|Home | About | Journals | Submit | Contact Us | Français|
Mice were infected with Mycoplasma pneumoniae and monitored for the synthesis and distribution of the unique adenosine diphosphate–ribosylating and vacuolating Community Acquired Respiratory Distress Syndrome (CARDS) toxin in bronchiolar lavage fluid (BALF) and lung. We noted direct relationships between the concentration of CARDS toxin and numbers of mycoplasma genomes in BALF and the degree of histologic pulmonary inflammation. Immunostaining of lungs revealed extensive colonization by mycoplasmas, including the detection of CARDS toxin in the corresponding inflamed airways. Lung lesion scores were higher during the early stages of infection, decreased gradually by day 14 postinfection, and reached substantially lower values at day 35. Infected mouse immunoglobulin (Ig) M and IgG titers were positive for CARDS toxin as well as for the major adhesin P1 of M. pneumoniae. These data reinforce the proposed pathogenic role of CARDS toxin in M. pneumoniae–mediated pathologies.
Mycoplasma pneumoniae is an atypical bacterium that causes acute respiratory illnesses in humans, including pharyngitis, tracheobronchitis, and community-acquired pneumonia [1–3]. It also has been directly linked to reactive airway disease and asthma [4–8] and extrapulmonary manifestations [2, 9, 10]. M. pneumoniae has been isolated from the respiratory tract of up to 20%–25% of asthmatics experiencing acute exacerbations [6, 11]. While the wide-ranging clinical significance of M. pneumoniae infection is becoming more evident, the mechanisms by which mycoplasma-mediated host cell injury occurs in the respiratory tract remain unclear. Over the years the pathogenic potential of M. pneumoniae has been demonstrated in tracheal organ cultures and hamster animal models [12–15]. Our earlier reports described the specific attachment of virulent M. pneumoniae via a constellation of mycoplasma tip organelle-associated proteins to sialic acid–associated receptors on the respiratory epithelium and via other mycoplasma surface proteins that mediate binding to extracellular matrix proteins, like fibronectin and surfactant protein A [16–20]. We showed that viable and attached virulent mycoplasmas elicited abnormal host cell reactions at transcriptional and translational levels, with subsequent interruption of host metabolic pathways and generation of tissue cytopathology [13, 21]. In addition, microbiologic and histologic findings of experimental murine M. pneumoniae pneumonia have been detailed [22–26].
Using hamster tracheal organ cultures and hamster and murine animal models, we suggested that unidentified virulence factor(s) associated only with viable mycoplasmas mediates host cell injury [13, 21, 22, 27, 28]. Recently, we identified a novel M. pneumoniae cell–associated adenosine diphosphate–ribosylating and vacuolating cytotoxin, designated the Community Acquired Respiratory Distress Syndrome (CARDS) toxin, which alone reproduced the characteristic ciliostasis, cytoplasmic and nuclear vacuolization, and extensive respiratory epithelial cell fragmentation and sloughing  that had been observed in M. pneumoniae–infected tracheal organ cultures [12–15].
In order to understand the possible relationship of CARDS toxin in M. pneumoniae–mediated disease progression, we established specific CARDS toxin assays to enable detection and localization of mycoplasmas and CARDS toxin in biological fluids. Specifically, we developed a toxin antigen capture assay to quantify CARDS toxin protein levels in tissue samples ; a CARDS toxin gene-based quantitative real-time polymerase chain reaction (qPCR) assay to enable quantification of M. pneumoniae genomes; and immunostaining methodology that permitted identification and localization of mycoplasmas and CARDS toxin in the lungs. This report focuses on CARDS toxin–related events that for the first time to our knowledge provide fundamental insights concerning the synthesis and distribution of this unique toxin during M. pneumoniae airway infection.
M. pneumoniae strain M129 (ATCC 29342) was grown in SP4 broth at 37°C for 72 hours and concentrated in 2 mL fresh SP4 to 7–8 log10 colony-forming units (CFU) per mL.
Two-month-old female BALB/c mice were intranasally (IN) infected once (day 0) with 5.9–6.2 log10 CFU of M. pneumoniae in 50 μL of SP4 broth. Control mice were inoculated with sterile SP4 medium. Mycoplasma and murine virus–free mice (Charles River and Harlan) were housed in filter-top cages and allowed to acclimate to their new environment for 1 week. Animal guidelines were followed in accordance with the Institutional Animal Care and Research Advisory Committee at the University of Texas Southwestern Medical Center at Dallas.
Mouse tissue samples were obtained at 1, 4, 7, 14, and 35 days postinfection (PI). At each time point, 6 infected and 6 uninfected control mice were sacrificed for bronchiolar lavage fluid (BALF; 0.5 ml), serum samples, and lung specimens . Whole-lung specimens, including trachea and both lungs, were then collected and fixed with 10% neutral buffered formalin solution for histologic evaluation. Following fixation, lungs from each animal were cut coronally and processed for paraffin embedment. Sections were prepared at 5 μm thickness and stained with hematoxylin and eosin (H&E). Two control and 4 additional infected mice were sacrificed at 4, 7, and 14 days, and the lungs were air inflated and frozen in liquid nitrogen. Cryosections from these lungs were cut at 5 μm, fixed in acetone, and stained using CD4 and CD19 biotinylated antibodies (1:25; BD Pharmingen) with avidin-biotin–blocking reagents, streptavidin-horseradish peroxidase conjugate, and diaminobenzidine (DAB) chromogen (Vector Laboratories). Rabbit recombinant CARDS (rCARDS) toxin antibodies and rabbit whole-cell M. pneumoniae antibodies at 1:1000 and 1:1500 dilutions, respectively, were incubated with representative lung sections, which were then stained with DAB chromogen. Histopathological findings and grading of lung lesions were performed by a pathologist (J. J. C.), who was unaware of the infection status of animals from which specimens were taken. Lesions of peribronchiolar and bronchial infiltrates, bronchiolar and bronchial luminal exudates, perivascular infiltrate, and parenchymal pneumonia were evaluated . This method assigns values from 0 to 26 (the greater the score, the greater the inflammatory changes in the lung). Inflammatory cell infiltrates of lymphocytes and polymorphonuclear leukocytes were graded at few (grade 1), numerous (grade 2), or abundant (grade 3) in peribronchial/peribronchiolar, perivascular, and intra-alveolar pneumonitic exudate sites.
M. pneumoniae cells were quantified in SP4 cultures and BALF by CFU and qPCR. In the latter case, 2 separate duplex assays (targeting CARDS toxin [MPN372] and P1 adhesin [MPN141] gene sequences) were compared. Primers and TaqMan probe for detection of CARDS gene were designed; those for P1 were described previously . Standard curves were established using M. pneumoniae chromosomal DNA serially diluted from 107 to 5 copies per reaction.
Since CARDS toxin is expressed at low levels in mycoplasma broth cultures , we used purified rCARDS toxin for comparative serological studies with the recombinant immunodominant COOH epitope of adhesin P1 (rP1) . As described before , proteins were expressed and purified by nickel-affinity column chromatography and desalted in 50 mmol/L Tris pH 7.4.
CARDS toxin capture assay was performed as reported . Extracted BALF samples were mixed with 1% bovine serum albumin/phosphate-buffered saline with 0.05% Tween (BSA/PBS-T) to a final volume of 50 μL and added to single wells. In order to quantify the amount of CARDS toxin in a given sample, known amounts of purified rCARDS toxin (ranging from 7 ng to 7 fg per well) were diluted in 1% BSA/PBS-T to establish a standard curve. To determine the sensitivity of the CARDS toxin capture assay in the presence of BALF, we measured toxin concentrations diluted in BALF derived from uninfected mice, which were compared with the standard curve.
Immunoglobulin (Ig) M and IgG antibodies to M. pneumoniae CARDS toxin and adhesin P1 in infected mice were determined by enzyme-linked immunosorbent assay. Ninety-six well microtiter plates were individually coated with 50 μL/well of equimolar concentrations of recombinant proteins (P1 or CARDS toxin) in carbonate-bicarbonate buffer (pH 9.6) overnight at 4°C and washed 3 times with PBS. Plates were incubated 2 hours at room temperature with individual mouse serum samples (50 μL/well) at 1:200 dilution in 1% BSA/PBS, and experiments were performed as reported . Values of uninfected control mouse sera were compared with test samples.
Sigma Stat 2003 software (SPSS Science) was used. If data were normally distributed, the t test compared values of different groups of animals at identical points. When data were not normally distributed, the Mann–Whitney rank-sum test was applied for comparisons. Bonferroni correction was employed where multiple comparisons were made. The Spearman rank-order test was used for correlations, as all data taken together were not normally distributed. A comparison was considered statistically significant if the P value was < .05.
The number of M. pneumoniae cells in BALF was estimated by CFU and real-time PCR. BALF was culture-positive in 100% of IN-infected mice up to 14 days PI and in 75% of mice at 35 days PI. Mean titers of positive BALF cultures were 5.9 log10 CFU/mL during the first 4 days of infection and 5.3 log10 CFU/mL at day 7 PI. At days 14 and 35 PI, mean CFU titers decreased to 2.95 and 1.95 log10 CFU/mL, respectively. All control mice had negative BALF cultures. CARDS toxin real-time PCR values paralleled CFU throughout the experiment (Figure 1). In addition, real-time PCR was 100% positive in infected mice and negative in uninfected mice. To further analyze the sensitivity and specificity of the CARDS toxin qPCR assay, we compared its efficiency with adhesin P1–based qPCR and found that both PCR targets were equally specific (data not shown).
Using the CARDS toxin capture assay, we detected M. pneumoniae–synthesized CARDS toxin in BALF, with highest levels observed at day 1 PI (384 ± 70 pg/mL; Figure 1) and lower levels at days 4 and 7 PI (~293 ± 42 and 303 ± 37 pg/mL, respectively). At days 14 and 35 PI, CARDS toxin values were still detectable but at markedly reduced levels (4 ± 1.54 and 0.18 ± 0.1 pg/mL, respectively). We observed direct correlation between the microbial loads and toxin concentration in BALF, which is consistent with our recent comparison of clinical strains of M. pneumoniae .
Histopathological scores (HPS) using the Cimoli et al  methodology are shown in Figure 2A. Temporal shifts in lymphocyte and neutrophil populations at the different study times over the 35-day infectious time frame are not well defined using this grading method. As indicated in the Methods section, we used an alternative grading system to define the lymphocytes and neutrophil populations in the infected animals (Figure 2B).
H&E-stained lung specimens from sterile SP4 broth IN-inoculated control mice were normal in appearance at all time points. In infected mice at day 1 PI, bronchi and bronchioles were lined by intact respiratory epithelium, but ulceration was identified at terminal and respiratory bronchiole junctions (Figure 3A, arrows). A mixed population of lymphocytes and polymorphonuclear (PMN) cells was evident in the walls of pulmonary vessels and airways, with numerous intra-alveolar neutrophils and macrophages present in the pneumonitic exudates in the alveolar spaces (Figure 3A). The graded cell findings reflected in Figure 2B support the rapid recruitment of both lymphocytic and PMN cells into the infected sites. By day 4 PI, PMN recruitment had waned in the perivascular spaces, but the PMN exudate in the pneumonitic response was at its peak (Figure 2B). Multilayered collections of small lymphocytes with thin rims of cytoplasm and abundant nuclear chromatin that encircled large and small airways and pulmonary arteries and veins also peaked in number at 4 days PI (Figure 3B). A consistent finding at 4 days PI was the distinct formation of separate aggregates of 2 subsets of lymphocytes (CD4 and CD19) in the inflamed foci (Figure 4A and and4B).4B). At day 7 PI, sites of pneumonia had largely disappeared, but lymphocytes remained elevated in number (Figure 2B) and concentrated around perivascular spaces, with fewer numbers in the walls of airways (Figure 3C). Many lymphocytes were larger in size and displayed abundant cytoplasm and enlarged nuclei, that is, more blastic in appearance (Figure 3C). These reactive lymphocytes were focally embedded in infiltrates of small lymphocytes. Multiple and comparable microscopic sites of these different lymphocyte phenotypes were photographed on serial or sequentially cut sections immunostained with CD4 and CD19. Although some foci appeared to contain more cells of 1 phenotype, overall, CD4+ T cells and CD19+ B cells were both evident in similar numbers at all 3 time points. At day 14 PI, lymphoid infiltrates in airway walls were decreased in number (Figure 2B), whereas those in the perivascular spaces were persistently higher in number throughout the 35-day study period (Figure 2B). At day 35 PI, perivascular lymphoid aggregates were infrequent and showed depleted numbers of lymphocytes and occasional collections of pigmented mononuclear cells (Figures 2B and and3D3D).
To localize M. pneumoniae cells during infection, we initially used rabbit antisera raised against whole-cell lysates of broth-grown M. pneumoniae. We identified mycoplasmas on the surfaces of respiratory epithelia at day 4 PI (Figure 5B, arrows) and not in uninfected lung tissue (Figure 5A). However, we also noted background staining below epithelial cell surfaces in both infected and uninfected lung tissue (Figure 5A and 5B). This staining pattern is likely due to the presence of antibodies to mycoplasma membrane protein epitopes that cross-react with mammalian proteins, such as fibrinogen, keratin, myosin, and collagen . Importantly, when we used highly purified anti–CARDS toxin antibodies, no background labeling occurred (Figure 5C), and CARDS toxin, along with M. pneumoniae organisms, was readily evident on respiratory epithelium surfaces of infected lungs (Figure 5D).
Closer inspection of infected lung tissues at day 4 PI revealed mycoplasmas (Figure 6A) and CARDS toxin (Figure 6B) in peribronchiolar alveolar spaces that contained inflammatory exudates of edema, neutrophils, and alveolar macrophages/monocytes. Notably, lung epithelial cells associated with CARDS toxin distribution showed cellular damage and focal loss of cilia. Alveolar walls in these sites revealed increased interstitial cellularity, but necrosis was not observed. Some mycoplasmas appeared in close proximity to both neutrophils and alveolar macrophages/monocytes (Figure 6A and 6B, arrows). The number of mycoplasmas, abundance of CARDS toxin, and co-localization of mycoplasmas and toxin within alveolar spaces and inflammatory cells are clearly discernible (Figure 6A and and6B,6B, arrows).
At days 1 and 4 PI, we observed little to no detectable levels of antibodies reactive against CARDS toxin or adhesin P1 when compared with SP4 broth–inoculated, uninfected controls. However, serum IgM levels against CARDS toxin and P1 peaked at day 7 in 75% of infected mice and approached background values at 35 days PI (Figure 7). IgG titers to CARDS toxin and P1 increased significantly between days 7 and 35 PI (Figure 7).
Using a murine M. pneumoniae acute pulmonary infection model, we monitored the synthesis and distribution of CARDS toxin following M. pneumoniae IN infection over a 35-day period. We readily detected mycoplasmas and CARDS toxin in BALF during the initial stages of M. pneumoniae colonization of the airways (Figure 1). We observed a direct relationship between numbers of mycoplasmas, amounts of CARDS toxin, and lung histopathology using CFU, real-time PCR, antigen capture assay, and HPS (Figures 1–3). Immunostaining demonstrated CD4 and CD19 positively stained subsets of lymphocytes in inflamed foci (Figure 4) and localization of mycoplasmas and CARDS toxin with respiratory epithelial cells and macrophages (Figures 5 and and6).6). Also, the striking anti–CARDS toxin IgM and IgG seroconversion not only reinforced the synthesis and detection of CARDS toxin during airway infection but also established its ability to elicit a strong immunogenic response (Figure 7) consistent with observations in humans .
During the course of murine IN infection, M. pneumoniae CFU in BALF were highest at days 1–4 PI (5.9 log10 CFU/mL) and decreased to ~1.95 log10 CFU/mL at 35 days PI (Figure 1). Similarly, CARDS toxin concentrations in BALF were highest at day 1 PI and remained near this level during days 4 and 7 PI, indicating that CARDS toxin synthesis and stability are relatively steady and sustainable during acute stages of infection. However, toxin concentrations declined dramatically at 14 and 35 days PI, which paralleled reduced mycoplasma cell loads, suggesting that CARDS toxin may serve as a surrogate marker to track active M. pneumoniae infection and disease progression. Interestingly, and in contrast to these in vivo dynamics, the expression of CARDS toxin in broth-grown mycoplasmas falls precipitously during the first 24–72 hours (during early-to-late log phases) and remains at very low levels throughout in vitro growth . This suggests that toxin expression is tightly regulated by host conditions that signal transcriptional activation and synthesis of toxin, consistent with the important role that CARDS toxin plays during infection.
Another fundamental finding was the apparent linkage between mycoplasma genome numbers, CARDS toxin levels, and degree of histologic lung inflammation, which correlates well with our recent report in which we compared 3 different strains of M. pneumoniae and observed a direct link between load of mycoplasma, CARDS toxin levels, and cytokine responses . M. pneumoniae infection induced an inflammatory phase that lasted ~3 weeks, with peak inflammatory changes occurring at day 4 PI, consistent with early CARDS toxin levels in BALF. Relatively low titers of mycoplasmas persisted, along with diminished amounts of CARDS toxin and pulmonary histologic inflammation at days 14 and 35 PI (Figures 1–3). Interestingly, mice that received IN rCARDS toxin alone elicited cellular inflammation similar to M. pneumoniae infection, further suggesting the central role of CARDS toxin in pulmonary inflammation . However, the induction of additional cytokines in response to M. pneumoniae infection in BALF, when compared with CARDS toxin alone [22, 26, 34, 35], suggests that other M. pneumoniae factors contribute to the inflammatory process.
Detailed microscopic analysis of lung sections confirmed the development of pneumonia, alveolar wall and space edema, inflammatory cell infiltrates, early and persistent lymphocyte-dominant (CD4+ T and CD19+ B cells [Figure 4]) peribronchial and perivascular regions, and immunostained mycoplasmas plus CARDS toxin in both airway and alveolar compartments (Figures 2–6). The early neutrophilic pneumonitis induced in this experimental infection model is accompanied by a robust and simultaneous infiltration of perivascular lymphocytes, as we reported with rCARDS toxin alone . Consistent features of M. pneumoniae infection in humans are bronchiolar luminal purulent exudates, lymphoplasmacytic infiltrate in bronchiolar walls, peribronchiolar septal widening, and alveolar type 2 hyperplasia . In our mouse model, these histopathological features are prominent. Additionally, the pneumonitis evident in alveoli subjacent to infected bronchioles, which occurred primarily in the first week, along with the subsequent retention of the prominent perivascular lymphocytic infiltrates over time, indicates that temporal changes occur in histopathology as infection and intoxication events evolve (Figure 3A–D). In contrast to mycoplasmal pneumonia, the dominant feature in most bacterial pneumonias is bronchiolar and alveolar filling with purulent exudates, and bronchiolar lymphocytic infiltration is not a prominent or diagnostic feature . M. pneumoniae is well known for its ability to act as a polyclonal activator of lymphocytes and autoantibodies [38, 39]. B cells, CD4+ T cells, and plasma cells infiltrate the lungs, which is followed by further augmentation of the immune response, namely, abundance of lymphocytes, production of immunoglobulins, and release of proinflammatory cytokines [40, 41]. We observed considerable numbers of CD4+ and CD19+ cells around the site of inflammation (Figure 4). Previous studies have shown that the percentage of CD4+ and CD19+ cells varies during M. pneumoniae infection. It was suggested that the redistribution of CD4+ T cells to the site of infection may explain the decreased proportion of these cells in the blood [40, 42, 43]. However, further studies are needed to determine how M. pneumoniae and CARDS toxin alter the immune system and influence the concentration and distribution of CD4 and CD19 cells.
While we observed peribronchial and perivascular infiltration in lungs of M. pneumoniae–infected mice, even at day 35, we speculate that mycoplasmas persist and continue to synthesize low levels of CARDS toxin, resulting in chronic infection and inflammation that lead to other acute and chronic airway diseases, like asthma, and extrapulmonary manifestations. In humans, M. pneumoniae is reported to persist in the respiratory tract for many months, even after therapy with appropriate antibiotics [44–48]. We have observed similar persistence of M. pneumoniae in the mouse model and in mammalian cell lines after prolonged antibiotic treatment [24, 49]. It will be important to investigate the chronicity of M. pneumoniae infection in our mouse model, along with CARDS toxin expression and stability, especially since chronic infections have been postulated to lead to a range of pathologies accompanied by dysfunctional changes and tissue remodeling in the respiratory tract . Clearly, M. pneumoniae cells and CARDS toxin localize to the ciliated respiratory epithelium (Figure 5) and are also associated with lung macrophages (Figure 6). The binding of CARDS toxin and mycoplasma cells selectively to surfactant protein A probably facilitates these interactions [20, 30].
Seroconversion against CARDS toxin in experimentally infected mice (Figure 7) parallels what is observed during human M. pneumoniae infections . We also noted in M. pneumoniae–infected mice that antibody titers against CARDS toxin are at least equivalent to the “immunodominant” M. pneumoniae adhesin P1, suggesting that CARDS toxin may be used alone, or along with P1, for diagnostic and prognostic applications. Also, these observations suggest that CARDS toxin could serve as an effective vaccine candidate.
In summary, our studies provide unequivocal evidence that CARDS toxin is synthesized during experimental infection, which directly correlates with mycoplasma replication and persistence, airway inflammation, and lung histopathology [29, 30, 35]. We also demonstrate that CARDS toxin expression colocalizes with mycoplasma cells that colonize respiratory epithelial cell surfaces. Further, we present additional evidence for the immunodominant and diagnostic properties of CARDS toxin. Other studies are in progress to advance our understanding of the mechanisms that govern CARDS toxin expression and its mode of action in vivo, which should aid in the development of preventative therapies for the spectrum of acute and chronic airway and extrapulmonary diseases associated with this common respiratory bacterial pathogen.
We thank Vicki T. Winter and Shellye R. Lampkin of the Department of Pathology for their technical assistance and Rose Garza for her assistance in assembling the manuscript.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.
This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (U19AI070412); and the Kleberg Foundation.
All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.