The animal experimentation was approved by the ethics committee on the use of animals from Centro de Ciências da Saúde, Universidade Federal do Rio de Janeiro, Brazil (#01200.001568/2013-87- CEAU).

The ability to form biofilms by 37 SDSD isolates from bovine clinical mastitis obtained between 2011 and 2013 [collection II, (Alves-Barroco et al., 2021)] was evaluated on polystyrene and glass surface (borosilicate test tubes) according to previously described protocols (Genteluci et al., 2015; Alves-Barroco et al., 2019; Xu et al., 2022). For a comparative analysis, 18 SDSD isolates from bovine clinical mastitis (collection I) (Rato et al., 2013; Alves-Barroco et al., 2019), obtained between 2002 and 2003, were included in the study. Sample collection design followed the international (Directive 2010/63/EU of the European parliament, on the protection of animals used for scientific purposes) and national (Decreto-Lei n° 113/2013) welfare regulations and guidelines (ARRIVE) was previously approved by the Portuguese “Direção Geral de Alimentação e Veterinária (DGAV)” (authorization document 0421/000/000/2013). In addition, the two authors have a level C FELASA certification (Federation of European Laboratory Animal Science Associations).

To evaluate the biofilm production on glass surfaces, the bovine SDSD isolates were streaked on blood agar plates and incubated at 37 °C for 18 h in a 5% (v/v) CO₂ incubator. About 5 colonies were transferred to Tryptic Soy Broth (TSB; Becton, Dickinson and Company, Le Pont de Claix, France) supplemented with 0.5% (w/v) glucose and incubated at 37 °C until the middle of the exponential growth phase. The pure culture (OD₅₇₀ = 0.6) was diluted 1:40 in a glass test tube (16 x 1,05 x 100 mm, NORMAX, Portugal) containing TSB supplemented with glucose (final volume 4 mL) and incubated at 37 °C for 20 h. Then, the supernatant was removed, and the glass tube washed with sterile saline solution [0.85% (w/v) NaCl]. The tubes were incubated at 65 °C for 1 h for drying. Biofilms were resuspended in 4 mL of saline and the OD at 600 nm measured. The isolates were defined as non-biofilm producers: OD₆₀₀ ≤ 0.099, weak: OD₆₀₀ between 0.1–0.299, moderate: between 0.3–0.599, or strong biofilm producers: OD₆₀₀>0.600.

On polystyrene surfaces, after growing to the middle of the exponential phase, the bacterial culture (OD₅₇₀ = 0.6) was diluted 1:2 in TSB supplemented with glucose in a 96 well plate (final volume 200 µL/well). The 96 well plate was sealed and incubated at 37 °C for 20 h. The supernatant was removed, and the wells washed with saline to remove non-adherent bacteria. Then, the plates were incubated at 65 °C for 1 h for drying and fixing biofilm. The biofilm was stained with crystal violet 1% (w/v) for 1 min. The wells were washed with sterile distilled water until the dye from the negative control-wells was completely removed. The OD₅₇₀ of the stained biofilm was directly measured in a plate reader (Infinite M200, Tecan, Männedorf, Switzerland). Interpretation of biofilm formation was performed according to the criteria previously described (Alves-Barroco et al., 2019) and the isolates were therefore categorized as follows: non producer: OD ≤ OD_ctrl, (OD_ctrl = 0.060); weak producer: OD_ctrl < OD ≤ 2 × OD_ctrl; moderate producer: 2 × OD_ctrl < OD ≤ 4 × OD_ctrl. strong producer: OD > 4 × OD_ctrl

This assay was based on previously described protocols (Roma-Rodrigues et al., 2015) with few modifications. Bacteria were grown at 37°C in Todd Hewitt Broth (THB; Oxoid; Basingstoke, UK) supplemented with 0.5% (w/v) yeast extract until the middle exponential growth phase. The infection was started by adding bacterial suspension (containing 10⁶ bacterial cells) in Dulbecco’s Modified Eagle’s Medium (DMEM; ThermoFisher Scientific; Waltham, MA, USA) supplemented with 10% (v/v) fetal bovine serum (ThermoFisher Scientific) to 10⁴ human epidermal keratinocyte (HEK) cells (ATCC-PCS-200-010, ATCC, Manassas, VA, USA). The infected culture was incubated at 37 °C, 5% (v/v) CO₂, and 99% relative humidity. After 2 h and 4 h of incubation, HEK cells were washed with phosphate buffer saline (PBS, 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 1.8 mM KH₂PO₄, pH 7.4) (Sigma-Aldrich, St. Louis, MO, USA), to remove non-adhered bacteria, and then fixed with 2% (v/v) glutaraldehyde (Sigma-Aldrich) in PBS for 2 h at room temperature. The HEK cells were washed with PBS (three times), post-fixed with 1.0% (w/v) osmium tetroxide (Sigma-Aldrich) at 4°C for 1 h, and then processed as previously described (García-Pérez et al., 2003). The infected HEK cells were visualized using a scanning electron microscope (JEOL JSM-5400).

Viability of HEK cells was determined using MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt] assay as previously described (Fernandes et al., 2017). HEK cells were seeded in 96-well plates (ThermoFisher Scientific) and grown at 37 °C, 5% (v/v) CO₂, and 99% relative humidity for 24 h, before incubations in the same conditions for 2 h, 4 h, and 6 h with SDSD cells or bacterial growth supernatant. After the incubation period, the culture medium was removed and, after washing HEK cells with PBS, fresh medium containing 10% MTS reagent (ThermoFisher Scientific) was added to each well. The 96-well plate was incubated at 37°C, in the same atmosphere, for 60 min. The absorbance (Abs) was measured in a microplate reader at 490 nm (Infinite M200, Tecan, Männedorf, Switzerland). The following equation was applied: cell viability (%) = 100 x [mean Abs of SDSD cells (or mean Abs bacterial growth supernatant)/mean Abs of control group without SDSD cells or bacterial growth supernatant].

Exponentially growing SDSD cells in TSB supplemented with 0.5% (w/v) glucose were harvested by centrifugation and diluted in the fresh broth. A volume of 10 μL [containing 10⁴ colony forming units (CFU)] was injected into the lumen of a 1 cm segment of the polyurethane catheter (C-953-J-UDLM; Cook Inc., Bloomington, IN, USA). Then, the catheter was placed into the well of a 24-well plate and incubated for 72 h. To count the SDSD cells adhered, the catheter was washed with PBS twice, to remove non-adherent bacteria, and placed in fresh broth. After sonication (15 min; 38.5–40.5 kHz, in ice), the CFU/mL was determined using Todd Hewitt Agar (THA). Biofilm was assessed by counting the SDSD cells adhered to the catheter.

The in vivo assays using a mouse foreign-body model were performed as described in Genteluci et al., 2015. Briefly, young adult BALB/c mice (age between 8 to 10 weeks), obtained from NCAL-UFRJ (https://ccs.ufrj.br/paginas/sobre-o-ccs/coordenacoes/cambe), were anesthetized, and a subcutaneous incision was created to introduce a 1 cm segment of a polyurethane intravenous catheter containing 10 μL of the bacterial suspension (10⁶ CFU). The catheter was implanted subcutaneously (at least 1.5 cm from the incision). after 72 hours of infection, the animals were euthanized, and the catheter segments removed. After that, the catheter segments were washed with 0.15 M NaCl to remove any planktonic bacteria and placed in a tube containing 1 mL of saline. After sonication (15 min, 38.5–40.5 kHz, in ice), CFU/mL was determined using THA. Biofilm was assessed by counting the SDSD cells adhered to the catheter.

To investigate whether animal passage can increase the ability of SDSD to accumulate biofilms in vitro, cells collected from the catheter implanted in the mice were inoculated in TSB containing 0.5% (w/v) glucose at 37 °C for 18 h. Then, aliquots (with and without animal passage) were obtained to assess biofilm formation on the glass surface following the protocol described above.

Expression levels of genes associated with biofilm formation in other streptococci were evaluated using sessile cells recovered from in vivo and in vitro biofilms. RNA was extracted using NucleoSpin RNAII kit (Macherey-Nagel, Dueren, Germany) according to the manufacturer’s instructions to comparatively quantify transcripts of genes encoding BrpA-like (biofilm regulatory protein), FbpA (fibronectin-binding protein A), HtrA (serine protease), and SagA (streptolysin S precursor). The cDNA was synthesized using the SuperScript first-strand synthesis system (Invitrogen) according to the manufacturer’s instructions. The RT-qPCR reaction mixture (20 μL) contained NZY qPCR Green Master Mix (NZYTech, Lisbon, Portugal), 1 μL cDNA, and 0.5 μM of the forward and reverse primers described in Table 1 . PCR conditions included an initial denaturation for 10 min at 95°C, followed by 30 cycles of amplification consisting of denaturation for 15 s at 95°C, and annealing for 30 s at 58°C and extension for 45 s at 60 °C. The critical Ct was defined as the cycle in which fluorescence becomes detectable above the background fluorescence. The expression levels were normalized using the 16S rRNA gene as an internal standard. Each assay was performed with at least three independent RNA samples.

GraphPad Prism version 7.0 was used for statistical analysis. All data were expressed as mean ± SEM from at least three independent (biological) experiments. The statistical significance was determined for each data set using the student’s t-test, and statistical significance was considered when p < 0.05.

For comparison purposes, biofilm developments between SDSD isolates recovered from catheter implanted in mice (in vivo) and in vitro assay were both measured by CFU. In the case of biofilm developments by SDSD on glass surfaces before and after animal passage, biofilm growth was measured by OD determination.

As a first approach, the ability of 37 SDSD isolates (collection II) to form biofilms on glass and polystyrene surfaces was evaluated. For a comparative analysis, the results obtained for 18 SDSD isolates also obtained from bovines (collection I), shown in Alves-Barroco et al. (2019), were included in the study. Overall, despite some differences, the results obtained point to a high in vitro biofilm-forming ability by most SDSD isolates with isolates from the collection II showing a greater ability to accumulate biofilms than those from the collection I ( Figure 1 ).

HEK cells have an important role in host defense, providing a physical and immunological barrier against pathogenic bacteria. The adherence and/or invasion of Group G streptococci in HEK cells was associated with the severity of skin infections, e.g., necrotizing fasciitis (Siemens et al., 2015). We previously reported for the first time that bovine SDSD isolates are capable to adhere and internalize several human cells, including HEK cells (Roma-Rodrigues et al., 2016); Alves-Barroco et al., 2018. Here, the ability of SDSD isolates to form biofilms on HEK cells was analyzed by scanning electron microscope ( Figure 2 ). Our results demonstrated the formation of an extracellular polymeric matrix by the growth of SDSD biofilms after 2 h of infection of HEK cells ( Figure 2A ). After 4 h of infection, it was possible to visualize a typical biofilm architecture ( Figure 2B ).

Similar results were observed by Matsue and co-workers in investigations with other streptococci. They showed that after 2 h incubation of HEK cells with Streptococcus dysgalactiae subsp. equisimilis (SDSE), the percentage of adhered bacterial was on average 70%. Furthermore, the adherence of SDSE on HEK cells was about 10 times higher than that on polystyrene surfaces (Matsue et al., 2020). Previous studies have also shown the formation of biofilms by Streptococcus pyogenes in HEK cells. Visually, S. pyogenes biofilms formed on HEK cells were similar to biofilms on abiotic surfaces; however, S. pyogenes biofilms on HEK cells were more resistant to antimicrobial therapy (Marks et al., 2014). Marks and co-workers demonstrated that during coculture, the S. pyogenes biofilm extended about 20–30 µm above the HEK cells; however, S. pyogenes biofilms did not induce HEK cell death, as the keratinocytes layer remained intact during the experiment (Marks et al., 2014). Contrary to what was observed for S. pyogenes (Marks et al., 2014), SDSD biofilm induced a decline in viability of the HEK cells over time ( Figure 3 ). After 6 h incubation with VSD13 isolate grown in biofilm condition or with its filtered culture supernatant, the viability of HEK cells was 32% and 86%, respectively. Our results also showed that the development of biofilms on the HEK cell monolayers exhibited greater cytotoxicity than extracellular products from bacterial growth ( Figure 3 ). Thus, these data could indicate that SDSD biofilm may develop during the process of skin/soft tissue infections, suggesting that it might be important in the pathogenesis of human SDSD cellulitis.

Together, they suggest an important role for SDSD biofilm formation on HEK cells, which may contribute to the development of deeper tissue infections and bacterial dissemination. In fact, in recent years, the association of SDSD with human infections such as cellulitis has been reported (Chennapragada et al., 2018; Nathan et al., 2021), and one case of cellulitis rapidly progressed to septic shock (Nathan et al., 2021). Indeed, we have recently reported for the first time the resistance to conventional antibiotics associated with biofilm formation by SDSD (Alves-Barroco et al., 2022), which may complicate the treatment of some infections associated with this subspecies.

To reduce the number of sacrificed animals and to compare in vivo and in vitro biofilm formation and accumulation, SDSD isolates were selected based on their in vitro ability to form strong (n=2; isolates VSD9 and VSD22; collection I and II, respectively), moderate (n=1; isolate VSD16), or weak (n=2; isolates VSD13 and VSD45; collection I and II, respectively) biofilms on the glass surface. All SDSD isolates tested, including weak biofilm producers in vitro, were able to develop biofilm in vivo ( Figure 4 ).

The results showed an important increased ability to develop biofilm on catheter implanted in mice compared with the respective biofilm formed in vitro, except for the SDSD isolate VSD22, which already produced a very strong biofilm in vitro. Overall, the results suggest that the capability of SDSD isolated from bovines to develop strong biofilm in vivo is independent of the ability to form biofilms in vitro on abiotic surfaces. A possible limitation of this study is the fact that we used a collection of SDSD isolates from 2002 to 2013. However, it is important to emphasize that our results were not influenced by SDSD collection period, being similar for the isolates obtained in 2002-2003 or 2011-2013. Indeed, our data contribute for a better understanding of the pathogenic mechanisms of diseases not only in animals but also in humans, such as cellulitis and prosthetic joint infections that happened during these periods (Koh et al., 2009; Park et al., 2012).

The quantification of in vitro versus in vivo biofilms (CFU/mL) varied, respectively, from 9.8 x10³ to 4.0 x10⁷ for VSD9 isolate, from 8.0 x10³ to 7.8 x10⁶ for VSD13 isolate, from 1.8 x10³ to 3.4 x10⁶ for VSD16 isolate, and from 7.4 x10⁴ to 6.4 x10⁶ for VSD45 isolate. An increased ability (approx. 2.2 times) to form biofilm in vitro (glass surface) after animal passage was observed for the SDSD isolate VSD45 (weak biofilm producer in vitro, Figure 5 ), but not for the other isolates tested.

Once the role of biofilms in infections was recognized, different in vitro and in vivo models of infections were developed. The in vitro models, although more simplistic, contributed to the current knowledge of the biofilm. These models are also used to investigate the role of genes involved in biofilm formation and to screen antimicrobial agents capable of disintegrating bacterial biofilms. However, in vitro models ignore important parameters and host factors. The in vivo models have contributed to a better understanding of bacterial adhesion, invasion and cytotoxicity factors, as well as the mechanisms involved in host inflammatory responses. There is no gold-standard model since each model can provide a specific answer. Information about models of biofilm-related infections and their applications is reviewed in Lebeaux et al., 2013.

The foreign-body mouse model used in the present work has also been successfully applied in previous studies with different bacterial species to analyze the ability of bacteria to form biofilms (Ferreira et al., 2012; Marks et al., 2014; Genteluci et al., 2015). Genteluci and co-workers showed that SDSE isolates can form biofilms in vivo, regardless of their ability to form biofilms in vitro. Marks and co-workers showed that S. pyogenes non-biofilm producers on abiotic surfaces ​can form biofilms on epidermal cells with characteristics similar to an in vivo colonization or infection (Marks et al., 2014). Although far from a complete understanding of the multiple factors that control the interactions between the pathogen and the host, animal models provide a better understanding of the biofilm within the context of the host. Taking all our data together, it can be concluded that in vivo growth increases the SDSD ability to form biofilms, possibly reflecting an important impact during some SDSD infections in bovine and in human hosts.

In vivo colonization by the group of pyogenic streptococci requires a series of interactions between the pathogen and the host, involving differential gene expression in both (Alves-Barroco et al., 2020). Our data revealed that in vivo, a similar number of sessile cells were recovered from SDSD isolates previously classified as weak and strong biofilm producers in vitro. This difference might be explained by changes in gene expression profiles associated with the regulation of biofilm formation (Marks et al., 2014). To test this hypothesis, we compared the expression of some biofilm-associated genes in sessile cells of SDSD isolates grown in vivo and in vitro; except for VSD22 isolate, as no difference was observed between in vivo and in vitro biofilm formation. Indeed, a remarkably increased expression was observed for the genes encoding BrpA-like (biofilm regulatory protein) and FbpA (fibronectin-binding protein A) for all bacterial biofilms collected from catheters recovered from the mice model ( Figures 6A , B ). The brpA-like gene was upregulated ~182, 112, 335, and 144-fold for VSD9, VSD13, VAS16, and VSD45, respectively, while fbpA was upregulated ~369, 822, 1419, and 708-fold for VSD9, VSD13, VAS16, and VSD45, respectively. The mRNA expression of htrA (encoding serine protease) was more dramatically increased for VSD9 (796-fold) and VSD13 (1441-fold) isolates, and mRNA expression of sagA (encoding streptolysin S percussor) for VSD13 isolate (~9.8-fold) ( Figures 6C , D ).

The role of the biofilm regulatory protein A (BrpA) in autolysis and cell division of the Streptococcus mutans has been shown. In vitro, the BrpA-deficient mutant of S. mutans maintained its adherence property, but the ability to form biofilms was considerably affected. Additionally, the deficiency of BrpA impaired cell envelope stress responses and acid and oxidative stress tolerance. (Bitoun et al., 2012; Bitoun et al., 2014). The biofilm-producing SDSD isolates carry a brpA-like gene, which expression level was associated with the ability to form biofilms jn vitro (Alves-Barroco et al., 2019). Therefore, our data showing a parallel increase of biofilm accumulation and brpA-like gene expression in the in vivo model corroborate a role played by this gene in the development of biofilm by SDSD.

Studies estimated that initial adherence to host cells is mainly mediated by adhesins, such as fibronectin-binding proteins, that allow adhesion to provide biofilm development on host tissues and/or bacterial internalization into host cells (Cue et al., 2000; Šmitran et al., 2018). Several fibronectin-binding proteins are produced by S. dysgalactiae, with different binding affinities and properties (Alves-Barroco et al., 2020). These proteins provide adherence to human cells such as fibroblasts and keratinocytes cells, contributing to biofilm development in vivo and consequently to persistent infections (Collin and Olsén, 2003; Brandt and Spellerberg, 2009). In this work, we observed an increased expression of fibronectin-binding protein A (fbpA) gene for sessile cells of SDSD grown in vivo compared with in vitro ( Figure 6B ). These results corroborate studies showing the important role of fibronectin-binding proteins in biofilm formation (Brandt and Spellerberg, 2009; O'Neill et al., 2009).

The high-temperature requirement protein A (HtrA, also known as DegP), is a serine protease widely distributed among streptococci (Alves-Barroco et al., 2020). HtrA homologs are responsible for the degradation of abnormal proteins in response to environmental stress. These proteins have also been identified in Gram-positive isolates (Kim and Kim, 2005). In Streptococcus mutans, the deletion of the htrA gene causes decreased ability to respond to environmental stress (Diaz-Torres and Russell, 2001). In S. pyogenes, the deletion of htrA gene affected the expression of several virulence genes (Lyon and Caparon, 2004). In addition to the proteolytic properties, this enzyme can adhere to the extracellular matrix of host tissues (Lyon and Caparon, 2004; Kim and Kim, 2005). Herein, differential expression of the htrA gene was observed between in vitro and in vivo biofilms. Isolates VSD9, VSD13 and VSD45 exhibited an increased htrA gene expression in vivo ( Figure 6C ). Interestingly, the VSD16 isolate exhibited a decreased expression of this gene ( Figure 6C ). Unlike other VSD isolates, such as VSD9 and VSD10, which produce extracellular matrices mostly composed of protein, VSD16 biofilm shows mucus-like material in the biofilm extracellular matrix (Alves-Barroco et al., 2019), suggesting that the matrix might be formed mostly by extracellular DNA (eDNA) or by complexes of eDNA and proteins (Alves-Barroco et al., 2019). Taken together, these results indicate that the VSD16 isolate may have a different biofilm formation pathway.

The sagA gene encodes the mature streptolysin S (SLS) toxin responsible mainly for the β-hemolytic activity among the pyogenic group of streptococci (Datta et al., 2005; Molloy et al., 2011). The SLS operon encodes the sagA gene (the structural propeptide), followed by genes that provide the conversion of SagA propeptide into SLS (sagB to D), leader cleavage (sagE), and transport across the membrane (sagF to I). The S. pyogenes SLS causes host soft-tissue damages, impacts phagocytes, and contributes to translocation across the epithelial barrier (Molloy et al., 2011). SLS also promotes programmed cell death and enhances inflammation in HEK cells (Flaherty et al., 2015). Studies revealed that SLS promotes host-associated biofilm formation (Vajjala et al., 2019), besides inducing mitochondrial damage and consequently macrophage death (Tsao et al., 2019). Datta and co-works reported that all SLS operon is required for the functional expression of streptolysin S (Datta et al., 2005). The loss of SagB-I observed in SDSD isolates from bovine origin is associated with loss of β-hemolytic activity (Alves-Barroco et al., 2021); however, the sagA gene has been maintained in the bovine SDSD genome, which possibly indicates an additional function to the product of this gene (Alves-Barroco et al., 2021). Some studies suggested that SagA plays an important role in the regulation of several virulence determinants, including M proteins (Datta et al., 2005; Molloy et al., 2011). The mechanisms by which sagA mRNA regulates virulence in Streptococcus have been the subject of investigations. In the present study, a high and significant increase in sagA expression was observed in vivo for the sessile cells of VSD13 isolate ( Figure 6D ), suggesting that the regulation of sagA expression may differ in a strain-specific manner.

The multiple factors that control the interactions between the pathogen and the host is still far from being fully understood. Nonetheless, the animal models provide better knowledge of biofilm within the host context. Importantly, the in vivo biofilm growth of SDSD isolates trigger distinct stress pathways that lead to the upregulation of the brpA-like, fbpA, htrA, and sagA genes that may play an important role in the host colonization and infection. Although previous studies suggest that SDSD may have different host preferences (Porcellato et al., 2021), together, the results presented in the present work suggest that SDSD isolates from bovine origin are able to infect other hosts, and may have a potential zoonotic capability.