Effects of Calcium on the Interactions of Acinetobacter baumannii with Human Respiratory Epithelial Cells

Background Investigating the factors that influence Acinetobacter baumannii (Ab) adhesion/invasion into host cells is important to understand its pathogenicity. Metal cations have been shown to play an important role in regulating the biofilm formation and increasing the virulence of Ab; however, the effects of calcium on host-bacterial interactions have yet to be clarified. Here, the dynamic process of the interactions between Ab and human respiratory epithelial cells and the effects of calcium on host-bacterial interactions were explored using the technologies of microscopic imaging, quantitative PCR and real time cellular analysis (RTCA). Results The concentration of calcium, multiplicity of infection and co-culture time were demonstrated to have effects on host-bacterial interactions. A unique "double peak" phenomenon changed to a sharp "single peak" phenomenon during the process of Ab infection under the effects of calcium were determined based on the time-dependent cell response profiles. Moreover, calcium can increase Ab adhesion/invasion of epithelial cells by regulating the expression of Ab-related genes ( ompA , bfmRS , abaI ). Conclusions Effective control of calcium concentrations can provide new ideas for the prevention and treatment of multi-drug resistant Ab. effects on interaction between and calcium amount into (the

the biggest challenges to medical care. Adhesion of Ab to epithelial cells is considered an essential first step in colonization and infection. Colonization or infection with Ab occurs mostly in critically ill patients and causes severe pneumonia or bloodstream infections that result in increased in morbidity and mortality in these patients, which is a troublesome problem for clinical diagnosis, treatment and prevention [1,5,6]. Moreover, it is important to emphasize that colonization with Ab is more common than infection, even in susceptible populations [5,[7][8][9]. Although the pathogenicity of Ab is generally low, bacterial colonization is the greatest risk factor for infection in hospitals. Once the balance between host and microbe tips towards the development of an infection, the result can be severe and possibly lead to an increased probability of Ab hospital outbreaks. Therefore, studies on controlling Ab infection and colonization are of great practical significance and value.
In view of the seriousness of the antibiotic resistance of and mass colonization by Ab on biotic surfaces, the pathogenicity of Ab has attracted wide attention, especially in regard to the host-bacteria interaction and the process and molecular mechanism of Ab adhesion/invasion into human respiratory epithelial cells. Host-bacterial interactions are influenced by many factors including biofilm formation and endotoxin production. In addition, biofilm formation represents an important factor associated with virulence and is affected by bacterial fimbriae, outer membrane proteins, adhesins, metal ions, quorum sensing, and complex regulatory networks (e.g., two-component regulatory systems), among others [10]. Investigating the factors that influence Ab adhesion/invasion into epithelial cells is important to understand its pathogenicity.
Metal cations have been shown to play an important role in regulating biofilm formation and the differential expression of Ab-related genes, as well as in increasing the virulence of Ab and its ability to adhere to epithelial cells; however, the effects of calcium on host-bacterial interactions have yet to be clarified [11,12]. Studies have shown that bacterial infection can lead to a destabilization of the cellular calcium homeostasis and the activation of the calpain system ultimately triggering cell death, which suggests that changing the concentration of calcium in the environment may have a significant impact on the pathogenicity of Ab [13]. Lee et al. showed that bacterial attachment and biofilm formation on human respiratory epithelial cells and plastic surfaces were markedly reduced in the presence of the chelating agent EDTA (low levels of metal cations such as Ca 2+ and Mg 2+ ) through the analysis of a group of multidrug-resistant Ab clinical isolates [14]. These data suggested that high concentrations of calcium may promote biofilm formation by Ab and enhance its ability to adhere to respiratory epithelial cells.
However, this work did not specify the effects of calcium. It is important to explore the effects of calcium on host-bacterial interactions and to elucidate the functional mechanism, thus making it possible to take effective measures to control bacterial biofilm formation, adhesion and invasion and to ultimately provide new ideas for addressing the challenges of colonization and infection with multidrug-resistant Ab.
A label-free and noninvasive detection system (RTCA S16 system, ACEA Biosciences Inc.) based on dynamical and quantitative monitoring of cellular impedance in real time can produce specific time-dependent cell response profile (TCRPs) patterns. This approach can provide biological information related to cellular physiological function for studying hostbacterial interactions [15]. Thus, in this work, we developed microscopic imaging, quantitative PCR (qPCR) and TCRP methods for continuously monitoring the interactions between Ab and human respiratory epithelial cells and the effects of calcium on hostbacterial interactions. Our research can be used for the study of calcium-mediated signalling pathways in human respiratory epithelial cells infected with Ab, which provides a basis for studying the pathogenicity of Ab [16].

Optimum multiplicity of infection (MOI) of Ab to human respiratory epithelial cells
The effects of Ab on the morphology and proliferation of human respiratory epithelial cells at different MOIs and co-culture time points were determined by inverted microscopy and are shown in Table 1 and Fig. 1A. After host-bacterial co-culture for 2 h, there were relatively few bacteria and almost none were adhered to epithelial cells. The differences among the groups were also relatively small (especially in the control group and the MOI 1 and MOI 10 experimental groups). After co-culture for approximately 4 h, the differences among the groups became more obvious. As the MOI increased, the effects of Ab on the epithelial cells were also increased (the differences between the MOI 50 and MOI 100 experimental groups were not obvious). Additionally, the changes in cell morphology and proliferation at 4 h of co-culture were more typical compared to other co-culture time points. Thereby, the co-culture time point of 4 h was selected as suitable for subsequent studies. In addition, at 6-8 h of co-culture, the quantity of bacteria was relatively large, and they were not suitable for further research and analysis.
Bacterial invasion (including strong adhesion) to epithelial cells at different MOIs was determined by qPCR. Host cells and bacteria were co-cultured for 4 h and 1 × 10 5 cells were isolated for qPCR. The qPCR results (threshold cycles, Ct values) are shown in Fig.   1B. With increased MOI, the invasion of Ab into human respiratory epithelial cells gradually increased (the smaller the Ct value was, the more bacterial invasion occurred).
There was no invasion in the control group. The differences in the Ct values among all the experimental groups were statistically significant (P < 0.05). The MOI 100 group (Ct values: 23.56 ± 0.04) < the MOI 50 group (24.31 ± 0.05) < the MOI 10 group (27.36 ± 0.05) < the MOI 1 group (30.26±0.11). The differences in the Ct values between the MOI 10 and MOI 50 groups or between the MOI 1 and MOI 10 groups were approximately 3, while the differences between the MOI 50 and MOI 100 groups were relatively small (an approximately 0.8 difference in the Ct values).
The TCRPs were determined by real time cellular analysis. Ab infection induced the cell index (CI) to rise and then fall (Fig. 1C). As the MOI decreased, the time to reach the peak CI was gradually delayed and the peak CI was also higher. A unique "double peak" phenomenon emerged during the process of infection (the smaller the MOI was, the more obvious the phenomenon was). The bacterial concentration of the MOI 50 experimental group was most appropriate, and the phenomenon characteristics were more typical compared to the MOI 100 group. Therefore, the bacterial concentration (1 × 10 8 CFU/ml) corresponding to the MOI 50 was suitable for subsequent studies based on the above results.
The effects of Ab on different incubation states of human respiratory epithelial cells Ab had an almost identical effect on the different incubation states of epithelial cells ( Fig.   2A). In the early stages of host-bacterial co-culture, epithelial cells can adhere to the plastic surface, although bubble-like dead cells (nuclear pyknosis, cell swelling and dissolution, Additional file 1: Fig. S1) can be seen (after approximately 4 h). However, as the co-culture time is extended and the bacteria increase in quantity, the changes in cell morphology and proliferation, as well as the bacterial aggregation phenomenon became more typical (full field of bacteria after 8 h). It was observed that Ab had an effect on the adherent growth of epithelial cells. After 24 h, bacteria dominated the entire cell culture dish, while the host cells were all dead.
The dynamic processes (as determined based on the TCRPs) of the interactions between Ab and epithelial cells in different states of cell incubation were also similar (Fig. 2B). For instance, initial Ab infection had little effect on cell growth, and the "double peak" phenomenon emerged during Ab infection (this was not typical when the CI was about to enter the platform period). Almost identical cell growth curves were observed for the 0.45% sodium chloride solution (NaCl) and sterile distilled water groups. Compared to the untreated group, slight decreases or increases in the CI of the 0.45% NaCl and sterile distilled water groups may have been associated with the dilution of nutrients or metabolites in the culture medium. Therefore, the effect of 0.45% NaCl in the bacterial suspension on the adherent growth of epithelial cells could be considered negligible.

Effects of calcium on Ab proliferation and biofilm formation
As shown in Fig. 3 we found that high concentrations of calcium could contribute to the proliferation of Ab and that this effect was more pronounced with time. Additionally, the bacterial biofilm formation was more obvious as the concentration of calcium increased (Additional file 2: Table S1). Ab deficient in ompA may reduce biofilm formation.
Effects of calcium on the morphology and proliferation of human respiratory epithelial cells It was difficult to distinguish differences among the groups using an inverted microscopy, so we next used TCRPs to evaluate the effects of calcium on the proliferation of human  Table S2). The higher the calcium concentration was (≤4.4 mmol/L) and the longer the culture time was (≤24 h), the more significant the promoting effect on the CI values was.

Effects of calcium on host-bacterial interactions
The role of calcium in bacterial proliferation was the same as has been described above.
With increasing co-culture time and increased calcium concentrations, the bacterial aggregation phenomenon (biofilm formation) became more obvious (Additional file 5: Fig.   S3). After 8h, there was an obvious increase in the quantity of bacteria. After approximately 24h, the bacteria encompassed the full field, while massive numbers of host cells were dead. It is known that Ab can have effects on the adherent growth of epithelial cells; however, the effects of calcium on the interactions between Ab and epithelial cells could not be determined using an inverted microscopy.
Host and bacteria were co-cultured in calcium-supplemented medium for 2 h, 4 h and 6 h.
The effects of calcium on bacterial invasion (including strong adhesion) of epithelial cells (1 × 10 5 cells as the standard) were determined using qPCR and shown in Fig. 4. The results were negative for control group II. The Ct values were compared by univariate ANOVA with repeated measures and the SNK test. The time as a factor was significant with a P = 0.00, which meant that the amount of Ab epithelial cell invasion changed over time.
Moreover, treatment (calcium concentrations) as a factor was effective (P = 0.00), meaning that the amount of Ab epithelial cell invasion differed depending on grouping, and this difference among the groups was statistically significant (P < 0.05). There was a positive interaction between the time and treatment factors (P = 0.00). The role of time as a factor varied depending on the group. As a result, the higher the calcium concentration was and the longer the co-culture time was, the more frequently Ab epithelial cell invasion was observed.
The TCRPs showed that either initial Ab infection or the action of calcium can induce the CI to rise. With the increasing calcium concentration in host and bacteria culture medium (≤ 4.4 mmol/L), the faster the CI rose, the higher the peak CI was (6-8h), and the more significant cell growth stimulation was. By contrast, the CI declined rapidly with prolonged host-bacterial interactions (the higher the calcium concentration was, the faster the CI declined). Meanwhile, a sharp "single peak" phenomenon occurred during the infection

Effects of calcium on the expression of Ab-related genes
Calcium can affect the expression of Ab-related genes. The recA gene was used as an internal reference control. Both negative controls (I and II) had no amplification. Relative changes in the expression levels of target genes (ompA, bfmRS and abaI) between experimental groups and control I group were calculated by the 2 -△△Ct method. The trends were different between the abiotic and cellular environments that the bacteria were cultured in (Fig. 6).
There was no significant difference in the expression level of ompA among the groups cultured in the abiotic environment (P > 0.05), whereas expression of this gene showed significant differences in the cellular environment (P < 0.05). In the cellular environment, relative changes in expression levels between group c and control I, as well as between group d and control I were small (P > 0.05); however, ompA expression in group b was approximately 4-fold greater than that of the control I group.
In the abiotic environment, there were significant differences in the expression of bfmRS between the experimental groups and the control I group (P < 0.05). With the increase in calcium concentration in the culture medium, bfmRS expression in the experimental groups showed a decreasing trend; its expression in group d was approximately 0.31-fold lower than that of the control I group. In the cellular environment there was no significant difference in the expression levels between group a and the control I group (P > 0.05).
Contrary to the expression trends in the abiotic environment, expression levels in groups b and c were approximately 2.3-fold higher than that of the control I group, while the level in group d was approximately 4-fold higher.
In the abiotic environment abaI displayed similar levels of expression among groups b, c and the control I group (P > 0.05). The expression of abaI in group a was approximately 0.5-fold lower than that of the control I group, while its expression in group d was Interestingly, RTCA detection in the present study found unique "double peak" (calciumfree, Fig. 1C and 2B) and sharp "single peak" (with calcium, Fig. 4) phenomena during bacterial infection of epithelial cells. The sharp "single peak" may suggest that during the initial stages of bacterial infection, calcium stimulation promoted the rapid growth of cells (the observed peak). When the amount of bacteria reached the critical point, calcium in turn enhanced bacterial adhesion/invasion of epithelial cells, thereby rapidly decreasing the cell index (CI). The mechanism of the "double peak" phenomenon needs to be further studied and analysed. Possible mechanisms include that cell death was caused by bacteria and cells stimulating, adapting and interacting with each other. Specifically, the bacteria were added to the cell culture environment as exogenous foreign bodies, which stimulated the proliferation of the cells in the initial stage of infection (the appearance of the first peak), but then the bacteria inhibited host cell proliferation due to its own rapid proliferation (the decrease in the first peak). With prolonged host-bacterial interactions, the cells became tolerant to the bacterial inhibition (adaptation) and continued to proliferate (the appearance of the second peak), after which the two competed with one another in the nutrient-rich environment. Additionally, the increasing toxic effects of the bacteria on the epithelial cells resulted in a rapid reduction in the CI (the decrease in the second peak).
Ab-ompA, a highly conserved outer membrane protein, is also an important virulence factor that plays an important role in bacterial infection (which is also supported by our research) and causes an upregulation of epithelial cellular immune response signalling pathways [16,17]. It has been shown that Ab-ompA secretes and transmits virulence factors through outer membrane vesicles. Translocation of ompA-containing vesicles into host cells can result in host cell apoptosis, whereas ompA mutants fail to induce cell death [22]. Thus, blocking or inhibiting the expression of Ab-ompA will greatly reduce Ab adhesion/invasion into host cells [23]. Our studies showed that compared with abiotic environments, Ab-ompA expression is obviously changed in cellular environments. Finally, the bacterialDNA was extracted from 1 × 10 5 cells based on procedures described by Chen et al [26]. qPCR was performed with SYBR Premix Ex Taq II (TaKaRa Bio Inc., Detailed RTCA (real time cellular analysis) experimental procedures have been previously described [15,19,20]. Briefly, 50 μl of medium was added to 16-well E-plates (specific cell culture plates for RTCA) to obtain background readings, which were followed by the addition of 100 μl of a cell suspension (2×10 5

Ab biofilm assays
Mutant strains of Ab deficient in ompA (Ab-ompA -) were constructed and modified according to the conjugative transfer and parental conjugation methods [29,30]. Briefly, a modified pMO130-Tel R with the fragments corresponding to the regions up and downstream of the ompA gene generated from genome of Abwas constructed. The resultant plasmid was then transformed into the E.coli S17-1 λ pir competent cells. Then, trans-conjugation was performed between E.coli S17-1 λ pir donor strain and Ab recipient strain to transfer and integrate pMO130-Tel R -ompA Up/Down into the chromosome of Ab.Donor and recipient strains were plated onto LB agar containing tellurite (30 mg/L) and gentamicin (25 mg/L) and incubated at 37 °C overnight. 0.45 mol/L catechol solution was sprayed on the surface of the plate, and yellow clones were picked for PCR identification.
The second selection was then performed by incubating Ab on 10% sucrose with no salt LB agar plates, overnight at 37 °C to identify sucrose-resistant sensitive clones, then analyzed by PCR and sequencing to confirm that the target gene was excised, resulting in an unmarked in-frame deletion. A single clone with no ompA sequence was saved as Ab-ompA -. Ab and Ab-ompAisolates were added separately to 96-well plates containing RPMI In addition, a 400 μl aliquot of each of sterile distilled water and 0.45% NaCl was used as the normal control II. All the liquid in the wells was carefully discarded after a 2 h, 4 h or 6 h incubation, as described above. Before digestion, the cells were washed three times with PBS. Other procedures, including bacterial DNA extraction and qPCR were carried out as RTCA experimental procedures were as described above. After the cells (20,000 cells per well) were incubated for 48 h, the two part experiment was carried out as follows:

1.
A 20 μl aliquot of each of the abovementioned bacterial suspensions and different concentrations of CaCl2 was added to the wells; the bacterial suspensions and EDTA was used as the normal control I; sterile distilled water and 0.45% NaCl was used as the normal control II.

2.
A 20 μl aliquot of each of different concentrations of CaCl2 and the bacterial suspension/0.45% NaCl/sterile distilled water was added to the wells; a total of 20 μl of sterile distilled water was added as the normal control. Wells without intervention were used as the blank control.
The effects of calcium on the host-bacterial interactions were evaluated based on the TCRP results, qPCR and microscopic imaging.

Statistical analysis
The data are presented as the means ± SD. One-way analysis of variance (ANOVA) and the    for the E-Plate were obtained and the plate was incubated at room temperature for 30 min). The "double peak" phenomenon emerged during Ab infection. The smaller the MOI was, the later and higher the peak CI was, the more significant the phenomenon was. Representative curves are an average of three replicate wells.   The effects of calcium on Ab infection TCRPs of human respiratory epithelial cells.
A total of 20,000 cells per well were seeded into E-plates. t: treatment time point (48 h, 80-90% cell confluence). With increased calcium concentrations (bacteriainfected cells), the CI increased or decreased faster and reached a higher peak value. A sharp "single peak" phenomenon occurred during the infection.