|Home | About | Journals | Submit | Contact Us | Français|
We developed a human CD46-expressing transgenic (Tg) mouse model of subcutaneous (s.c.) infection into both hind footpads with clinically isolated 11 group A streptococcus (GAS) serotype M1 strains. When the severity levels of foot lesions at 72 h and the mortality rates by 336 h were compared after s.c. infection with 1 × 107 CFU of each GAS strain, the GAS472 strain, isolated from the blood of a patient suffering from streptococcal toxic shock syndrome (STSS), induced the highest severity levels and mortality rates. GAS472 led to a 100% mortality rate in CD46 Tg mice after only 168 h postinfection through the supervention of severe necrotizing fasciitis (NF) of the feet. In contrast, GAS472 led to a 10% mortality rate in non-Tg mice through the supervention of partial necrotizing cutaneous lesions of the feet. The footpad skin sections of CD46 Tg mice showed hemorrhaging and necrotic striated muscle layers in the dermis, along with the exfoliation of epidermis with intracellular edema until 48 h after s.c. infection with GAS472. Thereafter, the bacteria proliferated, reaching a 90-fold or 7-fold increase in the livers of CD46 Tg mice or non-Tg mice, respectively, for 24 h between 48 and 72 h after s.c. infection with GAS472. As a result, the infected CD46 Tg mice appeared to suffer severe liver injuries. These findings suggest that human CD46 enhanced the progression of NF in the feet and the exponential growth of bacteria in deep tissues, leading to death.
Group A streptococci (GAS), which are among the most common human pathogens, can cause a variety of uncomplicated superficial skin infections such as impetigo/pyoderma or throat infections, including streptococcal pharyngitis (sore throat) and tonsillitis (3). In addition, patients suffering from acute and complicated GAS infections, in particular streptococcal toxic shock syndrome (STSS) associated with severe necrotizing fasciitis (NF), have high mortality rates (4). M proteins, which attach to the cell wall, are one of the important virulence factors that GAS possess (15). The various GAS strains have more than 100 different antigenically distinguishable M proteins (46), and GAS serotype M1 strains are regarded as a highly virulent group (3).
Membrane cofactor protein (MCP; CD46), which is expressed in every cell type except erythrocytes, is implicated as a receptor for at least six human pathogens (four viruses and two bacteria), including measles virus (9, 31, 35), herpesvirus 6 (40), adenovirus groups B (17, 41) and D (48), pathogenic Neisseria (23), and GAS (37). Human CD46-expressing transgenic (Tg) mice are susceptible to streptococcal disease (30). When CD46 Tg and non-Tg mice were infected intravenously (i.v.) with GAS, the bacteremia levels, frequency of arthritis, and mortality rate were higher in CD46 Tg mice than in non-Tg mice (30). Unfortunately, this animal model does not reflect the natural infection process in the human host. Consequently, we have attempted to establish a CD46 Tg mouse model of skin and soft tissue infection with GAS that closely represents the human disease. If the disease pathogenesis in an animal model is in many ways very similar to that observed in humans, it will enable researchers to address a variety of questions regarding the development of urgently needed diagnostic methods and therapies. To the best of our knowledge, this is the first report to create the CD46 Tg mouse model of subcutaneous (s.c.) infection with GAS472 (a serotype M1 strain isolated from the blood of a patient suffering from STSS) into both hind footpads, thereby causing an acute disease that mimics severe NF in human.
Beta-hemolytic GAS serotype M1 strains collected in Japan in 2006 are described in Table Table1.1. GAS strains were preserved in 10% (wt/vol) skim milk and stored at −85°C until use. The frozen GAS strains were streaked onto sheep blood agar plates (Nippon Becton Dickinson, Tokyo, Japan) and cultured overnight at 37°C in a humidified 5% CO2 incubator. Prior to their use in infection experiments, GAS strains were grown on Todd-Hewitt broth containing 0.2% (wt/vol) yeast extract (THY; Difco and BBL, Detroit, MI) in 5% CO2 at 37°C without shaking.
The human CD46-expressing Tg mice were a generous gift from J. P. Atkinson, Washington University. CD46 Tg mice possess two copies of human CD46 gene per diploid genome, and the expression levels of CD46 RNA and protein were comparable to those observed in matched human tissues, including muscle, fat, and skin (25). Using immunohistochemistry, human CD46 was found on epithelial cells, endothelial cells, glial cells, and hepatocytes; in the glomerulus (in kidney) and adrenal gland; as well as on B cells, T cells, neutrophils, and macrophages (22). The CD46 Tg mice used in this study had the C57BL/6 genetic background (H-2b haplotype). In previous studies, C57BL/6 (21, 22, 25, 34) or B6C3F1 (H-2bk) (30, 42) mice have been used as non-Tg control mice. Therefore, to begin this study, we used both C57BL/6 and B6C3F1 mice obtained from Charles River Japan (Yokohama, Japan) as non-Tg control mice. The expression of many pathogenic traits of GAS has been shown to depend on the growth phase (33). Thus, 7-week-old mice were s.c. infected with a total of 1 × 107 CFU of stationary-growth-phase GAS into both hind footpads (5 × 106 CFU per footpad). After s.c. infection, the survival rates were observed every 24 h for 336 h postinfection, or the number of bacteria in the blood and tissue samples including the popliteal lymph nodes (PLN), spleen, and liver were determined by plating onto sheep blood agar (11, 27). All mice were bred at the animal facility at the Kitasato Institute, and all mouse experiments were performed in accordance with institutional protocol guidelines under an approved protocol.
Macroscopic images were obtained with a digital camera (D80; Nikon, Tokyo, Japan). For histological examination, a portion of each footpad and a portion of each liver were fixed with 10% (vol/vol) buffered formalin and then embedded in paraffin. Tissue sections approximately 5 μm thick were prepared and mounted on glass slides. The slides were stained with hematoxylin and eosin (H&E) and Giemsa. Alternatively, another portion of each liver was also fixed with Zamboni's fixative (2% [wt/vol] paraformaldehyde and 15% [vol/vol] saturated picric acid in 0.1 M phosphate buffer, pH 7.3) for at least 8 h at 4°C (44), and the 4-μm cryostat sections obtained were stained with Alexa Fluor 594-conjugated phalloidin (Molecular Probes, Eugene, OR) at a dilution of 1:1,000. The sections were mounted in Permafluor and were observed by confocal laser microscopy (TCS NT; Leica Biosystems, Nussloch, Germany).
The bacterial cells of GAS472 grown in THY broth were washed once with Dulbecco's modified Eagle's medium (Sigma-Aldrich, St. Louis, MO) and suspended in Dulbecco's modified Eagle's medium. The bacterial cell suspension (1 × 108 CFU) was mixed with an equal volume of whole blood from CD46 Tg or C57BL/6 mice collected by cardiac puncture with a heparin-coated 23-gauge needle. The mixture was then placed in a 37°C humidified 5% CO2 incubator. At 0, 1, 2, and 6 h of incubation, bacterial cells were harvested by centrifugation and washed three times with cold distilled water. Gene expression patterns of GAS472 were determined by quantitative real-time reverse transcription (RT)-PCR analysis. Total RNA extracted from bacterial cells using the RNeasy Protect bacteria minikit with on-column RNase-free DNase I treatment (Qiagen Sciences, Germantown, MD) was subjected to the RT reaction from 100 ng of total RNA in a 20-μl reaction volume using the iScript cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA). The following real-time PCR assays were performed with RNA templates to ensure that contaminating genomic DNA was absent before the RT reaction. Quantitative real-time PCR in triplicate on 1 μl of template cDNA per 10-μl reaction was performed in a CFX96 real-time PCR detection system (Bio-Rad Laboratories) using the iQ SYBR green supermix (Bio-Rad Laboratories). Gene-specific primer pairs (16S rRNA, 5′-GAGAGACTAACGCATGTTAGTA-3′ and 5′-TAGTTACCGTCACTTGGTGG-3′; sagA, 5′-TGCTGCTGCTGCTGTACTAC-3′ and 5′-GCTTCCGCTACCACCTTGAG-3′) were designed using Beacon Designer 7.5. The PCR conditions included an initial denaturation step at 95°C for 3 min, followed by 50 amplification cycles of denaturation at 95°C for 10 s, annealing at 52°C for 10 s, and extension at 72°C for 30 s.
For a comparison of the DNA sequences of the csrRS (also called covRS) locus between virulent (GAS472 and GAS467) and avirulent (RE386 and RE344) strains of GAS, we used 12 primers (p1 to p12) as previously described by Walker et al. (47). We used an ABI 3130 automatic sequencer (Applied Biosystems, Foster City, CA) for the direct sequence of the amplified PCR products (3,018 bp) with the primer pair p1/p12. The determined DNA sequences were assembled by the use of a component of Vector NTI version10.3.1 (Invitrogen, Carlsbad, CA).
Significant differences between the means ± standard deviations (SDs) of different groups were examined using a two-tailed unpaired Student's t test (see Fig. Fig.6)6) or a one-tailed paired Student's t test (see Fig. Fig.8).8). A P value of <0.05 was regarded as statistically significant.
As shown in Fig. Fig.1,1, we classified 11 GAS strains into two groups associated with the severity levels of foot lesions in CD46 Tg mice 72 h after infection with GAS strains. In the first group, the feet developed well-defined necrosis (GAS472, GAS467, and RE025). The second group showed swelled feet (the skin appeared red and blotchy) and occasionally blisters (RE335, GAS465, GAS453, RE303, RE386, RE137, RE157, and RE344). All infected CD46 Tg mice were still alive in this time period. In contrast, even in the worst cases, the necrotizing cutaneous lesions developed partially in the feet of non-Tg mice 72 h after infection with GAS472 (Fig. 2A and B). Alternatively, the feet merely swelled up or developed minor wounds in non-Tg mice 72 h after infection with GAS467 (Fig. 2C and D) or RE335 (Fig. 2E and F). These results demonstrate that GAS472 had the strongest virulence traits among the 11 GAS strains essential for causing necrotizing lesions. Accordingly, we selected and used GAS472 for the following histological study.
The footpad skin sections of CD46 Tg and non-Tg mice exhibited acute inflammation (neutrophil infiltration) in the hypodermis 6 h after infection with GAS472. At 24 h postinfection, although increased inflammatory cell infiltration was recognized in the dermis of CD46 Tg and non-Tg mice, higher numbers of inflammatory cells were observed in the non-Tg mice (Fig. 3B and E). In addition, exfoliation of epidermis with intracellular edema and hemorrhaging in the dermis developed in CD46 Tg mice (Fig. (Fig.3B).3B). At 48 h postinfection, the necrotic striated muscle layers were recognized in the dermis of CD46 Tg mice (Fig. (Fig.3C).3C). In contrast, although the exfoliation of epidermis was observed, the necrosis could not yet be detected in the dermis of non-Tg mice at this stage (Fig. (Fig.3F3F).
These findings suggest that s.c. infection with GAS472 rapidly led to severe NF in the feet of CD46 Tg mice.
As shown in Fig. Fig.4,4, we observed differences in susceptibility of CD46 Tg mice after infection with 11 GAS strains. As a result, the mortality rates of the 336-h observation period were ranked in the following order: GAS472 (100%) > RE335 (83%) > GAS467 (67%) > RE025 (50%) > RE157 = GAS453 (43%) > GAS465 = RE137 (29%) > RE303 = RE386 = RE344 (0%). Therefore, we chose GAS472, RE335, and GAS467 as the three most virulent of the 11 GAS strains concerning the susceptibility of CD46 Tg mice. We then newly infected CD46 Tg and non-Tg mice with GAS472, RE335, or GAS467 at the same time. As shown in Fig. Fig.5,5, the mortality rates among non-Tg mice were extremely low (0 or 10%). The differences in mortality rates between C57BL/6 and B6C3F1 mice after infection with each GAS strain were almost negligible. Thus, GAS472 was regarded as the most virulent strain in terms of the progression of NF and the susceptibility of CD46 Tg mice among 11 GAS strains.
As shown in Fig. Fig.6,6, CD46 Tg mice had a significantly higher number of viable bacteria in each blood, spleen, and liver sample compared with non-Tg mice 48 h after infection with GAS472. Thereafter, the viable bacterial numbers in the blood and in every tissue sample increased 72 h after infection in CD46 Tg mice. Especially, the number of bacteria increased 90-fold in the livers of CD46 Tg mice for 24 h from 48 h to 72 h postinfection. In contrast, the number of bacteria increased sevenfold in the livers of non-Tg mice for 24 h from 48 h to 72 h postinfection. In other words, the estimated bacterial doubling times during these 24 h were 3.7 h and 8.5 h in the livers of CD46 Tg and non-Tg mice, respectively.
Macroscopic observations showed the development of granulomatous nodules in the CD46 Tg mouse liver 72 h after s.c. infection with GAS472 (Fig. (Fig.7A).7A). Simultaneously, microscopic observations revealed the development of full porous cytoplasmic F-actin bundles (Fig. (Fig.7C)7C) in the liver sections of CD46 Tg mice after infection. Multifocal areas of necrosis were associated with inflammatory infiltrate (Fig. (Fig.7D),7D), and clusters of streptococci were visible at the edges of the necrotic areas (Fig. (Fig.7E)7E) in the liver sections of CD46 Tg mice after infection. Based on these results, we considered it likely that s.c. infection with GAS472 induced hepatic damage in CD46 Tg mice along with NF in their feet.
It was demonstrated that during 6 h of incubation, the growth of GAS was enhanced in the cell culture medium mixed with an equal volume of mouse whole blood containing soluble human CD46 (30). Then, we examined the changes in expression of sagA (streptolysin S; SLS) of GAS472 during 6 h of incubation under the above experimental conditions. As shown in Fig. Fig.8,8, ~3-, 30-, and 30-fold increases in expression of sagA were detected during the 1-, 2-, and 6-h incubations, respectively, with CD46 Tg mouse blood. In contrast, sagA increases in expression of about three-, seven-, and seven-fold were detected during the 1-, 2-, and 6-h incubations, respectively, with non-Tg mouse blood. Thus, progressive increases in the expression of sagA detected during the 2-h and 6-h incubations were CD46 Tg mouse blood dependent. The differences were not significant (P values of >0.05), although the expression levels of the CD46 Tg mouse blood-dependent group were higher than those of the non-Tg mouse blood-dependent group in all three of the experiments. Although we also examined for changes in gene expression of emm (M protein), csrS (sensor histidine kinase), scpA (C5a peptidase), scpC (serine protease), ska (streptokinase), speB (cystein protease), and speA (streptococcal pyrogenic exotoxin A) using real-time RT-PCR with each gene-specific primer pair, significant changes in gene expression were not detected during the 6-h incubation with CD46 Tg or non-Tg mouse blood.
In the meantime, it was also demonstrated that CsrRS (a member of the two-component regulatory system) influenced the transcription of 15% of all chromosomal genes of GAS, including negative controls on the gene expression of five virulence factors (capsule, cystein protease, streptokinase, SLS, and streptodornase) (18). Since recent studies showed that the mutation of csrRS was important for the invasive phenotype of GAS (2, 45), we compared the DNA sequences of csrRS among two virulent strains (GAS472 and GAS467), two avirulent strains (RE386 and RE344), and a type strain of GAS serotype M1 (MGAS5005). No DNA sequence variations were identified inside the coding regions of csrRS among GAS472 (accession no. AB513958), GAS467 (accession no. AB513957), RE386 (accession no. AB513956), RE344 (accession no. AB513955), and MGAS5005 (accession no. NC_007297). Instead, the following two variations were found in the 5′ noncoding regions of csrRS: variation 1, c.−55_−56insT (a single nucleotide insertion of T in GAS472, GAS467, RE386, and RE344 at the position between −55 and −56 from the TTG initiation codon of csrS in MGAS5005); and variation 2, c.−232delA (a single nucleotide deletion of A in GAS472 at position −232 from the ATG initiation codon of csrR in MGAS5005).
In human patients, NF due to GAS infection is defined pathologically by a deep-seated infection of the s.c. tissue that results in the progressive destruction of fascia and fat, with relative sparing of skeletal muscle (4, 13, 16, 28). NF is rare in young children but more common in otherwise healthy adults, in whom a minor skin wound often precedes the disease, as was reported in a case study of a 25-year-old man (13). That study showed extensive bands of liquefactive necrosis and patchy inflammation of the deep dermis and subcutis in specimens of the patient's skin and soft tissue (13). Moreover, a focal abscess with numerous degenerating neutrophils formed in the deep dermis in that case (13). These symptoms closely resemble the skin specimens of CD46 Tg mice after s.c. infection with GAS472, as described in Fig. Fig.33.
In previously published mouse models of skin and soft tissue infection with GAS, the bacteria were inoculated s.c. into the backs (8, 32, 36), haunches (6), or flanks (1, 5, 7, 38). Compared to such models, the s.c. inoculation into both hind footpads in the present study was a simple procedure, and the obtained data showed a narrow dispersion of reproducibility. A recent study demonstrated that infection in epithelial cells with GAS led to the shedding of human CD46 at the same time as bacterium-induced apoptosis and cell death occurred (30). In the present study, the amount of bacterial load in the PLN of CD46 Tg mice was the same as that of non-Tg mice 48 h after s.c. infection (Fig. (Fig.6B),6B), indicating that the same amounts of bacteria resided in the feet of CD46 Tg and non-Tg mice for 48 h after s.c. infection. However, since the progression of NF until 48 h after s.c. infection was dramatically elevated in the feet of CD46 Tg mice compared to those of non-Tg mice (Fig. (Fig.3),3), human CD46-mediated GAS infection was thought to be necessary for the observed acceleration of disease progression in the mouse feet. Interestingly, in many cases, non-Tg mouse feet were completely extinguished by persistent necrosis of skin, soft tissue, and bone up to 336 h after infection with GAS472, although the mice were still alive (Fig. (Fig.9A).9A). In addition, the majority of CD46 Tg mouse feet were also lost by persistent necrosis no later than 336 h after infection with even avirulent GAS strains other than GAS472, GAS467, and RE025 (Fig. (Fig.9B).9B). Possibly, such persistent necrosis due to GAS infection cannot lead to death in CD46 Tg and non-Tg mice.
We classified 11 clinical isolates into the virulent or avirulent strains of GAS based on the severity levels of foot lesions at 72 h and the mortality rates by 336 h after s.c. infection. RE303 and RE344 did not show the capacity to induce lethal disease in CD46 Tg mice (Fig. (Fig.4),4), despite the fact that they were recovered from patients with invasive diseases (Table (Table1).1). In both cases, the patients were immunocompromised due to preexisting diseases. Supposedly, even avirulent GAS strains are able to induce invasive diseases in the patients suffering from preexisting medical conditions.
According to a previous report, GAS infection was commonly associated with the early onset of shock and organ failure in humans (4). Another recent study demonstrated that GAS might bind available extracellular CD46 as a strategy to survive and avoid host defenses and that the lethal disease and arthritis were much more frequent in CD46 Tg mice than in non-Tg mice after i.v. infection with GAS (30). Our data clearly indicate that the proliferation rate during the 24 h between 48 and 72 h after s.c. infection with GAS472 in the liver was much higher in CD46 Tg mice than in non-Tg mice (Fig. (Fig.6D).6D). Indeed, bacterial infection induced serious liver damage in CD46 Tg mice (Fig. (Fig.7).7). Probably, human CD46 potentially ensures bacterial growth in the deep tissues. Thereafter, systemic GAS infection might lead to organ failure and ultimately to the death of the mice.
The M proteins appear as long, hairlike filaments on the bacterial surface and are composed of two α-helical chains that are predominantly arranged in a coiled-coil conformation (14). All of the M proteins contain a conserved C-terminal region near the cell surface and a hypervariable N-terminal region that provides the basis for serological typing (14). A function of M proteins was shown to be critical for protecting organisms from phagocytosis (14). An in vitro invasion study using A549 human lung epithelial cells demonstrated that the M protein-promoted ingestion of GAS was dependent on the coengagement of CD46 and α5β1 integrin (39). In addition, CD46-mediated invasion was also found to depend on the extracellular matrix protein fibronectin (39). CD46 has been generally recognized as a regulator of the complement system that facilitates factor I-mediated inactivation of the activated complement proteins C3b and C4b (29). In the meanwhile, the interaction between GAS and CD46 triggered cell signaling pathways that directly induced an immunosuppressive phenotype in T cells (26). In fact, by immunohistochemistry, M protein/fibrinogen complexes were identified in tissue biopsies from a patient suffering from NF and STSS (19). Recently, it was shown that the binding of CD46 to GAS serotype M18 strains was independent of Emm (M protein) or Enn (immunoglobulin A-binding protein) (12). There may be some biological advantages to GAS having Emm-dependent and Emm-independent CD46-binding abilities. In any case, whether or not the M1 protein directly responded to human CD46 in the host cell, we suppose that the human CD46-dependent regulation of virulence gene expression in GAS serotype M1 strains should be necessary for the induction of both local inflammation and deep tissue damage in the CD46 Tg mouse model.
It is well known that ScpC can cleave interleukin-8, a chemokine that mediates neutrophil transmission and activation (10, 20). The ScpC mutant strain did not degrade interleukin-8 and thus failed to prevent recruitment of immune cells; also, it caused disease after soft tissue infection (43), whereas the increased expression of ScpC and streptolysin O (a membrane-damaging extracellular toxin) caused by csrS mutation played a prominent role in the mouse model of GAS infection (2). Meanwhile, since GAS serotype M1 strains isolated from invasive diseases in human had mutation accumulation at the csrRS locus, resulting in an SpeB-negative phenotype (45), the SpeB expression levels of GAS serotype M1 strains were thought to be inversely correlated with disease severity in humans (24). In fact, an SpeB-negative phenotype caused by the csrRS mutation was a more virulent strain of GAS than the wild-type strain in the mouse model (47). Consequently, we had expected that GAS472 gained any mutation inside the coding regions of csrRS during infection, resulting in the high-virulence phenotype. Against our expectation, no DNA sequence variation was found inside the coding regions of csrRS in virulent strains (GAS472 and GAS467) or avirulent strains (RE344 and RE386) compared to a standard strain (MGAS5005). Unexpectedly, however, it became clear that the high-virulence phenotypes were not due to the mutation of csrRS. Therefore, the most significant questions—regarding genetic differences and virulence genes in different strains and how the virulent strains cause serious and lethal disease in humans—remain unanswered.
We are grateful to John P. Atkinson for kindly providing the CD46 Tg mouse strain.
This work was financially supported by a grant under the category “Research Projects for Emerging and Re-emerging Infectious Diseases” (H-19-002 and H-20-002) from the Ministry of Health, Labor and Welfare of Japan.
We declare that no conflict of interest exists.
Editor: J. N. Weiser
Published ahead of print on 8 September 2009.