TI - Results . AB - Our objective was to test whether AECII in a culture system can boost the immune function of AM . Through co-culturing macrophages with lung epithelial cells , macrophages were significantly activated , demonstrating multiple functional activities including enhanced migration and actin reorganization ( Figure S1A , B ) . To investigate whether secreted ( soluble ) substances are major mediators , we collected a conditioned medium from primary mouse AECII following PAO1 infection for 1 h and added to primary AM culture . After 24 h , AM activity was detected by three measurements : migration ( Figure 1A ) , phagocytosis ( Figure 1B ) and superoxide production (Figure 1C) . Our data indicates that the three AM function parameters were significantly increased by the AECII conditioned medium compared to the control medium from AECII without infection ( P lt 005 , Student's t test ) . The activated AM also showed more significant morphological changes with membrane projections such as lamellipodia and filopodia by confocal microscopy ( Figure 1D ) . Moreover , we demonstrated apparent morphology alterations by scanning electron microscopy ( SEM ) (Figure 1E) . Although we found that AECII have a role in activating AM , it is possible that AM themselves can be activated by P.aeruginosa , which enhances the function of resting AM through an autocrine loop . Thus , we evaluated whether the activated AM can bolster immunity by stimulating other naive AM . We co-incubated resting AM with a conditioned medium that was derived from AM following P.aeruginosa time course infection . As expected , we have found that the AECII conditioned medium indeed showed greater increases in AM migration and superoxide production compared to the AM-conditioned medium , particularly at earlier times ( 6 h , Figure S1C ) . However , AM also secreted comparable MCP-1 possibly through autocrine activation at 24 h . To investigate how AECII can activate AM , we studied the possible involvement of various cytokines and chemokines . Thus we performed an in vitro experiment with isolated primary mouse AECII to measure cytokines in the conditioned medium of AECII infected by P.aeruginosa , which showed a significant increase in MCP-1 , IL-1beta , MIP-2alpha and TNF-alpha ( Figure 2A , B ) . We also showed that MCP-1 can transmit cellular signals to AM (4) , and may be a dominant cytokine in cultured AECII cells . The secretion of MCP-1 is time dependent ( Figure 2B ) . Also , we detected increased expression of MCP-1 and CCR2 ( MCP-1 receptor ) expression on AECII cells by western blotting ( Figure 2C ) . We further confirmed that the MCP-1-secreting cells expressed SP-C ( a marker of AECII ) by co-localization with confocal microscopy ( Figure 2D ) . Similar results showing cytokine secretion by AECII were also found in murine lung type II cell line MLE-12 cells [11] (Figure S2A) . We then infected mice to investigate MCP-1 secretion in response to P.aeruginosa infection in "in vivo" conditions [12] . C57BL6 mice ( 8 weeks female , 5 mice per group ) were infected with PAO1 intranasally , and broncho-alveolar lavage ( BAL ) was performed 18 h post infection . MCP-1 level was found to be increased as measured by immunohistochemistry in lung tissue and PAO1 infected mice showed a significant increase in MCP-1 compared to uninfected controls ( Figure 2G ) . Furthermore , the MCP-1 expression was demonstrated on AECII cells of human lung tissues from CF patients with P.aeruginosa infection (Figure 2E) , suggesting that MCP-1 actually participated in host defense in infected humans . We showed that lipid rafts in AECII are reorganized into aggregates ( platforms for cell signaling ) following acute infection by P.aeruginosa[12] , which is consistent with several studies regarding P.aeruginosa or other microorganisms [13] -[15] . P.aeruginosa has type three secretion system ( T3SS ) exoenzymes ExoS , ExoT , ExoU and ExoY , among which ExoS and ExoT are similar in the sequence and function for ADP ribosyltransferase and GTPase-activation protein activities [16] . The raft aggregates are differentially induced by various T3SS mutants or a pili deficient mutant [12] . Since there were some rafts outside the P.aeruginosa , we further quantified raft aggregates using Image J software and showed that P.aeruginosa infection induced strong raft aggregates , which were also associated with P.aeruginosa (Figure 2F) . In addition , controls without infection did not show raft aggregates , thus the aggregates of lipid rafts may be a specific response to P.aeruginosa . Next , we used various inhibitors to define the raft pathway in P.aeruginosa adhesion and internalization into AECII . Among the various inhibitors we have tested , we found that the lipid raft inhibitor (mbetaCD) , Lyn tyrosine kinase inhibitor (PP2) , Akt inhibitor and NF-kappaB inhibitor effectively blocked adhesion and internalization of P.aeruginosa , whereas the ceramide inhibitor ( blocking sphingolipid pathway ) only reduced internalization without affecting adhesion ( Figure 2G , Table 1 and Figure S2B ) . This data suggests that ceramide and cholesterol may differentially impact P.aeruginosa adhesion and internalization . To further dissect the mechanism , we intranasally instilled cholesterol chelator mbetaCD into mouse lungs and found a reduction in bacterial internalization and MCP-1 secretion was also partially hindered ( Figure 2G ) . The expression of CCR2 ( MCP-1 receptor ) was increased in AM treated with AECII conditioned medium as determined by western blots ( Figure 3A ) , indicating a functional involvement of this receptor in the pathway . Next , blocking of MCP-1 with neutralizing antibodies for 30 min before infection also down-regulated CCR2 compared to no antibody controls using fluorescent microscopy ( Figure 3B ) . To further confirm the role of MCP-1 , we added commercially available purified MCP-1 peptide ( Calbiochem ) directly to AM and found direct correlation between MCP-1 presence and AM activation , including superoxide production (Figure 3C) . Furthermore , by blocking with MCP-1 neutralizing antibodies for 30 min before PAO1 infection , we observed that the mice showed more severe infection with increased bacterial burden ( CFU/1 mg lung homogenates ) compared to that in the wt mice ( Figure 3D ) . In addition , the MCP-1 antibodies reduced superoxide production (Figure 3E) and phagocytosis by AM (Figure 3F) against the Ig isotype control . To ascertain the immune activity of AECII , we determined the immune symbolic markers , such as IL-12R and MHC Class II antigen by immunofluorescence . We showed that activated AECII expressed enhanced levels of MHC class II antigen as well as IL-12R and IL-17R ( Figure S3A , B ) . To further determine the role of MCP-1 as an immune stimulator , MCP-1 deficient mice ( MCP-/- mice ) were used to study MCP-1 effects on AM immune function and physiological role during P.aeruginosa infection . This can also distinguish the relative contributions of MCP-1 to immune defense against P.aeruginosa infection . We found that MCP-1-/- mice showed increased bacterial burden in the lung ( Figure 4A ) . A significant decrease in the superoxide production in AM was also noted ( Figure 4B ) . Furthermore , the lung from MCP-1-/- mice showed increased lung cell death ( showing decreased mitochondrial potential ) following infection by P.aeruginosa (Figure 4C) . Lung injury was also present as wet/dry ratio was increased by infection in the mice ( data not shown ) . We also showed that isolated AM from MCP-1-/- mice after bacterial infection had increased ability to phagocytose opsonized-E coli particles (Figure 4D) , indicating that the AM have not been saturated in uptaking bacteria ( probably due to reduced migration ) during in vivo infection . Interestingly , we noted that the lung injury is correlated with the increase in a number of pro-inflammatory factors such as MIP-2alpha ( but probably also due to the loss of MCP-1 ) than the wild-type control ( Figure 4E ) . Also , there was an increased neutrophil infiltration in the BAL ( not shown ) . These data suggest that MCP-1 may be crucial in maintaining a fine balance between pro-inflammatory and anti-inflammatory cytokines , which is important for combating bacterial invasion as well as minimizing acute lung injury . Also , we instilled the conditioned medium into MCP-1-/- mouse lungs , which increased the host defense compared to the control medium ( Figure 4F ) . These data critically confirm that MCP-1 is a potent immune regulator and inflammation controller in the airway spaces . Because raft signaling is partially dependent on T3SS , we attempted to define whether deletion in a single exoenzyme impacts the MCP-1 secretion by AECII . PAO1 wt and several toxin deficient strains ( PAO1 DeltaExoS , DeltaExoT , PA14 DeltaExoU and DeltaExoY ) were employed to examine the effect on AECII cells . Our data indicates that the various T3SS mutant strains influenced secretion of MCP-1 compared to PAO1 wt ( ) , the highest inducer being DeltaExoS as determined by RT-PCR and ELISA ( Figure 5A , B ) . Consequently , the conditioned medium from mutant strain DeltaExoS induced greater AM activation ( phagocytosis ) than PAO1 wt ( Figure 5C ) . By contrast , DeltaExoS strain resulted in lesser superoxide in AM than PAO1 wt strain (Figure 5D) . One possible explanation for this result is that loss of a particular exoenzyme may reduce the bacterial invasive ability , permitting the host to mount a better defense . The controls with normal medium , DMSO ( for inhibitor dilution ) and PAO1 supernatant induce measurable but lesser response than that of live PAO1 (Figure 5D) . We next examined whether live P.aeruginosa is necessary to induce MCP-1 secretion by AECII and to turn the cells immunologically more potent . Using dead PAO1 ( heating at 60degC for 1 h ) , supernatant , LPS and live PAO1 wt ( in the same amount of bacteria and same number of cells ) , we compared differences in MCP-1 secretion during infection of AECII . We found that live PAO1 increased MCP-1 secretion more so than did dead PAO1 , supernatants , and LPS . The conditioned medium from live infection also induced the strongest activation of AM (Figure 5E) . Thus , it is likely that live bacterial infection initiated stronger stimulation of AECII secretion than dead bacterium or its components . Since Src tyrosine kinase Lyn is located in the inner cytoplasm membrane of the cell , in the general proximity to lipid rafts , we investigated the role of several members of this family and discovered that Lyn ( p53/56 ) played a major role in P.aeruginosa infection . We have demonstrated that Lyn was activated in A549 cells following exposure to P.aeruginosa[12] . Lyn PHOSphorylation was measured by immunoblotting with phosphor-Src antibodies on immunoprecipitated Lyn protein from cell lysates . Pre-treatment of AECII with 5 nM PP2 ( synthetic inhibitor for Lyn ) blocked MCP-1 secretion (Figure S2B) . To elaborate the role of Lyn in MCP-1 production , Lyn siRNA transfection was performed to assess the effect of Lyn on P.aeruginosa invasion ( transfection efficiency >95% ) [17] , and our data show that MCP-1 was markedly decreased by Lyn siRNA (Figure 6A) . Importantly , the role of Lyn was confirmed using Lyn-/- mice , demonstrating significantly reduced MCP-1 by PAO1 and PAK infection versus . Lyn wt mice as determined by mRNA expression (Figure 6B) . Reduction of MCP-1 secretion was also observed in Lyn-/- mice as determined by ELISA (Figure 6C) , whiles there is no difference between wt and KO mice in non-treated condition ( data not shown ) . To probe downstream effectors in this pathway , we investigated NF-kappaB activation in response to P.aeruginosa infection . We evaluated this signaling process and our data showed that blocking of Lyn using dominant negative transfection of LynK275D construct [12] significantly reduced the NF-kappaB nuclear translocation compared to Lyn wt control transfectants ( not shown ) . Using NF-kappaB luciferase plasmid , we found a pronounced increase in luciferase activity with conditioned medium obtained from PAO1 infected AECII cells , but this was blocked by LynK275D transfection (Figure 6D) . We also observed similar Lyn-K275D-based inhibition in functional roles of AM . These findings suggest that Lyn signaling is linked to both the secretion of AECII and its functional role of AM . Finally , we also demonstrated that Lyn overexpression and increased PHOSphorylation in the human lung tissue of CF patients against the normal control ( Figure 6E ) . Our data indicate that Lyn activation might have occurred during CF disease process , thus it is reasonable to propose that Lyn may serve as a therapeutic target for controlling the disease . The AECII - AM cross-talk is presented in a simplified model ( Figure 7 ) . Bacterial infection activates AECII cells that secrete chemokines (MCP-1) . MCP-1 transmits signals to activate AM , which migrate towards infection sites and eliminate the pathogen . Lyn and lipid rafts may regulate MCP-1 and AM activation . In addition , the host defense may be balanced through the fine regulation by both Lyn and MCP-1 .