Inactivation of the CovR/S Virulence Regulator Impairs Infection in an Improved Murine Model of Streptococcus pyogenes Naso-Pharyngeal Infection

Streptococcus pyogenes is a leading cause of pharyngeal infection, with an estimated 616 million cases per year. The human nasopharynx represents the major reservoir for all S. pyogenes infection, including severe invasive disease. To investigate bacterial and host factors that influence S. pyogenes infection, we have devised an improved murine model of nasopharyngeal colonization, with an optimized dosing volume to avoid fulminant infections and a sensitive host strain. In addition we have utilized a refined technique for longitudinal monitoring of bacterial burden that is non-invasive thereby reducing the numbers of animals required. The model was used to demonstrate that the two component regulatory system, CovR/S, is required for optimum infection and transmission from the nasopharynx. There is a fitness cost conferred by covR/S mutation that is specific to the nasopharynx. This may explain why S. pyogenes with altered covR/S have not become prevalent in community infections despite possessing a selective advantage in invasive infection.


Introduction
Streptococcus pyogenes is estimated to cause 616 million cases of pharyngeal infection per year, and 663,000 cases of invasive disease [1]. As the human nasopharynx represents the major reservoir for all types of S. pyogenes infection, it is essential to develop a better understanding of the factors that influence upper respiratory tract infection.
Despite their limitations, mice play an important role in infectious diseases research [2]. The mouse nasopharynx has structural similarities to the nasal turbinate system in humans [3], although mice lack tonsils [4]. Instead, mice possess nasal associated lymphoid tissue (NALT), which shares some similarity to the tonsils [5] and has been shown to be a target for infection by S. pyogenes [6]. Indeed, mice have been used by several groups to investigate S. pyogenes in the upper respiratory tract, although there is no consensus on which is the most appropriate strain, sex or age of animal to use [6,7,8,9,10]. Furthermore, the maximum dose volume posited for establishing infection by previous studies ranges from 5 ml, as determined by administration of colored dye [8,11], 10 ml as determined by radioactive microspheres [10,12], to 20 ml volumes [6]. This is an important consideration, as aspiration of the bacteria into the lungs has the potential to trigger a more invasive disease and systemic infection.
It is known that phenotypic differences can exist between nasopharyngeal and invasive S. pyogenes isolates, and these have been ascribed to altered activity or mutation of the streptococcal two component regulatory system, covR/S [13]. As a result of signalling from the sensor kinase, CovS, CovR represses a range of virulence factors concerned with resistance to phagocytosis, such as the capsule synthesis operon hasABC, the DNase, sda, and the CXC chemokine protease, SpyCEP [13,14,15,16,17]. Mutations in covR/S de-repress these virulence genes, conferring a selective advantage to S. pyogenes in mouse models of invasive infection, leading to greater mortality [15,18]. However, the impact of such mutations on nasopharyngeal infection is unclear. Isolates of S. pyogenes with covS mutations bind less well to skin cells in vitro and in vivo than those without the mutation [19]. Furthermore, S. pyogenes with mutations in covS lack competitiveness in the saliva relative to wild type [20].
In this work, we set out to produce a longitudinally monitored murine model of nasopharyngeal infection, by examining the effect of mouse strain, age and sex on S. pyogenes carriage. We evaluated S. pyogenes pharyngitis isolates from patients rather than a previously-described mouse-pathogenic strain that lacks a functional copy of the multigene activator, mga [21,22]. The improved nasopharyngeal infection model was used to evaluate the impact of the S. pyogenes CovR/S two component regulatory system on longevity and transmission of S. pyogenes upper respiratory tract infection.

Ethics Statement
In vivo experiments were performed in accordance with the Animals (scientific Procedures) Act 1986, subject to protocols set out in PPL 70/7379 that were approved by the Imperial College Ethical Review Process (ERP) panel and the UK Home Office.

Bacterial Strains
The bacterial strains used in this study are given in Table 1 [16,23,24]. Streptococcal strains were cultured in Todd Hewitt Yeast broth (THY) or on Columbia Blood Agar (CBA), while Luria Bertani (LB) medium was used for culturing C. rodentium ICC180. All strains were grown at 37uC. For animal experiments, S. pyogenes was grown without shaking with 5% CO 2 overnight, centrifuged at 18646g, (Sorvall RTH 750 Rotor), washed twice in phosphate buffered saline (PBS), and re-suspended in PBS to produce an inoculum of 1-7610 8 colony forming units (cfu) per 5 ml. Numbers of viable bacteria within the inoculum were retrospectively assessed by plating of 10 26 -10 28 dilutions of the inoculum onto agar.

Animals
Male and female 5-10 week old CD1, C57BL/6, A/J, BALB/c, FVB/n specific-pathogen free mice (Harlan, UK) were maintained in individually HEPA filtered cages with sterile bedding and free access to sterilized food and water. GLP Mini Fun Tunnels (Lillico), or Des. Res. Mini Mouse Houses (Lillico) were provided in each cage for environmental enrichment.

Intranasal Infection
Pilot experiments were conducted using ,10 9 cfu in doses of 2.5 ml-20 ml bioluminescent C. rodentium ICC180 [24] to determine the correct dosing strategy to deliver bacteria to the murine nasopharynx without lung involvement. Bioluminescence (as photons s 21 cm 22 steridian [sr] 21 ) from living animals was performed as previously described [24] using an IVISH 100 system (Perkin Elmer).
For streptococcal infection of the nasopharynx, 1-7610 8 cfu of S. pyogenes was administered intranasally using a pipette to mice in a volume of 2.5 ml per nostril under 2-5% isoflurane anaesthesia. Mice were weighed daily; reduction by 20% of original weight was a defined humane endpoint. Intramuscular Infection 6610 8 cfu S. pyogenes were administered to mice under isoflurane anaesthesia via injection with a 27 gauge needle into the right lateral thigh. Numbers of viable bacteria within the inoculum were assessed by retrospective plating of 10 26 -10 28 dilutions. At 72 hours, the right thigh muscle and ipsilateral inguinal node were extracted, weighed and homogenized into PBS and then plated out onto CBA for bacterial enumeration.  Figure 1. Bioluminescence imaging of bacterial distribution after intranasal inoculation with bioluminescent C. rodentium. 10 9 colony forming units (cfu) of bioluminescent C. rodentium were administered intranasally to 8 week old CD1 outbred female mice in 20 ml (n = 3), 10 ml (n = 2), 5 ml (n = 3) and 2.5 ml (n = 2) of PBS. Images were acquired using an IVIS spectrum system, and are displayed as images of peak bioluminescence, with variations in colour representing light intensity at a given location. Red represents the most intense light emission, while blue corresponds to the weakest signal. The colour bar indicates relative signal intensity (as photons s 21 cm 2 sr 21 ). Two representative mice shown for each group. doi:10.1371/journal.pone.0061655.g001

Nasal Sampling
The level of shedding of S. pyogenes from the nasopharynx was assessed longitudinally using direct nasal sampling. The nares of each mouse were gently pressed onto the surface of a CBA plate ten times. Exhaled particulates were streaked out, and the plates were then incubated overnight at 37uC with 5% CO 2 for bacterial enumeration. In preliminary experiments using naive, noninfected mice, a-hemolytic streptococci, staphylococci, pseudomonads and corynebacteria were recovered, but no b hemolytic Group A streptococci were found to naturally colonize the mouse nasopharyngeal tract.
For experiments where mice had been infected with S. pyogenes, b-hemolytic colonies were counted for each mouse and confirmed as S. pyogenes through Gram staining, catalase testing, oxidase testing, and Lancefield grouping. No other b-haemolytic bacteria were recovered from the mouse nasopharynx. Kaplan-Meier plots were created to analyse the duration of S. pyogenes shedding. Nasal samples were taken for 21 days post inoculation. Mice were determined to have stopped shedding upon the first instance of a nasal sample turning up negative.
In some studies, the nasopharynx was dissected and removed at fixed time points for microbiological culture. To ensure complete extraction of the nasopharynx, the skin and mandibles were removed to expose the cranium. This was then sectioned along the coronal plane at the bregma. The brain tissue within the cranium anterior to this incision was removed, exposing the posterior aspect of the nasal cavity, known as the cribriform plate. The orbits were removed from this tissue via sagittal incisions lateral to the premaxilla. The remaining tissue comprised the entire nasal cavity, and NALT. This was homogenized into PBS and serial dilutions plated onto CBA to quantify S. pyogenes from the whole nasopharynx.

Settle Plates
To detect the presence of airborne bacteria within cages of infected mice, CBA plates were placed in the upper rack of the individually HEPA filtered cages (n = 4 plates per cage) and exposed for defined time periods throughout each experiment.
Plates then incubated overnight at 37uC and the numbers of S. pyogenes colonies, (identified by Gram staining, catalase testing, and Lancefield grouping) were enumerated.

Histopathology
The head of each mouse was removed at the atlanto-occipital joint and sagittally hemi-sected. One half was fixed in formalin and processed routinely to paraffin wax, while the other was homogenised and plated to assess S. pyogenes numbers in the nasopharynx. Paraffin sections were cut at 6 mm and stained with Haematoxylin and Eosin and Gram stains. The degree of nasal damage and inflammation was scored as : No significant abnormality (Intact nasal mucosa and absence of inflammation), Mild (Focal erosion of the mucosa with local neutrophil exocytosis across the affected epithelium), Moderate (Focal necrosis and ulceration of the mucosa with local neutrophilic exocytosis and surface neutrophilic exudation), Marked (Extensive necrosis and ulceration of the mucosa with widespread neutrophilic exocytosis and surface neutrophilic exudation) and Severe (Extensive necrosis and ulceration of the mucosa with widespread neutrophilic exocytosis and surface neutrophilic exudation and with extension of necrosis into underlying stroma). Slides were reviewed and scored by an experienced histopathologist (KS).

Statistics
For statistical analysis of Kaplan-Meier curves, the Mantel-Cox Logrank test was applied. For statistical analysis of colony count comparisons, a non-parametric Kruskal-Wallis test and Dunn's post-test were used. P values less than 0.05 were defined as significant. Statistics were performed using Prism Graphpad version 5.02. Data are presented as median, 6 interquartile range.

Volume of Inoculum Determines Distribution within the Respiratory Tract
Bioluminescence imaging demonstrated that dose volumes of 20 ml volume delivered bacteria to the lungs, whereas this was not shown with lower volumes (Figure 1). Any dose volume above 10 ml was deposited in the trachea. Dose volumes of 5 ml and 2 ml did not distribute bacteria to the lungs. As the optimal dose volume for nasopharyngeal deposition without lung involvement or significant nasal clearance was 5 ml (2.5 ml per nostril), this was the volume used in subsequent experiments.  To select an appropriate bacterial strain for development of the model, BALB/c mice were intranasally inoculated with four clinical S. pyogenes strains of different emm genotypes. An emm75 pharyngitis strain was found to have shed for the longest period using direct nasal sampling ( Table 2) and was used in subsequent experiments. Importantly, a significant correlation was found between the numbers of colonies recovered from nasal shedding of S. pyogenes and bacterial numbers from dissected and homogenised nasal tissue on the same day ( Figure 2, r 2 .0.95, n = 36). This longitudinal method of monitoring was therefore employed in subsequent experiments consistent with the principles of the 3Rs (Replacement, Refinement, and Reduction) by reducing the numbers of animals used.

Longevity of Nasopharyngeal Shedding is Greater in FVB/ n Mice
To determine the most appropriate mouse strain for model development, male mice of different host backgrounds were infected intranasally with emm75 S. pyogenes, and observed longitudinally for 72 h using direct nasal sampling. Data were used to create Kaplan Meier plots to analyse the duration of shedding for each strain of mouse (Figure 3 A). FVB/n mice carried S. pyogenes longer and shed significantly more S. pyogenes on the final day of the time course than all other strains tested   weeks of age) and post pubertal (n = 10 per group, 5 per cage, 10 weeks of age) male and female FVB/n mice were intranasally inoculated with S. pyogenes. There was no significant difference in infection duration between 5 week old males and females over a 21 day period (Figure 4, Logrank p.0.05), whereas 10 week old males shed S. pyogenes for significantly longer than 10 week old female mice ( Figure 5, Logrank test p,0.05). However, the older males had to be housed separately to prevent intraspecific aggression and for this reason further experiments were conducted using female mice. DcovR/S Mutation is Detrimental to Long Term S. pyogenes Infection of the Nasopharynx Two groups of FVB/n female mice (n = 20, 5 per cage) were infected intranasally with 5 ml of 10 8 cfu of emm75 S. pyogenes or an isogenic DcovR/S strain and observed over 21 days. Kaplan Meier analysis of daily nasal samples demonstrated that the DcovR/S strain was shed from the nasopharynx for a shorter length of time compared to its wild type counterpart. (Figure 6, Mantel Cox Logrank p,0.05).
We considered the possibility that the reduced longevity of infection reflected a general fitness defect in the DcovR/S strain. However, following intramuscular infection of groups of mice with each strain, both the DcovR/S strain and wild-type strain survived equally well within the thigh muscle (n = 12 per group, 6 per cage, Figure 7 A). Furthermore the DcovR/S strain disseminated to the ipsilateral inguinal lymph node in greater numbers than the wildtype (Figure 7 B,

Transmission of S. pyogenes within a Mouse Cage is Dependent on the Proportion of Infected Donor Mice Present
Preliminary work showed that S. pyogenes shed by infected mice could lead to the infection of uninfected mice in the same cage. Transmissibility of S. pyogenes in the nasopharynx was formally investigated through introducing infected donor mice into cages of uninfected recipient mice (n = 8 per cage). The effect of varying the ratio of donor to recipient (D:R) mice present in a cage on transmission was evaluated.
Within 4 hours of donor introduction transmission occurred to all recipients. A higher D:R ratio resulted in greater counts of S. pyogenes cultured from recipient mice in those cages. Recipient mice in the cage with a D:R ratio of 4:4 had significantly more bacteria in the nasopharynx compared to mice in the cage with a D:R ratio of 2:6 ( Figure 9 p,0.05).

DcovR/S Mutation is Detrimental to the Transmission of S. pyogenes in a Mouse Model
Mice were observed over the first three days of infection, when shedding of the wild-type and the DcovR/S strain were shown to be similar, to determine whether a mutation in covR/S affects transmission.
Female FVB/n mice infected with either the wild-type or the DcovR/S strains were introduced into cages at a D:R ratio of 3:5. After inoculation, the infected mice were separated for six hours before being introduced to the recipients, to prevent passive inoculum transfer. Direct nasal samples were taken from recipient mice over three days after the introduction of the donor mice.
Recipient mice housed with donor mice carrying wild type S. pyogenes acquired significantly more bacteria over the time course than the recipient mice housed with donor mice carrying the DcovR/S strain (Figure 10 A, AUC analysis, with Mann Whitney U p,0.05).
There was no statistically significant difference in the abundance of airborne S. pyogenes in the cages of mice infected with the wild-type compared with the cages of mice infected with the DcovR/S strain. (Figure 10 B, AUC analysis with Mann Whitney U p.0.05).

Discussion
To facilitate the investigation of bacterial and host factors that influence S. pyogenes in nasopharyngeal infection, an improved, new model of nasopharyngeal colonisation was devised, with an optimised dosing volume to avoid fulminant infections. A noninvasive method of longitudinal monitoring was developed that does not require culling of mice at multiple time points, thus reducing the numbers of animals used.
An emm75 strain of S. pyogenes was found in preliminary experiments to be carried better than other emm types by BALB/c mice, although previous studies have however found that the BALB/c strain is more resistant to infection than other strains [9]. A number of mouse strains were therefore tested in this study, of which the FVB/n was found to be the most susceptible to S. pyogenes intranasal infection. Emm types 1, 2, 3, 4, 6, 12, 22 and 89 were also successfully carried by FVB/n mice (data not shown), although in some cases causing a far more severe disease than the emm75 strain, making them unsuitable for long term infection studies.
We found that gender had an impact on susceptibility to carriage in post pubertal mice only. Data were consistent with other published work demonstrating an increase in susceptibility to infection in male mice [9,25]. However, the intraspecific aggression expressed by males of this strain made them difficult to house in groups, and thus the older individuals were housed individually. In previous studies, housing mice singly has been demonstrated to increase immune responses [26,27,28], and would theoretically increase their resistance to infection.
Histological analysis during infection revealed that nasal shedding of S. pyogenes was associated with on-going inflammation that subsided in the second and third weeks of infection. Previous studies have demonstrated bacterial infection of the mouse NALT [6]. However this study focussed on the site of infection in the deeper nasal passages, which demonstrated a suppurative rhinitis. Studies have shown that S. pyogenes distributes to the ethmoid sinuses in humans during rhinosinusitis [29,30].
The model was used to demonstrate that a functional covR/S is required for optimum infection and transmission from the nasopharynx. The failure of the DcovR/S strain to survive in the mouse nasopharynx was not due to a consistent fitness burden, since in invasive infection, the covR/S strain disseminated in significantly greater numbers than the wild type bacteria to the inguinal lymph node.
During preliminary experiments, we became aware that individual mice occasionally became re-infected after clearing the initial infection. To address transmission, and the factors that may influence this, co-mingling experiments were conducted. These demonstrated that transmission occurred as early as 4 hours post introduction, and the numbers of bacteria recovered from the nares of the recipients increased as the number of infected donor mice in the cage was increased. Importantly, such transmission events may not necessarily constitute a productive infection, as S. pyogenes did not reach the same abundance in recipient mice as observed in donor mice.
Co-mingling was then used to examine the impact of covR/S on transmission. Despite the fact that the infection burden (as measured by direct nasal sampling) was similar between the two donor groups in the first 72 h, recipients housed with donor mice carrying the DcovR/S were demonstrated to have a shorter shedding duration than the recipient mice housed with the donors carrying the wild type strain. Furthermore, despite some differences between the experimental groups, settle plates placed in each cage revealed no significant difference in aerosolization between the two strains. This suggests that the impairment of the covR/S primarily affects the survival of S. pyogenes in the nasopharynx after the initial transmission event has taken place.
There is thus a fitness cost conferred by covR/S mutation specific to the nasopharynx that may explain why such bacteria have not Figure 10. Transmission of S. pyogenes is hampered by loss of covR/S regulation. Naïve five week old female FVB/n recipients co-mingled at a D:R ratio of 3:5 with female FVB/n mice infected with either the emm75 wild type strain or it's isogenic DcovR/S strain (5 610 8 cfu per dose) and sampled after the introduction of the donor mice. Donor mice had .5000 cfu recovered from direct nasal sampling throughout the experiment. The DcovR/S strain transmitted significantly less well to recipients compared to the wild type strain (A, n = 15 recipients per group, AUC analysis, followed by Mann-Whitney U test). Line indicates median, error bars indicate interquartile range. Settle plates exposed to the air in the cages revealed no significant differences in the bacteria deposited on the surface of the plates by mice infected with the strain, or the DcovR/S strain (B, n = 4 plates per cage, AUC analysis followed by Mann-Whitney U test p.0.05) Data is shown for individual animals with medians indicated by black line. doi:10.1371/journal.pone.0061655.g010 become prevalent in community S. pyogenes pharyngitis despite being advantageous in invasive infection.
The model described represents a refinement of previous systems to study upper respiratory tract infection by S. pyogenes; the model is non-invasive and allows longitudinal monitoring of bacterial infection, using the same mice throughout the study. Such a model will facilitate research which might otherwise require prohibitively large numbers of animals and could be of importance in future evaluation of vaccines, antimicrobials, as well as the factors that influence transmission.