Prevalence of virulence factor genes in Staphylococcus aureus isolated from infected skin lesions as determined by real-time polymerase chain reaction analysis for 61 total bacterial isolates and divided into 44 isolates associated with low white blood cell (WBC) count and 17 isolates associated with high WBC count on initial culture. A, Prevalence for low WBC count group vs high WBC count group. B, Prevalence for low WBC count group vs community. C, Prevalence for high WBC count group vs community. *P < .05.
Representative Staphylococcus aureus–infected skin lesions. A, Lesion infected with isolate CPM010 contained no white blood cells (WBCs) in the low-power field (LPF) and was positive for exfoliative toxins A and B (ETA and ETB). B, Lesion infected with isolate CPM044 contained more than 100 WBCs/LPF and was negative for ETA, ETB, and Panton-Valentine leukocidin (PVL). C, Lesion infected with isolate CPM015 contained more than 100 WBCs/LPF and was positive for PVL and toxic shock syndrome toxin.
Mertz PM, Cardenas TCP, Snyder RV, Kinney MA, Davis SC, Plano LRW. Staphylococcus aureus Virulence Factors Associated With Infected Skin LesionsInfluence on the Local Immune Response. Arch Dermatol. 2007;143(10):1259-1263. doi:10.1001/archderm.143.10.1259
To evaluate Staphylococcus aureus isolates from infected skin lesions for their potential to produce immune system–modulating toxins and to correlate these with white blood cell (WBC) counts associated with these lesions.
Specimens were obtained for bacterial culture and gram staining from 105 infected skin lesions, and the number of WBCs per low-power field (LPF) was determined. Chromosomal DNA was prepared from 84 bacterial isolates and subjected to real-time polymerase chain reaction analysis to determine the presence of genes encoding potential immunomodulating toxins. Bacterial populations were divided into 2 groups: those associated with low WBC counts (0-5 WBCs/LPF) and those with high WBC counts (> 5 WBCs/LPF). We applied χ2 statistical analyses to compare the toxin gene profiles associated with WBC counts on initial swab for culture.
Samples were obtained from patients at a single geographic location.
A higher than expected percentage of bacteria capable of producing the exfoliative toxins A and/or B (ETA and/or ETB) and Panton-Valentine leukocidin (PVL) was seen in all skin lesions infected with S aureus without regard to WBC count with initial cultures. Comparison of the toxins associated with the low WBC group vs the high WBC group showed that low WBC counts were associated with ETA and ETB, while high WBC counts were associated with PVL and toxic shock syndrome toxin. There were no differences in the clinical appearance of the lesions between groups.
Staphylococcus aureus virulence factors ETA, ETB, and PVL are associated with WBC counts from infected skin lesions. The exact role they play in affecting the WBC counts remains to be determined.
Staphylococcus aureus is a major cause of multiple types of infections both in and outside of the hospital setting. These infections range from superficial skin infections to deeper infections of hair follicles, abscesses, and deep tissue infections, and even to systemic infections including those of the heart, lungs, bones, and blood.1Staphylococcus aureus carries a large repertoire of virulence factors, including over 40 secreted proteins and enzymes that it uses to establish and maintain infections.2,3 Some of these virulence factors are know to cause or be associated with specific diseases, for example, toxic shock syndrome toxin (TSST) and toxic shock syndrome; Panton-Valentine leukocidin (PVL) and necrotizing pneumonia and skin diseases4,5; the exfoliative toxins A and B (ETA and ETB) and scalded skin syndrome, impetigo, skin infections, and atopic dermatitis6,7; and the family of staphylococcal enterotoxins A and B (SEA and SEB) and food poisoning.3
Also among these virulence factors are several toxins that have the potential to manipulate components of the immune response. These include leukotoxins that specifically affect white blood cells (WBC) (PVL and leukocidin D-E [LukD-E]); toxins that function as superantigens that can manipulate the immune system by hyperstimulating the release of cytokines (SEA, SEB, and TSST); and hemolysins that are active to lyse red blood cells and WBCs as well as other cell types, as described by Foster.8
To our knowledge, an association of WBCs in infected skin lesions and the presence of bacteria capable of producing potentially immunomodulating toxins has not been reported. Our hypothesis is that the presence of immunomodulating or cytotoxic virulence factors produced by certain bacterial pathogens in infected skin lesions will have an effect on the number of WBCs associated with secondarily infected skin lesions. In the present study, S aureus cultures isolated from secondarily infected skin lesions were evaluated for their potential ability to produce immune system–modulating toxins by identification of the encoding genes using polymerase chain reaction (PCR). The association of these virulence factors with the level of localized WBC response was determined.
Swab specimens of infected skin lesions were obtained for bacterial culture and gram staining, and photographs were taken from 105 patients at a single geographical location after institutional review board approval and appropriate informed patient consent were obtained. Specimens were processed by a centralized laboratory. Slides were prepared for light microscopy, and bacterial and WBC counts per low-power field (LPF) were determined. Bacterial isolates including S aureus and Streptococcus species were obtained from 85 of 105 lesions. Seventy-two isolates were determined to be S aureus by standard identification methods and PCR detection of S aureus–specific gyrA gene. These 72 isolates were considered in the analysis of virulence factors associated with S aureus from infected skin lesions. Sixty-one of the 72 bacteria were isolated without coinfecting streptococci and were divided into 2 groups according to corresponding WBC count in the lesions and considered in the analysis of the effect of S aureus infection on the presence of localized WBCs.
Chromosomal DNA samples were obtained from each of the 84 bacterial isolates collected. DNA was prepared using a DNeasy tissue kit (Qiagen, Valencia, California) according to the manufacturer's recommendations and with the addition of lysostaphin to aid in the lysis of Staphylococcus species. The DNA was collected in 10mM Tris–hydrochloric acid. Concentrations of DNA were determined using a Quant-iT dsDNA BR or HS Assay Kit and Qubit Fluorometer (Invitrogen, Carlsbad, California). The quality of the chromosomal DNA was confirmed by agarose gel electrophoresis using 0.8% gels, with ethidium bromide staining. Samples of DNA were diluted for consistent template amounts for each PCR reaction, aliquoted to prevent contamination, and stored at −20°C.
Real-time PCR analyses were preformed on all bacterial isolates to determine the presence or absence of selected virulence or control genes. Seventy-two isolates were determined to be S aureus by standard identification methods and PCR detection of the S aureus–specific gyrA gene. The selected genes encoding α-hemolysin (hla), PVL (pvl), LukD-E (lukD-E), SEA (sea), SEB (seb), TSST (tst), ETA (eta), ETB (etb), and methicillin resistance (mecA) and chemotaxis-inhibiting protein (CHIP) (chs) were determined using real-time PCR with confirming melt curves and agarose gel visualization of selected DNA products for confirmation. The gene encoding S aureus DNA gyrase A (gyrA) was included as a positive control for PCR reactions. Oligonucleotide primer pairs used were either those used currently in our laboratory or were designed for this study using the Beacon Designer for Real-Time PCR Assay Design (Bio-Rad Laboratories, Hercules, California)
(eTable 1).9- 12 Individual reaction conditions were optimized for each oligonucleotide primer pair with gradient control templates to establish best reaction conditions. Real-time PCR reactions were carried out in a 96-well plate format using iCycler iQ PCR plates with Microseal “B” Film optically clear adhesives (Bio-Rad Laboratories). Thermal cycling was done in an iCycler, and detection was performed by an iCycler iQ Optical Module. Each unknown sample reaction plate contained duplicate samples and appropriate positive and negative controls for each toxin gene. The PCR reactions were done in 25-μL volumes containing iQ SYBR Green Supermix (Bio-Rad Laboratories), appropriate forward and reverse primers at a final 0.2μM concentration each with 1 to 10 ng of DNA template and nuclease-free water. Cycling reaction conditions were those optimized for each individual toxin primer pair. The real-time cycle threshold (Ct) curves were monitored, and data were collected at the end of each run. Corresponding melting temperature curves were determined for each sample for PCR product confirmation. The test for the presence of toxin genes was considered positive for an isolate with a Ct within 6 cycles of control Ct and an appropriate melting curve. Selected samples were further confirmed by agarose gel electrophoresis on 4% agarose gels. Control chromosomal DNA samples were obtained from our standard laboratory controls (CLP001-eta, CLP007-etb, and CPM003-seb) or strains provided by the Network on Microbial Resistance in S aureus (NRS384-pvl, -mecA, and -chs [CHIP]; NRS385-hla; NRS383-sea and -gyrA; and NRS178-lukD-E).
The prevalence of specific toxin genes associated with the 72 total S aureus isolates from infected skin lesions was calculated without regard to associated WBC counts determined at the time of culture. The prevalence of the toxin genes carried by the 61 bacterial isolates from the S aureus only–infected lesions was determined in the low (≤ 5 WBCs/LPF) and high (> 5 WBCs/LPF) WBC groups and compared using χ2 analysis. Significance was defined as P < .05.
Staphylococcus aureus, either exclusively or in combination with Streptococcus pyogenes (Lansfield group A streptococci [GAS]), group G streptococci (GGS), or group C streptococci (GCS) was isolated from 72 of 105 skin lesions determined to be infected by clinical evaluation (69%). Sixty-one of these lesions were determined to be infected with S aureus as a single causative organism (58%). White blood cell counts on the initial swab specimen taken for culture were determined per LPF for each lesion. Bacterial isolates were collected and correlated with the respective WBC count and separated into 2 groups, those associated with a low WBC count (0-5 WBCs/LPF) and those associated with a high WBC count (> 5 WBCs/LPF). The presence or absence of S aureus genes associated with virulence and with the potential to modulate the host's immune response were determined by real-time PCR using oligonucleotide primers specific for the genes of interest, and the results for all S aureus isolates are reported in
eTable 2. All 72 isolates were analyzed for the presence of S aureus virulence factor genes associated with skin infections without consideration of a localized immune response. The numbers of S aureus carrying the genes for the exfoliative toxins ETA and ETB and PVL were increased (P < .001 for ETA and ETB and P = .049 for PVL) compared with the expected community prevalence, while the numbers of isolates encoding TSST (P < .001) and SEA (P = .005) were significantly decreased (Table).13- 18 Coding for all other genes showed no significant difference in prevalence vs that previously reported for each. These data indicate that S aureus associated with skin lesions are more likely to carry the genes for ETA, ETB, and PVL and less likely to carry the genes for SEA and TSST.
To determine if an association exists between S aureus virulence factors and the local immune response as indicated by the number of WBCs in an infected skin lesion, the 61 bacterial isolates from S aureus only–infected lesions were divided into 2 groups, those associated with low WBC counts and those associated with high WBC counts and compared. Comparison of the presence of virulence genes carried by the bacteria isolated from the lesions in the low WBC vs the high WBC group demonstrated a difference in these bacterial populations (Figure 1A). Staphylococcus aureus capable of producing ETA (P = .006) and ETB (P = .002) were significantly associated with infections with low WBC counts in the lesions, while those organisms capable of producing PVL (P = .03) were associated with infections with high WBC counts on initial gram stain of swabs for culture. Analysis of the low WBC population alone (44 isolates) showed that the number of S aureus carrying the genes for ETA and ETB were significantly increased (P < .001 for both) in this subset, while the number of isolates encoding SEA (P = .03) were significantly decreased compared with the expected community prevalence Figure 1B. Similar analysis of the high WBC population alone (17 isolates) revealed no statistically significant differences in the prevalence of virulence factors in this population compared with the expected prevalence Figure 1C.
Comparisons of the clinical appearance of individual skin lesions associated with each WBC subset showed no apparent differences between the low and high WBC groups (Figure 2) despite the significant difference in associated infecting bacterial populations. Physical examination findings of all lesions were consistent with a significant bacterial skin infection, and so all lesions were selected for culture and gram staining. Of the 105 total lesions sampled, 21 (20%) failed to have any associated bacteria by culture. Of the 80% of bacteria-positive lesions, 85% grew S aureus either alone or in combination with a Streptococcus species. Overall, of the skin lesions identified clinically as infected, 69% (n = 72) were associated with growth of any S aureus, and 58% (n = 61) were infected exclusively with S aureus.
Staphylococcus aureus is a common cause of local skin infections as well as many other serious infections. Infections caused by S aureus have been linked to a number of potential virulence factors made by the organism. These associations range from superficial skin infections such as bullous impetigo, associated with ETA and ETB, to the multiple-organ system failure associated with toxic shock syndrome and the organisms capable of making TSST. Many of the toxins and virulence factors of S aureus are known to have effects on the immune response, either directly by lysing WBCs (as in the actions of leukotoxins such as PVL) or indirectly by the actions of superantigens (such as TSST and several enterotoxins) stimulating the release of abnormally high concentrations of cytokines and chemokines and potentially leading to multiple-organ system failure. Other virulence factors (eg, CHIP and Staphylococcus complement inhibitor13,19) function to prevent the influx of WBCs into an infected site.8
In the present study we describe for the first time to our knowledge a direct association between the number of WBCs in a local skin infection and the toxin genes that the infecting S aureus carry. In this population taken as a whole, and as has been previously reported,6,7 there is a significant association of exfoliative toxin genes and pvl with skin infections. In addition, we demonstrate for the first time to our knowledge that this association can be categorized into 2 distinct infection groups: those with few to no WBCs in the skin lesions at the time the bacterial cultures were obtained (≤ 5 WBCs/LPF) and those with many WBCs associated with the lesions. We determined that the lesions with the fewest WBCs were more likely to be infected with S aureus capable of producing ETA or ETB, while the lesions with the most WBCs were more likely to be infected with an organism capable of producing PVL.
Comparisons of the physical appearance of the lesions associated with each WBC subset showed no apparent differences between the low WBC and high WBC groups. The appearances of all lesions were determined to be consistent with significant infections when the initial lesions were chosen for obtaining swabs for culture, and therefore appearance alone is not a good indicator of inflammatory cell involvement. Common dogma holds that for a superficial bacterial skin infection to be deemed significant it must be associated with a high WBC count from the lesion at the time the bacterial cultures are obtained.20 Findings of the present study indicate that the particular bacterial population involved in an infection might contribute greatly to the severity and quality of the local immune response and that the population of infecting bacteria with its associated repertoire of potential virulence factors contributes to the robustness of the immune response.
The relationship between the genes encoding ETA and ETB and the lack of WBCs in some of the lesions is not understood, and the exact mechanisms by which these toxins may be acting to either eliminate WBCs from the lesions or prevent their accumulation is likewise not known. The action of the exfoliative toxins to cause exfoliation of the skin by the cleavage of the cadherin family member protein desmoglein 1,21 thereby disrupting the desmosome and destroying the cell-cell integrity in the upper epidermis, does not directly explain these findings. However, cadherin family member proteins are involved in a variety of cell signaling processes.22- 24 It is possible that the cleavage of desmoglein 1 is directly or indirectly involved in modifying the signaling pathways that affect the local immune response. It has also been shown that ETA is capable of stimulating a specific population of murine T cells,25 which could also contribute to cytokine release and affect the local immune response.
An additional possibility is that the associated toxin genes and their products are not involved at all but serve as genetic markers for other virulence factors that are responsible. Such a situation exists for PVL and its association with necrotizing infections: under certain conditions it is clearly a major virulence factor contributing to disease,26 while in other infections it serves as a genetic marker for other virulence factors or specific bacterial populations.27 Given that these bacteria have the potential to express a number of immunomodulating factors,8 we argue that the measurement of the local immune response (WBC count) is not always an accurate estimate of the seriousness of an infection—a significant infection can exist without local infiltration of large numbers of WBCs. These findings warrant further testing using well-defined bacterial populations and controlled animal models.
Corresponding Author: Lisa R. W. Plano, MD, PhD, University of Miami, Leonard M. Miller School of Medicine, Department of Microbiology and Immunology, RMSB 3066 (R-138), 1600 NW 10th Ave, Miami, FL 33136 (email@example.com)
Financial Disclosure: None reported.
Accepted for Publication: July 1, 2007.
Author Contributions: Ms Mertz and Dr Plano had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Mertz and Plano. Acquisition of data: Cardenas, Snyder, Kinney, Davis, and Plano. Analysis and interpretation of data: Mertz, Cardenas, Snyder, and Plano. Drafting of the manuscript: Mertz and Plano. Critical revision of the manuscript for important intellectual content: Davis and Plano. Statistical analysis: Cardenas and Plano. Obtained funding: Plano. Administrative, technical, or material support: Cardenas. Study supervision: Mertz, Davis, and Plano.
Funding/Support: This study was supported in part by National Institutes of Health grant AR048882 and the National Institute of Arthritis and Musculoskeletal and Skin Disease grant R01-AR04882 (Dr Plano).
Additional Contributions: Yvette Piovanetti, MD, Maria D. Ordriozola, MD, Kamara Mertz-Rivera, MS, CCRC, and Juan B Rivera, AA, collected the clinical samples.