Differential modulation of pulmonary caspases: Is this the key to Ureaplasma-driven chronic inflammation?

Although accepted agents in chorioamnionitis and preterm birth, the role of Ureaplasma species (spp.) in inflammation-driven morbidities of prematurity, including the development of bronchopulmonary dysplasia, remains controversial. To add to scarce in vitro data addressing the pro-inflammatory capacity of Ureaplasma spp., pulmonary epithelial-like A549 cells and human pulmonary microvascular endothelial cells (HPMEC) were incubated with Ureaplasma (U.) urealyticum, U. parvum, and Escherichia coli lipopolysaccharide (LPS). Ureaplasma isolates down-regulated caspase mRNA levels in A549 cells (caspase 8: p<0.001, 9: p<0.001, vs. broth), while increasing caspase protein expression, enzyme activity, and cell death in HPMEC (active caspase 3: p<0.05, caspase 8: p<0.05, active caspase 9: p<0.05, viability: p<0.05). LPS, contrarily, induced caspase mRNA expression in HPMEC (caspase 3: p<0.01, 4: p<0.001, 5: p<0.001, 8: p<0.001, vs. control), but not in A549 cells, and did not affect enzyme activity or protein levels in either cell line. LPS, but neither Ureaplasma isolate, enhanced mRNA expression of pro-inflammatory interleukin (IL)-6 in both A549 (p<0.05, vs. control) and HPMEC (p<0.001) as well as tumor necrosis factor-α (p<0.01), IL-1β (p<0.001), and IL-8 (p<0.05) in HPMEC. We are therefore the first to demonstrate a differential modulation of pulmonary caspases by Ureaplasma spp. in vitro. Ureaplasma-driven enhanced protein expression and activity of caspases in pulmonary endothelial cells result in cell death and may cause structural damage. Down-regulated caspase mRNA in pulmonary epithelial cells, contrarily, may indicate Ureaplasma-induced inhibition of apoptosis and prevent effective immune responses. Both may ultimately contribute to chronic Ureaplasma colonization and long-term pulmonary inflammation.


Introduction
Ureaplasma species (spp.) commonly colonize the adult urogenital tract and are generally considered of low virulence [1]. Transmission from mother to infant is frequent and can occur in utero, intrapartum, or postpartum [1]. Intraamniotic Ureaplasma infection is an accepted risk factor for chorioamnionitis and premature birth [2][3][4], and Ureaplasma spp. are known to a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 cause sepsis, meningitis, and pneumonia in neonates [5][6][7][8] as well as severe invasive infections in immunocompromised adults such as lung transplant patients [9]. Ureaplasma spp. can be detected in the respiratory tract in 65% of preterm infants < 26 weeks of gestation [10]. Fetal or neonatal respiratory tract colonization with Ureaplasma spp. has been associated with bronchopulmonary inflammation and altered lung development, which may ultimately culminate in chronic lung diseases such as bronchopulmonary dysplasia (BPD) in preterm infants [11][12][13]. Inflammation is considered a key factor in the multifactorial pathogenesis of BPD development [13,14]. Animal models support a potential causality between Ureaplasma spp. and development of BPD, demonstrating pulmonary inflammation accompanied by structural lung tissue impairment upon fetal Ureaplasma exposure [15,16]. Clinical studies, however, are contradictory [17], and in vitro data on the pro-inflammatory capacity of Ureaplasma spp. are generally scarce.
In pulmonary inflammation, lung epithelial and endothelial cells both deserve attention. They usually maintain intrapulmonary homeostasis as well as an immunological balance [18,19]. Pulmonary epithelial cells maintain the air-blood barrier and comprise alveolar type I and II cells [20]. While type I cells primarily enable gas exchange, type II cells produce surfactant and are crucial for tissue repair [19]. Opposed to epithelial cells, lung endothelial cells are more permeable [21] and contribute to inflammatory processes by signal transduction and initiation of inflammatory cell migration into the alveolar space [19,22].
Inflammation is classically understood to be initiated by early pro-inflammatory cytokines, including tumor necrosis factor (TNF)-α, interleukin (IL)-1β, IL-6, and IL-8 [23]. However, recent studies have suggested an interlinkage of apoptosis and inflammation, proposing a crucial role for caspases in both [24,25]. Whereas caspases 3, 8, and 9 are among those primarily involved in apoptosis, caspases 4 and 5 mediate inflammatory responses and pyroptosis as an inflammatory form of cell death [26,27]. Regulation of caspase activity is complex and involves the transcription and translation processes as well as specific cleavage and activation of synthesized inactive caspase proenzymes [27].
We could recently demonstrate caspase modulation and a pro-apoptotic capacity of Ureaplasma spp. in human brain microvascular endothelial cells (HBMEC) [28]. In the present study, we addressed Ureaplasma-induced caspase expression as well as cytokine responses in the well described epithelial type II cell-like line A549 and human pulmonary microvascular endothelial cell line HPMEC-ST1.6R (HPMEC) [19,[29][30][31].
For stimulation assays, 3.5×10 5 A549 cells and 2.5×10 5 HPMEC per well were seeded in uncoated six well plates (Greiner, Frickenhausen, Germany). 24 h later, cells were stimulated as indicated in 1 mL fresh growth medium without antibiotics.

Stimulation assays
Ureaplasma suspensions of 10 9 −10 10 CCU/mL were concentrated by centrifugation, and 10 9 −10 10 CCU in 250 μl broth were added to each well. CCU were determined by 10-fold titration, as described elsewhere [33], and corresponding genome equivalents were identified, amounting to 5×10 7 −6×10 8 copy numbers/mL (Institute of Medical Microbiology and Hospital Hygiene, Duesseldorf, Germany) [34]. Viability was verified by simultaneous culture on selective agar plates (medco Diagnostika GmbH, Ottobrunn, Germany). Lipopolysaccharide (LPS, Escherichia coli serotype 055:B5, Sigma-Aldrich) was added to a subgroup of cells at a concentration of 100 ng/mL. All doses had been determined by preliminary assays analogous to previous approaches [33,[35][36][37][38][39][40]. According to results of preliminary time kinetic experiments [35], exposure times of 4 and 30 h were chosen for RNA analysis, whereas flow cytometry was performed after 24 h. Experiments were repeated at least 3 times (n � 3). Unstimulated cells served as negative controls. To adjust for potential confounding effects of Ureaplasma medium, cells exposed to Ureaplasma isolates were additionally compared to broth control and results were considered relevant if comparisons to both negative controls showed p values < 0.05.

RNA extraction and RT-PCR
For RNA extraction, cells were treated as indicated and total RNA was isolated using NucleoSpin RNA Kit (Macherey-Nagel, Dueren, Germany). For quantification of total RNA, a Qubit 2.0 Fluorometer (Thermo Fisher Scientific) was employed. Total RNA was eluted in 60 μL RNase-free H 2 O (Macherey-Nagel) and stored at -80˚C until reverse transcription. For RT-PCR, 0.5-1 μg of total RNA was reverse transcribed using High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). First strand cDNA was diluted 1 to 10 with deionized, nuclease-free H 2 O (Sigma-Aldrich) and stored at -20˚C upon analysis.

Quantitative real time RT-PCR (qRT-PCR)
For quantitative detection of mRNA, 10 μL of diluted first strand cDNA were analyzed in duplicates of 25 μL reactions using 12.5 μL iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA), 0.5 μL deionized H 2 O, and 1 μL of a 10 μM solution of forward and reverse primers (Sigma-Aldrich) as indicated in Table 1. Primers were designed using PerlPrimer software v1.1.20 [41]. Amplicons were designed to be 100-200 bp in size. A BLAST search was performed for every primer to confirm specificity with E values < 1. At least one primer of each pair spanned an intron/exon boundary to ensure mRNA-specific amplification. PCRs were performed on an Applied Biosystems 7500 Real-Time PCR System (Thermo Fisher Scientific), using a 2-step PCR protocol after an initial denaturation at 95˚C for 10 min with 40 cycles of 95˚C for 15 s and 60˚C for 1 min. A melt curve analysis was performed at the end of every run to verify single PCR products. Levels of mRNA were normalized to those of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Mean fold changes in mRNA expression were calculated by the ΔΔC T method by Livak and Schmittgen [42].

Flow cytometry
Harvested cells were separated by centrifugation and stained with Fixable Viability Dye eFluor 780 (eBioScience, San Diego, CA). Centrifugation and resuspension in Phosphate Buffered Saline (PBS, Sigma-Aldrich) were followed by fixation (fixation buffer, BioLegend). After another centrifugation step, cells were permeabilized (permeabilization wash buffer, BioLegend) and stained with antibodies to cleaved caspase 3 (Alexa Fluor 647 conjugated, rabbit, Cell Signaling Technology, Danvers, MA), caspase 8 (unconjugated, mouse, Abcam, Cambridge, UK), and cleaved caspase 9 (PE conjugated, rabbit, Cell Signaling Technology). To detect the unconjugated caspase 8 antibody, cells were once more separated by centrifugation and stained with an Alexa Fluor 405 conjugated secondary antibody (goat anti-mouse, Life Technologies, Thermo Fisher Scientific). After centrifugation and resuspension of the cells in PBS containing 1% human serum (Biochrom), samples were read on a FACSCanto II flow cytometer (BD Biosciences, San Jose, CA). At least 10,000 events were acquired and analyzed with FACSDiva v6.1.3 software (BD Biosciences). Events were gated via forward and side scatter, and doublets were excluded employing a SSC-width versus FSC-area dot plot (the gating strategy is described in S1 Fig). For viability assessment, all events were included and the percentage of viability dye positive cells, considered dead, was determined.

Statistical analysis
Data were analyzed by a one way ANOVA with Tukey's multiple comparisons test using Prism 6 software (GraphPad Software, San Diego, CA). The significance threshold for p values was set at < 0.05. Data are shown as means ± standard deviation (SD).

Ureaplasma-driven cell death in pulmonary epithelial and endothelial cells
Flow cytometry was used to identify viability dye positive, dead cells. In A549 cells, we generally counted low numbers of dead cells, without any significant influences on cell viability yielded by Ureaplasma spp. (Fig 1A). Exposure of HPMEC to both Ureaplasma isolates, however, caused a distinct increase in numbers of dead cells after 24 h (Fig 1B), which was significant for serovar 8 (2.5 ± 0.3-fold, p = 0.0347, vs. broth). Broth itself induced a mild increase in dead cells. LPS did not affect viability of A549 cells or HPMEC (Fig 1).

Ureaplasma-driven caspase responses in pulmonary endothelial cells
With the exception of a mild reduction of caspase 9 mRNA upon 30 h of serovar 8 exposure (0.7 ± 0.2-fold, p = 0.0374, vs. broth), Ureaplasma isolates did not influence caspase mRNA levels in HPMEC (Fig 3A-3E). Variations in comparison to unstimulated control cells were not significant compared to broth control, since broth itself had moderate and inconsistent effects on caspase mRNA expression. LPS stimulation of HPMEC for 4 h did not evoke any Data are shown as means ± SD and were obtained from n � 3 independent experiments. Cells stimulated with LPS were compared vs. control, cells exposed to Ureaplasma isolates vs. control and vs. broth. � p < 0.05, �� p < 0.01, and ��� p < 0.001 compared to untreated controls; † p < 0.05, † † p < 0.01, and † † † p < 0.001 compared to cells treated with broth. SV8: Ureaplasma urealyticum serovar 8, SV3: Ureaplasma parvum serovar 3.

Ureaplasma-driven cytokine responses in pulmonary epithelial and endothelial cells
In A549 cells, no significant TNF-α, IL-1β, IL-6, and IL-8 mRNA responses were measurable upon Ureaplasma exposure if compared to broth and control, since broth itself had a certain cytokine-inducing effect (Fig 4A-4D). Exposure of A549 cells to LPS for 4 h induced a significant increase of IL-6 mRNA (2.1 ± 0.6-fold, p = 0.0430, vs. control), whereas LPS did not influence mRNA levels of TNF-α, IL-1β, and IL-8, regardless of the duration of stimulation ( Fig  4A-4D).

Caspase and cytokine responses upon co-stimulation of pulmonary epithelial and endothelial cells
Co-stimulation of A549 cells or HPMEC with LPS and Ureaplasma isolates did not significantly aggravate effects observed after mono-stimulation with one or the other stimulus regarding caspase 3, 4, 5, 8, and 9 mRNA expression, cleaved caspase 3, caspase 8, and cleaved caspase 9 protein levels, or TNF-α, IL-1β, IL-6, and IL-8 mRNA levels (S2 Fig-S4 Fig).

Discussion
This is the first in vitro study assessing influences of Ureaplasma spp. on pulmonary epi-and endothelial cells. Results indicate Ureaplasma-driven suppression of apoptosis in pulmonary epithelial cells and a simultaneous pro-apoptotic capacity of Ureaplasma spp. in pulmonary endothelial cells. Influencing caspase levels and activity may be one of the mechanisms Ureaplasma spp. employ to relevantly interfere with immunological processes.
We could demonstrate that exposure of A549 cells to Ureaplasma spp. resulted in a distinct reduction of caspase mRNA and lacking protein responses (Fig 2), whereas in HPMEC, Ureaplasma isolates did not influence mRNA expression, but promoted caspase activation and protein increase (Fig 3). These divergent results regarding mRNA and protein levels can be explained by the meticulous multistep process in caspase protein production and activation ( Fig 5A) [27,43]. The Ureaplasma-induced caspase mRNA suppression we could demonstrate in A549 cells may result in impaired protein production. Consequently, we observed no relevant pathogen-induced caspase protein increase in A549 cells. Caspase mRNA synthesis in HPMEC appeared mostly unaffected by Ureaplasma-stimulation. These cells, consequently, were able to exercise an unconfined caspase response regarding protein abundance and enzyme activity.

Ureaplasma-driven modulation of pulmonary caspases
Ureaplasma-associated down-regulation of caspase 8 and 9 mRNA in A549 cells and consecutively absent protein responses may therefore suppress apoptosis (Fig 5B). Apoptosis as an immune defense mechanism is particularly relevant in elimination of intracellular pathogens [45]. Given the ability of Ureaplasma spp. to invade host cells [46,47], down-regulation of caspases may function as an escape mechanism Ureaplasma spp. use to prevent eradication. In line with this hypothesis, we did not observe an impact of Ureaplasma isolates on cell viability in A549 cells (Fig 1). Impaired apoptosis and subsequently reduced pathogen eradication may ultimately facilitate chronic Ureaplasma colonization. Ureaplasma spp. are known to cause long-lasting pulmonary infections [48], and BPD-development in preterm infants, to name Under engagement of several additional proteins not mentioned here, initiator caspases 4 and 5 for pyroptosis or 8 and 9 for apoptosis are produced and activated. These subsequently activate effector caspases 1 or 3. Caspases are activated by cleavage or dimerization, often followed by a maturational process. Control mechanisms confine programmed cell death [26,27,[43][44][45]. The potential influence of Ureaplasma spp. on apoptosis according to our results is illustrated in (b). Pathways affected by Ureaplasma isolates are marked in red. In A549 cells, Ureaplasma-triggered reduction of caspase 8 and 9 mRNA may result in an absent increase in protein production and active caspases. Ultimately, effector caspase 3 remains unactivated and apoptosis is impaired. In HPMEC, Ureaplasma spp. seem to enhance caspase 8 protein and caspase 9 activity. Both caspases may subsequently activate effector caspase 3 and induce apoptosis. It remains to be determined if the caspase 3 activation observed is directly due to Ureaplasma or a consequence of initiator caspase activation. Hypothesized underlying mechanisms Ureaplasma isolates may engage are indicated with a dashed line. I-inhibit / down-regulate; activate / up-regulate. CASP: caspase; IFN: interferon; LPS: lipopolysaccharide; NK: natural killer; TLR: toll-like receptor; TNF: tumor necrosis factor. Illustrations: https://smart. servier.com/.
https://doi.org/10.1371/journal.pone.0216569.g005 one example, has been associated with long-term respiratory tract Ureaplasma colonization in particular [49]. Our findings may therefore be of considerable clinical relevance.
Whereas several intracellular pathogens appear to be able to suppress apoptosis, the capacity for down-regulation or inactivation of caspases has been recognized for only a few of them [50,51]. Our data are the first to indicate Ureaplasma-associated down-regulation of apoptotic caspases, which may be relevant not only in pulmonary cells.
Ureaplasma-driven enhanced caspase 8 protein levels and enzyme activity of caspases 3 and 9 in HPMEC (Figs 3 and 5B), in contrast, provide first in vitro evidence for a pro-apoptotic effect of Ureaplasma spp. in pulmonary endothelial cells. We could confirm Ureaplasmadriven cell death in HPMEC by flow cytometric viability assessment (Fig 1B).
Apoptosis has a dual role with both beneficial as well as harmful effects. It contributes to immune defense and repair processes on the one hand, but facilitates tissue damage on the other [45]. A meticulous balance between growth and apoptosis is essential for normal lung development in utero as well as after birth [52]. Ureaplasma-driven increase or inhibition of apoptosis may disturb this physiological balance and may begin to impair structural lung development even prenatally, with severe implications as, for example, seen in BPD pathogenesis. Indeed, early structural lung tissue impairment such as pulmonary fibrosis could be demonstrated in Ureaplasma-infected preterm infants [53], and increased apoptotic activity was demonstrated within the lungs of preterm infants having developed BPD [54].
Underlying mechanisms Ureaplasma spp. use to modulate caspases remain to be determined. Ureaplasma spp. are known to engage toll-like receptor (TLR) signaling and induce TNF-α [33,55]. TLR as well as TNF-α are relevant initiators of apoptosis and inflammatory cell death [45] (Fig 5A). Another contributing factor may be intracellular Ureaplasma invasion with subsequent ammonia production and pH shift. Ureaplasma spp. can cause fatal hyperammonemia in lung transplant patients [56], but data regarding the influence of ammonia or pH on apoptosis and caspases are inconclusive [57][58][59]. Last but not least, Ureaplasma spp. might interfere with inhibition or degradation of caspases (Fig 5).
Ureaplasma isolates did not relevantly modulate inflammatory responses in this study. We found only non-significantly reduced inflammatory caspase levels in A549 cells (Fig 2), and, opposed to LPS, Ureaplasma spp. did not evoke distinct cytokine reactions in both pulmonary cell lines. This is in line with data from animal studies describing only mild pulmonary inflammatory reactions upon Ureaplasma exposure [48].
With this study, we describe distinct differences in the inflammatory responses of pulmonary epithelial and endothelial cells (Figs 1-4). Alveolar epithelial cells fulfill certain immunological functions and interact with, but do not belong to professional immunological cells [18,60]. This possibly reduces their ability to effectively target infections. Other studies also reported insignificant cytokine induction in A549 cells [61,62], and pathogen persistence is predominantly described in epithelial cells [63,64]. Contrarily, opposed to earlier perceptions, pulmonary endothelial cells are increasingly regarded relevant in immune recruitment and production of pro-inflammatory proteins [65,66]. This is reflected by our observation of Ureaplasma-or LPS-induced caspase and cytokine enhancements in HPMEC. Of note, we could recently demonstrate Ureaplasma-driven apoptosis in HBMEC [28], possibly indicating a higher vulnerability of endothelial cells in general for Ureaplasma-induced cell death. In any case, as many studies suggest a cross-talk between epithelial and endothelial pulmonary cells [67], affection of the one is likely to bear consequences for the other.
In this study, U. urealyticum serovar 8 appeared to evoke stronger reactions than U. parvum serovar 3 (Figs 1B, 2F, 3F and 3G). It is uncertain in how far this relates to in vivo conditions, since literature often does not distinguish between serovars. A relevant quantity of clinically isolated Ureaplasma was furthermore revealed to be hybrids of different serovars [3,68]. U. urealyticum were nonetheless reported to cause male urethritis and endometritis in pregnancy more frequently than U. parvum, and relevant genetic differences among serovars were described [69,70]. Compared to U. parvum, more frequent horizontal gene transfer and a higher phospholipase activity, both virulence determinants, were reported in U. urealyticum [70,71]. However, several studies did not find relevant differences between serovars and many authors conclude that Ureaplasma strains do no generally differ in their pathogenicity, but that host characteristics may determine individual inflammatory responses [1,3,70].
Our results contradict the one previous study describing apoptosis in Ureaplasma-exposed A549 cells [36]. These authors, however, used heat-inactivated Ureaplasma isolates. Although this is common practice, heat exposure may not only kill Ureaplasma cells, but is also likely to destroy their immunogenic surface proteins, impairing host immune responses [72]. In our experience, heat-inactivated Ureaplasma spp. tend to lose their pathogenicity. The use of viable Ureaplasma isolates is therefore advantageous and a strength of this study. It does, however, require a complex culture medium. Although we implemented growth of Ureaplasma isolates in yeast-free medium and furthermore ruled out endotoxin contamination, broth itself bore a certain immunogenic effect (Figs 1-4). A reason for this might be cell affection by the altered composition of A549 / HPMEC growth medium after addition of Ureaplasma broth. Particularly urea has been shown to be potentially cytotoxic [73] and may therefore have induced at least parts of the broth effects observed in this study. Another potential limitation of this study is the use of well-established, but adult cell lines. Although Ureaplasma infections may affect immunocompromised adults, preterm and term neonates are those with the highest susceptibility [9,17]. Future studies should therefore be conducted with cell lines of neonatal origin. Furthermore, in vitro settings can never fully represent in vivo conditions, where complex interactions between numerous mediators and cell types have to be taken into consideration. In this study, we concentrated on a purposive selection of caspases and cytokines, and we can correlate mRNA levels and enzyme activity or protein levels for only a few of them. Nonetheless, we provide first evidence for Ureaplasma-driven, cell-type specific interference with caspases and cell death. These findings once more suggest a profound clinical relevance of Ureaplasma spp. and should now encourage further research.

Conclusions
Pathogen-triggered inflammation usually evokes fierce immune reactions, condoning tissue damage, but ultimately resulting in pathogen eradication. The clinical significance of respiratory tract Ureaplasma colonization, however, may be based less on fulminant and temporary inflammatory reactions, but rather on chronic, subclinical infections, in which even mild inflammatory effects cause long-term sequelae. A key pathomechanism Ureaplasma spp. seem to employ is an interference with the caspase system. On the one hand, Ureaplasma-driven increases of caspase protein expression and activity in pulmonary endothelial cells may cause apoptosis and thus relevantly contribute to structural lung impairment. On the other hand, Ureaplasma spp. down-regulate caspase mRNA levels in pulmonary epithelial cells, thereby potentially suppressing programmed cell death as an important immune defense mechanism. Combined, these processes may facilitate chronic infections, long-term lung injury, and possibly inflammatory lung diseases such as BPD. This study provides additional evidence for the growing perception that Ureaplasma spp. are no innocent bystanders, but most likely much more relevant than contemplated to date. Glaser.