Non-haemolytic and non-pigmented group b streptococcus, an infrequent cause of early onset neonatal sepsis

Streptococcus agalactiae (Group B streptococcus, GBS) is the most common bacterium causing early-onset sepsis (EOS) in neonates (Verani, McGee and Schrag 2010; Simonsen et al. 2014; Verani et al. 2014). Detection of GBS vaginorectal colonization in pregnant women and administration of intrapartum antibiotic prophylaxis has achieved a significant decrease in the incidence of EOS (Verani, McGee and Schrag 2010; Verani et al. 2014). Selective broth enrichment culture performed at 35–37 weeks of gestation is the currently recommended standard for prenatal GBS screening (Verani, McGee and Schrag 2010). For detection of GBS colonization, in vaginorectal samples, selective broth is subcultured onto either blood agar, granada-based media or chromogenic media (depending of each microbiology laboratory) and inspected for ß-haemolytic or pigmented colonies (Verani, McGee and Schrag 2010). Nevertheless, β-haemolysis is difficult to observe in some GBS strains; and haemolysis may be so weak that it is only apparent after removal of colonies from the agar surface (Edwards and Nizet 2011).

Granada media exploits the unique ability of GBS to synthesize granadaene, a specific orange-red non-isoprenoid polyenic pigment (Rosa-Fraile et al. 2006), reducing detection and identification of GBS to one single step. Expression of the pigment is invariably linked in GBS to the expression of the ß-haemolysin (Rosa-Fraile, Dramsi and Spellerberg 2014). Moreover, in GBS, pigment and β-haemolysin have been reported as identical molecules (Whidbey et al. 2013) that are encoded by the cyl gene cluster (Spellerberg et al. 1999, 2000). In accordance with these findings, non-haemolytic (NH) as well as non-pigmented (NP) GBS strains have been shown to harbour mutations in the cyl genes. GBS β-haemolysin is a broad-spectrum cytolysin lysing not only erythrocytes, but also destroying many eukaryotic cells. It is therefore referred to as β-haemolysin/cytolysin (βH/C) (Doran et al. 2002) and considered as a key virulence factor for GBS. In vitro studies as well as animal models have demonstrated NH GBS strains to be less virulent than haemolytic strains (Whidbey et al. 2013; Randis et al. 2014). It has also been demonstrated that the GBS haemolysin/pigment is a lipid toxin (ornithine rhamnopolyene) able to cause cell death and promote foetal injury (Whidbey et al. 2013; Whidbey et al. 2015).

Nevertheless, around 5% of human-colonizing GBS isolates are reported NH and NP (Brimil et al. 2006; Nickmans et al. 2012; Verhoeven et al. 2014) complicating the reliable detection of GBS. While alternative chromogenic media not requiring pigment production, or PCR methods based on the detection of the CAMP-factor gene are positive for NH GBS strains, none of these methods have demonstrated a 100% sensitivity and specificity (El Aila et al. 2011; Joubrel et al. 2014). Considerable effort is thus recommended to detect NH–NP GBS isolates (Verani, McGee and Schrag 2010; Nickmans et al. 2012); however, if these strains are relevant for the development of EOS is currently unknown, since studies on the rates of NH GBS strains among EOS cases have to our knowledge not been published.

In line with these findings, the isolation of NH–NP GBS strains from blood cultures occurs rarely (Sigge et al. 2008). It may be explained by the reduced virulence of these strains, since βH/C promotes the invasion of host cell barriers and it is considered important for the development of EOS (Whidbey et al. 2013, 2015; Randis et al. 2014)

If NH–NP GBS strains are less virulent, they should cause EOS infrequently. We therefore evaluated the incidence of NH–NP strains among 199 GBS strains isolated from blood cultures of newborns (first week of life) with EOS.

Strains were collected within the framework of the DEVANI study (DEVANI project 2010; Rodriguez-Granger et al. 2012; DEVANI final report 2013), and originated from seven European countries (Belgium, Bulgaria, Czech Republic, Denmark, Italy, Germany, Spain and United Kingdom). These strains represented the totality of GBS strains recovered, during the DEVANI study, from blood cultures of sick newborns with GBS EOS, and were kindly provided by the Institute Superiore de Sanita, Rome. In our laboratory, all these GBS strains causing EOS were systematically characterized at the species level as GBS using latex agglutination (Oxoid) and MALDI-TOF mass spectrometry (Bruker Daltonics, GmbH, Leipzig, Germany). ß-haemolysis was evaluated on blood agar and pigment production by granada agar and granada biphasic broth (bioMerieux). Media were incubated at 36ºC, blood agar and granada broth aerobically, and granada agar anaerobically, and inspected for ß-haemolysis and orange-red colonies after 48 h. Overall, among the 199 GBS strains causing EOS we found only two (1%) NH–NP strains. One strain was from Denmark and the other from Germany, and in both cases, the newborns recover uneventfully. Both NH strains were serotype III, sequence type ST 19 and give the expected reaction (pink-red colonies) in bioMerieux chromID Strepto B plates.

As expected and previously reported (Rosa-Fraile, Dramsi and Spellerberg 2014) there was a perfect concordance between haemolysis and pigment production, the two NH strains did not produce pigmented colonies. Analysis of MALDI-TOF profiles did not reveal any differences between haemolytic and NH strains.

To further investigate the reason for loss of haemolysin and pigment production in these two strains, a PCR covering the genes of the cyl cluster was performed (Sigge et al. 2008). PCR and subsequent nucleotide sequencing revealed an insertion of IS1381 into the haemolysin transporter gene cylA (Gottschalk et al. 2006). In both strains, insertion occurred after nucleotide 467 of cylA, a site that has previously been described as an IS1381 integration site in GBS (Sigge et al. 2008).

The rate of NH–NP GBS strains in the EOS cases of our study (1%) appears to be lower than the rates published for colonizing isolates, which are around 3–6% (Brimil et al.2006; Nickmans et al. 2012; Verhoeven et al. 2014). Our results therefore appear consistent with the assumption that the ß-haemolysin toxin is important for the development of invasive GBS disease. Though these numbers are still too small for solid statistical analysis, they may prompt further investigations into this matter. Current guidelines (Verani, McGee and Schrag 2010) for prevention of EOS recommend that vaginorectal samples in which pigmented colonies of GBS are not detected in granada media should be retested using a different approach as for example alternative chromogenic media, NAATs or latex agglutination to pick up NH–NP strains. Nevertheless, the low frequency of NH–NP GBS strains causing EOS found in this study suggests that the importance of NH strains for the development of EOS could be overestimated.

This research was partially presented in 2014 at the XIX International Lancefield Symposium, Buenos Aires, Argentina, abstract 039.

Collection of GBS strains was supported by the European Commission Seventh Framework (DEVANI Project grant agreement number 200481). The contributing members of the DEVANI (Design of a Vaccine Against Neonatal Infections) Study Group were John TELFORD (Project Coordinator). Novartis Vaccine and Development, Siena. Italy; Graziella Orefici (Scientific Coordinator). Istituto Superiore di Sanità, Rome. Italy; Lucilla Baldassarri. Istituto Superiore di Sanità, Rome. Italy; Androulla Efstratieu. Health Protection Agency. Colindale, UK; Pierrette Melin. Centre Hospitalier Universitaire Liege. Belgium; Manuel de la Rosa-Fraile. Hospital Universitario Virgen de las Nieves, Granada. Spain; Paula Krizova. National Institute of Public Health, Prague. Czech Republic; Reinhard Berner. University Hospital, Freiburg. Germany; Antoaneta Detcheva. National Center of Infectious and Parasitic Diseases, Sofia. Bulgaria; Mogens Kilian. Aarhus University, Aarhus. Denmark.

Conflict of interest. None declared.

REFERENCES