Myeloid PTEN Promotes Inflammation but Impairs Bactericidal Activities during Murine Pneumococcal Pneumonia

Infectious diseases are a major burden for our society, with respiratory tract infections being a leading cause of morbidity and mortality worldwide. Streptococcus pneumoniae is the most frequent causative pathogen of community acquired pneumonia, affecting >500,000 people in the United States annually (1, 2). The worldwide increase in antibiotic resistance among S. pneumoniae strains underlines the urgent need for a better understanding of molecular mechanisms associated with pneumococcal pneumonia (3).

The phosphatase and tensin homolog deleted on chromosome 10 (PTEN) is a well-described tumor suppressor gene and multifunctional phosphatase that antagonizes PI3K’s enzymatic activity by dephosphorylating phosphatidylinositol (3,5)-trisphosphate to generate phosphatidylinositol (4,5)-biphosphate. PI3K’s enzymatic activity is warranted by two distinct subclasses, namely class Ia (p110α, β, δ) and class Ib (p110γ) (4). PTEN efficiently limits PI3K activity and downstream Akt signaling. PI3K/PTEN has been shown to play a prominent role in a variety of cellular mechanisms, such as survival, migration, and proliferation (5). However, little is known about the biological role of PTEN in inflammation and in particular during infectious diseases. The PI3K/Akt signaling axis is crucial for site-directed migration and diapedesis of immune effector cells, such as neutrophils and monocytes, to the site of inflammation and infection (6–10). In contrast to the conception that PI3K mediates proinflammatory signals, several studies indicate that the PI3K/PTEN pathway modulates the inflammatory response to bacterial cell wall components (11–15). Making use of a combination of pharmacologic and genetic means, we and others could previously show that the PI3K pathway provides beneficial anti-inflammatory properties in mouse models of endotoxemia and sepsis. These observations have been further supported by Martin et al. (16), who intriguingly demonstrated that TLR-induced PI3K/Akt activation phosphorylated, and thereby inactivated, downstream glycogen synthase kinase (GSK) 3β, which in turn resulted in diminished NF-κB–driven proinflammatory gene expression in monocytic cells.

The precise function of PI3K signaling during infections with viable bacteria is less well understood. Maus et al. (17) published a report that investigated the contribution of the γ-catalytic subunit of PI3K (p110γ) during pneumococcal pneumonia. They hereby demonstrated that p110γ, which is generally thought to be responsible for the G-protein coupled receptor-induced chemotactic response of neutrophils and monocytes, was not required for neutrophil influx following pulmonary inoculation of pneumolysin or whole bacteria in vivo (17). However, p110γ-deficient mice displayed an impaired bacterial clearance and delayed recruitment of exudate macrophages (17).

Still, the role of myeloid cell-associated PTEN upon infection with clinically relevant pathogens in healthy mice, such as community-acquired pneumonia by S. pneumoniae, is incompletely understood. The increasing number of patients and clinical trials applying drugs targeting the PI3K/PTEN pathway highlights the urgent need for better comprehension of PI3K/PTEN signaling during clinically relevant infections (18). To elucidate the contribution of myeloid cell-associated PI3K/PTEN during S. pneumoniae infection, we therefore made use of a conditional knockout strategy to specifically eliminate PTEN expression in myeloid cells (we hereafter refer to these mice as PTEN^MC-KO animals and PTEN^MC-Wt littermate controls, respectively). Pneumococcal pneumonia was then induced in PTEN^MC-KO mice, which displayed enhanced PI3K activity, and littermate PTEN^MC-Wt controls, after which the inflammatory response was investigated.

Floxed PTEN mice were kindly provided by T.W. Mak (Cancer Institute at Princess Margaret Hospital, University Health Network, Toronto, Ontario, Canada) (19); lysozyme M (LysM) Cre recombinase transgenic mice were a kind gift from R. Johnson (University of California San Diego, La Jolla, CA) (20). PI3Kγ (p110γ) mice were obtained from J.M. Penninger (Institute of Molecular Biotechnology of Austrian Academy of Sciences, Vienna, Austria) (9). Intercrossed mice were backcrossed to a C57BL/6J background for at least eight generations. Littermate-controlled experiments were performed using 8–12-wk-old male mice. For genotyping, murine tissue was lysed in PCR-lysis buffer, and direct PCR was performed using GoTaq DNA Polymerase (Promega, Madison, WI). All animal studies were approved and comply with institutional guidelines (BMWF-66.009/0103-C/GT/2007).

Thioglycollate-elicited peritoneal macrophages were isolated from PTEN^MC-KO and PTEN^MC-Wt controls as described previously (21). Alveolar macrophages were isolated by bilateral bronchoalveolar lavage (BAL) as described elsewhere (22, 23). Bone marrow was isolated from femurs and tibias of healthy PTEN^MC-KO and PTEN^MC-Wt mice. Bone marrow cells were incubated with conditioned media from L929 cells (20% in RPMI 1640) for 10 d to allow differentiation and maturation of macrophages.

Macrophage cell lysates were separated by SDS-PAGE, blotted to membrane (Immobilon PVDF Transfer Membrane, Millipore, Bedford, MA), probed with rabbit primary Abs against PTEN, protein kinase B, phospho-protein kinase B (Ser473), phospho-GSK3β (Ser9) (Cell Signaling Technology, Beverly, MA), and β-actin (Sigma-Aldrich, St. Louis, MO). For detection, a goat anti-rabbit secondary Ab conjugated with HRP (Amersham Biosciences, Piscataway, NJ) was used.

RNA from cell culture was extracted using Qiagen’s RNEasy kit (Qiagen, Valencia, CA), and real-time PCR was conducted according to the LightCycler FastStart DNA MasterPLUS SYBR Green I system using the Roche Light cycler II sequence detector (Roche Diagnostic Systems, Somerville, NJ). Cycling conditions were set at 1 cycle at 95°C for 10 min, 50 cycles at 95°C for 5 s, 68°C for 5 s, and 72°C for 10 s. To confirm specificity of the reaction products, the melting profile of each sample was analyzed using the LightCycler Software 3.5 (Roche Diagnostic Systems). Mouse gene-specific primer sequences for inducible NO synthase (iNOS) were: 5′-ACC TCA CTG TGG CCT TGG TC-3′ (forward) and 5′-GGG TCC TCA GGG AGC TGG AA-3′ (reverse). NO release was measured by the generation of nitrite (Sigma-Aldrich) in supernatants of PTEN^MC-KO and PTEN^MC-Wt littermate control macrophages incubated with heat-killed S. pneumonia (ATCC 6303) for 24 h. Assays were performed according to the manufacturer’s protocol (24).

Primary alveolar macrophages (AMs) were incubated with FITC-labeled heat-killed S. pneumoniae (ATCC 6303) at a multiplicity of infection of 100 for 30 min at 37°C. After washing steps, lysosomes were stained with Lysotracker red and nuclei with DAPI (Invitrogen, Carlsbad, CA), followed by visualization using confocal laser scanning microscopy (LSM 510, Zeiss, Oberkochen, Germany). The ratio of engulfed bacteria (as determined by overlay of green bacteria and red lysosomes) were quantified by an independent researcher from 300–400 counted cells per well and are expressed as percentage of cells that contain bacteria. In addition, an FACS-based phagocytosis assay was performed exactly as described earlier (25). In brief, primary AMs were allowed to adhere overnight before being incubated with FITC-labeled S. pneumoniae at 37°C or 4°C, respectively. Uptake of bacteria was quantified by FACS, and the phagocytosis index was calculated as follows: (mean fluorescence × % positive cells at 37°C) − (mean fluorescence × % positive cells at 4°C). Bacterial killing was performed as described (25). In brief, AMs were isolated, plated at a density of 2 × 10⁵ cells/well, and allowed to adhere. S. pneumoniae were added at a multiplicity of infection of 100, and plates were placed at 37°C for 10 min. Each well was then washed five times with ice-cold PBS to remove extracellular bacteria. To determine bacterial uptake after 10 min, triplicate of wells were lysed with sterile H₂O and designated as t = 0. Prewarmed SF-RPMI 1640 was added to remaining wells, and plates were placed at 37°C for 10, 30, 60, or 90 min, after which cells were again washed five times with ice-cold PBS and lysed as described above. Cell lysates were plated in serial-fold dilutions on blood agar plates, and bacterial counts were enumerated after 16 h. Bacterial killing was expressed as the percentage of killed bacteria in relation to t = 0 (percent killing = 100 − [(number of CFUs at time x/number of CFUs at time 0) × 100]).

Pneumococcal pneumonia was induced as described previously (26–28). Briefly, S. pneumoniae serotype 3 was obtained from American Type Culture Collection (ATCC 6303, Rockville, MD) and grown to log-phase. Mice were short-term anesthetized with isoflurane (Forene, Abbott Laboratories, Abbott Part, IL) and 50 μl bacterial suspension (~5 × 10⁴ CFUs) was inoculated intranasally. For survival analysis, infected mice were observed every 3 h. At indicated time points, mice were sacrificed; BAL was performed, and blood and lungs were collected and processed as described (27, 29). CFUs were determined from serial dilutions of lung homogenates, blood, and BAL fluid (BALF), plated on blood agar plates, and incubated at 37°C for 16 h before colonies were counted. Cytokines and chemokines were quantified in lung homogenates and BALF. TNF-α, IL-6, keratinocyte-derived chemokine (KC), MCP-1, and MIP-2 were measured using ELISAs (R&D Systems, Minneapolis, MN), as was myeloperoxidase (MPO) (HyCult Biotechnology, Uden, The Netherlands) and IL-10 (Bender Medsystems, Vienna, Austria). Detection limits were: 15 ng/ml for TNF-α; 16 pg/ml for IL-6; 12 pg/ml for KC; 4 pg/ml for MCP-1; 94 pg/ml for MIP-2; and 15 pg/ml for IL-10. Differential cell counts were determined using counting chambers and cytospin preparations stained with Giemsa.

Lungs were fixed in formalin and embedded in paraffin; 4-μm sections were stained in H&E and analyzed by a blinded pathologist. The lung was scored with respect to the following parameters: interstitial inflammation, edema, endothelitis, bronchitis, pleuritis, and thrombi formation. Each parameter was graded on a scale of 0–3 (0: absent; 1: mild; 2: moderate; and 3: severe). The total lung inflammation score was expressed as the sum of the scores for each parameter, the maximum being 18. Granulocyte immunostaining was performed on paraffin-embedded lungs as described (27). After Ag retrieval using pepsin, tissue sections were incubated with FITC rat anti-mouse Ly-6G (BD Biosciences, San Jose, CA) or corresponding isotype control IgG (Emfret Analytics, Würzburg, Germany), followed by rabbit anti-FITC Ab (Zymed, Invitrogen) in normal mouse serum. Finally, slides were incubated with polyclonal anti-rabbit-HRP Ab (Immunologic, Duiven, The Netherlands) and visualized using 3,3-diaminobenzidine-tetrahydrochloride (Vector Laboratories, Burlingame, CA). Counterstaining was done with hemalaun solution.

Data were analyzed with GraphPad Prism 4 software (GraphPad, San Diego, CA) using unpaired Student t test or one-way ANOVA followed by post hoc tests when appropriate. Bacterial killing data were calculated by two-way ANOVA. Survival data were analyzed by Kaplan-Meier followed by log-rank test. Criteria for significance for all experiments were p < 0.05.

To investigate the role of myeloid cell-derived PTEN during the inflammatory response to bacterial infections, we generated conditional pten knockout mice in which PTEN expression was controlled by LysM, a myeloid cell-specific promoter (20). Depending on the presence of LysM Cre recombinase, double-floxed (pten^flox/flox) mice are referred to as PTEN^MC-KO or PTEN^MC-Wt mice, respectively (Fig. 1A). Deletion of PTEN was confirmed in primary macrophages (Fig. 1B), and resulting downstream effects of constitutively active PI3K were reflected by greatly elevated baseline levels of phospho-Akt and phospho-GSK3β in PTEN^MC-KO macrophages (Fig. 1C). These findings also indicate that alternative lipid phosphatases, such as SHIP1 and SHIP2, which are known to influence phosphatidylinositol (3,4,5)-triphosphate plasma membrane content (30), do not sufficiently limit PI3K activity in macrophages or compensate for the loss of PTEN.

To characterize PTEN-associated properties of macrophages during bacterial infection, we next examined kinase phosphorylation downstream of PI3K in PTEN^MC-Wt and PTEN^MC-KO macrophages upon stimulation with S. pneumoniae. Bacterial challenge led to Akt and GSK3β phosphorylation in PTEN^MC-Wt cells with highest levels 60 min postactivation (Fig. 1D). Although Akt phosphorylation only modestly increased over baseline levels upon S. pneumoniae stimulation in PTEN-deficient macrophages, GSK3β phosphorylation was found markedly enhanced in these cells (Fig. 1D). These data illustrate that the innate immune response to S. pneumoniae involves downstream PI3K pathway activation.

To test whether PTEN deficiency in myeloid cells might impact the outcome of pneumococcal pneumonia in vivo, we infected PTEN^MC-KO and PTEN^MC-Wt mice with 5 × 10⁴ CFU S. pneumoniae and monitored survival over 7 d. Already 2 d postinfection, we found PTEN^MC-KO mice to show less severe signs of disease, which was quantified using a clinical severity score (data not shown). PTEN^MC-Wt animals rapidly displayed signs of serious infection, and all mice succumbed within 107 h postinduction of pneumonia, at a time when >50% of PTEN^MC-KO animals were still alive (Fig. 2). Together, myeloid-cell specific PTEN deficiency was associated with less severe signs of disease and significantly improved survival during pneumococcal pneumonia in vivo.

In our effort to elucidate the detrimental role of PTEN during pneumococcal pneumonia, we investigated PTEN’s contribution to basic functional properties of myeloid cells, such as cytokine secretion or bacterial phagocytosis and killing. Upon stimulation of bone marrow-derived macrophages (BMDMs) and primary AMs from PTEN^MC-KO and PTEN^MC-Wt mice with S. pneumoniae, we observed a significantly reduced TNF-α release by PTEN^MC-KO macrophages compared with PTEN^MC-Wt cells, whereas KC levels did not differ significantly (Fig. 3A, 3B). To study effector molecules importantly associated with bactericidal mechanisms, we next investigated the pathogen-induced expression of iNOS in AMs and BMDMs. Surprisingly, we hereby discovered that PTEN^MC-KO BMDMs stimulated with S. pneumoniae expressed significantly higher iNOS levels 8 h postinduction (data not shown), whereas iNOS expression was diminished in PTEN-deficient AMs (Fig. 3C). We furthermore confirmed that decreased iNOS expression correlated with suppressed NO production by quantifying nitrite in supernatants of AMs (Fig. 3C).

We then isolated primary AMs from PTEN^MC-Wt and PTEN^MC-KO mice and explored their ability to phagocytose S. pneumoniae using confocal microscopy as well as an FACS-based phagocytosis assay. By quantifying the proportion of macrophages that contained intracellular bacteria, we discovered a significantly increased uptake of S. pneumoniae by PTEN^MC-KO cells (p < 0.05 versus wild-type cells) (Fig. 4A, 4B). Furthermore, performing a killing assay enabled us to demonstrate enhanced bacterial killing by AMs from PTEN^MC-KO mice (Fig. 4C).

Hence, these data illustrate that PTEN activity strongly impacted functional properties attributed to macrophages. Although constitutively active PI3K signaling counteracted the proinflammatory TNF-α and iNOS response, it augmented the phagocytic and bactericidal properties of primary AMs upon challenge with S. pneumoniae in vitro.

To discern the in vivo impact of myeloid-cell associated PTEN on inflammation during bacterial infection, we then asked how above described findings would translate into the immediate host response during bacterial pneumonia in vivo. For this purpose, we infected PTEN^MC-KO and PTEN^MC-Wt mice with S. pneumoniae and studied the early inflammatory response after 6 h. In line with in vitro data depicted in Fig. 3, PTEN^MC-KO animals exhibited a diminished proinflammatory cytokine response, illustrated by significantly lower TNF-α concentrations in BALF of these mice (p < 0.01 versus PTEN^MC-Wt mice) (Fig. 5A). At the same time, the anti-inflammatory cytokine IL-10 was found significantly increased in lungs of PTEN^MC-KO mice as compared with wild-type animals (Fig. 5B), indicating that the constitutive activation of myeloid cell-derived PI3K pathways dampened the inflammatory response in vivo.

Host defense against respiratory tract infections critically depends on the effective influx of neutrophils. Because PI3K activation has been repeatedly shown to promote neutrophil chemotaxis (10), and in light of a recent publication that highlighted the role of myeloid PTEN as a suppressor of neutrophil migration (31), we expected PTEN^MC-KO mice to exhibit an enhanced recruitment of neutrophils to the alveolar compartment. When enumerating the number of cells attracted to the alveolar compartment 6 h postinduction of pneumococcal pneumonia, we surprisingly found a significantly impaired influx of neutrophils in the absence of PTEN (Fig. 5C, 5D). Despite this reduced neutrophil attraction and diminished TNF-α response, bacterial outgrowth in lungs and BALF of PTEN^MC-KO mice was not affected at this early time point (Fig. 5E, 5F). Together, PTEN deficiency dampened the early inflammatory response during bacterial pneumonia.

Having established PTEN’s proinflammatory contribution to the induction-phase of S. pneumoniae infection in vivo, we then explored PTEN’s impact over the course of pneumococcal pneumonia and investigated mice 48 h postinfection. At this later time point, we continued to observe reduced levels of proinflammatory mediators in BALF and lungs of mice lacking PTEN. As depicted in Table I, BALF concentrations of TNF-α, KC, and MCP-1 were significantly decreased, whereas IL-6 showed a modest but nonsignificant reduction in PTEN^MC-KO versus PTEN^MC-Wt mice. In contrast to the alveolar compartment, TNF-α and IL-6 levels did not differ in lung homogenates (data not shown), whereas KC concentrations were considerably decreased, and IL-10 levels significantly increased in lungs from PTEN^MC-KO mice as compared with PTEN^MC-Wt animals (Fig. 6A, 6B). In line with this diminished chemokine release, PTEN^MC-KO animals displayed an impaired ability to attract neutrophils to the site of infection, as illustrated by a significantly decreased proportion of neutrophils in BALF (Table II) as well as reduced MPO levels in lungs from PTEN^MC-KO mice (Fig. 6C) 48 h postinfection. The reduction in proinflammatory mediators and neutrophil influx was accompanied by modestly increased bacterial counts in the alveolar compartment of PTEN^MC-KO mice 48 h postinfection with S. pneumoniae (Fig. 6D), whereas CFU counts in lung homogenates did not differ between the mouse strains (Fig. 6E).

Because PI3K activation has been repeatedly shown to promote migration of neutrophils (7, 8, 21, 32), the continual observation of impaired pulmonary neutrophil recruitment in PTEN^MC-KO mice was unanticipated (Figs. 5C, 5D, 6C). Based on a recent report that illustrated PTEN’s suppressive function on neutrophil migration during Escherichia coli peritonitis and sterile peritoneal inflammation (31), we wondered whether these contradicting results were due to pathogen- or organ-specific differences that have never been investigated before. To answer this question, we injected S. pneumoniae i.p. in PTEN^MC-KO mice and PTEN^MC-Wt littermate controls and enumerated peritoneal neutrophil counts after 6 h. In contrast to the diminished alveolar neutrophil influx during pneumonia, we found an enhanced peritoneal recruitment of neutrophils in animals lacking myeloid cell-associated PTEN (Fig. 7A). These findings indicate that organ-specific differences could explain our observation of diminished neutrophil migration in PTEN^MC-KO mice suffering from pneumococcal pneumonia in vivo and argue against a fundamental cellular defect of PTEN-deficient neutrophils. To understand organ-specific differences in more detail, we additionally studied the inflammatory cytokine and chemokine response of primary peritoneal macrophages upon S. pneumoniae stimulation in vitro. Comparable to our observations from BMDMs and AMs (Fig. 3), we discovered reduced TNF-α secretion by peritoneal PTEN^MC-KO macrophages (Fig. 7B). However, in strong contrast to BMDMs or AMs, peritoneal macrophages that lacked PTEN released significantly more KC than PTEN^MC-Wt cells (Fig. 7C). This enhanced chemokine release by PTEN^MC-KO peritoneal macrophages provides a potential explanation for the augmented PMN influx into the peritoneal cavity of PTEN^MC-KO animals.

In an attempt to identify the contributing factors that ultimately led to improved outcome of PTEN^MC-KO mice suffering from pneumococcal pneumonia, we repeated the pneumonia study and sacrificed mice after 65 h (i.e., right before animals started to succumb to infection). We hereby observed significantly decreased IL-6 levels and modestly reduced TNF-α and KC concentrations in BALF of PTEN^MC-KO (Table III). In line with above-described findings at 6 h and 48 h postinfection, we continued to detect significantly higher IL-10 levels in lung homogenates of mice lacking myeloid PTEN (Fig. 8A). When analyzing the cellular composition in BALF, we found a predominance of monocytes/macrophages in PTEN^MC-KO mice, whereas neutrophil numbers exceeded monocytes/macrophages in PTEN^MC-Wt littermates (Fig. 8B, 8C). In accordance, histological evaluation of lung slides disclosed significantly more pronounced signs of inflammation in PTEN^MC-Wt mice than PTEN^MC-KO animals (Fig. 8F, 8G). However, when enumerating bacterial counts in BALF and lungs, we did not discover any differences between wild-type and PTEN-deficient mice 65 h postinfection (Fig. 8D, 8E). Furthermore, blood cultures did not reveal any differences between groups (data not shown).

Hence, these data indicate that the continuous activation of PI3K, as seen in PTEN^MC-KO mice, beneficially modulated the inflammatory response to S. pneumoniae, thus allowing for accelerated resolution of pneumonia without impairing bacterial clearance in vivo.

The role of PI3K pathways in the inflammatory response is a controversial matter, as published reports suggested either proinflammatory or anti-inflammatory properties (15, 33, 34). We and others demonstrated earlier that PI3K activation exerts protective immunomodulatory effects in murine models of endotoxemia, sepsis, and viral infection (16, 35–37). These findings have been partly attributed to PI3K’s ability to modulate the transcriptional activity of NF-κB and to efficiently limit proinflammatory signaling cascades induced via MAPK pathways (11, 34). Given that PTEN is a key regulator of PI3K activity, we hypothesized that PTEN might act as a critical modulator of the inflammatory response during bacterial infection. To investigate this idea in a clinically relevant model, we studied the role of PTEN during S. pneumoniae pneumonia in vivo and discovered that myeloid cell-specific PTEN deficiency exerted beneficial effects. PTEN deficiency was associated with diminished TNF-α and increased IL-10 responses, enhanced macrophage phagocytosis, reduced neutrophil migration to lungs, and, ultimately, improved survival.

Modulation of PI3K activity by cell type-specific pten gene ablation disclosed a markedly reduced TNF-α response by various primary macrophage subsets that were stimulated with S. pneumoniae. These findings correlated with earlier observations by us and other investigators who showed that LPS-challenged PTEN-deficient macrophages displayed a profoundly reduced TNF-α release and diminished activation of MAPKs (15, 38). We concurrently discovered the enhanced phosphorylation of GSK3β in PTEN-deficient macrophages. GSK3β is a constitutively active serine/threonine kinase and downstream target of PI3K that can be inactivated through phosphorylation by Akt (39). The biological significance of GSK3β during inflammation was discovered by Martin et al. (16), who revealed that GSK3β inhibition led to a diminished inflammatory response toward various TLR agonists, which was illustrated by reduced TNF-α and enhanced IL-10 releases. In striking agreement with this report, we hereby observed decreased levels of proinflammatory cytokines, such as TNF-α and elevated concentrations of the anti-inflammatory cytokine IL-10 in lungs of S. pneumoniae-infected PTEN^MC-KO animals. Therefore, enhanced phosphorylation and consecutive inactivation of GSK3β activity in macrophages of PTEN^MC-KO animals might explain our in vivo findings of a dampened inflammatory response in these mice.

The most unanticipated finding of our studies was the diminished pulmonary neutrophil influx in PTEN^MC-KO mice suffering from pneumococcal pneumonia. Importantly, this observation was not related to constitutively reduced neutrophil numbers in this specific mouse strain, because we previously showed that blood neutrophil numbers were even higher in healthy PTEN^MC-KO animals as compared with wild-type mice (15). In support of the general notion of PI3K being a key player in cell migration (6–8, 10, 19), the migratory capacity of PTEN-deficient neutrophils was found enhanced in earlier reports (31). Furthermore, increased neutrophil recruitment was observed in PTEN^MC-KO animals using models of thioglycollate or E. coli-induced peritonitis in vivo (31). Although we were able to confirm these data, as we also observed enhanced peritoneal neutrophil influx following thioglycollate administration in PTEN^MC-KO mice (data not shown), we consistently observed reduced alveolar neutrophil migration during pneumococcal pneumonia in PTEN^MC-KO animals. To exclude the possibility of pathogen-specific differences, we challenged mice i.p. with S. pneumoniae and observed enhanced peritoneal neutrophil recruitment in PTEN^MC-KO mice (Fig. 7A). Because neutrophil attraction to sites of infection critically depends on the effective release of chemokines, such as KC (40, 41), we investigated if organ-specific differences in neutrophil migration in PTEN^MC-KO animals were a consequence of altered chemokine release by resident macrophages. Indeed, when measuring KC concentrations in supernatants of alveolar and peritoneal macrophages that were stimulated with S. pneumoniae in vitro, we discovered an enhanced KC release by peritoneal but not alveolar PTEN^MC-KO macrophages. Beside macrophages, respiratory epithelial cells are a major source of KC within the lungs in vivo, and release of this chemokine is largely triggered by macrophage-derived proinflammatory cytokines, such as TNF-α (42). The fact that we identified reduced KC levels in lung-homogenates from infected PTEN^MC-KO mice in vivo (Table I) might therefore result from the attenuated macrophage-associated (e.g., TNF-α–mediated) activation of airway epithelial cells in vivo, ultimately resulting in reduced neutrophil recruitment. In contrast to our observations, Li et al. (43) disclosed increased pulmonary KC concentrations and enhanced neutrophil migration in PTEN-deficient mice suffering from E. coli pneumonia. It therefore seems that PTEN differentially regulates the attraction of neutrophils depending on either the affected organ and/or the inducing agent.

Unlike neutrophils, cells of monocytic origin (infiltrating monocytes/macrophages) were recruited in increased numbers to lungs of healthy and S. pneumoniae-infected PTEN^MC-KO mice (data not shown) (Table I). This result is in agreement with a report by Maus et al. (17), who showed that monocyte/macrophage recruitment in pneumoccocal pneumonia critically depended on proper p110γ signal transduction. We recently demonstrated that AMs crucially contribute to host defense during murine pneumococcal pneumonia, as they exert an important role in the resolution of pneumococcal pneumonia by virtue of their capacity to eliminate apoptotic neutrophils (29). This idea is strengthened by data obtained in p110γ-deficient mice in which the recruitment of monocytes/macrophages was found to be impaired (17 and data not shown). PI3Kγ-KO mice showed substantial lung infiltrates and tissue injury during pneumococcal pneumonia (data not shown), whereas PTEN^MC-KO mice that embraced significantly increased numbers of (alveolar) macrophages showed less severe signs of tissue damage (Fig. 8). We in addition discovered an increased phagocytic and killing potential of AMs in the absence of PTEN. The enhanced phagocytic properties of PTEN-deficient macrophages might have compensated for the reduced number of infiltrating neutrophils. However, the precise role of neutrophils during pneumococcal pneumonia has been challenged recently by a report showing unaltered bacterial clearance and improved survival in neutrophil-depleted animals that were infected with S. pneumoniae (44). It seems that enhanced neutrophil numbers prolong inflammation and ultimately fuel tissue damage, thus resulting in worsened outcome. It is therefore tempting to hypothesize that lower neutrophil counts and simultaneously increased numbers of macrophages in PTEN^MC-KO animals improved bacterial clearance and augmented the resolution of inflammation, thus contributing to diminished tissue damage and favorable outcome in these animals.

In conclusion, our findings demonstrate that enhanced PI3K activity in PTEN-deficient mice resulted in an improved outcome during pneumococcal pneumonia. These findings implicate a crucial role for PTEN in the homeostasis of pro- and anti-inflammatory mechanisms evoked during a relevant bacterial infection. Thus, interfering with PI3K signaling might have tremendous implications on the course of pneumococcal pneumonia: although blockade of PTEN might be beneficial, reduced PI3K activity might prove detrimental.

We thank L. Erhart and R. Raim for excellent technical assistance.

Disclosures The authors have no financial conflicts of interest.