoeri, methicillin-susceptible S. aureus, and Finegoldia magna. AAST of S. halichoeri by broth microdilution revealed susceptibility to all drugs tested for viridians group streptococci. The patient received treatment for the polymicrobial infection with intravenous ceftriaxone and oral rifampin for 6 weeks.
The third case-patient was a 68-year-old man with a pertinent history of poorly controlled type 2 diabetes mellitus with peripheral neuropathy and suspected arterial insufficiency, seeking treatment for a severe diabetic foot infection and altered mental status. The patient fulfilled 3 of 4 criteria for systemic inflammatory response syndrome: fever of 101.2°F, tachycardia (112 bpm), and leukocytosis (leukocytes 18 k/μL). He had elevated CRP (77 mg/L),erythrocyte sediment rate within reference range (21 mm/h), and elevated blood glucose (248 mg/dL; reference fasting glucose <100 mg/dL). Aside from confusion, the patient had edema, erythema, and pain of the right second toe. No previous history of chronic wound or drainage in this area was noted and the wound was provisionally attributed to a nail grinder. Radiographs showed no signs of osteomyelitis. The patient was started on broad-spectrum intravenous therapy with vancomycin and ampicillin/sulbactam, and he showed marked improvement in the edema and erythema of the toe. Blood cultures collected at the time of examination were negative, however cultures of a right second-toe abscess were positive for gram-positive cocci in chains. The organism grew as small, white, nonhemolytic colonies that were weakly positive for slide coagulase but tube coagulase negative. Catalasepositivity was observed when the organism grew on SBA and chocolate agar but not when grown on tryptic soy agar. The organism was typed as Lancefield group B by latex agglutination and was initially identified by the VITEK 2 system (bioMérieux, https://www.biomerieux.com) as Streptococcus suis. Because of the discrepancy between the Lancefield typing and VITEK 2 results, we performed 16S rRNA sequencing, which identified the organism as S. halichoeri. Unfortunately, the isolate was not available for further testing in this study. The patient was transitioned to a 14-day course of oral trimethoprim/ sulfamethoxazole and amoxicillin/clavulanate, and by 4 months after debridement, the patient's wound had healed.
Of the 9 patients with positive blood cultures, 3 were also positive for the organism in urine, foot wound, or knee fluid cultures.

Microbiologic Characteristics of S. halichoeri Isolates
We performed a retrospective review of the microbiology results from the laboratory information system on 45 isolates of S. halichoeri. The laboratory performed Gram stains on all isolates, which were reported as gram-positive cocci in chains or clusters. All but 1 isolate failed to exhibit hemolysis on SBA plates. Most isolates (n = 29, 64%) were identified by MALDI-TOF mass spectrometry with the remainder (n = 18, 40%) by partial 16S rRNA gene sequencing. We selected 5 isolates available in our strain repository (including isolates from case-patients 1 and 2) for further characterization with biochemical and molecular methods ( Table 1). Four of these isolates were catalase-positive when grown on SBA, but all 5 isolates were catalase-negative when grown on Mueller-Hinton agar lacking blood supplementation. All 5 isolates were PYR-positive, bile-esculin and esculin positive, hippurate-negative, and weakly reactive with Lancefield group B antiserum. Antimicrobial susceptibility testing results were available for 22 clinical isolates ( Table 2). All isolates were susceptible to ceftriaxone, daptomycin, levofloxacin, linezolid, meropenem, penicillin, and vancomycin. Six isolates (28%) were resistant to clindamycin and nonsusceptible to erythromycin (5 resistant, 1 intermediate).

Mass Spectra of S. halichoeri Isolates
We generated MALDI-TOF MSPs from the 5 selected S. halichoeri isolates and compared them with Bruker Biotyper database entries from related streptococcal species. The MSP dendrogram ( Figure 2) shows that the S. halichoeri clinical isolates are readily distinguishable from other streptococci and cluster closely together with the S. halichoeri subsp. halichoeri-type strain. This suggests that the proposed subspecies are not currently distinguishable by MALDI-TOF mass spectrometry.

Discussion
The genus Streptococcus is composed of >130 species of gram-positive, catalase-negative bacteria that are found in a variety of environments. Several streptococcal species cause zoonotic infections in humans, including S. canis, Streptococcus equi subsp. zooepidemicus, S. iniae, and S. suis (10)(11)(12)(13). S. halichoeri is one of the more recently identified zoonotic streptococci, first described in 2004 after isolation from gray seals (H. grypus) (1). It was originally described as a gram-positive, catalase-negative, nonhemolytic organism, occurring in pairs or short chains, and expressing Lancefield group B antigen; this combination did not correspond to any previously identified Streptococcus species. 16S rRNA gene sequencing confirmed its placement in the genus Streptococcus, but the >3% divergence from described taxa suggested it was a novel species (1). S. halichoeri was also isolated from a European badger with pyogranulomatous pleuropneumonia (9) and from several canine and fur-producing animal species, including companion animals (14,15). These observations indicate a much broader host range than originally described for this organism.
The clinical laboratories at the Marshfield Clinic in Wisconsin have isolated this organism from both canine (n = 5) and feline (n = 1) clinical specimens (16). All isolates were nonhemolytic, had streptococcal morphology on Gram stain, were catalase and PYR-positive, and typed as Lancefield group B. Although initially reported as "Streptococcus species, unable to identify further," they were subsequently identified as S. halichoeri by MALDI-TOF mass spectrometry and 16S rRNA gene sequencing.
The first reported human infection with S. halichoeri was in a man with diabetes who sought treatment for an empyema; he reported handling fish before his illness (2). S. halichoeri was isolated from pleural fluid cultures as tiny, white, nonhemolytic colonies of grampositive cocci in chains that could not be identified by standard phenotypic methods or 16S rRNA sequencing. The organism was positive for Lancefield group B antigen and was ultimately identified as S. halichoeri by MALDI-TOF mass spectrometry. The isolate was sensitive to penicillin and levofloxacin, and the patient's condition was successfully treated with 4 weeks of intravenous ceftriaxone after 10 days of levofloxacin.
Another reported human case was of an 84-yearold man with a history of diabetes who sought treatment for bacterial cellulitis of the left thigh following prior treatment for left-sided endocarditis and cellulitis at the same site due to S. agalactiae (17). Blood and wound culture isolates were initially identified as S. pyogenes by VITEK 2 and S. agalactiae by API 20 STREP (bioMérieux). The discordance in identification necessitated analysis at a reference laboratory where the isolates were identified as S. halichoeri by MALDI-TOF mass spectrometry but as S. suis by VITEK 2. Ultimately, both were confirmed as S. halichoeri by sodA and 16S rRNA gene sequencing. The isolates were sensitive to penicillin, levofloxacin, and linezolid but resistant to erythromycin, clindamycin, and tetracycline. The patient's condition was successfully treated with 15 days of amoxicillin.
Finally, a 2016 publication with limited clinical information highlighted 6 additional isolates recov-ered from human specimens (3). Most (4/6) isolates were from blood and were shown to have homogenous phenotypes, sequence at multiple loci, and genomic similarities consistent with a difference at the subspecies level from the originally described S. halichoeri isolate, leading to proposal of the new taxon, S. halichoeri subsp. hominis. AST data on the 6 human isolates revealed susceptibility to most drugs tested, although 1 isolate was erythromycin resistant and 2 were tetracycline resistant.
It is unclear if S. halichoeri represents an emerging zoonosis, a rare but still underdiagnosed infection, or if it is simply being recognized only now because of improved identification methods (e.g., MALDI-TOF mass spectrometry). The source for infection with S. halichoeri is certainly not limited to marine life, because the organism has been isolated from both domestic and wild terrestrial mammals (14). Patients in the cases presented here had no known exposures to animals; therefore, reservoirs for infection by S. halichoeri might be more widespread than once thought.
S. halichoeri was originally described as catalase-negative, but our testing confirmed reports that it can be catalase-positive when grown on bloodsupplemented media. This observation is not unprecedented for Streptococcus-like organisms; some streptococci and enterococci can be weakly catalasepositive when grown on media containing blood (4). However, some S. halichoeri isolates are strongly catalase-positive under these conditions, leading to the risk of misidentification (Table 1). A 2020 study identified genes in some S. halichoeri isolates highly similar to known catalase genes, potentially  explaining these phenotypic observations (14). Whereas phenotypic methods may be misleading, MALDI-TOF mass spectrometry is a rapid and reliable method for identifying S. halichoeri ( Figure  2). As the MALDI-TOF mass spectrometry method becomes more widely available, we may begin to identify more infections cause by S. halichoeri and other underrecognized bacteria (6). In the absence of MALDI-TOF mass spectrometry, laboratories should consider further testing on atypical isolates from deep wound and sterile body sites when they are phenotypically identified as coagulase-negative staphylococci but morphologically consistent with streptococci. S. halichoeri subsp. hominis isolates were described as positive for bile-esculin, esculin, and acid from sucrose fermentation, but S. halichoeri subsp. halichoeri was negative for all 3 tests (3), suggesting possible phenotypic distinction of the subspecies. The 5 clinical isolates we characterized were positive for bile-esculin and esculin; 4 of them (isolates 018, 116, 229, and 583) were most closely related to S. halichoeri subsp. hominis by full-length 16S sequencing. Surprisingly, isolate 076 was positive for bile-esculin and esculin, like S. halichoeri subsp. hominis, but was nearly identical to the full-length 16S rRNA sequences from S. halichoeri subsp. halichoeri from both gray seals and badgers. This apparent discrepancy suggests that isolate 076 was a phenotypic variant of the S. halichoeri subsp. halichoeri. However, because of limited published data, the reliability of esculin or other phenotypic tests for distinguishing these proposed subspecies is unknown. Unfortunately, we do not have information on recent zoonotic exposures for isolate 076.
There are several parallels between the patients reported in our study, 2 of whom had diabetes, and the patients from the 2 previously published cases, both of whom also had diabetes (2,17). These similarities may indicate an opportunistic nature of this organism, in which establishing infection requires diabetes or an immunocompromised state. A similar association with diabetes mellitus has been observed among fishmongers who have contracted zoonotic infections through occupational exposures (13). Diabetes mellitus is also one of the most common coexisting health conditions associated with group B Streptococcus infection (18,19). One patient from this study and 1 from another study (2) had purulent infections, which may indicate pathogenic potential similar to the pyogenic streptococci, to which S. halichoeri is closely related (20).
All isolates in this study, as well as those in previous reports, were susceptible to β-lactams and levofloxacin, which may be considered antimicrobials of choice; daptomycin, linezolid, and vancomycin also showed very good antimicrobial activity. Our AST analysis showed that 5 (23%) of 22 isolates were resistant to erythromycin and 4 (18%) were nonsusceptible to doxycycline (when CLSI Figure 2. Matrix-assisted laser desorption/ionization time of flight mass spectrometry profile dendrogram of selected pyogenic and zoonotic streptococci used in study of human infections caused by unusual strains of Streptococcus halichoeri, United States. We compared main spectra from S. halichoeri isolates to Bruker Biotyper (https://www.bruker.com) database entries from related streptococcal species. Distance level indicates relative similarity of mass spectral profiles. T indicates type strains. breakpoints for S. pneumoniae were applied) (5), similar to observations made elsewhere (3). This analysis highlights the relative antimicrobial susceptibility of S. halichoeri isolates, but the exceptions point out the need to monitor the susceptibility patterns of this emerging pathogen. Given its phenotypic similarities with coagulase-negative staphylococci and the viridans group streptococci, it is likely that S. halichoeri continues to be disregarded by some as a skin commensal rather than a true cause of infection. In addition, our study highlights the importance of recognizing S. halichoeri infections and the role of MALDI-TOF mass spectrometry and 16S rRNA gene sequencing in accurately identifying this pathogen.