Selective ablation of lung epithelial IKK2 impairs pulmonary Th17 responses and delays the clearance of Pneumocystis

Abstract

Pneumocystis is an atypical fungal pathogen which causes severe, often fatal pneumonia (PcP) in immunocompromised patients. Healthy humans and animals also encounter this pathogen, but generate a protective CD4⁺ T cell dependent immune response that clears the pathogen with little evidence of disease. Pneumocystis organisms attach tightly to respiratory epithelial cells, and in vitro studies have demonstrated that this interaction triggers NF-κB-dependent epithelial cell responses. However, the contribution of respiratory epithelial cells to the normal host response to Pneumocystis remains unknown. Inhibitor of κB Kinase 2 (IKK2) is the upstream kinase that is critical for inducible NF-κB activation. To determine whether IKK2-dependent lung epithelial cell responses contribute to the anti-Pneumocystis immune response in vivo, transgenic mice with lung epithelial cell-specific deletion of IKK2 (IKK2^ΔLEC) were generated. Compared to wild type mice, IKK2^ΔLEC mice exhibited a delayed onset of Th17 and B cell responses in the lung, and delayed fungal clearance. Importantly, delayed Pneumocystis clearance in IKK2^ΔLEC mice was associated with an exacerbated immune response, impaired pulmonary function, and altered lung histology. These data demonstrate that IKK2-dependent lung epithelial cell responses are important regulators of pulmonary adaptive immune responses, and are required for optimal host defense against Pneumocystis infection. LECs likely set the threshold for initiation of the pulmonary immune response, and serve to prevent exacerbated lung inflammation by promoting the rapid control of respiratory fungal infection.

Introduction

Pneumocystis carinii (Pc) is an opportunistic fungal pathogen with specific tropism for the mammalian lung. Pneumocystis organisms recovered from different mammalian hosts are genetically distinct, and attempts at cross-species transmission have not been successful [1–3]. Furthermore, the requirements for Pneumocystis growth in vitro have not been determined, making the study of life cycle and biology a significant challenge.

The environmental reservoir for human Pneumocystis is unknown, but Pneumocystis organisms have been found in lungs of healthy individuals [4]. In addition, most children become seropositive for anti-Pneumocystis antibodies at a young age [5, 6], making them a potential reservoir for infection [7]. Studies performed in experimental models of infection have found that Pneumocystis is capable of proliferating and establishing short-term infection in immunocompetent mice. While infected immunocompetent mice can transmit Pneumocystis infection to other mice, a cell-mediated adaptive immune response clears the pathogen rapidly with minimal health consequences [8]. These studies suggest that most people at some point in their lives become infected with Pneumocystis without presenting with any obvious or long-term clinical manifestations. The individual’s normal adaptive immune system resolves infection and confers protective immunity.

Although most people are exposed to Pneumocystis [4, 6, 9], it only causes the disease known as Pneumocystis pneumonia (PcP) in immunocompromised hosts. Typically the onset of PcP correlates with CD4⁺ T cell counts below 200 cells/μl [10], emphasizing the key role of this lymphocyte subset in lung defense against Pneumocystis infection. Populations at risk for PcP are AIDS patients, cancer patients undergoing chemotherapy, organ recipients, and persons with other primary or acquired immunodeficiency. Animal studies have clearly demonstrated that CD4⁺ T cells are critical for host defense against Pc infection [11–13]. However, the specific mechanisms through which an appropriate CD4⁺ T cell response is initiated, as well as the specific process by which the organisms are cleared remain only partially understood. A recent study determined that the ultimate effector mechanism for CD4⁺ T cell-dependent removal of Pneumocystis from the lung in vivo is macrophage phagocytosis [14].

One of the earliest events during Pneumocystis lung infection is the tight attachment of Pneumocystis to alveolar epithelial cells (AECs). This early interaction is necessary for Pc growth and for the establishment of pulmonary infection. In vitro studies have shown that the interaction of Pneumocystis with AECs activates the NF-κB signaling cascade, resulting in the production of chemokines and cytokines that may accelerate the development of adaptive immunity in immunocompetent hosts, and/or contribute to PcP-related immunopathogenesis in compromised hosts [15–18]. AECs have also been shown to produce chemokines in vivo during Pc infection, and pulmonary chemokine expression is associated with both protective immune responses and the development of PcP-related immunopathogenesis [18, 19]. However, the specific contributions of NF-κB-dependent AEC responses to either host defense against Pc infection, or the development on immunopathogenesis, remain unexplored.

In order to study the role of NF-κB-dependent AEC responses during Pc infection in vivo, the cre-lox system was used to generate tissue specific knock-out mice. Inhibitor of κB Kinase 2 (IKK2) is an important signaling kinase that is critical for inducible activation of the NF-κB pathway, and blockade of IKK2 activity effectively inhibits NF-κB activation [20]. Therefore, conditional ablation of IKK2 has been used study the role of inducible NF-κB activation in normal immune responses, as well as in inflammatory disease models. Transgenic mice in which the IKK2 gene was flanked by loxp recombination sites were crossed with mice expressing Cre recombinase under the control of the surfactant protein C (Sftpc) promoter to create mice which had specific and exclusive deletion of IKK2 in lung epithelial cells. These mice were used to determine how IKK2-dependent AEC responses contribute to host defense against Pc infection.

Material and Methods

Mice

Lung Epithelial Cell-specific IKK2 deficient mice (IKK2^ΔLEC) on the C57BL/6 background were generated by crossing mice with loxP-flanked Ikk2 alleles (IKK2^fl/fl; provided by Dr. Manolis Pasparakis) [21] with mice expressing Cre recombinase under the control of the Sftpc promoter (Sftpc-cre; generated by Dr. Brigid Hogan) [22, 23]. C57BL/6 IKK2^fl/fl litter mates and wild-type mice purchased from Jackson Laboratories were used as controls. Mice were housed using micro-isolator technology at the University of Rochester vivarium according to approved Institutional Animal Care and Use Committee protocols.

DNA isolation and genotyping

Genomic DNA was isolated from tail snips or isolated cell populations using the DNeasy Blood and Tissue Kit (Qiagen, Germantown, MD) or the Kapa Mouse Genotyping Kit (Kapa Biosystems, Boston, MA) following manufacturer’s instructions. Primers for the Cre gene (forward 5′-GAT CTT CGG CTA TAC GTA ACA GGG -3′; and reverse 5′-GAT CTC GAT GCA ACG AGT GAT GAG -3′), as well as primers for the IKK2 region of interest (IKK2 forward 5′-CAC CAT ACT AGC TGA ACT GC -3′; and IKK2 reverse 5′-AGG TAA GTG CTG AGA TGA CG -3′) were purchased from Integrated DNA Technologies (Coralville, IA). PCR reactions were performed using PuReTaq Ready-To-Go PCR Beads (GE Healthcare, Buckinghamshire, UK) and KAPA2G Fast Genotyping PCR mix.

Alveolar Macrophage isolation

Lungs from WT and IKK2^ΔLEC mice were lavaged with three, 1ml aliquots of Isolation Buffer (1X PBS, 1% glucose, 0.35mg/ml Gentamicin, 0.2mM EGTA). The lavage fluid was centrifuged for 10min at 250 ×g. The cell pellet was used for protein extraction or DNA isolation.

Primary murine lung epithelial cell isolation

Primary alveolar epithelial cells were isolated following published protocols [17]. Briefly, mice were euthanized with sodium pentobarbital. Using sterile technique, the peritoneal cavity was exposed to cut portal vein for exsanguination. The chest cavity was opened and the lungs were perfused with 10mls of heparinized 1X HBSS. A 20 gauge catheter was inserted into the trachea and the lungs were filled with 2mls of Dispase. Immediately, 0.45mls of 1% low-melting-point agarose was also introduced into the lungs, and a small bag of ice was placed in the chest to cool the lungs for 2 min. Lungs were removed into tubes containing 2mls of Dispase and were incubated for 45min at room temperature. Lungs were transferred to a 100mm dish with DMEM supplemented with 25nM HEPES and 0.01% DNAse I, and minced into small pieces. The suspension was swirled for 10 min at room temperature followed by filtration through 100, 40 and 25μm filters. Cells were pelleted and resuspended in DMEM supplemented with 10% FBS, and incubated with biotinylated anti-CD45 and anti-CD16/32 for 30 min at 37°C to bind hematopoietic cells. Cells were pelleted and resuspended in serum free media and incubated with streptavidin-coated magnetic beads for 30 min at room temperature. Sample tubes were placed against a magnet for 15 min to remove unbound cells in the supernatant. Cell suspensions were incubated at 37°C for 4–16 hrs in DMEM supplemented with 10% FBS and 10ng/ml of Keratinocyte Growth Factor (KGF, Calbiochem) to allow mesenchymal cells to adhere to tissue culture dish. Following incubation, the non-adherent epithelial cells are removed and cultured. Type II alveolar epithelial cell purity, as assessed by papanicolaou staining and Sftpc expression, was typically greater than 90%.

Protein Extraction and Western Blots

Protein from primary alveolar epithelial cells and from alveolar macrophages was extracted using mammalian protein extraction reagent (M-PER) following the manufacturer’s instructions (Thermo Scientific/Pierce, Rockford, IL). Total protein was run in NuPAGE 4–12% Bis-Tris polyacrylamide gels from Invitrogen (Life Technologies, Grand Island, NY). Upon transfer to nitrocellulose membrane, the protein of interest was detected using primary goat anti-IKK2 polyclonal antibody and secondary donkey anti-goat IgG horseradish peroxidase (HRP). Antibody to actin was used as a control. All antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA).

Pneumocystis isolation and enumeration

Pneumocystis organisms were propagated in CB.17 severe combined immunodeficient (SCID) mice. Heavily infected mice were euthanized and organisms were isolated from the lungs following our published protocol [15]. Briefly, lungs were perfused and aseptically removed into a glass tissue grinder in Pc Isolation Buffer (1X HBSS, 0.5% glutathione, 20mM HEPES buffer, 1% penicillin and streptomycin; pH 7.2). Following homogenization the preparation was passed sequentially through a series of needles decreasing in size from 18 to 22 to 26 gauge. The homogenate was centrifuged at 52 ×g to remove larger tissue pieces in the pellet. The supernatant was then removed and centrifuged at 2,000 ×g for 20 min to pellet Pc organisms. The pellet was resuspended in sterile water for 35 seconds to lyse red blood cells, followed by addition of 2X PBS. The preparation was incubated for 30 min at 37°C with 10μg/ml of DNAse containing media, and then centrifuged at 500 ×g for 20 min. The supernatant was passed through a 20-μm filter and then centrifuged at 2,000 ×g for 20 min. The pellet was resuspended in serum-free media and plated for 1.5 h at 37°C in tissue culture dishes coated with anti-Ly-6G/6C, anti-CD16/CD32 and anti-CD45 antibodies to remove remaining hematopoietic cells. Non-adherent Pneumocystis organisms were removed, centrifuged, and passed through a 26-gauge needle three times to disperse clumps. Cysts were enumerated by bright field examination of Gomori’s methenamine silver (GMS) stained slides. Typical preparations consist of a mixed population of 10% cysts and 90% trophic forms.

Pneumocystis infection

Experimental mice were anesthetized with isoflurane, and Pneumocystis organisms in a volume less than 0.1 mL were inoculated directly into the trachea. Mice received 5e5 or 1e6 Pc (based on cyst count) for long term and short term experiments, respectively.

Pulmonary Physiology

At each time point live, anesthetized mice were intubated, placed in a whole body plethysmograph (BUXCO Electronics, inc., Wilmington, NC), and connected to a rodent ventilator (Harvard Apparatus, Southnatic, MA). Dynamic lung compliance and lung resistance measurements were collected and analyzed using Biosystems XA software package.

Sample collection

Following pulmonary function measurement, the peritoneal and chest cavities were opened. Blood was obtained from a cardiac puncture using a 1mL syringe for isolation of sera. Mice were exsanguinated by cutting portal vein and lungs were perfused with 10 mLs of HBSS. The left lobes of the lung (or whole lungs in some cases) were lavaged with three, 1 mL aliquots of 1X HBSS. Lavaged lungs were quickly frozen in liquid nitrogen and stored at −80°C for later analysis. In some experiments the right lobe was inflation fixed with aqueous buffered zinc formalin Z-FIX (Anatech LTD, Battle Creek, MI), embedded in paraffin, and 4μm sections were cut for histological analyses. Recovered BAL fluid was centrifuged at 250 ×g, and the cell pellet was resuspended in 500μl of 1X HBSS. Total cells were counted using a hemocytometer. Supernatant was centrifuged at higher speed 12,000 ×g, transferred into fresh vials, and stored at −80°C for cytokine and chemokine analyses by ELISA.

Lung leukocyte isolation

Lungs were excised, placed in a petri dish, and quickly minced using scissors. One-half of each lung was flash-frozen, and stored at −80°C for later DNA analyses. The remaining tissue was dispersed through a 70 μm nylon filter using a syringe plunger. Cells were resuspended in RPMI 1640 media supplemented with 20% FBS, 2% L-glutamine, 1% penicillin and streptomycin, 1% Sodium Pyruvate, 20mM HEPES, and 2-mercaptoethanol. All reagents were purchased from GIBCO, Life Technologies Corporation. After centrifugation at 300 xg for 10 min, the cells were incubated in media containing collagenase and DNAse for 1 h at 37°C to dissociate extracellular matrix. Cells were passed through a 40 μm filter and subjected to red blood cell lysis. After washing with media, cells are ready for re-stimulation and/or staining.

Flow cytometry

Cellular fractions obtained from BAL and lung homogenates were incubated with Fc block (BD Biosciences, San Diego, CA) followed by incubation with fluorescently labeled antibodies against CD4 (RM4-5), CD8α (53–6.7), CD11c (HL3), CD19 (1D3), and GR-1 (RB6-8C5) (all from BD Biosciences). For intracellular cytokine staining cells were re-stimulated with 50–100 ng/ml PMA and 1 μM ionomycin for 4 h in the presence of GolgiPlug (BD Biosciences), a protein transport inhibitor containing Brefeldin A. Cells were washed and incubated with Live/Dead Fixable Aqua stain (Molecular Probes, Eugene, OR) for 30 min followed by surface staining. Cytofix/Cytoperm kit (BD Biosciences) reagents and directions were followed for subsequent fixation and permeabilization steps. Anti-cytokine antibodies for IL-17 (TC11-18H10.1, Biolegend) and IFNγ (XMG1.2, EBiosciences) were used for intracellular staining. Data was collected using a BD LSR II Flow Cytometer with BD FACSDiva Software (BD Biosciences) and analyses were performed using FlowJo (Tree Star, Ashland, OR).

Pneumocystis burden

Pneumocystis burden was determined as described [14]. Lung homogenates from experimental mice were subjected to three cycles of freezing and thawing followed by boiling for 20 min. Samples were then centrifuged at 13,000 rpm for 20 min to remove cell debris. The supernatant was removed and Pc burden was determined by quantitative real-time PCR (QPCR) for the single copy Pc kexin gene. QPCR was performed using TaqMan primer/fluorogenic probe chemistry detected with an Applied Biosystems Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA) as previously described [14].

Cytokine and chemokine measurement

Concentrations of IFN-γ, TNF-α, IL-1β, IL-4, IL-10, IL-17, MCP-1, MIP-2 and RANTES were determined by ELISA using kits purchased from R&D Systems (Minneapolis, MN) and used according to manufacturer’s instructions.

Detection of anti-Pneumocystis serum antibodies

Pneumocystis specific antibodies were measured in sera samples using an ELISA technique as previously described [8, 24]. Briefly, flat-bottom microtiter plates (ICM Biomedical, Aurora, OH) were coated overnight at room temperature with Pneumocystis antigen (in 0.05M Carbonate Buffer pH 9.6) obtained from infected SCID mouse lungs. After a 2 hour blocking step with 5% powdered milk in PBS, sera from experimental mice was diluted 1:50 in PBS + 0.1% Tween and incubated on plates for 2h at 37°C. After washing, secondary goat anti-mouse IgG and IgM conjugated to alkaline phosphate (Jackson ImmunoReaseach Laboratories, West Grove, PA) was incubated on plates for 1h at 37°C. After washing, the chromogenic substrate was added and the assay was developed. Optical density was read in a SpectraMax M5 microplate reader and analyzed with SoftPro Max 5.2 software (Molecular Devices, Sunnyvale, CA). Controls included hyperimmune sera of immunocompetent mice immunized with whole Pneumocystis. Samples were also tested against non-infected normal lung antigen preparations for data normalization.

Statistics

All statistical analyses and graphs were made using Sigma Plot version 10 (Systat Software Inc, San Jose, CA) and GraphPad Prism version 5.00 (GraphPad Software, San Diego, CA). Data is presented as mean ± 1 standard error measurement (SEM). T-tests were performed and statistical significance was determined for p values lower than 0.05 (*P< 0.05). For analysis of Pc burden in the lungs of experimental mice multiple t tests with Holm-Sidak correction were performed. P ≤ 0.05 was considered statistically significant.

Results

Generation of IKK2^ΔLEC transgenic mice

Mice with lung epithelial cell-specific loss of IKK2 were generated by crossing mice with loxP-flanked IKK2 alleles (IKK2^fl/fl) with mice expressing Cre recombinase under the control of the human surfactant protein C promoter (Sftpc-cre). The resulting offspring, designated IKK2^ΔLEC, had floxed IKK2 genes and carried at least one copy of the Cre recombinase gene (IKK2^{fl/fl-cre+/−}). The new strain was viable, fertile, and appeared phenotypically normal.

The floxed IKK2 gene has loxP sites flanking a 2 kilobase region containing exon 6 and exon 7. Deletion of this region results in an IKK2 null phenotype. Disruption of the IKK2 gene in lung epithelial cells was confirmed by PCR amplification using primers external to the loxP-flanked region (Figure 1A). Genomic DNA was isolated from primary alveolar epithelial cells (AECs) from IKK2^ΔLEC and IKK2^fl/fl mice. As expected, a large PCR product (~2.9kb) was amplified from the genomic DNA of IKK2^fl/fl AECs, which contained an intact IKK2 gene. In contrast, a much smaller PCR product (~500bp) was amplified from the genomic DNA of IKK2^ΔLEC AECs, indicating successful deletion of exons 6 and 7 of the IKK2 gene (Figure 1B). The IKK2 gene was not deleted from the genome of alveolar macrophages or spleen cells of IKK2^ΔLEC mice (Figure 1B and data not shown), demonstrating the specificity of cre-mediated deletion.

Figure 1. Generation of mice with selective deletion of IKK2 in LEC (IKK2^ΔLEC mice).

(A) Schematic IKK2 gene region and primer genotyping strategy for conditional deletion of IKK2 in LECs. (B) DNA and (C) protein were extracted from primary alveolar epithelial cells (AEC) and alveolar macrophages (AM) isolated from IKK2^fl/fl and IKK2^ΔLEC mice. Genomic deletion of IKK2 was confirmed by PCR and Western immunoblotting with anti-IKK2 antibody.

To confirm the PCR results, primary lung epithelial cells and AMs were isolated from IKK2^fl/fl and IKK2^ΔLEC mice and used for total protein isolation. Western immunoblotting demonstrated that IKK2 protein was present in IKK2^fl/fl AECs, while IKK2 protein was absent from the IKK2^ΔLEC AECs (Figure 1C). To demonstrate the specificity of this deletion, protein was also isolated from alveolar macrophages of both strains. Importantly, alveolar macrophages from both IKK2^fl/fl and IKK2^ΔLEC mice expressed IKK2 protein. These results demonstrate that IKK2^ΔLEC mice have lung epithelial cell-specific deficiency of IKK2, and are consistent with published work demonstrating that the sftpc promoter drives cre recombinase expression in lung epithelial precursor cells during embryogenesis [22, 23, 25].

Delayed Pneumocystis clearance in IKK2 ^ΔLEC mice

Immunocompetent wild-type mice exposed to Pneumocystis develop a mild infection, and are able to clear Pneumocystis organisms from the lung without obvious health effects [8]. To determine whether loss of IKK2 signaling in LECs affects the course of infection or the kinetics of pathogen clearance, wild-type and IKK2^ΔLEC mice were inoculated with freshly isolated Pneumocystis organisms. At different time points spanning the establishment, immune response, and clearance phases of infection, the Pneumocystis lung burden was measured. The Pneumocystis burden remained similar in both wild-type and IKK2^ΔLEC mice for the first 18 days post-infection, with little reduction in organism number (Figure 2). However, wild-type mice showed a one log reduction in total burden by day 21 and a two log reduction by day 25 (Figure 2A). In contrast, IKK2^ΔLEC mice showed little reduction in Pneumocystis burden at days 21 and 25, and harbored one to two logs more Pneumocystis than wild-type mice at these time points. IKK2^ΔLEC mice were able to eventually clear the pathogen by 32 days post-infection. The same trend was confirmed by Pneumocystis cyst counts of GMS stained lung homogenate slides (Figure 2B). These data demonstrate that IKK2-dependent LEC responses are required to promote the efficient clearance of Pneumocystis from the lungs of immunocompetent mice.

Figure 2. Delayed Pneumocystis clearance in IKK2^ΔLEC mice.

(A) DNA was isolated from lung homogenates of infected mice, collected at various times after infection. Quantitative real-time PCR was used to quantify Pneumocystis kexin gene copies to measure fungal burden in the lung. Results shown are the pool of 4 independent experiments (n= 4–14 at each time point) (** denotes P = 0.040 and ** denotes P = 0.008 compared to WT at same time point as determined by multiple t tests with Holm-Sidak correction) (B) Pneumocystis cysts were enumerated by bright field, microscopic examination of individual lung homogenates smears stained with Gomori’s methenamine silver (GMS). Data represents the mean ± 1 standard error measurement (SEM).

B cell recruitment to the lung is delayed in Pneumocystis-infected IKK2 ^ΔLEC mice

The clearance of Pneumocystis is dependent upon CD4⁺ T cells [26], B cells [27–29] and macrophage phagocytosis [14, 30]. To determine whether obvious immune response defects were associated with delayed Pneumocystis clearance in IKK2 ^ΔLEC mice, BAL cells recovered from infected wild-type and IKK2 ^ΔLEC mice were enumerated and analyzed by flow cytometry and differential staining. The only cellular deficiency noted in Pc-infected IKK2^ΔLEC mice was a delay in B lymphocyte recruitment to the lung relative to infected wild-type mice. B cells appeared in the lungs of infected wild-type mice at day 14, peaked at day 18, and decreased by day 21 (Figure 3A). In contrast, very low numbers of B cells were present in the lungs of infected IKK2^ΔLEC mice until day 18, peaked at day 21, and were decreased by day 25. Despite the differences in B cell numbers there was no significant effect on the Pneumocystis-specific antibody response in the sera of the two strains of mice (Figure 3B). Interestingly, IKK2^ΔLEC mice had a 5-fold greater number of BAL cells than wild-type mice at day 21 post-infection (Table I), which may be related to the persistence of Pneumocystis infection in the deficient mice (Figure 2). Differential staining analyses of immune cell phenotype revealed that the main contributors to the high numbers of BAL cells in the IKK2^ΔLEC mice were eosinophils, macrophages, and CD4⁺ T cells. There was no significant difference in numbers of CD8⁺ T cells or neutrophils between the groups at any time point.

Figure 3. B cell recruitment to the lung airways is delayed in IKK2^ΔLEC mice.

(A) CD19⁺ B cells present in the BAL of experimental mice were detected by flow cytometry. Data shows mean ± 1 standard error measurement (SEM) for each group (n=3–8 mice per time point). (*P< 0.05 compared to WT) (B) Pc specific antibody levels in the sera were detected by ELISA. Data shows average optical densities ± 1 SEM at each time point of a representative experiment. Controls include sera from uninfected mice and hyperimmune sera from immunocompetent mice immunized with whole Pneumocystis. Samples were also tested against non-infected normal lung antigen preparations for data normalization.

Table I. Cellular composition of BAL fluid of Pc infected mice.

Analyses of cells recovered in BAL fluid of experimental mice. Total cell numbers were counted in a hemocytometer. Cell suspensions were fluorescently labeled with antibodies specific for CD4, CD8, and CD11c and analyzed by flow cytometry. Cytospin slides were stained to count neutrophils (PMNs) and eosinophils, based in their morphologic features. Data shows mean ± 1 standard error measurement for each group (n=3–5 mice per time point). Data shown is from a representative experiment of at least 3 independent repeats.

Differential analyses of BAL cells recovered from Pc infected mice wild type and IKK2^ΔLEC mice.

Cell Type	Day 14		Day 21		Day 25
	WT	IKK2ΔLEC	WT	IKK2ΔLEC	WT	IKK2ΔLEC
Total Cells (105)	4.65±0.14	6.20±0.08	5.23±2.61	*23.3±1.75	2.45±0.60	4.25±1.03
Macrophages (105)	0.99±0.37	0.87±0.24	0.63±0.27	*3.11±0.75	1.07±0.41	0.94±0.52
B cells (105)	0.21±0.11	0.01±0.00	0.07±0.02	*0.31±0.01	0.03±0.01	0.12±0.07
CD4+ T cells (105)	0.61±0.19	0.42±0.04	0.21±0.08	*1.28±0.26	0.24±0.05	0.24±0.90
CD8+ T cells (105)	0.57±0.14	0.51±0.05	0.19±0.07	0.41±0.05	0.26±0.03	0.11±0.04
PMNs (105)	0.76±0.18	0.93±0.14	0.32±0.29	0.22±0.22	0.22±0.11	0.16±0.16
Eosinophils (105)	0.31±0.18	2.11±0.16	0.80±0.67	*14.37±1.70	0.10±0.10	1.34±0.89

^*P< 0.05 compared to WT

Lung cytokine profiles indicate altered T helper polarization in Pc-infected IKK2^ΔLEC mice

To further investigate the nature of the pulmonary immune response in IKK2^ΔLEC mice, cytokine levels were measured in lung samples taken from experimentally infected mice. Compared to infected wild-type mice, IFN-γ levels were lower in the infected IKK2^ΔLEC group at day 11 (Figure 4A). However, at later time points levels were comparable among the groups, with the exception of day 21 when IKK2^ΔLEC mice had higher lung IFN-γ levels. TNFα levels trended higher in the IKK2^ΔLEC group at days 14 through 21 (Figure 4B), which may be related to higher lung Pneumocystis burden in these mice. Levels of the Th2-associated cytokines IL-4 and IL-10 were also measured. Pneumocystis-infected IKK2^ΔLEC mice had two-fold higher IL-4 and IL-10 protein in the lungs than infected wild-type mice at days 14 through 25 (Figure 4C, D). These data demonstrate that mice with LEC-specific IKK2 deficiency have a higher concentration of cytokines which are associated with a Th2 response, and are consistent with the lung eosinophilia observed in these mice.

Figure 4. IKK2^ΔLEC mice have altered lung cytokine and chemokine production during Pc infection.

Protein levels of cytokines and chemokines were measured in the lung homogenates of experimental mice by ELISA: (A) IFN-γ, (B) TNF-α, (C) interleukin-4 (IL-4), (D) IL-10, (E) IL-1β, (F) IL-17, (G) MCP-1, (H) MIP-2 and (I) RANTES. Values shown are mean ± 1 SEM for a representative of 3 independent experiments (n≥3 mice per group per time point. (*P< 0.05 compared to WT).

IL-1β and IL-17 cytokine levels were also analyzed. IL-1β is important for promoting IL-17 production in CD4⁺ T cells [31, 32]. IL-17 is produced mainly by Th17 cells which are involved in protective immunity against fungal infections, including those caused by Pneumocystis [33]. Interestingly, IL-1β levels in the wild-type mice were elevated at day 7 post-infection and later went down to nearly basal levels. In contrast, IL-1β production was delayed in IKK2^ΔLEC mice, with peak levels observed at day 14 post-infection (Figure 4E). IL-17 levels were significantly lower in the lungs of IKK2^ΔLEC mice at days 14 and 21 post-infection compared to wild-type (Figure 4F). These data suggest that the loss of IKK2 in LECs alters the nature of the normal pulmonary immune response to Pneumocystis infection.

Chemokine levels were also measured in lung homogenate samples of experimental mice. CCL2 (MCP-1) attracts mainly T lymphocytes and monocytes, and is elevated in the lung during the immune response to Pneumocystis infection. Infected IKK2^ΔLEC mice had higher lung levels of CCL2 than wild-type mice at days 14 through 21 post-infection (Figure 4G). CXCL2 (MIP-2), which primarily attracts polymorphonuclear cells and is elevated during Pc infection, was also present in higher levels from day 14 to 25 post-infection in the IKK2^ΔLEC mice relative to wild-type (Figure 4H). This increase in chemokine secretion correlates with the higher number of cells present in the lungs of IKK2^ΔLEC mice. CCL5 (RANTES) is chemotactic for T cells, and is another chemokine which has been associated with the host response to Pneumocystis infection. Despite the fact that we saw higher number of T cells in Pneumocystis-infected IKK2^ΔLEC mice, RANTES lung levels were similar in both strains of mice (Figure 4I), indicating that CCL5 levels are not affected by the IKK2 deletion. Overall, these data demonstrate that IKK2-dependent LEC responses regulate the normal immune response to Pneumocystis infection.

IKK^ΔLEC mice Exhibit Delayed Pulmonary Th17 responses

To further investigate how loss of epithelial IKK2 signaling alters the normal immune response to Pneumocystis infection, the phenotype of T helper cells recruited to the lung was determined using intracellular cytokine staining. Leukocytes were isolated from the lavage fluid and lungs of Pneumocystis-infected wild-type and IKK2^ΔLEC mice. Following re-stimulation, lung leukocytes were stained for CD4, CD8, IFN-γ, IL-4, and IL-17. Interestingly, both the ratio and number of IL-17 producing CD4⁺ T cells was significantly lower in the IKK2^ΔLEC mice in both BAL and lung homogenate samples at all the time points assessed (Figure 5). Importantly, by day 25 post-infection the percentage and number of IL-17 positive cells increased in the lungs of infected IKK2^ΔLEC mice, suggesting that this phenomenon is one of delay rather than impairment of Th17 differentiation. To further confirm the ability of CD4⁺ T cells from IKK2^ΔLEC mice to adopt the Th17 phenotype, an in vitro culture system was used to induce differentiation of naïve CD4⁺ cells into Th17. Naïve CD4⁺ T cells isolated from the spleens of wild-type or IKK2^ΔLEC mice were equally able to adopt a Th17 phenotype following in vitro differentiation (data not shown). These data show that while the pulmonary Th17 response is delayed in Pneumocystis-infected IKK2^ΔLEC mice, there is no intrinsic deficiency in the CD4⁺ T cells of IKK2^ΔLEC mice that prevents Th17 differentiation.

Figure 5. Accumulation of pulmonary IL-17⁺ CD4⁺ T cells is compromised in IKK2^ΔLEC mice.

Cells were isolated from BAL (A–C) or freshly homogenized lungs (D–F) of Pneumocystis-infected mice and stimulated with PMA and ionomycin as described. Cells were incubated with fluorescently labeled anti-CD4 and anti-IL-17 monoclonal antibodies. (A and D) Representative dot plot panels showing the percentage of CD4⁺ cells that are also IL-17⁺. Each plot is pooled data from 3–4 mice per time point from a representative experiment. Bar graphs show mean ± 1 SEM for percentage of CD4 cells that are IL-17 positive (B and E) and absolute number of CD4⁺IL-17⁺ cells recovered (C and F) for three combined experiments (n=3–14 mice per time point). (*P< 0.05 compared to WT).

To detect the presence of Th1 and CD4⁺ T regulatory cells (Tregs), IFNγ producing cells and Foxp3⁺ CD4⁺ T cells were analyzed in BAL fluid and lung homogenates of Pneumocystis-infected wild-type and IKK2^ΔLEC mice. There was no difference in ratios or cell numbers of Th1 cells in the BAL or lung homogenate of infected wild-type or IKK2^ΔLEC mice (Figure 6 and data not shown). There was also no difference in the numbers of Foxp3⁺ or IL-4⁺ Th2 cells in infected mice of either strain (data not shown), despite the fact that IL-4 protein was increased in lung homogenates of infected IKK2^ΔLEC mice.

Figure 6. IKK2 deficiency does not affect accumulation of CD4⁺ IFNγ⁺ T cells BAL of Pneumocystis infected mice.

Cells were isolated from BAL (A–C) of infected mice and re-stimulated with PMA and ionomycin as described. Cells were fluorescently labeled with antibody against surface CD4 and for intracellular IFNγ. Cellular events were gated on live CD4⁺ cells. (A) Representative dot plot panels show the percentage of CD4⁺ cells that are also IFNγ ⁺. Each plot includes data pooled from 3–4 mice per time point for one representative experiment. Bar graphs show mean ± 1 SEM for (B) percentage of CD4 cells that are IFNγ positive and (C) absolute number of CD4⁺IFNγ⁺ cells recovered for three combined experiments (n=3–14 mice). (*P< 0.05 compared to WT).

Altered pulmonary physiology in Pneumocystis-infected IKK2^ΔLEC mice

To determine whether loss of IKK2 signaling in LECs affects pulmonary function following Pneumocystis-infection, dynamic lung compliance and lung resistance measurements were taken on infected wild-type and IKK2^ΔLEC mice using a whole body plethysmograph. As expected, wild-type mice cleared the Pneumocystis infection with little evidence of respiratory distress or impaired pulmonary function (Figure 7A, B). In contrast, a nearly 30% reduction in lung compliance as well as a significant increase in lung resistance was observed in IKK2^ΔLEC mice at days 18 and 21 post-infection. These observations of pulmonary dysfunction in IKK2^ΔLEC mice coincide temporally with the elevated lung cytokine levels and higher BAL cellularity relative to wild-type mice, suggesting that an exaggerated host response contributes to the impaired pulmonary function in IKK2^ΔLEC mice.

Figure 7. Epithelial IKK2 deficiency alters pulmonary physiology during Pneumocystis infection.

Mice were anesthetized and placed in a whole body plethysmograph to measure (A) dynamic lung compliance as well as (B) lung resistance. Results have been normalized to control uninfected mice at each time point. Data shows mean ± SEM for each group (n=3–8 mice per time point) for two pooled experiments (*P< 0.05 compared to WT). (C) Formalin-fixed, paraffin lung sections were stained with Hematoxylin and Eosin (H&E). Representative 100X bright field images from lungs collected at day 21 after infection are shown. Arrows highlight areas of inflammation.

To visualize the architecture of the lungs in Pneumocystis-infected wild-type and IKK2^ΔLEC mice, inflation fixed sections were stained with Hematoxylin and Eosin (H&E). Wild-type lungs were cleared of most cell infiltrates by day 21 (Figure 7C), coincident with the clearance of organisms from the lung. In contrast, cell infiltrates remained evident in the lungs of IKK2^ΔLEC mice at this time, and perivascular and septal thickening were also observed (Figure 7C). To determine whether the lungs of IKK2^ΔLEC mice show fibrotic changes, tissue sections were stained for alpha-smooth muscle actin (α-SMA). While α-SMA staining in infected wild-type mice varied little from uninfected control mice throughout the course of infection and clearance, the lungs of IKK2^ΔLEC mice displayed transiently elevated perivascular α-SMA staining compared to the wild type (Figure 8A–B). This was especially evident at days 14, 18, 21 post-infection. By day 25 the level of α-SMA staining in IKK2^ΔLEC mice had decreased significantly, and was similar to infected wild-type. These data suggest that loss of IKK2 signaling in LECs not only affects the immune response to Pneumocystis infection, but may also alter the normal resolution of infection.

Figure 8. Enhanced alpha-smooth muscle actin staining in Pneumocystis-infected IKK2^ΔLEC mice.

A) Formalin fixed, paraffin embedded lung sections were stained with antibodies against α-SMA (red stain) and counterstained with DAPI. Representative pictures, from lungs collected at different times after Pneumocystis infection, were taken with an Axioplan 2 Zeiss microscope. Similar observations were made in the rest of the lungs from WT and IKK2^ΔLEC groups (n=3–5). B) Area occupied by α-SMA⁺ cells around 5 randomly chosen medium size blood vessels (150–200 m diameter) was determined with the automated outline tool of the Axioplan Zeiss microscope. Slides were examined in a blinded fashion by the observer. Data from an experiment of three performed is shown. Paired t test was used to determine the statistically significant differences. *, p = 0.05, ***, p = 0.0005.

Discussion

A transgenic mouse line with lung epithelial cell-specific deletion of IKK2 was successfully generated and characterized. While several studies have demonstrated the ability of lung epithelial cells to respond to various infectious and non-infectious stimuli in vitro, the physiological relevance of the in vivo epithelial response has been more difficult to ascertain because of the many cell types, including alveolar macrophages, which are present in the lung. Likewise, the cell type-specific role of the NF-κB pathway has also been difficult to elucidate because NF-κB regulates a great number of cell processes in most cell types. Global knockout of either the NF-κB p65 subunit or IKK2 results in embryonic lethality caused by TNF-mediated hepatocyte apoptosis [20, 34]. Thus, the IKK2^ΔLEC mice are a valuable tool for not only studying the in vivo role of AECs during lung infection and disease, but also for gaining insight into the cell type-specific contributions of inducible NF-κB signaling. These mice will not only be useful for our studies of PcP, but will also be valuable for understanding the roles of lung epithelial cells and NF-κB in other models of acute and chronic lung disease.

The IKK2^ΔLEC mouse model was used to study the contribution of IKK2-dependent lung epithelial cell responses to the normal protective immune response against a respiratory fungal infection. We found that IKK2 is involved in the early host response to Pneumocystis infection and likely sets the threshold for immune activation and host defense which leads to organism clearance. IKK2-dependent LEC responses were not absolutely required for the clearance of this fungal pathogen from the lung, but mice with IKK2 deficient LECs exhibited delayed clearance kinetics. The most notable immune deficit associated with impaired Pneumocystis clearance in IKK2^ΔLEC mice was a delayed lung TH17 response. Th17 responses are important for host defense against fungal pathogens [35] including Pneumocystis [33, 36], and mice treated with anti-IL-17 antibody also exhibit delayed Pneumocystis clearance which was similar to the results found in this study [33, 36]. Therefore, it is likely that the impaired onset of pulmonary Th17 responses contributed to the delayed fungal clearance observed in IKK2^ΔLEC mice. Past studies have focused on the mechanisms by which lymphocyte and dendritic cell interactions promote a cytokine environment that is important for the Th17 arm of adaptive host defense. The present study demonstrates that lung epithelial cells are also key players in the initiation of the Th17 response. We speculate that their contributions could be through cytokine secretion and possibly direct interactions with dendritic cells or T cells. Epithelial cells produce chemokines that recruit IL-23-producing dendritic cells to promote the generation of Th17 responses [37]. IL-17 promotes maturation and class switching in B cells, and is also important for neutrophil recruitment and activation [38]. Further studies are required to determine the mechanism by which Th17 cells and associated cytokines contribute to Pneumocystis clearance.

Another notable defect associated with delayed Pneumocystis clearance in IKK2^ΔLEC mice was a delayed lung recruitment of B lymphocytes. B cells appeared in the lungs of Pneumocystis-infected wild-type mice by 2 weeks post-infection and declined in number as the Pneumocystis was cleared. In contrast, the appearance of B cells in the lung was delayed several days in the infected IKK2^ΔLEC mice. The overall antibody response was not greatly affected in the IKK2^ΔLEC mice, and was similar to responses observed in wild-type mice here and in prior reports [8, 39]. B cells as well as CD4⁺ T cells are critical for Pneumocystis clearance and host defense, and B cells may have antibody-independent functions that modulate CD4⁺ T cell activation or expansion during Pneumocystis infection [27]. Some of these functions include antigen presentation, cytokine secretion and regulation of T cell activity [40]. The significance of the delayed B cell response in Pneumocystis-infected IKK2^ΔLEC mice is not known, but these studies suggest that IKK2-dependent epithelial responses regulate Th17 and B cell recruitment to the lung during Pneumocystis infection, and that epithelial cells are important regulators of pulmonary adaptive immune responses.

In addition to the finding of delayed Pneumocystis clearance in IKK^ΔLEC mice, it was also notable that these mice mounted an exaggerated pulmonary immune response leading to pulmonary dysfunction. A nearly 30% reduction in lung compliance as well as a significant increase in lung resistance was observed in IKK2^ΔLEC mice at days 18 and 21 post-infection. These observations of pulmonary dysfunction in IKK2^ΔLEC mice coincided temporally with the elevated lung cytokine levels, and higher BAL cellularity relative to wild-type mice. The higher number of immune cells recruited to the lungs of IKK^ΔLEC mice clearly did not enhance fungal clearance, but instead their presence correlated with impaired lung function in these mice. Many of these cells were eosinophils, which have been found in higher numbers in lungs of AIDS patients [41], and may contribute to lung damage and the observed fibrotic changes. In a study by Swain et al., an increase in eosinophils was associated with enhanced lung injury in a mouse model of PcP [42]. The thickening of the perivascular regions in the lungs of infected IKK^ΔLEC mice resembles a phenomenon known as transitory pulmonary hypertension, which has been described in mouse models of PcP [43]. We speculate that Pneumocystis grew longer and to higher numbers in IKK^ΔLEC mice before the protective immune response was initiated, which consequently led to the exacerbated response. The immune impairment in IKK^ΔLEC mice is not permanent, likely because other resident immune cells, such as alveolar macrophages, produce strong cytokine and chemokine responses to activate the immune response in the absence if epithelial IKK2-dependent responses.

The Sftpc-Cre mice used in this study have been previously described and characterized [23, 25], and detailed studies have determined that Cre recombinase is expressed and active in the precursor cells that give rise to nearly all airway and alveolar epithelial cells. Thus the IKK2^ΔLEC mice we generated lack IKK2 in all lung epithelial cells. However, we suspect that the critical loss of IKK2 that creates the observed consequences on host defense is within the alveolar epithelial cell. Pneumocystis resides within the alveolar space and is most often found tightly attached to alveolar epithelial cells. The interaction of Pneumocystis with AECs is required for fungal growth and the progression of disease, and in vitro studies have demonstrated that the Pneumocystis:AEC interaction leads to NF-κB dependent AEC chemokine responses. Therefore, despite the fact that IKK2 is deleted from all lung epithelial cells, it is likely that the observed effects are directly related to IKK2 deficiency in AECs. We have shown that AECs of IKK2^ΔLEC mice are IKK2 deficient, while IKK2 expression is preserved in AMs. While Pneumocystis colonization of airway epithelial cells is not typically reported, it is possible that loss of IKK2 in this cell population also contributes to the impaired host defense.

Although Th1 and Th2 helper phenotypes have been both reported to play a role in host defense against Pneumocystis [44, 45], neither population is solely required because both IL-4^−/− and IFN-γ^−/− mice are able to clear this infection [46]. In the present study the IL-4 and IL-10 cytokine profiles found in lung homogenate samples suggest a Th2 biased response in Pneumocystis-infected IKK2^ΔLEC mice. Studies have shown that a Th2 lung environment induces alternative macrophage activation which results in increased phagocytosis of Pneumocystis [14, 47, 48]. To the contrary, we did not see faster clearance of Pneumocystis in the IKK2^ΔLEC mice. One explanation for this observation is the finding of elevated IL-10 levels in IKK2^ΔLEC mice. The presence of IL-10 delays Pneumocystis clearance in immunocompetent mice, and IL-10 deficient mice exhibit accelerated clearance [49]. Another possibility is that macrophage phenotype is altered in IKK2^ΔLEC mice. Distinct subpopulations of alternatively activated or M2 macrophages exist, and they have distinct functional capabilities [50]. Some macrophages, such as those associated with tumors and those found in burn patients, adopt an M2b phenotype that is more potently immunosuppressive than the M2a macrophages that are reported to promote anti-Pneumocystis host defense [48, 51]. Our finding of elevated lung IL-10 and MCP-1 levels in IKK2^ΔLEC mice suggest that that a pro-M2b environment does exist in these mice which display delayed Pneumocystis clearance. Further studies are required to determine the role of AECs in modulation of adaptive immunity and macrophage phenotype.

This study demonstrates that IKK2-dependent lung epithelial cell responses modulate the pulmonary immune response to respiratory fungal infection. While this work focused on the normal immune response, the conditional knockout mice will also be useful for studies of lung disease models in which inflammation and the immune response are major components of lung injury. The NF-κB signaling pathway is a popular therapeutic target for blocking inflammatory-mediated disease, and many inhibitors are currently under evaluation. Our findings suggest that NF-κB blockade could have both positive and negative effects on disease depending upon the nature of the disease and which cell types are primarily affected by the inhibitor. While prior studies have demonstrated a prominent role for NF-κB-dependent macrophage responses in the generation of lung inflammation, a recent study has reported that IKK2-dependent lung epithelial cell responses do not play a major role in the generation of airway inflammation following cigarette smoke exposure [52]. However, this study focused on innate pulmonary responses to the toxicants present in cigarette smoke, and did not assess the ability of IKK2-dependent epithelial cell responses to regulate adaptive immunity. Furthermore, these investigators were primarily concerned with airway disease, whereas PcP is mainly a disease of the alveolus. Therefore, it is conceivable that IKK2 dependent epithelial cell responses play an active role in the regulation of adaptive immune responses, but are dispensable for the generation of non-infectious airway inflammation. Further investigation into the cell-type specific role of NF-κB during lung disease is critical to the design of optimal therapeutic interventions.